diff options
Diffstat (limited to 'arch/powerpc/platforms/cell')
68 files changed, 7081 insertions, 2336 deletions
diff --git a/arch/powerpc/platforms/cell/Kconfig b/arch/powerpc/platforms/cell/Kconfig index 2f169991896..9978f594cac 100644 --- a/arch/powerpc/platforms/cell/Kconfig +++ b/arch/powerpc/platforms/cell/Kconfig @@ -2,29 +2,58 @@ config PPC_CELL bool default n -config PPC_CELL_NATIVE +config PPC_CELL_COMMON bool select PPC_CELL select PPC_DCR_MMIO - select PPC_OF_PLATFORM_PCI - select PPC_INDIRECT_IO + select PPC_INDIRECT_PIO + select PPC_INDIRECT_MMIO select PPC_NATIVE + select PPC_RTAS + select IRQ_EDGE_EOI_HANDLER + +config PPC_CELL_NATIVE + bool + select PPC_CELL_COMMON select MPIC - select IBM_NEW_EMAC_EMAC4 - select IBM_NEW_EMAC_RGMII - select IBM_NEW_EMAC_ZMII #test only - select IBM_NEW_EMAC_TAH #test only + select PPC_IO_WORKAROUNDS + select IBM_EMAC_EMAC4 + select IBM_EMAC_RGMII + select IBM_EMAC_ZMII #test only + select IBM_EMAC_TAH #test only default n config PPC_IBM_CELL_BLADE bool "IBM Cell Blade" - depends on PPC_MULTIPLATFORM && PPC64 + depends on PPC64 && PPC_BOOK3S select PPC_CELL_NATIVE - select PPC_RTAS + select PPC_OF_PLATFORM_PCI + select PCI select MMIO_NVRAM select PPC_UDBG_16550 select UDBG_RTAS_CONSOLE +config PPC_CELLEB + bool "Toshiba's Cell Reference Set 'Celleb' Architecture" + depends on PPC64 && PPC_BOOK3S + select PPC_CELL_NATIVE + select PPC_OF_PLATFORM_PCI + select PCI + select HAS_TXX9_SERIAL + select PPC_UDBG_BEAT + select USB_OHCI_BIG_ENDIAN_MMIO + select USB_EHCI_BIG_ENDIAN_MMIO + +config PPC_CELL_QPACE + bool "IBM Cell - QPACE" + depends on PPC64 && PPC_BOOK3S + select PPC_CELL_COMMON + +config AXON_MSI + bool + depends on PPC_IBM_CELL_BLADE && PCI_MSI + default y + menu "Cell Broadband Engine options" depends on PPC_CELL @@ -54,13 +83,6 @@ config SPU_FS_64K_LS uses 4K pages. This can improve performances of applications using multiple SPEs by lowering the TLB pressure on them. -config SPU_TRACE - tristate "SPU event tracing support" - depends on SPU_FS && MARKERS - help - This option allows reading a trace of spu-related events through - the sputrace file in procfs. - config SPU_BASE bool default n @@ -70,33 +92,48 @@ config CBE_RAS depends on PPC_CELL_NATIVE default y +config PPC_IBM_CELL_RESETBUTTON + bool "IBM Cell Blade Pinhole reset button" + depends on CBE_RAS && PPC_IBM_CELL_BLADE + default y + help + Support Pinhole Resetbutton on IBM Cell blades. + This adds a method to trigger system reset via front panel pinhole button. + +config PPC_IBM_CELL_POWERBUTTON + tristate "IBM Cell Blade power button" + depends on PPC_IBM_CELL_BLADE && INPUT_EVDEV + default y + help + Support Powerbutton on IBM Cell blades. + This will enable the powerbutton as an input device. + config CBE_THERM tristate "CBE thermal support" default m - depends on CBE_RAS + depends on CBE_RAS && SPU_BASE -config CBE_CPUFREQ - tristate "CBE frequency scaling" - depends on CBE_RAS && CPU_FREQ - default m +config PPC_PMI + tristate + default y + depends on CPU_FREQ_CBE_PMI || PPC_IBM_CELL_POWERBUTTON help - This adds the cpufreq driver for Cell BE processors. - For details, take a look at <file:Documentation/cpu-freq/>. - If you don't have such processor, say N + PMI (Platform Management Interrupt) is a way to + communicate with the BMC (Baseboard Management Controller). + It is used in some IBM Cell blades. -config CBE_CPUFREQ_PMI - tristate "CBE frequency scaling using PMI interface" - depends on CBE_CPUFREQ && PPC_PMI && EXPERIMENTAL - default n +config CBE_CPUFREQ_SPU_GOVERNOR + tristate "CBE frequency scaling based on SPU usage" + depends on SPU_FS && CPU_FREQ + default m help - Select this, if you want to use the PMI interface - to switch frequencies. Using PMI, the - processor will not only be able to run at lower speed, - but also at lower core voltage. + This governor checks for spu usage to adjust the cpu frequency. + If no spu is running on a given cpu, that cpu will be throttled to + the minimal possible frequency. endmenu config OPROFILE_CELL def_bool y - depends on PPC_CELL_NATIVE && (OPROFILE = m || OPROFILE = y) + depends on PPC_CELL_NATIVE && (OPROFILE = m || OPROFILE = y) && SPU_BASE diff --git a/arch/powerpc/platforms/cell/Makefile b/arch/powerpc/platforms/cell/Makefile index c89964c6fb1..fe053e7c73e 100644 --- a/arch/powerpc/platforms/cell/Makefile +++ b/arch/powerpc/platforms/cell/Makefile @@ -1,22 +1,22 @@ -obj-$(CONFIG_PPC_CELL_NATIVE) += interrupt.o iommu.o setup.o \ - cbe_regs.o spider-pic.o \ - pervasive.o pmu.o io-workarounds.o +obj-$(CONFIG_PPC_CELL_COMMON) += cbe_regs.o interrupt.o pervasive.o + +obj-$(CONFIG_PPC_CELL_NATIVE) += iommu.o setup.o spider-pic.o \ + pmu.o spider-pci.o obj-$(CONFIG_CBE_RAS) += ras.o obj-$(CONFIG_CBE_THERM) += cbe_thermal.o -obj-$(CONFIG_CBE_CPUFREQ_PMI) += cbe_cpufreq_pmi.o -obj-$(CONFIG_CBE_CPUFREQ) += cbe-cpufreq.o -cbe-cpufreq-y += cbe_cpufreq_pervasive.o cbe_cpufreq.o +obj-$(CONFIG_CBE_CPUFREQ_SPU_GOVERNOR) += cpufreq_spudemand.o + +obj-$(CONFIG_PPC_IBM_CELL_POWERBUTTON) += cbe_powerbutton.o ifeq ($(CONFIG_SMP),y) obj-$(CONFIG_PPC_CELL_NATIVE) += smp.o +obj-$(CONFIG_PPC_CELL_QPACE) += smp.o endif # needed only when building loadable spufs.ko -spu-priv1-$(CONFIG_PPC_CELL_NATIVE) += spu_priv1_mmio.o - -spu-manage-$(CONFIG_PPC_CELLEB) += spu_manage.o -spu-manage-$(CONFIG_PPC_CELL_NATIVE) += spu_manage.o +spu-priv1-$(CONFIG_PPC_CELL_COMMON) += spu_priv1_mmio.o +spu-manage-$(CONFIG_PPC_CELL_COMMON) += spu_manage.o obj-$(CONFIG_SPU_BASE) += spu_callbacks.o spu_base.o \ spu_notify.o \ @@ -25,4 +25,22 @@ obj-$(CONFIG_SPU_BASE) += spu_callbacks.o spu_base.o \ $(spu-manage-y) \ spufs/ -obj-$(CONFIG_PCI_MSI) += axon_msi.o +obj-$(CONFIG_AXON_MSI) += axon_msi.o + +# qpace setup +obj-$(CONFIG_PPC_CELL_QPACE) += qpace_setup.o + +# celleb stuff +ifeq ($(CONFIG_PPC_CELLEB),y) +obj-y += celleb_setup.o \ + celleb_pci.o celleb_scc_epci.o \ + celleb_scc_pciex.o \ + celleb_scc_uhc.o \ + spider-pci.o beat.o beat_htab.o \ + beat_hvCall.o beat_interrupt.o \ + beat_iommu.o + +obj-$(CONFIG_PPC_UDBG_BEAT) += beat_udbg.o +obj-$(CONFIG_SERIAL_TXX9) += celleb_scc_sio.o +obj-$(CONFIG_SPU_BASE) += beat_spu_priv1.o +endif diff --git a/arch/powerpc/platforms/cell/axon_msi.c b/arch/powerpc/platforms/cell/axon_msi.c index d95e71dee91..85825b5401e 100644 --- a/arch/powerpc/platforms/cell/axon_msi.c +++ b/arch/powerpc/platforms/cell/axon_msi.c @@ -13,7 +13,10 @@ #include <linux/kernel.h> #include <linux/pci.h> #include <linux/msi.h> +#include <linux/export.h> #include <linux/of_platform.h> +#include <linux/debugfs.h> +#include <linux/slab.h> #include <asm/dcr.h> #include <asm/machdep.h> @@ -64,55 +67,92 @@ struct axon_msic { - struct irq_host *irq_host; + struct irq_domain *irq_domain; __le32 *fifo_virt; dma_addr_t fifo_phys; dcr_host_t dcr_host; u32 read_offset; +#ifdef DEBUG + u32 __iomem *trigger; +#endif }; +#ifdef DEBUG +void axon_msi_debug_setup(struct device_node *dn, struct axon_msic *msic); +#else +static inline void axon_msi_debug_setup(struct device_node *dn, + struct axon_msic *msic) { } +#endif + + static void msic_dcr_write(struct axon_msic *msic, unsigned int dcr_n, u32 val) { - pr_debug("axon_msi: dcr_write(0x%x, 0x%x)\n", val, dcr_n); + pr_devel("axon_msi: dcr_write(0x%x, 0x%x)\n", val, dcr_n); dcr_write(msic->dcr_host, dcr_n, val); } static void axon_msi_cascade(unsigned int irq, struct irq_desc *desc) { - struct axon_msic *msic = get_irq_data(irq); + struct irq_chip *chip = irq_desc_get_chip(desc); + struct axon_msic *msic = irq_get_handler_data(irq); u32 write_offset, msi; int idx; + int retry = 0; write_offset = dcr_read(msic->dcr_host, MSIC_WRITE_OFFSET_REG); - pr_debug("axon_msi: original write_offset 0x%x\n", write_offset); + pr_devel("axon_msi: original write_offset 0x%x\n", write_offset); /* write_offset doesn't wrap properly, so we have to mask it */ write_offset &= MSIC_FIFO_SIZE_MASK; - while (msic->read_offset != write_offset) { + while (msic->read_offset != write_offset && retry < 100) { idx = msic->read_offset / sizeof(__le32); msi = le32_to_cpu(msic->fifo_virt[idx]); msi &= 0xFFFF; - pr_debug("axon_msi: woff %x roff %x msi %x\n", + pr_devel("axon_msi: woff %x roff %x msi %x\n", write_offset, msic->read_offset, msi); + if (msi < nr_irqs && irq_get_chip_data(msi) == msic) { + generic_handle_irq(msi); + msic->fifo_virt[idx] = cpu_to_le32(0xffffffff); + } else { + /* + * Reading the MSIC_WRITE_OFFSET_REG does not + * reliably flush the outstanding DMA to the + * FIFO buffer. Here we were reading stale + * data, so we need to retry. + */ + udelay(1); + retry++; + pr_devel("axon_msi: invalid irq 0x%x!\n", msi); + continue; + } + + if (retry) { + pr_devel("axon_msi: late irq 0x%x, retry %d\n", + msi, retry); + retry = 0; + } + msic->read_offset += MSIC_FIFO_ENTRY_SIZE; msic->read_offset &= MSIC_FIFO_SIZE_MASK; + } - if (msi < NR_IRQS && irq_map[msi].host == msic->irq_host) - generic_handle_irq(msi); - else - pr_debug("axon_msi: invalid irq 0x%x!\n", msi); + if (retry) { + printk(KERN_WARNING "axon_msi: irq timed out\n"); + + msic->read_offset += MSIC_FIFO_ENTRY_SIZE; + msic->read_offset &= MSIC_FIFO_SIZE_MASK; } - desc->chip->eoi(irq); + chip->irq_eoi(&desc->irq_data); } static struct axon_msic *find_msi_translator(struct pci_dev *dev) { - struct irq_host *irq_host; + struct irq_domain *irq_domain; struct device_node *dn, *tmp; const phandle *ph; struct axon_msic *msic = NULL; @@ -123,7 +163,7 @@ static struct axon_msic *find_msi_translator(struct pci_dev *dev) return NULL; } - for (; dn; tmp = of_get_parent(dn), of_node_put(dn), dn = tmp) { + for (; dn; dn = of_get_next_parent(dn)) { ph = of_get_property(dn, "msi-translator", NULL); if (ph) break; @@ -144,14 +184,14 @@ static struct axon_msic *find_msi_translator(struct pci_dev *dev) goto out_error; } - irq_host = irq_find_host(dn); - if (!irq_host) { - dev_dbg(&dev->dev, "axon_msi: no irq_host found for node %s\n", + irq_domain = irq_find_host(dn); + if (!irq_domain) { + dev_dbg(&dev->dev, "axon_msi: no irq_domain found for node %s\n", dn->full_name); goto out_error; } - msic = irq_host->host_data; + msic = irq_domain->host_data; out_error: of_node_put(dn); @@ -169,7 +209,7 @@ static int axon_msi_check_device(struct pci_dev *dev, int nvec, int type) static int setup_msi_msg_address(struct pci_dev *dev, struct msi_msg *msg) { - struct device_node *dn, *tmp; + struct device_node *dn; struct msi_desc *entry; int len; const u32 *prop; @@ -182,7 +222,7 @@ static int setup_msi_msg_address(struct pci_dev *dev, struct msi_msg *msg) entry = list_first_entry(&dev->msi_list, struct msi_desc, list); - for (; dn; tmp = of_get_parent(dn), of_node_put(dn), dn = tmp) { + for (; dn; dn = of_get_next_parent(dn)) { if (entry->msi_attrib.is_64) { prop = of_get_property(dn, "msi-address-64", &len); if (prop) @@ -236,11 +276,8 @@ static int axon_msi_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) if (rc) return rc; - /* We rely on being able to stash a virq in a u16 */ - BUILD_BUG_ON(NR_IRQS > 65536); - list_for_each_entry(entry, &dev->msi_list, list) { - virq = irq_create_direct_mapping(msic->irq_host); + virq = irq_create_direct_mapping(msic->irq_domain); if (virq == NO_IRQ) { dev_warn(&dev->dev, "axon_msi: virq allocation failed!\n"); @@ -248,7 +285,7 @@ static int axon_msi_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) } dev_dbg(&dev->dev, "axon_msi: allocated virq 0x%x\n", virq); - set_irq_msi(virq, entry); + irq_set_msi_desc(virq, entry); msg.data = virq; write_msi_msg(virq, &msg); } @@ -266,53 +303,51 @@ static void axon_msi_teardown_msi_irqs(struct pci_dev *dev) if (entry->irq == NO_IRQ) continue; - set_irq_msi(entry->irq, NULL); + irq_set_msi_desc(entry->irq, NULL); irq_dispose_mapping(entry->irq); } } static struct irq_chip msic_irq_chip = { - .mask = mask_msi_irq, - .unmask = unmask_msi_irq, - .shutdown = unmask_msi_irq, - .typename = "AXON-MSI", + .irq_mask = mask_msi_irq, + .irq_unmask = unmask_msi_irq, + .irq_shutdown = mask_msi_irq, + .name = "AXON-MSI", }; -static int msic_host_map(struct irq_host *h, unsigned int virq, +static int msic_host_map(struct irq_domain *h, unsigned int virq, irq_hw_number_t hw) { - set_irq_chip_and_handler(virq, &msic_irq_chip, handle_simple_irq); + irq_set_chip_data(virq, h->host_data); + irq_set_chip_and_handler(virq, &msic_irq_chip, handle_simple_irq); return 0; } -static struct irq_host_ops msic_host_ops = { +static const struct irq_domain_ops msic_host_ops = { .map = msic_host_map, }; -static int axon_msi_shutdown(struct of_device *device) +static void axon_msi_shutdown(struct platform_device *device) { - struct axon_msic *msic = device->dev.platform_data; + struct axon_msic *msic = dev_get_drvdata(&device->dev); u32 tmp; - pr_debug("axon_msi: disabling %s\n", - msic->irq_host->of_node->full_name); + pr_devel("axon_msi: disabling %s\n", + msic->irq_domain->of_node->full_name); tmp = dcr_read(msic->dcr_host, MSIC_CTRL_REG); tmp &= ~MSIC_CTRL_ENABLE & ~MSIC_CTRL_IRQ_ENABLE; msic_dcr_write(msic, MSIC_CTRL_REG, tmp); - - return 0; } -static int axon_msi_probe(struct of_device *device, - const struct of_device_id *device_id) +static int axon_msi_probe(struct platform_device *device) { - struct device_node *dn = device->node; + struct device_node *dn = device->dev.of_node; struct axon_msic *msic; unsigned int virq; int dcr_base, dcr_len; - pr_debug("axon_msi: setting up dn %s\n", dn->full_name); + pr_devel("axon_msi: setting up dn %s\n", dn->full_name); msic = kzalloc(sizeof(struct axon_msic), GFP_KERNEL); if (!msic) { @@ -328,7 +363,7 @@ static int axon_msi_probe(struct of_device *device, printk(KERN_ERR "axon_msi: couldn't parse dcr properties on %s\n", dn->full_name); - goto out; + goto out_free_msic; } msic->dcr_host = dcr_map(dn, dcr_base, dcr_len); @@ -346,26 +381,25 @@ static int axon_msi_probe(struct of_device *device, goto out_free_msic; } - msic->irq_host = irq_alloc_host(of_node_get(dn), IRQ_HOST_MAP_NOMAP, - NR_IRQS, &msic_host_ops, 0); - if (!msic->irq_host) { - printk(KERN_ERR "axon_msi: couldn't allocate irq_host for %s\n", + virq = irq_of_parse_and_map(dn, 0); + if (virq == NO_IRQ) { + printk(KERN_ERR "axon_msi: irq parse and map failed for %s\n", dn->full_name); goto out_free_fifo; } + memset(msic->fifo_virt, 0xff, MSIC_FIFO_SIZE_BYTES); - msic->irq_host->host_data = msic; - - virq = irq_of_parse_and_map(dn, 0); - if (virq == NO_IRQ) { - printk(KERN_ERR "axon_msi: irq parse and map failed for %s\n", + /* We rely on being able to stash a virq in a u16, so limit irqs to < 65536 */ + msic->irq_domain = irq_domain_add_nomap(dn, 65536, &msic_host_ops, msic); + if (!msic->irq_domain) { + printk(KERN_ERR "axon_msi: couldn't allocate irq_domain for %s\n", dn->full_name); - goto out_free_host; + goto out_free_fifo; } - set_irq_data(virq, msic); - set_irq_chained_handler(virq, axon_msi_cascade); - pr_debug("axon_msi: irq 0x%x setup for axon_msi\n", virq); + irq_set_handler_data(virq, msic); + irq_set_chained_handler(virq, axon_msi_cascade); + pr_devel("axon_msi: irq 0x%x setup for axon_msi\n", virq); /* Enable the MSIC hardware */ msic_dcr_write(msic, MSIC_BASE_ADDR_HI_REG, msic->fifo_phys >> 32); @@ -375,18 +409,21 @@ static int axon_msi_probe(struct of_device *device, MSIC_CTRL_IRQ_ENABLE | MSIC_CTRL_ENABLE | MSIC_CTRL_FIFO_SIZE); - device->dev.platform_data = msic; + msic->read_offset = dcr_read(msic->dcr_host, MSIC_WRITE_OFFSET_REG) + & MSIC_FIFO_SIZE_MASK; + + dev_set_drvdata(&device->dev, msic); ppc_md.setup_msi_irqs = axon_msi_setup_msi_irqs; ppc_md.teardown_msi_irqs = axon_msi_teardown_msi_irqs; ppc_md.msi_check_device = axon_msi_check_device; + axon_msi_debug_setup(dn, msic); + printk(KERN_DEBUG "axon_msi: setup MSIC on %s\n", dn->full_name); return 0; -out_free_host: - kfree(msic->irq_host); out_free_fifo: dma_free_coherent(&device->dev, MSIC_FIFO_SIZE_BYTES, msic->fifo_virt, msic->fifo_phys); @@ -404,17 +441,62 @@ static const struct of_device_id axon_msi_device_id[] = { {} }; -static struct of_platform_driver axon_msi_driver = { - .match_table = axon_msi_device_id, +static struct platform_driver axon_msi_driver = { .probe = axon_msi_probe, .shutdown = axon_msi_shutdown, - .driver = { - .name = "axon-msi" + .driver = { + .name = "axon-msi", + .owner = THIS_MODULE, + .of_match_table = axon_msi_device_id, }, }; static int __init axon_msi_init(void) { - return of_register_platform_driver(&axon_msi_driver); + return platform_driver_register(&axon_msi_driver); } subsys_initcall(axon_msi_init); + + +#ifdef DEBUG +static int msic_set(void *data, u64 val) +{ + struct axon_msic *msic = data; + out_le32(msic->trigger, val); + return 0; +} + +static int msic_get(void *data, u64 *val) +{ + *val = 0; + return 0; +} + +DEFINE_SIMPLE_ATTRIBUTE(fops_msic, msic_get, msic_set, "%llu\n"); + +void axon_msi_debug_setup(struct device_node *dn, struct axon_msic *msic) +{ + char name[8]; + u64 addr; + + addr = of_translate_address(dn, of_get_property(dn, "reg", NULL)); + if (addr == OF_BAD_ADDR) { + pr_devel("axon_msi: couldn't translate reg property\n"); + return; + } + + msic->trigger = ioremap(addr, 0x4); + if (!msic->trigger) { + pr_devel("axon_msi: ioremap failed\n"); + return; + } + + snprintf(name, sizeof(name), "msic_%d", of_node_to_nid(dn)); + + if (!debugfs_create_file(name, 0600, powerpc_debugfs_root, + msic, &fops_msic)) { + pr_devel("axon_msi: debugfs_create_file failed!\n"); + return; + } +} +#endif /* DEBUG */ diff --git a/arch/powerpc/platforms/cell/beat.c b/arch/powerpc/platforms/cell/beat.c new file mode 100644 index 00000000000..affcf566d46 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat.c @@ -0,0 +1,264 @@ +/* + * Simple routines for Celleb/Beat + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <linux/export.h> +#include <linux/init.h> +#include <linux/err.h> +#include <linux/rtc.h> +#include <linux/interrupt.h> +#include <linux/irqreturn.h> +#include <linux/reboot.h> + +#include <asm/hvconsole.h> +#include <asm/time.h> +#include <asm/machdep.h> +#include <asm/firmware.h> + +#include "beat_wrapper.h" +#include "beat.h" +#include "beat_interrupt.h" + +static int beat_pm_poweroff_flag; + +void beat_restart(char *cmd) +{ + beat_shutdown_logical_partition(!beat_pm_poweroff_flag); +} + +void beat_power_off(void) +{ + beat_shutdown_logical_partition(0); +} + +u64 beat_halt_code = 0x1000000000000000UL; +EXPORT_SYMBOL(beat_halt_code); + +void beat_halt(void) +{ + beat_shutdown_logical_partition(beat_halt_code); +} + +int beat_set_rtc_time(struct rtc_time *rtc_time) +{ + u64 tim; + tim = mktime(rtc_time->tm_year+1900, + rtc_time->tm_mon+1, rtc_time->tm_mday, + rtc_time->tm_hour, rtc_time->tm_min, rtc_time->tm_sec); + if (beat_rtc_write(tim)) + return -1; + return 0; +} + +void beat_get_rtc_time(struct rtc_time *rtc_time) +{ + u64 tim; + + if (beat_rtc_read(&tim)) + tim = 0; + to_tm(tim, rtc_time); + rtc_time->tm_year -= 1900; + rtc_time->tm_mon -= 1; +} + +#define BEAT_NVRAM_SIZE 4096 + +ssize_t beat_nvram_read(char *buf, size_t count, loff_t *index) +{ + unsigned int i; + unsigned long len; + char *p = buf; + + if (*index >= BEAT_NVRAM_SIZE) + return -ENODEV; + i = *index; + if (i + count > BEAT_NVRAM_SIZE) + count = BEAT_NVRAM_SIZE - i; + + for (; count != 0; count -= len) { + len = count; + if (len > BEAT_NVRW_CNT) + len = BEAT_NVRW_CNT; + if (beat_eeprom_read(i, len, p)) + return -EIO; + + p += len; + i += len; + } + *index = i; + return p - buf; +} + +ssize_t beat_nvram_write(char *buf, size_t count, loff_t *index) +{ + unsigned int i; + unsigned long len; + char *p = buf; + + if (*index >= BEAT_NVRAM_SIZE) + return -ENODEV; + i = *index; + if (i + count > BEAT_NVRAM_SIZE) + count = BEAT_NVRAM_SIZE - i; + + for (; count != 0; count -= len) { + len = count; + if (len > BEAT_NVRW_CNT) + len = BEAT_NVRW_CNT; + if (beat_eeprom_write(i, len, p)) + return -EIO; + + p += len; + i += len; + } + *index = i; + return p - buf; +} + +ssize_t beat_nvram_get_size(void) +{ + return BEAT_NVRAM_SIZE; +} + +int beat_set_xdabr(unsigned long dabr, unsigned long dabrx) +{ + if (beat_set_dabr(dabr, dabrx)) + return -1; + return 0; +} + +int64_t beat_get_term_char(u64 vterm, u64 *len, u64 *t1, u64 *t2) +{ + u64 db[2]; + s64 ret; + + ret = beat_get_characters_from_console(vterm, len, (u8 *)db); + if (ret == 0) { + *t1 = db[0]; + *t2 = db[1]; + } + return ret; +} +EXPORT_SYMBOL(beat_get_term_char); + +int64_t beat_put_term_char(u64 vterm, u64 len, u64 t1, u64 t2) +{ + u64 db[2]; + + db[0] = t1; + db[1] = t2; + return beat_put_characters_to_console(vterm, len, (u8 *)db); +} +EXPORT_SYMBOL(beat_put_term_char); + +void beat_power_save(void) +{ + beat_pause(0); +} + +#ifdef CONFIG_KEXEC +void beat_kexec_cpu_down(int crash, int secondary) +{ + beatic_deinit_IRQ(); +} +#endif + +static irqreturn_t beat_power_event(int virq, void *arg) +{ + printk(KERN_DEBUG "Beat: power button pressed\n"); + beat_pm_poweroff_flag = 1; + ctrl_alt_del(); + return IRQ_HANDLED; +} + +static irqreturn_t beat_reset_event(int virq, void *arg) +{ + printk(KERN_DEBUG "Beat: reset button pressed\n"); + beat_pm_poweroff_flag = 0; + ctrl_alt_del(); + return IRQ_HANDLED; +} + +static struct beat_event_list { + const char *typecode; + irq_handler_t handler; + unsigned int virq; +} beat_event_list[] = { + { "power", beat_power_event, 0 }, + { "reset", beat_reset_event, 0 }, +}; + +static int __init beat_register_event(void) +{ + u64 path[4], data[2]; + int rc, i; + unsigned int virq; + + for (i = 0; i < ARRAY_SIZE(beat_event_list); i++) { + struct beat_event_list *ev = &beat_event_list[i]; + + if (beat_construct_event_receive_port(data) != 0) { + printk(KERN_ERR "Beat: " + "cannot construct event receive port for %s\n", + ev->typecode); + return -EINVAL; + } + + virq = irq_create_mapping(NULL, data[0]); + if (virq == NO_IRQ) { + printk(KERN_ERR "Beat: failed to get virtual IRQ" + " for event receive port for %s\n", + ev->typecode); + beat_destruct_event_receive_port(data[0]); + return -EIO; + } + ev->virq = virq; + + rc = request_irq(virq, ev->handler, 0, + ev->typecode, NULL); + if (rc != 0) { + printk(KERN_ERR "Beat: failed to request virtual IRQ" + " for event receive port for %s\n", + ev->typecode); + beat_destruct_event_receive_port(data[0]); + return rc; + } + + path[0] = 0x1000000065780000ul; /* 1,ex */ + path[1] = 0x627574746f6e0000ul; /* button */ + path[2] = 0; + strncpy((char *)&path[2], ev->typecode, 8); + path[3] = 0; + data[1] = 0; + + beat_create_repository_node(path, data); + } + return 0; +} + +static int __init beat_event_init(void) +{ + if (!firmware_has_feature(FW_FEATURE_BEAT)) + return -EINVAL; + + beat_pm_poweroff_flag = 0; + return beat_register_event(); +} + +device_initcall(beat_event_init); diff --git a/arch/powerpc/platforms/cell/beat.h b/arch/powerpc/platforms/cell/beat.h new file mode 100644 index 00000000000..bfcb8e351ae --- /dev/null +++ b/arch/powerpc/platforms/cell/beat.h @@ -0,0 +1,39 @@ +/* + * Guest OS Interfaces. + * + * (C) Copyright 2006 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef _CELLEB_BEAT_H +#define _CELLEB_BEAT_H + +int64_t beat_get_term_char(uint64_t, uint64_t *, uint64_t *, uint64_t *); +int64_t beat_put_term_char(uint64_t, uint64_t, uint64_t, uint64_t); +int64_t beat_repository_encode(int, const char *, uint64_t[4]); +void beat_restart(char *); +void beat_power_off(void); +void beat_halt(void); +int beat_set_rtc_time(struct rtc_time *); +void beat_get_rtc_time(struct rtc_time *); +ssize_t beat_nvram_get_size(void); +ssize_t beat_nvram_read(char *, size_t, loff_t *); +ssize_t beat_nvram_write(char *, size_t, loff_t *); +int beat_set_xdabr(unsigned long, unsigned long); +void beat_power_save(void); +void beat_kexec_cpu_down(int, int); + +#endif /* _CELLEB_BEAT_H */ diff --git a/arch/powerpc/platforms/cell/beat_htab.c b/arch/powerpc/platforms/cell/beat_htab.c new file mode 100644 index 00000000000..d4d245c0d78 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_htab.c @@ -0,0 +1,445 @@ +/* + * "Cell Reference Set" HTAB support. + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This code is based on arch/powerpc/platforms/pseries/lpar.c: + * Copyright (C) 2001 Todd Inglett, IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#undef DEBUG_LOW + +#include <linux/kernel.h> +#include <linux/spinlock.h> + +#include <asm/mmu.h> +#include <asm/page.h> +#include <asm/pgtable.h> +#include <asm/machdep.h> +#include <asm/udbg.h> + +#include "beat_wrapper.h" + +#ifdef DEBUG_LOW +#define DBG_LOW(fmt...) do { udbg_printf(fmt); } while (0) +#else +#define DBG_LOW(fmt...) do { } while (0) +#endif + +static DEFINE_RAW_SPINLOCK(beat_htab_lock); + +static inline unsigned int beat_read_mask(unsigned hpte_group) +{ + unsigned long rmask = 0; + u64 hpte_v[5]; + + beat_read_htab_entries(0, hpte_group + 0, hpte_v); + if (!(hpte_v[0] & HPTE_V_BOLTED)) + rmask |= 0x8000; + if (!(hpte_v[1] & HPTE_V_BOLTED)) + rmask |= 0x4000; + if (!(hpte_v[2] & HPTE_V_BOLTED)) + rmask |= 0x2000; + if (!(hpte_v[3] & HPTE_V_BOLTED)) + rmask |= 0x1000; + beat_read_htab_entries(0, hpte_group + 4, hpte_v); + if (!(hpte_v[0] & HPTE_V_BOLTED)) + rmask |= 0x0800; + if (!(hpte_v[1] & HPTE_V_BOLTED)) + rmask |= 0x0400; + if (!(hpte_v[2] & HPTE_V_BOLTED)) + rmask |= 0x0200; + if (!(hpte_v[3] & HPTE_V_BOLTED)) + rmask |= 0x0100; + hpte_group = ~hpte_group & (htab_hash_mask * HPTES_PER_GROUP); + beat_read_htab_entries(0, hpte_group + 0, hpte_v); + if (!(hpte_v[0] & HPTE_V_BOLTED)) + rmask |= 0x80; + if (!(hpte_v[1] & HPTE_V_BOLTED)) + rmask |= 0x40; + if (!(hpte_v[2] & HPTE_V_BOLTED)) + rmask |= 0x20; + if (!(hpte_v[3] & HPTE_V_BOLTED)) + rmask |= 0x10; + beat_read_htab_entries(0, hpte_group + 4, hpte_v); + if (!(hpte_v[0] & HPTE_V_BOLTED)) + rmask |= 0x08; + if (!(hpte_v[1] & HPTE_V_BOLTED)) + rmask |= 0x04; + if (!(hpte_v[2] & HPTE_V_BOLTED)) + rmask |= 0x02; + if (!(hpte_v[3] & HPTE_V_BOLTED)) + rmask |= 0x01; + return rmask; +} + +static long beat_lpar_hpte_insert(unsigned long hpte_group, + unsigned long vpn, unsigned long pa, + unsigned long rflags, unsigned long vflags, + int psize, int apsize, int ssize) +{ + unsigned long lpar_rc; + u64 hpte_v, hpte_r, slot; + + if (vflags & HPTE_V_SECONDARY) + return -1; + + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW("hpte_insert(group=%lx, va=%016lx, pa=%016lx, " + "rflags=%lx, vflags=%lx, psize=%d)\n", + hpte_group, va, pa, rflags, vflags, psize); + + hpte_v = hpte_encode_v(vpn, psize, apsize, MMU_SEGSIZE_256M) | + vflags | HPTE_V_VALID; + hpte_r = hpte_encode_r(pa, psize, apsize) | rflags; + + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" hpte_v=%016lx, hpte_r=%016lx\n", hpte_v, hpte_r); + + if (rflags & _PAGE_NO_CACHE) + hpte_r &= ~HPTE_R_M; + + raw_spin_lock(&beat_htab_lock); + lpar_rc = beat_read_mask(hpte_group); + if (lpar_rc == 0) { + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" full\n"); + raw_spin_unlock(&beat_htab_lock); + return -1; + } + + lpar_rc = beat_insert_htab_entry(0, hpte_group, lpar_rc << 48, + hpte_v, hpte_r, &slot); + raw_spin_unlock(&beat_htab_lock); + + /* + * Since we try and ioremap PHBs we don't own, the pte insert + * will fail. However we must catch the failure in hash_page + * or we will loop forever, so return -2 in this case. + */ + if (unlikely(lpar_rc != 0)) { + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" lpar err %lx\n", lpar_rc); + return -2; + } + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" -> slot: %lx\n", slot); + + /* We have to pass down the secondary bucket bit here as well */ + return (slot ^ hpte_group) & 15; +} + +static long beat_lpar_hpte_remove(unsigned long hpte_group) +{ + DBG_LOW("hpte_remove(group=%lx)\n", hpte_group); + return -1; +} + +static unsigned long beat_lpar_hpte_getword0(unsigned long slot) +{ + unsigned long dword0; + unsigned long lpar_rc; + u64 dword[5]; + + lpar_rc = beat_read_htab_entries(0, slot & ~3UL, dword); + + dword0 = dword[slot&3]; + + BUG_ON(lpar_rc != 0); + + return dword0; +} + +static void beat_lpar_hptab_clear(void) +{ + unsigned long size_bytes = 1UL << ppc64_pft_size; + unsigned long hpte_count = size_bytes >> 4; + int i; + u64 dummy0, dummy1; + + /* TODO: Use bulk call */ + for (i = 0; i < hpte_count; i++) + beat_write_htab_entry(0, i, 0, 0, -1UL, -1UL, &dummy0, &dummy1); +} + +/* + * NOTE: for updatepp ops we are fortunate that the linux "newpp" bits and + * the low 3 bits of flags happen to line up. So no transform is needed. + * We can probably optimize here and assume the high bits of newpp are + * already zero. For now I am paranoid. + */ +static long beat_lpar_hpte_updatepp(unsigned long slot, + unsigned long newpp, + unsigned long vpn, + int psize, int apsize, + int ssize, int local) +{ + unsigned long lpar_rc; + u64 dummy0, dummy1; + unsigned long want_v; + + want_v = hpte_encode_avpn(vpn, psize, MMU_SEGSIZE_256M); + + DBG_LOW(" update: " + "avpnv=%016lx, slot=%016lx, psize: %d, newpp %016lx ... ", + want_v & HPTE_V_AVPN, slot, psize, newpp); + + raw_spin_lock(&beat_htab_lock); + dummy0 = beat_lpar_hpte_getword0(slot); + if ((dummy0 & ~0x7FUL) != (want_v & ~0x7FUL)) { + DBG_LOW("not found !\n"); + raw_spin_unlock(&beat_htab_lock); + return -1; + } + + lpar_rc = beat_write_htab_entry(0, slot, 0, newpp, 0, 7, &dummy0, + &dummy1); + raw_spin_unlock(&beat_htab_lock); + if (lpar_rc != 0 || dummy0 == 0) { + DBG_LOW("not found !\n"); + return -1; + } + + DBG_LOW("ok %lx %lx\n", dummy0, dummy1); + + BUG_ON(lpar_rc != 0); + + return 0; +} + +static long beat_lpar_hpte_find(unsigned long vpn, int psize) +{ + unsigned long hash; + unsigned long i, j; + long slot; + unsigned long want_v, hpte_v; + + hash = hpt_hash(vpn, mmu_psize_defs[psize].shift, MMU_SEGSIZE_256M); + want_v = hpte_encode_avpn(vpn, psize, MMU_SEGSIZE_256M); + + for (j = 0; j < 2; j++) { + slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; + for (i = 0; i < HPTES_PER_GROUP; i++) { + hpte_v = beat_lpar_hpte_getword0(slot); + + if (HPTE_V_COMPARE(hpte_v, want_v) + && (hpte_v & HPTE_V_VALID) + && (!!(hpte_v & HPTE_V_SECONDARY) == j)) { + /* HPTE matches */ + if (j) + slot = -slot; + return slot; + } + ++slot; + } + hash = ~hash; + } + + return -1; +} + +static void beat_lpar_hpte_updateboltedpp(unsigned long newpp, + unsigned long ea, + int psize, int ssize) +{ + unsigned long vpn; + unsigned long lpar_rc, slot, vsid; + u64 dummy0, dummy1; + + vsid = get_kernel_vsid(ea, MMU_SEGSIZE_256M); + vpn = hpt_vpn(ea, vsid, MMU_SEGSIZE_256M); + + raw_spin_lock(&beat_htab_lock); + slot = beat_lpar_hpte_find(vpn, psize); + BUG_ON(slot == -1); + + lpar_rc = beat_write_htab_entry(0, slot, 0, newpp, 0, 7, + &dummy0, &dummy1); + raw_spin_unlock(&beat_htab_lock); + + BUG_ON(lpar_rc != 0); +} + +static void beat_lpar_hpte_invalidate(unsigned long slot, unsigned long vpn, + int psize, int apsize, + int ssize, int local) +{ + unsigned long want_v; + unsigned long lpar_rc; + u64 dummy1, dummy2; + unsigned long flags; + + DBG_LOW(" inval : slot=%lx, va=%016lx, psize: %d, local: %d\n", + slot, va, psize, local); + want_v = hpte_encode_avpn(vpn, psize, MMU_SEGSIZE_256M); + + raw_spin_lock_irqsave(&beat_htab_lock, flags); + dummy1 = beat_lpar_hpte_getword0(slot); + + if ((dummy1 & ~0x7FUL) != (want_v & ~0x7FUL)) { + DBG_LOW("not found !\n"); + raw_spin_unlock_irqrestore(&beat_htab_lock, flags); + return; + } + + lpar_rc = beat_write_htab_entry(0, slot, 0, 0, HPTE_V_VALID, 0, + &dummy1, &dummy2); + raw_spin_unlock_irqrestore(&beat_htab_lock, flags); + + BUG_ON(lpar_rc != 0); +} + +void __init hpte_init_beat(void) +{ + ppc_md.hpte_invalidate = beat_lpar_hpte_invalidate; + ppc_md.hpte_updatepp = beat_lpar_hpte_updatepp; + ppc_md.hpte_updateboltedpp = beat_lpar_hpte_updateboltedpp; + ppc_md.hpte_insert = beat_lpar_hpte_insert; + ppc_md.hpte_remove = beat_lpar_hpte_remove; + ppc_md.hpte_clear_all = beat_lpar_hptab_clear; +} + +static long beat_lpar_hpte_insert_v3(unsigned long hpte_group, + unsigned long vpn, unsigned long pa, + unsigned long rflags, unsigned long vflags, + int psize, int apsize, int ssize) +{ + unsigned long lpar_rc; + u64 hpte_v, hpte_r, slot; + + if (vflags & HPTE_V_SECONDARY) + return -1; + + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW("hpte_insert(group=%lx, vpn=%016lx, pa=%016lx, " + "rflags=%lx, vflags=%lx, psize=%d)\n", + hpte_group, vpn, pa, rflags, vflags, psize); + + hpte_v = hpte_encode_v(vpn, psize, apsize, MMU_SEGSIZE_256M) | + vflags | HPTE_V_VALID; + hpte_r = hpte_encode_r(pa, psize, apsize) | rflags; + + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" hpte_v=%016lx, hpte_r=%016lx\n", hpte_v, hpte_r); + + if (rflags & _PAGE_NO_CACHE) + hpte_r &= ~HPTE_R_M; + + /* insert into not-volted entry */ + lpar_rc = beat_insert_htab_entry3(0, hpte_group, hpte_v, hpte_r, + HPTE_V_BOLTED, 0, &slot); + /* + * Since we try and ioremap PHBs we don't own, the pte insert + * will fail. However we must catch the failure in hash_page + * or we will loop forever, so return -2 in this case. + */ + if (unlikely(lpar_rc != 0)) { + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" lpar err %lx\n", lpar_rc); + return -2; + } + if (!(vflags & HPTE_V_BOLTED)) + DBG_LOW(" -> slot: %lx\n", slot); + + /* We have to pass down the secondary bucket bit here as well */ + return (slot ^ hpte_group) & 15; +} + +/* + * NOTE: for updatepp ops we are fortunate that the linux "newpp" bits and + * the low 3 bits of flags happen to line up. So no transform is needed. + * We can probably optimize here and assume the high bits of newpp are + * already zero. For now I am paranoid. + */ +static long beat_lpar_hpte_updatepp_v3(unsigned long slot, + unsigned long newpp, + unsigned long vpn, + int psize, int apsize, + int ssize, int local) +{ + unsigned long lpar_rc; + unsigned long want_v; + unsigned long pss; + + want_v = hpte_encode_avpn(vpn, psize, MMU_SEGSIZE_256M); + pss = (psize == MMU_PAGE_4K) ? -1UL : mmu_psize_defs[psize].penc[psize]; + + DBG_LOW(" update: " + "avpnv=%016lx, slot=%016lx, psize: %d, newpp %016lx ... ", + want_v & HPTE_V_AVPN, slot, psize, newpp); + + lpar_rc = beat_update_htab_permission3(0, slot, want_v, pss, 7, newpp); + + if (lpar_rc == 0xfffffff7) { + DBG_LOW("not found !\n"); + return -1; + } + + DBG_LOW("ok\n"); + + BUG_ON(lpar_rc != 0); + + return 0; +} + +static void beat_lpar_hpte_invalidate_v3(unsigned long slot, unsigned long vpn, + int psize, int apsize, + int ssize, int local) +{ + unsigned long want_v; + unsigned long lpar_rc; + unsigned long pss; + + DBG_LOW(" inval : slot=%lx, vpn=%016lx, psize: %d, local: %d\n", + slot, vpn, psize, local); + want_v = hpte_encode_avpn(vpn, psize, MMU_SEGSIZE_256M); + pss = (psize == MMU_PAGE_4K) ? -1UL : mmu_psize_defs[psize].penc[psize]; + + lpar_rc = beat_invalidate_htab_entry3(0, slot, want_v, pss); + + /* E_busy can be valid output: page may be already replaced */ + BUG_ON(lpar_rc != 0 && lpar_rc != 0xfffffff7); +} + +static int64_t _beat_lpar_hptab_clear_v3(void) +{ + return beat_clear_htab3(0); +} + +static void beat_lpar_hptab_clear_v3(void) +{ + _beat_lpar_hptab_clear_v3(); +} + +void __init hpte_init_beat_v3(void) +{ + if (_beat_lpar_hptab_clear_v3() == 0) { + ppc_md.hpte_invalidate = beat_lpar_hpte_invalidate_v3; + ppc_md.hpte_updatepp = beat_lpar_hpte_updatepp_v3; + ppc_md.hpte_updateboltedpp = beat_lpar_hpte_updateboltedpp; + ppc_md.hpte_insert = beat_lpar_hpte_insert_v3; + ppc_md.hpte_remove = beat_lpar_hpte_remove; + ppc_md.hpte_clear_all = beat_lpar_hptab_clear_v3; + } else { + ppc_md.hpte_invalidate = beat_lpar_hpte_invalidate; + ppc_md.hpte_updatepp = beat_lpar_hpte_updatepp; + ppc_md.hpte_updateboltedpp = beat_lpar_hpte_updateboltedpp; + ppc_md.hpte_insert = beat_lpar_hpte_insert; + ppc_md.hpte_remove = beat_lpar_hpte_remove; + ppc_md.hpte_clear_all = beat_lpar_hptab_clear; + } +} diff --git a/arch/powerpc/platforms/cell/beat_hvCall.S b/arch/powerpc/platforms/cell/beat_hvCall.S new file mode 100644 index 00000000000..96c80190712 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_hvCall.S @@ -0,0 +1,285 @@ +/* + * Beat hypervisor call I/F + * + * (C) Copyright 2007 TOSHIBA CORPORATION + * + * This code is based on arch/powerpc/platforms/pseries/hvCall.S. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <asm/ppc_asm.h> + +/* Not implemented on Beat, now */ +#define HCALL_INST_PRECALL +#define HCALL_INST_POSTCALL + + .text + +#define HVSC .long 0x44000022 + +/* Note: takes only 7 input parameters at maximum */ +_GLOBAL(beat_hcall_norets) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + mr r11,r3 + mr r3,r4 + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +/* Note: takes 8 input parameters at maximum */ +_GLOBAL(beat_hcall_norets8) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + mr r11,r3 + mr r3,r4 + mr r4,r5 + mr r5,r6 + mr r6,r7 + mr r7,r8 + mr r8,r9 + ld r10,STK_PARAM(R10)(r1) + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +/* Note: takes only 6 input parameters, 1 output parameters at maximum */ +_GLOBAL(beat_hcall1) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + std r4,STK_PARAM(R4)(r1) /* save ret buffer */ + + mr r11,r3 + mr r3,r5 + mr r4,r6 + mr r5,r7 + mr r6,r8 + mr r7,r9 + mr r8,r10 + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +/* Note: takes only 6 input parameters, 2 output parameters at maximum */ +_GLOBAL(beat_hcall2) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + std r4,STK_PARAM(R4)(r1) /* save ret buffer */ + + mr r11,r3 + mr r3,r5 + mr r4,r6 + mr r5,r7 + mr r6,r8 + mr r7,r9 + mr r8,r10 + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +/* Note: takes only 6 input parameters, 3 output parameters at maximum */ +_GLOBAL(beat_hcall3) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + std r4,STK_PARAM(R4)(r1) /* save ret buffer */ + + mr r11,r3 + mr r3,r5 + mr r4,r6 + mr r5,r7 + mr r6,r8 + mr r7,r9 + mr r8,r10 + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +/* Note: takes only 6 input parameters, 4 output parameters at maximum */ +_GLOBAL(beat_hcall4) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + std r4,STK_PARAM(R4)(r1) /* save ret buffer */ + + mr r11,r3 + mr r3,r5 + mr r4,r6 + mr r5,r7 + mr r6,r8 + mr r7,r9 + mr r8,r10 + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +/* Note: takes only 6 input parameters, 5 output parameters at maximum */ +_GLOBAL(beat_hcall5) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + std r4,STK_PARAM(R4)(r1) /* save ret buffer */ + + mr r11,r3 + mr r3,r5 + mr r4,r6 + mr r5,r7 + mr r6,r8 + mr r7,r9 + mr r8,r10 + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + std r8, 32(r12) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ + +/* Note: takes only 6 input parameters, 6 output parameters at maximum */ +_GLOBAL(beat_hcall6) + HMT_MEDIUM + + mfcr r0 + stw r0,8(r1) + + HCALL_INST_PRECALL + + std r4,STK_PARAM(R4)(r1) /* save ret buffer */ + + mr r11,r3 + mr r3,r5 + mr r4,r6 + mr r5,r7 + mr r6,r8 + mr r7,r9 + mr r8,r10 + + HVSC /* invoke the hypervisor */ + + HCALL_INST_POSTCALL + + ld r12,STK_PARAM(R4)(r1) + std r4, 0(r12) + std r5, 8(r12) + std r6, 16(r12) + std r7, 24(r12) + std r8, 32(r12) + std r9, 40(r12) + + lwz r0,8(r1) + mtcrf 0xff,r0 + + blr /* return r3 = status */ diff --git a/arch/powerpc/platforms/cell/beat_interrupt.c b/arch/powerpc/platforms/cell/beat_interrupt.c new file mode 100644 index 00000000000..9e5dfbcc00a --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_interrupt.c @@ -0,0 +1,253 @@ +/* + * Celleb/Beat Interrupt controller + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <linux/init.h> +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/percpu.h> +#include <linux/types.h> + +#include <asm/machdep.h> + +#include "beat_interrupt.h" +#include "beat_wrapper.h" + +#define MAX_IRQS NR_IRQS +static DEFINE_RAW_SPINLOCK(beatic_irq_mask_lock); +static uint64_t beatic_irq_mask_enable[(MAX_IRQS+255)/64]; +static uint64_t beatic_irq_mask_ack[(MAX_IRQS+255)/64]; + +static struct irq_domain *beatic_host; + +/* + * In this implementation, "virq" == "IRQ plug number", + * "(irq_hw_number_t)hwirq" == "IRQ outlet number". + */ + +/* assumption: locked */ +static inline void beatic_update_irq_mask(unsigned int irq_plug) +{ + int off; + unsigned long masks[4]; + + off = (irq_plug / 256) * 4; + masks[0] = beatic_irq_mask_enable[off + 0] + & beatic_irq_mask_ack[off + 0]; + masks[1] = beatic_irq_mask_enable[off + 1] + & beatic_irq_mask_ack[off + 1]; + masks[2] = beatic_irq_mask_enable[off + 2] + & beatic_irq_mask_ack[off + 2]; + masks[3] = beatic_irq_mask_enable[off + 3] + & beatic_irq_mask_ack[off + 3]; + if (beat_set_interrupt_mask(irq_plug&~255UL, + masks[0], masks[1], masks[2], masks[3]) != 0) + panic("Failed to set mask IRQ!"); +} + +static void beatic_mask_irq(struct irq_data *d) +{ + unsigned long flags; + + raw_spin_lock_irqsave(&beatic_irq_mask_lock, flags); + beatic_irq_mask_enable[d->irq/64] &= ~(1UL << (63 - (d->irq%64))); + beatic_update_irq_mask(d->irq); + raw_spin_unlock_irqrestore(&beatic_irq_mask_lock, flags); +} + +static void beatic_unmask_irq(struct irq_data *d) +{ + unsigned long flags; + + raw_spin_lock_irqsave(&beatic_irq_mask_lock, flags); + beatic_irq_mask_enable[d->irq/64] |= 1UL << (63 - (d->irq%64)); + beatic_update_irq_mask(d->irq); + raw_spin_unlock_irqrestore(&beatic_irq_mask_lock, flags); +} + +static void beatic_ack_irq(struct irq_data *d) +{ + unsigned long flags; + + raw_spin_lock_irqsave(&beatic_irq_mask_lock, flags); + beatic_irq_mask_ack[d->irq/64] &= ~(1UL << (63 - (d->irq%64))); + beatic_update_irq_mask(d->irq); + raw_spin_unlock_irqrestore(&beatic_irq_mask_lock, flags); +} + +static void beatic_end_irq(struct irq_data *d) +{ + s64 err; + unsigned long flags; + + err = beat_downcount_of_interrupt(d->irq); + if (err != 0) { + if ((err & 0xFFFFFFFF) != 0xFFFFFFF5) /* -11: wrong state */ + panic("Failed to downcount IRQ! Error = %16llx", err); + + printk(KERN_ERR "IRQ over-downcounted, plug %d\n", d->irq); + } + raw_spin_lock_irqsave(&beatic_irq_mask_lock, flags); + beatic_irq_mask_ack[d->irq/64] |= 1UL << (63 - (d->irq%64)); + beatic_update_irq_mask(d->irq); + raw_spin_unlock_irqrestore(&beatic_irq_mask_lock, flags); +} + +static struct irq_chip beatic_pic = { + .name = "CELL-BEAT", + .irq_unmask = beatic_unmask_irq, + .irq_mask = beatic_mask_irq, + .irq_eoi = beatic_end_irq, +}; + +/* + * Dispose binding hardware IRQ number (hw) and Virtuql IRQ number (virq), + * update flags. + * + * Note that the number (virq) is already assigned at upper layer. + */ +static void beatic_pic_host_unmap(struct irq_domain *h, unsigned int virq) +{ + beat_destruct_irq_plug(virq); +} + +/* + * Create or update binding hardware IRQ number (hw) and Virtuql + * IRQ number (virq). This is called only once for a given mapping. + * + * Note that the number (virq) is already assigned at upper layer. + */ +static int beatic_pic_host_map(struct irq_domain *h, unsigned int virq, + irq_hw_number_t hw) +{ + int64_t err; + + err = beat_construct_and_connect_irq_plug(virq, hw); + if (err < 0) + return -EIO; + + irq_set_status_flags(virq, IRQ_LEVEL); + irq_set_chip_and_handler(virq, &beatic_pic, handle_fasteoi_irq); + return 0; +} + +/* + * Translate device-tree interrupt spec to irq_hw_number_t style (ulong), + * to pass away to irq_create_mapping(). + * + * Called from irq_create_of_mapping() only. + * Note: We have only 1 entry to translate. + */ +static int beatic_pic_host_xlate(struct irq_domain *h, struct device_node *ct, + const u32 *intspec, unsigned int intsize, + irq_hw_number_t *out_hwirq, + unsigned int *out_flags) +{ + const u64 *intspec2 = (const u64 *)intspec; + + *out_hwirq = *intspec2; + *out_flags |= IRQ_TYPE_LEVEL_LOW; + return 0; +} + +static int beatic_pic_host_match(struct irq_domain *h, struct device_node *np) +{ + /* Match all */ + return 1; +} + +static const struct irq_domain_ops beatic_pic_host_ops = { + .map = beatic_pic_host_map, + .unmap = beatic_pic_host_unmap, + .xlate = beatic_pic_host_xlate, + .match = beatic_pic_host_match, +}; + +/* + * Get an IRQ number + * Note: returns VIRQ + */ +static inline unsigned int beatic_get_irq_plug(void) +{ + int i; + uint64_t pending[4], ub; + + for (i = 0; i < MAX_IRQS; i += 256) { + beat_detect_pending_interrupts(i, pending); + __asm__ ("cntlzd %0,%1":"=r"(ub): + "r"(pending[0] & beatic_irq_mask_enable[i/64+0] + & beatic_irq_mask_ack[i/64+0])); + if (ub != 64) + return i + ub + 0; + __asm__ ("cntlzd %0,%1":"=r"(ub): + "r"(pending[1] & beatic_irq_mask_enable[i/64+1] + & beatic_irq_mask_ack[i/64+1])); + if (ub != 64) + return i + ub + 64; + __asm__ ("cntlzd %0,%1":"=r"(ub): + "r"(pending[2] & beatic_irq_mask_enable[i/64+2] + & beatic_irq_mask_ack[i/64+2])); + if (ub != 64) + return i + ub + 128; + __asm__ ("cntlzd %0,%1":"=r"(ub): + "r"(pending[3] & beatic_irq_mask_enable[i/64+3] + & beatic_irq_mask_ack[i/64+3])); + if (ub != 64) + return i + ub + 192; + } + + return NO_IRQ; +} +unsigned int beatic_get_irq(void) +{ + unsigned int ret; + + ret = beatic_get_irq_plug(); + if (ret != NO_IRQ) + beatic_ack_irq(irq_get_irq_data(ret)); + return ret; +} + +/* + */ +void __init beatic_init_IRQ(void) +{ + int i; + + memset(beatic_irq_mask_enable, 0, sizeof(beatic_irq_mask_enable)); + memset(beatic_irq_mask_ack, 255, sizeof(beatic_irq_mask_ack)); + for (i = 0; i < MAX_IRQS; i += 256) + beat_set_interrupt_mask(i, 0L, 0L, 0L, 0L); + + /* Set out get_irq function */ + ppc_md.get_irq = beatic_get_irq; + + /* Allocate an irq host */ + beatic_host = irq_domain_add_nomap(NULL, ~0, &beatic_pic_host_ops, NULL); + BUG_ON(beatic_host == NULL); + irq_set_default_host(beatic_host); +} + +void beatic_deinit_IRQ(void) +{ + int i; + + for (i = 1; i < nr_irqs; i++) + beat_destruct_irq_plug(i); +} diff --git a/arch/powerpc/platforms/cell/beat_interrupt.h b/arch/powerpc/platforms/cell/beat_interrupt.h new file mode 100644 index 00000000000..a7e52f91a07 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_interrupt.h @@ -0,0 +1,30 @@ +/* + * Celleb/Beat Interrupt controller + * + * (C) Copyright 2006 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef ASM_BEAT_PIC_H +#define ASM_BEAT_PIC_H +#ifdef __KERNEL__ + +extern void beatic_init_IRQ(void); +extern unsigned int beatic_get_irq(void); +extern void beatic_deinit_IRQ(void); + +#endif +#endif /* ASM_BEAT_PIC_H */ diff --git a/arch/powerpc/platforms/cell/beat_iommu.c b/arch/powerpc/platforms/cell/beat_iommu.c new file mode 100644 index 00000000000..3ce68556893 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_iommu.c @@ -0,0 +1,115 @@ +/* + * Support for IOMMU on Celleb platform. + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/dma-mapping.h> +#include <linux/pci.h> +#include <linux/of_platform.h> + +#include <asm/machdep.h> + +#include "beat_wrapper.h" + +#define DMA_FLAGS 0xf800000000000000UL /* r/w permitted, coherency required, + strongest order */ + +static int __init find_dma_window(u64 *io_space_id, u64 *ioid, + u64 *base, u64 *size, u64 *io_page_size) +{ + struct device_node *dn; + const unsigned long *dma_window; + + for_each_node_by_type(dn, "ioif") { + dma_window = of_get_property(dn, "toshiba,dma-window", NULL); + if (dma_window) { + *io_space_id = (dma_window[0] >> 32) & 0xffffffffUL; + *ioid = dma_window[0] & 0x7ffUL; + *base = dma_window[1]; + *size = dma_window[2]; + *io_page_size = 1 << dma_window[3]; + of_node_put(dn); + return 1; + } + } + return 0; +} + +static unsigned long celleb_dma_direct_offset; + +static void __init celleb_init_direct_mapping(void) +{ + u64 lpar_addr, io_addr; + u64 io_space_id, ioid, dma_base, dma_size, io_page_size; + + if (!find_dma_window(&io_space_id, &ioid, &dma_base, &dma_size, + &io_page_size)) { + pr_info("No dma window found !\n"); + return; + } + + for (lpar_addr = 0; lpar_addr < dma_size; lpar_addr += io_page_size) { + io_addr = lpar_addr + dma_base; + (void)beat_put_iopte(io_space_id, io_addr, lpar_addr, + ioid, DMA_FLAGS); + } + + celleb_dma_direct_offset = dma_base; +} + +static void celleb_dma_dev_setup(struct device *dev) +{ + set_dma_ops(dev, &dma_direct_ops); + set_dma_offset(dev, celleb_dma_direct_offset); +} + +static void celleb_pci_dma_dev_setup(struct pci_dev *pdev) +{ + celleb_dma_dev_setup(&pdev->dev); +} + +static int celleb_of_bus_notify(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct device *dev = data; + + /* We are only intereted in device addition */ + if (action != BUS_NOTIFY_ADD_DEVICE) + return 0; + + celleb_dma_dev_setup(dev); + + return 0; +} + +static struct notifier_block celleb_of_bus_notifier = { + .notifier_call = celleb_of_bus_notify +}; + +static int __init celleb_init_iommu(void) +{ + celleb_init_direct_mapping(); + ppc_md.pci_dma_dev_setup = celleb_pci_dma_dev_setup; + bus_register_notifier(&platform_bus_type, &celleb_of_bus_notifier); + + return 0; +} + +machine_arch_initcall(celleb_beat, celleb_init_iommu); diff --git a/arch/powerpc/platforms/cell/beat_spu_priv1.c b/arch/powerpc/platforms/cell/beat_spu_priv1.c new file mode 100644 index 00000000000..13f52589d3a --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_spu_priv1.c @@ -0,0 +1,205 @@ +/* + * spu hypervisor abstraction for Beat + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <asm/types.h> +#include <asm/spu.h> +#include <asm/spu_priv1.h> + +#include "beat_wrapper.h" + +static inline void _int_mask_set(struct spu *spu, int class, u64 mask) +{ + spu->shadow_int_mask_RW[class] = mask; + beat_set_irq_mask_for_spe(spu->spe_id, class, mask); +} + +static inline u64 _int_mask_get(struct spu *spu, int class) +{ + return spu->shadow_int_mask_RW[class]; +} + +static void int_mask_set(struct spu *spu, int class, u64 mask) +{ + _int_mask_set(spu, class, mask); +} + +static u64 int_mask_get(struct spu *spu, int class) +{ + return _int_mask_get(spu, class); +} + +static void int_mask_and(struct spu *spu, int class, u64 mask) +{ + u64 old_mask; + old_mask = _int_mask_get(spu, class); + _int_mask_set(spu, class, old_mask & mask); +} + +static void int_mask_or(struct spu *spu, int class, u64 mask) +{ + u64 old_mask; + old_mask = _int_mask_get(spu, class); + _int_mask_set(spu, class, old_mask | mask); +} + +static void int_stat_clear(struct spu *spu, int class, u64 stat) +{ + beat_clear_interrupt_status_of_spe(spu->spe_id, class, stat); +} + +static u64 int_stat_get(struct spu *spu, int class) +{ + u64 int_stat; + beat_get_interrupt_status_of_spe(spu->spe_id, class, &int_stat); + return int_stat; +} + +static void cpu_affinity_set(struct spu *spu, int cpu) +{ + return; +} + +static u64 mfc_dar_get(struct spu *spu) +{ + u64 dar; + beat_get_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, mfc_dar_RW), &dar); + return dar; +} + +static u64 mfc_dsisr_get(struct spu *spu) +{ + u64 dsisr; + beat_get_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, mfc_dsisr_RW), &dsisr); + return dsisr; +} + +static void mfc_dsisr_set(struct spu *spu, u64 dsisr) +{ + beat_set_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, mfc_dsisr_RW), dsisr); +} + +static void mfc_sdr_setup(struct spu *spu) +{ + return; +} + +static void mfc_sr1_set(struct spu *spu, u64 sr1) +{ + beat_set_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, mfc_sr1_RW), sr1); +} + +static u64 mfc_sr1_get(struct spu *spu) +{ + u64 sr1; + beat_get_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, mfc_sr1_RW), &sr1); + return sr1; +} + +static void mfc_tclass_id_set(struct spu *spu, u64 tclass_id) +{ + beat_set_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, mfc_tclass_id_RW), tclass_id); +} + +static u64 mfc_tclass_id_get(struct spu *spu) +{ + u64 tclass_id; + beat_get_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, mfc_tclass_id_RW), &tclass_id); + return tclass_id; +} + +static void tlb_invalidate(struct spu *spu) +{ + beat_set_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, tlb_invalidate_entry_W), 0ul); +} + +static void resource_allocation_groupID_set(struct spu *spu, u64 id) +{ + beat_set_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, resource_allocation_groupID_RW), + id); +} + +static u64 resource_allocation_groupID_get(struct spu *spu) +{ + u64 id; + beat_get_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, resource_allocation_groupID_RW), + &id); + return id; +} + +static void resource_allocation_enable_set(struct spu *spu, u64 enable) +{ + beat_set_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, resource_allocation_enable_RW), + enable); +} + +static u64 resource_allocation_enable_get(struct spu *spu) +{ + u64 enable; + beat_get_spe_privileged_state_1_registers( + spu->spe_id, + offsetof(struct spu_priv1, resource_allocation_enable_RW), + &enable); + return enable; +} + +const struct spu_priv1_ops spu_priv1_beat_ops = { + .int_mask_and = int_mask_and, + .int_mask_or = int_mask_or, + .int_mask_set = int_mask_set, + .int_mask_get = int_mask_get, + .int_stat_clear = int_stat_clear, + .int_stat_get = int_stat_get, + .cpu_affinity_set = cpu_affinity_set, + .mfc_dar_get = mfc_dar_get, + .mfc_dsisr_get = mfc_dsisr_get, + .mfc_dsisr_set = mfc_dsisr_set, + .mfc_sdr_setup = mfc_sdr_setup, + .mfc_sr1_set = mfc_sr1_set, + .mfc_sr1_get = mfc_sr1_get, + .mfc_tclass_id_set = mfc_tclass_id_set, + .mfc_tclass_id_get = mfc_tclass_id_get, + .tlb_invalidate = tlb_invalidate, + .resource_allocation_groupID_set = resource_allocation_groupID_set, + .resource_allocation_groupID_get = resource_allocation_groupID_get, + .resource_allocation_enable_set = resource_allocation_enable_set, + .resource_allocation_enable_get = resource_allocation_enable_get, +}; diff --git a/arch/powerpc/platforms/cell/beat_syscall.h b/arch/powerpc/platforms/cell/beat_syscall.h new file mode 100644 index 00000000000..8580dc7e179 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_syscall.h @@ -0,0 +1,164 @@ +/* + * Beat hypervisor call numbers + * + * (C) Copyright 2004-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef BEAT_BEAT_syscall_H +#define BEAT_BEAT_syscall_H + +#ifdef __ASSEMBLY__ +#define __BEAT_ADD_VENDOR_ID(__x, __v) ((__v)<<60|(__x)) +#else +#define __BEAT_ADD_VENDOR_ID(__x, __v) ((u64)(__v)<<60|(__x)) +#endif +#define HV_allocate_memory __BEAT_ADD_VENDOR_ID(0, 0) +#define HV_construct_virtual_address_space __BEAT_ADD_VENDOR_ID(2, 0) +#define HV_destruct_virtual_address_space __BEAT_ADD_VENDOR_ID(10, 0) +#define HV_get_virtual_address_space_id_of_ppe __BEAT_ADD_VENDOR_ID(4, 0) +#define HV_query_logical_partition_address_region_info \ + __BEAT_ADD_VENDOR_ID(6, 0) +#define HV_release_memory __BEAT_ADD_VENDOR_ID(13, 0) +#define HV_select_virtual_address_space __BEAT_ADD_VENDOR_ID(7, 0) +#define HV_load_range_registers __BEAT_ADD_VENDOR_ID(68, 0) +#define HV_set_ppe_l2cache_rmt_entry __BEAT_ADD_VENDOR_ID(70, 0) +#define HV_set_ppe_tlb_rmt_entry __BEAT_ADD_VENDOR_ID(71, 0) +#define HV_set_spe_tlb_rmt_entry __BEAT_ADD_VENDOR_ID(72, 0) +#define HV_get_io_address_translation_fault_info __BEAT_ADD_VENDOR_ID(14, 0) +#define HV_get_iopte __BEAT_ADD_VENDOR_ID(16, 0) +#define HV_preload_iopt_cache __BEAT_ADD_VENDOR_ID(17, 0) +#define HV_put_iopte __BEAT_ADD_VENDOR_ID(15, 0) +#define HV_connect_event_ports __BEAT_ADD_VENDOR_ID(21, 0) +#define HV_construct_event_receive_port __BEAT_ADD_VENDOR_ID(18, 0) +#define HV_destruct_event_receive_port __BEAT_ADD_VENDOR_ID(19, 0) +#define HV_destruct_event_send_port __BEAT_ADD_VENDOR_ID(22, 0) +#define HV_get_state_of_event_send_port __BEAT_ADD_VENDOR_ID(25, 0) +#define HV_request_to_connect_event_ports __BEAT_ADD_VENDOR_ID(20, 0) +#define HV_send_event_externally __BEAT_ADD_VENDOR_ID(23, 0) +#define HV_send_event_locally __BEAT_ADD_VENDOR_ID(24, 0) +#define HV_construct_and_connect_irq_plug __BEAT_ADD_VENDOR_ID(28, 0) +#define HV_destruct_irq_plug __BEAT_ADD_VENDOR_ID(29, 0) +#define HV_detect_pending_interrupts __BEAT_ADD_VENDOR_ID(26, 0) +#define HV_end_of_interrupt __BEAT_ADD_VENDOR_ID(27, 0) +#define HV_assign_control_signal_notification_port __BEAT_ADD_VENDOR_ID(45, 0) +#define HV_end_of_control_signal_processing __BEAT_ADD_VENDOR_ID(48, 0) +#define HV_get_control_signal __BEAT_ADD_VENDOR_ID(46, 0) +#define HV_set_irq_mask_for_spe __BEAT_ADD_VENDOR_ID(61, 0) +#define HV_shutdown_logical_partition __BEAT_ADD_VENDOR_ID(44, 0) +#define HV_connect_message_ports __BEAT_ADD_VENDOR_ID(35, 0) +#define HV_destruct_message_port __BEAT_ADD_VENDOR_ID(36, 0) +#define HV_receive_message __BEAT_ADD_VENDOR_ID(37, 0) +#define HV_get_message_port_info __BEAT_ADD_VENDOR_ID(34, 0) +#define HV_request_to_connect_message_ports __BEAT_ADD_VENDOR_ID(33, 0) +#define HV_send_message __BEAT_ADD_VENDOR_ID(32, 0) +#define HV_get_logical_ppe_id __BEAT_ADD_VENDOR_ID(69, 0) +#define HV_pause __BEAT_ADD_VENDOR_ID(9, 0) +#define HV_destruct_shared_memory_handle __BEAT_ADD_VENDOR_ID(51, 0) +#define HV_get_shared_memory_info __BEAT_ADD_VENDOR_ID(52, 0) +#define HV_permit_sharing_memory __BEAT_ADD_VENDOR_ID(50, 0) +#define HV_request_to_attach_shared_memory __BEAT_ADD_VENDOR_ID(49, 0) +#define HV_enable_logical_spe_execution __BEAT_ADD_VENDOR_ID(55, 0) +#define HV_construct_logical_spe __BEAT_ADD_VENDOR_ID(53, 0) +#define HV_disable_logical_spe_execution __BEAT_ADD_VENDOR_ID(56, 0) +#define HV_destruct_logical_spe __BEAT_ADD_VENDOR_ID(54, 0) +#define HV_sense_spe_execution_status __BEAT_ADD_VENDOR_ID(58, 0) +#define HV_insert_htab_entry __BEAT_ADD_VENDOR_ID(101, 0) +#define HV_read_htab_entries __BEAT_ADD_VENDOR_ID(95, 0) +#define HV_write_htab_entry __BEAT_ADD_VENDOR_ID(94, 0) +#define HV_assign_io_address_translation_fault_port \ + __BEAT_ADD_VENDOR_ID(100, 0) +#define HV_set_interrupt_mask __BEAT_ADD_VENDOR_ID(73, 0) +#define HV_get_logical_partition_id __BEAT_ADD_VENDOR_ID(74, 0) +#define HV_create_repository_node2 __BEAT_ADD_VENDOR_ID(90, 0) +#define HV_create_repository_node __BEAT_ADD_VENDOR_ID(90, 0) /* alias */ +#define HV_get_repository_node_value2 __BEAT_ADD_VENDOR_ID(91, 0) +#define HV_get_repository_node_value __BEAT_ADD_VENDOR_ID(91, 0) /* alias */ +#define HV_modify_repository_node_value2 __BEAT_ADD_VENDOR_ID(92, 0) +#define HV_modify_repository_node_value __BEAT_ADD_VENDOR_ID(92, 0) /* alias */ +#define HV_remove_repository_node2 __BEAT_ADD_VENDOR_ID(93, 0) +#define HV_remove_repository_node __BEAT_ADD_VENDOR_ID(93, 0) /* alias */ +#define HV_cancel_shared_memory __BEAT_ADD_VENDOR_ID(104, 0) +#define HV_clear_interrupt_status_of_spe __BEAT_ADD_VENDOR_ID(206, 0) +#define HV_construct_spe_irq_outlet __BEAT_ADD_VENDOR_ID(80, 0) +#define HV_destruct_spe_irq_outlet __BEAT_ADD_VENDOR_ID(81, 0) +#define HV_disconnect_ipspc_service __BEAT_ADD_VENDOR_ID(88, 0) +#define HV_execute_ipspc_command __BEAT_ADD_VENDOR_ID(86, 0) +#define HV_get_interrupt_status_of_spe __BEAT_ADD_VENDOR_ID(205, 0) +#define HV_get_spe_privileged_state_1_registers __BEAT_ADD_VENDOR_ID(208, 0) +#define HV_permit_use_of_ipspc_service __BEAT_ADD_VENDOR_ID(85, 0) +#define HV_reinitialize_logical_spe __BEAT_ADD_VENDOR_ID(82, 0) +#define HV_request_ipspc_service __BEAT_ADD_VENDOR_ID(84, 0) +#define HV_stop_ipspc_command __BEAT_ADD_VENDOR_ID(87, 0) +#define HV_set_spe_privileged_state_1_registers __BEAT_ADD_VENDOR_ID(204, 0) +#define HV_get_status_of_ipspc_service __BEAT_ADD_VENDOR_ID(203, 0) +#define HV_put_characters_to_console __BEAT_ADD_VENDOR_ID(0x101, 1) +#define HV_get_characters_from_console __BEAT_ADD_VENDOR_ID(0x102, 1) +#define HV_get_base_clock __BEAT_ADD_VENDOR_ID(0x111, 1) +#define HV_set_base_clock __BEAT_ADD_VENDOR_ID(0x112, 1) +#define HV_get_frame_cycle __BEAT_ADD_VENDOR_ID(0x114, 1) +#define HV_disable_console __BEAT_ADD_VENDOR_ID(0x115, 1) +#define HV_disable_all_console __BEAT_ADD_VENDOR_ID(0x116, 1) +#define HV_oneshot_timer __BEAT_ADD_VENDOR_ID(0x117, 1) +#define HV_set_dabr __BEAT_ADD_VENDOR_ID(0x118, 1) +#define HV_get_dabr __BEAT_ADD_VENDOR_ID(0x119, 1) +#define HV_start_hv_stats __BEAT_ADD_VENDOR_ID(0x21c, 1) +#define HV_stop_hv_stats __BEAT_ADD_VENDOR_ID(0x21d, 1) +#define HV_get_hv_stats __BEAT_ADD_VENDOR_ID(0x21e, 1) +#define HV_get_hv_error_stats __BEAT_ADD_VENDOR_ID(0x221, 1) +#define HV_get_stats __BEAT_ADD_VENDOR_ID(0x224, 1) +#define HV_get_heap_stats __BEAT_ADD_VENDOR_ID(0x225, 1) +#define HV_get_memory_stats __BEAT_ADD_VENDOR_ID(0x227, 1) +#define HV_get_memory_detail __BEAT_ADD_VENDOR_ID(0x228, 1) +#define HV_set_priority_of_irq_outlet __BEAT_ADD_VENDOR_ID(0x122, 1) +#define HV_get_physical_spe_by_reservation_id __BEAT_ADD_VENDOR_ID(0x128, 1) +#define HV_get_spe_context __BEAT_ADD_VENDOR_ID(0x129, 1) +#define HV_set_spe_context __BEAT_ADD_VENDOR_ID(0x12a, 1) +#define HV_downcount_of_interrupt __BEAT_ADD_VENDOR_ID(0x12e, 1) +#define HV_peek_spe_context __BEAT_ADD_VENDOR_ID(0x12f, 1) +#define HV_read_bpa_register __BEAT_ADD_VENDOR_ID(0x131, 1) +#define HV_write_bpa_register __BEAT_ADD_VENDOR_ID(0x132, 1) +#define HV_map_context_table_of_spe __BEAT_ADD_VENDOR_ID(0x137, 1) +#define HV_get_slb_for_logical_spe __BEAT_ADD_VENDOR_ID(0x138, 1) +#define HV_set_slb_for_logical_spe __BEAT_ADD_VENDOR_ID(0x139, 1) +#define HV_init_pm __BEAT_ADD_VENDOR_ID(0x150, 1) +#define HV_set_pm_signal __BEAT_ADD_VENDOR_ID(0x151, 1) +#define HV_get_pm_signal __BEAT_ADD_VENDOR_ID(0x152, 1) +#define HV_set_pm_config __BEAT_ADD_VENDOR_ID(0x153, 1) +#define HV_get_pm_config __BEAT_ADD_VENDOR_ID(0x154, 1) +#define HV_get_inner_trace_data __BEAT_ADD_VENDOR_ID(0x155, 1) +#define HV_set_ext_trace_buffer __BEAT_ADD_VENDOR_ID(0x156, 1) +#define HV_get_ext_trace_buffer __BEAT_ADD_VENDOR_ID(0x157, 1) +#define HV_set_pm_interrupt __BEAT_ADD_VENDOR_ID(0x158, 1) +#define HV_get_pm_interrupt __BEAT_ADD_VENDOR_ID(0x159, 1) +#define HV_kick_pm __BEAT_ADD_VENDOR_ID(0x160, 1) +#define HV_construct_pm_context __BEAT_ADD_VENDOR_ID(0x164, 1) +#define HV_destruct_pm_context __BEAT_ADD_VENDOR_ID(0x165, 1) +#define HV_be_slow __BEAT_ADD_VENDOR_ID(0x170, 1) +#define HV_assign_ipspc_server_connection_status_notification_port \ + __BEAT_ADD_VENDOR_ID(0x173, 1) +#define HV_get_raid_of_physical_spe __BEAT_ADD_VENDOR_ID(0x174, 1) +#define HV_set_physical_spe_to_rag __BEAT_ADD_VENDOR_ID(0x175, 1) +#define HV_release_physical_spe_from_rag __BEAT_ADD_VENDOR_ID(0x176, 1) +#define HV_rtc_read __BEAT_ADD_VENDOR_ID(0x190, 1) +#define HV_rtc_write __BEAT_ADD_VENDOR_ID(0x191, 1) +#define HV_eeprom_read __BEAT_ADD_VENDOR_ID(0x192, 1) +#define HV_eeprom_write __BEAT_ADD_VENDOR_ID(0x193, 1) +#define HV_insert_htab_entry3 __BEAT_ADD_VENDOR_ID(0x104, 1) +#define HV_invalidate_htab_entry3 __BEAT_ADD_VENDOR_ID(0x105, 1) +#define HV_update_htab_permission3 __BEAT_ADD_VENDOR_ID(0x106, 1) +#define HV_clear_htab3 __BEAT_ADD_VENDOR_ID(0x107, 1) +#endif diff --git a/arch/powerpc/platforms/cell/beat_udbg.c b/arch/powerpc/platforms/cell/beat_udbg.c new file mode 100644 index 00000000000..350735bc888 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_udbg.c @@ -0,0 +1,98 @@ +/* + * udbg function for Beat + * + * (C) Copyright 2006 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <linux/kernel.h> +#include <linux/console.h> + +#include <asm/machdep.h> +#include <asm/prom.h> +#include <asm/udbg.h> + +#include "beat.h" + +#define celleb_vtermno 0 + +static void udbg_putc_beat(char c) +{ + unsigned long rc; + + if (c == '\n') + udbg_putc_beat('\r'); + + rc = beat_put_term_char(celleb_vtermno, 1, (uint64_t)c << 56, 0); +} + +/* Buffered chars getc */ +static u64 inbuflen; +static u64 inbuf[2]; /* must be 2 u64s */ + +static int udbg_getc_poll_beat(void) +{ + /* The interface is tricky because it may return up to 16 chars. + * We save them statically for future calls to udbg_getc(). + */ + char ch, *buf = (char *)inbuf; + int i; + long rc; + if (inbuflen == 0) { + /* get some more chars. */ + inbuflen = 0; + rc = beat_get_term_char(celleb_vtermno, &inbuflen, + inbuf+0, inbuf+1); + if (rc != 0) + inbuflen = 0; /* otherwise inbuflen is garbage */ + } + if (inbuflen <= 0 || inbuflen > 16) { + /* Catch error case as well as other oddities (corruption) */ + inbuflen = 0; + return -1; + } + ch = buf[0]; + for (i = 1; i < inbuflen; i++) /* shuffle them down. */ + buf[i-1] = buf[i]; + inbuflen--; + return ch; +} + +static int udbg_getc_beat(void) +{ + int ch; + for (;;) { + ch = udbg_getc_poll_beat(); + if (ch == -1) { + /* This shouldn't be needed...but... */ + volatile unsigned long delay; + for (delay = 0; delay < 2000000; delay++) + ; + } else { + return ch; + } + } +} + +/* call this from early_init() for a working debug console on + * vterm capable LPAR machines + */ +void __init udbg_init_debug_beat(void) +{ + udbg_putc = udbg_putc_beat; + udbg_getc = udbg_getc_beat; + udbg_getc_poll = udbg_getc_poll_beat; +} diff --git a/arch/powerpc/platforms/cell/beat_wrapper.h b/arch/powerpc/platforms/cell/beat_wrapper.h new file mode 100644 index 00000000000..c1109969f24 --- /dev/null +++ b/arch/powerpc/platforms/cell/beat_wrapper.h @@ -0,0 +1,290 @@ +/* + * Beat hypervisor call I/F + * + * (C) Copyright 2007 TOSHIBA CORPORATION + * + * This code is based on arch/powerpc/platforms/pseries/plpar_wrapper.h. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef BEAT_HCALL +#include <linux/string.h> +#include "beat_syscall.h" + +/* defined in hvCall.S */ +extern s64 beat_hcall_norets(u64 opcode, ...); +extern s64 beat_hcall_norets8(u64 opcode, u64 arg1, u64 arg2, u64 arg3, + u64 arg4, u64 arg5, u64 arg6, u64 arg7, u64 arg8); +extern s64 beat_hcall1(u64 opcode, u64 retbuf[1], ...); +extern s64 beat_hcall2(u64 opcode, u64 retbuf[2], ...); +extern s64 beat_hcall3(u64 opcode, u64 retbuf[3], ...); +extern s64 beat_hcall4(u64 opcode, u64 retbuf[4], ...); +extern s64 beat_hcall5(u64 opcode, u64 retbuf[5], ...); +extern s64 beat_hcall6(u64 opcode, u64 retbuf[6], ...); + +static inline s64 beat_downcount_of_interrupt(u64 plug_id) +{ + return beat_hcall_norets(HV_downcount_of_interrupt, plug_id); +} + +static inline s64 beat_set_interrupt_mask(u64 index, + u64 val0, u64 val1, u64 val2, u64 val3) +{ + return beat_hcall_norets(HV_set_interrupt_mask, index, + val0, val1, val2, val3); +} + +static inline s64 beat_destruct_irq_plug(u64 plug_id) +{ + return beat_hcall_norets(HV_destruct_irq_plug, plug_id); +} + +static inline s64 beat_construct_and_connect_irq_plug(u64 plug_id, + u64 outlet_id) +{ + return beat_hcall_norets(HV_construct_and_connect_irq_plug, plug_id, + outlet_id); +} + +static inline s64 beat_detect_pending_interrupts(u64 index, u64 *retbuf) +{ + return beat_hcall4(HV_detect_pending_interrupts, retbuf, index); +} + +static inline s64 beat_pause(u64 style) +{ + return beat_hcall_norets(HV_pause, style); +} + +static inline s64 beat_read_htab_entries(u64 htab_id, u64 index, u64 *retbuf) +{ + return beat_hcall5(HV_read_htab_entries, retbuf, htab_id, index); +} + +static inline s64 beat_insert_htab_entry(u64 htab_id, u64 group, + u64 bitmask, u64 hpte_v, u64 hpte_r, u64 *slot) +{ + u64 dummy[3]; + s64 ret; + + ret = beat_hcall3(HV_insert_htab_entry, dummy, htab_id, group, + bitmask, hpte_v, hpte_r); + *slot = dummy[0]; + return ret; +} + +static inline s64 beat_write_htab_entry(u64 htab_id, u64 slot, + u64 hpte_v, u64 hpte_r, u64 mask_v, u64 mask_r, + u64 *ret_v, u64 *ret_r) +{ + u64 dummy[2]; + s64 ret; + + ret = beat_hcall2(HV_write_htab_entry, dummy, htab_id, slot, + hpte_v, hpte_r, mask_v, mask_r); + *ret_v = dummy[0]; + *ret_r = dummy[1]; + return ret; +} + +static inline s64 beat_insert_htab_entry3(u64 htab_id, u64 group, + u64 hpte_v, u64 hpte_r, u64 mask_v, u64 value_v, u64 *slot) +{ + u64 dummy[1]; + s64 ret; + + ret = beat_hcall1(HV_insert_htab_entry3, dummy, htab_id, group, + hpte_v, hpte_r, mask_v, value_v); + *slot = dummy[0]; + return ret; +} + +static inline s64 beat_invalidate_htab_entry3(u64 htab_id, u64 group, + u64 va, u64 pss) +{ + return beat_hcall_norets(HV_invalidate_htab_entry3, + htab_id, group, va, pss); +} + +static inline s64 beat_update_htab_permission3(u64 htab_id, u64 group, + u64 va, u64 pss, u64 ptel_mask, u64 ptel_value) +{ + return beat_hcall_norets(HV_update_htab_permission3, + htab_id, group, va, pss, ptel_mask, ptel_value); +} + +static inline s64 beat_clear_htab3(u64 htab_id) +{ + return beat_hcall_norets(HV_clear_htab3, htab_id); +} + +static inline void beat_shutdown_logical_partition(u64 code) +{ + (void)beat_hcall_norets(HV_shutdown_logical_partition, code); +} + +static inline s64 beat_rtc_write(u64 time_from_epoch) +{ + return beat_hcall_norets(HV_rtc_write, time_from_epoch); +} + +static inline s64 beat_rtc_read(u64 *time_from_epoch) +{ + u64 dummy[1]; + s64 ret; + + ret = beat_hcall1(HV_rtc_read, dummy); + *time_from_epoch = dummy[0]; + return ret; +} + +#define BEAT_NVRW_CNT (sizeof(u64) * 6) + +static inline s64 beat_eeprom_write(u64 index, u64 length, u8 *buffer) +{ + u64 b[6]; + + if (length > BEAT_NVRW_CNT) + return -1; + memcpy(b, buffer, sizeof(b)); + return beat_hcall_norets8(HV_eeprom_write, index, length, + b[0], b[1], b[2], b[3], b[4], b[5]); +} + +static inline s64 beat_eeprom_read(u64 index, u64 length, u8 *buffer) +{ + u64 b[6]; + s64 ret; + + if (length > BEAT_NVRW_CNT) + return -1; + ret = beat_hcall6(HV_eeprom_read, b, index, length); + memcpy(buffer, b, length); + return ret; +} + +static inline s64 beat_set_dabr(u64 value, u64 style) +{ + return beat_hcall_norets(HV_set_dabr, value, style); +} + +static inline s64 beat_get_characters_from_console(u64 termno, u64 *len, + u8 *buffer) +{ + u64 dummy[3]; + s64 ret; + + ret = beat_hcall3(HV_get_characters_from_console, dummy, termno, len); + *len = dummy[0]; + memcpy(buffer, dummy + 1, *len); + return ret; +} + +static inline s64 beat_put_characters_to_console(u64 termno, u64 len, + u8 *buffer) +{ + u64 b[2]; + + memcpy(b, buffer, len); + return beat_hcall_norets(HV_put_characters_to_console, termno, len, + b[0], b[1]); +} + +static inline s64 beat_get_spe_privileged_state_1_registers( + u64 id, u64 offsetof, u64 *value) +{ + u64 dummy[1]; + s64 ret; + + ret = beat_hcall1(HV_get_spe_privileged_state_1_registers, dummy, id, + offsetof); + *value = dummy[0]; + return ret; +} + +static inline s64 beat_set_irq_mask_for_spe(u64 id, u64 class, u64 mask) +{ + return beat_hcall_norets(HV_set_irq_mask_for_spe, id, class, mask); +} + +static inline s64 beat_clear_interrupt_status_of_spe(u64 id, u64 class, + u64 mask) +{ + return beat_hcall_norets(HV_clear_interrupt_status_of_spe, + id, class, mask); +} + +static inline s64 beat_set_spe_privileged_state_1_registers( + u64 id, u64 offsetof, u64 value) +{ + return beat_hcall_norets(HV_set_spe_privileged_state_1_registers, + id, offsetof, value); +} + +static inline s64 beat_get_interrupt_status_of_spe(u64 id, u64 class, u64 *val) +{ + u64 dummy[1]; + s64 ret; + + ret = beat_hcall1(HV_get_interrupt_status_of_spe, dummy, id, class); + *val = dummy[0]; + return ret; +} + +static inline s64 beat_put_iopte(u64 ioas_id, u64 io_addr, u64 real_addr, + u64 ioid, u64 flags) +{ + return beat_hcall_norets(HV_put_iopte, ioas_id, io_addr, real_addr, + ioid, flags); +} + +static inline s64 beat_construct_event_receive_port(u64 *port) +{ + u64 dummy[1]; + s64 ret; + + ret = beat_hcall1(HV_construct_event_receive_port, dummy); + *port = dummy[0]; + return ret; +} + +static inline s64 beat_destruct_event_receive_port(u64 port) +{ + s64 ret; + + ret = beat_hcall_norets(HV_destruct_event_receive_port, port); + return ret; +} + +static inline s64 beat_create_repository_node(u64 path[4], u64 data[2]) +{ + s64 ret; + + ret = beat_hcall_norets(HV_create_repository_node2, + path[0], path[1], path[2], path[3], data[0], data[1]); + return ret; +} + +static inline s64 beat_get_repository_node_value(u64 lpid, u64 path[4], + u64 data[2]) +{ + s64 ret; + + ret = beat_hcall2(HV_get_repository_node_value2, data, + lpid, path[0], path[1], path[2], path[3]); + return ret; +} + +#endif diff --git a/arch/powerpc/platforms/cell/cbe_cpufreq.c b/arch/powerpc/platforms/cell/cbe_cpufreq.c deleted file mode 100644 index ec7c8f45a21..00000000000 --- a/arch/powerpc/platforms/cell/cbe_cpufreq.c +++ /dev/null @@ -1,208 +0,0 @@ -/* - * cpufreq driver for the cell processor - * - * (C) Copyright IBM Deutschland Entwicklung GmbH 2005-2007 - * - * Author: Christian Krafft <krafft@de.ibm.com> - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2, or (at your option) - * any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - */ - -#include <linux/cpufreq.h> -#include <linux/of_platform.h> - -#include <asm/machdep.h> -#include <asm/prom.h> -#include <asm/cell-regs.h> -#include "cbe_cpufreq.h" - -static DEFINE_MUTEX(cbe_switch_mutex); - - -/* the CBE supports an 8 step frequency scaling */ -static struct cpufreq_frequency_table cbe_freqs[] = { - {1, 0}, - {2, 0}, - {3, 0}, - {4, 0}, - {5, 0}, - {6, 0}, - {8, 0}, - {10, 0}, - {0, CPUFREQ_TABLE_END}, -}; - -/* - * hardware specific functions - */ - -static int set_pmode(unsigned int cpu, unsigned int slow_mode) -{ - int rc; - - if (cbe_cpufreq_has_pmi) - rc = cbe_cpufreq_set_pmode_pmi(cpu, slow_mode); - else - rc = cbe_cpufreq_set_pmode(cpu, slow_mode); - - pr_debug("register contains slow mode %d\n", cbe_cpufreq_get_pmode(cpu)); - - return rc; -} - -/* - * cpufreq functions - */ - -static int cbe_cpufreq_cpu_init(struct cpufreq_policy *policy) -{ - const u32 *max_freqp; - u32 max_freq; - int i, cur_pmode; - struct device_node *cpu; - - cpu = of_get_cpu_node(policy->cpu, NULL); - - if (!cpu) - return -ENODEV; - - pr_debug("init cpufreq on CPU %d\n", policy->cpu); - - /* - * Let's check we can actually get to the CELL regs - */ - if (!cbe_get_cpu_pmd_regs(policy->cpu) || - !cbe_get_cpu_mic_tm_regs(policy->cpu)) { - pr_info("invalid CBE regs pointers for cpufreq\n"); - return -EINVAL; - } - - max_freqp = of_get_property(cpu, "clock-frequency", NULL); - - of_node_put(cpu); - - if (!max_freqp) - return -EINVAL; - - /* we need the freq in kHz */ - max_freq = *max_freqp / 1000; - - pr_debug("max clock-frequency is at %u kHz\n", max_freq); - pr_debug("initializing frequency table\n"); - - /* initialize frequency table */ - for (i=0; cbe_freqs[i].frequency!=CPUFREQ_TABLE_END; i++) { - cbe_freqs[i].frequency = max_freq / cbe_freqs[i].index; - pr_debug("%d: %d\n", i, cbe_freqs[i].frequency); - } - - /* if DEBUG is enabled set_pmode() measures the latency - * of a transition */ - policy->cpuinfo.transition_latency = 25000; - - cur_pmode = cbe_cpufreq_get_pmode(policy->cpu); - pr_debug("current pmode is at %d\n",cur_pmode); - - policy->cur = cbe_freqs[cur_pmode].frequency; - -#ifdef CONFIG_SMP - policy->cpus = per_cpu(cpu_sibling_map, policy->cpu); -#endif - - cpufreq_frequency_table_get_attr(cbe_freqs, policy->cpu); - - /* this ensures that policy->cpuinfo_min - * and policy->cpuinfo_max are set correctly */ - return cpufreq_frequency_table_cpuinfo(policy, cbe_freqs); -} - -static int cbe_cpufreq_cpu_exit(struct cpufreq_policy *policy) -{ - cpufreq_frequency_table_put_attr(policy->cpu); - return 0; -} - -static int cbe_cpufreq_verify(struct cpufreq_policy *policy) -{ - return cpufreq_frequency_table_verify(policy, cbe_freqs); -} - -static int cbe_cpufreq_target(struct cpufreq_policy *policy, - unsigned int target_freq, - unsigned int relation) -{ - int rc; - struct cpufreq_freqs freqs; - unsigned int cbe_pmode_new; - - cpufreq_frequency_table_target(policy, - cbe_freqs, - target_freq, - relation, - &cbe_pmode_new); - - freqs.old = policy->cur; - freqs.new = cbe_freqs[cbe_pmode_new].frequency; - freqs.cpu = policy->cpu; - - mutex_lock(&cbe_switch_mutex); - cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); - - pr_debug("setting frequency for cpu %d to %d kHz, " \ - "1/%d of max frequency\n", - policy->cpu, - cbe_freqs[cbe_pmode_new].frequency, - cbe_freqs[cbe_pmode_new].index); - - rc = set_pmode(policy->cpu, cbe_pmode_new); - - cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); - mutex_unlock(&cbe_switch_mutex); - - return rc; -} - -static struct cpufreq_driver cbe_cpufreq_driver = { - .verify = cbe_cpufreq_verify, - .target = cbe_cpufreq_target, - .init = cbe_cpufreq_cpu_init, - .exit = cbe_cpufreq_cpu_exit, - .name = "cbe-cpufreq", - .owner = THIS_MODULE, - .flags = CPUFREQ_CONST_LOOPS, -}; - -/* - * module init and destoy - */ - -static int __init cbe_cpufreq_init(void) -{ - if (!machine_is(cell)) - return -ENODEV; - - return cpufreq_register_driver(&cbe_cpufreq_driver); -} - -static void __exit cbe_cpufreq_exit(void) -{ - cpufreq_unregister_driver(&cbe_cpufreq_driver); -} - -module_init(cbe_cpufreq_init); -module_exit(cbe_cpufreq_exit); - -MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Christian Krafft <krafft@de.ibm.com>"); diff --git a/arch/powerpc/platforms/cell/cbe_cpufreq.h b/arch/powerpc/platforms/cell/cbe_cpufreq.h deleted file mode 100644 index c1d86bfa92f..00000000000 --- a/arch/powerpc/platforms/cell/cbe_cpufreq.h +++ /dev/null @@ -1,24 +0,0 @@ -/* - * cbe_cpufreq.h - * - * This file contains the definitions used by the cbe_cpufreq driver. - * - * (C) Copyright IBM Deutschland Entwicklung GmbH 2005-2007 - * - * Author: Christian Krafft <krafft@de.ibm.com> - * - */ - -#include <linux/cpufreq.h> -#include <linux/types.h> - -int cbe_cpufreq_set_pmode(int cpu, unsigned int pmode); -int cbe_cpufreq_get_pmode(int cpu); - -int cbe_cpufreq_set_pmode_pmi(int cpu, unsigned int pmode); - -#if defined(CONFIG_CBE_CPUFREQ_PMI) || defined(CONFIG_CBE_CPUFREQ_PMI_MODULE) -extern bool cbe_cpufreq_has_pmi; -#else -#define cbe_cpufreq_has_pmi (0) -#endif diff --git a/arch/powerpc/platforms/cell/cbe_cpufreq_pervasive.c b/arch/powerpc/platforms/cell/cbe_cpufreq_pervasive.c deleted file mode 100644 index 70fa7aef5ed..00000000000 --- a/arch/powerpc/platforms/cell/cbe_cpufreq_pervasive.c +++ /dev/null @@ -1,115 +0,0 @@ -/* - * pervasive backend for the cbe_cpufreq driver - * - * This driver makes use of the pervasive unit to - * engage the desired frequency. - * - * (C) Copyright IBM Deutschland Entwicklung GmbH 2005-2007 - * - * Author: Christian Krafft <krafft@de.ibm.com> - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2, or (at your option) - * any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - */ - -#include <linux/io.h> -#include <linux/kernel.h> -#include <linux/time.h> -#include <asm/machdep.h> -#include <asm/hw_irq.h> -#include <asm/cell-regs.h> - -#include "cbe_cpufreq.h" - -/* to write to MIC register */ -static u64 MIC_Slow_Fast_Timer_table[] = { - [0 ... 7] = 0x007fc00000000000ull, -}; - -/* more values for the MIC */ -static u64 MIC_Slow_Next_Timer_table[] = { - 0x0000240000000000ull, - 0x0000268000000000ull, - 0x000029C000000000ull, - 0x00002D0000000000ull, - 0x0000300000000000ull, - 0x0000334000000000ull, - 0x000039C000000000ull, - 0x00003FC000000000ull, -}; - - -int cbe_cpufreq_set_pmode(int cpu, unsigned int pmode) -{ - struct cbe_pmd_regs __iomem *pmd_regs; - struct cbe_mic_tm_regs __iomem *mic_tm_regs; - u64 flags; - u64 value; -#ifdef DEBUG - long time; -#endif - - local_irq_save(flags); - - mic_tm_regs = cbe_get_cpu_mic_tm_regs(cpu); - pmd_regs = cbe_get_cpu_pmd_regs(cpu); - -#ifdef DEBUG - time = jiffies; -#endif - - out_be64(&mic_tm_regs->slow_fast_timer_0, MIC_Slow_Fast_Timer_table[pmode]); - out_be64(&mic_tm_regs->slow_fast_timer_1, MIC_Slow_Fast_Timer_table[pmode]); - - out_be64(&mic_tm_regs->slow_next_timer_0, MIC_Slow_Next_Timer_table[pmode]); - out_be64(&mic_tm_regs->slow_next_timer_1, MIC_Slow_Next_Timer_table[pmode]); - - value = in_be64(&pmd_regs->pmcr); - /* set bits to zero */ - value &= 0xFFFFFFFFFFFFFFF8ull; - /* set bits to next pmode */ - value |= pmode; - - out_be64(&pmd_regs->pmcr, value); - -#ifdef DEBUG - /* wait until new pmode appears in status register */ - value = in_be64(&pmd_regs->pmsr) & 0x07; - while (value != pmode) { - cpu_relax(); - value = in_be64(&pmd_regs->pmsr) & 0x07; - } - - time = jiffies - time; - time = jiffies_to_msecs(time); - pr_debug("had to wait %lu ms for a transition using " \ - "pervasive unit\n", time); -#endif - local_irq_restore(flags); - - return 0; -} - - -int cbe_cpufreq_get_pmode(int cpu) -{ - int ret; - struct cbe_pmd_regs __iomem *pmd_regs; - - pmd_regs = cbe_get_cpu_pmd_regs(cpu); - ret = in_be64(&pmd_regs->pmsr) & 0x07; - - return ret; -} - diff --git a/arch/powerpc/platforms/cell/cbe_cpufreq_pmi.c b/arch/powerpc/platforms/cell/cbe_cpufreq_pmi.c deleted file mode 100644 index 69288f65314..00000000000 --- a/arch/powerpc/platforms/cell/cbe_cpufreq_pmi.c +++ /dev/null @@ -1,149 +0,0 @@ -/* - * pmi backend for the cbe_cpufreq driver - * - * (C) Copyright IBM Deutschland Entwicklung GmbH 2005-2007 - * - * Author: Christian Krafft <krafft@de.ibm.com> - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2, or (at your option) - * any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - */ - -#include <linux/kernel.h> -#include <linux/types.h> -#include <linux/timer.h> -#include <linux/of_platform.h> - -#include <asm/processor.h> -#include <asm/prom.h> -#include <asm/pmi.h> -#include <asm/cell-regs.h> - -#ifdef DEBUG -#include <asm/time.h> -#endif - -#include "cbe_cpufreq.h" - -static u8 pmi_slow_mode_limit[MAX_CBE]; - -bool cbe_cpufreq_has_pmi = false; -EXPORT_SYMBOL_GPL(cbe_cpufreq_has_pmi); - -/* - * hardware specific functions - */ - -int cbe_cpufreq_set_pmode_pmi(int cpu, unsigned int pmode) -{ - int ret; - pmi_message_t pmi_msg; -#ifdef DEBUG - long time; -#endif - pmi_msg.type = PMI_TYPE_FREQ_CHANGE; - pmi_msg.data1 = cbe_cpu_to_node(cpu); - pmi_msg.data2 = pmode; - -#ifdef DEBUG - time = jiffies; -#endif - pmi_send_message(pmi_msg); - -#ifdef DEBUG - time = jiffies - time; - time = jiffies_to_msecs(time); - pr_debug("had to wait %lu ms for a transition using " \ - "PMI\n", time); -#endif - ret = pmi_msg.data2; - pr_debug("PMI returned slow mode %d\n", ret); - - return ret; -} -EXPORT_SYMBOL_GPL(cbe_cpufreq_set_pmode_pmi); - - -static void cbe_cpufreq_handle_pmi(pmi_message_t pmi_msg) -{ - u8 node, slow_mode; - - BUG_ON(pmi_msg.type != PMI_TYPE_FREQ_CHANGE); - - node = pmi_msg.data1; - slow_mode = pmi_msg.data2; - - pmi_slow_mode_limit[node] = slow_mode; - - pr_debug("cbe_handle_pmi: node: %d max_freq: %d\n", node, slow_mode); -} - -static int pmi_notifier(struct notifier_block *nb, - unsigned long event, void *data) -{ - struct cpufreq_policy *policy = data; - struct cpufreq_frequency_table *cbe_freqs; - u8 node; - - cbe_freqs = cpufreq_frequency_get_table(policy->cpu); - node = cbe_cpu_to_node(policy->cpu); - - pr_debug("got notified, event=%lu, node=%u\n", event, node); - - if (pmi_slow_mode_limit[node] != 0) { - pr_debug("limiting node %d to slow mode %d\n", - node, pmi_slow_mode_limit[node]); - - cpufreq_verify_within_limits(policy, 0, - - cbe_freqs[pmi_slow_mode_limit[node]].frequency); - } - - return 0; -} - -static struct notifier_block pmi_notifier_block = { - .notifier_call = pmi_notifier, -}; - -static struct pmi_handler cbe_pmi_handler = { - .type = PMI_TYPE_FREQ_CHANGE, - .handle_pmi_message = cbe_cpufreq_handle_pmi, -}; - - - -static int __init cbe_cpufreq_pmi_init(void) -{ - cbe_cpufreq_has_pmi = pmi_register_handler(&cbe_pmi_handler) == 0; - - if (!cbe_cpufreq_has_pmi) - return -ENODEV; - - cpufreq_register_notifier(&pmi_notifier_block, CPUFREQ_POLICY_NOTIFIER); - - return 0; -} - -static void __exit cbe_cpufreq_pmi_exit(void) -{ - cpufreq_unregister_notifier(&pmi_notifier_block, CPUFREQ_POLICY_NOTIFIER); - pmi_unregister_handler(&cbe_pmi_handler); -} - -module_init(cbe_cpufreq_pmi_init); -module_exit(cbe_cpufreq_pmi_exit); - -MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Christian Krafft <krafft@de.ibm.com>"); diff --git a/arch/powerpc/platforms/cell/cbe_powerbutton.c b/arch/powerpc/platforms/cell/cbe_powerbutton.c new file mode 100644 index 00000000000..2bb8031303f --- /dev/null +++ b/arch/powerpc/platforms/cell/cbe_powerbutton.c @@ -0,0 +1,118 @@ +/* + * driver for powerbutton on IBM cell blades + * + * (C) Copyright IBM Corp. 2005-2008 + * + * Author: Christian Krafft <krafft@de.ibm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include <linux/input.h> +#include <linux/module.h> +#include <linux/platform_device.h> +#include <asm/pmi.h> +#include <asm/prom.h> + +static struct input_dev *button_dev; +static struct platform_device *button_pdev; + +static void cbe_powerbutton_handle_pmi(pmi_message_t pmi_msg) +{ + BUG_ON(pmi_msg.type != PMI_TYPE_POWER_BUTTON); + + input_report_key(button_dev, KEY_POWER, 1); + input_sync(button_dev); + input_report_key(button_dev, KEY_POWER, 0); + input_sync(button_dev); +} + +static struct pmi_handler cbe_pmi_handler = { + .type = PMI_TYPE_POWER_BUTTON, + .handle_pmi_message = cbe_powerbutton_handle_pmi, +}; + +static int __init cbe_powerbutton_init(void) +{ + int ret = 0; + struct input_dev *dev; + + if (!of_machine_is_compatible("IBM,CBPLUS-1.0")) { + printk(KERN_ERR "%s: Not a cell blade.\n", __func__); + ret = -ENODEV; + goto out; + } + + dev = input_allocate_device(); + if (!dev) { + ret = -ENOMEM; + printk(KERN_ERR "%s: Not enough memory.\n", __func__); + goto out; + } + + set_bit(EV_KEY, dev->evbit); + set_bit(KEY_POWER, dev->keybit); + + dev->name = "Power Button"; + dev->id.bustype = BUS_HOST; + + /* this makes the button look like an acpi power button + * no clue whether anyone relies on that though */ + dev->id.product = 0x02; + dev->phys = "LNXPWRBN/button/input0"; + + button_pdev = platform_device_register_simple("power_button", 0, NULL, 0); + if (IS_ERR(button_pdev)) { + ret = PTR_ERR(button_pdev); + goto out_free_input; + } + + dev->dev.parent = &button_pdev->dev; + ret = input_register_device(dev); + if (ret) { + printk(KERN_ERR "%s: Failed to register device\n", __func__); + goto out_free_pdev; + } + + button_dev = dev; + + ret = pmi_register_handler(&cbe_pmi_handler); + if (ret) { + printk(KERN_ERR "%s: Failed to register with pmi.\n", __func__); + goto out_free_pdev; + } + + goto out; + +out_free_pdev: + platform_device_unregister(button_pdev); +out_free_input: + input_free_device(dev); +out: + return ret; +} + +static void __exit cbe_powerbutton_exit(void) +{ + pmi_unregister_handler(&cbe_pmi_handler); + platform_device_unregister(button_pdev); + input_free_device(button_dev); +} + +module_init(cbe_powerbutton_init); +module_exit(cbe_powerbutton_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Christian Krafft <krafft@de.ibm.com>"); diff --git a/arch/powerpc/platforms/cell/cbe_regs.c b/arch/powerpc/platforms/cell/cbe_regs.c index dbc338f187a..1428d583c23 100644 --- a/arch/powerpc/platforms/cell/cbe_regs.c +++ b/arch/powerpc/platforms/cell/cbe_regs.c @@ -8,7 +8,7 @@ #include <linux/percpu.h> #include <linux/types.h> -#include <linux/module.h> +#include <linux/export.h> #include <linux/of_device.h> #include <linux/of_platform.h> @@ -45,8 +45,8 @@ static struct cbe_thread_map unsigned int cbe_id; } cbe_thread_map[NR_CPUS]; -static cpumask_t cbe_local_mask[MAX_CBE] = { [0 ... MAX_CBE-1] = CPU_MASK_NONE }; -static cpumask_t cbe_first_online_cpu = CPU_MASK_NONE; +static cpumask_t cbe_local_mask[MAX_CBE] = { [0 ... MAX_CBE-1] = {CPU_BITS_NONE} }; +static cpumask_t cbe_first_online_cpu = { CPU_BITS_NONE }; static struct cbe_regs_map *cbe_find_map(struct device_node *np) { @@ -159,7 +159,8 @@ EXPORT_SYMBOL_GPL(cbe_cpu_to_node); u32 cbe_node_to_cpu(int node) { - return find_first_bit( (unsigned long *) &cbe_local_mask[node], sizeof(cpumask_t)); + return cpumask_first(&cbe_local_mask[node]); + } EXPORT_SYMBOL_GPL(cbe_node_to_cpu); @@ -268,9 +269,9 @@ void __init cbe_regs_init(void) thread->regs = map; thread->cbe_id = cbe_id; map->be_node = thread->be_node; - cpu_set(i, cbe_local_mask[cbe_id]); + cpumask_set_cpu(i, &cbe_local_mask[cbe_id]); if(thread->thread_id == 0) - cpu_set(i, cbe_first_online_cpu); + cpumask_set_cpu(i, &cbe_first_online_cpu); } } diff --git a/arch/powerpc/platforms/cell/cbe_thermal.c b/arch/powerpc/platforms/cell/cbe_thermal.c index 4852bf312d8..2c15ff09448 100644 --- a/arch/powerpc/platforms/cell/cbe_thermal.c +++ b/arch/powerpc/platforms/cell/cbe_thermal.c @@ -46,7 +46,7 @@ */ #include <linux/module.h> -#include <linux/sysdev.h> +#include <linux/device.h> #include <linux/kernel.h> #include <linux/cpu.h> #include <asm/spu.h> @@ -59,8 +59,8 @@ #define TEMP_MIN 65 #define TEMP_MAX 125 -#define SYSDEV_PREFIX_ATTR(_prefix,_name,_mode) \ -struct sysdev_attribute attr_ ## _prefix ## _ ## _name = { \ +#define DEVICE_PREFIX_ATTR(_prefix,_name,_mode) \ +struct device_attribute attr_ ## _prefix ## _ ## _name = { \ .attr = { .name = __stringify(_name), .mode = _mode }, \ .show = _prefix ## _show_ ## _name, \ .store = _prefix ## _store_ ## _name, \ @@ -76,35 +76,36 @@ static inline u8 temp_to_reg(u8 temp) return ((temp - TEMP_MIN) >> 1) & 0x3f; } -static struct cbe_pmd_regs __iomem *get_pmd_regs(struct sys_device *sysdev) +static struct cbe_pmd_regs __iomem *get_pmd_regs(struct device *dev) { struct spu *spu; - spu = container_of(sysdev, struct spu, sysdev); + spu = container_of(dev, struct spu, dev); return cbe_get_pmd_regs(spu_devnode(spu)); } /* returns the value for a given spu in a given register */ -static u8 spu_read_register_value(struct sys_device *sysdev, union spe_reg __iomem *reg) +static u8 spu_read_register_value(struct device *dev, union spe_reg __iomem *reg) { union spe_reg value; struct spu *spu; - spu = container_of(sysdev, struct spu, sysdev); + spu = container_of(dev, struct spu, dev); value.val = in_be64(®->val); return value.spe[spu->spe_id]; } -static ssize_t spu_show_temp(struct sys_device *sysdev, char *buf) +static ssize_t spu_show_temp(struct device *dev, struct device_attribute *attr, + char *buf) { u8 value; struct cbe_pmd_regs __iomem *pmd_regs; - pmd_regs = get_pmd_regs(sysdev); + pmd_regs = get_pmd_regs(dev); - value = spu_read_register_value(sysdev, &pmd_regs->ts_ctsr1); + value = spu_read_register_value(dev, &pmd_regs->ts_ctsr1); return sprintf(buf, "%d\n", reg_to_temp(value)); } @@ -124,7 +125,7 @@ static ssize_t show_throttle(struct cbe_pmd_regs __iomem *pmd_regs, char *buf, i static ssize_t store_throttle(struct cbe_pmd_regs __iomem *pmd_regs, const char *buf, size_t size, int pos) { u64 reg_value; - int temp; + unsigned int temp; u64 new_value; int ret; @@ -146,42 +147,48 @@ static ssize_t store_throttle(struct cbe_pmd_regs __iomem *pmd_regs, const char return size; } -static ssize_t spu_show_throttle_end(struct sys_device *sysdev, char *buf) +static ssize_t spu_show_throttle_end(struct device *dev, + struct device_attribute *attr, char *buf) { - return show_throttle(get_pmd_regs(sysdev), buf, 0); + return show_throttle(get_pmd_regs(dev), buf, 0); } -static ssize_t spu_show_throttle_begin(struct sys_device *sysdev, char *buf) +static ssize_t spu_show_throttle_begin(struct device *dev, + struct device_attribute *attr, char *buf) { - return show_throttle(get_pmd_regs(sysdev), buf, 8); + return show_throttle(get_pmd_regs(dev), buf, 8); } -static ssize_t spu_show_throttle_full_stop(struct sys_device *sysdev, char *buf) +static ssize_t spu_show_throttle_full_stop(struct device *dev, + struct device_attribute *attr, char *buf) { - return show_throttle(get_pmd_regs(sysdev), buf, 16); + return show_throttle(get_pmd_regs(dev), buf, 16); } -static ssize_t spu_store_throttle_end(struct sys_device *sysdev, const char *buf, size_t size) +static ssize_t spu_store_throttle_end(struct device *dev, + struct device_attribute *attr, const char *buf, size_t size) { - return store_throttle(get_pmd_regs(sysdev), buf, size, 0); + return store_throttle(get_pmd_regs(dev), buf, size, 0); } -static ssize_t spu_store_throttle_begin(struct sys_device *sysdev, const char *buf, size_t size) +static ssize_t spu_store_throttle_begin(struct device *dev, + struct device_attribute *attr, const char *buf, size_t size) { - return store_throttle(get_pmd_regs(sysdev), buf, size, 8); + return store_throttle(get_pmd_regs(dev), buf, size, 8); } -static ssize_t spu_store_throttle_full_stop(struct sys_device *sysdev, const char *buf, size_t size) +static ssize_t spu_store_throttle_full_stop(struct device *dev, + struct device_attribute *attr, const char *buf, size_t size) { - return store_throttle(get_pmd_regs(sysdev), buf, size, 16); + return store_throttle(get_pmd_regs(dev), buf, size, 16); } -static ssize_t ppe_show_temp(struct sys_device *sysdev, char *buf, int pos) +static ssize_t ppe_show_temp(struct device *dev, char *buf, int pos) { struct cbe_pmd_regs __iomem *pmd_regs; u64 value; - pmd_regs = cbe_get_cpu_pmd_regs(sysdev->id); + pmd_regs = cbe_get_cpu_pmd_regs(dev->id); value = in_be64(&pmd_regs->ts_ctsr2); value = (value >> pos) & 0x3f; @@ -192,56 +199,64 @@ static ssize_t ppe_show_temp(struct sys_device *sysdev, char *buf, int pos) /* shows the temperature of the DTS on the PPE, * located near the linear thermal sensor */ -static ssize_t ppe_show_temp0(struct sys_device *sysdev, char *buf) +static ssize_t ppe_show_temp0(struct device *dev, + struct device_attribute *attr, char *buf) { - return ppe_show_temp(sysdev, buf, 32); + return ppe_show_temp(dev, buf, 32); } /* shows the temperature of the second DTS on the PPE */ -static ssize_t ppe_show_temp1(struct sys_device *sysdev, char *buf) +static ssize_t ppe_show_temp1(struct device *dev, + struct device_attribute *attr, char *buf) { - return ppe_show_temp(sysdev, buf, 0); + return ppe_show_temp(dev, buf, 0); } -static ssize_t ppe_show_throttle_end(struct sys_device *sysdev, char *buf) +static ssize_t ppe_show_throttle_end(struct device *dev, + struct device_attribute *attr, char *buf) { - return show_throttle(cbe_get_cpu_pmd_regs(sysdev->id), buf, 32); + return show_throttle(cbe_get_cpu_pmd_regs(dev->id), buf, 32); } -static ssize_t ppe_show_throttle_begin(struct sys_device *sysdev, char *buf) +static ssize_t ppe_show_throttle_begin(struct device *dev, + struct device_attribute *attr, char *buf) { - return show_throttle(cbe_get_cpu_pmd_regs(sysdev->id), buf, 40); + return show_throttle(cbe_get_cpu_pmd_regs(dev->id), buf, 40); } -static ssize_t ppe_show_throttle_full_stop(struct sys_device *sysdev, char *buf) +static ssize_t ppe_show_throttle_full_stop(struct device *dev, + struct device_attribute *attr, char *buf) { - return show_throttle(cbe_get_cpu_pmd_regs(sysdev->id), buf, 48); + return show_throttle(cbe_get_cpu_pmd_regs(dev->id), buf, 48); } -static ssize_t ppe_store_throttle_end(struct sys_device *sysdev, const char *buf, size_t size) +static ssize_t ppe_store_throttle_end(struct device *dev, + struct device_attribute *attr, const char *buf, size_t size) { - return store_throttle(cbe_get_cpu_pmd_regs(sysdev->id), buf, size, 32); + return store_throttle(cbe_get_cpu_pmd_regs(dev->id), buf, size, 32); } -static ssize_t ppe_store_throttle_begin(struct sys_device *sysdev, const char *buf, size_t size) +static ssize_t ppe_store_throttle_begin(struct device *dev, + struct device_attribute *attr, const char *buf, size_t size) { - return store_throttle(cbe_get_cpu_pmd_regs(sysdev->id), buf, size, 40); + return store_throttle(cbe_get_cpu_pmd_regs(dev->id), buf, size, 40); } -static ssize_t ppe_store_throttle_full_stop(struct sys_device *sysdev, const char *buf, size_t size) +static ssize_t ppe_store_throttle_full_stop(struct device *dev, + struct device_attribute *attr, const char *buf, size_t size) { - return store_throttle(cbe_get_cpu_pmd_regs(sysdev->id), buf, size, 48); + return store_throttle(cbe_get_cpu_pmd_regs(dev->id), buf, size, 48); } -static struct sysdev_attribute attr_spu_temperature = { +static struct device_attribute attr_spu_temperature = { .attr = {.name = "temperature", .mode = 0400 }, .show = spu_show_temp, }; -static SYSDEV_PREFIX_ATTR(spu, throttle_end, 0600); -static SYSDEV_PREFIX_ATTR(spu, throttle_begin, 0600); -static SYSDEV_PREFIX_ATTR(spu, throttle_full_stop, 0600); +static DEVICE_PREFIX_ATTR(spu, throttle_end, 0600); +static DEVICE_PREFIX_ATTR(spu, throttle_begin, 0600); +static DEVICE_PREFIX_ATTR(spu, throttle_full_stop, 0600); static struct attribute *spu_attributes[] = { @@ -257,19 +272,19 @@ static struct attribute_group spu_attribute_group = { .attrs = spu_attributes, }; -static struct sysdev_attribute attr_ppe_temperature0 = { +static struct device_attribute attr_ppe_temperature0 = { .attr = {.name = "temperature0", .mode = 0400 }, .show = ppe_show_temp0, }; -static struct sysdev_attribute attr_ppe_temperature1 = { +static struct device_attribute attr_ppe_temperature1 = { .attr = {.name = "temperature1", .mode = 0400 }, .show = ppe_show_temp1, }; -static SYSDEV_PREFIX_ATTR(ppe, throttle_end, 0600); -static SYSDEV_PREFIX_ATTR(ppe, throttle_begin, 0600); -static SYSDEV_PREFIX_ATTR(ppe, throttle_full_stop, 0600); +static DEVICE_PREFIX_ATTR(ppe, throttle_end, 0600); +static DEVICE_PREFIX_ATTR(ppe, throttle_begin, 0600); +static DEVICE_PREFIX_ATTR(ppe, throttle_full_stop, 0600); static struct attribute *ppe_attributes[] = { &attr_ppe_temperature0.attr, @@ -292,7 +307,7 @@ static int __init init_default_values(void) { int cpu; struct cbe_pmd_regs __iomem *pmd_regs; - struct sys_device *sysdev; + struct device *dev; union ppe_spe_reg tpr; union spe_reg str1; u64 str2; @@ -334,14 +349,14 @@ static int __init init_default_values(void) for_each_possible_cpu (cpu) { pr_debug("processing cpu %d\n", cpu); - sysdev = get_cpu_sysdev(cpu); + dev = get_cpu_device(cpu); - if (!sysdev) { - pr_info("invalid sysdev pointer for cbe_thermal\n"); + if (!dev) { + pr_info("invalid dev pointer for cbe_thermal\n"); return -EINVAL; } - pmd_regs = cbe_get_cpu_pmd_regs(sysdev->id); + pmd_regs = cbe_get_cpu_pmd_regs(dev->id); if (!pmd_regs) { pr_info("invalid CBE regs pointer for cbe_thermal\n"); @@ -364,8 +379,8 @@ static int __init thermal_init(void) int rc = init_default_values(); if (rc == 0) { - spu_add_sysdev_attr_group(&spu_attribute_group); - cpu_add_sysdev_attr_group(&ppe_attribute_group); + spu_add_dev_attr_group(&spu_attribute_group); + cpu_add_dev_attr_group(&ppe_attribute_group); } return rc; @@ -374,8 +389,8 @@ module_init(thermal_init); static void __exit thermal_exit(void) { - spu_remove_sysdev_attr_group(&spu_attribute_group); - cpu_remove_sysdev_attr_group(&ppe_attribute_group); + spu_remove_dev_attr_group(&spu_attribute_group); + cpu_remove_dev_attr_group(&ppe_attribute_group); } module_exit(thermal_exit); diff --git a/arch/powerpc/platforms/cell/celleb_pci.c b/arch/powerpc/platforms/cell/celleb_pci.c new file mode 100644 index 00000000000..173568140a3 --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_pci.c @@ -0,0 +1,500 @@ +/* + * Support for PCI on Celleb platform. + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This code is based on arch/powerpc/kernel/rtas_pci.c: + * Copyright (C) 2001 Dave Engebretsen, IBM Corporation + * Copyright (C) 2003 Anton Blanchard <anton@au.ibm.com>, IBM + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#undef DEBUG + +#include <linux/kernel.h> +#include <linux/threads.h> +#include <linux/pci.h> +#include <linux/string.h> +#include <linux/init.h> +#include <linux/bootmem.h> +#include <linux/pci_regs.h> +#include <linux/of.h> +#include <linux/of_device.h> +#include <linux/slab.h> + +#include <asm/io.h> +#include <asm/irq.h> +#include <asm/prom.h> +#include <asm/pci-bridge.h> +#include <asm/ppc-pci.h> + +#include "celleb_pci.h" + +#define MAX_PCI_DEVICES 32 +#define MAX_PCI_FUNCTIONS 8 +#define MAX_PCI_BASE_ADDRS 3 /* use 64 bit address */ + +/* definition for fake pci configuration area for GbE, .... ,and etc. */ + +struct celleb_pci_resource { + struct resource r[MAX_PCI_BASE_ADDRS]; +}; + +struct celleb_pci_private { + unsigned char *fake_config[MAX_PCI_DEVICES][MAX_PCI_FUNCTIONS]; + struct celleb_pci_resource *res[MAX_PCI_DEVICES][MAX_PCI_FUNCTIONS]; +}; + +static inline u8 celleb_fake_config_readb(void *addr) +{ + u8 *p = addr; + return *p; +} + +static inline u16 celleb_fake_config_readw(void *addr) +{ + __le16 *p = addr; + return le16_to_cpu(*p); +} + +static inline u32 celleb_fake_config_readl(void *addr) +{ + __le32 *p = addr; + return le32_to_cpu(*p); +} + +static inline void celleb_fake_config_writeb(u32 val, void *addr) +{ + u8 *p = addr; + *p = val; +} + +static inline void celleb_fake_config_writew(u32 val, void *addr) +{ + __le16 val16; + __le16 *p = addr; + val16 = cpu_to_le16(val); + *p = val16; +} + +static inline void celleb_fake_config_writel(u32 val, void *addr) +{ + __le32 val32; + __le32 *p = addr; + val32 = cpu_to_le32(val); + *p = val32; +} + +static unsigned char *get_fake_config_start(struct pci_controller *hose, + int devno, int fn) +{ + struct celleb_pci_private *private = hose->private_data; + + if (private == NULL) + return NULL; + + return private->fake_config[devno][fn]; +} + +static struct celleb_pci_resource *get_resource_start( + struct pci_controller *hose, + int devno, int fn) +{ + struct celleb_pci_private *private = hose->private_data; + + if (private == NULL) + return NULL; + + return private->res[devno][fn]; +} + + +static void celleb_config_read_fake(unsigned char *config, int where, + int size, u32 *val) +{ + char *p = config + where; + + switch (size) { + case 1: + *val = celleb_fake_config_readb(p); + break; + case 2: + *val = celleb_fake_config_readw(p); + break; + case 4: + *val = celleb_fake_config_readl(p); + break; + } +} + +static void celleb_config_write_fake(unsigned char *config, int where, + int size, u32 val) +{ + char *p = config + where; + + switch (size) { + case 1: + celleb_fake_config_writeb(val, p); + break; + case 2: + celleb_fake_config_writew(val, p); + break; + case 4: + celleb_fake_config_writel(val, p); + break; + } +} + +static int celleb_fake_pci_read_config(struct pci_bus *bus, + unsigned int devfn, int where, int size, u32 *val) +{ + char *config; + struct pci_controller *hose = pci_bus_to_host(bus); + unsigned int devno = devfn >> 3; + unsigned int fn = devfn & 0x7; + + /* allignment check */ + BUG_ON(where % size); + + pr_debug(" fake read: bus=0x%x, ", bus->number); + config = get_fake_config_start(hose, devno, fn); + + pr_debug("devno=0x%x, where=0x%x, size=0x%x, ", devno, where, size); + if (!config) { + pr_debug("failed\n"); + return PCIBIOS_DEVICE_NOT_FOUND; + } + + celleb_config_read_fake(config, where, size, val); + pr_debug("val=0x%x\n", *val); + + return PCIBIOS_SUCCESSFUL; +} + + +static int celleb_fake_pci_write_config(struct pci_bus *bus, + unsigned int devfn, int where, int size, u32 val) +{ + char *config; + struct pci_controller *hose = pci_bus_to_host(bus); + struct celleb_pci_resource *res; + unsigned int devno = devfn >> 3; + unsigned int fn = devfn & 0x7; + + /* allignment check */ + BUG_ON(where % size); + + config = get_fake_config_start(hose, devno, fn); + + if (!config) + return PCIBIOS_DEVICE_NOT_FOUND; + + if (val == ~0) { + int i = (where - PCI_BASE_ADDRESS_0) >> 3; + + switch (where) { + case PCI_BASE_ADDRESS_0: + case PCI_BASE_ADDRESS_2: + if (size != 4) + return PCIBIOS_DEVICE_NOT_FOUND; + res = get_resource_start(hose, devno, fn); + if (!res) + return PCIBIOS_DEVICE_NOT_FOUND; + celleb_config_write_fake(config, where, size, + (res->r[i].end - res->r[i].start)); + return PCIBIOS_SUCCESSFUL; + case PCI_BASE_ADDRESS_1: + case PCI_BASE_ADDRESS_3: + case PCI_BASE_ADDRESS_4: + case PCI_BASE_ADDRESS_5: + break; + default: + break; + } + } + + celleb_config_write_fake(config, where, size, val); + pr_debug(" fake write: where=%x, size=%d, val=%x\n", + where, size, val); + + return PCIBIOS_SUCCESSFUL; +} + +static struct pci_ops celleb_fake_pci_ops = { + .read = celleb_fake_pci_read_config, + .write = celleb_fake_pci_write_config, +}; + +static inline void celleb_setup_pci_base_addrs(struct pci_controller *hose, + unsigned int devno, unsigned int fn, + unsigned int num_base_addr) +{ + u32 val; + unsigned char *config; + struct celleb_pci_resource *res; + + config = get_fake_config_start(hose, devno, fn); + res = get_resource_start(hose, devno, fn); + + if (!config || !res) + return; + + switch (num_base_addr) { + case 3: + val = (res->r[2].start & 0xfffffff0) + | PCI_BASE_ADDRESS_MEM_TYPE_64; + celleb_config_write_fake(config, PCI_BASE_ADDRESS_4, 4, val); + val = res->r[2].start >> 32; + celleb_config_write_fake(config, PCI_BASE_ADDRESS_5, 4, val); + /* FALLTHROUGH */ + case 2: + val = (res->r[1].start & 0xfffffff0) + | PCI_BASE_ADDRESS_MEM_TYPE_64; + celleb_config_write_fake(config, PCI_BASE_ADDRESS_2, 4, val); + val = res->r[1].start >> 32; + celleb_config_write_fake(config, PCI_BASE_ADDRESS_3, 4, val); + /* FALLTHROUGH */ + case 1: + val = (res->r[0].start & 0xfffffff0) + | PCI_BASE_ADDRESS_MEM_TYPE_64; + celleb_config_write_fake(config, PCI_BASE_ADDRESS_0, 4, val); + val = res->r[0].start >> 32; + celleb_config_write_fake(config, PCI_BASE_ADDRESS_1, 4, val); + break; + } + + val = PCI_COMMAND_IO | PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER; + celleb_config_write_fake(config, PCI_COMMAND, 2, val); +} + +static int __init celleb_setup_fake_pci_device(struct device_node *node, + struct pci_controller *hose) +{ + unsigned int rlen; + int num_base_addr = 0; + u32 val; + const u32 *wi0, *wi1, *wi2, *wi3, *wi4; + unsigned int devno, fn; + struct celleb_pci_private *private = hose->private_data; + unsigned char **config = NULL; + struct celleb_pci_resource **res = NULL; + const char *name; + const unsigned long *li; + int size, result; + + if (private == NULL) { + printk(KERN_ERR "PCI: " + "memory space for pci controller is not assigned\n"); + goto error; + } + + name = of_get_property(node, "model", &rlen); + if (!name) { + printk(KERN_ERR "PCI: model property not found.\n"); + goto error; + } + + wi4 = of_get_property(node, "reg", &rlen); + if (wi4 == NULL) + goto error; + + devno = ((wi4[0] >> 8) & 0xff) >> 3; + fn = (wi4[0] >> 8) & 0x7; + + pr_debug("PCI: celleb_setup_fake_pci() %s devno=%x fn=%x\n", name, + devno, fn); + + size = 256; + config = &private->fake_config[devno][fn]; + *config = zalloc_maybe_bootmem(size, GFP_KERNEL); + if (*config == NULL) { + printk(KERN_ERR "PCI: " + "not enough memory for fake configuration space\n"); + goto error; + } + pr_debug("PCI: fake config area assigned 0x%016lx\n", + (unsigned long)*config); + + size = sizeof(struct celleb_pci_resource); + res = &private->res[devno][fn]; + *res = zalloc_maybe_bootmem(size, GFP_KERNEL); + if (*res == NULL) { + printk(KERN_ERR + "PCI: not enough memory for resource data space\n"); + goto error; + } + pr_debug("PCI: res assigned 0x%016lx\n", (unsigned long)*res); + + wi0 = of_get_property(node, "device-id", NULL); + wi1 = of_get_property(node, "vendor-id", NULL); + wi2 = of_get_property(node, "class-code", NULL); + wi3 = of_get_property(node, "revision-id", NULL); + if (!wi0 || !wi1 || !wi2 || !wi3) { + printk(KERN_ERR "PCI: Missing device tree properties.\n"); + goto error; + } + + celleb_config_write_fake(*config, PCI_DEVICE_ID, 2, wi0[0] & 0xffff); + celleb_config_write_fake(*config, PCI_VENDOR_ID, 2, wi1[0] & 0xffff); + pr_debug("class-code = 0x%08x\n", wi2[0]); + + celleb_config_write_fake(*config, PCI_CLASS_PROG, 1, wi2[0] & 0xff); + celleb_config_write_fake(*config, PCI_CLASS_DEVICE, 2, + (wi2[0] >> 8) & 0xffff); + celleb_config_write_fake(*config, PCI_REVISION_ID, 1, wi3[0]); + + while (num_base_addr < MAX_PCI_BASE_ADDRS) { + result = of_address_to_resource(node, + num_base_addr, &(*res)->r[num_base_addr]); + if (result) + break; + num_base_addr++; + } + + celleb_setup_pci_base_addrs(hose, devno, fn, num_base_addr); + + li = of_get_property(node, "interrupts", &rlen); + if (!li) { + printk(KERN_ERR "PCI: interrupts not found.\n"); + goto error; + } + val = li[0]; + celleb_config_write_fake(*config, PCI_INTERRUPT_PIN, 1, 1); + celleb_config_write_fake(*config, PCI_INTERRUPT_LINE, 1, val); + +#ifdef DEBUG + pr_debug("PCI: %s irq=%ld\n", name, li[0]); + for (i = 0; i < 6; i++) { + celleb_config_read_fake(*config, + PCI_BASE_ADDRESS_0 + 0x4 * i, 4, + &val); + pr_debug("PCI: %s fn=%d base_address_%d=0x%x\n", + name, fn, i, val); + } +#endif + + celleb_config_write_fake(*config, PCI_HEADER_TYPE, 1, + PCI_HEADER_TYPE_NORMAL); + + return 0; + +error: + if (mem_init_done) { + if (config && *config) + kfree(*config); + if (res && *res) + kfree(*res); + + } else { + if (config && *config) { + size = 256; + free_bootmem(__pa(*config), size); + } + if (res && *res) { + size = sizeof(struct celleb_pci_resource); + free_bootmem(__pa(*res), size); + } + } + + return 1; +} + +static int __init phb_set_bus_ranges(struct device_node *dev, + struct pci_controller *phb) +{ + const int *bus_range; + unsigned int len; + + bus_range = of_get_property(dev, "bus-range", &len); + if (bus_range == NULL || len < 2 * sizeof(int)) + return 1; + + phb->first_busno = bus_range[0]; + phb->last_busno = bus_range[1]; + + return 0; +} + +static void __init celleb_alloc_private_mem(struct pci_controller *hose) +{ + hose->private_data = + zalloc_maybe_bootmem(sizeof(struct celleb_pci_private), + GFP_KERNEL); +} + +static int __init celleb_setup_fake_pci(struct device_node *dev, + struct pci_controller *phb) +{ + struct device_node *node; + + phb->ops = &celleb_fake_pci_ops; + celleb_alloc_private_mem(phb); + + for (node = of_get_next_child(dev, NULL); + node != NULL; node = of_get_next_child(dev, node)) + celleb_setup_fake_pci_device(node, phb); + + return 0; +} + +static struct celleb_phb_spec celleb_fake_pci_spec __initdata = { + .setup = celleb_setup_fake_pci, +}; + +static struct of_device_id celleb_phb_match[] __initdata = { + { + .name = "pci-pseudo", + .data = &celleb_fake_pci_spec, + }, { + .name = "epci", + .data = &celleb_epci_spec, + }, { + .name = "pcie", + .data = &celleb_pciex_spec, + }, { + }, +}; + +int __init celleb_setup_phb(struct pci_controller *phb) +{ + struct device_node *dev = phb->dn; + const struct of_device_id *match; + const struct celleb_phb_spec *phb_spec; + int rc; + + match = of_match_node(celleb_phb_match, dev); + if (!match) + return 1; + + phb_set_bus_ranges(dev, phb); + phb->buid = 1; + + phb_spec = match->data; + rc = (*phb_spec->setup)(dev, phb); + if (rc) + return 1; + + if (phb_spec->ops) + iowa_register_bus(phb, phb_spec->ops, + phb_spec->iowa_init, + phb_spec->iowa_data); + return 0; +} + +int celleb_pci_probe_mode(struct pci_bus *bus) +{ + return PCI_PROBE_DEVTREE; +} diff --git a/arch/powerpc/platforms/cell/celleb_pci.h b/arch/powerpc/platforms/cell/celleb_pci.h new file mode 100644 index 00000000000..a801fcc5f38 --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_pci.h @@ -0,0 +1,46 @@ +/* + * pci prototypes for Celleb platform + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef _CELLEB_PCI_H +#define _CELLEB_PCI_H + +#include <linux/pci.h> + +#include <asm/pci-bridge.h> +#include <asm/prom.h> +#include <asm/ppc-pci.h> +#include <asm/io-workarounds.h> + +struct iowa_bus; + +struct celleb_phb_spec { + int (*setup)(struct device_node *, struct pci_controller *); + struct ppc_pci_io *ops; + int (*iowa_init)(struct iowa_bus *, void *); + void *iowa_data; +}; + +extern int celleb_setup_phb(struct pci_controller *); +extern int celleb_pci_probe_mode(struct pci_bus *); + +extern struct celleb_phb_spec celleb_epci_spec; +extern struct celleb_phb_spec celleb_pciex_spec; + +#endif /* _CELLEB_PCI_H */ diff --git a/arch/powerpc/platforms/cell/celleb_scc.h b/arch/powerpc/platforms/cell/celleb_scc.h new file mode 100644 index 00000000000..b596a711c34 --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_scc.h @@ -0,0 +1,232 @@ +/* + * SCC (Super Companion Chip) definitions + * + * (C) Copyright 2004-2006 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#ifndef _CELLEB_SCC_H +#define _CELLEB_SCC_H + +#define PCI_VENDOR_ID_TOSHIBA_2 0x102f +#define PCI_DEVICE_ID_TOSHIBA_SCC_PCIEXC_BRIDGE 0x01b0 +#define PCI_DEVICE_ID_TOSHIBA_SCC_EPCI_BRIDGE 0x01b1 +#define PCI_DEVICE_ID_TOSHIBA_SCC_BRIDGE 0x01b2 +#define PCI_DEVICE_ID_TOSHIBA_SCC_GBE 0x01b3 +#define PCI_DEVICE_ID_TOSHIBA_SCC_ATA 0x01b4 +#define PCI_DEVICE_ID_TOSHIBA_SCC_USB2 0x01b5 +#define PCI_DEVICE_ID_TOSHIBA_SCC_USB 0x01b6 +#define PCI_DEVICE_ID_TOSHIBA_SCC_ENCDEC 0x01b7 + +#define SCC_EPCI_REG 0x0000d000 + +/* EPCI registers */ +#define SCC_EPCI_CNF10_REG 0x010 +#define SCC_EPCI_CNF14_REG 0x014 +#define SCC_EPCI_CNF18_REG 0x018 +#define SCC_EPCI_PVBAT 0x100 +#define SCC_EPCI_VPMBAT 0x104 +#define SCC_EPCI_VPIBAT 0x108 +#define SCC_EPCI_VCSR 0x110 +#define SCC_EPCI_VIENAB 0x114 +#define SCC_EPCI_VISTAT 0x118 +#define SCC_EPCI_VRDCOUNT 0x124 +#define SCC_EPCI_BAM0 0x12c +#define SCC_EPCI_BAM1 0x134 +#define SCC_EPCI_BAM2 0x13c +#define SCC_EPCI_IADR 0x164 +#define SCC_EPCI_CLKRST 0x800 +#define SCC_EPCI_INTSET 0x804 +#define SCC_EPCI_STATUS 0x808 +#define SCC_EPCI_ABTSET 0x80c +#define SCC_EPCI_WATRP 0x810 +#define SCC_EPCI_DUMYRADR 0x814 +#define SCC_EPCI_SWRESP 0x818 +#define SCC_EPCI_CNTOPT 0x81c +#define SCC_EPCI_ECMODE 0xf00 +#define SCC_EPCI_IOM_AC_NUM 5 +#define SCC_EPCI_IOM_ACTE(n) (0xf10 + (n) * 4) +#define SCC_EPCI_IOT_AC_NUM 4 +#define SCC_EPCI_IOT_ACTE(n) (0xf30 + (n) * 4) +#define SCC_EPCI_MAEA 0xf50 +#define SCC_EPCI_MAEC 0xf54 +#define SCC_EPCI_CKCTRL 0xff0 + +/* bits for SCC_EPCI_VCSR */ +#define SCC_EPCI_VCSR_FRE 0x00020000 +#define SCC_EPCI_VCSR_FWE 0x00010000 +#define SCC_EPCI_VCSR_DR 0x00000400 +#define SCC_EPCI_VCSR_SR 0x00000008 +#define SCC_EPCI_VCSR_AT 0x00000004 + +/* bits for SCC_EPCI_VIENAB/SCC_EPCI_VISTAT */ +#define SCC_EPCI_VISTAT_PMPE 0x00000008 +#define SCC_EPCI_VISTAT_PMFE 0x00000004 +#define SCC_EPCI_VISTAT_PRA 0x00000002 +#define SCC_EPCI_VISTAT_PRD 0x00000001 +#define SCC_EPCI_VISTAT_ALL 0x0000000f + +#define SCC_EPCI_VIENAB_PMPEE 0x00000008 +#define SCC_EPCI_VIENAB_PMFEE 0x00000004 +#define SCC_EPCI_VIENAB_PRA 0x00000002 +#define SCC_EPCI_VIENAB_PRD 0x00000001 +#define SCC_EPCI_VIENAB_ALL 0x0000000f + +/* bits for SCC_EPCI_CLKRST */ +#define SCC_EPCI_CLKRST_CKS_MASK 0x00030000 +#define SCC_EPCI_CLKRST_CKS_2 0x00000000 +#define SCC_EPCI_CLKRST_CKS_4 0x00010000 +#define SCC_EPCI_CLKRST_CKS_8 0x00020000 +#define SCC_EPCI_CLKRST_PCICRST 0x00000400 +#define SCC_EPCI_CLKRST_BC 0x00000200 +#define SCC_EPCI_CLKRST_PCIRST 0x00000100 +#define SCC_EPCI_CLKRST_PCKEN 0x00000001 + +/* bits for SCC_EPCI_INTSET/SCC_EPCI_STATUS */ +#define SCC_EPCI_INT_2M 0x01000000 +#define SCC_EPCI_INT_RERR 0x00200000 +#define SCC_EPCI_INT_SERR 0x00100000 +#define SCC_EPCI_INT_PRTER 0x00080000 +#define SCC_EPCI_INT_SER 0x00040000 +#define SCC_EPCI_INT_PER 0x00020000 +#define SCC_EPCI_INT_PAI 0x00010000 +#define SCC_EPCI_INT_1M 0x00000100 +#define SCC_EPCI_INT_PME 0x00000010 +#define SCC_EPCI_INT_INTD 0x00000008 +#define SCC_EPCI_INT_INTC 0x00000004 +#define SCC_EPCI_INT_INTB 0x00000002 +#define SCC_EPCI_INT_INTA 0x00000001 +#define SCC_EPCI_INT_DEVINT 0x0000000f +#define SCC_EPCI_INT_ALL 0x003f001f +#define SCC_EPCI_INT_ALLERR 0x003f0000 + +/* bits for SCC_EPCI_CKCTRL */ +#define SCC_EPCI_CKCTRL_CRST0 0x00010000 +#define SCC_EPCI_CKCTRL_CRST1 0x00020000 +#define SCC_EPCI_CKCTRL_OCLKEN 0x00000100 +#define SCC_EPCI_CKCTRL_LCLKEN 0x00000001 + +#define SCC_EPCI_IDSEL_AD_TO_SLOT(ad) ((ad) - 10) +#define SCC_EPCI_MAX_DEVNU SCC_EPCI_IDSEL_AD_TO_SLOT(32) + +/* bits for SCC_EPCI_CNTOPT */ +#define SCC_EPCI_CNTOPT_O2PMB 0x00000002 + +/* SCC PCIEXC SMMIO registers */ +#define PEXCADRS 0x000 +#define PEXCWDATA 0x004 +#define PEXCRDATA 0x008 +#define PEXDADRS 0x010 +#define PEXDCMND 0x014 +#define PEXDWDATA 0x018 +#define PEXDRDATA 0x01c +#define PEXREQID 0x020 +#define PEXTIDMAP 0x024 +#define PEXINTMASK 0x028 +#define PEXINTSTS 0x02c +#define PEXAERRMASK 0x030 +#define PEXAERRSTS 0x034 +#define PEXPRERRMASK 0x040 +#define PEXPRERRSTS 0x044 +#define PEXPRERRID01 0x048 +#define PEXPRERRID23 0x04c +#define PEXVDMASK 0x050 +#define PEXVDSTS 0x054 +#define PEXRCVCPLIDA 0x060 +#define PEXLENERRIDA 0x068 +#define PEXPHYPLLST 0x070 +#define PEXDMRDEN0 0x100 +#define PEXDMRDADR0 0x104 +#define PEXDMRDENX 0x110 +#define PEXDMRDADRX 0x114 +#define PEXECMODE 0xf00 +#define PEXMAEA(n) (0xf50 + (8 * n)) +#define PEXMAEC(n) (0xf54 + (8 * n)) +#define PEXCCRCTRL 0xff0 + +/* SCC PCIEXC bits and shifts for PEXCADRS */ +#define PEXCADRS_BYTE_EN_SHIFT 20 +#define PEXCADRS_CMD_SHIFT 16 +#define PEXCADRS_CMD_READ (0xa << PEXCADRS_CMD_SHIFT) +#define PEXCADRS_CMD_WRITE (0xb << PEXCADRS_CMD_SHIFT) + +/* SCC PCIEXC shifts for PEXDADRS */ +#define PEXDADRS_BUSNO_SHIFT 20 +#define PEXDADRS_DEVNO_SHIFT 15 +#define PEXDADRS_FUNCNO_SHIFT 12 + +/* SCC PCIEXC bits and shifts for PEXDCMND */ +#define PEXDCMND_BYTE_EN_SHIFT 4 +#define PEXDCMND_IO_READ 0x2 +#define PEXDCMND_IO_WRITE 0x3 +#define PEXDCMND_CONFIG_READ 0xa +#define PEXDCMND_CONFIG_WRITE 0xb + +/* SCC PCIEXC bits for PEXPHYPLLST */ +#define PEXPHYPLLST_PEXPHYAPLLST 0x00000001 + +/* SCC PCIEXC bits for PEXECMODE */ +#define PEXECMODE_ALL_THROUGH 0x00000000 +#define PEXECMODE_ALL_8BIT 0x00550155 +#define PEXECMODE_ALL_16BIT 0x00aa02aa + +/* SCC PCIEXC bits for PEXCCRCTRL */ +#define PEXCCRCTRL_PEXIPCOREEN 0x00040000 +#define PEXCCRCTRL_PEXIPCONTEN 0x00020000 +#define PEXCCRCTRL_PEXPHYPLLEN 0x00010000 +#define PEXCCRCTRL_PCIEXCAOCKEN 0x00000100 + +/* SCC PCIEXC port configuration registers */ +#define PEXTCERRCHK 0x21c +#define PEXTAMAPB0 0x220 +#define PEXTAMAPL0 0x224 +#define PEXTAMAPB(n) (PEXTAMAPB0 + 8 * (n)) +#define PEXTAMAPL(n) (PEXTAMAPL0 + 8 * (n)) +#define PEXCHVC0P 0x500 +#define PEXCHVC0NP 0x504 +#define PEXCHVC0C 0x508 +#define PEXCDVC0P 0x50c +#define PEXCDVC0NP 0x510 +#define PEXCDVC0C 0x514 +#define PEXCHVCXP 0x518 +#define PEXCHVCXNP 0x51c +#define PEXCHVCXC 0x520 +#define PEXCDVCXP 0x524 +#define PEXCDVCXNP 0x528 +#define PEXCDVCXC 0x52c +#define PEXCTTRG 0x530 +#define PEXTSCTRL 0x700 +#define PEXTSSTS 0x704 +#define PEXSKPCTRL 0x708 + +/* UHC registers */ +#define SCC_UHC_CKRCTRL 0xff0 +#define SCC_UHC_ECMODE 0xf00 + +/* bits for SCC_UHC_CKRCTRL */ +#define SCC_UHC_F48MCKLEN 0x00000001 +#define SCC_UHC_P_SUSPEND 0x00000002 +#define SCC_UHC_PHY_SUSPEND_SEL 0x00000004 +#define SCC_UHC_HCLKEN 0x00000100 +#define SCC_UHC_USBEN 0x00010000 +#define SCC_UHC_USBCEN 0x00020000 +#define SCC_UHC_PHYEN 0x00040000 + +/* bits for SCC_UHC_ECMODE */ +#define SCC_UHC_ECMODE_BY_BYTE 0x00000555 +#define SCC_UHC_ECMODE_BY_WORD 0x00000aaa + +#endif /* _CELLEB_SCC_H */ diff --git a/arch/powerpc/platforms/cell/celleb_scc_epci.c b/arch/powerpc/platforms/cell/celleb_scc_epci.c new file mode 100644 index 00000000000..844c0facb4f --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_scc_epci.c @@ -0,0 +1,429 @@ +/* + * Support for SCC external PCI + * + * (C) Copyright 2004-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#undef DEBUG + +#include <linux/kernel.h> +#include <linux/threads.h> +#include <linux/pci.h> +#include <linux/init.h> +#include <linux/pci_regs.h> +#include <linux/bootmem.h> + +#include <asm/io.h> +#include <asm/irq.h> +#include <asm/prom.h> +#include <asm/pci-bridge.h> +#include <asm/ppc-pci.h> + +#include "celleb_scc.h" +#include "celleb_pci.h" + +#define MAX_PCI_DEVICES 32 +#define MAX_PCI_FUNCTIONS 8 + +#define iob() __asm__ __volatile__("eieio; sync":::"memory") + +static inline PCI_IO_ADDR celleb_epci_get_epci_base( + struct pci_controller *hose) +{ + /* + * Note: + * Celleb epci uses cfg_addr as a base address for + * epci control registers. + */ + + return hose->cfg_addr; +} + +static inline PCI_IO_ADDR celleb_epci_get_epci_cfg( + struct pci_controller *hose) +{ + /* + * Note: + * Celleb epci uses cfg_data as a base address for + * configuration area for epci devices. + */ + + return hose->cfg_data; +} + +static inline void clear_and_disable_master_abort_interrupt( + struct pci_controller *hose) +{ + PCI_IO_ADDR epci_base; + PCI_IO_ADDR reg; + epci_base = celleb_epci_get_epci_base(hose); + reg = epci_base + PCI_COMMAND; + out_be32(reg, in_be32(reg) | (PCI_STATUS_REC_MASTER_ABORT << 16)); +} + +static int celleb_epci_check_abort(struct pci_controller *hose, + PCI_IO_ADDR addr) +{ + PCI_IO_ADDR reg; + PCI_IO_ADDR epci_base; + u32 val; + + iob(); + epci_base = celleb_epci_get_epci_base(hose); + + reg = epci_base + PCI_COMMAND; + val = in_be32(reg); + + if (val & (PCI_STATUS_REC_MASTER_ABORT << 16)) { + out_be32(reg, + (val & 0xffff) | (PCI_STATUS_REC_MASTER_ABORT << 16)); + + /* clear PCI Controller error, FRE, PMFE */ + reg = epci_base + SCC_EPCI_STATUS; + out_be32(reg, SCC_EPCI_INT_PAI); + + reg = epci_base + SCC_EPCI_VCSR; + val = in_be32(reg) & 0xffff; + val |= SCC_EPCI_VCSR_FRE; + out_be32(reg, val); + + reg = epci_base + SCC_EPCI_VISTAT; + out_be32(reg, SCC_EPCI_VISTAT_PMFE); + return PCIBIOS_DEVICE_NOT_FOUND; + } + + return PCIBIOS_SUCCESSFUL; +} + +static PCI_IO_ADDR celleb_epci_make_config_addr(struct pci_bus *bus, + struct pci_controller *hose, unsigned int devfn, int where) +{ + PCI_IO_ADDR addr; + + if (bus != hose->bus) + addr = celleb_epci_get_epci_cfg(hose) + + (((bus->number & 0xff) << 16) + | ((devfn & 0xff) << 8) + | (where & 0xff) + | 0x01000000); + else + addr = celleb_epci_get_epci_cfg(hose) + + (((devfn & 0xff) << 8) | (where & 0xff)); + + pr_debug("EPCI: config_addr = 0x%p\n", addr); + + return addr; +} + +static int celleb_epci_read_config(struct pci_bus *bus, + unsigned int devfn, int where, int size, u32 *val) +{ + PCI_IO_ADDR epci_base; + PCI_IO_ADDR addr; + struct pci_controller *hose = pci_bus_to_host(bus); + + /* allignment check */ + BUG_ON(where % size); + + if (!celleb_epci_get_epci_cfg(hose)) + return PCIBIOS_DEVICE_NOT_FOUND; + + if (bus->number == hose->first_busno && devfn == 0) { + /* EPCI controller self */ + + epci_base = celleb_epci_get_epci_base(hose); + addr = epci_base + where; + + switch (size) { + case 1: + *val = in_8(addr); + break; + case 2: + *val = in_be16(addr); + break; + case 4: + *val = in_be32(addr); + break; + default: + return PCIBIOS_DEVICE_NOT_FOUND; + } + + } else { + + clear_and_disable_master_abort_interrupt(hose); + addr = celleb_epci_make_config_addr(bus, hose, devfn, where); + + switch (size) { + case 1: + *val = in_8(addr); + break; + case 2: + *val = in_le16(addr); + break; + case 4: + *val = in_le32(addr); + break; + default: + return PCIBIOS_DEVICE_NOT_FOUND; + } + } + + pr_debug("EPCI: " + "addr=0x%p, devfn=0x%x, where=0x%x, size=0x%x, val=0x%x\n", + addr, devfn, where, size, *val); + + return celleb_epci_check_abort(hose, NULL); +} + +static int celleb_epci_write_config(struct pci_bus *bus, + unsigned int devfn, int where, int size, u32 val) +{ + PCI_IO_ADDR epci_base; + PCI_IO_ADDR addr; + struct pci_controller *hose = pci_bus_to_host(bus); + + /* allignment check */ + BUG_ON(where % size); + + if (!celleb_epci_get_epci_cfg(hose)) + return PCIBIOS_DEVICE_NOT_FOUND; + + if (bus->number == hose->first_busno && devfn == 0) { + /* EPCI controller self */ + + epci_base = celleb_epci_get_epci_base(hose); + addr = epci_base + where; + + switch (size) { + case 1: + out_8(addr, val); + break; + case 2: + out_be16(addr, val); + break; + case 4: + out_be32(addr, val); + break; + default: + return PCIBIOS_DEVICE_NOT_FOUND; + } + + } else { + + clear_and_disable_master_abort_interrupt(hose); + addr = celleb_epci_make_config_addr(bus, hose, devfn, where); + + switch (size) { + case 1: + out_8(addr, val); + break; + case 2: + out_le16(addr, val); + break; + case 4: + out_le32(addr, val); + break; + default: + return PCIBIOS_DEVICE_NOT_FOUND; + } + } + + return celleb_epci_check_abort(hose, addr); +} + +struct pci_ops celleb_epci_ops = { + .read = celleb_epci_read_config, + .write = celleb_epci_write_config, +}; + +/* to be moved in FW */ +static int __init celleb_epci_init(struct pci_controller *hose) +{ + u32 val; + PCI_IO_ADDR reg; + PCI_IO_ADDR epci_base; + int hwres = 0; + + epci_base = celleb_epci_get_epci_base(hose); + + /* PCI core reset(Internal bus and PCI clock) */ + reg = epci_base + SCC_EPCI_CKCTRL; + val = in_be32(reg); + if (val == 0x00030101) + hwres = 1; + else { + val &= ~(SCC_EPCI_CKCTRL_CRST0 | SCC_EPCI_CKCTRL_CRST1); + out_be32(reg, val); + + /* set PCI core clock */ + val = in_be32(reg); + val |= (SCC_EPCI_CKCTRL_OCLKEN | SCC_EPCI_CKCTRL_LCLKEN); + out_be32(reg, val); + + /* release PCI core reset (internal bus) */ + val = in_be32(reg); + val |= SCC_EPCI_CKCTRL_CRST0; + out_be32(reg, val); + + /* set PCI clock select */ + reg = epci_base + SCC_EPCI_CLKRST; + val = in_be32(reg); + val &= ~SCC_EPCI_CLKRST_CKS_MASK; + val |= SCC_EPCI_CLKRST_CKS_2; + out_be32(reg, val); + + /* set arbiter */ + reg = epci_base + SCC_EPCI_ABTSET; + out_be32(reg, 0x0f1f001f); /* temporary value */ + + /* buffer on */ + reg = epci_base + SCC_EPCI_CLKRST; + val = in_be32(reg); + val |= SCC_EPCI_CLKRST_BC; + out_be32(reg, val); + + /* PCI clock enable */ + val = in_be32(reg); + val |= SCC_EPCI_CLKRST_PCKEN; + out_be32(reg, val); + + /* release PCI core reset (all) */ + reg = epci_base + SCC_EPCI_CKCTRL; + val = in_be32(reg); + val |= (SCC_EPCI_CKCTRL_CRST0 | SCC_EPCI_CKCTRL_CRST1); + out_be32(reg, val); + + /* set base translation registers. (already set by Beat) */ + + /* set base address masks. (already set by Beat) */ + } + + /* release interrupt masks and clear all interrupts */ + reg = epci_base + SCC_EPCI_INTSET; + out_be32(reg, 0x013f011f); /* all interrupts enable */ + reg = epci_base + SCC_EPCI_VIENAB; + val = SCC_EPCI_VIENAB_PMPEE | SCC_EPCI_VIENAB_PMFEE; + out_be32(reg, val); + reg = epci_base + SCC_EPCI_STATUS; + out_be32(reg, 0xffffffff); + reg = epci_base + SCC_EPCI_VISTAT; + out_be32(reg, 0xffffffff); + + /* disable PCI->IB address translation */ + reg = epci_base + SCC_EPCI_VCSR; + val = in_be32(reg); + val &= ~(SCC_EPCI_VCSR_DR | SCC_EPCI_VCSR_AT); + out_be32(reg, val); + + /* set base addresses. (no need to set?) */ + + /* memory space, bus master enable */ + reg = epci_base + PCI_COMMAND; + val = PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER; + out_be32(reg, val); + + /* endian mode setup */ + reg = epci_base + SCC_EPCI_ECMODE; + val = 0x00550155; + out_be32(reg, val); + + /* set control option */ + reg = epci_base + SCC_EPCI_CNTOPT; + val = in_be32(reg); + val |= SCC_EPCI_CNTOPT_O2PMB; + out_be32(reg, val); + + /* XXX: temporay: set registers for address conversion setup */ + reg = epci_base + SCC_EPCI_CNF10_REG; + out_be32(reg, 0x80000008); + reg = epci_base + SCC_EPCI_CNF14_REG; + out_be32(reg, 0x40000008); + + reg = epci_base + SCC_EPCI_BAM0; + out_be32(reg, 0x80000000); + reg = epci_base + SCC_EPCI_BAM1; + out_be32(reg, 0xe0000000); + + reg = epci_base + SCC_EPCI_PVBAT; + out_be32(reg, 0x80000000); + + if (!hwres) { + /* release external PCI reset */ + reg = epci_base + SCC_EPCI_CLKRST; + val = in_be32(reg); + val |= SCC_EPCI_CLKRST_PCIRST; + out_be32(reg, val); + } + + return 0; +} + +static int __init celleb_setup_epci(struct device_node *node, + struct pci_controller *hose) +{ + struct resource r; + + pr_debug("PCI: celleb_setup_epci()\n"); + + /* + * Note: + * Celleb epci uses cfg_addr and cfg_data member of + * pci_controller structure in irregular way. + * + * cfg_addr is used to map for control registers of + * celleb epci. + * + * cfg_data is used for configuration area of devices + * on Celleb epci buses. + */ + + if (of_address_to_resource(node, 0, &r)) + goto error; + hose->cfg_addr = ioremap(r.start, resource_size(&r)); + if (!hose->cfg_addr) + goto error; + pr_debug("EPCI: cfg_addr map 0x%016llx->0x%016lx + 0x%016llx\n", + r.start, (unsigned long)hose->cfg_addr, resource_size(&r)); + + if (of_address_to_resource(node, 2, &r)) + goto error; + hose->cfg_data = ioremap(r.start, resource_size(&r)); + if (!hose->cfg_data) + goto error; + pr_debug("EPCI: cfg_data map 0x%016llx->0x%016lx + 0x%016llx\n", + r.start, (unsigned long)hose->cfg_data, resource_size(&r)); + + hose->ops = &celleb_epci_ops; + celleb_epci_init(hose); + + return 0; + +error: + if (hose->cfg_addr) + iounmap(hose->cfg_addr); + + if (hose->cfg_data) + iounmap(hose->cfg_data); + return 1; +} + +struct celleb_phb_spec celleb_epci_spec __initdata = { + .setup = celleb_setup_epci, + .ops = &spiderpci_ops, + .iowa_init = &spiderpci_iowa_init, + .iowa_data = (void *)0, +}; diff --git a/arch/powerpc/platforms/cell/celleb_scc_pciex.c b/arch/powerpc/platforms/cell/celleb_scc_pciex.c new file mode 100644 index 00000000000..4278acfa2ed --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_scc_pciex.c @@ -0,0 +1,539 @@ +/* + * Support for Celleb PCI-Express. + * + * (C) Copyright 2007-2008 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#undef DEBUG + +#include <linux/kernel.h> +#include <linux/pci.h> +#include <linux/string.h> +#include <linux/slab.h> +#include <linux/init.h> +#include <linux/bootmem.h> +#include <linux/delay.h> +#include <linux/interrupt.h> + +#include <asm/io.h> +#include <asm/irq.h> +#include <asm/iommu.h> +#include <asm/byteorder.h> + +#include "celleb_scc.h" +#include "celleb_pci.h" + +#define PEX_IN(base, off) in_be32((void __iomem *)(base) + (off)) +#define PEX_OUT(base, off, data) out_be32((void __iomem *)(base) + (off), (data)) + +static void scc_pciex_io_flush(struct iowa_bus *bus) +{ + (void)PEX_IN(bus->phb->cfg_addr, PEXDMRDEN0); +} + +/* + * Memory space access to device on PCIEX + */ +#define PCIEX_MMIO_READ(name, ret) \ +static ret scc_pciex_##name(const PCI_IO_ADDR addr) \ +{ \ + ret val = __do_##name(addr); \ + scc_pciex_io_flush(iowa_mem_find_bus(addr)); \ + return val; \ +} + +#define PCIEX_MMIO_READ_STR(name) \ +static void scc_pciex_##name(const PCI_IO_ADDR addr, void *buf, \ + unsigned long count) \ +{ \ + __do_##name(addr, buf, count); \ + scc_pciex_io_flush(iowa_mem_find_bus(addr)); \ +} + +PCIEX_MMIO_READ(readb, u8) +PCIEX_MMIO_READ(readw, u16) +PCIEX_MMIO_READ(readl, u32) +PCIEX_MMIO_READ(readq, u64) +PCIEX_MMIO_READ(readw_be, u16) +PCIEX_MMIO_READ(readl_be, u32) +PCIEX_MMIO_READ(readq_be, u64) +PCIEX_MMIO_READ_STR(readsb) +PCIEX_MMIO_READ_STR(readsw) +PCIEX_MMIO_READ_STR(readsl) + +static void scc_pciex_memcpy_fromio(void *dest, const PCI_IO_ADDR src, + unsigned long n) +{ + __do_memcpy_fromio(dest, src, n); + scc_pciex_io_flush(iowa_mem_find_bus(src)); +} + +/* + * I/O port access to devices on PCIEX. + */ + +static inline unsigned long get_bus_address(struct pci_controller *phb, + unsigned long port) +{ + return port - ((unsigned long)(phb->io_base_virt) - _IO_BASE); +} + +static u32 scc_pciex_read_port(struct pci_controller *phb, + unsigned long port, int size) +{ + unsigned int byte_enable; + unsigned int cmd, shift; + unsigned long addr; + u32 data, ret; + + BUG_ON(((port & 0x3ul) + size) > 4); + + addr = get_bus_address(phb, port); + shift = addr & 0x3ul; + byte_enable = ((1 << size) - 1) << shift; + cmd = PEXDCMND_IO_READ | (byte_enable << PEXDCMND_BYTE_EN_SHIFT); + PEX_OUT(phb->cfg_addr, PEXDADRS, (addr & ~0x3ul)); + PEX_OUT(phb->cfg_addr, PEXDCMND, cmd); + data = PEX_IN(phb->cfg_addr, PEXDRDATA); + ret = (data >> (shift * 8)) & (0xFFFFFFFF >> ((4 - size) * 8)); + + pr_debug("PCIEX:PIO READ:port=0x%lx, addr=0x%lx, size=%d, be=%x," + " cmd=%x, data=%x, ret=%x\n", port, addr, size, byte_enable, + cmd, data, ret); + + return ret; +} + +static void scc_pciex_write_port(struct pci_controller *phb, + unsigned long port, int size, u32 val) +{ + unsigned int byte_enable; + unsigned int cmd, shift; + unsigned long addr; + u32 data; + + BUG_ON(((port & 0x3ul) + size) > 4); + + addr = get_bus_address(phb, port); + shift = addr & 0x3ul; + byte_enable = ((1 << size) - 1) << shift; + cmd = PEXDCMND_IO_WRITE | (byte_enable << PEXDCMND_BYTE_EN_SHIFT); + data = (val & (0xFFFFFFFF >> (4 - size) * 8)) << (shift * 8); + PEX_OUT(phb->cfg_addr, PEXDADRS, (addr & ~0x3ul)); + PEX_OUT(phb->cfg_addr, PEXDCMND, cmd); + PEX_OUT(phb->cfg_addr, PEXDWDATA, data); + + pr_debug("PCIEX:PIO WRITE:port=0x%lx, addr=%lx, size=%d, val=%x," + " be=%x, cmd=%x, data=%x\n", port, addr, size, val, + byte_enable, cmd, data); +} + +static u8 __scc_pciex_inb(struct pci_controller *phb, unsigned long port) +{ + return (u8)scc_pciex_read_port(phb, port, 1); +} + +static u16 __scc_pciex_inw(struct pci_controller *phb, unsigned long port) +{ + u32 data; + if ((port & 0x3ul) < 3) + data = scc_pciex_read_port(phb, port, 2); + else { + u32 d1 = scc_pciex_read_port(phb, port, 1); + u32 d2 = scc_pciex_read_port(phb, port + 1, 1); + data = d1 | (d2 << 8); + } + return (u16)data; +} + +static u32 __scc_pciex_inl(struct pci_controller *phb, unsigned long port) +{ + unsigned int mod = port & 0x3ul; + u32 data; + if (mod == 0) + data = scc_pciex_read_port(phb, port, 4); + else { + u32 d1 = scc_pciex_read_port(phb, port, 4 - mod); + u32 d2 = scc_pciex_read_port(phb, port + 1, mod); + data = d1 | (d2 << (mod * 8)); + } + return data; +} + +static void __scc_pciex_outb(struct pci_controller *phb, + u8 val, unsigned long port) +{ + scc_pciex_write_port(phb, port, 1, (u32)val); +} + +static void __scc_pciex_outw(struct pci_controller *phb, + u16 val, unsigned long port) +{ + if ((port & 0x3ul) < 3) + scc_pciex_write_port(phb, port, 2, (u32)val); + else { + u32 d1 = val & 0x000000FF; + u32 d2 = (val & 0x0000FF00) >> 8; + scc_pciex_write_port(phb, port, 1, d1); + scc_pciex_write_port(phb, port + 1, 1, d2); + } +} + +static void __scc_pciex_outl(struct pci_controller *phb, + u32 val, unsigned long port) +{ + unsigned int mod = port & 0x3ul; + if (mod == 0) + scc_pciex_write_port(phb, port, 4, val); + else { + u32 d1 = val & (0xFFFFFFFFul >> (mod * 8)); + u32 d2 = val >> ((4 - mod) * 8); + scc_pciex_write_port(phb, port, 4 - mod, d1); + scc_pciex_write_port(phb, port + 1, mod, d2); + } +} + +#define PCIEX_PIO_FUNC(size, name) \ +static u##size scc_pciex_in##name(unsigned long port) \ +{ \ + struct iowa_bus *bus = iowa_pio_find_bus(port); \ + u##size data = __scc_pciex_in##name(bus->phb, port); \ + scc_pciex_io_flush(bus); \ + return data; \ +} \ +static void scc_pciex_ins##name(unsigned long p, void *b, unsigned long c) \ +{ \ + struct iowa_bus *bus = iowa_pio_find_bus(p); \ + __le##size *dst = b; \ + for (; c != 0; c--, dst++) \ + *dst = cpu_to_le##size(__scc_pciex_in##name(bus->phb, p)); \ + scc_pciex_io_flush(bus); \ +} \ +static void scc_pciex_out##name(u##size val, unsigned long port) \ +{ \ + struct iowa_bus *bus = iowa_pio_find_bus(port); \ + __scc_pciex_out##name(bus->phb, val, port); \ +} \ +static void scc_pciex_outs##name(unsigned long p, const void *b, \ + unsigned long c) \ +{ \ + struct iowa_bus *bus = iowa_pio_find_bus(p); \ + const __le##size *src = b; \ + for (; c != 0; c--, src++) \ + __scc_pciex_out##name(bus->phb, le##size##_to_cpu(*src), p); \ +} +#define __le8 u8 +#define cpu_to_le8(x) (x) +#define le8_to_cpu(x) (x) +PCIEX_PIO_FUNC(8, b) +PCIEX_PIO_FUNC(16, w) +PCIEX_PIO_FUNC(32, l) + +static struct ppc_pci_io scc_pciex_ops = { + .readb = scc_pciex_readb, + .readw = scc_pciex_readw, + .readl = scc_pciex_readl, + .readq = scc_pciex_readq, + .readw_be = scc_pciex_readw_be, + .readl_be = scc_pciex_readl_be, + .readq_be = scc_pciex_readq_be, + .readsb = scc_pciex_readsb, + .readsw = scc_pciex_readsw, + .readsl = scc_pciex_readsl, + .memcpy_fromio = scc_pciex_memcpy_fromio, + .inb = scc_pciex_inb, + .inw = scc_pciex_inw, + .inl = scc_pciex_inl, + .outb = scc_pciex_outb, + .outw = scc_pciex_outw, + .outl = scc_pciex_outl, + .insb = scc_pciex_insb, + .insw = scc_pciex_insw, + .insl = scc_pciex_insl, + .outsb = scc_pciex_outsb, + .outsw = scc_pciex_outsw, + .outsl = scc_pciex_outsl, +}; + +static int __init scc_pciex_iowa_init(struct iowa_bus *bus, void *data) +{ + dma_addr_t dummy_page_da; + void *dummy_page_va; + + dummy_page_va = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!dummy_page_va) { + pr_err("PCIEX:Alloc dummy_page_va failed\n"); + return -1; + } + + dummy_page_da = dma_map_single(bus->phb->parent, dummy_page_va, + PAGE_SIZE, DMA_FROM_DEVICE); + if (dma_mapping_error(bus->phb->parent, dummy_page_da)) { + pr_err("PCIEX:Map dummy page failed.\n"); + kfree(dummy_page_va); + return -1; + } + + PEX_OUT(bus->phb->cfg_addr, PEXDMRDADR0, dummy_page_da); + + return 0; +} + +/* + * config space access + */ +#define MK_PEXDADRS(bus_no, dev_no, func_no, addr) \ + ((uint32_t)(((addr) & ~0x3UL) | \ + ((bus_no) << PEXDADRS_BUSNO_SHIFT) | \ + ((dev_no) << PEXDADRS_DEVNO_SHIFT) | \ + ((func_no) << PEXDADRS_FUNCNO_SHIFT))) + +#define MK_PEXDCMND_BYTE_EN(addr, size) \ + ((((0x1 << (size))-1) << ((addr) & 0x3)) << PEXDCMND_BYTE_EN_SHIFT) +#define MK_PEXDCMND(cmd, addr, size) ((cmd) | MK_PEXDCMND_BYTE_EN(addr, size)) + +static uint32_t config_read_pciex_dev(unsigned int __iomem *base, + uint64_t bus_no, uint64_t dev_no, uint64_t func_no, + uint64_t off, uint64_t size) +{ + uint32_t ret; + uint32_t addr, cmd; + + addr = MK_PEXDADRS(bus_no, dev_no, func_no, off); + cmd = MK_PEXDCMND(PEXDCMND_CONFIG_READ, off, size); + PEX_OUT(base, PEXDADRS, addr); + PEX_OUT(base, PEXDCMND, cmd); + ret = (PEX_IN(base, PEXDRDATA) + >> ((off & (4-size)) * 8)) & ((0x1 << (size * 8)) - 1); + return ret; +} + +static void config_write_pciex_dev(unsigned int __iomem *base, uint64_t bus_no, + uint64_t dev_no, uint64_t func_no, uint64_t off, uint64_t size, + uint32_t data) +{ + uint32_t addr, cmd; + + addr = MK_PEXDADRS(bus_no, dev_no, func_no, off); + cmd = MK_PEXDCMND(PEXDCMND_CONFIG_WRITE, off, size); + PEX_OUT(base, PEXDADRS, addr); + PEX_OUT(base, PEXDCMND, cmd); + PEX_OUT(base, PEXDWDATA, + (data & ((0x1 << (size * 8)) - 1)) << ((off & (4-size)) * 8)); +} + +#define MK_PEXCADRS_BYTE_EN(off, len) \ + ((((0x1 << (len)) - 1) << ((off) & 0x3)) << PEXCADRS_BYTE_EN_SHIFT) +#define MK_PEXCADRS(cmd, addr, size) \ + ((cmd) | MK_PEXCADRS_BYTE_EN(addr, size) | ((addr) & ~0x3)) +static uint32_t config_read_pciex_rc(unsigned int __iomem *base, + uint32_t where, uint32_t size) +{ + PEX_OUT(base, PEXCADRS, MK_PEXCADRS(PEXCADRS_CMD_READ, where, size)); + return (PEX_IN(base, PEXCRDATA) + >> ((where & (4 - size)) * 8)) & ((0x1 << (size * 8)) - 1); +} + +static void config_write_pciex_rc(unsigned int __iomem *base, uint32_t where, + uint32_t size, uint32_t val) +{ + uint32_t data; + + data = (val & ((0x1 << (size * 8)) - 1)) << ((where & (4 - size)) * 8); + PEX_OUT(base, PEXCADRS, MK_PEXCADRS(PEXCADRS_CMD_WRITE, where, size)); + PEX_OUT(base, PEXCWDATA, data); +} + +/* Interfaces */ +/* Note: Work-around + * On SCC PCIEXC, one device is seen on all 32 dev_no. + * As SCC PCIEXC can have only one device on the bus, we look only one dev_no. + * (dev_no = 1) + */ +static int scc_pciex_read_config(struct pci_bus *bus, unsigned int devfn, + int where, int size, unsigned int *val) +{ + struct pci_controller *phb = pci_bus_to_host(bus); + + if (bus->number == phb->first_busno && PCI_SLOT(devfn) != 1) { + *val = ~0; + return PCIBIOS_DEVICE_NOT_FOUND; + } + + if (bus->number == 0 && PCI_SLOT(devfn) == 0) + *val = config_read_pciex_rc(phb->cfg_addr, where, size); + else + *val = config_read_pciex_dev(phb->cfg_addr, bus->number, + PCI_SLOT(devfn), PCI_FUNC(devfn), where, size); + + return PCIBIOS_SUCCESSFUL; +} + +static int scc_pciex_write_config(struct pci_bus *bus, unsigned int devfn, + int where, int size, unsigned int val) +{ + struct pci_controller *phb = pci_bus_to_host(bus); + + if (bus->number == phb->first_busno && PCI_SLOT(devfn) != 1) + return PCIBIOS_DEVICE_NOT_FOUND; + + if (bus->number == 0 && PCI_SLOT(devfn) == 0) + config_write_pciex_rc(phb->cfg_addr, where, size, val); + else + config_write_pciex_dev(phb->cfg_addr, bus->number, + PCI_SLOT(devfn), PCI_FUNC(devfn), where, size, val); + return PCIBIOS_SUCCESSFUL; +} + +static struct pci_ops scc_pciex_pci_ops = { + scc_pciex_read_config, + scc_pciex_write_config, +}; + +static void pciex_clear_intr_all(unsigned int __iomem *base) +{ + PEX_OUT(base, PEXAERRSTS, 0xffffffff); + PEX_OUT(base, PEXPRERRSTS, 0xffffffff); + PEX_OUT(base, PEXINTSTS, 0xffffffff); +} + +#if 0 +static void pciex_disable_intr_all(unsigned int *base) +{ + PEX_OUT(base, PEXINTMASK, 0x0); + PEX_OUT(base, PEXAERRMASK, 0x0); + PEX_OUT(base, PEXPRERRMASK, 0x0); + PEX_OUT(base, PEXVDMASK, 0x0); +} +#endif + +static void pciex_enable_intr_all(unsigned int __iomem *base) +{ + PEX_OUT(base, PEXINTMASK, 0x0000e7f1); + PEX_OUT(base, PEXAERRMASK, 0x03ff01ff); + PEX_OUT(base, PEXPRERRMASK, 0x0001010f); + PEX_OUT(base, PEXVDMASK, 0x00000001); +} + +static void pciex_check_status(unsigned int __iomem *base) +{ + uint32_t err = 0; + uint32_t intsts, aerr, prerr, rcvcp, lenerr; + uint32_t maea, maec; + + intsts = PEX_IN(base, PEXINTSTS); + aerr = PEX_IN(base, PEXAERRSTS); + prerr = PEX_IN(base, PEXPRERRSTS); + rcvcp = PEX_IN(base, PEXRCVCPLIDA); + lenerr = PEX_IN(base, PEXLENERRIDA); + + if (intsts || aerr || prerr || rcvcp || lenerr) + err = 1; + + pr_info("PCEXC interrupt!!\n"); + pr_info("PEXINTSTS :0x%08x\n", intsts); + pr_info("PEXAERRSTS :0x%08x\n", aerr); + pr_info("PEXPRERRSTS :0x%08x\n", prerr); + pr_info("PEXRCVCPLIDA :0x%08x\n", rcvcp); + pr_info("PEXLENERRIDA :0x%08x\n", lenerr); + + /* print detail of Protection Error */ + if (intsts & 0x00004000) { + uint32_t i, n; + for (i = 0; i < 4; i++) { + n = 1 << i; + if (prerr & n) { + maea = PEX_IN(base, PEXMAEA(i)); + maec = PEX_IN(base, PEXMAEC(i)); + pr_info("PEXMAEC%d :0x%08x\n", i, maec); + pr_info("PEXMAEA%d :0x%08x\n", i, maea); + } + } + } + + if (err) + pciex_clear_intr_all(base); +} + +static irqreturn_t pciex_handle_internal_irq(int irq, void *dev_id) +{ + struct pci_controller *phb = dev_id; + + pr_debug("PCIEX:pciex_handle_internal_irq(irq=%d)\n", irq); + + BUG_ON(phb->cfg_addr == NULL); + + pciex_check_status(phb->cfg_addr); + + return IRQ_HANDLED; +} + +static __init int celleb_setup_pciex(struct device_node *node, + struct pci_controller *phb) +{ + struct resource r; + int virq; + + /* SMMIO registers; used inside this file */ + if (of_address_to_resource(node, 0, &r)) { + pr_err("PCIEXC:Failed to get config resource.\n"); + return 1; + } + phb->cfg_addr = ioremap(r.start, resource_size(&r)); + if (!phb->cfg_addr) { + pr_err("PCIEXC:Failed to remap SMMIO region.\n"); + return 1; + } + + /* Not use cfg_data, cmd and data regs are near address reg */ + phb->cfg_data = NULL; + + /* set pci_ops */ + phb->ops = &scc_pciex_pci_ops; + + /* internal interrupt handler */ + virq = irq_of_parse_and_map(node, 1); + if (!virq) { + pr_err("PCIEXC:Failed to map irq\n"); + goto error; + } + if (request_irq(virq, pciex_handle_internal_irq, + 0, "pciex", (void *)phb)) { + pr_err("PCIEXC:Failed to request irq\n"); + goto error; + } + + /* enable all interrupts */ + pciex_clear_intr_all(phb->cfg_addr); + pciex_enable_intr_all(phb->cfg_addr); + /* MSI: TBD */ + + return 0; + +error: + phb->cfg_data = NULL; + if (phb->cfg_addr) + iounmap(phb->cfg_addr); + phb->cfg_addr = NULL; + return 1; +} + +struct celleb_phb_spec celleb_pciex_spec __initdata = { + .setup = celleb_setup_pciex, + .ops = &scc_pciex_ops, + .iowa_init = &scc_pciex_iowa_init, +}; diff --git a/arch/powerpc/platforms/cell/celleb_scc_sio.c b/arch/powerpc/platforms/cell/celleb_scc_sio.c new file mode 100644 index 00000000000..c8eb5719382 --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_scc_sio.c @@ -0,0 +1,99 @@ +/* + * setup serial port in SCC + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <linux/tty.h> +#include <linux/serial.h> +#include <linux/serial_core.h> +#include <linux/console.h> + +#include <asm/io.h> +#include <asm/prom.h> + +/* sio irq0=0xb00010022 irq0=0xb00010023 irq2=0xb00010024 + mmio=0xfff000-0x1000,0xff2000-0x1000 */ +static int txx9_serial_bitmap __initdata; + +static struct { + uint32_t offset; + uint32_t index; +} txx9_scc_tab[3] __initdata = { + { 0x300, 0 }, /* 0xFFF300 */ + { 0x400, 0 }, /* 0xFFF400 */ + { 0x800, 1 } /* 0xFF2800 */ +}; + +static int __init txx9_serial_init(void) +{ + extern int early_serial_txx9_setup(struct uart_port *port); + struct device_node *node; + int i; + struct uart_port req; + struct of_phandle_args irq; + struct resource res; + + for_each_compatible_node(node, "serial", "toshiba,sio-scc") { + for (i = 0; i < ARRAY_SIZE(txx9_scc_tab); i++) { + if (!(txx9_serial_bitmap & (1<<i))) + continue; + + if (of_irq_parse_one(node, i, &irq)) + continue; + if (of_address_to_resource(node, + txx9_scc_tab[i].index, &res)) + continue; + + memset(&req, 0, sizeof(req)); + req.line = i; + req.iotype = UPIO_MEM; + req.mapbase = res.start + txx9_scc_tab[i].offset; +#ifdef CONFIG_SERIAL_TXX9_CONSOLE + req.membase = ioremap(req.mapbase, 0x24); +#endif + req.irq = irq_create_of_mapping(&irq); + req.flags |= UPF_IOREMAP | UPF_BUGGY_UART + /*HAVE_CTS_LINE*/; + req.uartclk = 83300000; + early_serial_txx9_setup(&req); + } + } + + return 0; +} + +static int __init txx9_serial_config(char *ptr) +{ + int i; + + for (;;) { + switch (get_option(&ptr, &i)) { + default: + return 0; + case 2: + txx9_serial_bitmap |= 1 << i; + break; + case 1: + txx9_serial_bitmap |= 1 << i; + return 0; + } + } +} +__setup("txx9_serial=", txx9_serial_config); + +console_initcall(txx9_serial_init); diff --git a/arch/powerpc/platforms/cell/celleb_scc_uhc.c b/arch/powerpc/platforms/cell/celleb_scc_uhc.c new file mode 100644 index 00000000000..d63b720bfe3 --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_scc_uhc.c @@ -0,0 +1,95 @@ +/* + * SCC (Super Companion Chip) UHC setup + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <linux/kernel.h> +#include <linux/pci.h> + +#include <asm/delay.h> +#include <asm/io.h> +#include <asm/machdep.h> + +#include "celleb_scc.h" + +#define UHC_RESET_WAIT_MAX 10000 + +static inline int uhc_clkctrl_ready(u32 val) +{ + const u32 mask = SCC_UHC_USBCEN | SCC_UHC_USBCEN; + return((val & mask) == mask); +} + +/* + * UHC(usb host controller) enable function. + * affect to both of OHCI and EHCI core module. + */ +static void enable_scc_uhc(struct pci_dev *dev) +{ + void __iomem *uhc_base; + u32 __iomem *uhc_clkctrl; + u32 __iomem *uhc_ecmode; + u32 val = 0; + int i; + + if (!machine_is(celleb_beat) && + !machine_is(celleb_native)) + return; + + uhc_base = ioremap(pci_resource_start(dev, 0), + pci_resource_len(dev, 0)); + if (!uhc_base) { + printk(KERN_ERR "failed to map UHC register base.\n"); + return; + } + uhc_clkctrl = uhc_base + SCC_UHC_CKRCTRL; + uhc_ecmode = uhc_base + SCC_UHC_ECMODE; + + /* setup for normal mode */ + val |= SCC_UHC_F48MCKLEN; + out_be32(uhc_clkctrl, val); + val |= SCC_UHC_PHY_SUSPEND_SEL; + out_be32(uhc_clkctrl, val); + udelay(10); + val |= SCC_UHC_PHYEN; + out_be32(uhc_clkctrl, val); + udelay(50); + + /* disable reset */ + val |= SCC_UHC_HCLKEN; + out_be32(uhc_clkctrl, val); + val |= (SCC_UHC_USBCEN | SCC_UHC_USBEN); + out_be32(uhc_clkctrl, val); + i = 0; + while (!uhc_clkctrl_ready(in_be32(uhc_clkctrl))) { + udelay(10); + if (i++ > UHC_RESET_WAIT_MAX) { + printk(KERN_ERR "Failed to disable UHC reset %x\n", + in_be32(uhc_clkctrl)); + break; + } + } + + /* Endian Conversion Mode for Master ALL area */ + out_be32(uhc_ecmode, SCC_UHC_ECMODE_BY_BYTE); + + iounmap(uhc_base); +} + +DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_TOSHIBA_2, + PCI_DEVICE_ID_TOSHIBA_SCC_USB, enable_scc_uhc); diff --git a/arch/powerpc/platforms/cell/celleb_setup.c b/arch/powerpc/platforms/cell/celleb_setup.c new file mode 100644 index 00000000000..1d5a4d8ddad --- /dev/null +++ b/arch/powerpc/platforms/cell/celleb_setup.c @@ -0,0 +1,243 @@ +/* + * Celleb setup code + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This code is based on arch/powerpc/platforms/cell/setup.c: + * Copyright (C) 1995 Linus Torvalds + * Adapted from 'alpha' version by Gary Thomas + * Modified by Cort Dougan (cort@cs.nmt.edu) + * Modified by PPC64 Team, IBM Corp + * Modified by Cell Team, IBM Deutschland Entwicklung GmbH + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#undef DEBUG + +#include <linux/cpu.h> +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/export.h> +#include <linux/mm.h> +#include <linux/stddef.h> +#include <linux/unistd.h> +#include <linux/reboot.h> +#include <linux/init.h> +#include <linux/delay.h> +#include <linux/irq.h> +#include <linux/seq_file.h> +#include <linux/root_dev.h> +#include <linux/console.h> +#include <linux/of_platform.h> + +#include <asm/mmu.h> +#include <asm/processor.h> +#include <asm/io.h> +#include <asm/prom.h> +#include <asm/machdep.h> +#include <asm/cputable.h> +#include <asm/irq.h> +#include <asm/time.h> +#include <asm/spu_priv1.h> +#include <asm/firmware.h> +#include <asm/rtas.h> +#include <asm/cell-regs.h> + +#include "beat_interrupt.h" +#include "beat_wrapper.h" +#include "beat.h" +#include "celleb_pci.h" +#include "interrupt.h" +#include "pervasive.h" +#include "ras.h" + +static char celleb_machine_type[128] = "Celleb"; + +static void celleb_show_cpuinfo(struct seq_file *m) +{ + struct device_node *root; + const char *model = ""; + + root = of_find_node_by_path("/"); + if (root) + model = of_get_property(root, "model", NULL); + /* using "CHRP" is to trick anaconda into installing FCx into Celleb */ + seq_printf(m, "machine\t\t: %s %s\n", celleb_machine_type, model); + of_node_put(root); +} + +static int __init celleb_machine_type_hack(char *ptr) +{ + strlcpy(celleb_machine_type, ptr, sizeof(celleb_machine_type)); + return 0; +} + +__setup("celleb_machine_type_hack=", celleb_machine_type_hack); + +static void celleb_progress(char *s, unsigned short hex) +{ + printk("*** %04x : %s\n", hex, s ? s : ""); +} + +static void __init celleb_setup_arch_common(void) +{ + /* init to some ~sane value until calibrate_delay() runs */ + loops_per_jiffy = 50000000; + +#ifdef CONFIG_DUMMY_CONSOLE + conswitchp = &dummy_con; +#endif +} + +static struct of_device_id celleb_bus_ids[] __initdata = { + { .type = "scc", }, + { .type = "ioif", }, /* old style */ + {}, +}; + +static int __init celleb_publish_devices(void) +{ + /* Publish OF platform devices for southbridge IOs */ + of_platform_bus_probe(NULL, celleb_bus_ids, NULL); + + return 0; +} +machine_device_initcall(celleb_beat, celleb_publish_devices); +machine_device_initcall(celleb_native, celleb_publish_devices); + + +/* + * functions for Celleb-Beat + */ +static void __init celleb_setup_arch_beat(void) +{ +#ifdef CONFIG_SPU_BASE + spu_priv1_ops = &spu_priv1_beat_ops; + spu_management_ops = &spu_management_of_ops; +#endif + + celleb_setup_arch_common(); +} + +static int __init celleb_probe_beat(void) +{ + unsigned long root = of_get_flat_dt_root(); + + if (!of_flat_dt_is_compatible(root, "Beat")) + return 0; + + powerpc_firmware_features |= FW_FEATURE_CELLEB_ALWAYS + | FW_FEATURE_BEAT | FW_FEATURE_LPAR; + hpte_init_beat_v3(); + + return 1; +} + + +/* + * functions for Celleb-native + */ +static void __init celleb_init_IRQ_native(void) +{ + iic_init_IRQ(); + spider_init_IRQ(); +} + +static void __init celleb_setup_arch_native(void) +{ +#ifdef CONFIG_SPU_BASE + spu_priv1_ops = &spu_priv1_mmio_ops; + spu_management_ops = &spu_management_of_ops; +#endif + + cbe_regs_init(); + +#ifdef CONFIG_CBE_RAS + cbe_ras_init(); +#endif + +#ifdef CONFIG_SMP + smp_init_cell(); +#endif + + cbe_pervasive_init(); + + /* XXX: nvram initialization should be added */ + + celleb_setup_arch_common(); +} + +static int __init celleb_probe_native(void) +{ + unsigned long root = of_get_flat_dt_root(); + + if (of_flat_dt_is_compatible(root, "Beat") || + !of_flat_dt_is_compatible(root, "TOSHIBA,Celleb")) + return 0; + + powerpc_firmware_features |= FW_FEATURE_CELLEB_ALWAYS; + hpte_init_native(); + + return 1; +} + + +/* + * machine definitions + */ +define_machine(celleb_beat) { + .name = "Cell Reference Set (Beat)", + .probe = celleb_probe_beat, + .setup_arch = celleb_setup_arch_beat, + .show_cpuinfo = celleb_show_cpuinfo, + .restart = beat_restart, + .power_off = beat_power_off, + .halt = beat_halt, + .get_rtc_time = beat_get_rtc_time, + .set_rtc_time = beat_set_rtc_time, + .calibrate_decr = generic_calibrate_decr, + .progress = celleb_progress, + .power_save = beat_power_save, + .nvram_size = beat_nvram_get_size, + .nvram_read = beat_nvram_read, + .nvram_write = beat_nvram_write, + .set_dabr = beat_set_xdabr, + .init_IRQ = beatic_init_IRQ, + .get_irq = beatic_get_irq, + .pci_probe_mode = celleb_pci_probe_mode, + .pci_setup_phb = celleb_setup_phb, +#ifdef CONFIG_KEXEC + .kexec_cpu_down = beat_kexec_cpu_down, +#endif +}; + +define_machine(celleb_native) { + .name = "Cell Reference Set (native)", + .probe = celleb_probe_native, + .setup_arch = celleb_setup_arch_native, + .show_cpuinfo = celleb_show_cpuinfo, + .restart = rtas_restart, + .power_off = rtas_power_off, + .halt = rtas_halt, + .get_boot_time = rtas_get_boot_time, + .get_rtc_time = rtas_get_rtc_time, + .set_rtc_time = rtas_set_rtc_time, + .calibrate_decr = generic_calibrate_decr, + .progress = celleb_progress, + .pci_probe_mode = celleb_pci_probe_mode, + .pci_setup_phb = celleb_setup_phb, + .init_IRQ = celleb_init_IRQ_native, +}; diff --git a/arch/powerpc/platforms/cell/cpufreq_spudemand.c b/arch/powerpc/platforms/cell/cpufreq_spudemand.c new file mode 100644 index 00000000000..82607d621ac --- /dev/null +++ b/arch/powerpc/platforms/cell/cpufreq_spudemand.c @@ -0,0 +1,171 @@ +/* + * spu aware cpufreq governor for the cell processor + * + * © Copyright IBM Corporation 2006-2008 + * + * Author: Christian Krafft <krafft@de.ibm.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2, or (at your option) + * any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include <linux/cpufreq.h> +#include <linux/sched.h> +#include <linux/module.h> +#include <linux/timer.h> +#include <linux/workqueue.h> +#include <linux/atomic.h> +#include <asm/machdep.h> +#include <asm/spu.h> + +#define POLL_TIME 100000 /* in µs */ +#define EXP 753 /* exp(-1) in fixed-point */ + +struct spu_gov_info_struct { + unsigned long busy_spus; /* fixed-point */ + struct cpufreq_policy *policy; + struct delayed_work work; + unsigned int poll_int; /* µs */ +}; +static DEFINE_PER_CPU(struct spu_gov_info_struct, spu_gov_info); + +static int calc_freq(struct spu_gov_info_struct *info) +{ + int cpu; + int busy_spus; + + cpu = info->policy->cpu; + busy_spus = atomic_read(&cbe_spu_info[cpu_to_node(cpu)].busy_spus); + + CALC_LOAD(info->busy_spus, EXP, busy_spus * FIXED_1); + pr_debug("cpu %d: busy_spus=%d, info->busy_spus=%ld\n", + cpu, busy_spus, info->busy_spus); + + return info->policy->max * info->busy_spus / FIXED_1; +} + +static void spu_gov_work(struct work_struct *work) +{ + struct spu_gov_info_struct *info; + int delay; + unsigned long target_freq; + + info = container_of(work, struct spu_gov_info_struct, work.work); + + /* after cancel_delayed_work_sync we unset info->policy */ + BUG_ON(info->policy == NULL); + + target_freq = calc_freq(info); + __cpufreq_driver_target(info->policy, target_freq, CPUFREQ_RELATION_H); + + delay = usecs_to_jiffies(info->poll_int); + schedule_delayed_work_on(info->policy->cpu, &info->work, delay); +} + +static void spu_gov_init_work(struct spu_gov_info_struct *info) +{ + int delay = usecs_to_jiffies(info->poll_int); + INIT_DEFERRABLE_WORK(&info->work, spu_gov_work); + schedule_delayed_work_on(info->policy->cpu, &info->work, delay); +} + +static void spu_gov_cancel_work(struct spu_gov_info_struct *info) +{ + cancel_delayed_work_sync(&info->work); +} + +static int spu_gov_govern(struct cpufreq_policy *policy, unsigned int event) +{ + unsigned int cpu = policy->cpu; + struct spu_gov_info_struct *info, *affected_info; + int i; + int ret = 0; + + info = &per_cpu(spu_gov_info, cpu); + + switch (event) { + case CPUFREQ_GOV_START: + if (!cpu_online(cpu)) { + printk(KERN_ERR "cpu %d is not online\n", cpu); + ret = -EINVAL; + break; + } + + if (!policy->cur) { + printk(KERN_ERR "no cpu specified in policy\n"); + ret = -EINVAL; + break; + } + + /* initialize spu_gov_info for all affected cpus */ + for_each_cpu(i, policy->cpus) { + affected_info = &per_cpu(spu_gov_info, i); + affected_info->policy = policy; + } + + info->poll_int = POLL_TIME; + + /* setup timer */ + spu_gov_init_work(info); + + break; + + case CPUFREQ_GOV_STOP: + /* cancel timer */ + spu_gov_cancel_work(info); + + /* clean spu_gov_info for all affected cpus */ + for_each_cpu (i, policy->cpus) { + info = &per_cpu(spu_gov_info, i); + info->policy = NULL; + } + + break; + } + + return ret; +} + +static struct cpufreq_governor spu_governor = { + .name = "spudemand", + .governor = spu_gov_govern, + .owner = THIS_MODULE, +}; + +/* + * module init and destoy + */ + +static int __init spu_gov_init(void) +{ + int ret; + + ret = cpufreq_register_governor(&spu_governor); + if (ret) + printk(KERN_ERR "registration of governor failed\n"); + return ret; +} + +static void __exit spu_gov_exit(void) +{ + cpufreq_unregister_governor(&spu_governor); +} + + +module_init(spu_gov_init); +module_exit(spu_gov_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Christian Krafft <krafft@de.ibm.com>"); + diff --git a/arch/powerpc/platforms/cell/interrupt.c b/arch/powerpc/platforms/cell/interrupt.c index 04f74f9f9ab..8a106b4172e 100644 --- a/arch/powerpc/platforms/cell/interrupt.c +++ b/arch/powerpc/platforms/cell/interrupt.c @@ -31,10 +31,11 @@ #include <linux/interrupt.h> #include <linux/irq.h> -#include <linux/module.h> +#include <linux/export.h> #include <linux/percpu.h> #include <linux/types.h> #include <linux/ioport.h> +#include <linux/kernel_stat.h> #include <asm/io.h> #include <asm/pgtable.h> @@ -53,9 +54,9 @@ struct iic { struct device_node *node; }; -static DEFINE_PER_CPU(struct iic, iic); +static DEFINE_PER_CPU(struct iic, cpu_iic); #define IIC_NODE_COUNT 2 -static struct irq_host *iic_host; +static struct irq_domain *iic_host; /* Convert between "pending" bits and hw irq number */ static irq_hw_number_t iic_pending_to_hwnum(struct cbe_iic_pending_bits bits) @@ -71,36 +72,38 @@ static irq_hw_number_t iic_pending_to_hwnum(struct cbe_iic_pending_bits bits) return (node << IIC_IRQ_NODE_SHIFT) | (class << 4) | unit; } -static void iic_mask(unsigned int irq) +static void iic_mask(struct irq_data *d) { } -static void iic_unmask(unsigned int irq) +static void iic_unmask(struct irq_data *d) { } -static void iic_eoi(unsigned int irq) +static void iic_eoi(struct irq_data *d) { - struct iic *iic = &__get_cpu_var(iic); + struct iic *iic = &__get_cpu_var(cpu_iic); out_be64(&iic->regs->prio, iic->eoi_stack[--iic->eoi_ptr]); BUG_ON(iic->eoi_ptr < 0); } static struct irq_chip iic_chip = { - .typename = " CELL-IIC ", - .mask = iic_mask, - .unmask = iic_unmask, - .eoi = iic_eoi, + .name = "CELL-IIC", + .irq_mask = iic_mask, + .irq_unmask = iic_unmask, + .irq_eoi = iic_eoi, }; -static void iic_ioexc_eoi(unsigned int irq) +static void iic_ioexc_eoi(struct irq_data *d) { } static void iic_ioexc_cascade(unsigned int irq, struct irq_desc *desc) { - struct cbe_iic_regs __iomem *node_iic = (void __iomem *)desc->handler_data; + struct irq_chip *chip = irq_desc_get_chip(desc); + struct cbe_iic_regs __iomem *node_iic = + (void __iomem *)irq_desc_get_handler_data(desc); unsigned int base = (irq & 0xffffff00) | IIC_IRQ_TYPE_IOEXC; unsigned long bits, ack; int cascade; @@ -127,15 +130,15 @@ static void iic_ioexc_cascade(unsigned int irq, struct irq_desc *desc) if (ack) out_be64(&node_iic->iic_is, ack); } - desc->chip->eoi(irq); + chip->irq_eoi(&desc->irq_data); } static struct irq_chip iic_ioexc_chip = { - .typename = " CELL-IOEX", - .mask = iic_mask, - .unmask = iic_unmask, - .eoi = iic_ioexc_eoi, + .name = "CELL-IOEX", + .irq_mask = iic_mask, + .irq_unmask = iic_unmask, + .irq_eoi = iic_ioexc_eoi, }; /* Get an IRQ number from the pending state register of the IIC */ @@ -145,9 +148,9 @@ static unsigned int iic_get_irq(void) struct iic *iic; unsigned int virq; - iic = &__get_cpu_var(iic); + iic = &__get_cpu_var(cpu_iic); *(unsigned long *) &pending = - in_be64((unsigned long __iomem *) &iic->regs->pending_destr); + in_be64((u64 __iomem *) &iic->regs->pending_destr); if (!(pending.flags & CBE_IIC_IRQ_VALID)) return NO_IRQ; virq = irq_linear_revmap(iic_host, iic_pending_to_hwnum(pending)); @@ -160,12 +163,12 @@ static unsigned int iic_get_irq(void) void iic_setup_cpu(void) { - out_be64(&__get_cpu_var(iic).regs->prio, 0xff); + out_be64(&__get_cpu_var(cpu_iic).regs->prio, 0xff); } u8 iic_get_target_id(int cpu) { - return per_cpu(iic, cpu).target_id; + return per_cpu(cpu_iic, cpu).target_id; } EXPORT_SYMBOL_GPL(iic_get_target_id); @@ -173,83 +176,77 @@ EXPORT_SYMBOL_GPL(iic_get_target_id); #ifdef CONFIG_SMP /* Use the highest interrupt priorities for IPI */ -static inline int iic_ipi_to_irq(int ipi) +static inline int iic_msg_to_irq(int msg) { - return IIC_IRQ_TYPE_IPI + 0xf - ipi; + return IIC_IRQ_TYPE_IPI + 0xf - msg; } -void iic_cause_IPI(int cpu, int mesg) +void iic_message_pass(int cpu, int msg) { - out_be64(&per_cpu(iic, cpu).regs->generate, (0xf - mesg) << 4); + out_be64(&per_cpu(cpu_iic, cpu).regs->generate, (0xf - msg) << 4); } -struct irq_host *iic_get_irq_host(int node) +struct irq_domain *iic_get_irq_host(int node) { return iic_host; } EXPORT_SYMBOL_GPL(iic_get_irq_host); -static irqreturn_t iic_ipi_action(int irq, void *dev_id) -{ - int ipi = (int)(long)dev_id; - - smp_message_recv(ipi); - - return IRQ_HANDLED; -} -static void iic_request_ipi(int ipi, const char *name) +static void iic_request_ipi(int msg) { int virq; - virq = irq_create_mapping(iic_host, iic_ipi_to_irq(ipi)); + virq = irq_create_mapping(iic_host, iic_msg_to_irq(msg)); if (virq == NO_IRQ) { printk(KERN_ERR - "iic: failed to map IPI %s\n", name); + "iic: failed to map IPI %s\n", smp_ipi_name[msg]); return; } - if (request_irq(virq, iic_ipi_action, IRQF_DISABLED, name, - (void *)(long)ipi)) - printk(KERN_ERR - "iic: failed to request IPI %s\n", name); + + /* + * If smp_request_message_ipi encounters an error it will notify + * the error. If a message is not needed it will return non-zero. + */ + if (smp_request_message_ipi(virq, msg)) + irq_dispose_mapping(virq); } void iic_request_IPIs(void) { - iic_request_ipi(PPC_MSG_CALL_FUNCTION, "IPI-call"); - iic_request_ipi(PPC_MSG_RESCHEDULE, "IPI-resched"); -#ifdef CONFIG_DEBUGGER - iic_request_ipi(PPC_MSG_DEBUGGER_BREAK, "IPI-debug"); -#endif /* CONFIG_DEBUGGER */ + iic_request_ipi(PPC_MSG_CALL_FUNCTION); + iic_request_ipi(PPC_MSG_RESCHEDULE); + iic_request_ipi(PPC_MSG_TICK_BROADCAST); + iic_request_ipi(PPC_MSG_DEBUGGER_BREAK); } #endif /* CONFIG_SMP */ -static int iic_host_match(struct irq_host *h, struct device_node *node) +static int iic_host_match(struct irq_domain *h, struct device_node *node) { return of_device_is_compatible(node, "IBM,CBEA-Internal-Interrupt-Controller"); } -static int iic_host_map(struct irq_host *h, unsigned int virq, +static int iic_host_map(struct irq_domain *h, unsigned int virq, irq_hw_number_t hw) { switch (hw & IIC_IRQ_TYPE_MASK) { case IIC_IRQ_TYPE_IPI: - set_irq_chip_and_handler(virq, &iic_chip, handle_percpu_irq); + irq_set_chip_and_handler(virq, &iic_chip, handle_percpu_irq); break; case IIC_IRQ_TYPE_IOEXC: - set_irq_chip_and_handler(virq, &iic_ioexc_chip, - handle_fasteoi_irq); + irq_set_chip_and_handler(virq, &iic_ioexc_chip, + handle_edge_eoi_irq); break; default: - set_irq_chip_and_handler(virq, &iic_chip, handle_fasteoi_irq); + irq_set_chip_and_handler(virq, &iic_chip, handle_edge_eoi_irq); } return 0; } -static int iic_host_xlate(struct irq_host *h, struct device_node *ct, - u32 *intspec, unsigned int intsize, +static int iic_host_xlate(struct irq_domain *h, struct device_node *ct, + const u32 *intspec, unsigned int intsize, irq_hw_number_t *out_hwirq, unsigned int *out_flags) { @@ -288,7 +285,7 @@ static int iic_host_xlate(struct irq_host *h, struct device_node *ct, return 0; } -static struct irq_host_ops iic_host_ops = { +static const struct irq_domain_ops iic_host_ops = { .match = iic_host_match, .map = iic_host_map, .xlate = iic_host_xlate, @@ -300,7 +297,7 @@ static void __init init_one_iic(unsigned int hw_cpu, unsigned long addr, /* XXX FIXME: should locate the linux CPU number from the HW cpu * number properly. We are lucky for now */ - struct iic *iic = &per_cpu(iic, hw_cpu); + struct iic *iic = &per_cpu(cpu_iic, hw_cpu); iic->regs = ioremap(addr, sizeof(struct cbe_iic_thread_regs)); BUG_ON(iic->regs == NULL); @@ -360,8 +357,8 @@ static int __init setup_iic(void) * irq_data is a generic pointer that gets passed back * to us later, so the forced cast is fine. */ - set_irq_data(cascade, (void __force *)node_iic); - set_irq_chained_handler(cascade , iic_ioexc_cascade); + irq_set_handler_data(cascade, (void __force *)node_iic); + irq_set_chained_handler(cascade, iic_ioexc_cascade); out_be64(&node_iic->iic_ir, (1 << 12) /* priority */ | (node << 4) /* dest node */ | @@ -381,8 +378,8 @@ static int __init setup_iic(void) void __init iic_init_IRQ(void) { /* Setup an irq host data structure */ - iic_host = irq_alloc_host(NULL, IRQ_HOST_MAP_LINEAR, IIC_SOURCE_COUNT, - &iic_host_ops, IIC_IRQ_INVALID); + iic_host = irq_domain_add_linear(NULL, IIC_SOURCE_COUNT, &iic_host_ops, + NULL); BUG_ON(iic_host == NULL); irq_set_default_host(iic_host); diff --git a/arch/powerpc/platforms/cell/interrupt.h b/arch/powerpc/platforms/cell/interrupt.h index 942dc39d604..4f60ae6ca35 100644 --- a/arch/powerpc/platforms/cell/interrupt.h +++ b/arch/powerpc/platforms/cell/interrupt.h @@ -75,7 +75,7 @@ enum { }; extern void iic_init_IRQ(void); -extern void iic_cause_IPI(int cpu, int mesg); +extern void iic_message_pass(int cpu, int msg); extern void iic_request_IPIs(void); extern void iic_setup_cpu(void); diff --git a/arch/powerpc/platforms/cell/io-workarounds.c b/arch/powerpc/platforms/cell/io-workarounds.c deleted file mode 100644 index 979d4b67efb..00000000000 --- a/arch/powerpc/platforms/cell/io-workarounds.c +++ /dev/null @@ -1,343 +0,0 @@ -/* - * Copyright (C) 2006 Benjamin Herrenschmidt <benh@kernel.crashing.org> - * IBM, Corp. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License version 2 as - * published by the Free Software Foundation. - */ -#undef DEBUG - -#include <linux/kernel.h> -#include <linux/mm.h> -#include <linux/pci.h> -#include <asm/io.h> -#include <asm/machdep.h> -#include <asm/pci-bridge.h> -#include <asm/ppc-pci.h> - - -#define SPIDER_PCI_REG_BASE 0xd000 -#define SPIDER_PCI_VCI_CNTL_STAT 0x0110 -#define SPIDER_PCI_DUMMY_READ 0x0810 -#define SPIDER_PCI_DUMMY_READ_BASE 0x0814 - -/* Undefine that to re-enable bogus prefetch - * - * Without that workaround, the chip will do bogus prefetch past - * page boundary from system memory. This setting will disable that, - * though the documentation is unclear as to the consequences of doing - * so, either purely performances, or possible misbehaviour... It's not - * clear wether the chip can handle unaligned accesses at all without - * prefetching enabled. - * - * For now, things appear to be behaving properly with that prefetching - * disabled and IDE, possibly because IDE isn't doing any unaligned - * access. - */ -#define SPIDER_DISABLE_PREFETCH - -#define MAX_SPIDERS 3 - -static struct spider_pci_bus { - void __iomem *regs; - unsigned long mmio_start; - unsigned long mmio_end; - unsigned long pio_vstart; - unsigned long pio_vend; -} spider_pci_busses[MAX_SPIDERS]; -static int spider_pci_count; - -static struct spider_pci_bus *spider_pci_find(unsigned long vaddr, - unsigned long paddr) -{ - int i; - - for (i = 0; i < spider_pci_count; i++) { - struct spider_pci_bus *bus = &spider_pci_busses[i]; - if (paddr && paddr >= bus->mmio_start && paddr < bus->mmio_end) - return bus; - if (vaddr && vaddr >= bus->pio_vstart && vaddr < bus->pio_vend) - return bus; - } - return NULL; -} - -static void spider_io_flush(const volatile void __iomem *addr) -{ - struct spider_pci_bus *bus; - int token; - - /* Get platform token (set by ioremap) from address */ - token = PCI_GET_ADDR_TOKEN(addr); - - /* Fast path if we have a non-0 token, it indicates which bus we - * are on. - * - * If the token is 0, that means either that the ioremap was done - * before we initialized this layer, or it's a PIO operation. We - * fallback to a low path in this case. Hopefully, internal devices - * which are ioremap'ed early should use in_XX/out_XX functions - * instead of the PCI ones and thus not suffer from the slowdown. - * - * Also note that currently, the workaround will not work for areas - * that are not mapped with PTEs (bolted in the hash table). This - * is the case for ioremaps done very early at boot (before - * mem_init_done) and includes the mapping of the ISA IO space. - * - * Fortunately, none of the affected devices is expected to do DMA - * and thus there should be no problem in practice. - * - * In order to improve performances, we only do the PTE search for - * addresses falling in the PHB IO space area. That means it will - * not work for hotplug'ed PHBs but those don't exist with Spider. - */ - if (token && token <= spider_pci_count) - bus = &spider_pci_busses[token - 1]; - else { - unsigned long vaddr, paddr; - pte_t *ptep; - - /* Fixup physical address */ - vaddr = (unsigned long)PCI_FIX_ADDR(addr); - - /* Check if it's in allowed range for PIO */ - if (vaddr < PHB_IO_BASE || vaddr > PHB_IO_END) - return; - - /* Try to find a PTE. If not, clear the paddr, we'll do - * a vaddr only lookup (PIO only) - */ - ptep = find_linux_pte(init_mm.pgd, vaddr); - if (ptep == NULL) - paddr = 0; - else - paddr = pte_pfn(*ptep) << PAGE_SHIFT; - - bus = spider_pci_find(vaddr, paddr); - if (bus == NULL) - return; - } - - /* Now do the workaround - */ - (void)in_be32(bus->regs + SPIDER_PCI_DUMMY_READ); -} - -static u8 spider_readb(const volatile void __iomem *addr) -{ - u8 val = __do_readb(addr); - spider_io_flush(addr); - return val; -} - -static u16 spider_readw(const volatile void __iomem *addr) -{ - u16 val = __do_readw(addr); - spider_io_flush(addr); - return val; -} - -static u32 spider_readl(const volatile void __iomem *addr) -{ - u32 val = __do_readl(addr); - spider_io_flush(addr); - return val; -} - -static u64 spider_readq(const volatile void __iomem *addr) -{ - u64 val = __do_readq(addr); - spider_io_flush(addr); - return val; -} - -static u16 spider_readw_be(const volatile void __iomem *addr) -{ - u16 val = __do_readw_be(addr); - spider_io_flush(addr); - return val; -} - -static u32 spider_readl_be(const volatile void __iomem *addr) -{ - u32 val = __do_readl_be(addr); - spider_io_flush(addr); - return val; -} - -static u64 spider_readq_be(const volatile void __iomem *addr) -{ - u64 val = __do_readq_be(addr); - spider_io_flush(addr); - return val; -} - -static void spider_readsb(const volatile void __iomem *addr, void *buf, - unsigned long count) -{ - __do_readsb(addr, buf, count); - spider_io_flush(addr); -} - -static void spider_readsw(const volatile void __iomem *addr, void *buf, - unsigned long count) -{ - __do_readsw(addr, buf, count); - spider_io_flush(addr); -} - -static void spider_readsl(const volatile void __iomem *addr, void *buf, - unsigned long count) -{ - __do_readsl(addr, buf, count); - spider_io_flush(addr); -} - -static void spider_memcpy_fromio(void *dest, const volatile void __iomem *src, - unsigned long n) -{ - __do_memcpy_fromio(dest, src, n); - spider_io_flush(src); -} - - -static void __iomem * spider_ioremap(unsigned long addr, unsigned long size, - unsigned long flags) -{ - struct spider_pci_bus *bus; - void __iomem *res = __ioremap(addr, size, flags); - int busno; - - pr_debug("spider_ioremap(0x%lx, 0x%lx, 0x%lx) -> 0x%p\n", - addr, size, flags, res); - - bus = spider_pci_find(0, addr); - if (bus != NULL) { - busno = bus - spider_pci_busses; - pr_debug(" found bus %d, setting token\n", busno); - PCI_SET_ADDR_TOKEN(res, busno + 1); - } - pr_debug(" result=0x%p\n", res); - - return res; -} - -static void __init spider_pci_setup_chip(struct spider_pci_bus *bus) -{ -#ifdef SPIDER_DISABLE_PREFETCH - u32 val = in_be32(bus->regs + SPIDER_PCI_VCI_CNTL_STAT); - pr_debug(" PVCI_Control_Status was 0x%08x\n", val); - out_be32(bus->regs + SPIDER_PCI_VCI_CNTL_STAT, val | 0x8); -#endif - - /* Configure the dummy address for the workaround */ - out_be32(bus->regs + SPIDER_PCI_DUMMY_READ_BASE, 0x80000000); -} - -static void __init spider_pci_add_one(struct pci_controller *phb) -{ - struct spider_pci_bus *bus = &spider_pci_busses[spider_pci_count]; - struct device_node *np = phb->dn; - struct resource rsrc; - void __iomem *regs; - - if (spider_pci_count >= MAX_SPIDERS) { - printk(KERN_ERR "Too many spider bridges, workarounds" - " disabled for %s\n", np->full_name); - return; - } - - /* Get the registers for the beast */ - if (of_address_to_resource(np, 0, &rsrc)) { - printk(KERN_ERR "Failed to get registers for spider %s" - " workarounds disabled\n", np->full_name); - return; - } - - /* Mask out some useless bits in there to get to the base of the - * spider chip - */ - rsrc.start &= ~0xfffffffful; - - /* Map them */ - regs = ioremap(rsrc.start + SPIDER_PCI_REG_BASE, 0x1000); - if (regs == NULL) { - printk(KERN_ERR "Failed to map registers for spider %s" - " workarounds disabled\n", np->full_name); - return; - } - - spider_pci_count++; - - /* We assume spiders only have one MMIO resource */ - bus->mmio_start = phb->mem_resources[0].start; - bus->mmio_end = phb->mem_resources[0].end + 1; - - bus->pio_vstart = (unsigned long)phb->io_base_virt; - bus->pio_vend = bus->pio_vstart + phb->pci_io_size; - - bus->regs = regs; - - printk(KERN_INFO "PCI: Spider MMIO workaround for %s\n",np->full_name); - - pr_debug(" mmio (P) = 0x%016lx..0x%016lx\n", - bus->mmio_start, bus->mmio_end); - pr_debug(" pio (V) = 0x%016lx..0x%016lx\n", - bus->pio_vstart, bus->pio_vend); - pr_debug(" regs (P) = 0x%016lx (V) = 0x%p\n", - rsrc.start + SPIDER_PCI_REG_BASE, bus->regs); - - spider_pci_setup_chip(bus); -} - -static struct ppc_pci_io __initdata spider_pci_io = { - .readb = spider_readb, - .readw = spider_readw, - .readl = spider_readl, - .readq = spider_readq, - .readw_be = spider_readw_be, - .readl_be = spider_readl_be, - .readq_be = spider_readq_be, - .readsb = spider_readsb, - .readsw = spider_readsw, - .readsl = spider_readsl, - .memcpy_fromio = spider_memcpy_fromio, -}; - -static int __init spider_pci_workaround_init(void) -{ - struct pci_controller *phb; - - /* Find spider bridges. We assume they have been all probed - * in setup_arch(). If that was to change, we would need to - * update this code to cope with dynamically added busses - */ - list_for_each_entry(phb, &hose_list, list_node) { - struct device_node *np = phb->dn; - const char *model = of_get_property(np, "model", NULL); - - /* If no model property or name isn't exactly "pci", skip */ - if (model == NULL || strcmp(np->name, "pci")) - continue; - /* If model is not "Spider", skip */ - if (strcmp(model, "Spider")) - continue; - spider_pci_add_one(phb); - } - - /* No Spider PCI found, exit */ - if (spider_pci_count == 0) - return 0; - - /* Setup IO callbacks. We only setup MMIO reads. PIO reads will - * fallback to MMIO reads (though without a token, thus slower) - */ - ppc_pci_io = spider_pci_io; - - /* Setup ioremap callback */ - ppc_md.ioremap = spider_ioremap; - - return 0; -} -machine_arch_initcall(cell, spider_pci_workaround_init); diff --git a/arch/powerpc/platforms/cell/iommu.c b/arch/powerpc/platforms/cell/iommu.c index d75ccded7f1..2b90ff8a93b 100644 --- a/arch/powerpc/platforms/cell/iommu.c +++ b/arch/powerpc/platforms/cell/iommu.c @@ -28,13 +28,14 @@ #include <linux/notifier.h> #include <linux/of.h> #include <linux/of_platform.h> +#include <linux/slab.h> +#include <linux/memblock.h> #include <asm/prom.h> #include <asm/iommu.h> #include <asm/machdep.h> #include <asm/pci-bridge.h> #include <asm/udbg.h> -#include <asm/lmb.h> #include <asm/firmware.h> #include <asm/cell-regs.h> @@ -74,7 +75,7 @@ #define IOC_IO_ExcpStat_V 0x8000000000000000ul #define IOC_IO_ExcpStat_SPF_Mask 0x6000000000000000ul #define IOC_IO_ExcpStat_SPF_S 0x6000000000000000ul -#define IOC_IO_ExcpStat_SPF_P 0x4000000000000000ul +#define IOC_IO_ExcpStat_SPF_P 0x2000000000000000ul #define IOC_IO_ExcpStat_ADDR_Mask 0x00000007fffff000ul #define IOC_IO_ExcpStat_RW_Mask 0x0000000000000800ul #define IOC_IO_ExcpStat_IOID_Mask 0x00000000000007fful @@ -100,16 +101,6 @@ #define IOSTE_PS_1M 0x0000000000000005ul /* - 1MB */ #define IOSTE_PS_16M 0x0000000000000007ul /* - 16MB */ -/* Page table entries */ -#define IOPTE_PP_W 0x8000000000000000ul /* protection: write */ -#define IOPTE_PP_R 0x4000000000000000ul /* protection: read */ -#define IOPTE_M 0x2000000000000000ul /* coherency required */ -#define IOPTE_SO_R 0x1000000000000000ul /* ordering: writes */ -#define IOPTE_SO_RW 0x1800000000000000ul /* ordering: r & w */ -#define IOPTE_RPN_Mask 0x07fffffffffff000ul /* RPN */ -#define IOPTE_H 0x0000000000000800ul /* cache hint */ -#define IOPTE_IOID_Mask 0x00000000000007fful /* ioid */ - /* IOMMU sizing */ #define IO_SEGMENT_SHIFT 28 @@ -150,8 +141,8 @@ static int cbe_nr_iommus; static void invalidate_tce_cache(struct cbe_iommu *iommu, unsigned long *pte, long n_ptes) { - unsigned long __iomem *reg; - unsigned long val; + u64 __iomem *reg; + u64 val; long n; reg = iommu->xlate_regs + IOC_IOPT_CacheInvd; @@ -172,8 +163,9 @@ static void invalidate_tce_cache(struct cbe_iommu *iommu, unsigned long *pte, } } -static void tce_build_cell(struct iommu_table *tbl, long index, long npages, - unsigned long uaddr, enum dma_data_direction direction) +static int tce_build_cell(struct iommu_table *tbl, long index, long npages, + unsigned long uaddr, enum dma_data_direction direction, + struct dma_attrs *attrs) { int i; unsigned long *io_pte, base_pte; @@ -192,17 +184,21 @@ static void tce_build_cell(struct iommu_table *tbl, long index, long npages, */ const unsigned long prot = 0xc48; base_pte = - ((prot << (52 + 4 * direction)) & (IOPTE_PP_W | IOPTE_PP_R)) - | IOPTE_M | IOPTE_SO_RW | (window->ioid & IOPTE_IOID_Mask); + ((prot << (52 + 4 * direction)) & + (CBE_IOPTE_PP_W | CBE_IOPTE_PP_R)) | + CBE_IOPTE_M | CBE_IOPTE_SO_RW | + (window->ioid & CBE_IOPTE_IOID_Mask); #else - base_pte = IOPTE_PP_W | IOPTE_PP_R | IOPTE_M | IOPTE_SO_RW | - (window->ioid & IOPTE_IOID_Mask); + base_pte = CBE_IOPTE_PP_W | CBE_IOPTE_PP_R | CBE_IOPTE_M | + CBE_IOPTE_SO_RW | (window->ioid & CBE_IOPTE_IOID_Mask); #endif + if (unlikely(dma_get_attr(DMA_ATTR_WEAK_ORDERING, attrs))) + base_pte &= ~CBE_IOPTE_SO_RW; io_pte = (unsigned long *)tbl->it_base + (index - tbl->it_offset); - for (i = 0; i < npages; i++, uaddr += IOMMU_PAGE_SIZE) - io_pte[i] = base_pte | (__pa(uaddr) & IOPTE_RPN_Mask); + for (i = 0; i < npages; i++, uaddr += tbl->it_page_shift) + io_pte[i] = base_pte | (__pa(uaddr) & CBE_IOPTE_RPN_Mask); mb(); @@ -210,6 +206,7 @@ static void tce_build_cell(struct iommu_table *tbl, long index, long npages, pr_debug("tce_build_cell(index=%lx,n=%lx,dir=%d,base_pte=%lx)\n", index, npages, direction, base_pte); + return 0; } static void tce_free_cell(struct iommu_table *tbl, long index, long npages) @@ -227,8 +224,9 @@ static void tce_free_cell(struct iommu_table *tbl, long index, long npages) #else /* spider bridge does PCI reads after freeing - insert a mapping * to a scratch page instead of an invalid entry */ - pte = IOPTE_PP_R | IOPTE_M | IOPTE_SO_RW | __pa(window->iommu->pad_page) - | (window->ioid & IOPTE_IOID_Mask); + pte = CBE_IOPTE_PP_R | CBE_IOPTE_M | CBE_IOPTE_SO_RW | + __pa(window->iommu->pad_page) | + (window->ioid & CBE_IOPTE_IOID_Mask); #endif io_pte = (unsigned long *)tbl->it_base + (index - tbl->it_offset); @@ -243,17 +241,18 @@ static void tce_free_cell(struct iommu_table *tbl, long index, long npages) static irqreturn_t ioc_interrupt(int irq, void *data) { - unsigned long stat; + unsigned long stat, spf; struct cbe_iommu *iommu = data; stat = in_be64(iommu->xlate_regs + IOC_IO_ExcpStat); + spf = stat & IOC_IO_ExcpStat_SPF_Mask; /* Might want to rate limit it */ printk(KERN_ERR "iommu: DMA exception 0x%016lx\n", stat); printk(KERN_ERR " V=%d, SPF=[%c%c], RW=%s, IOID=0x%04x\n", !!(stat & IOC_IO_ExcpStat_V), - (stat & IOC_IO_ExcpStat_SPF_S) ? 'S' : ' ', - (stat & IOC_IO_ExcpStat_SPF_P) ? 'P' : ' ', + (spf == IOC_IO_ExcpStat_SPF_S) ? 'S' : ' ', + (spf == IOC_IO_ExcpStat_SPF_P) ? 'P' : ' ', (stat & IOC_IO_ExcpStat_RW_Mask) ? "Read" : "Write", (unsigned int)(stat & IOC_IO_ExcpStat_IOID_Mask)); printk(KERN_ERR " page=0x%016lx\n", @@ -316,7 +315,7 @@ static void cell_iommu_setup_stab(struct cbe_iommu *iommu, segments = max(dbase + dsize, fbase + fsize) >> IO_SEGMENT_SHIFT; pr_debug("%s: iommu[%d]: segments: %lu\n", - __FUNCTION__, iommu->nid, segments); + __func__, iommu->nid, segments); /* set up the segment table */ stab_size = segments * sizeof(unsigned long); @@ -343,7 +342,7 @@ static unsigned long *cell_iommu_alloc_ptab(struct cbe_iommu *iommu, (1 << 12) / sizeof(unsigned long)); ptab_size = segments * pages_per_segment * sizeof(unsigned long); - pr_debug("%s: iommu[%d]: ptab_size: %lu, order: %d\n", __FUNCTION__, + pr_debug("%s: iommu[%d]: ptab_size: %lu, order: %d\n", __func__, iommu->nid, ptab_size, get_order(ptab_size)); page = alloc_pages_node(iommu->nid, GFP_KERNEL, get_order(ptab_size)); BUG_ON(!page); @@ -355,7 +354,7 @@ static unsigned long *cell_iommu_alloc_ptab(struct cbe_iommu *iommu, n_pte_pages = (pages_per_segment * sizeof(unsigned long)) >> 12; pr_debug("%s: iommu[%d]: stab at %p, ptab at %p, n_pte_pages: %lu\n", - __FUNCTION__, iommu->nid, iommu->stab, ptab, + __func__, iommu->nid, iommu->stab, ptab, n_pte_pages); /* initialise the STEs */ @@ -394,7 +393,7 @@ static void cell_iommu_enable_hardware(struct cbe_iommu *iommu) if (cell_iommu_find_ioc(iommu->nid, &xlate_base)) panic("%s: missing IOC register mappings for node %d\n", - __FUNCTION__, iommu->nid); + __func__, iommu->nid); iommu->xlate_regs = ioremap(xlate_base, IOC_Reg_Size); iommu->cmd_regs = iommu->xlate_regs + IOC_IOCmd_Offset; @@ -413,8 +412,7 @@ static void cell_iommu_enable_hardware(struct cbe_iommu *iommu) IIC_IRQ_IOEX_ATI | (iommu->nid << IIC_IRQ_NODE_SHIFT)); BUG_ON(virq == NO_IRQ); - ret = request_irq(virq, ioc_interrupt, IRQF_DISABLED, - iommu->name, iommu); + ret = request_irq(virq, ioc_interrupt, 0, iommu->name, iommu); BUG_ON(ret); /* set the IOC segment table origin register (and turn on the iommu) */ @@ -432,7 +430,7 @@ static void cell_iommu_setup_hardware(struct cbe_iommu *iommu, { cell_iommu_setup_stab(iommu, base, size, 0, 0); iommu->ptab = cell_iommu_alloc_ptab(iommu, base, size, 0, 0, - IOMMU_PAGE_SHIFT); + IOMMU_PAGE_SHIFT_4K); cell_iommu_enable_hardware(iommu); } @@ -478,7 +476,7 @@ cell_iommu_setup_window(struct cbe_iommu *iommu, struct device_node *np, ioid = cell_iommu_get_ioid(np); - window = kmalloc_node(sizeof(*window), GFP_KERNEL, iommu->nid); + window = kzalloc_node(sizeof(*window), GFP_KERNEL, iommu->nid); BUG_ON(window == NULL); window->offset = offset; @@ -489,8 +487,10 @@ cell_iommu_setup_window(struct cbe_iommu *iommu, struct device_node *np, window->table.it_blocksize = 16; window->table.it_base = (unsigned long)iommu->ptab; window->table.it_index = iommu->nid; - window->table.it_offset = (offset >> IOMMU_PAGE_SHIFT) + pte_offset; - window->table.it_size = size >> IOMMU_PAGE_SHIFT; + window->table.it_page_shift = IOMMU_PAGE_SHIFT_4K; + window->table.it_offset = + (offset >> window->table.it_page_shift) + pte_offset; + window->table.it_size = size >> window->table.it_page_shift; iommu_init_table(&window->table, iommu->nid); @@ -519,8 +519,7 @@ cell_iommu_setup_window(struct cbe_iommu *iommu, struct device_node *np, __set_bit(0, window->table.it_map); tce_build_cell(&window->table, window->table.it_offset, 1, - (unsigned long)iommu->pad_page, DMA_TO_DEVICE); - window->table.it_hint = window->table.it_blocksize; + (unsigned long)iommu->pad_page, DMA_TO_DEVICE, NULL); return window; } @@ -538,43 +537,134 @@ static struct cbe_iommu *cell_iommu_for_node(int nid) static unsigned long cell_dma_direct_offset; static unsigned long dma_iommu_fixed_base; -struct dma_mapping_ops dma_iommu_fixed_ops; -static void cell_dma_dev_setup_iommu(struct device *dev) +/* iommu_fixed_is_weak is set if booted with iommu_fixed=weak */ +static int iommu_fixed_is_weak; + +static struct iommu_table *cell_get_iommu_table(struct device *dev) { struct iommu_window *window; struct cbe_iommu *iommu; - struct dev_archdata *archdata = &dev->archdata; /* Current implementation uses the first window available in that * node's iommu. We -might- do something smarter later though it may * never be necessary */ - iommu = cell_iommu_for_node(archdata->numa_node); + iommu = cell_iommu_for_node(dev_to_node(dev)); if (iommu == NULL || list_empty(&iommu->windows)) { - printk(KERN_ERR "iommu: missing iommu for %s (node %d)\n", - archdata->of_node ? archdata->of_node->full_name : "?", - archdata->numa_node); - return; + dev_err(dev, "iommu: missing iommu for %s (node %d)\n", + of_node_full_name(dev->of_node), dev_to_node(dev)); + return NULL; } window = list_entry(iommu->windows.next, struct iommu_window, list); - archdata->dma_data = &window->table; + return &window->table; +} + +/* A coherent allocation implies strong ordering */ + +static void *dma_fixed_alloc_coherent(struct device *dev, size_t size, + dma_addr_t *dma_handle, gfp_t flag, + struct dma_attrs *attrs) +{ + if (iommu_fixed_is_weak) + return iommu_alloc_coherent(dev, cell_get_iommu_table(dev), + size, dma_handle, + device_to_mask(dev), flag, + dev_to_node(dev)); + else + return dma_direct_ops.alloc(dev, size, dma_handle, flag, + attrs); +} + +static void dma_fixed_free_coherent(struct device *dev, size_t size, + void *vaddr, dma_addr_t dma_handle, + struct dma_attrs *attrs) +{ + if (iommu_fixed_is_weak) + iommu_free_coherent(cell_get_iommu_table(dev), size, vaddr, + dma_handle); + else + dma_direct_ops.free(dev, size, vaddr, dma_handle, attrs); +} + +static dma_addr_t dma_fixed_map_page(struct device *dev, struct page *page, + unsigned long offset, size_t size, + enum dma_data_direction direction, + struct dma_attrs *attrs) +{ + if (iommu_fixed_is_weak == dma_get_attr(DMA_ATTR_WEAK_ORDERING, attrs)) + return dma_direct_ops.map_page(dev, page, offset, size, + direction, attrs); + else + return iommu_map_page(dev, cell_get_iommu_table(dev), page, + offset, size, device_to_mask(dev), + direction, attrs); +} + +static void dma_fixed_unmap_page(struct device *dev, dma_addr_t dma_addr, + size_t size, enum dma_data_direction direction, + struct dma_attrs *attrs) +{ + if (iommu_fixed_is_weak == dma_get_attr(DMA_ATTR_WEAK_ORDERING, attrs)) + dma_direct_ops.unmap_page(dev, dma_addr, size, direction, + attrs); + else + iommu_unmap_page(cell_get_iommu_table(dev), dma_addr, size, + direction, attrs); +} + +static int dma_fixed_map_sg(struct device *dev, struct scatterlist *sg, + int nents, enum dma_data_direction direction, + struct dma_attrs *attrs) +{ + if (iommu_fixed_is_weak == dma_get_attr(DMA_ATTR_WEAK_ORDERING, attrs)) + return dma_direct_ops.map_sg(dev, sg, nents, direction, attrs); + else + return iommu_map_sg(dev, cell_get_iommu_table(dev), sg, nents, + device_to_mask(dev), direction, attrs); +} + +static void dma_fixed_unmap_sg(struct device *dev, struct scatterlist *sg, + int nents, enum dma_data_direction direction, + struct dma_attrs *attrs) +{ + if (iommu_fixed_is_weak == dma_get_attr(DMA_ATTR_WEAK_ORDERING, attrs)) + dma_direct_ops.unmap_sg(dev, sg, nents, direction, attrs); + else + iommu_unmap_sg(cell_get_iommu_table(dev), sg, nents, direction, + attrs); +} + +static int dma_fixed_dma_supported(struct device *dev, u64 mask) +{ + return mask == DMA_BIT_MASK(64); } +static int dma_set_mask_and_switch(struct device *dev, u64 dma_mask); + +struct dma_map_ops dma_iommu_fixed_ops = { + .alloc = dma_fixed_alloc_coherent, + .free = dma_fixed_free_coherent, + .map_sg = dma_fixed_map_sg, + .unmap_sg = dma_fixed_unmap_sg, + .dma_supported = dma_fixed_dma_supported, + .set_dma_mask = dma_set_mask_and_switch, + .map_page = dma_fixed_map_page, + .unmap_page = dma_fixed_unmap_page, +}; + static void cell_dma_dev_setup_fixed(struct device *dev); static void cell_dma_dev_setup(struct device *dev) { - struct dev_archdata *archdata = &dev->archdata; - /* Order is important here, these are not mutually exclusive */ if (get_dma_ops(dev) == &dma_iommu_fixed_ops) cell_dma_dev_setup_fixed(dev); else if (get_pci_dma_ops() == &dma_iommu_ops) - cell_dma_dev_setup_iommu(dev); + set_iommu_table_base(dev, cell_get_iommu_table(dev)); else if (get_pci_dma_ops() == &dma_direct_ops) - archdata->dma_data = (void *)cell_dma_direct_offset; + set_dma_offset(dev, cell_dma_direct_offset); else BUG(); } @@ -609,7 +699,7 @@ static int __init cell_iommu_get_window(struct device_node *np, unsigned long *base, unsigned long *size) { - const void *dma_window; + const __be32 *dma_window; unsigned long index; /* Use ibm,dma-window if available, else, hard code ! */ @@ -640,7 +730,7 @@ static struct cbe_iommu * __init cell_iommu_alloc(struct device_node *np) nid, np->full_name); /* XXX todo: If we can have multiple windows on the same IOMMU, which - * isn't the case today, we probably want here to check wether the + * isn't the case today, we probably want here to check whether the * iommu for that node is already setup. * However, there might be issue with getting the size right so let's * ignore that for now. We might want to completely get rid of the @@ -685,7 +775,7 @@ static void __init cell_iommu_init_one(struct device_node *np, /* Setup the iommu_table */ cell_iommu_setup_window(iommu, np, base, size, - offset >> IOMMU_PAGE_SHIFT); + offset >> IOMMU_PAGE_SHIFT_4K); } static void __init cell_disable_iommus(void) @@ -756,10 +846,10 @@ static int __init cell_iommu_init_disabled(void) /* If we found a DMA window, we check if it's big enough to enclose * all of physical memory. If not, we force enable IOMMU */ - if (np && size < lmb_end_of_DRAM()) { + if (np && size < memblock_end_of_DRAM()) { printk(KERN_WARNING "iommu: force-enabled, dma window" - " (%ldMB) smaller than total memory (%ldMB)\n", - size >> 20, lmb_end_of_DRAM() >> 20); + " (%ldMB) smaller than total memory (%lldMB)\n", + size >> 20, memblock_end_of_DRAM() >> 20); return -ENODEV; } @@ -807,7 +897,7 @@ static u64 cell_iommu_get_fixed_address(struct device *dev) const u32 *ranges = NULL; int i, len, best, naddr, nsize, pna, range_size; - np = of_node_get(dev->archdata.of_node); + np = of_node_get(dev->of_node); while (1) { naddr = of_n_addr_cells(np); nsize = of_n_size_cells(np); @@ -882,13 +972,12 @@ static int dma_set_mask_and_switch(struct device *dev, u64 dma_mask) static void cell_dma_dev_setup_fixed(struct device *dev) { - struct dev_archdata *archdata = &dev->archdata; u64 addr; addr = cell_iommu_get_fixed_address(dev) + dma_iommu_fixed_base; - archdata->dma_data = (void *)addr; + set_dma_offset(dev, addr); - dev_dbg(dev, "iommu: fixed addr = %lx\n", addr); + dev_dbg(dev, "iommu: fixed addr = %llx\n", addr); } static void insert_16M_pte(unsigned long addr, unsigned long *ptab, @@ -903,7 +992,7 @@ static void insert_16M_pte(unsigned long addr, unsigned long *ptab, pr_debug("iommu: addr %lx ptab %p segment %lx offset %lx\n", addr, ptab, segment, offset); - ptab[offset] = base_pte | (__pa(addr) & IOPTE_RPN_Mask); + ptab[offset] = base_pte | (__pa(addr) & CBE_IOPTE_RPN_Mask); } static void cell_iommu_setup_fixed_ptab(struct cbe_iommu *iommu, @@ -918,8 +1007,15 @@ static void cell_iommu_setup_fixed_ptab(struct cbe_iommu *iommu, pr_debug("iommu: mapping 0x%lx pages from 0x%lx\n", fsize, fbase); - base_pte = IOPTE_PP_W | IOPTE_PP_R | IOPTE_M | IOPTE_SO_RW - | (cell_iommu_get_ioid(np) & IOPTE_IOID_Mask); + base_pte = CBE_IOPTE_PP_W | CBE_IOPTE_PP_R | CBE_IOPTE_M | + (cell_iommu_get_ioid(np) & CBE_IOPTE_IOID_Mask); + + if (iommu_fixed_is_weak) + pr_info("IOMMU: Using weak ordering for fixed mapping\n"); + else { + pr_info("IOMMU: Using strong ordering for fixed mapping\n"); + base_pte |= CBE_IOPTE_SO_RW; + } for (uaddr = 0; uaddr < fsize; uaddr += (1 << 24)) { /* Don't touch the dynamic region */ @@ -943,16 +1039,15 @@ static int __init cell_iommu_fixed_mapping_init(void) /* The fixed mapping is only supported on axon machines */ np = of_find_node_by_name(NULL, "axon"); + of_node_put(np); + if (!np) { pr_debug("iommu: fixed mapping disabled, no axons found\n"); return -1; } /* We must have dma-ranges properties for fixed mapping to work */ - for (np = NULL; (np = of_find_all_nodes(np));) { - if (of_find_property(np, "dma-ranges", NULL)) - break; - } + np = of_find_node_with_property(NULL, "dma-ranges"); of_node_put(np); if (!np) { @@ -972,9 +1067,9 @@ static int __init cell_iommu_fixed_mapping_init(void) } fbase = _ALIGN_UP(fbase, 1 << IO_SEGMENT_SHIFT); - fsize = lmb_phys_mem_size(); + fsize = memblock_phys_mem_size(); - if ((fbase + fsize) <= 0x800000000) + if ((fbase + fsize) <= 0x800000000ul) hbase = 0; /* use the device tree window */ else { /* If we're over 32 GB we need to cheat. We can't map all of @@ -1029,16 +1124,13 @@ static int __init cell_iommu_fixed_mapping_init(void) cell_iommu_setup_stab(iommu, dbase, dsize, fbase, fsize); iommu->ptab = cell_iommu_alloc_ptab(iommu, dbase, dsize, 0, 0, - IOMMU_PAGE_SHIFT); + IOMMU_PAGE_SHIFT_4K); cell_iommu_setup_fixed_ptab(iommu, np, dbase, dsize, fbase, fsize); cell_iommu_enable_hardware(iommu); cell_iommu_setup_window(iommu, np, dbase, dsize, 0); } - dma_iommu_fixed_ops = dma_direct_ops; - dma_iommu_fixed_ops.set_dma_mask = dma_set_mask_and_switch; - dma_iommu_ops.set_dma_mask = dma_set_mask_and_switch; set_pci_dma_ops(&dma_iommu_ops); @@ -1049,13 +1141,47 @@ static int iommu_fixed_disabled; static int __init setup_iommu_fixed(char *str) { + struct device_node *pciep; + if (strcmp(str, "off") == 0) iommu_fixed_disabled = 1; + /* If we can find a pcie-endpoint in the device tree assume that + * we're on a triblade or a CAB so by default the fixed mapping + * should be set to be weakly ordered; but only if the boot + * option WASN'T set for strong ordering + */ + pciep = of_find_node_by_type(NULL, "pcie-endpoint"); + + if (strcmp(str, "weak") == 0 || (pciep && strcmp(str, "strong") != 0)) + iommu_fixed_is_weak = 1; + + of_node_put(pciep); + return 1; } __setup("iommu_fixed=", setup_iommu_fixed); +static u64 cell_dma_get_required_mask(struct device *dev) +{ + struct dma_map_ops *dma_ops; + + if (!dev->dma_mask) + return 0; + + if (!iommu_fixed_disabled && + cell_iommu_get_fixed_address(dev) != OF_BAD_ADDR) + return DMA_BIT_MASK(64); + + dma_ops = get_dma_ops(dev); + if (dma_ops->get_required_mask) + return dma_ops->get_required_mask(dev); + + WARN_ONCE(1, "no get_required_mask in %p ops", dma_ops); + + return DMA_BIT_MASK(64); +} + static int __init cell_iommu_init(void) { struct device_node *np; @@ -1066,12 +1192,13 @@ static int __init cell_iommu_init(void) * Note: should we make sure we have the IOMMU actually disabled ? */ if (iommu_is_off || - (!iommu_force_on && lmb_end_of_DRAM() <= 0x80000000ull)) + (!iommu_force_on && memblock_end_of_DRAM() <= 0x80000000ull)) if (cell_iommu_init_disabled() == 0) goto bail; /* Setup various ppc_md. callbacks */ ppc_md.pci_dma_dev_setup = cell_pci_dma_dev_setup; + ppc_md.dma_get_required_mask = cell_dma_get_required_mask; ppc_md.tce_build = tce_build_cell; ppc_md.tce_free = tce_free_cell; @@ -1101,7 +1228,7 @@ static int __init cell_iommu_init(void) /* Register callbacks on OF platform device addition/removal * to handle linking them to the right DMA operations */ - bus_register_notifier(&of_platform_bus_type, &cell_of_bus_notifier); + bus_register_notifier(&platform_bus_type, &cell_of_bus_notifier); return 0; } diff --git a/arch/powerpc/platforms/cell/pervasive.c b/arch/powerpc/platforms/cell/pervasive.c index 0304589c0a8..d17e98bc0c1 100644 --- a/arch/powerpc/platforms/cell/pervasive.c +++ b/arch/powerpc/platforms/cell/pervasive.c @@ -38,17 +38,13 @@ #include "pervasive.h" -static int sysreset_hack; - static void cbe_power_save(void) { unsigned long ctrl, thread_switch_control; - /* - * We need to hard disable interrupts, the local_irq_enable() done by - * our caller upon return will hard re-enable. - */ - hard_irq_disable(); + /* Ensure our interrupt state is properly tracked */ + if (!prep_irq_for_idle()) + return; ctrl = mfspr(SPRN_CTRLF); @@ -65,7 +61,7 @@ static void cbe_power_save(void) break; default: printk(KERN_WARNING "%s: unknown configuration\n", - __FUNCTION__); + __func__); break; } mtspr(SPRN_TSC_CELL, thread_switch_control); @@ -83,13 +79,13 @@ static void cbe_power_save(void) */ ctrl &= ~(CTRL_RUNLATCH | CTRL_TE); mtspr(SPRN_CTRLT, ctrl); + + /* Re-enable interrupts in MSR */ + __hard_irq_enable(); } static int cbe_system_reset_exception(struct pt_regs *regs) { - int cpu; - struct cbe_pmd_regs __iomem *pmd; - switch (regs->msr & SRR1_WAKEMASK) { case SRR1_WAKEEE: do_IRQ(regs); @@ -98,19 +94,7 @@ static int cbe_system_reset_exception(struct pt_regs *regs) timer_interrupt(regs); break; case SRR1_WAKEMT: - /* - * The BMC can inject user triggered system reset exceptions, - * but cannot set the system reset reason in srr1, - * so check an extra register here. - */ - if (sysreset_hack && (cpu = smp_processor_id()) == 0) { - pmd = cbe_get_cpu_pmd_regs(cpu); - if (in_be64(&pmd->ras_esc_0) & 0xffff) { - out_be64(&pmd->ras_esc_0, 0); - return 0; - } - } - break; + return cbe_sysreset_hack(); #ifdef CONFIG_CBE_RAS case SRR1_WAKESYSERR: cbe_system_error_exception(regs); @@ -134,8 +118,6 @@ void __init cbe_pervasive_init(void) if (!cpu_has_feature(CPU_FTR_PAUSE_ZERO)) return; - sysreset_hack = machine_is_compatible("IBM,CBPLUS-1.0"); - for_each_possible_cpu(cpu) { struct cbe_pmd_regs __iomem *regs = cbe_get_cpu_pmd_regs(cpu); if (!regs) @@ -144,12 +126,6 @@ void __init cbe_pervasive_init(void) /* Enable Pause(0) control bit */ out_be64(®s->pmcr, in_be64(®s->pmcr) | CBE_PMD_PAUSE_ZERO_CONTROL); - - /* Enable JTAG system-reset hack */ - if (sysreset_hack) - out_be32(®s->fir_mode_reg, - in_be32(®s->fir_mode_reg) | - CBE_PMD_FIR_MODE_M8); } ppc_md.power_save = cbe_power_save; diff --git a/arch/powerpc/platforms/cell/pervasive.h b/arch/powerpc/platforms/cell/pervasive.h index 7b50947f804..fd4d7b7092b 100644 --- a/arch/powerpc/platforms/cell/pervasive.h +++ b/arch/powerpc/platforms/cell/pervasive.h @@ -30,4 +30,13 @@ extern void cbe_system_error_exception(struct pt_regs *regs); extern void cbe_maintenance_exception(struct pt_regs *regs); extern void cbe_thermal_exception(struct pt_regs *regs); +#ifdef CONFIG_PPC_IBM_CELL_RESETBUTTON +extern int cbe_sysreset_hack(void); +#else +static inline int cbe_sysreset_hack(void) +{ + return 1; +} +#endif /* CONFIG_PPC_IBM_CELL_RESETBUTTON */ + #endif diff --git a/arch/powerpc/platforms/cell/pmu.c b/arch/powerpc/platforms/cell/pmu.c index 69ed0d7f164..348a27b1251 100644 --- a/arch/powerpc/platforms/cell/pmu.c +++ b/arch/powerpc/platforms/cell/pmu.c @@ -24,6 +24,7 @@ #include <linux/interrupt.h> #include <linux/types.h> +#include <linux/export.h> #include <asm/io.h> #include <asm/irq_regs.h> #include <asm/machdep.h> @@ -381,7 +382,7 @@ static int __init cbe_init_pm_irq(void) unsigned int irq; int rc, node; - for_each_node(node) { + for_each_online_node(node) { irq = irq_create_mapping(NULL, IIC_IRQ_IOEX_PMI | (node << IIC_IRQ_NODE_SHIFT)); if (irq == NO_IRQ) { @@ -391,7 +392,7 @@ static int __init cbe_init_pm_irq(void) } rc = request_irq(irq, cbe_pm_irq, - IRQF_DISABLED, "cbe-pmu-0", NULL); + 0, "cbe-pmu-0", NULL); if (rc) { printk("ERROR: Request for irq on node %d failed\n", node); diff --git a/arch/powerpc/platforms/cell/qpace_setup.c b/arch/powerpc/platforms/cell/qpace_setup.c new file mode 100644 index 00000000000..6e3409d590a --- /dev/null +++ b/arch/powerpc/platforms/cell/qpace_setup.c @@ -0,0 +1,148 @@ +/* + * linux/arch/powerpc/platforms/cell/qpace_setup.c + * + * Copyright (C) 1995 Linus Torvalds + * Adapted from 'alpha' version by Gary Thomas + * Modified by Cort Dougan (cort@cs.nmt.edu) + * Modified by PPC64 Team, IBM Corp + * Modified by Cell Team, IBM Deutschland Entwicklung GmbH + * Modified by Benjamin Krill <ben@codiert.org>, IBM Corp. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/export.h> +#include <linux/delay.h> +#include <linux/irq.h> +#include <linux/console.h> +#include <linux/of_platform.h> + +#include <asm/mmu.h> +#include <asm/processor.h> +#include <asm/io.h> +#include <asm/kexec.h> +#include <asm/pgtable.h> +#include <asm/prom.h> +#include <asm/rtas.h> +#include <asm/dma.h> +#include <asm/machdep.h> +#include <asm/time.h> +#include <asm/cputable.h> +#include <asm/irq.h> +#include <asm/spu.h> +#include <asm/spu_priv1.h> +#include <asm/udbg.h> +#include <asm/cell-regs.h> + +#include "interrupt.h" +#include "pervasive.h" +#include "ras.h" + +static void qpace_show_cpuinfo(struct seq_file *m) +{ + struct device_node *root; + const char *model = ""; + + root = of_find_node_by_path("/"); + if (root) + model = of_get_property(root, "model", NULL); + seq_printf(m, "machine\t\t: CHRP %s\n", model); + of_node_put(root); +} + +static void qpace_progress(char *s, unsigned short hex) +{ + printk("*** %04x : %s\n", hex, s ? s : ""); +} + +static const struct of_device_id qpace_bus_ids[] __initconst = { + { .type = "soc", }, + { .compatible = "soc", }, + { .type = "spider", }, + { .type = "axon", }, + { .type = "plb5", }, + { .type = "plb4", }, + { .type = "opb", }, + { .type = "ebc", }, + {}, +}; + +static int __init qpace_publish_devices(void) +{ + int node; + + /* Publish OF platform devices for southbridge IOs */ + of_platform_bus_probe(NULL, qpace_bus_ids, NULL); + + /* There is no device for the MIC memory controller, thus we create + * a platform device for it to attach the EDAC driver to. + */ + for_each_online_node(node) { + if (cbe_get_cpu_mic_tm_regs(cbe_node_to_cpu(node)) == NULL) + continue; + platform_device_register_simple("cbe-mic", node, NULL, 0); + } + + return 0; +} +machine_subsys_initcall(qpace, qpace_publish_devices); + +static void __init qpace_setup_arch(void) +{ +#ifdef CONFIG_SPU_BASE + spu_priv1_ops = &spu_priv1_mmio_ops; + spu_management_ops = &spu_management_of_ops; +#endif + + cbe_regs_init(); + +#ifdef CONFIG_CBE_RAS + cbe_ras_init(); +#endif + +#ifdef CONFIG_SMP + smp_init_cell(); +#endif + + /* init to some ~sane value until calibrate_delay() runs */ + loops_per_jiffy = 50000000; + + cbe_pervasive_init(); +#ifdef CONFIG_DUMMY_CONSOLE + conswitchp = &dummy_con; +#endif +} + +static int __init qpace_probe(void) +{ + unsigned long root = of_get_flat_dt_root(); + + if (!of_flat_dt_is_compatible(root, "IBM,QPACE")) + return 0; + + hpte_init_native(); + + return 1; +} + +define_machine(qpace) { + .name = "QPACE", + .probe = qpace_probe, + .setup_arch = qpace_setup_arch, + .show_cpuinfo = qpace_show_cpuinfo, + .restart = rtas_restart, + .power_off = rtas_power_off, + .halt = rtas_halt, + .get_boot_time = rtas_get_boot_time, + .get_rtc_time = rtas_get_rtc_time, + .set_rtc_time = rtas_set_rtc_time, + .calibrate_decr = generic_calibrate_decr, + .progress = qpace_progress, + .init_IRQ = iic_init_IRQ, +}; diff --git a/arch/powerpc/platforms/cell/ras.c b/arch/powerpc/platforms/cell/ras.c index e43024c0392..e865d748179 100644 --- a/arch/powerpc/platforms/cell/ras.c +++ b/arch/powerpc/platforms/cell/ras.c @@ -11,9 +11,13 @@ #include <linux/types.h> #include <linux/kernel.h> +#include <linux/slab.h> #include <linux/smp.h> #include <linux/reboot.h> +#include <linux/kexec.h> +#include <linux/crash_dump.h> +#include <asm/kexec.h> #include <asm/reg.h> #include <asm/io.h> #include <asm/prom.h> @@ -35,16 +39,16 @@ static void dump_fir(int cpu) /* Todo: do some nicer parsing of bits and based on them go down * to other sub-units FIRs and not only IIC */ - printk(KERN_ERR "Global Checkstop FIR : 0x%016lx\n", + printk(KERN_ERR "Global Checkstop FIR : 0x%016llx\n", in_be64(&pregs->checkstop_fir)); - printk(KERN_ERR "Global Recoverable FIR : 0x%016lx\n", + printk(KERN_ERR "Global Recoverable FIR : 0x%016llx\n", in_be64(&pregs->checkstop_fir)); - printk(KERN_ERR "Global MachineCheck FIR : 0x%016lx\n", + printk(KERN_ERR "Global MachineCheck FIR : 0x%016llx\n", in_be64(&pregs->spec_att_mchk_fir)); if (iregs == NULL) return; - printk(KERN_ERR "IOC FIR : 0x%016lx\n", + printk(KERN_ERR "IOC FIR : 0x%016llx\n", in_be64(&iregs->ioc_fir)); } @@ -110,9 +114,8 @@ static int __init cbe_ptcal_enable_on_node(int nid, int order) int ret = -ENOMEM; unsigned long addr; -#ifdef CONFIG_CRASH_DUMP - rtas_call(ptcal_stop_tok, 1, 1, NULL, nid); -#endif + if (is_kdump_kernel()) + rtas_call(ptcal_stop_tok, 1, 1, NULL, nid); area = kmalloc(sizeof(*area), GFP_KERNEL); if (!area) @@ -120,19 +123,31 @@ static int __init cbe_ptcal_enable_on_node(int nid, int order) area->nid = nid; area->order = order; - area->pages = alloc_pages_node(area->nid, GFP_KERNEL, area->order); + area->pages = alloc_pages_exact_node(area->nid, + GFP_KERNEL|__GFP_THISNODE, + area->order); - if (!area->pages) + if (!area->pages) { + printk(KERN_WARNING "%s: no page on node %d\n", + __func__, area->nid); goto out_free_area; + } - addr = __pa(page_address(area->pages)); + /* + * We move the ptcal area to the middle of the allocated + * page, in order to avoid prefetches in memcpy and similar + * functions stepping on it. + */ + addr = __pa(page_address(area->pages)) + (PAGE_SIZE >> 1); + printk(KERN_DEBUG "%s: enabling PTCAL on node %d address=0x%016lx\n", + __func__, area->nid, addr); ret = -EIO; if (rtas_call(ptcal_start_tok, 3, 1, NULL, area->nid, (unsigned int)(addr >> 32), (unsigned int)(addr & 0xffffffff))) { printk(KERN_ERR "%s: error enabling PTCAL on node %d!\n", - __FUNCTION__, nid); + __func__, nid); goto out_free_pages; } @@ -159,10 +174,12 @@ static int __init cbe_ptcal_enable(void) return -ENODEV; size = of_get_property(np, "ibm,cbe-ptcal-size", NULL); - if (!size) + if (!size) { + of_node_put(np); return -ENODEV; + } - pr_debug("%s: enabling PTCAL, size = 0x%x\n", __FUNCTION__, *size); + pr_debug("%s: enabling PTCAL, size = 0x%x\n", __func__, *size); order = get_order(*size); of_node_put(np); @@ -180,7 +197,7 @@ static int __init cbe_ptcal_enable(void) const u32 *nid = of_get_property(np, "node-id", NULL); if (!nid) { printk(KERN_ERR "%s: node %s is missing node-id?\n", - __FUNCTION__, np->full_name); + __func__, np->full_name); continue; } cbe_ptcal_enable_on_node(*nid, order); @@ -195,13 +212,13 @@ static int cbe_ptcal_disable(void) struct ptcal_area *area, *tmp; int ret = 0; - pr_debug("%s: disabling PTCAL\n", __FUNCTION__); + pr_debug("%s: disabling PTCAL\n", __func__); list_for_each_entry_safe(area, tmp, &ptcal_list, list) { /* disable ptcal on this node */ if (rtas_call(ptcal_stop_tok, 1, 1, NULL, area->nid)) { printk(KERN_ERR "%s: error disabling PTCAL " - "on node %d!\n", __FUNCTION__, + "on node %d!\n", __func__, area->nid); ret = -EIO; continue; @@ -226,10 +243,61 @@ static int cbe_ptcal_notify_reboot(struct notifier_block *nb, return cbe_ptcal_disable(); } +static void cbe_ptcal_crash_shutdown(void) +{ + cbe_ptcal_disable(); +} + static struct notifier_block cbe_ptcal_reboot_notifier = { .notifier_call = cbe_ptcal_notify_reboot }; +#ifdef CONFIG_PPC_IBM_CELL_RESETBUTTON +static int sysreset_hack; + +static int __init cbe_sysreset_init(void) +{ + struct cbe_pmd_regs __iomem *regs; + + sysreset_hack = of_machine_is_compatible("IBM,CBPLUS-1.0"); + if (!sysreset_hack) + return 0; + + regs = cbe_get_cpu_pmd_regs(0); + if (!regs) + return 0; + + /* Enable JTAG system-reset hack */ + out_be32(®s->fir_mode_reg, + in_be32(®s->fir_mode_reg) | + CBE_PMD_FIR_MODE_M8); + + return 0; +} +device_initcall(cbe_sysreset_init); + +int cbe_sysreset_hack(void) +{ + struct cbe_pmd_regs __iomem *regs; + + /* + * The BMC can inject user triggered system reset exceptions, + * but cannot set the system reset reason in srr1, + * so check an extra register here. + */ + if (sysreset_hack && (smp_processor_id() == 0)) { + regs = cbe_get_cpu_pmd_regs(0); + if (!regs) + return 0; + if (in_be64(®s->ras_esc_0) & 0x0000ffff) { + out_be64(®s->ras_esc_0, 0); + return 0; + } + } + return 1; +} +#endif /* CONFIG_PPC_IBM_CELL_RESETBUTTON */ + int __init cbe_ptcal_init(void) { int ret; @@ -241,12 +309,20 @@ int __init cbe_ptcal_init(void) return -ENODEV; ret = register_reboot_notifier(&cbe_ptcal_reboot_notifier); - if (ret) { - printk(KERN_ERR "Can't disable PTCAL, so not enabling\n"); - return ret; - } + if (ret) + goto out1; + + ret = crash_shutdown_register(&cbe_ptcal_crash_shutdown); + if (ret) + goto out2; return cbe_ptcal_enable(); + +out2: + unregister_reboot_notifier(&cbe_ptcal_reboot_notifier); +out1: + printk(KERN_ERR "Can't disable PTCAL, so not enabling\n"); + return ret; } arch_initcall(cbe_ptcal_init); diff --git a/arch/powerpc/platforms/cell/setup.c b/arch/powerpc/platforms/cell/setup.c index 5c531e8f9f6..6ae25fb6201 100644 --- a/arch/powerpc/platforms/cell/setup.c +++ b/arch/powerpc/platforms/cell/setup.c @@ -18,8 +18,8 @@ #include <linux/kernel.h> #include <linux/mm.h> #include <linux/stddef.h> +#include <linux/export.h> #include <linux/unistd.h> -#include <linux/slab.h> #include <linux/user.h> #include <linux/reboot.h> #include <linux/init.h> @@ -35,7 +35,6 @@ #include <asm/mmu.h> #include <asm/processor.h> #include <asm/io.h> -#include <asm/kexec.h> #include <asm/pgtable.h> #include <asm/prom.h> #include <asm/rtas.h> @@ -53,6 +52,7 @@ #include <asm/udbg.h> #include <asm/mpic.h> #include <asm/cell-regs.h> +#include <asm/io-workarounds.h> #include "interrupt.h" #include "pervasive.h" @@ -117,12 +117,59 @@ static void cell_fixup_pcie_rootcomplex(struct pci_dev *dev) } DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, cell_fixup_pcie_rootcomplex); +static int cell_setup_phb(struct pci_controller *phb) +{ + const char *model; + struct device_node *np; + + int rc = rtas_setup_phb(phb); + if (rc) + return rc; + + np = phb->dn; + model = of_get_property(np, "model", NULL); + if (model == NULL || strcmp(np->name, "pci")) + return 0; + + /* Setup workarounds for spider */ + if (strcmp(model, "Spider")) + return 0; + + iowa_register_bus(phb, &spiderpci_ops, &spiderpci_iowa_init, + (void *)SPIDER_PCI_REG_BASE); + return 0; +} + +static const struct of_device_id cell_bus_ids[] __initconst = { + { .type = "soc", }, + { .compatible = "soc", }, + { .type = "spider", }, + { .type = "axon", }, + { .type = "plb5", }, + { .type = "plb4", }, + { .type = "opb", }, + { .type = "ebc", }, + {}, +}; + static int __init cell_publish_devices(void) { + struct device_node *root = of_find_node_by_path("/"); + struct device_node *np; int node; /* Publish OF platform devices for southbridge IOs */ - of_platform_bus_probe(NULL, NULL, NULL); + of_platform_bus_probe(NULL, cell_bus_ids, NULL); + + /* On spider based blades, we need to manually create the OF + * platform devices for the PCI host bridges + */ + for_each_child_of_node(root, np) { + if (np->type == NULL || (strcmp(np->type, "pci") != 0 && + strcmp(np->type, "pciex") != 0)) + continue; + of_platform_device_create(np, NULL, NULL); + } /* There is no device for the MIC memory controller, thus we create * a platform device for it to attach the EDAC driver to. @@ -132,26 +179,15 @@ static int __init cell_publish_devices(void) continue; platform_device_register_simple("cbe-mic", node, NULL, 0); } + return 0; } machine_subsys_initcall(cell, cell_publish_devices); -static void cell_mpic_cascade(unsigned int irq, struct irq_desc *desc) -{ - struct mpic *mpic = desc->handler_data; - unsigned int virq; - - virq = mpic_get_one_irq(mpic); - if (virq != NO_IRQ) - generic_handle_irq(virq); - desc->chip->eoi(irq); -} - static void __init mpic_init_IRQ(void) { struct device_node *dn; struct mpic *mpic; - unsigned int virq; for (dn = NULL; (dn = of_find_node_by_name(dn, "interrupt-controller"));) { @@ -161,19 +197,11 @@ static void __init mpic_init_IRQ(void) /* The MPIC driver will get everything it needs from the * device-tree, just pass 0 to all arguments */ - mpic = mpic_alloc(dn, 0, 0, 0, 0, " MPIC "); + mpic = mpic_alloc(dn, 0, MPIC_SECONDARY | MPIC_NO_RESET, + 0, 0, " MPIC "); if (mpic == NULL) continue; mpic_init(mpic); - - virq = irq_of_parse_and_map(dn, 0); - if (virq == NO_IRQ) - continue; - - printk(KERN_INFO "%s : hooking up to IRQ %d\n", - dn->full_name, virq); - set_irq_data(virq, mpic); - set_irq_chained_handler(virq, cell_mpic_cascade); } } @@ -213,7 +241,7 @@ static void __init cell_setup_arch(void) /* Find and initialize PCI host bridges */ init_pci_config_tokens(); - find_and_init_phbs(); + cbe_pervasive_init(); #ifdef CONFIG_DUMMY_CONSOLE conswitchp = &dummy_con; @@ -249,10 +277,5 @@ define_machine(cell) { .calibrate_decr = generic_calibrate_decr, .progress = cell_progress, .init_IRQ = cell_init_irq, - .pci_setup_phb = rtas_setup_phb, -#ifdef CONFIG_KEXEC - .machine_kexec = default_machine_kexec, - .machine_kexec_prepare = default_machine_kexec_prepare, - .machine_crash_shutdown = default_machine_crash_shutdown, -#endif + .pci_setup_phb = cell_setup_phb, }; diff --git a/arch/powerpc/platforms/cell/smp.c b/arch/powerpc/platforms/cell/smp.c index efb3964457b..c8017a7bcab 100644 --- a/arch/powerpc/platforms/cell/smp.c +++ b/arch/powerpc/platforms/cell/smp.c @@ -15,7 +15,6 @@ #undef DEBUG #include <linux/kernel.h> -#include <linux/module.h> #include <linux/sched.h> #include <linux/smp.h> #include <linux/interrupt.h> @@ -24,11 +23,11 @@ #include <linux/spinlock.h> #include <linux/cache.h> #include <linux/err.h> -#include <linux/sysdev.h> +#include <linux/device.h> #include <linux/cpu.h> #include <asm/ptrace.h> -#include <asm/atomic.h> +#include <linux/atomic.h> #include <asm/irq.h> #include <asm/page.h> #include <asm/pgtable.h> @@ -36,13 +35,12 @@ #include <asm/prom.h> #include <asm/smp.h> #include <asm/paca.h> -#include <asm/time.h> #include <asm/machdep.h> #include <asm/cputable.h> #include <asm/firmware.h> -#include <asm/system.h> #include <asm/rtas.h> #include <asm/cputhreads.h> +#include <asm/code-patching.h> #include "interrupt.h" #include <asm/udbg.h> @@ -54,13 +52,11 @@ #endif /* - * The primary thread of each non-boot processor is recorded here before - * smp init. + * The Primary thread of each non-boot processor was started from the OF client + * interface by prom_hold_cpus and is spinning on secondary_hold_spinloop. */ static cpumask_t of_spin_map; -extern void generic_secondary_smp_init(unsigned long); - /** * smp_startup_cpu() - start the given cpu * @@ -72,15 +68,15 @@ extern void generic_secondary_smp_init(unsigned long); * 0 - failure * 1 - success */ -static inline int __devinit smp_startup_cpu(unsigned int lcpu) +static inline int smp_startup_cpu(unsigned int lcpu) { int status; - unsigned long start_here = __pa((u32)*((unsigned long *) - generic_secondary_smp_init)); + unsigned long start_here = + __pa(ppc_function_entry(generic_secondary_smp_init)); unsigned int pcpu; int start_cpu; - if (cpu_isset(lcpu, of_spin_map)) + if (cpumask_test_cpu(lcpu, &of_spin_map)) /* Already started by OF and sitting in spin loop */ return 1; @@ -106,66 +102,30 @@ static inline int __devinit smp_startup_cpu(unsigned int lcpu) return 1; } -static void smp_iic_message_pass(int target, int msg) -{ - unsigned int i; - - if (target < NR_CPUS) { - iic_cause_IPI(target, msg); - } else { - for_each_online_cpu(i) { - if (target == MSG_ALL_BUT_SELF - && i == smp_processor_id()) - continue; - iic_cause_IPI(i, msg); - } - } -} - static int __init smp_iic_probe(void) { iic_request_IPIs(); - return cpus_weight(cpu_possible_map); + return cpumask_weight(cpu_possible_mask); } -static void __devinit smp_iic_setup_cpu(int cpu) +static void smp_cell_setup_cpu(int cpu) { if (cpu != boot_cpuid) iic_setup_cpu(); -} - -static DEFINE_SPINLOCK(timebase_lock); -static unsigned long timebase = 0; - -static void __devinit cell_give_timebase(void) -{ - spin_lock(&timebase_lock); - rtas_call(rtas_token("freeze-time-base"), 0, 1, NULL); - timebase = get_tb(); - spin_unlock(&timebase_lock); - - while (timebase) - barrier(); - rtas_call(rtas_token("thaw-time-base"), 0, 1, NULL); -} -static void __devinit cell_take_timebase(void) -{ - while (!timebase) - barrier(); - spin_lock(&timebase_lock); - set_tb(timebase >> 32, timebase & 0xffffffff); - timebase = 0; - spin_unlock(&timebase_lock); + /* + * change default DABRX to allow user watchpoints + */ + mtspr(SPRN_DABRX, DABRX_KERNEL | DABRX_USER); } -static void __devinit smp_cell_kick_cpu(int nr) +static int smp_cell_kick_cpu(int nr) { BUG_ON(nr < 0 || nr >= NR_CPUS); if (!smp_startup_cpu(nr)) - return; + return -ENOENT; /* * The processor is currently spinning, waiting for the @@ -173,27 +133,16 @@ static void __devinit smp_cell_kick_cpu(int nr) * the processor will continue on to secondary_start */ paca[nr].cpu_start = 1; -} -static int smp_cell_cpu_bootable(unsigned int nr) -{ - /* Special case - we inhibit secondary thread startup - * during boot if the user requests it. Odd-numbered - * cpus are assumed to be secondary threads. - */ - if (system_state < SYSTEM_RUNNING && - cpu_has_feature(CPU_FTR_SMT) && - !smt_enabled_at_boot && cpu_thread_in_core(nr) != 0) - return 0; - - return 1; + return 0; } + static struct smp_ops_t bpa_iic_smp_ops = { - .message_pass = smp_iic_message_pass, + .message_pass = iic_message_pass, .probe = smp_iic_probe, .kick_cpu = smp_cell_kick_cpu, - .setup_cpu = smp_iic_setup_cpu, - .cpu_bootable = smp_cell_cpu_bootable, + .setup_cpu = smp_cell_setup_cpu, + .cpu_bootable = smp_generic_cpu_bootable, }; /* This is called very early */ @@ -208,23 +157,18 @@ void __init smp_init_cell(void) /* Mark threads which are still spinning in hold loops. */ if (cpu_has_feature(CPU_FTR_SMT)) { for_each_present_cpu(i) { - if (i % 2 == 0) - /* - * Even-numbered logical cpus correspond to - * primary threads. - */ - cpu_set(i, of_spin_map); + if (cpu_thread_in_core(i) == 0) + cpumask_set_cpu(i, &of_spin_map); } - } else { - of_spin_map = cpu_present_map; - } + } else + cpumask_copy(&of_spin_map, cpu_present_mask); - cpu_clear(boot_cpuid, of_spin_map); + cpumask_clear_cpu(boot_cpuid, &of_spin_map); /* Non-lpar has additional take/give timebase */ if (rtas_token("freeze-time-base") != RTAS_UNKNOWN_SERVICE) { - smp_ops->give_timebase = cell_give_timebase; - smp_ops->take_timebase = cell_take_timebase; + smp_ops->give_timebase = rtas_give_timebase; + smp_ops->take_timebase = rtas_take_timebase; } DBG(" <- smp_init_cell()\n"); diff --git a/arch/powerpc/platforms/cell/spider-pci.c b/arch/powerpc/platforms/cell/spider-pci.c new file mode 100644 index 00000000000..f1f7878893f --- /dev/null +++ b/arch/powerpc/platforms/cell/spider-pci.c @@ -0,0 +1,184 @@ +/* + * IO workarounds for PCI on Celleb/Cell platform + * + * (C) Copyright 2006-2007 TOSHIBA CORPORATION + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#undef DEBUG + +#include <linux/kernel.h> +#include <linux/of_platform.h> +#include <linux/slab.h> +#include <linux/io.h> + +#include <asm/ppc-pci.h> +#include <asm/pci-bridge.h> +#include <asm/io-workarounds.h> + +#define SPIDER_PCI_DISABLE_PREFETCH + +struct spiderpci_iowa_private { + void __iomem *regs; +}; + +static void spiderpci_io_flush(struct iowa_bus *bus) +{ + struct spiderpci_iowa_private *priv; + u32 val; + + priv = bus->private; + val = in_be32(priv->regs + SPIDER_PCI_DUMMY_READ); + iosync(); +} + +#define SPIDER_PCI_MMIO_READ(name, ret) \ +static ret spiderpci_##name(const PCI_IO_ADDR addr) \ +{ \ + ret val = __do_##name(addr); \ + spiderpci_io_flush(iowa_mem_find_bus(addr)); \ + return val; \ +} + +#define SPIDER_PCI_MMIO_READ_STR(name) \ +static void spiderpci_##name(const PCI_IO_ADDR addr, void *buf, \ + unsigned long count) \ +{ \ + __do_##name(addr, buf, count); \ + spiderpci_io_flush(iowa_mem_find_bus(addr)); \ +} + +SPIDER_PCI_MMIO_READ(readb, u8) +SPIDER_PCI_MMIO_READ(readw, u16) +SPIDER_PCI_MMIO_READ(readl, u32) +SPIDER_PCI_MMIO_READ(readq, u64) +SPIDER_PCI_MMIO_READ(readw_be, u16) +SPIDER_PCI_MMIO_READ(readl_be, u32) +SPIDER_PCI_MMIO_READ(readq_be, u64) +SPIDER_PCI_MMIO_READ_STR(readsb) +SPIDER_PCI_MMIO_READ_STR(readsw) +SPIDER_PCI_MMIO_READ_STR(readsl) + +static void spiderpci_memcpy_fromio(void *dest, const PCI_IO_ADDR src, + unsigned long n) +{ + __do_memcpy_fromio(dest, src, n); + spiderpci_io_flush(iowa_mem_find_bus(src)); +} + +static int __init spiderpci_pci_setup_chip(struct pci_controller *phb, + void __iomem *regs) +{ + void *dummy_page_va; + dma_addr_t dummy_page_da; + +#ifdef SPIDER_PCI_DISABLE_PREFETCH + u32 val = in_be32(regs + SPIDER_PCI_VCI_CNTL_STAT); + pr_debug("SPIDER_IOWA:PVCI_Control_Status was 0x%08x\n", val); + out_be32(regs + SPIDER_PCI_VCI_CNTL_STAT, val | 0x8); +#endif /* SPIDER_PCI_DISABLE_PREFETCH */ + + /* setup dummy read */ + /* + * On CellBlade, we can't know that which XDR memory is used by + * kmalloc() to allocate dummy_page_va. + * In order to imporve the performance, the XDR which is used to + * allocate dummy_page_va is the nearest the spider-pci. + * We have to select the CBE which is the nearest the spider-pci + * to allocate memory from the best XDR, but I don't know that + * how to do. + * + * Celleb does not have this problem, because it has only one XDR. + */ + dummy_page_va = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!dummy_page_va) { + pr_err("SPIDERPCI-IOWA:Alloc dummy_page_va failed.\n"); + return -1; + } + + dummy_page_da = dma_map_single(phb->parent, dummy_page_va, + PAGE_SIZE, DMA_FROM_DEVICE); + if (dma_mapping_error(phb->parent, dummy_page_da)) { + pr_err("SPIDER-IOWA:Map dummy page filed.\n"); + kfree(dummy_page_va); + return -1; + } + + out_be32(regs + SPIDER_PCI_DUMMY_READ_BASE, dummy_page_da); + + return 0; +} + +int __init spiderpci_iowa_init(struct iowa_bus *bus, void *data) +{ + void __iomem *regs = NULL; + struct spiderpci_iowa_private *priv; + struct device_node *np = bus->phb->dn; + struct resource r; + unsigned long offset = (unsigned long)data; + + pr_debug("SPIDERPCI-IOWA:Bus initialize for spider(%s)\n", + np->full_name); + + priv = kzalloc(sizeof(struct spiderpci_iowa_private), GFP_KERNEL); + if (!priv) { + pr_err("SPIDERPCI-IOWA:" + "Can't allocate struct spiderpci_iowa_private"); + return -1; + } + + if (of_address_to_resource(np, 0, &r)) { + pr_err("SPIDERPCI-IOWA:Can't get resource.\n"); + goto error; + } + + regs = ioremap(r.start + offset, SPIDER_PCI_REG_SIZE); + if (!regs) { + pr_err("SPIDERPCI-IOWA:ioremap failed.\n"); + goto error; + } + priv->regs = regs; + bus->private = priv; + + if (spiderpci_pci_setup_chip(bus->phb, regs)) + goto error; + + return 0; + +error: + kfree(priv); + bus->private = NULL; + + if (regs) + iounmap(regs); + + return -1; +} + +struct ppc_pci_io spiderpci_ops = { + .readb = spiderpci_readb, + .readw = spiderpci_readw, + .readl = spiderpci_readl, + .readq = spiderpci_readq, + .readw_be = spiderpci_readw_be, + .readl_be = spiderpci_readl_be, + .readq_be = spiderpci_readq_be, + .readsb = spiderpci_readsb, + .readsw = spiderpci_readsw, + .readsl = spiderpci_readsl, + .memcpy_fromio = spiderpci_memcpy_fromio, +}; + diff --git a/arch/powerpc/platforms/cell/spider-pic.c b/arch/powerpc/platforms/cell/spider-pic.c index 3f4b4aef756..1f72f4ab635 100644 --- a/arch/powerpc/platforms/cell/spider-pic.c +++ b/arch/powerpc/platforms/cell/spider-pic.c @@ -62,15 +62,15 @@ enum { #define SPIDER_IRQ_INVALID 63 struct spider_pic { - struct irq_host *host; + struct irq_domain *host; void __iomem *regs; unsigned int node_id; }; static struct spider_pic spider_pics[SPIDER_CHIP_COUNT]; -static struct spider_pic *spider_virq_to_pic(unsigned int virq) +static struct spider_pic *spider_irq_data_to_pic(struct irq_data *d) { - return irq_map[virq].host->host_data; + return irq_data_get_irq_chip_data(d); } static void __iomem *spider_get_irq_config(struct spider_pic *pic, @@ -79,30 +79,30 @@ static void __iomem *spider_get_irq_config(struct spider_pic *pic, return pic->regs + TIR_CFGA + 8 * src; } -static void spider_unmask_irq(unsigned int virq) +static void spider_unmask_irq(struct irq_data *d) { - struct spider_pic *pic = spider_virq_to_pic(virq); - void __iomem *cfg = spider_get_irq_config(pic, irq_map[virq].hwirq); + struct spider_pic *pic = spider_irq_data_to_pic(d); + void __iomem *cfg = spider_get_irq_config(pic, irqd_to_hwirq(d)); out_be32(cfg, in_be32(cfg) | 0x30000000u); } -static void spider_mask_irq(unsigned int virq) +static void spider_mask_irq(struct irq_data *d) { - struct spider_pic *pic = spider_virq_to_pic(virq); - void __iomem *cfg = spider_get_irq_config(pic, irq_map[virq].hwirq); + struct spider_pic *pic = spider_irq_data_to_pic(d); + void __iomem *cfg = spider_get_irq_config(pic, irqd_to_hwirq(d)); out_be32(cfg, in_be32(cfg) & ~0x30000000u); } -static void spider_ack_irq(unsigned int virq) +static void spider_ack_irq(struct irq_data *d) { - struct spider_pic *pic = spider_virq_to_pic(virq); - unsigned int src = irq_map[virq].hwirq; + struct spider_pic *pic = spider_irq_data_to_pic(d); + unsigned int src = irqd_to_hwirq(d); /* Reset edge detection logic if necessary */ - if (get_irq_desc(virq)->status & IRQ_LEVEL) + if (irqd_is_level_type(d)) return; /* Only interrupts 47 to 50 can be set to edge */ @@ -113,13 +113,12 @@ static void spider_ack_irq(unsigned int virq) out_be32(pic->regs + TIR_EDC, 0x100 | (src & 0xf)); } -static int spider_set_irq_type(unsigned int virq, unsigned int type) +static int spider_set_irq_type(struct irq_data *d, unsigned int type) { unsigned int sense = type & IRQ_TYPE_SENSE_MASK; - struct spider_pic *pic = spider_virq_to_pic(virq); - unsigned int hw = irq_map[virq].hwirq; + struct spider_pic *pic = spider_irq_data_to_pic(d); + unsigned int hw = irqd_to_hwirq(d); void __iomem *cfg = spider_get_irq_config(pic, hw); - struct irq_desc *desc = get_irq_desc(virq); u32 old_mask; u32 ic; @@ -147,15 +146,9 @@ static int spider_set_irq_type(unsigned int virq, unsigned int type) return -EINVAL; } - /* Update irq_desc */ - desc->status &= ~(IRQ_TYPE_SENSE_MASK | IRQ_LEVEL); - desc->status |= type & IRQ_TYPE_SENSE_MASK; - if (type & (IRQ_TYPE_LEVEL_HIGH | IRQ_TYPE_LEVEL_LOW)) - desc->status |= IRQ_LEVEL; - /* Configure the source. One gross hack that was there before and * that I've kept around is the priority to the BE which I set to - * be the same as the interrupt source number. I don't know wether + * be the same as the interrupt source number. I don't know whether * that's supposed to make any kind of sense however, we'll have to * decide that, but for now, I'm not changing the behaviour. */ @@ -168,26 +161,27 @@ static int spider_set_irq_type(unsigned int virq, unsigned int type) } static struct irq_chip spider_pic = { - .typename = " SPIDER ", - .unmask = spider_unmask_irq, - .mask = spider_mask_irq, - .ack = spider_ack_irq, - .set_type = spider_set_irq_type, + .name = "SPIDER", + .irq_unmask = spider_unmask_irq, + .irq_mask = spider_mask_irq, + .irq_ack = spider_ack_irq, + .irq_set_type = spider_set_irq_type, }; -static int spider_host_map(struct irq_host *h, unsigned int virq, +static int spider_host_map(struct irq_domain *h, unsigned int virq, irq_hw_number_t hw) { - set_irq_chip_and_handler(virq, &spider_pic, handle_level_irq); + irq_set_chip_data(virq, h->host_data); + irq_set_chip_and_handler(virq, &spider_pic, handle_level_irq); /* Set default irq type */ - set_irq_type(virq, IRQ_TYPE_NONE); + irq_set_irq_type(virq, IRQ_TYPE_NONE); return 0; } -static int spider_host_xlate(struct irq_host *h, struct device_node *ct, - u32 *intspec, unsigned int intsize, +static int spider_host_xlate(struct irq_domain *h, struct device_node *ct, + const u32 *intspec, unsigned int intsize, irq_hw_number_t *out_hwirq, unsigned int *out_flags) { @@ -200,14 +194,15 @@ static int spider_host_xlate(struct irq_host *h, struct device_node *ct, return 0; } -static struct irq_host_ops spider_host_ops = { +static const struct irq_domain_ops spider_host_ops = { .map = spider_host_map, .xlate = spider_host_xlate, }; static void spider_irq_cascade(unsigned int irq, struct irq_desc *desc) { - struct spider_pic *pic = desc->handler_data; + struct irq_chip *chip = irq_desc_get_chip(desc); + struct spider_pic *pic = irq_desc_get_handler_data(desc); unsigned int cs, virq; cs = in_be32(pic->regs + TIR_CS) >> 24; @@ -215,15 +210,17 @@ static void spider_irq_cascade(unsigned int irq, struct irq_desc *desc) virq = NO_IRQ; else virq = irq_linear_revmap(pic->host, cs); + if (virq != NO_IRQ) generic_handle_irq(virq); - desc->chip->eoi(irq); + + chip->irq_eoi(&desc->irq_data); } /* For hooking up the cascace we have a problem. Our device-tree is * crap and we don't know on which BE iic interrupt we are hooked on at * least not the "standard" way. We can reconstitute it based on two - * informations though: which BE node we are connected to and wether + * informations though: which BE node we are connected to and whether * we are connected to IOIF0 or IOIF1. Right now, we really only care * about the IBM cell blade and we know that its firmware gives us an * interrupt-map property which is pretty strange. @@ -235,15 +232,12 @@ static unsigned int __init spider_find_cascade_and_node(struct spider_pic *pic) int imaplen, intsize, unit; struct device_node *iic; - /* First, we check wether we have a real "interrupts" in the device + /* First, we check whether we have a real "interrupts" in the device * tree in case the device-tree is ever fixed */ - struct of_irq oirq; - if (of_irq_map_one(pic->host->of_node, 0, &oirq) == 0) { - virq = irq_create_of_mapping(oirq.controller, oirq.specifier, - oirq.size); + virq = irq_of_parse_and_map(pic->host->of_node, 0); + if (virq) return virq; - } /* Now do the horrible hacks */ tmp = of_get_property(pic->host->of_node, "#interrupt-cells", NULL); @@ -258,8 +252,10 @@ static unsigned int __init spider_find_cascade_and_node(struct spider_pic *pic) return NO_IRQ; imap += intsize + 1; tmp = of_get_property(iic, "#interrupt-cells", NULL); - if (tmp == NULL) + if (tmp == NULL) { + of_node_put(iic); return NO_IRQ; + } intsize = *tmp; /* Assume unit is last entry of interrupt specifier */ unit = imap[intsize - 1]; @@ -300,12 +296,10 @@ static void __init spider_init_one(struct device_node *of_node, int chip, panic("spider_pic: can't map registers !"); /* Allocate a host */ - pic->host = irq_alloc_host(of_node_get(of_node), IRQ_HOST_MAP_LINEAR, - SPIDER_SRC_COUNT, &spider_host_ops, - SPIDER_IRQ_INVALID); + pic->host = irq_domain_add_linear(of_node, SPIDER_SRC_COUNT, + &spider_host_ops, pic); if (pic->host == NULL) panic("spider_pic: can't allocate irq host !"); - pic->host->host_data = pic; /* Go through all sources and disable them */ for (i = 0; i < SPIDER_SRC_COUNT; i++) { @@ -323,8 +317,8 @@ static void __init spider_init_one(struct device_node *of_node, int chip, virq = spider_find_cascade_and_node(pic); if (virq == NO_IRQ) return; - set_irq_data(virq, pic); - set_irq_chained_handler(virq, spider_irq_cascade); + irq_set_handler_data(virq, pic); + irq_set_chained_handler(virq, spider_irq_cascade); printk(KERN_INFO "spider_pic: node %d, addr: 0x%lx %s\n", pic->node_id, addr, of_node->full_name); diff --git a/arch/powerpc/platforms/cell/spu_base.c b/arch/powerpc/platforms/cell/spu_base.c index 712001f6b7d..f85db3a69b4 100644 --- a/arch/powerpc/platforms/cell/spu_base.c +++ b/arch/powerpc/platforms/cell/spu_base.c @@ -32,11 +32,13 @@ #include <linux/io.h> #include <linux/mutex.h> #include <linux/linux_logo.h> +#include <linux/syscore_ops.h> #include <asm/spu.h> #include <asm/spu_priv1.h> #include <asm/spu_csa.h> #include <asm/xmon.h> #include <asm/prom.h> +#include <asm/kexec.h> const struct spu_management_ops *spu_management_ops; EXPORT_SYMBOL_GPL(spu_management_ops); @@ -114,7 +116,7 @@ static inline void mm_needs_global_tlbie(struct mm_struct *mm) int nr = (NR_CPUS > 1) ? NR_CPUS : NR_CPUS + 1; /* Global TLBIE broadcast required with SPEs. */ - __cpus_setall(&mm->cpu_vm_mask, nr); + bitmap_fill(cpumask_bits(mm_cpumask(mm)), nr); } void spu_associate_mm(struct spu *spu, struct mm_struct *mm) @@ -141,13 +143,17 @@ static void spu_restart_dma(struct spu *spu) if (!test_bit(SPU_CONTEXT_SWITCH_PENDING, &spu->flags)) out_be64(&priv2->mfc_control_RW, MFC_CNTL_RESTART_DMA_COMMAND); + else { + set_bit(SPU_CONTEXT_FAULT_PENDING, &spu->flags); + mb(); + } } static inline void spu_load_slb(struct spu *spu, int slbe, struct spu_slb *slb) { struct spu_priv2 __iomem *priv2 = spu->priv2; - pr_debug("%s: adding SLB[%d] 0x%016lx 0x%016lx\n", + pr_debug("%s: adding SLB[%d] 0x%016llx 0x%016llx\n", __func__, slbe, slb->vsid, slb->esid); out_be64(&priv2->slb_index_W, slbe); @@ -165,7 +171,7 @@ static int __spu_trap_data_seg(struct spu *spu, unsigned long ea) struct spu_slb slb; int psize; - pr_debug("%s\n", __FUNCTION__); + pr_debug("%s\n", __func__); slb.esid = (ea & ESID_MASK) | SLB_ESID_V; @@ -215,22 +221,34 @@ static int __spu_trap_data_seg(struct spu *spu, unsigned long ea) extern int hash_page(unsigned long ea, unsigned long access, unsigned long trap); //XXX static int __spu_trap_data_map(struct spu *spu, unsigned long ea, u64 dsisr) { - pr_debug("%s, %lx, %lx\n", __FUNCTION__, dsisr, ea); - - /* Handle kernel space hash faults immediately. - User hash faults need to be deferred to process context. */ - if ((dsisr & MFC_DSISR_PTE_NOT_FOUND) - && REGION_ID(ea) != USER_REGION_ID - && hash_page(ea, _PAGE_PRESENT, 0x300) == 0) { - spu_restart_dma(spu); - return 0; + int ret; + + pr_debug("%s, %llx, %lx\n", __func__, dsisr, ea); + + /* + * Handle kernel space hash faults immediately. User hash + * faults need to be deferred to process context. + */ + if ((dsisr & MFC_DSISR_PTE_NOT_FOUND) && + (REGION_ID(ea) != USER_REGION_ID)) { + + spin_unlock(&spu->register_lock); + ret = hash_page(ea, _PAGE_PRESENT, 0x300); + spin_lock(&spu->register_lock); + + if (!ret) { + spu_restart_dma(spu); + return 0; + } } - spu->class_0_pending = 0; - spu->dar = ea; - spu->dsisr = dsisr; + spu->class_1_dar = ea; + spu->class_1_dsisr = dsisr; + + spu->stop_callback(spu, 1); - spu->stop_callback(spu); + spu->class_1_dar = 0; + spu->class_1_dsisr = 0; return 0; } @@ -318,13 +336,13 @@ spu_irq_class_0(int irq, void *data) stat = spu_int_stat_get(spu, 0) & mask; spu->class_0_pending |= stat; - spu->dsisr = spu_mfc_dsisr_get(spu); - spu->dar = spu_mfc_dar_get(spu); - spin_unlock(&spu->register_lock); - - spu->stop_callback(spu); + spu->class_0_dar = spu_mfc_dar_get(spu); + spu->stop_callback(spu, 0); + spu->class_0_pending = 0; + spu->class_0_dar = 0; spu_int_stat_clear(spu, 0, stat); + spin_unlock(&spu->register_lock); return IRQ_HANDLED; } @@ -347,13 +365,12 @@ spu_irq_class_1(int irq, void *data) spu_mfc_dsisr_set(spu, 0ul); spu_int_stat_clear(spu, 1, stat); + pr_debug("%s: %lx %lx %lx %lx\n", __func__, mask, stat, + dar, dsisr); + if (stat & CLASS1_SEGMENT_FAULT_INTR) __spu_trap_data_seg(spu, dar); - spin_unlock(&spu->register_lock); - pr_debug("%s: %lx %lx %lx %lx\n", __FUNCTION__, mask, stat, - dar, dsisr); - if (stat & CLASS1_STORAGE_FAULT_INTR) __spu_trap_data_map(spu, dar, dsisr); @@ -363,6 +380,11 @@ spu_irq_class_1(int irq, void *data) if (stat & CLASS1_LS_COMPARE_SUSPEND_ON_PUT_INTR) ; + spu->class_1_dsisr = 0; + spu->class_1_dar = 0; + + spin_unlock(&spu->register_lock); + return stat ? IRQ_HANDLED : IRQ_NONE; } @@ -381,14 +403,12 @@ spu_irq_class_2(int irq, void *data) mask = spu_int_mask_get(spu, 2); /* ignore interrupts we're not waiting for */ stat &= mask; - /* mailbox interrupts are level triggered. mask them now before * acknowledging */ if (stat & mailbox_intrs) spu_int_mask_and(spu, 2, ~(stat & mailbox_intrs)); /* acknowledge all interrupts before the callbacks */ spu_int_stat_clear(spu, 2, stat); - spin_unlock(&spu->register_lock); pr_debug("class 2 interrupt %d, %lx, %lx\n", irq, stat, mask); @@ -396,10 +416,10 @@ spu_irq_class_2(int irq, void *data) spu->ibox_callback(spu); if (stat & CLASS2_SPU_STOP_INTR) - spu->stop_callback(spu); + spu->stop_callback(spu, 2); if (stat & CLASS2_SPU_HALT_INTR) - spu->stop_callback(spu); + spu->stop_callback(spu, 2); if (stat & CLASS2_SPU_DMA_TAG_GROUP_COMPLETE_INTR) spu->mfc_callback(spu); @@ -408,6 +428,9 @@ spu_irq_class_2(int irq, void *data) spu->wbox_callback(spu); spu->stats.class2_intr++; + + spin_unlock(&spu->register_lock); + return stat ? IRQ_HANDLED : IRQ_NONE; } @@ -419,8 +442,7 @@ static int spu_request_irqs(struct spu *spu) snprintf(spu->irq_c0, sizeof (spu->irq_c0), "spe%02d.0", spu->number); ret = request_irq(spu->irqs[0], spu_irq_class_0, - IRQF_DISABLED, - spu->irq_c0, spu); + 0, spu->irq_c0, spu); if (ret) goto bail0; } @@ -428,8 +450,7 @@ static int spu_request_irqs(struct spu *spu) snprintf(spu->irq_c1, sizeof (spu->irq_c1), "spe%02d.1", spu->number); ret = request_irq(spu->irqs[1], spu_irq_class_1, - IRQF_DISABLED, - spu->irq_c1, spu); + 0, spu->irq_c1, spu); if (ret) goto bail1; } @@ -437,8 +458,7 @@ static int spu_request_irqs(struct spu *spu) snprintf(spu->irq_c2, sizeof (spu->irq_c2), "spe%02d.2", spu->number); ret = request_irq(spu->irqs[2], spu_irq_class_2, - IRQF_DISABLED, - spu->irq_c2, spu); + 0, spu->irq_c2, spu); if (ret) goto bail2; } @@ -499,41 +519,32 @@ void spu_init_channels(struct spu *spu) } EXPORT_SYMBOL_GPL(spu_init_channels); -static int spu_shutdown(struct sys_device *sysdev) -{ - struct spu *spu = container_of(sysdev, struct spu, sysdev); - - spu_free_irqs(spu); - spu_destroy_spu(spu); - return 0; -} - -static struct sysdev_class spu_sysdev_class = { +static struct bus_type spu_subsys = { .name = "spu", - .shutdown = spu_shutdown, + .dev_name = "spu", }; -int spu_add_sysdev_attr(struct sysdev_attribute *attr) +int spu_add_dev_attr(struct device_attribute *attr) { struct spu *spu; mutex_lock(&spu_full_list_mutex); list_for_each_entry(spu, &spu_full_list, full_list) - sysdev_create_file(&spu->sysdev, attr); + device_create_file(&spu->dev, attr); mutex_unlock(&spu_full_list_mutex); return 0; } -EXPORT_SYMBOL_GPL(spu_add_sysdev_attr); +EXPORT_SYMBOL_GPL(spu_add_dev_attr); -int spu_add_sysdev_attr_group(struct attribute_group *attrs) +int spu_add_dev_attr_group(struct attribute_group *attrs) { struct spu *spu; int rc = 0; mutex_lock(&spu_full_list_mutex); list_for_each_entry(spu, &spu_full_list, full_list) { - rc = sysfs_create_group(&spu->sysdev.kobj, attrs); + rc = sysfs_create_group(&spu->dev.kobj, attrs); /* we're in trouble here, but try unwinding anyway */ if (rc) { @@ -542,7 +553,7 @@ int spu_add_sysdev_attr_group(struct attribute_group *attrs) list_for_each_entry_continue_reverse(spu, &spu_full_list, full_list) - sysfs_remove_group(&spu->sysdev.kobj, attrs); + sysfs_remove_group(&spu->dev.kobj, attrs); break; } } @@ -551,45 +562,45 @@ int spu_add_sysdev_attr_group(struct attribute_group *attrs) return rc; } -EXPORT_SYMBOL_GPL(spu_add_sysdev_attr_group); +EXPORT_SYMBOL_GPL(spu_add_dev_attr_group); -void spu_remove_sysdev_attr(struct sysdev_attribute *attr) +void spu_remove_dev_attr(struct device_attribute *attr) { struct spu *spu; mutex_lock(&spu_full_list_mutex); list_for_each_entry(spu, &spu_full_list, full_list) - sysdev_remove_file(&spu->sysdev, attr); + device_remove_file(&spu->dev, attr); mutex_unlock(&spu_full_list_mutex); } -EXPORT_SYMBOL_GPL(spu_remove_sysdev_attr); +EXPORT_SYMBOL_GPL(spu_remove_dev_attr); -void spu_remove_sysdev_attr_group(struct attribute_group *attrs) +void spu_remove_dev_attr_group(struct attribute_group *attrs) { struct spu *spu; mutex_lock(&spu_full_list_mutex); list_for_each_entry(spu, &spu_full_list, full_list) - sysfs_remove_group(&spu->sysdev.kobj, attrs); + sysfs_remove_group(&spu->dev.kobj, attrs); mutex_unlock(&spu_full_list_mutex); } -EXPORT_SYMBOL_GPL(spu_remove_sysdev_attr_group); +EXPORT_SYMBOL_GPL(spu_remove_dev_attr_group); -static int spu_create_sysdev(struct spu *spu) +static int spu_create_dev(struct spu *spu) { int ret; - spu->sysdev.id = spu->number; - spu->sysdev.cls = &spu_sysdev_class; - ret = sysdev_register(&spu->sysdev); + spu->dev.id = spu->number; + spu->dev.bus = &spu_subsys; + ret = device_register(&spu->dev); if (ret) { printk(KERN_ERR "Can't register SPU %d with sysfs\n", spu->number); return ret; } - sysfs_add_device_to_node(&spu->sysdev, spu->node); + sysfs_add_device_to_node(&spu->dev, spu->node); return 0; } @@ -625,7 +636,7 @@ static int __init create_spu(void *data) if (ret) goto out_destroy; - ret = spu_create_sysdev(spu); + ret = spu_create_dev(spu); if (ret) goto out_free_irqs; @@ -682,9 +693,10 @@ static unsigned long long spu_acct_time(struct spu *spu, } -static ssize_t spu_stat_show(struct sys_device *sysdev, char *buf) +static ssize_t spu_stat_show(struct device *dev, + struct device_attribute *attr, char *buf) { - struct spu *spu = container_of(sysdev, struct spu, sysdev); + struct spu *spu = container_of(dev, struct spu, dev); return sprintf(buf, "%s %llu %llu %llu %llu " "%llu %llu %llu %llu %llu %llu %llu %llu\n", @@ -703,7 +715,92 @@ static ssize_t spu_stat_show(struct sys_device *sysdev, char *buf) spu->stats.libassist); } -static SYSDEV_ATTR(stat, 0644, spu_stat_show, NULL); +static DEVICE_ATTR(stat, 0444, spu_stat_show, NULL); + +#ifdef CONFIG_KEXEC + +struct crash_spu_info { + struct spu *spu; + u32 saved_spu_runcntl_RW; + u32 saved_spu_status_R; + u32 saved_spu_npc_RW; + u64 saved_mfc_sr1_RW; + u64 saved_mfc_dar; + u64 saved_mfc_dsisr; +}; + +#define CRASH_NUM_SPUS 16 /* Enough for current hardware */ +static struct crash_spu_info crash_spu_info[CRASH_NUM_SPUS]; + +static void crash_kexec_stop_spus(void) +{ + struct spu *spu; + int i; + u64 tmp; + + for (i = 0; i < CRASH_NUM_SPUS; i++) { + if (!crash_spu_info[i].spu) + continue; + + spu = crash_spu_info[i].spu; + + crash_spu_info[i].saved_spu_runcntl_RW = + in_be32(&spu->problem->spu_runcntl_RW); + crash_spu_info[i].saved_spu_status_R = + in_be32(&spu->problem->spu_status_R); + crash_spu_info[i].saved_spu_npc_RW = + in_be32(&spu->problem->spu_npc_RW); + + crash_spu_info[i].saved_mfc_dar = spu_mfc_dar_get(spu); + crash_spu_info[i].saved_mfc_dsisr = spu_mfc_dsisr_get(spu); + tmp = spu_mfc_sr1_get(spu); + crash_spu_info[i].saved_mfc_sr1_RW = tmp; + + tmp &= ~MFC_STATE1_MASTER_RUN_CONTROL_MASK; + spu_mfc_sr1_set(spu, tmp); + + __delay(200); + } +} + +static void crash_register_spus(struct list_head *list) +{ + struct spu *spu; + int ret; + + list_for_each_entry(spu, list, full_list) { + if (WARN_ON(spu->number >= CRASH_NUM_SPUS)) + continue; + + crash_spu_info[spu->number].spu = spu; + } + + ret = crash_shutdown_register(&crash_kexec_stop_spus); + if (ret) + printk(KERN_ERR "Could not register SPU crash handler"); +} + +#else +static inline void crash_register_spus(struct list_head *list) +{ +} +#endif + +static void spu_shutdown(void) +{ + struct spu *spu; + + mutex_lock(&spu_full_list_mutex); + list_for_each_entry(spu, &spu_full_list, full_list) { + spu_free_irqs(spu); + spu_destroy_spu(spu); + } + mutex_unlock(&spu_full_list_mutex); +} + +static struct syscore_ops spu_syscore_ops = { + .shutdown = spu_shutdown, +}; static int __init init_spu_base(void) { @@ -717,8 +814,8 @@ static int __init init_spu_base(void) if (!spu_management_ops) goto out; - /* create sysdev class for spus */ - ret = sysdev_class_register(&spu_sysdev_class); + /* create system subsystem for spus */ + ret = subsys_system_register(&spu_subsys, NULL); if (ret) goto out; @@ -726,34 +823,26 @@ static int __init init_spu_base(void) if (ret < 0) { printk(KERN_WARNING "%s: Error initializing spus\n", - __FUNCTION__); - goto out_unregister_sysdev_class; + __func__); + goto out_unregister_subsys; } - if (ret > 0) { - /* - * We cannot put the forward declaration in - * <linux/linux_logo.h> because of conflicting session type - * conflicts for const and __initdata with different compiler - * versions - */ - extern const struct linux_logo logo_spe_clut224; - + if (ret > 0) fb_append_extra_logo(&logo_spe_clut224, ret); - } mutex_lock(&spu_full_list_mutex); xmon_register_spus(&spu_full_list); crash_register_spus(&spu_full_list); mutex_unlock(&spu_full_list_mutex); - spu_add_sysdev_attr(&attr_stat); + spu_add_dev_attr(&dev_attr_stat); + register_syscore_ops(&spu_syscore_ops); spu_init_affinity(); return 0; - out_unregister_sysdev_class: - sysdev_class_unregister(&spu_sysdev_class); + out_unregister_subsys: + bus_unregister(&spu_subsys); out: return ret; } diff --git a/arch/powerpc/platforms/cell/spu_callbacks.c b/arch/powerpc/platforms/cell/spu_callbacks.c index dceb8b6a938..b0ec78e8ad6 100644 --- a/arch/powerpc/platforms/cell/spu_callbacks.c +++ b/arch/powerpc/platforms/cell/spu_callbacks.c @@ -5,7 +5,7 @@ #undef DEBUG #include <linux/kallsyms.h> -#include <linux/module.h> +#include <linux/export.h> #include <linux/syscalls.h> #include <asm/spu.h> @@ -54,19 +54,18 @@ long spu_sys_callback(struct spu_syscall_block *s) long (*syscall)(u64 a1, u64 a2, u64 a3, u64 a4, u64 a5, u64 a6); if (s->nr_ret >= ARRAY_SIZE(spu_syscall_table)) { - pr_debug("%s: invalid syscall #%ld", __FUNCTION__, s->nr_ret); + pr_debug("%s: invalid syscall #%lld", __func__, s->nr_ret); return -ENOSYS; } syscall = spu_syscall_table[s->nr_ret]; -#ifdef DEBUG - print_symbol(KERN_DEBUG "SPU-syscall %s:", (unsigned long)syscall); - printk("syscall%ld(%lx, %lx, %lx, %lx, %lx, %lx)\n", - s->nr_ret, - s->parm[0], s->parm[1], s->parm[2], - s->parm[3], s->parm[4], s->parm[5]); -#endif + pr_debug("SPU-syscall " + "%pSR:syscall%lld(%llx, %llx, %llx, %llx, %llx, %llx)\n", + syscall, + s->nr_ret, + s->parm[0], s->parm[1], s->parm[2], + s->parm[3], s->parm[4], s->parm[5]); return syscall(s->parm[0], s->parm[1], s->parm[2], s->parm[3], s->parm[4], s->parm[5]); diff --git a/arch/powerpc/platforms/cell/spu_fault.c b/arch/powerpc/platforms/cell/spu_fault.c index c8b1cd42905..641e7273d75 100644 --- a/arch/powerpc/platforms/cell/spu_fault.c +++ b/arch/powerpc/platforms/cell/spu_fault.c @@ -22,7 +22,7 @@ */ #include <linux/sched.h> #include <linux/mm.h> -#include <linux/module.h> +#include <linux/export.h> #include <asm/spu.h> #include <asm/spu_csa.h> @@ -39,60 +39,56 @@ int spu_handle_mm_fault(struct mm_struct *mm, unsigned long ea, unsigned long is_write; int ret; -#if 0 - if (!IS_VALID_EA(ea)) { + if (mm == NULL) return -EFAULT; - } -#endif /* XXX */ - if (mm == NULL) { - return -EFAULT; - } - if (mm->pgd == NULL) { + + if (mm->pgd == NULL) return -EFAULT; - } down_read(&mm->mmap_sem); + ret = -EFAULT; vma = find_vma(mm, ea); if (!vma) - goto bad_area; - if (vma->vm_start <= ea) - goto good_area; - if (!(vma->vm_flags & VM_GROWSDOWN)) - goto bad_area; - if (expand_stack(vma, ea)) - goto bad_area; -good_area: + goto out_unlock; + + if (ea < vma->vm_start) { + if (!(vma->vm_flags & VM_GROWSDOWN)) + goto out_unlock; + if (expand_stack(vma, ea)) + goto out_unlock; + } + is_write = dsisr & MFC_DSISR_ACCESS_PUT; if (is_write) { if (!(vma->vm_flags & VM_WRITE)) - goto bad_area; + goto out_unlock; } else { if (dsisr & MFC_DSISR_ACCESS_DENIED) - goto bad_area; + goto out_unlock; if (!(vma->vm_flags & (VM_READ | VM_EXEC))) - goto bad_area; + goto out_unlock; } + ret = 0; - *flt = handle_mm_fault(mm, vma, ea, is_write); + *flt = handle_mm_fault(mm, vma, ea, is_write ? FAULT_FLAG_WRITE : 0); if (unlikely(*flt & VM_FAULT_ERROR)) { if (*flt & VM_FAULT_OOM) { ret = -ENOMEM; - goto bad_area; + goto out_unlock; } else if (*flt & VM_FAULT_SIGBUS) { ret = -EFAULT; - goto bad_area; + goto out_unlock; } BUG(); } + if (*flt & VM_FAULT_MAJOR) current->maj_flt++; else current->min_flt++; - up_read(&mm->mmap_sem); - return ret; -bad_area: +out_unlock: up_read(&mm->mmap_sem); - return -EFAULT; + return ret; } EXPORT_SYMBOL_GPL(spu_handle_mm_fault); diff --git a/arch/powerpc/platforms/cell/spu_manage.c b/arch/powerpc/platforms/cell/spu_manage.c index d351bdebf5f..c3327f3d8cf 100644 --- a/arch/powerpc/platforms/cell/spu_manage.c +++ b/arch/powerpc/platforms/cell/spu_manage.c @@ -21,9 +21,8 @@ #include <linux/interrupt.h> #include <linux/list.h> -#include <linux/module.h> +#include <linux/export.h> #include <linux/ptrace.h> -#include <linux/slab.h> #include <linux/wait.h> #include <linux/mm.h> #include <linux/io.h> @@ -92,7 +91,7 @@ static int __init spu_map_interrupts_old(struct spu *spu, tmp = of_get_property(np->parent->parent, "node-id", NULL); if (!tmp) { - printk(KERN_WARNING "%s: can't find node-id\n", __FUNCTION__); + printk(KERN_WARNING "%s: can't find node-id\n", __func__); nid = spu->node; } else nid = tmp[0]; @@ -178,21 +177,20 @@ out: static int __init spu_map_interrupts(struct spu *spu, struct device_node *np) { - struct of_irq oirq; + struct of_phandle_args oirq; int ret; int i; for (i=0; i < 3; i++) { - ret = of_irq_map_one(np, i, &oirq); + ret = of_irq_parse_one(np, i, &oirq); if (ret) { pr_debug("spu_new: failed to get irq %d\n", i); goto err; } ret = -EINVAL; - pr_debug(" irq %d no 0x%x on %s\n", i, oirq.specifier[0], - oirq.controller->full_name); - spu->irqs[i] = irq_create_of_mapping(oirq.controller, - oirq.specifier, oirq.size); + pr_debug(" irq %d no 0x%x on %s\n", i, oirq.args[0], + oirq.np->full_name); + spu->irqs[i] = irq_create_of_mapping(&oirq); if (spu->irqs[i] == NO_IRQ) { pr_debug("spu_new: failed to map it !\n"); goto err; @@ -201,7 +199,7 @@ static int __init spu_map_interrupts(struct spu *spu, struct device_node *np) return 0; err: - pr_debug("failed to map irq %x for spu %s\n", *oirq.specifier, + pr_debug("failed to map irq %x for spu %s\n", *oirq.args, spu->name); for (; i >= 0; i--) { if (spu->irqs[i] != NO_IRQ) @@ -223,7 +221,7 @@ static int spu_map_resource(struct spu *spu, int nr, return ret; if (phys) *phys = resource.start; - len = resource.end - resource.start + 1; + len = resource_size(&resource); *virt = ioremap(resource.start, len); if (!*virt) return -EINVAL; @@ -296,7 +294,7 @@ static int __init of_enumerate_spus(int (*fn)(void *data)) ret = fn(node); if (ret) { printk(KERN_WARNING "%s: Error initializing %s\n", - __FUNCTION__, node->name); + __func__, node->name); break; } n++; @@ -327,7 +325,7 @@ static int __init of_create_spu(struct spu *spu, void *data) if (!legacy_map) { legacy_map = 1; printk(KERN_WARNING "%s: Legacy device tree found, " - "trying to map old style\n", __FUNCTION__); + "trying to map old style\n", __func__); } ret = spu_map_device_old(spu); if (ret) { @@ -342,7 +340,7 @@ static int __init of_create_spu(struct spu *spu, void *data) if (!legacy_irq) { legacy_irq = 1; printk(KERN_WARNING "%s: Legacy device tree found, " - "trying old style irq\n", __FUNCTION__); + "trying old style irq\n", __func__); } ret = spu_map_interrupts_old(spu, spe); if (ret) { @@ -457,7 +455,7 @@ neighbour_spu(int cbe, struct device_node *target, struct device_node *avoid) continue; vic_handles = of_get_property(spu_dn, "vicinity", &lenp); for (i=0; i < (lenp / sizeof(phandle)); i++) { - if (vic_handles[i] == target->linux_phandle) + if (vic_handles[i] == target->phandle) return spu; } } @@ -499,7 +497,7 @@ static void init_affinity_node(int cbe) if (strcmp(name, "spe") == 0) { spu = devnode_spu(cbe, vic_dn); - avoid_ph = last_spu_dn->linux_phandle; + avoid_ph = last_spu_dn->phandle; } else { /* * "mic-tm" and "bif0" nodes do not have @@ -514,7 +512,7 @@ static void init_affinity_node(int cbe) last_spu->has_mem_affinity = 1; spu->has_mem_affinity = 1; } - avoid_ph = vic_dn->linux_phandle; + avoid_ph = vic_dn->phandle; } list_add_tail(&spu->aff_list, &last_spu->aff_list); diff --git a/arch/powerpc/platforms/cell/spu_notify.c b/arch/powerpc/platforms/cell/spu_notify.c index 34d156959f3..afdf857c318 100644 --- a/arch/powerpc/platforms/cell/spu_notify.c +++ b/arch/powerpc/platforms/cell/spu_notify.c @@ -21,7 +21,8 @@ #undef DEBUG -#include <linux/module.h> +#include <linux/export.h> +#include <linux/notifier.h> #include <asm/spu.h> #include "spufs/spufs.h" diff --git a/arch/powerpc/platforms/cell/spu_priv1_mmio.c b/arch/powerpc/platforms/cell/spu_priv1_mmio.c index 67fa7247b80..66d33724f16 100644 --- a/arch/powerpc/platforms/cell/spu_priv1_mmio.c +++ b/arch/powerpc/platforms/cell/spu_priv1_mmio.c @@ -20,14 +20,13 @@ #include <linux/interrupt.h> #include <linux/list.h> -#include <linux/module.h> #include <linux/ptrace.h> -#include <linux/slab.h> #include <linux/wait.h> #include <linux/mm.h> #include <linux/io.h> #include <linux/mutex.h> #include <linux/device.h> +#include <linux/sched.h> #include <asm/spu.h> #include <asm/spu_priv1.h> @@ -75,8 +74,19 @@ static u64 int_stat_get(struct spu *spu, int class) static void cpu_affinity_set(struct spu *spu, int cpu) { - u64 target = iic_get_target_id(cpu); - u64 route = target << 48 | target << 32 | target << 16; + u64 target; + u64 route; + + if (nr_cpus_node(spu->node)) { + const struct cpumask *spumask = cpumask_of_node(spu->node), + *cpumask = cpumask_of_node(cpu_to_node(cpu)); + + if (!cpumask_intersects(spumask, cpumask)) + return; + } + + target = iic_get_target_id(cpu); + route = target << 48 | target << 32 | target << 16; out_be64(&spu->priv1->int_route_RW, route); } diff --git a/arch/powerpc/platforms/cell/spu_syscalls.c b/arch/powerpc/platforms/cell/spu_syscalls.c index 75530d99eda..5e6e0bad6db 100644 --- a/arch/powerpc/platforms/cell/spu_syscalls.c +++ b/arch/powerpc/platforms/cell/spu_syscalls.c @@ -25,6 +25,7 @@ #include <linux/module.h> #include <linux/syscalls.h> #include <linux/rcupdate.h> +#include <linux/binfmts.h> #include <asm/spu.h> @@ -65,12 +66,10 @@ static inline void spufs_calls_put(struct spufs_calls *calls) { } #endif /* CONFIG_SPU_FS_MODULE */ -asmlinkage long sys_spu_create(const char __user *name, - unsigned int flags, mode_t mode, int neighbor_fd) +SYSCALL_DEFINE4(spu_create, const char __user *, name, unsigned int, flags, + umode_t, mode, int, neighbor_fd) { long ret; - struct file *neighbor; - int fput_needed; struct spufs_calls *calls; calls = spufs_calls_get(); @@ -78,11 +77,11 @@ asmlinkage long sys_spu_create(const char __user *name, return -ENOSYS; if (flags & SPU_CREATE_AFFINITY_SPU) { + struct fd neighbor = fdget(neighbor_fd); ret = -EBADF; - neighbor = fget_light(neighbor_fd, &fput_needed); - if (neighbor) { - ret = calls->create_thread(name, flags, mode, neighbor); - fput_light(neighbor, fput_needed); + if (neighbor.file) { + ret = calls->create_thread(name, flags, mode, neighbor.file); + fdput(neighbor); } } else ret = calls->create_thread(name, flags, mode, NULL); @@ -94,8 +93,7 @@ asmlinkage long sys_spu_create(const char __user *name, asmlinkage long sys_spu_run(int fd, __u32 __user *unpc, __u32 __user *ustatus) { long ret; - struct file *filp; - int fput_needed; + struct fd arg; struct spufs_calls *calls; calls = spufs_calls_get(); @@ -103,16 +101,17 @@ asmlinkage long sys_spu_run(int fd, __u32 __user *unpc, __u32 __user *ustatus) return -ENOSYS; ret = -EBADF; - filp = fget_light(fd, &fput_needed); - if (filp) { - ret = calls->spu_run(filp, unpc, ustatus); - fput_light(filp, fput_needed); + arg = fdget(fd); + if (arg.file) { + ret = calls->spu_run(arg.file, unpc, ustatus); + fdput(arg); } spufs_calls_put(calls); return ret; } +#ifdef CONFIG_COREDUMP int elf_coredump_extra_notes_size(void) { struct spufs_calls *calls; @@ -129,7 +128,7 @@ int elf_coredump_extra_notes_size(void) return ret; } -int elf_coredump_extra_notes_write(struct file *file, loff_t *foffset) +int elf_coredump_extra_notes_write(struct coredump_params *cprm) { struct spufs_calls *calls; int ret; @@ -138,12 +137,13 @@ int elf_coredump_extra_notes_write(struct file *file, loff_t *foffset) if (!calls) return 0; - ret = calls->coredump_extra_notes_write(file, foffset); + ret = calls->coredump_extra_notes_write(cprm); spufs_calls_put(calls); return ret; } +#endif void notify_spus_active(void) { @@ -172,7 +172,7 @@ EXPORT_SYMBOL_GPL(register_spu_syscalls); void unregister_spu_syscalls(struct spufs_calls *calls) { BUG_ON(spufs_calls->owner != calls->owner); - rcu_assign_pointer(spufs_calls, NULL); + RCU_INIT_POINTER(spufs_calls, NULL); synchronize_rcu(); } EXPORT_SYMBOL_GPL(unregister_spu_syscalls); diff --git a/arch/powerpc/platforms/cell/spufs/.gitignore b/arch/powerpc/platforms/cell/spufs/.gitignore new file mode 100644 index 00000000000..a09ee8d84d6 --- /dev/null +++ b/arch/powerpc/platforms/cell/spufs/.gitignore @@ -0,0 +1,2 @@ +spu_save_dump.h +spu_restore_dump.h diff --git a/arch/powerpc/platforms/cell/spufs/Makefile b/arch/powerpc/platforms/cell/spufs/Makefile index 99610a6361f..52a7d2596d3 100644 --- a/arch/powerpc/platforms/cell/spufs/Makefile +++ b/arch/powerpc/platforms/cell/spufs/Makefile @@ -1,10 +1,12 @@ obj-$(CONFIG_SPU_FS) += spufs.o -spufs-y += inode.o file.o context.o syscalls.o coredump.o +spufs-y += inode.o file.o context.o syscalls.o spufs-y += sched.o backing_ops.o hw_ops.o run.o gang.o spufs-y += switch.o fault.o lscsa_alloc.o +spufs-$(CONFIG_COREDUMP) += coredump.o -obj-$(CONFIG_SPU_TRACE) += sputrace.o +# magic for the trace events +CFLAGS_sched.o := -I$(src) # Rules to build switch.o with the help of SPU tool chain SPU_CROSS := spu- @@ -12,10 +14,8 @@ SPU_CC := $(SPU_CROSS)gcc SPU_AS := $(SPU_CROSS)gcc SPU_LD := $(SPU_CROSS)ld SPU_OBJCOPY := $(SPU_CROSS)objcopy -SPU_CFLAGS := -O2 -Wall -I$(srctree)/include \ - -I$(objtree)/include2 -D__KERNEL__ -SPU_AFLAGS := -c -D__ASSEMBLY__ -I$(srctree)/include \ - -I$(objtree)/include2 -D__KERNEL__ +SPU_CFLAGS := -O2 -Wall -I$(srctree)/include -D__KERNEL__ +SPU_AFLAGS := -c -D__ASSEMBLY__ -I$(srctree)/include -D__KERNEL__ SPU_LDFLAGS := -N -Ttext=0x0 $(obj)/switch.o: $(obj)/spu_save_dump.h $(obj)/spu_restore_dump.h diff --git a/arch/powerpc/platforms/cell/spufs/backing_ops.c b/arch/powerpc/platforms/cell/spufs/backing_ops.c index 64eb15b2204..6e8a9ef8590 100644 --- a/arch/powerpc/platforms/cell/spufs/backing_ops.c +++ b/arch/powerpc/platforms/cell/spufs/backing_ops.c @@ -21,7 +21,6 @@ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ -#include <linux/module.h> #include <linux/errno.h> #include <linux/sched.h> #include <linux/kernel.h> diff --git a/arch/powerpc/platforms/cell/spufs/context.c b/arch/powerpc/platforms/cell/spufs/context.c index 0ad83aeb70b..9c6790d17ed 100644 --- a/arch/powerpc/platforms/cell/spufs/context.c +++ b/arch/powerpc/platforms/cell/spufs/context.c @@ -22,12 +22,13 @@ #include <linux/fs.h> #include <linux/mm.h> -#include <linux/module.h> #include <linux/slab.h> -#include <asm/atomic.h> +#include <linux/atomic.h> +#include <linux/sched.h> #include <asm/spu.h> #include <asm/spu_csa.h> #include "spufs.h" +#include "sputrace.h" atomic_t nr_spu_contexts = ATOMIC_INIT(0); @@ -35,6 +36,8 @@ atomic_t nr_spu_contexts = ATOMIC_INIT(0); struct spu_context *alloc_spu_context(struct spu_gang *gang) { struct spu_context *ctx; + struct timespec ts; + ctx = kzalloc(sizeof *ctx, GFP_KERNEL); if (!ctx) goto out; @@ -64,6 +67,8 @@ struct spu_context *alloc_spu_context(struct spu_gang *gang) __spu_update_sched_info(ctx); spu_set_timeslice(ctx); ctx->stats.util_state = SPU_UTIL_IDLE_LOADED; + ktime_get_ts(&ts); + ctx->stats.tstamp = timespec_to_ns(&ts); atomic_inc(&nr_spu_contexts); goto out; @@ -78,6 +83,7 @@ void destroy_spu_context(struct kref *kref) { struct spu_context *ctx; ctx = container_of(kref, struct spu_context, kref); + spu_context_nospu_trace(destroy_spu_context__enter, ctx); mutex_lock(&ctx->state_mutex); spu_deactivate(ctx); mutex_unlock(&ctx->state_mutex); @@ -88,6 +94,7 @@ void destroy_spu_context(struct kref *kref) kref_put(ctx->prof_priv_kref, ctx->prof_priv_release); BUG_ON(!list_empty(&ctx->rq)); atomic_dec(&nr_spu_contexts); + kfree(ctx->switch_log); kfree(ctx); } @@ -128,17 +135,17 @@ void spu_unmap_mappings(struct spu_context *ctx) if (ctx->local_store) unmap_mapping_range(ctx->local_store, 0, LS_SIZE, 1); if (ctx->mfc) - unmap_mapping_range(ctx->mfc, 0, 0x1000, 1); + unmap_mapping_range(ctx->mfc, 0, SPUFS_MFC_MAP_SIZE, 1); if (ctx->cntl) - unmap_mapping_range(ctx->cntl, 0, 0x1000, 1); + unmap_mapping_range(ctx->cntl, 0, SPUFS_CNTL_MAP_SIZE, 1); if (ctx->signal1) - unmap_mapping_range(ctx->signal1, 0, PAGE_SIZE, 1); + unmap_mapping_range(ctx->signal1, 0, SPUFS_SIGNAL_MAP_SIZE, 1); if (ctx->signal2) - unmap_mapping_range(ctx->signal2, 0, PAGE_SIZE, 1); + unmap_mapping_range(ctx->signal2, 0, SPUFS_SIGNAL_MAP_SIZE, 1); if (ctx->mss) - unmap_mapping_range(ctx->mss, 0, 0x1000, 1); + unmap_mapping_range(ctx->mss, 0, SPUFS_MSS_MAP_SIZE, 1); if (ctx->psmap) - unmap_mapping_range(ctx->psmap, 0, 0x20000, 1); + unmap_mapping_range(ctx->psmap, 0, SPUFS_PS_MAP_SIZE, 1); mutex_unlock(&ctx->mapping_lock); } @@ -150,6 +157,8 @@ int spu_acquire_saved(struct spu_context *ctx) { int ret; + spu_context_nospu_trace(spu_acquire_saved__enter, ctx); + ret = spu_acquire(ctx); if (ret) return ret; diff --git a/arch/powerpc/platforms/cell/spufs/coredump.c b/arch/powerpc/platforms/cell/spufs/coredump.c index 0c6a96b82b2..be6212ddbf0 100644 --- a/arch/powerpc/platforms/cell/spufs/coredump.c +++ b/arch/powerpc/platforms/cell/spufs/coredump.c @@ -22,10 +22,13 @@ #include <linux/elf.h> #include <linux/file.h> +#include <linux/fdtable.h> #include <linux/fs.h> +#include <linux/gfp.h> #include <linux/list.h> -#include <linux/module.h> #include <linux/syscalls.h> +#include <linux/coredump.h> +#include <linux/binfmts.h> #include <asm/uaccess.h> @@ -41,50 +44,12 @@ static ssize_t do_coredump_read(int num, struct spu_context *ctx, void *buffer, return spufs_coredump_read[num].read(ctx, buffer, size, off); data = spufs_coredump_read[num].get(ctx); - ret = snprintf(buffer, size, "0x%.16lx", data); + ret = snprintf(buffer, size, "0x%.16llx", data); if (ret >= size) return size; return ++ret; /* count trailing NULL */ } -/* - * These are the only things you should do on a core-file: use only these - * functions to write out all the necessary info. - */ -static int spufs_dump_write(struct file *file, const void *addr, int nr, loff_t *foffset) -{ - unsigned long limit = current->signal->rlim[RLIMIT_CORE].rlim_cur; - ssize_t written; - - if (*foffset + nr > limit) - return -EIO; - - written = file->f_op->write(file, addr, nr, &file->f_pos); - *foffset += written; - - if (written != nr) - return -EIO; - - return 0; -} - -static int spufs_dump_align(struct file *file, char *buf, loff_t new_off, - loff_t *foffset) -{ - int rc, size; - - size = min((loff_t)PAGE_SIZE, new_off - *foffset); - memset(buf, 0, size); - - rc = 0; - while (rc == 0 && new_off > *foffset) { - size = min((loff_t)PAGE_SIZE, new_off - *foffset); - rc = spufs_dump_write(file, buf, size, foffset); - } - - return rc; -} - static int spufs_ctx_note_size(struct spu_context *ctx, int dfd) { int i, sz, total = 0; @@ -105,6 +70,17 @@ static int spufs_ctx_note_size(struct spu_context *ctx, int dfd) return total; } +static int match_context(const void *v, struct file *file, unsigned fd) +{ + struct spu_context *ctx; + if (file->f_op != &spufs_context_fops) + return 0; + ctx = SPUFS_I(file_inode(file))->i_ctx; + if (ctx->flags & SPU_CREATE_NOSCHED) + return 0; + return fd + 1; +} + /* * The additional architecture-specific notes for Cell are various * context files in the spu context. @@ -114,31 +90,18 @@ static int spufs_ctx_note_size(struct spu_context *ctx, int dfd) * internal functionality to dump them without needing to actually * open the files. */ +/* + * descriptor table is not shared, so files can't change or go away. + */ static struct spu_context *coredump_next_context(int *fd) { - struct fdtable *fdt = files_fdtable(current->files); struct file *file; - struct spu_context *ctx = NULL; - - for (; *fd < fdt->max_fds; (*fd)++) { - if (!FD_ISSET(*fd, fdt->open_fds)) - continue; - - file = fcheck(*fd); - - if (!file || file->f_op != &spufs_context_fops) - continue; - - ctx = SPUFS_I(file->f_dentry->d_inode)->i_ctx; - if (ctx->flags & SPU_CREATE_NOSCHED) - continue; - - /* start searching the next fd next time we're called */ - (*fd)++; - break; - } - - return ctx; + int n = iterate_fd(current->files, *fd, match_context, NULL); + if (!n) + return NULL; + *fd = n - 1; + file = fcheck(*fd); + return SPUFS_I(file_inode(file))->i_ctx; } int spufs_coredump_extra_notes_size(void) @@ -157,16 +120,19 @@ int spufs_coredump_extra_notes_size(void) break; size += rc; + + /* start searching the next fd next time */ + fd++; } return size; } static int spufs_arch_write_note(struct spu_context *ctx, int i, - struct file *file, int dfd, loff_t *foffset) + struct coredump_params *cprm, int dfd) { loff_t pos = 0; - int sz, rc, nread, total = 0; + int sz, rc, total = 0; const int bufsz = PAGE_SIZE; char *name; char fullname[80], *buf; @@ -184,42 +150,39 @@ static int spufs_arch_write_note(struct spu_context *ctx, int i, en.n_descsz = sz; en.n_type = NT_SPU; - rc = spufs_dump_write(file, &en, sizeof(en), foffset); - if (rc) - goto out; + if (!dump_emit(cprm, &en, sizeof(en))) + goto Eio; - rc = spufs_dump_write(file, fullname, en.n_namesz, foffset); - if (rc) - goto out; + if (!dump_emit(cprm, fullname, en.n_namesz)) + goto Eio; - rc = spufs_dump_align(file, buf, roundup(*foffset, 4), foffset); - if (rc) - goto out; + if (!dump_align(cprm, 4)) + goto Eio; do { - nread = do_coredump_read(i, ctx, buf, bufsz, &pos); - if (nread > 0) { - rc = spufs_dump_write(file, buf, nread, foffset); - if (rc) - goto out; - total += nread; + rc = do_coredump_read(i, ctx, buf, bufsz, &pos); + if (rc > 0) { + if (!dump_emit(cprm, buf, rc)) + goto Eio; + total += rc; } - } while (nread == bufsz && total < sz); + } while (rc == bufsz && total < sz); - if (nread < 0) { - rc = nread; + if (rc < 0) goto out; - } - - rc = spufs_dump_align(file, buf, roundup(*foffset - total + sz, 4), - foffset); + if (!dump_skip(cprm, + roundup(cprm->written - total + sz, 4) - cprm->written)) + goto Eio; out: free_page((unsigned long)buf); return rc; +Eio: + free_page((unsigned long)buf); + return -EIO; } -int spufs_coredump_extra_notes_write(struct file *file, loff_t *foffset) +int spufs_coredump_extra_notes_write(struct coredump_params *cprm) { struct spu_context *ctx; int fd, j, rc; @@ -231,7 +194,7 @@ int spufs_coredump_extra_notes_write(struct file *file, loff_t *foffset) return rc; for (j = 0; spufs_coredump_read[j].name != NULL; j++) { - rc = spufs_arch_write_note(ctx, j, file, fd, foffset); + rc = spufs_arch_write_note(ctx, j, cprm, fd); if (rc) { spu_release_saved(ctx); return rc; @@ -239,6 +202,9 @@ int spufs_coredump_extra_notes_write(struct file *file, loff_t *foffset) } spu_release_saved(ctx); + + /* start searching the next fd next time */ + fd++; } return 0; diff --git a/arch/powerpc/platforms/cell/spufs/fault.c b/arch/powerpc/platforms/cell/spufs/fault.c index e46d300e21a..8cb6260cc80 100644 --- a/arch/powerpc/platforms/cell/spufs/fault.c +++ b/arch/powerpc/platforms/cell/spufs/fault.c @@ -21,7 +21,6 @@ */ #include <linux/sched.h> #include <linux/mm.h> -#include <linux/module.h> #include <asm/spu.h> #include <asm/spu_csa.h> @@ -83,13 +82,18 @@ int spufs_handle_class0(struct spu_context *ctx) return 0; if (stat & CLASS0_DMA_ALIGNMENT_INTR) - spufs_handle_event(ctx, ctx->csa.dar, SPE_EVENT_DMA_ALIGNMENT); + spufs_handle_event(ctx, ctx->csa.class_0_dar, + SPE_EVENT_DMA_ALIGNMENT); if (stat & CLASS0_INVALID_DMA_COMMAND_INTR) - spufs_handle_event(ctx, ctx->csa.dar, SPE_EVENT_INVALID_DMA); + spufs_handle_event(ctx, ctx->csa.class_0_dar, + SPE_EVENT_INVALID_DMA); if (stat & CLASS0_SPU_ERROR_INTR) - spufs_handle_event(ctx, ctx->csa.dar, SPE_EVENT_SPE_ERROR); + spufs_handle_event(ctx, ctx->csa.class_0_dar, + SPE_EVENT_SPE_ERROR); + + ctx->csa.class_0_pending = 0; return -EIO; } @@ -119,15 +123,15 @@ int spufs_handle_class1(struct spu_context *ctx) * in time, we can still expect to get the same fault * the immediately after the context restore. */ - ea = ctx->csa.dar; - dsisr = ctx->csa.dsisr; + ea = ctx->csa.class_1_dar; + dsisr = ctx->csa.class_1_dsisr; if (!(dsisr & (MFC_DSISR_PTE_NOT_FOUND | MFC_DSISR_ACCESS_DENIED))) return 0; spuctx_switch_state(ctx, SPU_UTIL_IOWAIT); - pr_debug("ctx %p: ea %016lx, dsisr %016lx state %d\n", ctx, ea, + pr_debug("ctx %p: ea %016llx, dsisr %016llx state %d\n", ctx, ea, dsisr, ctx->state); ctx->stats.hash_flt++; @@ -158,7 +162,7 @@ int spufs_handle_class1(struct spu_context *ctx) * time slicing will not preempt the context while the page fault * handler is running. Context switch code removes mappings. */ - ctx->csa.dar = ctx->csa.dsisr = 0; + ctx->csa.class_1_dar = ctx->csa.class_1_dsisr = 0; /* * If we handled the fault successfully and are in runnable diff --git a/arch/powerpc/platforms/cell/spufs/file.c b/arch/powerpc/platforms/cell/spufs/file.c index f7a7e8635fb..90986923a53 100644 --- a/arch/powerpc/platforms/cell/spufs/file.c +++ b/arch/powerpc/platforms/cell/spufs/file.c @@ -24,20 +24,21 @@ #include <linux/fs.h> #include <linux/ioctl.h> -#include <linux/module.h> +#include <linux/export.h> #include <linux/pagemap.h> #include <linux/poll.h> #include <linux/ptrace.h> #include <linux/seq_file.h> -#include <linux/marker.h> +#include <linux/slab.h> #include <asm/io.h> -#include <asm/semaphore.h> +#include <asm/time.h> #include <asm/spu.h> #include <asm/spu_info.h> #include <asm/uaccess.h> #include "spufs.h" +#include "sputrace.h" #define SPUFS_MMAP_4K (PAGE_SIZE == 0x1000) @@ -147,12 +148,12 @@ static int __fops ## _open(struct inode *inode, struct file *file) \ __simple_attr_check_format(__fmt, 0ull); \ return spufs_attr_open(inode, file, __get, __set, __fmt); \ } \ -static struct file_operations __fops = { \ - .owner = THIS_MODULE, \ +static const struct file_operations __fops = { \ .open = __fops ## _open, \ .release = spufs_attr_release, \ .read = spufs_attr_read, \ .write = spufs_attr_write, \ + .llseek = generic_file_llseek, \ }; @@ -217,32 +218,27 @@ spufs_mem_write(struct file *file, const char __user *buffer, loff_t pos = *ppos; int ret; - if (pos < 0) - return -EINVAL; if (pos > LS_SIZE) return -EFBIG; - if (size > LS_SIZE - pos) - size = LS_SIZE - pos; ret = spu_acquire(ctx); if (ret) return ret; local_store = ctx->ops->get_ls(ctx); - ret = copy_from_user(local_store + pos, buffer, size); + size = simple_write_to_buffer(local_store, LS_SIZE, ppos, buffer, size); spu_release(ctx); - if (ret) - return -EFAULT; - *ppos = pos + size; return size; } -static unsigned long spufs_mem_mmap_nopfn(struct vm_area_struct *vma, - unsigned long address) +static int +spufs_mem_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { struct spu_context *ctx = vma->vm_file->private_data; - unsigned long pfn, offset, addr0 = address; + unsigned long address = (unsigned long)vmf->virtual_address; + unsigned long pfn, offset; + #ifdef CONFIG_SPU_FS_64K_LS struct spu_state *csa = &ctx->csa; int psize; @@ -260,35 +256,56 @@ static unsigned long spufs_mem_mmap_nopfn(struct vm_area_struct *vma, } #endif /* CONFIG_SPU_FS_64K_LS */ - offset = (address - vma->vm_start) + (vma->vm_pgoff << PAGE_SHIFT); + offset = vmf->pgoff << PAGE_SHIFT; if (offset >= LS_SIZE) - return NOPFN_SIGBUS; + return VM_FAULT_SIGBUS; - pr_debug("spufs_mem_mmap_nopfn address=0x%lx -> 0x%lx, offset=0x%lx\n", - addr0, address, offset); + pr_debug("spufs_mem_mmap_fault address=0x%lx, offset=0x%lx\n", + address, offset); if (spu_acquire(ctx)) - return NOPFN_REFAULT; + return VM_FAULT_NOPAGE; if (ctx->state == SPU_STATE_SAVED) { - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - & ~_PAGE_NO_CACHE); + vma->vm_page_prot = pgprot_cached(vma->vm_page_prot); pfn = vmalloc_to_pfn(ctx->csa.lscsa->ls + offset); } else { - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE); + vma->vm_page_prot = pgprot_noncached_wc(vma->vm_page_prot); pfn = (ctx->spu->local_store_phys + offset) >> PAGE_SHIFT; } vm_insert_pfn(vma, address, pfn); spu_release(ctx); - return NOPFN_REFAULT; + return VM_FAULT_NOPAGE; } +static int spufs_mem_mmap_access(struct vm_area_struct *vma, + unsigned long address, + void *buf, int len, int write) +{ + struct spu_context *ctx = vma->vm_file->private_data; + unsigned long offset = address - vma->vm_start; + char *local_store; + + if (write && !(vma->vm_flags & VM_WRITE)) + return -EACCES; + if (spu_acquire(ctx)) + return -EINTR; + if ((offset + len) > vma->vm_end) + len = vma->vm_end - offset; + local_store = ctx->ops->get_ls(ctx); + if (write) + memcpy_toio(local_store + offset, buf, len); + else + memcpy_fromio(buf, local_store + offset, len); + spu_release(ctx); + return len; +} -static struct vm_operations_struct spufs_mem_mmap_vmops = { - .nopfn = spufs_mem_mmap_nopfn, +static const struct vm_operations_struct spufs_mem_mmap_vmops = { + .fault = spufs_mem_mmap_fault, + .access = spufs_mem_mmap_access, }; static int spufs_mem_mmap(struct file *file, struct vm_area_struct *vma) @@ -313,8 +330,7 @@ static int spufs_mem_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; vma->vm_flags |= VM_IO | VM_PFNMAP; - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE); + vma->vm_page_prot = pgprot_noncached_wc(vma->vm_page_prot); vma->vm_ops = &spufs_mem_mmap_vmops; return 0; @@ -335,7 +351,7 @@ static unsigned long spufs_get_unmapped_area(struct file *file, /* Else, try to obtain a 64K pages slice */ return slice_get_unmapped_area(addr, len, flags, - MMU_PAGE_64K, 1, 0); + MMU_PAGE_64K, 1); } #endif /* CONFIG_SPU_FS_64K_LS */ @@ -351,20 +367,22 @@ static const struct file_operations spufs_mem_fops = { #endif }; -static unsigned long spufs_ps_nopfn(struct vm_area_struct *vma, - unsigned long address, +static int spufs_ps_fault(struct vm_area_struct *vma, + struct vm_fault *vmf, unsigned long ps_offs, unsigned long ps_size) { struct spu_context *ctx = vma->vm_file->private_data; - unsigned long area, offset = address - vma->vm_start; + unsigned long area, offset = vmf->pgoff << PAGE_SHIFT; int ret = 0; - spu_context_nospu_trace(spufs_ps_nopfn__enter, ctx); + spu_context_nospu_trace(spufs_ps_fault__enter, ctx); - offset += vma->vm_pgoff << PAGE_SHIFT; if (offset >= ps_size) - return NOPFN_SIGBUS; + return VM_FAULT_SIGBUS; + + if (fatal_signal_pending(current)) + return VM_FAULT_SIGBUS; /* * Because we release the mmap_sem, the context may be destroyed while @@ -378,7 +396,7 @@ static unsigned long spufs_ps_nopfn(struct vm_area_struct *vma, * pages to hand out to the user, but we don't want to wait * with the mmap_sem held. * It is possible to drop the mmap_sem here, but then we need - * to return NOPFN_REFAULT because the mappings may have + * to return VM_FAULT_NOPAGE because the mappings may have * hanged. */ if (spu_acquire(ctx)) @@ -386,14 +404,15 @@ static unsigned long spufs_ps_nopfn(struct vm_area_struct *vma, if (ctx->state == SPU_STATE_SAVED) { up_read(¤t->mm->mmap_sem); - spu_context_nospu_trace(spufs_ps_nopfn__sleep, ctx); + spu_context_nospu_trace(spufs_ps_fault__sleep, ctx); ret = spufs_wait(ctx->run_wq, ctx->state == SPU_STATE_RUNNABLE); - spu_context_trace(spufs_ps_nopfn__wake, ctx, ctx->spu); + spu_context_trace(spufs_ps_fault__wake, ctx, ctx->spu); down_read(¤t->mm->mmap_sem); } else { area = ctx->spu->problem_phys + ps_offs; - vm_insert_pfn(vma, address, (area + offset) >> PAGE_SHIFT); - spu_context_trace(spufs_ps_nopfn__insert, ctx, ctx->spu); + vm_insert_pfn(vma, (unsigned long)vmf->virtual_address, + (area + offset) >> PAGE_SHIFT); + spu_context_trace(spufs_ps_fault__insert, ctx, ctx->spu); } if (!ret) @@ -401,18 +420,18 @@ static unsigned long spufs_ps_nopfn(struct vm_area_struct *vma, refault: put_spu_context(ctx); - return NOPFN_REFAULT; + return VM_FAULT_NOPAGE; } #if SPUFS_MMAP_4K -static unsigned long spufs_cntl_mmap_nopfn(struct vm_area_struct *vma, - unsigned long address) +static int spufs_cntl_mmap_fault(struct vm_area_struct *vma, + struct vm_fault *vmf) { - return spufs_ps_nopfn(vma, address, 0x4000, 0x1000); + return spufs_ps_fault(vma, vmf, 0x4000, SPUFS_CNTL_MAP_SIZE); } -static struct vm_operations_struct spufs_cntl_mmap_vmops = { - .nopfn = spufs_cntl_mmap_nopfn, +static const struct vm_operations_struct spufs_cntl_mmap_vmops = { + .fault = spufs_cntl_mmap_fault, }; /* @@ -424,8 +443,7 @@ static int spufs_cntl_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; vma->vm_flags |= VM_IO | VM_PFNMAP; - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE | _PAGE_GUARDED); + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); vma->vm_ops = &spufs_cntl_mmap_vmops; return 0; @@ -496,6 +514,7 @@ static const struct file_operations spufs_cntl_fops = { .release = spufs_cntl_release, .read = simple_attr_read, .write = simple_attr_write, + .llseek = generic_file_llseek, .mmap = spufs_cntl_mmap, }; @@ -523,6 +542,11 @@ spufs_regs_read(struct file *file, char __user *buffer, int ret; struct spu_context *ctx = file->private_data; + /* pre-check for file position: if we'd return EOF, there's no point + * causing a deschedule */ + if (*pos >= sizeof(ctx->csa.lscsa->gprs)) + return 0; + ret = spu_acquire_saved(ctx); if (ret) return ret; @@ -539,20 +563,18 @@ spufs_regs_write(struct file *file, const char __user *buffer, struct spu_lscsa *lscsa = ctx->csa.lscsa; int ret; - size = min_t(ssize_t, sizeof lscsa->gprs - *pos, size); - if (size <= 0) + if (*pos >= sizeof(lscsa->gprs)) return -EFBIG; - *pos += size; ret = spu_acquire_saved(ctx); if (ret) return ret; - ret = copy_from_user(lscsa->gprs + *pos - size, - buffer, size) ? -EFAULT : size; + size = simple_write_to_buffer(lscsa->gprs, sizeof(lscsa->gprs), pos, + buffer, size); spu_release_saved(ctx); - return ret; + return size; } static const struct file_operations spufs_regs_fops = { @@ -594,20 +616,18 @@ spufs_fpcr_write(struct file *file, const char __user * buffer, struct spu_lscsa *lscsa = ctx->csa.lscsa; int ret; - size = min_t(ssize_t, sizeof(lscsa->fpcr) - *pos, size); - if (size <= 0) + if (*pos >= sizeof(lscsa->fpcr)) return -EFBIG; ret = spu_acquire_saved(ctx); if (ret) return ret; - *pos += size; - ret = copy_from_user((char *)&lscsa->fpcr + *pos - size, - buffer, size) ? -EFAULT : size; + size = simple_write_to_buffer(&lscsa->fpcr, sizeof(lscsa->fpcr), pos, + buffer, size); spu_release_saved(ctx); - return ret; + return size; } static const struct file_operations spufs_fpcr_fops = { @@ -682,6 +702,7 @@ static ssize_t spufs_mbox_read(struct file *file, char __user *buf, static const struct file_operations spufs_mbox_fops = { .open = spufs_pipe_open, .read = spufs_mbox_read, + .llseek = no_llseek, }; static ssize_t spufs_mbox_stat_read(struct file *file, char __user *buf, @@ -711,6 +732,7 @@ static ssize_t spufs_mbox_stat_read(struct file *file, char __user *buf, static const struct file_operations spufs_mbox_stat_fops = { .open = spufs_pipe_open, .read = spufs_mbox_stat_read, + .llseek = no_llseek, }; /* low-level ibox access function */ @@ -831,6 +853,7 @@ static const struct file_operations spufs_ibox_fops = { .read = spufs_ibox_read, .poll = spufs_ibox_poll, .fasync = spufs_ibox_fasync, + .llseek = no_llseek, }; static ssize_t spufs_ibox_stat_read(struct file *file, char __user *buf, @@ -858,6 +881,7 @@ static ssize_t spufs_ibox_stat_read(struct file *file, char __user *buf, static const struct file_operations spufs_ibox_stat_fops = { .open = spufs_pipe_open, .read = spufs_ibox_stat_read, + .llseek = no_llseek, }; /* low-level mailbox write */ @@ -979,6 +1003,7 @@ static const struct file_operations spufs_wbox_fops = { .write = spufs_wbox_write, .poll = spufs_wbox_poll, .fasync = spufs_wbox_fasync, + .llseek = no_llseek, }; static ssize_t spufs_wbox_stat_read(struct file *file, char __user *buf, @@ -1006,6 +1031,7 @@ static ssize_t spufs_wbox_stat_read(struct file *file, char __user *buf, static const struct file_operations spufs_wbox_stat_fops = { .open = spufs_pipe_open, .read = spufs_wbox_stat_read, + .llseek = no_llseek, }; static int spufs_signal1_open(struct inode *inode, struct file *file) @@ -1097,23 +1123,23 @@ static ssize_t spufs_signal1_write(struct file *file, const char __user *buf, return 4; } -static unsigned long spufs_signal1_mmap_nopfn(struct vm_area_struct *vma, - unsigned long address) +static int +spufs_signal1_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { -#if PAGE_SIZE == 0x1000 - return spufs_ps_nopfn(vma, address, 0x14000, 0x1000); -#elif PAGE_SIZE == 0x10000 +#if SPUFS_SIGNAL_MAP_SIZE == 0x1000 + return spufs_ps_fault(vma, vmf, 0x14000, SPUFS_SIGNAL_MAP_SIZE); +#elif SPUFS_SIGNAL_MAP_SIZE == 0x10000 /* For 64k pages, both signal1 and signal2 can be used to mmap the whole * signal 1 and 2 area */ - return spufs_ps_nopfn(vma, address, 0x10000, 0x10000); + return spufs_ps_fault(vma, vmf, 0x10000, SPUFS_SIGNAL_MAP_SIZE); #else #error unsupported page size #endif } -static struct vm_operations_struct spufs_signal1_mmap_vmops = { - .nopfn = spufs_signal1_mmap_nopfn, +static const struct vm_operations_struct spufs_signal1_mmap_vmops = { + .fault = spufs_signal1_mmap_fault, }; static int spufs_signal1_mmap(struct file *file, struct vm_area_struct *vma) @@ -1122,8 +1148,7 @@ static int spufs_signal1_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; vma->vm_flags |= VM_IO | VM_PFNMAP; - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE | _PAGE_GUARDED); + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); vma->vm_ops = &spufs_signal1_mmap_vmops; return 0; @@ -1135,6 +1160,7 @@ static const struct file_operations spufs_signal1_fops = { .read = spufs_signal1_read, .write = spufs_signal1_write, .mmap = spufs_signal1_mmap, + .llseek = no_llseek, }; static const struct file_operations spufs_signal1_nosched_fops = { @@ -1142,6 +1168,7 @@ static const struct file_operations spufs_signal1_nosched_fops = { .release = spufs_signal1_release, .write = spufs_signal1_write, .mmap = spufs_signal1_mmap, + .llseek = no_llseek, }; static int spufs_signal2_open(struct inode *inode, struct file *file) @@ -1234,23 +1261,23 @@ static ssize_t spufs_signal2_write(struct file *file, const char __user *buf, } #if SPUFS_MMAP_4K -static unsigned long spufs_signal2_mmap_nopfn(struct vm_area_struct *vma, - unsigned long address) +static int +spufs_signal2_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { -#if PAGE_SIZE == 0x1000 - return spufs_ps_nopfn(vma, address, 0x1c000, 0x1000); -#elif PAGE_SIZE == 0x10000 +#if SPUFS_SIGNAL_MAP_SIZE == 0x1000 + return spufs_ps_fault(vma, vmf, 0x1c000, SPUFS_SIGNAL_MAP_SIZE); +#elif SPUFS_SIGNAL_MAP_SIZE == 0x10000 /* For 64k pages, both signal1 and signal2 can be used to mmap the whole * signal 1 and 2 area */ - return spufs_ps_nopfn(vma, address, 0x10000, 0x10000); + return spufs_ps_fault(vma, vmf, 0x10000, SPUFS_SIGNAL_MAP_SIZE); #else #error unsupported page size #endif } -static struct vm_operations_struct spufs_signal2_mmap_vmops = { - .nopfn = spufs_signal2_mmap_nopfn, +static const struct vm_operations_struct spufs_signal2_mmap_vmops = { + .fault = spufs_signal2_mmap_fault, }; static int spufs_signal2_mmap(struct file *file, struct vm_area_struct *vma) @@ -1259,8 +1286,7 @@ static int spufs_signal2_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; vma->vm_flags |= VM_IO | VM_PFNMAP; - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE | _PAGE_GUARDED); + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); vma->vm_ops = &spufs_signal2_mmap_vmops; return 0; @@ -1275,6 +1301,7 @@ static const struct file_operations spufs_signal2_fops = { .read = spufs_signal2_read, .write = spufs_signal2_write, .mmap = spufs_signal2_mmap, + .llseek = no_llseek, }; static const struct file_operations spufs_signal2_nosched_fops = { @@ -1282,6 +1309,7 @@ static const struct file_operations spufs_signal2_nosched_fops = { .release = spufs_signal2_release, .write = spufs_signal2_write, .mmap = spufs_signal2_mmap, + .llseek = no_llseek, }; /* @@ -1337,7 +1365,7 @@ static u64 spufs_signal1_type_get(struct spu_context *ctx) return ctx->ops->signal1_type_get(ctx); } DEFINE_SPUFS_ATTRIBUTE(spufs_signal1_type, spufs_signal1_type_get, - spufs_signal1_type_set, "%llu", SPU_ATTR_ACQUIRE); + spufs_signal1_type_set, "%llu\n", SPU_ATTR_ACQUIRE); static int spufs_signal2_type_set(void *data, u64 val) @@ -1359,17 +1387,17 @@ static u64 spufs_signal2_type_get(struct spu_context *ctx) return ctx->ops->signal2_type_get(ctx); } DEFINE_SPUFS_ATTRIBUTE(spufs_signal2_type, spufs_signal2_type_get, - spufs_signal2_type_set, "%llu", SPU_ATTR_ACQUIRE); + spufs_signal2_type_set, "%llu\n", SPU_ATTR_ACQUIRE); #if SPUFS_MMAP_4K -static unsigned long spufs_mss_mmap_nopfn(struct vm_area_struct *vma, - unsigned long address) +static int +spufs_mss_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - return spufs_ps_nopfn(vma, address, 0x0000, 0x1000); + return spufs_ps_fault(vma, vmf, 0x0000, SPUFS_MSS_MAP_SIZE); } -static struct vm_operations_struct spufs_mss_mmap_vmops = { - .nopfn = spufs_mss_mmap_nopfn, +static const struct vm_operations_struct spufs_mss_mmap_vmops = { + .fault = spufs_mss_mmap_fault, }; /* @@ -1381,8 +1409,7 @@ static int spufs_mss_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; vma->vm_flags |= VM_IO | VM_PFNMAP; - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE | _PAGE_GUARDED); + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); vma->vm_ops = &spufs_mss_mmap_vmops; return 0; @@ -1422,16 +1449,17 @@ static const struct file_operations spufs_mss_fops = { .open = spufs_mss_open, .release = spufs_mss_release, .mmap = spufs_mss_mmap, + .llseek = no_llseek, }; -static unsigned long spufs_psmap_mmap_nopfn(struct vm_area_struct *vma, - unsigned long address) +static int +spufs_psmap_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - return spufs_ps_nopfn(vma, address, 0x0000, 0x20000); + return spufs_ps_fault(vma, vmf, 0x0000, SPUFS_PS_MAP_SIZE); } -static struct vm_operations_struct spufs_psmap_mmap_vmops = { - .nopfn = spufs_psmap_mmap_nopfn, +static const struct vm_operations_struct spufs_psmap_mmap_vmops = { + .fault = spufs_psmap_mmap_fault, }; /* @@ -1443,8 +1471,7 @@ static int spufs_psmap_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; vma->vm_flags |= VM_IO | VM_PFNMAP; - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE | _PAGE_GUARDED); + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); vma->vm_ops = &spufs_psmap_mmap_vmops; return 0; @@ -1480,18 +1507,19 @@ static const struct file_operations spufs_psmap_fops = { .open = spufs_psmap_open, .release = spufs_psmap_release, .mmap = spufs_psmap_mmap, + .llseek = no_llseek, }; #if SPUFS_MMAP_4K -static unsigned long spufs_mfc_mmap_nopfn(struct vm_area_struct *vma, - unsigned long address) +static int +spufs_mfc_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - return spufs_ps_nopfn(vma, address, 0x3000, 0x1000); + return spufs_ps_fault(vma, vmf, 0x3000, SPUFS_MFC_MAP_SIZE); } -static struct vm_operations_struct spufs_mfc_mmap_vmops = { - .nopfn = spufs_mfc_mmap_nopfn, +static const struct vm_operations_struct spufs_mfc_mmap_vmops = { + .fault = spufs_mfc_mmap_fault, }; /* @@ -1503,8 +1531,7 @@ static int spufs_mfc_mmap(struct file *file, struct vm_area_struct *vma) return -EINVAL; vma->vm_flags |= VM_IO | VM_PFNMAP; - vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot) - | _PAGE_NO_CACHE | _PAGE_GUARDED); + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); vma->vm_ops = &spufs_mfc_mmap_vmops; return 0; @@ -1556,7 +1583,7 @@ void spufs_mfc_callback(struct spu *spu) wake_up_all(&ctx->mfc_wq); - pr_debug("%s %s\n", __FUNCTION__, spu->name); + pr_debug("%s %s\n", __func__, spu->name); if (ctx->mfc_fasync) { u32 free_elements, tagstatus; unsigned int mask; @@ -1630,7 +1657,7 @@ out: static int spufs_check_valid_dma(struct mfc_dma_command *cmd) { - pr_debug("queueing DMA %x %lx %x %x %x\n", cmd->lsa, + pr_debug("queueing DMA %x %llx %x %x %x\n", cmd->lsa, cmd->ea, cmd->size, cmd->tag, cmd->cmd); switch (cmd->cmd) { @@ -1647,7 +1674,7 @@ static int spufs_check_valid_dma(struct mfc_dma_command *cmd) } if ((cmd->lsa & 0xf) != (cmd->ea &0xf)) { - pr_debug("invalid DMA alignment, ea %lx lsa %x\n", + pr_debug("invalid DMA alignment, ea %llx lsa %x\n", cmd->ea, cmd->lsa); return -EIO; } @@ -1790,7 +1817,7 @@ static unsigned int spufs_mfc_poll(struct file *file,poll_table *wait) if (tagstatus & ctx->tagwait) mask |= POLLIN | POLLRDNORM; - pr_debug("%s: free %d tagstatus %d tagwait %d\n", __FUNCTION__, + pr_debug("%s: free %d tagstatus %d tagwait %d\n", __func__, free_elements, tagstatus, ctx->tagwait); return mask; @@ -1822,10 +1849,16 @@ out: return ret; } -static int spufs_mfc_fsync(struct file *file, struct dentry *dentry, - int datasync) +static int spufs_mfc_fsync(struct file *file, loff_t start, loff_t end, int datasync) { - return spufs_mfc_flush(file, NULL); + struct inode *inode = file_inode(file); + int err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (!err) { + mutex_lock(&inode->i_mutex); + err = spufs_mfc_flush(file, NULL); + mutex_unlock(&inode->i_mutex); + } + return err; } static int spufs_mfc_fasync(int fd, struct file *file, int on) @@ -1845,6 +1878,7 @@ static const struct file_operations spufs_mfc_fops = { .fsync = spufs_mfc_fsync, .fasync = spufs_mfc_fasync, .mmap = spufs_mfc_mmap, + .llseek = no_llseek, }; static int spufs_npc_set(void *data, u64 val) @@ -2220,6 +2254,7 @@ static ssize_t spufs_dma_info_read(struct file *file, char __user *buf, static const struct file_operations spufs_dma_info_fops = { .open = spufs_info_open, .read = spufs_dma_info_read, + .llseek = no_llseek, }; static ssize_t __spufs_proxydma_info_read(struct spu_context *ctx, @@ -2273,6 +2308,7 @@ static ssize_t spufs_proxydma_info_read(struct file *file, char __user *buf, static const struct file_operations spufs_proxydma_info_fops = { .open = spufs_info_open, .read = spufs_proxydma_info_read, + .llseek = no_llseek, }; static int spufs_show_tid(struct seq_file *s, void *private) @@ -2387,23 +2423,276 @@ static const struct file_operations spufs_stat_fops = { .release = single_release, }; +static inline int spufs_switch_log_used(struct spu_context *ctx) +{ + return (ctx->switch_log->head - ctx->switch_log->tail) % + SWITCH_LOG_BUFSIZE; +} + +static inline int spufs_switch_log_avail(struct spu_context *ctx) +{ + return SWITCH_LOG_BUFSIZE - spufs_switch_log_used(ctx); +} + +static int spufs_switch_log_open(struct inode *inode, struct file *file) +{ + struct spu_context *ctx = SPUFS_I(inode)->i_ctx; + int rc; + + rc = spu_acquire(ctx); + if (rc) + return rc; + + if (ctx->switch_log) { + rc = -EBUSY; + goto out; + } + + ctx->switch_log = kmalloc(sizeof(struct switch_log) + + SWITCH_LOG_BUFSIZE * sizeof(struct switch_log_entry), + GFP_KERNEL); + + if (!ctx->switch_log) { + rc = -ENOMEM; + goto out; + } + + ctx->switch_log->head = ctx->switch_log->tail = 0; + init_waitqueue_head(&ctx->switch_log->wait); + rc = 0; + +out: + spu_release(ctx); + return rc; +} + +static int spufs_switch_log_release(struct inode *inode, struct file *file) +{ + struct spu_context *ctx = SPUFS_I(inode)->i_ctx; + int rc; + + rc = spu_acquire(ctx); + if (rc) + return rc; + + kfree(ctx->switch_log); + ctx->switch_log = NULL; + spu_release(ctx); + + return 0; +} + +static int switch_log_sprint(struct spu_context *ctx, char *tbuf, int n) +{ + struct switch_log_entry *p; + + p = ctx->switch_log->log + ctx->switch_log->tail % SWITCH_LOG_BUFSIZE; + + return snprintf(tbuf, n, "%u.%09u %d %u %u %llu\n", + (unsigned int) p->tstamp.tv_sec, + (unsigned int) p->tstamp.tv_nsec, + p->spu_id, + (unsigned int) p->type, + (unsigned int) p->val, + (unsigned long long) p->timebase); +} -struct tree_descr spufs_dir_contents[] = { +static ssize_t spufs_switch_log_read(struct file *file, char __user *buf, + size_t len, loff_t *ppos) +{ + struct inode *inode = file_inode(file); + struct spu_context *ctx = SPUFS_I(inode)->i_ctx; + int error = 0, cnt = 0; + + if (!buf) + return -EINVAL; + + error = spu_acquire(ctx); + if (error) + return error; + + while (cnt < len) { + char tbuf[128]; + int width; + + if (spufs_switch_log_used(ctx) == 0) { + if (cnt > 0) { + /* If there's data ready to go, we can + * just return straight away */ + break; + + } else if (file->f_flags & O_NONBLOCK) { + error = -EAGAIN; + break; + + } else { + /* spufs_wait will drop the mutex and + * re-acquire, but since we're in read(), the + * file cannot be _released (and so + * ctx->switch_log is stable). + */ + error = spufs_wait(ctx->switch_log->wait, + spufs_switch_log_used(ctx) > 0); + + /* On error, spufs_wait returns without the + * state mutex held */ + if (error) + return error; + + /* We may have had entries read from underneath + * us while we dropped the mutex in spufs_wait, + * so re-check */ + if (spufs_switch_log_used(ctx) == 0) + continue; + } + } + + width = switch_log_sprint(ctx, tbuf, sizeof(tbuf)); + if (width < len) + ctx->switch_log->tail = + (ctx->switch_log->tail + 1) % + SWITCH_LOG_BUFSIZE; + else + /* If the record is greater than space available return + * partial buffer (so far) */ + break; + + error = copy_to_user(buf + cnt, tbuf, width); + if (error) + break; + cnt += width; + } + + spu_release(ctx); + + return cnt == 0 ? error : cnt; +} + +static unsigned int spufs_switch_log_poll(struct file *file, poll_table *wait) +{ + struct inode *inode = file_inode(file); + struct spu_context *ctx = SPUFS_I(inode)->i_ctx; + unsigned int mask = 0; + int rc; + + poll_wait(file, &ctx->switch_log->wait, wait); + + rc = spu_acquire(ctx); + if (rc) + return rc; + + if (spufs_switch_log_used(ctx) > 0) + mask |= POLLIN; + + spu_release(ctx); + + return mask; +} + +static const struct file_operations spufs_switch_log_fops = { + .open = spufs_switch_log_open, + .read = spufs_switch_log_read, + .poll = spufs_switch_log_poll, + .release = spufs_switch_log_release, + .llseek = no_llseek, +}; + +/** + * Log a context switch event to a switch log reader. + * + * Must be called with ctx->state_mutex held. + */ +void spu_switch_log_notify(struct spu *spu, struct spu_context *ctx, + u32 type, u32 val) +{ + if (!ctx->switch_log) + return; + + if (spufs_switch_log_avail(ctx) > 1) { + struct switch_log_entry *p; + + p = ctx->switch_log->log + ctx->switch_log->head; + ktime_get_ts(&p->tstamp); + p->timebase = get_tb(); + p->spu_id = spu ? spu->number : -1; + p->type = type; + p->val = val; + + ctx->switch_log->head = + (ctx->switch_log->head + 1) % SWITCH_LOG_BUFSIZE; + } + + wake_up(&ctx->switch_log->wait); +} + +static int spufs_show_ctx(struct seq_file *s, void *private) +{ + struct spu_context *ctx = s->private; + u64 mfc_control_RW; + + mutex_lock(&ctx->state_mutex); + if (ctx->spu) { + struct spu *spu = ctx->spu; + struct spu_priv2 __iomem *priv2 = spu->priv2; + + spin_lock_irq(&spu->register_lock); + mfc_control_RW = in_be64(&priv2->mfc_control_RW); + spin_unlock_irq(&spu->register_lock); + } else { + struct spu_state *csa = &ctx->csa; + + mfc_control_RW = csa->priv2.mfc_control_RW; + } + + seq_printf(s, "%c flgs(%lx) sflgs(%lx) pri(%d) ts(%d) spu(%02d)" + " %c %llx %llx %llx %llx %x %x\n", + ctx->state == SPU_STATE_SAVED ? 'S' : 'R', + ctx->flags, + ctx->sched_flags, + ctx->prio, + ctx->time_slice, + ctx->spu ? ctx->spu->number : -1, + !list_empty(&ctx->rq) ? 'q' : ' ', + ctx->csa.class_0_pending, + ctx->csa.class_0_dar, + ctx->csa.class_1_dsisr, + mfc_control_RW, + ctx->ops->runcntl_read(ctx), + ctx->ops->status_read(ctx)); + + mutex_unlock(&ctx->state_mutex); + + return 0; +} + +static int spufs_ctx_open(struct inode *inode, struct file *file) +{ + return single_open(file, spufs_show_ctx, SPUFS_I(inode)->i_ctx); +} + +static const struct file_operations spufs_ctx_fops = { + .open = spufs_ctx_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +const struct spufs_tree_descr spufs_dir_contents[] = { { "capabilities", &spufs_caps_fops, 0444, }, - { "mem", &spufs_mem_fops, 0666, }, - { "regs", &spufs_regs_fops, 0666, }, + { "mem", &spufs_mem_fops, 0666, LS_SIZE, }, + { "regs", &spufs_regs_fops, 0666, sizeof(struct spu_reg128[128]), }, { "mbox", &spufs_mbox_fops, 0444, }, { "ibox", &spufs_ibox_fops, 0444, }, { "wbox", &spufs_wbox_fops, 0222, }, - { "mbox_stat", &spufs_mbox_stat_fops, 0444, }, - { "ibox_stat", &spufs_ibox_stat_fops, 0444, }, - { "wbox_stat", &spufs_wbox_stat_fops, 0444, }, + { "mbox_stat", &spufs_mbox_stat_fops, 0444, sizeof(u32), }, + { "ibox_stat", &spufs_ibox_stat_fops, 0444, sizeof(u32), }, + { "wbox_stat", &spufs_wbox_stat_fops, 0444, sizeof(u32), }, { "signal1", &spufs_signal1_fops, 0666, }, { "signal2", &spufs_signal2_fops, 0666, }, { "signal1_type", &spufs_signal1_type, 0666, }, { "signal2_type", &spufs_signal2_type, 0666, }, { "cntl", &spufs_cntl_fops, 0666, }, - { "fpcr", &spufs_fpcr_fops, 0666, }, + { "fpcr", &spufs_fpcr_fops, 0666, sizeof(struct spu_reg128), }, { "lslr", &spufs_lslr_ops, 0444, }, { "mfc", &spufs_mfc_fops, 0666, }, { "mss", &spufs_mss_fops, 0666, }, @@ -2413,28 +2702,31 @@ struct tree_descr spufs_dir_contents[] = { { "decr_status", &spufs_decr_status_ops, 0666, }, { "event_mask", &spufs_event_mask_ops, 0666, }, { "event_status", &spufs_event_status_ops, 0444, }, - { "psmap", &spufs_psmap_fops, 0666, }, + { "psmap", &spufs_psmap_fops, 0666, SPUFS_PS_MAP_SIZE, }, { "phys-id", &spufs_id_ops, 0666, }, { "object-id", &spufs_object_id_ops, 0666, }, - { "mbox_info", &spufs_mbox_info_fops, 0444, }, - { "ibox_info", &spufs_ibox_info_fops, 0444, }, - { "wbox_info", &spufs_wbox_info_fops, 0444, }, - { "dma_info", &spufs_dma_info_fops, 0444, }, - { "proxydma_info", &spufs_proxydma_info_fops, 0444, }, + { "mbox_info", &spufs_mbox_info_fops, 0444, sizeof(u32), }, + { "ibox_info", &spufs_ibox_info_fops, 0444, sizeof(u32), }, + { "wbox_info", &spufs_wbox_info_fops, 0444, sizeof(u32), }, + { "dma_info", &spufs_dma_info_fops, 0444, + sizeof(struct spu_dma_info), }, + { "proxydma_info", &spufs_proxydma_info_fops, 0444, + sizeof(struct spu_proxydma_info)}, { "tid", &spufs_tid_fops, 0444, }, { "stat", &spufs_stat_fops, 0444, }, + { "switch_log", &spufs_switch_log_fops, 0444 }, {}, }; -struct tree_descr spufs_dir_nosched_contents[] = { +const struct spufs_tree_descr spufs_dir_nosched_contents[] = { { "capabilities", &spufs_caps_fops, 0444, }, - { "mem", &spufs_mem_fops, 0666, }, + { "mem", &spufs_mem_fops, 0666, LS_SIZE, }, { "mbox", &spufs_mbox_fops, 0444, }, { "ibox", &spufs_ibox_fops, 0444, }, { "wbox", &spufs_wbox_fops, 0222, }, - { "mbox_stat", &spufs_mbox_stat_fops, 0444, }, - { "ibox_stat", &spufs_ibox_stat_fops, 0444, }, - { "wbox_stat", &spufs_wbox_stat_fops, 0444, }, + { "mbox_stat", &spufs_mbox_stat_fops, 0444, sizeof(u32), }, + { "ibox_stat", &spufs_ibox_stat_fops, 0444, sizeof(u32), }, + { "wbox_stat", &spufs_wbox_stat_fops, 0444, sizeof(u32), }, { "signal1", &spufs_signal1_nosched_fops, 0222, }, { "signal2", &spufs_signal2_nosched_fops, 0222, }, { "signal1_type", &spufs_signal1_type, 0666, }, @@ -2443,7 +2735,7 @@ struct tree_descr spufs_dir_nosched_contents[] = { { "mfc", &spufs_mfc_fops, 0666, }, { "cntl", &spufs_cntl_fops, 0666, }, { "npc", &spufs_npc_ops, 0666, }, - { "psmap", &spufs_psmap_fops, 0666, }, + { "psmap", &spufs_psmap_fops, 0666, SPUFS_PS_MAP_SIZE, }, { "phys-id", &spufs_id_ops, 0666, }, { "object-id", &spufs_object_id_ops, 0666, }, { "tid", &spufs_tid_fops, 0444, }, @@ -2451,7 +2743,12 @@ struct tree_descr spufs_dir_nosched_contents[] = { {}, }; -struct spufs_coredump_reader spufs_coredump_read[] = { +const struct spufs_tree_descr spufs_dir_debug_contents[] = { + { ".ctx", &spufs_ctx_fops, 0444, }, + {}, +}; + +const struct spufs_coredump_reader spufs_coredump_read[] = { { "regs", __spufs_regs_read, NULL, sizeof(struct spu_reg128[128])}, { "fpcr", __spufs_fpcr_read, NULL, sizeof(struct spu_reg128) }, { "lslr", NULL, spufs_lslr_get, 19 }, diff --git a/arch/powerpc/platforms/cell/spufs/hw_ops.c b/arch/powerpc/platforms/cell/spufs/hw_ops.c index 64f8540b832..8655c4cbefc 100644 --- a/arch/powerpc/platforms/cell/spufs/hw_ops.c +++ b/arch/powerpc/platforms/cell/spufs/hw_ops.c @@ -18,7 +18,6 @@ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ -#include <linux/module.h> #include <linux/errno.h> #include <linux/sched.h> #include <linux/kernel.h> diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c index 6d1228c66c5..87ba7cf99cd 100644 --- a/arch/powerpc/platforms/cell/spufs/inode.c +++ b/arch/powerpc/platforms/cell/spufs/inode.c @@ -23,6 +23,7 @@ #include <linux/file.h> #include <linux/fs.h> +#include <linux/fsnotify.h> #include <linux/backing-dev.h> #include <linux/init.h> #include <linux/ioctl.h> @@ -35,17 +36,25 @@ #include <linux/parser.h> #include <asm/prom.h> -#include <asm/semaphore.h> #include <asm/spu.h> #include <asm/spu_priv1.h> #include <asm/uaccess.h> #include "spufs.h" +struct spufs_sb_info { + int debug; +}; + static struct kmem_cache *spufs_inode_cache; char *isolated_loader; static int isolated_loader_size; +static struct spufs_sb_info *spufs_get_sb_info(struct super_block *sb) +{ + return sb->s_fs_info; +} + static struct inode * spufs_alloc_inode(struct super_block *sb) { @@ -62,14 +71,19 @@ spufs_alloc_inode(struct super_block *sb) return &ei->vfs_inode; } -static void -spufs_destroy_inode(struct inode *inode) +static void spufs_i_callback(struct rcu_head *head) { + struct inode *inode = container_of(head, struct inode, i_rcu); kmem_cache_free(spufs_inode_cache, SPUFS_I(inode)); } +static void spufs_destroy_inode(struct inode *inode) +{ + call_rcu(&inode->i_rcu, spufs_i_callback); +} + static void -spufs_init_once(struct kmem_cache *cachep, void *p) +spufs_init_once(void *p) { struct spufs_inode_info *ei = p; @@ -77,7 +91,7 @@ spufs_init_once(struct kmem_cache *cachep, void *p) } static struct inode * -spufs_new_inode(struct super_block *sb, int mode) +spufs_new_inode(struct super_block *sb, umode_t mode) { struct inode *inode; @@ -85,10 +99,10 @@ spufs_new_inode(struct super_block *sb, int mode) if (!inode) goto out; + inode->i_ino = get_next_ino(); inode->i_mode = mode; - inode->i_uid = current->fsuid; - inode->i_gid = current->fsgid; - inode->i_blocks = 0; + inode->i_uid = current_fsuid(); + inode->i_gid = current_fsgid(); inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; out: return inode; @@ -102,16 +116,18 @@ spufs_setattr(struct dentry *dentry, struct iattr *attr) if ((attr->ia_valid & ATTR_SIZE) && (attr->ia_size != inode->i_size)) return -EINVAL; - return inode_setattr(inode, attr); + setattr_copy(inode, attr); + mark_inode_dirty(inode); + return 0; } static int spufs_new_file(struct super_block *sb, struct dentry *dentry, - const struct file_operations *fops, int mode, - struct spu_context *ctx) + const struct file_operations *fops, umode_t mode, + size_t size, struct spu_context *ctx) { - static struct inode_operations spufs_file_iops = { + static const struct inode_operations spufs_file_iops = { .setattr = spufs_setattr, }; struct inode *inode; @@ -125,6 +141,7 @@ spufs_new_file(struct super_block *sb, struct dentry *dentry, ret = 0; inode->i_op = &spufs_file_iops; inode->i_fop = fops; + inode->i_size = size; inode->i_private = SPUFS_I(inode)->i_ctx = get_spu_context(ctx); d_add(dentry, inode); out: @@ -132,15 +149,14 @@ out: } static void -spufs_delete_inode(struct inode *inode) +spufs_evict_inode(struct inode *inode) { struct spufs_inode_info *ei = SPUFS_I(inode); - + clear_inode(inode); if (ei->i_ctx) put_spu_context(ei->i_ctx); if (ei->i_gang) put_spu_gang(ei->i_gang); - clear_inode(inode); } static void spufs_prune_dir(struct dentry *dir) @@ -149,18 +165,18 @@ static void spufs_prune_dir(struct dentry *dir) mutex_lock(&dir->d_inode->i_mutex); list_for_each_entry_safe(dentry, tmp, &dir->d_subdirs, d_u.d_child) { - spin_lock(&dcache_lock); spin_lock(&dentry->d_lock); if (!(d_unhashed(dentry)) && dentry->d_inode) { - dget_locked(dentry); + dget_dlock(dentry); __d_drop(dentry); spin_unlock(&dentry->d_lock); simple_unlink(dir->d_inode, dentry); - spin_unlock(&dcache_lock); + /* XXX: what was dcache_lock protecting here? Other + * filesystems (IB, configfs) release dcache_lock + * before unlink */ dput(dentry); } else { spin_unlock(&dentry->d_lock); - spin_unlock(&dcache_lock); } } shrink_dcache_parent(dir); @@ -171,46 +187,31 @@ static void spufs_prune_dir(struct dentry *dir) static int spufs_rmdir(struct inode *parent, struct dentry *dir) { /* remove all entries */ + int res; spufs_prune_dir(dir); d_drop(dir); - - return simple_rmdir(parent, dir); + res = simple_rmdir(parent, dir); + /* We have to give up the mm_struct */ + spu_forget(SPUFS_I(dir->d_inode)->i_ctx); + return res; } -static int spufs_fill_dir(struct dentry *dir, struct tree_descr *files, - int mode, struct spu_context *ctx) +static int spufs_fill_dir(struct dentry *dir, + const struct spufs_tree_descr *files, umode_t mode, + struct spu_context *ctx) { - struct dentry *dentry, *tmp; - int ret; - while (files->name && files->name[0]) { - ret = -ENOMEM; - dentry = d_alloc_name(dir, files->name); + int ret; + struct dentry *dentry = d_alloc_name(dir, files->name); if (!dentry) - goto out; + return -ENOMEM; ret = spufs_new_file(dir->d_sb, dentry, files->ops, - files->mode & mode, ctx); + files->mode & mode, files->size, ctx); if (ret) - goto out; + return ret; files++; } return 0; -out: - /* - * remove all children from dir. dir->inode is not set so don't - * just simply use spufs_prune_dir() and panic afterwards :) - * dput() looks like it will do the right thing: - * - dec parent's ref counter - * - remove child from parent's child list - * - free child's inode if possible - * - free child - */ - list_for_each_entry_safe(dentry, tmp, &dir->d_subdirs, d_u.d_child) { - dput(dentry); - } - - shrink_dcache_parent(dir); - return ret; } static int spufs_dir_close(struct inode *inode, struct file *file) @@ -224,14 +225,11 @@ static int spufs_dir_close(struct inode *inode, struct file *file) parent = dir->d_parent->d_inode; ctx = SPUFS_I(dir->d_inode)->i_ctx; - mutex_lock(&parent->i_mutex); + mutex_lock_nested(&parent->i_mutex, I_MUTEX_PARENT); ret = spufs_rmdir(parent, dir); mutex_unlock(&parent->i_mutex); WARN_ON(ret); - /* We have to give up the mm_struct */ - spu_forget(ctx); - return dcache_dir_close(inode, file); } @@ -240,23 +238,22 @@ const struct file_operations spufs_context_fops = { .release = spufs_dir_close, .llseek = dcache_dir_lseek, .read = generic_read_dir, - .readdir = dcache_readdir, - .fsync = simple_sync_file, + .iterate = dcache_readdir, + .fsync = noop_fsync, }; EXPORT_SYMBOL_GPL(spufs_context_fops); static int spufs_mkdir(struct inode *dir, struct dentry *dentry, unsigned int flags, - int mode) + umode_t mode) { int ret; struct inode *inode; struct spu_context *ctx; - ret = -ENOSPC; inode = spufs_new_inode(dir->i_sb, mode | S_IFDIR); if (!inode) - goto out; + return -ENOSPC; if (dir->i_mode & S_ISGID) { inode->i_gid = dir->i_gid; @@ -264,58 +261,58 @@ spufs_mkdir(struct inode *dir, struct dentry *dentry, unsigned int flags, } ctx = alloc_spu_context(SPUFS_I(dir)->i_gang); /* XXX gang */ SPUFS_I(inode)->i_ctx = ctx; - if (!ctx) - goto out_iput; + if (!ctx) { + iput(inode); + return -ENOSPC; + } ctx->flags = flags; inode->i_op = &simple_dir_inode_operations; inode->i_fop = &simple_dir_operations; + + mutex_lock(&inode->i_mutex); + + dget(dentry); + inc_nlink(dir); + inc_nlink(inode); + + d_instantiate(dentry, inode); + if (flags & SPU_CREATE_NOSCHED) ret = spufs_fill_dir(dentry, spufs_dir_nosched_contents, mode, ctx); else ret = spufs_fill_dir(dentry, spufs_dir_contents, mode, ctx); + if (!ret && spufs_get_sb_info(dir->i_sb)->debug) + ret = spufs_fill_dir(dentry, spufs_dir_debug_contents, + mode, ctx); + if (ret) - goto out_free_ctx; + spufs_rmdir(dir, dentry); - d_instantiate(dentry, inode); - dget(dentry); - dir->i_nlink++; - dentry->d_inode->i_nlink++; - goto out; + mutex_unlock(&inode->i_mutex); -out_free_ctx: - spu_forget(ctx); - put_spu_context(ctx); -out_iput: - iput(inode); -out: return ret; } -static int spufs_context_open(struct dentry *dentry, struct vfsmount *mnt) +static int spufs_context_open(struct path *path) { int ret; struct file *filp; ret = get_unused_fd(); - if (ret < 0) { - dput(dentry); - mntput(mnt); - goto out; - } + if (ret < 0) + return ret; - filp = dentry_open(dentry, mnt, O_RDONLY); + filp = dentry_open(path, O_RDONLY, current_cred()); if (IS_ERR(filp)) { put_unused_fd(ret); - ret = PTR_ERR(filp); - goto out; + return PTR_ERR(filp); } filp->f_op = &spufs_context_fops; fd_install(ret, filp); -out: return ret; } @@ -350,7 +347,7 @@ spufs_assert_affinity(unsigned int flags, struct spu_gang *gang, return ERR_PTR(-EINVAL); neighbor = get_spu_context( - SPUFS_I(filp->f_dentry->d_inode)->i_ctx); + SPUFS_I(file_inode(filp))->i_ctx); if (!list_empty(&neighbor->aff_list) && !(neighbor->aff_head) && !list_is_last(&neighbor->aff_list, &gang->aff_list_head) && @@ -423,36 +420,33 @@ spufs_set_affinity(unsigned int flags, struct spu_context *ctx, static int spufs_create_context(struct inode *inode, struct dentry *dentry, - struct vfsmount *mnt, int flags, int mode, + struct vfsmount *mnt, int flags, umode_t mode, struct file *aff_filp) { int ret; int affinity; struct spu_gang *gang; struct spu_context *neighbor; + struct path path = {.mnt = mnt, .dentry = dentry}; - ret = -EPERM; if ((flags & SPU_CREATE_NOSCHED) && !capable(CAP_SYS_NICE)) - goto out_unlock; + return -EPERM; - ret = -EINVAL; if ((flags & (SPU_CREATE_NOSCHED | SPU_CREATE_ISOLATE)) == SPU_CREATE_ISOLATE) - goto out_unlock; + return -EINVAL; - ret = -ENODEV; if ((flags & SPU_CREATE_ISOLATE) && !isolated_loader) - goto out_unlock; + return -ENODEV; gang = NULL; neighbor = NULL; affinity = flags & (SPU_CREATE_AFFINITY_MEM | SPU_CREATE_AFFINITY_SPU); if (affinity) { gang = SPUFS_I(inode)->i_gang; - ret = -EINVAL; if (!gang) - goto out_unlock; + return -EINVAL; mutex_lock(&gang->aff_mutex); neighbor = spufs_assert_affinity(flags, gang, aff_filp); if (IS_ERR(neighbor)) { @@ -472,30 +466,18 @@ spufs_create_context(struct inode *inode, struct dentry *dentry, put_spu_context(neighbor); } - /* - * get references for dget and mntget, will be released - * in error path of *_open(). - */ - ret = spufs_context_open(dget(dentry), mntget(mnt)); - if (ret < 0) { + ret = spufs_context_open(&path); + if (ret < 0) WARN_ON(spufs_rmdir(inode, dentry)); - mutex_unlock(&inode->i_mutex); - spu_forget(SPUFS_I(dentry->d_inode)->i_ctx); - goto out; - } out_aff_unlock: if (affinity) mutex_unlock(&gang->aff_mutex); -out_unlock: - mutex_unlock(&inode->i_mutex); -out: - dput(dentry); return ret; } static int -spufs_mkgang(struct inode *dir, struct dentry *dentry, int mode) +spufs_mkgang(struct inode *dir, struct dentry *dentry, umode_t mode) { int ret; struct inode *inode; @@ -521,8 +503,8 @@ spufs_mkgang(struct inode *dir, struct dentry *dentry, int mode) inode->i_fop = &simple_dir_operations; d_instantiate(dentry, inode); - dir->i_nlink++; - dentry->d_inode->i_nlink++; + inc_nlink(dir); + inc_nlink(dentry->d_inode); return ret; out_iput: @@ -531,123 +513,98 @@ out: return ret; } -static int spufs_gang_open(struct dentry *dentry, struct vfsmount *mnt) +static int spufs_gang_open(struct path *path) { int ret; struct file *filp; ret = get_unused_fd(); - if (ret < 0) { - dput(dentry); - mntput(mnt); - goto out; - } + if (ret < 0) + return ret; - filp = dentry_open(dentry, mnt, O_RDONLY); + /* + * get references for dget and mntget, will be released + * in error path of *_open(). + */ + filp = dentry_open(path, O_RDONLY, current_cred()); if (IS_ERR(filp)) { put_unused_fd(ret); - ret = PTR_ERR(filp); - goto out; + return PTR_ERR(filp); } filp->f_op = &simple_dir_operations; fd_install(ret, filp); -out: return ret; } static int spufs_create_gang(struct inode *inode, struct dentry *dentry, - struct vfsmount *mnt, int mode) + struct vfsmount *mnt, umode_t mode) { + struct path path = {.mnt = mnt, .dentry = dentry}; int ret; ret = spufs_mkgang(inode, dentry, mode & S_IRWXUGO); - if (ret) - goto out; - - /* - * get references for dget and mntget, will be released - * in error path of *_open(). - */ - ret = spufs_gang_open(dget(dentry), mntget(mnt)); - if (ret < 0) { - int err = simple_rmdir(inode, dentry); - WARN_ON(err); + if (!ret) { + ret = spufs_gang_open(&path); + if (ret < 0) { + int err = simple_rmdir(inode, dentry); + WARN_ON(err); + } } - -out: - mutex_unlock(&inode->i_mutex); - dput(dentry); return ret; } static struct file_system_type spufs_type; -long spufs_create(struct nameidata *nd, unsigned int flags, mode_t mode, - struct file *filp) +long spufs_create(struct path *path, struct dentry *dentry, + unsigned int flags, umode_t mode, struct file *filp) { - struct dentry *dentry; + struct inode *dir = path->dentry->d_inode; int ret; - ret = -EINVAL; /* check if we are on spufs */ - if (nd->path.dentry->d_sb->s_type != &spufs_type) - goto out; + if (path->dentry->d_sb->s_type != &spufs_type) + return -EINVAL; /* don't accept undefined flags */ if (flags & (~SPU_CREATE_FLAG_ALL)) - goto out; + return -EINVAL; /* only threads can be underneath a gang */ - if (nd->path.dentry != nd->path.dentry->d_sb->s_root) { - if ((flags & SPU_CREATE_GANG) || - !SPUFS_I(nd->path.dentry->d_inode)->i_gang) - goto out; - } - - dentry = lookup_create(nd, 1); - ret = PTR_ERR(dentry); - if (IS_ERR(dentry)) - goto out_dir; + if (path->dentry != path->dentry->d_sb->s_root) + if ((flags & SPU_CREATE_GANG) || !SPUFS_I(dir)->i_gang) + return -EINVAL; - ret = -EEXIST; - if (dentry->d_inode) - goto out_dput; - - mode &= ~current->fs->umask; + mode &= ~current_umask(); if (flags & SPU_CREATE_GANG) - return spufs_create_gang(nd->path.dentry->d_inode, - dentry, nd->path.mnt, mode); + ret = spufs_create_gang(dir, dentry, path->mnt, mode); else - return spufs_create_context(nd->path.dentry->d_inode, - dentry, nd->path.mnt, flags, mode, + ret = spufs_create_context(dir, dentry, path->mnt, flags, mode, filp); + if (ret >= 0) + fsnotify_mkdir(dir, dentry); -out_dput: - dput(dentry); -out_dir: - mutex_unlock(&nd->path.dentry->d_inode->i_mutex); -out: return ret; } /* File system initialization */ enum { - Opt_uid, Opt_gid, Opt_mode, Opt_err, + Opt_uid, Opt_gid, Opt_mode, Opt_debug, Opt_err, }; -static match_table_t spufs_tokens = { - { Opt_uid, "uid=%d" }, - { Opt_gid, "gid=%d" }, - { Opt_mode, "mode=%o" }, - { Opt_err, NULL }, +static const match_table_t spufs_tokens = { + { Opt_uid, "uid=%d" }, + { Opt_gid, "gid=%d" }, + { Opt_mode, "mode=%o" }, + { Opt_debug, "debug" }, + { Opt_err, NULL }, }; static int -spufs_parse_options(char *options, struct inode *root) +spufs_parse_options(struct super_block *sb, char *options, struct inode *root) { char *p; substring_t args[MAX_OPT_ARGS]; @@ -663,18 +620,25 @@ spufs_parse_options(char *options, struct inode *root) case Opt_uid: if (match_int(&args[0], &option)) return 0; - root->i_uid = option; + root->i_uid = make_kuid(current_user_ns(), option); + if (!uid_valid(root->i_uid)) + return 0; break; case Opt_gid: if (match_int(&args[0], &option)) return 0; - root->i_gid = option; + root->i_gid = make_kgid(current_user_ns(), option); + if (!gid_valid(root->i_gid)) + return 0; break; case Opt_mode: if (match_octal(&args[0], &option)) return 0; root->i_mode = option | S_IFDIR; break; + case Opt_debug: + spufs_get_sb_info(sb)->debug = 1; + break; default: return 0; } @@ -731,15 +695,16 @@ spufs_create_root(struct super_block *sb, void *data) inode->i_op = &simple_dir_inode_operations; inode->i_fop = &simple_dir_operations; SPUFS_I(inode)->i_ctx = NULL; + inc_nlink(inode); ret = -EINVAL; - if (!spufs_parse_options(data, inode)) + if (!spufs_parse_options(sb, data, inode)) goto out_iput; ret = -ENOMEM; - sb->s_root = d_alloc_root(inode); + sb->s_root = d_make_root(inode); if (!sb->s_root) - goto out_iput; + goto out; return 0; out_iput: @@ -751,39 +716,45 @@ out: static int spufs_fill_super(struct super_block *sb, void *data, int silent) { - static struct super_operations s_ops = { + struct spufs_sb_info *info; + static const struct super_operations s_ops = { .alloc_inode = spufs_alloc_inode, .destroy_inode = spufs_destroy_inode, .statfs = simple_statfs, - .delete_inode = spufs_delete_inode, - .drop_inode = generic_delete_inode, + .evict_inode = spufs_evict_inode, .show_options = generic_show_options, }; save_mount_options(sb, data); + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return -ENOMEM; + sb->s_maxbytes = MAX_LFS_FILESIZE; sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; sb->s_magic = SPUFS_MAGIC; sb->s_op = &s_ops; + sb->s_fs_info = info; return spufs_create_root(sb, data); } -static int -spufs_get_sb(struct file_system_type *fstype, int flags, - const char *name, void *data, struct vfsmount *mnt) +static struct dentry * +spufs_mount(struct file_system_type *fstype, int flags, + const char *name, void *data) { - return get_sb_single(fstype, flags, data, spufs_fill_super, mnt); + return mount_single(fstype, flags, data, spufs_fill_super); } static struct file_system_type spufs_type = { .owner = THIS_MODULE, .name = "spufs", - .get_sb = spufs_get_sb, + .mount = spufs_mount, .kill_sb = kill_litter_super, }; +MODULE_ALIAS_FS("spufs"); static int __init spufs_init(void) { @@ -803,19 +774,19 @@ static int __init spufs_init(void) ret = spu_sched_init(); if (ret) goto out_cache; - ret = register_filesystem(&spufs_type); + ret = register_spu_syscalls(&spufs_calls); if (ret) goto out_sched; - ret = register_spu_syscalls(&spufs_calls); + ret = register_filesystem(&spufs_type); if (ret) - goto out_fs; + goto out_syscalls; spufs_init_isolated_loader(); return 0; -out_fs: - unregister_filesystem(&spufs_type); +out_syscalls: + unregister_spu_syscalls(&spufs_calls); out_sched: spu_sched_exit(); out_cache: diff --git a/arch/powerpc/platforms/cell/spufs/lscsa_alloc.c b/arch/powerpc/platforms/cell/spufs/lscsa_alloc.c index 0e9f325c9ff..147069938cf 100644 --- a/arch/powerpc/platforms/cell/spufs/lscsa_alloc.c +++ b/arch/powerpc/platforms/cell/spufs/lscsa_alloc.c @@ -22,6 +22,7 @@ #include <linux/kernel.h> #include <linux/mm.h> +#include <linux/slab.h> #include <linux/vmalloc.h> #include <asm/spu.h> @@ -35,10 +36,9 @@ static int spu_alloc_lscsa_std(struct spu_state *csa) struct spu_lscsa *lscsa; unsigned char *p; - lscsa = vmalloc(sizeof(struct spu_lscsa)); + lscsa = vzalloc(sizeof(struct spu_lscsa)); if (!lscsa) return -ENOMEM; - memset(lscsa, 0, sizeof(struct spu_lscsa)); csa->lscsa = lscsa; /* Set LS pages reserved to allow for user-space mapping. */ @@ -90,7 +90,7 @@ int spu_alloc_lscsa(struct spu_state *csa) */ for (i = 0; i < SPU_LSCSA_NUM_BIG_PAGES; i++) { /* XXX This is likely to fail, we should use a special pool - * similiar to what hugetlbfs does. + * similar to what hugetlbfs does. */ csa->lscsa_pages[i] = alloc_pages(GFP_KERNEL, SPU_64K_PAGE_ORDER); diff --git a/arch/powerpc/platforms/cell/spufs/run.c b/arch/powerpc/platforms/cell/spufs/run.c index cac69e11677..4ddf769a64e 100644 --- a/arch/powerpc/platforms/cell/spufs/run.c +++ b/arch/powerpc/platforms/cell/spufs/run.c @@ -11,7 +11,7 @@ #include "spufs.h" /* interrupt-level stop callback function. */ -void spufs_stop_callback(struct spu *spu) +void spufs_stop_callback(struct spu *spu, int irq) { struct spu_context *ctx = spu->ctx; @@ -24,9 +24,18 @@ void spufs_stop_callback(struct spu *spu) */ if (ctx) { /* Copy exception arguments into module specific structure */ - ctx->csa.class_0_pending = spu->class_0_pending; - ctx->csa.dsisr = spu->dsisr; - ctx->csa.dar = spu->dar; + switch(irq) { + case 0 : + ctx->csa.class_0_pending = spu->class_0_pending; + ctx->csa.class_0_dar = spu->class_0_dar; + break; + case 1 : + ctx->csa.class_1_dsisr = spu->class_1_dsisr; + ctx->csa.class_1_dar = spu->class_1_dar; + break; + case 2 : + break; + } /* ensure that the exception status has hit memory before a * thread waiting on the context's stop queue is woken */ @@ -34,11 +43,6 @@ void spufs_stop_callback(struct spu *spu) wake_up_all(&ctx->stop_wq); } - - /* Clear callback arguments from spu structure */ - spu->class_0_pending = 0; - spu->dsisr = 0; - spu->dar = 0; } int spu_stopped(struct spu_context *ctx, u32 *stat) @@ -46,17 +50,25 @@ int spu_stopped(struct spu_context *ctx, u32 *stat) u64 dsisr; u32 stopped; - *stat = ctx->ops->status_read(ctx); + stopped = SPU_STATUS_INVALID_INSTR | SPU_STATUS_SINGLE_STEP | + SPU_STATUS_STOPPED_BY_HALT | SPU_STATUS_STOPPED_BY_STOP; - if (test_bit(SPU_SCHED_NOTIFY_ACTIVE, &ctx->sched_flags)) +top: + *stat = ctx->ops->status_read(ctx); + if (*stat & stopped) { + /* + * If the spu hasn't finished stopping, we need to + * re-read the register to get the stopped value. + */ + if (*stat & SPU_STATUS_RUNNING) + goto top; return 1; + } - stopped = SPU_STATUS_INVALID_INSTR | SPU_STATUS_SINGLE_STEP | - SPU_STATUS_STOPPED_BY_HALT | SPU_STATUS_STOPPED_BY_STOP; - if (!(*stat & SPU_STATUS_RUNNING) && (*stat & stopped)) + if (test_bit(SPU_SCHED_NOTIFY_ACTIVE, &ctx->sched_flags)) return 1; - dsisr = ctx->csa.dsisr; + dsisr = ctx->csa.class_1_dsisr; if (dsisr & (MFC_DSISR_PTE_NOT_FOUND | MFC_DSISR_ACCESS_DENIED)) return 1; @@ -98,13 +110,16 @@ static int spu_setup_isolated(struct spu_context *ctx) != MFC_CNTL_PURGE_DMA_COMPLETE) { if (time_after(jiffies, timeout)) { printk(KERN_ERR "%s: timeout flushing MFC DMA queue\n", - __FUNCTION__); + __func__); ret = -EIO; goto out; } cond_resched(); } + /* clear purge status */ + out_be64(mfc_cntl, 0); + /* put the SPE in kernel mode to allow access to the loader */ sr1 = spu_mfc_sr1_get(ctx->spu); sr1 &= ~MFC_STATE1_PROBLEM_STATE_MASK; @@ -124,7 +139,7 @@ static int spu_setup_isolated(struct spu_context *ctx) status_loading) { if (time_after(jiffies, timeout)) { printk(KERN_ERR "%s: timeout waiting for loader\n", - __FUNCTION__); + __func__); ret = -EIO; goto out_drop_priv; } @@ -134,7 +149,7 @@ static int spu_setup_isolated(struct spu_context *ctx) if (!(status & SPU_STATUS_RUNNING)) { /* If isolated LOAD has failed: run SPU, we will get a stop-and * signal later. */ - pr_debug("%s: isolated LOAD failed\n", __FUNCTION__); + pr_debug("%s: isolated LOAD failed\n", __func__); ctx->ops->runcntl_write(ctx, SPU_RUNCNTL_RUNNABLE); ret = -EACCES; goto out_drop_priv; @@ -142,7 +157,7 @@ static int spu_setup_isolated(struct spu_context *ctx) if (!(status & SPU_STATUS_ISOLATED_STATE)) { /* This isn't allowed by the CBEA, but check anyway */ - pr_debug("%s: SPU fell out of isolated mode?\n", __FUNCTION__); + pr_debug("%s: SPU fell out of isolated mode?\n", __func__); ctx->ops->runcntl_write(ctx, SPU_RUNCNTL_STOP); ret = -EINVAL; goto out_drop_priv; @@ -194,11 +209,6 @@ static int spu_run_init(struct spu_context *ctx, u32 *npc) (SPU_RUNCNTL_RUNNABLE | SPU_RUNCNTL_ISOLATE); if (runcntl == 0) runcntl = SPU_RUNCNTL_RUNNABLE; - } - - if (ctx->flags & SPU_CREATE_NOSCHED) { - spuctx_switch_state(ctx, SPU_UTIL_USER); - ctx->ops->runcntl_write(ctx, runcntl); } else { unsigned long privcntl; @@ -207,9 +217,15 @@ static int spu_run_init(struct spu_context *ctx, u32 *npc) else privcntl = SPU_PRIVCNTL_MODE_NORMAL; - ctx->ops->npc_write(ctx, *npc); ctx->ops->privcntl_write(ctx, privcntl); - ctx->ops->runcntl_write(ctx, runcntl); + ctx->ops->npc_write(ctx, *npc); + } + + ctx->ops->runcntl_write(ctx, runcntl); + + if (ctx->flags & SPU_CREATE_NOSCHED) { + spuctx_switch_state(ctx, SPU_UTIL_USER); + } else { if (ctx->state == SPU_STATE_SAVED) { ret = spu_activate(ctx, 0); @@ -236,6 +252,7 @@ static int spu_run_fini(struct spu_context *ctx, u32 *npc, spuctx_switch_state(ctx, SPU_UTIL_IDLE_LOADED); clear_bit(SPU_SCHED_SPU_RUN, &ctx->sched_flags); + spu_switch_log_notify(NULL, ctx, SWITCH_LOG_EXIT, *status); spu_release(ctx); if (signal_pending(current)) @@ -282,7 +299,7 @@ static int spu_handle_restartsys(struct spu_context *ctx, long *spu_ret, break; default: printk(KERN_WARNING "%s: unexpected return code %ld\n", - __FUNCTION__, *spu_ret); + __func__, *spu_ret); ret = 0; } return ret; @@ -294,7 +311,7 @@ static int spu_process_callback(struct spu_context *ctx) u32 ls_pointer, npc; void __iomem *ls; long spu_ret; - int ret, ret2; + int ret; /* get syscall block from local store */ npc = ctx->ops->npc_read(ctx) & ~3; @@ -316,13 +333,15 @@ static int spu_process_callback(struct spu_context *ctx) if (spu_ret <= -ERESTARTSYS) { ret = spu_handle_restartsys(ctx, &spu_ret, &npc); } - ret2 = spu_acquire(ctx); + mutex_lock(&ctx->state_mutex); if (ret == -ERESTARTSYS) return ret; - if (ret2) - return -EINTR; } + /* need to re-get the ls, as it may have changed when we released the + * spu */ + ls = (void __iomem *)ctx->ops->get_ls(ctx); + /* write result, jump over indirect pointer */ memcpy_toio(ls + ls_pointer, &spu_ret, sizeof(spu_ret)); ctx->ops->npc_write(ctx, npc); @@ -339,13 +358,14 @@ long spufs_run_spu(struct spu_context *ctx, u32 *npc, u32 *event) if (mutex_lock_interruptible(&ctx->run_mutex)) return -ERESTARTSYS; - spu_enable_spu(ctx); ctx->event_return = 0; ret = spu_acquire(ctx); if (ret) goto out_unlock; + spu_enable_spu(ctx); + spu_update_sched_info(ctx); ret = spu_run_init(ctx, npc); diff --git a/arch/powerpc/platforms/cell/spufs/sched.c b/arch/powerpc/platforms/cell/spufs/sched.c index 00528ef84ad..4a0a64fe25d 100644 --- a/arch/powerpc/platforms/cell/spufs/sched.c +++ b/arch/powerpc/platforms/cell/spufs/sched.c @@ -22,11 +22,12 @@ #undef DEBUG -#include <linux/module.h> #include <linux/errno.h> #include <linux/sched.h> +#include <linux/sched/rt.h> #include <linux/kernel.h> #include <linux/mm.h> +#include <linux/slab.h> #include <linux/completion.h> #include <linux/vmalloc.h> #include <linux/smp.h> @@ -39,7 +40,6 @@ #include <linux/pid_namespace.h> #include <linux/proc_fs.h> #include <linux/seq_file.h> -#include <linux/marker.h> #include <asm/io.h> #include <asm/mmu_context.h> @@ -47,6 +47,8 @@ #include <asm/spu_csa.h> #include <asm/spu_priv1.h> #include "spufs.h" +#define CREATE_TRACE_POINTS +#include "sputrace.h" struct spu_prio_array { DECLARE_BITMAP(bitmap, MAX_PRIO); @@ -81,7 +83,6 @@ static struct timer_list spuloadavg_timer; #define MIN_SPU_TIMESLICE max(5 * HZ / (1000 * SPUSCHED_TICK), 1) #define DEF_SPU_TIMESLICE (100 * HZ / (1000 * SPUSCHED_TICK)) -#define MAX_USER_PRIO (MAX_PRIO - MAX_RT_PRIO) #define SCALE_PRIO(x, prio) \ max(x * (MAX_PRIO - prio) / (MAX_USER_PRIO / 2), MIN_SPU_TIMESLICE) @@ -139,7 +140,10 @@ void __spu_update_sched_info(struct spu_context *ctx) * runqueue. The context will be rescheduled on the proper node * if it is timesliced or preempted. */ - ctx->cpus_allowed = current->cpus_allowed; + cpumask_copy(&ctx->cpus_allowed, tsk_cpus_allowed(current)); + + /* Save the current cpu id for spu interrupt routing. */ + ctx->last_ran = raw_smp_processor_id(); } void spu_update_sched_info(struct spu_context *ctx) @@ -163,9 +167,9 @@ void spu_update_sched_info(struct spu_context *ctx) static int __node_allowed(struct spu_context *ctx, int node) { if (nr_cpus_node(node)) { - cpumask_t mask = node_to_cpumask(node); + const struct cpumask *mask = cpumask_of_node(node); - if (cpus_intersects(mask, ctx->cpus_allowed)) + if (cpumask_intersects(mask, &ctx->cpus_allowed)) return 1; } @@ -227,22 +231,26 @@ static void spu_bind_context(struct spu *spu, struct spu_context *ctx) ctx->stats.slb_flt_base = spu->stats.slb_flt; ctx->stats.class2_intr_base = spu->stats.class2_intr; + spu_associate_mm(spu, ctx->owner); + + spin_lock_irq(&spu->register_lock); spu->ctx = ctx; spu->flags = 0; ctx->spu = spu; ctx->ops = &spu_hw_ops; spu->pid = current->pid; spu->tgid = current->tgid; - spu_associate_mm(spu, ctx->owner); spu->ibox_callback = spufs_ibox_callback; spu->wbox_callback = spufs_wbox_callback; spu->stop_callback = spufs_stop_callback; spu->mfc_callback = spufs_mfc_callback; - mb(); + spin_unlock_irq(&spu->register_lock); + spu_unmap_mappings(ctx); + + spu_switch_log_notify(spu, ctx, SWITCH_LOG_START, 0); spu_restore(&ctx->csa, spu); spu->timestamp = jiffies; - spu_cpu_affinity_set(spu, raw_smp_processor_id()); spu_switch_notify(spu, ctx); ctx->state = SPU_STATE_RUNNABLE; @@ -305,11 +313,35 @@ static struct spu *aff_ref_location(struct spu_context *ctx, int mem_aff, */ node = cpu_to_node(raw_smp_processor_id()); for (n = 0; n < MAX_NUMNODES; n++, node++) { + /* + * "available_spus" counts how many spus are not potentially + * going to be used by other affinity gangs whose reference + * context is already in place. Although this code seeks to + * avoid having affinity gangs with a summed amount of + * contexts bigger than the amount of spus in the node, + * this may happen sporadically. In this case, available_spus + * becomes negative, which is harmless. + */ + int available_spus; + node = (node < MAX_NUMNODES) ? node : 0; if (!node_allowed(ctx, node)) continue; + + available_spus = 0; mutex_lock(&cbe_spu_info[node].list_mutex); list_for_each_entry(spu, &cbe_spu_info[node].spus, cbe_list) { + if (spu->ctx && spu->ctx->gang && !spu->ctx->aff_offset + && spu->ctx->gang->aff_ref_spu) + available_spus -= spu->ctx->gang->contexts; + available_spus++; + } + if (available_spus < ctx->gang->contexts) { + mutex_unlock(&cbe_spu_info[node].list_mutex); + continue; + } + + list_for_each_entry(spu, &cbe_spu_info[node].spus, cbe_list) { if ((!mem_aff || spu->has_mem_affinity) && sched_spu(spu)) { mutex_unlock(&cbe_spu_info[node].list_mutex); @@ -382,6 +414,9 @@ static int has_affinity(struct spu_context *ctx) if (list_empty(&ctx->aff_list)) return 0; + if (atomic_read(&ctx->gang->aff_sched_count) == 0) + ctx->gang->aff_ref_spu = NULL; + if (!gang->aff_ref_spu) { if (!(gang->aff_flags & AFF_MERGED)) aff_merge_remaining_ctxs(gang); @@ -400,6 +435,8 @@ static int has_affinity(struct spu_context *ctx) */ static void spu_unbind_context(struct spu *spu, struct spu_context *ctx) { + u32 status; + spu_context_trace(spu_unbind_context__enter, ctx, spu); spuctx_switch_state(ctx, SPU_UTIL_SYSTEM); @@ -407,30 +444,34 @@ static void spu_unbind_context(struct spu *spu, struct spu_context *ctx) if (spu->ctx->flags & SPU_CREATE_NOSCHED) atomic_dec(&cbe_spu_info[spu->node].reserved_spus); - if (ctx->gang){ - mutex_lock(&ctx->gang->aff_mutex); - if (has_affinity(ctx)) { - if (atomic_dec_and_test(&ctx->gang->aff_sched_count)) - ctx->gang->aff_ref_spu = NULL; - } - mutex_unlock(&ctx->gang->aff_mutex); - } + if (ctx->gang) + /* + * If ctx->gang->aff_sched_count is positive, SPU affinity is + * being considered in this gang. Using atomic_dec_if_positive + * allow us to skip an explicit check for affinity in this gang + */ + atomic_dec_if_positive(&ctx->gang->aff_sched_count); spu_switch_notify(spu, NULL); spu_unmap_mappings(ctx); spu_save(&ctx->csa, spu); + spu_switch_log_notify(spu, ctx, SWITCH_LOG_STOP, 0); + + spin_lock_irq(&spu->register_lock); spu->timestamp = jiffies; ctx->state = SPU_STATE_SAVED; spu->ibox_callback = NULL; spu->wbox_callback = NULL; spu->stop_callback = NULL; spu->mfc_callback = NULL; - spu_associate_mm(spu, NULL); spu->pid = 0; spu->tgid = 0; ctx->ops = &spu_backing_ops; spu->flags = 0; spu->ctx = NULL; + spin_unlock_irq(&spu->register_lock); + + spu_associate_mm(spu, NULL); ctx->stats.slb_flt += (spu->stats.slb_flt - ctx->stats.slb_flt_base); @@ -440,6 +481,9 @@ static void spu_unbind_context(struct spu *spu, struct spu_context *ctx) /* This maps the underlying spu state to idle */ spuctx_switch_state(ctx, SPU_UTIL_IDLE_LOADED); ctx->spu = NULL; + + if (spu_stopped(ctx, &status)) + wake_up_all(&ctx->stop_wq); } /** @@ -465,7 +509,7 @@ static void __spu_add_to_rq(struct spu_context *ctx) list_add_tail(&ctx->rq, &spu_prio->runq[ctx->prio]); set_bit(ctx->prio, spu_prio->bitmap); if (!spu_prio->nr_waiting++) - __mod_timer(&spusched_timer, jiffies + SPUSCHED_TICK); + mod_timer(&spusched_timer, jiffies + SPUSCHED_TICK); } } @@ -545,10 +589,7 @@ static struct spu *spu_get_idle(struct spu_context *ctx) goto found; mutex_unlock(&cbe_spu_info[node].list_mutex); - mutex_lock(&ctx->gang->aff_mutex); - if (atomic_dec_and_test(&ctx->gang->aff_sched_count)) - ctx->gang->aff_ref_spu = NULL; - mutex_unlock(&ctx->gang->aff_mutex); + atomic_dec(&ctx->gang->aff_sched_count); goto not_found; } mutex_unlock(&ctx->gang->aff_mutex); @@ -591,7 +632,7 @@ static struct spu *find_victim(struct spu_context *ctx) struct spu *spu; int node, n; - spu_context_nospu_trace(spu_find_vitim__enter, ctx); + spu_context_nospu_trace(spu_find_victim__enter, ctx); /* * Look for a possible preemption candidate on the local node first. @@ -613,9 +654,12 @@ static struct spu *find_victim(struct spu_context *ctx) if (tmp && tmp->prio > ctx->prio && !(tmp->flags & SPU_CREATE_NOSCHED) && - (!victim || tmp->prio > victim->prio)) + (!victim || tmp->prio > victim->prio)) { victim = spu->ctx; + } } + if (victim) + get_spu_context(victim); mutex_unlock(&cbe_spu_info[node].list_mutex); if (victim) { @@ -630,6 +674,7 @@ static struct spu *find_victim(struct spu_context *ctx) * look at another context or give up after X retries. */ if (!mutex_trylock(&victim->state_mutex)) { + put_spu_context(victim); victim = NULL; goto restart; } @@ -642,6 +687,7 @@ static struct spu *find_victim(struct spu_context *ctx) * restart the search. */ mutex_unlock(&victim->state_mutex); + put_spu_context(victim); victim = NULL; goto restart; } @@ -655,9 +701,11 @@ static struct spu *find_victim(struct spu_context *ctx) victim->stats.invol_ctx_switch++; spu->stats.invol_ctx_switch++; - spu_add_to_rq(victim); + if (test_bit(SPU_SCHED_SPU_RUN, &victim->sched_flags)) + spu_add_to_rq(victim); mutex_unlock(&victim->state_mutex); + put_spu_context(victim); return spu; } @@ -693,17 +741,33 @@ static void spu_schedule(struct spu *spu, struct spu_context *ctx) /* not a candidate for interruptible because it's called either from the scheduler thread or from spu_deactivate */ mutex_lock(&ctx->state_mutex); - __spu_schedule(spu, ctx); + if (ctx->state == SPU_STATE_SAVED) + __spu_schedule(spu, ctx); spu_release(ctx); } -static void spu_unschedule(struct spu *spu, struct spu_context *ctx) +/** + * spu_unschedule - remove a context from a spu, and possibly release it. + * @spu: The SPU to unschedule from + * @ctx: The context currently scheduled on the SPU + * @free_spu Whether to free the SPU for other contexts + * + * Unbinds the context @ctx from the SPU @spu. If @free_spu is non-zero, the + * SPU is made available for other contexts (ie, may be returned by + * spu_get_idle). If this is zero, the caller is expected to schedule another + * context to this spu. + * + * Should be called with ctx->state_mutex held. + */ +static void spu_unschedule(struct spu *spu, struct spu_context *ctx, + int free_spu) { int node = spu->node; mutex_lock(&cbe_spu_info[node].list_mutex); cbe_spu_info[node].nr_active--; - spu->alloc_state = SPU_FREE; + if (free_spu) + spu->alloc_state = SPU_FREE; spu_unbind_context(spu, ctx); ctx->stats.invol_ctx_switch++; spu->stats.invol_ctx_switch++; @@ -781,7 +845,7 @@ static struct spu_context *grab_runnable_context(int prio, int node) struct list_head *rq = &spu_prio->runq[best]; list_for_each_entry(ctx, rq, rq) { - /* XXX(hch): check for affinity here aswell */ + /* XXX(hch): check for affinity here as well */ if (__node_allowed(ctx, node)) { __spu_del_from_rq(ctx); goto found; @@ -803,7 +867,7 @@ static int __spu_deactivate(struct spu_context *ctx, int force, int max_prio) if (spu) { new = grab_runnable_context(max_prio, spu->node); if (new || force) { - spu_unschedule(spu, ctx); + spu_unschedule(spu, ctx, new == NULL); if (new) { if (new->flags & SPU_CREATE_NOSCHED) wake_up(&new->stop_wq); @@ -876,12 +940,13 @@ static noinline void spusched_tick(struct spu_context *ctx) new = grab_runnable_context(ctx->prio + 1, spu->node); if (new) { - spu_unschedule(spu, ctx); + spu_unschedule(spu, ctx, 0); if (test_bit(SPU_SCHED_SPU_RUN, &ctx->sched_flags)) spu_add_to_rq(ctx); } else { spu_context_nospu_trace(spusched_tick__newslice, ctx); - ctx->time_slice++; + if (!ctx->time_slice) + ctx->time_slice++; } out: spu_release(ctx); @@ -955,9 +1020,11 @@ static int spusched_thread(void *unused) struct spu_context *ctx = spu->ctx; if (ctx) { + get_spu_context(ctx); mutex_unlock(mtx); spusched_tick(ctx); mutex_lock(mtx); + put_spu_context(ctx); } } mutex_unlock(mtx); @@ -975,6 +1042,7 @@ void spuctx_switch_state(struct spu_context *ctx, struct timespec ts; struct spu *spu; enum spu_utilization_state old_state; + int node; ktime_get_ts(&ts); curtime = timespec_to_ns(&ts); @@ -996,6 +1064,11 @@ void spuctx_switch_state(struct spu_context *ctx, spu->stats.times[old_state] += delta; spu->stats.util_state = new_state; spu->stats.tstamp = curtime; + node = spu->node; + if (old_state == SPU_UTIL_USER) + atomic_dec(&cbe_spu_info[node].busy_spus); + if (new_state == SPU_UTIL_USER) + atomic_inc(&cbe_spu_info[node].busy_spus); } } @@ -1021,7 +1094,7 @@ static int show_spu_loadavg(struct seq_file *s, void *private) LOAD_INT(c), LOAD_FRAC(c), count_active_contexts(), atomic_read(&nr_spu_contexts), - current->nsproxy->pid_ns->last_pid); + task_active_pid_ns(current)->last_pid); return 0; } @@ -1063,10 +1136,9 @@ int __init spu_sched_init(void) mod_timer(&spuloadavg_timer, 0); - entry = create_proc_entry("spu_loadavg", 0, NULL); + entry = proc_create("spu_loadavg", 0, NULL, &spu_loadavg_fops); if (!entry) goto out_stop_kthread; - entry->proc_fops = &spu_loadavg_fops; pr_debug("spusched: tick: %d, min ticks: %d, default ticks: %d\n", SPUSCHED_TICK, MIN_SPU_TIMESLICE, DEF_SPU_TIMESLICE); diff --git a/arch/powerpc/platforms/cell/spufs/spu_restore.c b/arch/powerpc/platforms/cell/spufs/spu_restore.c index 21a9c952d88..72c905f1ee7 100644 --- a/arch/powerpc/platforms/cell/spufs/spu_restore.c +++ b/arch/powerpc/platforms/cell/spufs/spu_restore.c @@ -284,7 +284,7 @@ static inline void restore_complete(void) exit_instrs[3] = BR_INSTR; break; default: - /* SPU_Status[R]=1. No additonal instructions. */ + /* SPU_Status[R]=1. No additional instructions. */ break; } spu_sync(); diff --git a/arch/powerpc/platforms/cell/spufs/spufs.h b/arch/powerpc/platforms/cell/spufs/spufs.h index cdc515182f8..bcfd6f063ef 100644 --- a/arch/powerpc/platforms/cell/spufs/spufs.h +++ b/arch/powerpc/platforms/cell/spufs/spufs.h @@ -32,6 +32,12 @@ #include <asm/spu_csa.h> #include <asm/spu_info.h> +#define SPUFS_PS_MAP_SIZE 0x20000 +#define SPUFS_MFC_MAP_SIZE 0x1000 +#define SPUFS_CNTL_MAP_SIZE 0x1000 +#define SPUFS_SIGNAL_MAP_SIZE PAGE_SIZE +#define SPUFS_MSS_MAP_SIZE 0x1000 + /* The magic number for our file system */ enum { SPUFS_MAGIC = 0x23c9b64e, @@ -47,6 +53,29 @@ enum { SPU_SCHED_SPU_RUN, /* context is within spu_run */ }; +enum { + SWITCH_LOG_BUFSIZE = 4096, +}; + +enum { + SWITCH_LOG_START, + SWITCH_LOG_STOP, + SWITCH_LOG_EXIT, +}; + +struct switch_log { + wait_queue_head_t wait; + unsigned long head; + unsigned long tail; + struct switch_log_entry { + struct timespec tstamp; + s32 spu_id; + u32 type; + u32 val; + u64 timebase; + } log[]; +}; + struct spu_context { struct spu *spu; /* pointer to a physical SPU */ struct spu_state csa; /* SPU context save area. */ @@ -97,6 +126,7 @@ struct spu_context { cpumask_t cpus_allowed; int policy; int prio; + int last_ran; /* statistics */ struct { @@ -116,6 +146,9 @@ struct spu_context { unsigned long long libassist; } stats; + /* context switch log */ + struct switch_log *switch_log; + struct list_head aff_list; int aff_head; int aff_offset; @@ -200,17 +233,26 @@ struct spufs_inode_info { #define SPUFS_I(inode) \ container_of(inode, struct spufs_inode_info, vfs_inode) -extern struct tree_descr spufs_dir_contents[]; -extern struct tree_descr spufs_dir_nosched_contents[]; +struct spufs_tree_descr { + const char *name; + const struct file_operations *ops; + umode_t mode; + size_t size; +}; + +extern const struct spufs_tree_descr spufs_dir_contents[]; +extern const struct spufs_tree_descr spufs_dir_nosched_contents[]; +extern const struct spufs_tree_descr spufs_dir_debug_contents[]; /* system call implementation */ extern struct spufs_calls spufs_calls; +struct coredump_params; long spufs_run_spu(struct spu_context *ctx, u32 *npc, u32 *status); -long spufs_create(struct nameidata *nd, unsigned int flags, - mode_t mode, struct file *filp); +long spufs_create(struct path *nd, struct dentry *dentry, unsigned int flags, + umode_t mode, struct file *filp); /* ELF coredump callbacks for writing SPU ELF notes */ extern int spufs_coredump_extra_notes_size(void); -extern int spufs_coredump_extra_notes_write(struct file *file, loff_t *foffset); +extern int spufs_coredump_extra_notes_write(struct coredump_params *cprm); extern const struct file_operations spufs_context_fops; @@ -256,6 +298,8 @@ int spu_activate(struct spu_context *ctx, unsigned long flags); void spu_deactivate(struct spu_context *ctx); void spu_yield(struct spu_context *ctx); void spu_switch_notify(struct spu *spu, struct spu_context *ctx); +void spu_switch_log_notify(struct spu *spu, struct spu_context *ctx, + u32 type, u32 val); void spu_set_timeslice(struct spu_context *ctx); void spu_update_sched_info(struct spu_context *ctx); void __spu_update_sched_info(struct spu_context *ctx); @@ -270,7 +314,7 @@ extern char *isolated_loader; * we need to call spu_release(ctx) before sleeping, and * then spu_acquire(ctx) when awoken. * - * Returns with state_mutex re-acquired when successfull or + * Returns with state_mutex re-acquired when successful or * with -ERESTARTSYS and the state_mutex dropped when interrupted. */ @@ -302,7 +346,7 @@ size_t spu_ibox_read(struct spu_context *ctx, u32 *data); /* irq callback funcs. */ void spufs_ibox_callback(struct spu *spu); void spufs_wbox_callback(struct spu *spu); -void spufs_stop_callback(struct spu *spu); +void spufs_stop_callback(struct spu *spu, int irq); void spufs_mfc_callback(struct spu *spu); void spufs_dma_callback(struct spu *spu, int type); @@ -314,7 +358,7 @@ struct spufs_coredump_reader { u64 (*get)(struct spu_context *ctx); size_t size; }; -extern struct spufs_coredump_reader spufs_coredump_read[]; +extern const struct spufs_coredump_reader spufs_coredump_read[]; extern int spufs_coredump_num_notes; extern int spu_init_csa(struct spu_state *csa); @@ -329,9 +373,4 @@ extern void spu_free_lscsa(struct spu_state *csa); extern void spuctx_switch_state(struct spu_context *ctx, enum spu_utilization_state new_state); -#define spu_context_trace(name, ctx, spu) \ - trace_mark(name, "%p %p", ctx, spu); -#define spu_context_nospu_trace(name, ctx) \ - trace_mark(name, "%p", ctx); - #endif diff --git a/arch/powerpc/platforms/cell/spufs/sputrace.c b/arch/powerpc/platforms/cell/spufs/sputrace.c deleted file mode 100644 index 79aa773f3c9..00000000000 --- a/arch/powerpc/platforms/cell/spufs/sputrace.c +++ /dev/null @@ -1,242 +0,0 @@ -/* - * Copyright (C) 2007 IBM Deutschland Entwicklung GmbH - * Released under GPL v2. - * - * Partially based on net/ipv4/tcp_probe.c. - * - * Simple tracing facility for spu contexts. - */ -#include <linux/sched.h> -#include <linux/kernel.h> -#include <linux/module.h> -#include <linux/marker.h> -#include <linux/proc_fs.h> -#include <linux/wait.h> -#include <asm/atomic.h> -#include <asm/uaccess.h> -#include "spufs.h" - -struct spu_probe { - const char *name; - const char *format; - marker_probe_func *probe_func; -}; - -struct sputrace { - ktime_t tstamp; - int owner_tid; /* owner */ - int curr_tid; - const char *name; - int number; -}; - -static int bufsize __read_mostly = 16384; -MODULE_PARM_DESC(bufsize, "Log buffer size (number of records)"); -module_param(bufsize, int, 0); - - -static DEFINE_SPINLOCK(sputrace_lock); -static DECLARE_WAIT_QUEUE_HEAD(sputrace_wait); -static ktime_t sputrace_start; -static unsigned long sputrace_head, sputrace_tail; -static struct sputrace *sputrace_log; - -static int sputrace_used(void) -{ - return (sputrace_head - sputrace_tail) % bufsize; -} - -static inline int sputrace_avail(void) -{ - return bufsize - sputrace_used(); -} - -static int sputrace_sprint(char *tbuf, int n) -{ - const struct sputrace *t = sputrace_log + sputrace_tail % bufsize; - struct timespec tv = - ktime_to_timespec(ktime_sub(t->tstamp, sputrace_start)); - - return snprintf(tbuf, n, - "[%lu.%09lu] %d: %s (ctxthread = %d, spu = %d)\n", - (unsigned long) tv.tv_sec, - (unsigned long) tv.tv_nsec, - t->curr_tid, - t->name, - t->owner_tid, - t->number); -} - -static ssize_t sputrace_read(struct file *file, char __user *buf, - size_t len, loff_t *ppos) -{ - int error = 0, cnt = 0; - - if (!buf || len < 0) - return -EINVAL; - - while (cnt < len) { - char tbuf[128]; - int width; - - error = wait_event_interruptible(sputrace_wait, - sputrace_used() > 0); - if (error) - break; - - spin_lock(&sputrace_lock); - if (sputrace_head == sputrace_tail) { - spin_unlock(&sputrace_lock); - continue; - } - - width = sputrace_sprint(tbuf, sizeof(tbuf)); - if (width < len) - sputrace_tail = (sputrace_tail + 1) % bufsize; - spin_unlock(&sputrace_lock); - - if (width >= len) - break; - - error = copy_to_user(buf + cnt, tbuf, width); - if (error) - break; - cnt += width; - } - - return cnt == 0 ? error : cnt; -} - -static int sputrace_open(struct inode *inode, struct file *file) -{ - spin_lock(&sputrace_lock); - sputrace_head = sputrace_tail = 0; - sputrace_start = ktime_get(); - spin_unlock(&sputrace_lock); - - return 0; -} - -static const struct file_operations sputrace_fops = { - .owner = THIS_MODULE, - .open = sputrace_open, - .read = sputrace_read, -}; - -static void sputrace_log_item(const char *name, struct spu_context *ctx, - struct spu *spu) -{ - spin_lock(&sputrace_lock); - if (sputrace_avail() > 1) { - struct sputrace *t = sputrace_log + sputrace_head; - - t->tstamp = ktime_get(); - t->owner_tid = ctx->tid; - t->name = name; - t->curr_tid = current->pid; - t->number = spu ? spu->number : -1; - - sputrace_head = (sputrace_head + 1) % bufsize; - } else { - printk(KERN_WARNING - "sputrace: lost samples due to full buffer.\n"); - } - spin_unlock(&sputrace_lock); - - wake_up(&sputrace_wait); -} - -static void spu_context_event(void *probe_private, void *call_data, - const char *format, va_list *args) -{ - struct spu_probe *p = probe_private; - struct spu_context *ctx; - struct spu *spu; - - ctx = va_arg(*args, struct spu_context *); - spu = va_arg(*args, struct spu *); - - sputrace_log_item(p->name, ctx, spu); -} - -static void spu_context_nospu_event(void *probe_private, void *call_data, - const char *format, va_list *args) -{ - struct spu_probe *p = probe_private; - struct spu_context *ctx; - - ctx = va_arg(*args, struct spu_context *); - - sputrace_log_item(p->name, ctx, NULL); -} - -struct spu_probe spu_probes[] = { - { "spu_bind_context__enter", "%p %p", spu_context_event }, - { "spu_unbind_context__enter", "%p %p", spu_context_event }, - { "spu_get_idle__enter", "%p", spu_context_nospu_event }, - { "spu_get_idle__found", "%p %p", spu_context_event }, - { "spu_get_idle__not_found", "%p", spu_context_nospu_event }, - { "spu_find_victim__enter", "%p", spu_context_nospu_event }, - { "spusched_tick__preempt", "%p %p", spu_context_event }, - { "spusched_tick__newslice", "%p", spu_context_nospu_event }, - { "spu_yield__enter", "%p", spu_context_nospu_event }, - { "spu_deactivate__enter", "%p", spu_context_nospu_event }, - { "__spu_deactivate__unload", "%p %p", spu_context_event }, - { "spufs_ps_nopfn__enter", "%p", spu_context_nospu_event }, - { "spufs_ps_nopfn__sleep", "%p", spu_context_nospu_event }, - { "spufs_ps_nopfn__wake", "%p %p", spu_context_event }, - { "spufs_ps_nopfn__insert", "%p %p", spu_context_event }, - { "spu_acquire_saved__enter", "%p", spu_context_nospu_event }, - { "destroy_spu_context__enter", "%p", spu_context_nospu_event }, - { "spufs_stop_callback__enter", "%p %p", spu_context_event }, -}; - -static int __init sputrace_init(void) -{ - struct proc_dir_entry *entry; - int i, error = -ENOMEM; - - sputrace_log = kcalloc(sizeof(struct sputrace), - bufsize, GFP_KERNEL); - if (!sputrace_log) - goto out; - - entry = create_proc_entry("sputrace", S_IRUSR, NULL); - if (!entry) - goto out_free_log; - entry->proc_fops = &sputrace_fops; - - for (i = 0; i < ARRAY_SIZE(spu_probes); i++) { - struct spu_probe *p = &spu_probes[i]; - - error = marker_probe_register(p->name, p->format, - p->probe_func, p); - if (error) - printk(KERN_INFO "Unable to register probe %s\n", - p->name); - } - - return 0; - -out_free_log: - kfree(sputrace_log); -out: - return -ENOMEM; -} - -static void __exit sputrace_exit(void) -{ - int i; - - for (i = 0; i < ARRAY_SIZE(spu_probes); i++) - marker_probe_unregister(spu_probes[i].name, - spu_probes[i].probe_func, &spu_probes[i]); - - remove_proc_entry("sputrace", NULL); - kfree(sputrace_log); -} - -module_init(sputrace_init); -module_exit(sputrace_exit); - -MODULE_LICENSE("GPL"); diff --git a/arch/powerpc/platforms/cell/spufs/sputrace.h b/arch/powerpc/platforms/cell/spufs/sputrace.h new file mode 100644 index 00000000000..db2656aa410 --- /dev/null +++ b/arch/powerpc/platforms/cell/spufs/sputrace.h @@ -0,0 +1,39 @@ +#if !defined(_TRACE_SPUFS_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_SPUFS_H + +#include <linux/tracepoint.h> + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM spufs + +TRACE_EVENT(spufs_context, + TP_PROTO(struct spu_context *ctx, struct spu *spu, const char *name), + TP_ARGS(ctx, spu, name), + + TP_STRUCT__entry( + __field(const char *, name) + __field(int, owner_tid) + __field(int, number) + ), + + TP_fast_assign( + __entry->name = name; + __entry->owner_tid = ctx->tid; + __entry->number = spu ? spu->number : -1; + ), + + TP_printk("%s (ctxthread = %d, spu = %d)", + __entry->name, __entry->owner_tid, __entry->number) +); + +#define spu_context_trace(name, ctx, spu) \ + trace_spufs_context(ctx, spu, __stringify(name)) +#define spu_context_nospu_trace(name, ctx) \ + trace_spufs_context(ctx, NULL, __stringify(name)) + +#endif /* _TRACE_SPUFS_H */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH . +#define TRACE_INCLUDE_FILE sputrace +#include <trace/define_trace.h> diff --git a/arch/powerpc/platforms/cell/spufs/switch.c b/arch/powerpc/platforms/cell/spufs/switch.c index e9dc7a55d1b..dde35551e74 100644 --- a/arch/powerpc/platforms/cell/spufs/switch.c +++ b/arch/powerpc/platforms/cell/spufs/switch.c @@ -32,7 +32,7 @@ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ -#include <linux/module.h> +#include <linux/export.h> #include <linux/errno.h> #include <linux/hardirq.h> #include <linux/sched.h> @@ -132,6 +132,14 @@ static inline void disable_interrupts(struct spu_state *csa, struct spu *spu) spu_int_mask_set(spu, 2, 0ul); eieio(); spin_unlock_irq(&spu->register_lock); + + /* + * This flag needs to be set before calling synchronize_irq so + * that the update will be visible to the relevant handlers + * via a simple load. + */ + set_bit(SPU_CONTEXT_SWITCH_PENDING, &spu->flags); + clear_bit(SPU_CONTEXT_FAULT_PENDING, &spu->flags); synchronize_irq(spu->irqs[0]); synchronize_irq(spu->irqs[1]); synchronize_irq(spu->irqs[2]); @@ -166,9 +174,8 @@ static inline void set_switch_pending(struct spu_state *csa, struct spu *spu) /* Save, Step 7: * Restore, Step 5: * Set a software context switch pending flag. + * Done above in Step 3 - disable_interrupts(). */ - set_bit(SPU_CONTEXT_SWITCH_PENDING, &spu->flags); - mb(); } static inline void save_mfc_cntl(struct spu_state *csa, struct spu *spu) @@ -186,20 +193,21 @@ static inline void save_mfc_cntl(struct spu_state *csa, struct spu *spu) MFC_CNTL_SUSPEND_COMPLETE); /* fall through */ case MFC_CNTL_SUSPEND_COMPLETE: - if (csa) { + if (csa) csa->priv2.mfc_control_RW = - MFC_CNTL_SUSPEND_MASK | + in_be64(&priv2->mfc_control_RW) | MFC_CNTL_SUSPEND_DMA_QUEUE; - } break; case MFC_CNTL_NORMAL_DMA_QUEUE_OPERATION: out_be64(&priv2->mfc_control_RW, MFC_CNTL_SUSPEND_DMA_QUEUE); POLL_WHILE_FALSE((in_be64(&priv2->mfc_control_RW) & MFC_CNTL_SUSPEND_DMA_STATUS_MASK) == MFC_CNTL_SUSPEND_COMPLETE); - if (csa) { - csa->priv2.mfc_control_RW = 0; - } + if (csa) + csa->priv2.mfc_control_RW = + in_be64(&priv2->mfc_control_RW) & + ~MFC_CNTL_SUSPEND_DMA_QUEUE & + ~MFC_CNTL_SUSPEND_MASK; break; } } @@ -249,16 +257,21 @@ static inline void save_spu_status(struct spu_state *csa, struct spu *spu) } } -static inline void save_mfc_decr(struct spu_state *csa, struct spu *spu) +static inline void save_mfc_stopped_status(struct spu_state *csa, + struct spu *spu) { struct spu_priv2 __iomem *priv2 = spu->priv2; + const u64 mask = MFC_CNTL_DECREMENTER_RUNNING | + MFC_CNTL_DMA_QUEUES_EMPTY; /* Save, Step 12: * Read MFC_CNTL[Ds]. Update saved copy of * CSA.MFC_CNTL[Ds]. + * + * update: do the same with MFC_CNTL[Q]. */ - csa->priv2.mfc_control_RW |= - in_be64(&priv2->mfc_control_RW) & MFC_CNTL_DECREMENTER_RUNNING; + csa->priv2.mfc_control_RW &= ~mask; + csa->priv2.mfc_control_RW |= in_be64(&priv2->mfc_control_RW) & mask; } static inline void halt_mfc_decr(struct spu_state *csa, struct spu *spu) @@ -462,7 +475,9 @@ static inline void purge_mfc_queue(struct spu_state *csa, struct spu *spu) * Restore, Step 14. * Write MFC_CNTL[Pc]=1 (purge queue). */ - out_be64(&priv2->mfc_control_RW, MFC_CNTL_PURGE_DMA_REQUEST); + out_be64(&priv2->mfc_control_RW, + MFC_CNTL_PURGE_DMA_REQUEST | + MFC_CNTL_SUSPEND_MASK); eieio(); } @@ -725,10 +740,14 @@ static inline void set_switch_active(struct spu_state *csa, struct spu *spu) /* Save, Step 48: * Restore, Step 23. * Change the software context switch pending flag - * to context switch active. + * to context switch active. This implementation does + * not uses a switch active flag. * - * This implementation does not uses a switch active flag. + * Now that we have saved the mfc in the csa, we can add in the + * restart command if an exception occurred. */ + if (test_bit(SPU_CONTEXT_FAULT_PENDING, &spu->flags)) + csa->priv2.mfc_control_RW |= MFC_CNTL_RESTART_DMA_COMMAND; clear_bit(SPU_CONTEXT_SWITCH_PENDING, &spu->flags); mb(); } @@ -1690,6 +1709,13 @@ static inline void restore_mfc_sr1(struct spu_state *csa, struct spu *spu) eieio(); } +static inline void set_int_route(struct spu_state *csa, struct spu *spu) +{ + struct spu_context *ctx = spu->ctx; + + spu_cpu_affinity_set(spu, ctx->last_ran); +} + static inline void restore_other_spu_access(struct spu_state *csa, struct spu *spu) { @@ -1721,15 +1747,15 @@ static inline void restore_mfc_cntl(struct spu_state *csa, struct spu *spu) */ out_be64(&priv2->mfc_control_RW, csa->priv2.mfc_control_RW); eieio(); + /* - * FIXME: this is to restart a DMA that we were processing - * before the save. better remember the fault information - * in the csa instead. + * The queue is put back into the same state that was evident prior to + * the context switch. The suspend flag is added to the saved state in + * the csa, if the operational state was suspending or suspended. In + * this case, the code that suspended the mfc is responsible for + * continuing it. Note that SPE faults do not change the operational + * state of the spu. */ - if ((csa->priv2.mfc_control_RW & MFC_CNTL_SUSPEND_DMA_QUEUE_MASK)) { - out_be64(&priv2->mfc_control_RW, MFC_CNTL_RESTART_DMA_COMMAND); - eieio(); - } } static inline void enable_user_access(struct spu_state *csa, struct spu *spu) @@ -1788,7 +1814,7 @@ static int quiece_spu(struct spu_state *prev, struct spu *spu) save_spu_runcntl(prev, spu); /* Step 9. */ save_mfc_sr1(prev, spu); /* Step 10. */ save_spu_status(prev, spu); /* Step 11. */ - save_mfc_decr(prev, spu); /* Step 12. */ + save_mfc_stopped_status(prev, spu); /* Step 12. */ halt_mfc_decr(prev, spu); /* Step 13. */ save_timebase(prev, spu); /* Step 14. */ remove_other_spu_access(prev, spu); /* Step 15. */ @@ -1815,6 +1841,7 @@ static void save_csa(struct spu_state *prev, struct spu *spu) save_mfc_csr_ato(prev, spu); /* Step 24. */ save_mfc_tclass_id(prev, spu); /* Step 25. */ set_mfc_tclass_id(prev, spu); /* Step 26. */ + save_mfc_cmd(prev, spu); /* Step 26a - moved from 44. */ purge_mfc_queue(prev, spu); /* Step 27. */ wait_purge_complete(prev, spu); /* Step 28. */ setup_mfc_sr1(prev, spu); /* Step 30. */ @@ -1831,7 +1858,6 @@ static void save_csa(struct spu_state *prev, struct spu *spu) save_ppuint_mb(prev, spu); /* Step 41. */ save_ch_part1(prev, spu); /* Step 42. */ save_spu_mb(prev, spu); /* Step 43. */ - save_mfc_cmd(prev, spu); /* Step 44. */ reset_ch(prev, spu); /* Step 45. */ } @@ -2000,6 +2026,7 @@ static void restore_csa(struct spu_state *next, struct spu *spu) check_ppuint_mb_stat(next, spu); /* Step 67. */ spu_invalidate_slbs(spu); /* Modified Step 68. */ restore_mfc_sr1(next, spu); /* Step 69. */ + set_int_route(next, spu); /* NEW */ restore_other_spu_access(next, spu); /* Step 70. */ restore_spu_runcntl(next, spu); /* Step 71. */ restore_mfc_cntl(next, spu); /* Step 72. */ diff --git a/arch/powerpc/platforms/cell/spufs/syscalls.c b/arch/powerpc/platforms/cell/spufs/syscalls.c index 49c87769b1f..a87200a535f 100644 --- a/arch/powerpc/platforms/cell/spufs/syscalls.c +++ b/arch/powerpc/platforms/cell/spufs/syscalls.c @@ -1,8 +1,9 @@ #include <linux/file.h> #include <linux/fs.h> -#include <linux/module.h> +#include <linux/export.h> #include <linux/mount.h> #include <linux/namei.h> +#include <linux/slab.h> #include <asm/uaccess.h> @@ -46,7 +47,7 @@ static long do_spu_run(struct file *filp, if (filp->f_op != &spufs_context_fops) goto out; - i = SPUFS_I(filp->f_path.dentry->d_inode); + i = SPUFS_I(file_inode(filp)); ret = spufs_run_spu(i->i_ctx, &npc, &status); if (put_user(npc, unpc)) @@ -59,23 +60,17 @@ out: } static long do_spu_create(const char __user *pathname, unsigned int flags, - mode_t mode, struct file *neighbor) + umode_t mode, struct file *neighbor) { - char *tmp; + struct path path; + struct dentry *dentry; int ret; - tmp = getname(pathname); - ret = PTR_ERR(tmp); - if (!IS_ERR(tmp)) { - struct nameidata nd; - - ret = path_lookup(tmp, LOOKUP_PARENT| - LOOKUP_OPEN|LOOKUP_CREATE, &nd); - if (!ret) { - ret = spufs_create(&nd, flags, mode, neighbor); - path_put(&nd.path); - } - putname(tmp); + dentry = user_path_create(AT_FDCWD, pathname, &path, LOOKUP_DIRECTORY); + ret = PTR_ERR(dentry); + if (!IS_ERR(dentry)) { + ret = spufs_create(&path, dentry, flags, mode, neighbor); + done_path_create(&path, dentry); } return ret; @@ -84,8 +79,10 @@ static long do_spu_create(const char __user *pathname, unsigned int flags, struct spufs_calls spufs_calls = { .create_thread = do_spu_create, .spu_run = do_spu_run, - .coredump_extra_notes_size = spufs_coredump_extra_notes_size, - .coredump_extra_notes_write = spufs_coredump_extra_notes_write, .notify_spus_active = do_notify_spus_active, .owner = THIS_MODULE, +#ifdef CONFIG_COREDUMP + .coredump_extra_notes_size = spufs_coredump_extra_notes_size, + .coredump_extra_notes_write = spufs_coredump_extra_notes_write, +#endif }; |