diff options
Diffstat (limited to 'arch/arc/kernel')
| -rw-r--r-- | arch/arc/kernel/Makefile | 1 | ||||
| -rw-r--r-- | arch/arc/kernel/ctx_sw_asm.S | 4 | ||||
| -rw-r--r-- | arch/arc/kernel/devtree.c | 4 | ||||
| -rw-r--r-- | arch/arc/kernel/entry.S | 72 | ||||
| -rw-r--r-- | arch/arc/kernel/head.S | 78 | ||||
| -rw-r--r-- | arch/arc/kernel/irq.c | 18 | ||||
| -rw-r--r-- | arch/arc/kernel/kprobes.c | 2 | ||||
| -rw-r--r-- | arch/arc/kernel/perf_event.c | 326 | ||||
| -rw-r--r-- | arch/arc/kernel/process.c | 23 | ||||
| -rw-r--r-- | arch/arc/kernel/ptrace.c | 4 | ||||
| -rw-r--r-- | arch/arc/kernel/setup.c | 48 | ||||
| -rw-r--r-- | arch/arc/kernel/smp.c | 141 | ||||
| -rw-r--r-- | arch/arc/kernel/time.c | 46 | ||||
| -rw-r--r-- | arch/arc/kernel/troubleshoot.c | 10 | ||||
| -rw-r--r-- | arch/arc/kernel/vmlinux.lds.S | 2 |
15 files changed, 594 insertions, 185 deletions
diff --git a/arch/arc/kernel/Makefile b/arch/arc/kernel/Makefile index c242ef07ba7..8004b4fa646 100644 --- a/arch/arc/kernel/Makefile +++ b/arch/arc/kernel/Makefile @@ -19,6 +19,7 @@ obj-$(CONFIG_KPROBES) += kprobes.o obj-$(CONFIG_ARC_MISALIGN_ACCESS) += unaligned.o obj-$(CONFIG_KGDB) += kgdb.o obj-$(CONFIG_ARC_METAWARE_HLINK) += arc_hostlink.o +obj-$(CONFIG_PERF_EVENTS) += perf_event.o obj-$(CONFIG_ARC_FPU_SAVE_RESTORE) += fpu.o CFLAGS_fpu.o += -mdpfp diff --git a/arch/arc/kernel/ctx_sw_asm.S b/arch/arc/kernel/ctx_sw_asm.S index 65690e7fcc8..e248594097e 100644 --- a/arch/arc/kernel/ctx_sw_asm.S +++ b/arch/arc/kernel/ctx_sw_asm.S @@ -10,9 +10,9 @@ * -This is the more "natural" hand written assembler */ +#include <linux/linkage.h> #include <asm/entry.h> /* For the SAVE_* macros */ #include <asm/asm-offsets.h> -#include <asm/linkage.h> #define KSP_WORD_OFF ((TASK_THREAD + THREAD_KSP) / 4) @@ -62,4 +62,4 @@ __switch_to: ld.ab blink, [sp, 4] j [blink] -ARC_EXIT __switch_to +END(__switch_to) diff --git a/arch/arc/kernel/devtree.c b/arch/arc/kernel/devtree.c index b6dc4e21fd3..fffdb5e41b2 100644 --- a/arch/arc/kernel/devtree.c +++ b/arch/arc/kernel/devtree.c @@ -41,8 +41,8 @@ const struct machine_desc * __init setup_machine_fdt(void *dt) { const struct machine_desc *mdesc; unsigned long dt_root; - void *clk; - unsigned long len; + const void *clk; + int len; if (!early_init_dt_scan(dt)) return NULL; diff --git a/arch/arc/kernel/entry.S b/arch/arc/kernel/entry.S index 47d09d07f09..83a046a7cd0 100644 --- a/arch/arc/kernel/entry.S +++ b/arch/arc/kernel/entry.S @@ -141,7 +141,7 @@ VECTOR EV_Extension ; 0x130, Extn Intruction Excp (0x26) VECTOR reserved ; Reserved Exceptions .endr -#include <linux/linkage.h> /* ARC_{EXTRY,EXIT} */ +#include <linux/linkage.h> /* {EXTRY,EXIT} */ #include <asm/entry.h> /* SAVE_ALL_{INT1,INT2,SYS...} */ #include <asm/errno.h> #include <asm/arcregs.h> @@ -156,7 +156,7 @@ ARCFP_DATA int1_saved_reg int1_saved_reg: .zero 4 -/* Each Interrupt level needs it's own scratch */ +/* Each Interrupt level needs its own scratch */ #ifdef CONFIG_ARC_COMPACT_IRQ_LEVELS ARCFP_DATA int2_saved_reg @@ -184,7 +184,7 @@ reserved: ; processor restart ; --------------------------------------------- ; Level 2 ISR: Can interrupt a Level 1 ISR ; --------------------------------------------- -ARC_ENTRY handle_interrupt_level2 +ENTRY(handle_interrupt_level2) ; TODO-vineetg for SMP this wont work ; free up r9 as scratchpad @@ -225,14 +225,14 @@ ARC_ENTRY handle_interrupt_level2 b ret_from_exception -ARC_EXIT handle_interrupt_level2 +END(handle_interrupt_level2) #endif ; --------------------------------------------- ; Level 1 ISR ; --------------------------------------------- -ARC_ENTRY handle_interrupt_level1 +ENTRY(handle_interrupt_level1) /* free up r9 as scratchpad */ #ifdef CONFIG_SMP @@ -265,7 +265,7 @@ ARC_ENTRY handle_interrupt_level1 sr r8, [AUX_IRQ_LV12] ; clear bit in Sticky Status Reg b ret_from_exception -ARC_EXIT handle_interrupt_level1 +END(handle_interrupt_level1) ;################### Non TLB Exception Handling ############################# @@ -273,7 +273,7 @@ ARC_EXIT handle_interrupt_level1 ; Instruction Error Exception Handler ; --------------------------------------------- -ARC_ENTRY instr_service +ENTRY(instr_service) EXCEPTION_PROLOGUE @@ -284,13 +284,13 @@ ARC_ENTRY instr_service bl do_insterror_or_kprobe b ret_from_exception -ARC_EXIT instr_service +END(instr_service) ; --------------------------------------------- ; Memory Error Exception Handler ; --------------------------------------------- -ARC_ENTRY mem_service +ENTRY(mem_service) EXCEPTION_PROLOGUE @@ -301,13 +301,13 @@ ARC_ENTRY mem_service bl do_memory_error b ret_from_exception -ARC_EXIT mem_service +END(mem_service) ; --------------------------------------------- ; Machine Check Exception Handler ; --------------------------------------------- -ARC_ENTRY EV_MachineCheck +ENTRY(EV_MachineCheck) EXCEPTION_PROLOGUE @@ -331,13 +331,13 @@ ARC_ENTRY EV_MachineCheck j do_machine_check_fault -ARC_EXIT EV_MachineCheck +END(EV_MachineCheck) ; --------------------------------------------- ; Protection Violation Exception Handler ; --------------------------------------------- -ARC_ENTRY EV_TLBProtV +ENTRY(EV_TLBProtV) EXCEPTION_PROLOGUE @@ -385,12 +385,12 @@ ARC_ENTRY EV_TLBProtV b ret_from_exception -ARC_EXIT EV_TLBProtV +END(EV_TLBProtV) ; --------------------------------------------- ; Privilege Violation Exception Handler ; --------------------------------------------- -ARC_ENTRY EV_PrivilegeV +ENTRY(EV_PrivilegeV) EXCEPTION_PROLOGUE @@ -401,12 +401,12 @@ ARC_ENTRY EV_PrivilegeV bl do_privilege_fault b ret_from_exception -ARC_EXIT EV_PrivilegeV +END(EV_PrivilegeV) ; --------------------------------------------- ; Extension Instruction Exception Handler ; --------------------------------------------- -ARC_ENTRY EV_Extension +ENTRY(EV_Extension) EXCEPTION_PROLOGUE @@ -417,7 +417,7 @@ ARC_ENTRY EV_Extension bl do_extension_fault b ret_from_exception -ARC_EXIT EV_Extension +END(EV_Extension) ;######################### System Call Tracing ######################### @@ -473,7 +473,7 @@ trap_with_param: lr r0, [efa] mov r1, sp - ; Now that we have read EFA, its safe to do "fake" rtie + ; Now that we have read EFA, it is safe to do "fake" rtie ; and get out of CPU exception mode FAKE_RET_FROM_EXCPN r11 @@ -504,7 +504,7 @@ trap_with_param: ; (2) Break Points ;------------------------------------------------------------------ -ARC_ENTRY EV_Trap +ENTRY(EV_Trap) EXCEPTION_PROLOGUE @@ -534,9 +534,9 @@ ARC_ENTRY EV_Trap jl [r9] ; Entry into Sys Call Handler ; fall through to ret_from_system_call -ARC_EXIT EV_Trap +END(EV_Trap) -ARC_ENTRY ret_from_system_call +ENTRY(ret_from_system_call) st r0, [sp, PT_r0] ; sys call return value in pt_regs @@ -546,7 +546,7 @@ ARC_ENTRY ret_from_system_call ; ; If ret to user mode do we need to handle signals, schedule() et al. -ARC_ENTRY ret_from_exception +ENTRY(ret_from_exception) ; Pre-{IRQ,Trap,Exception} K/U mode from pt_regs->status32 ld r8, [sp, PT_status32] ; returning to User/Kernel Mode @@ -614,11 +614,13 @@ resume_user_mode_begin: resume_kernel_mode: -#ifdef CONFIG_PREEMPT - - ; This is a must for preempt_schedule_irq() + ; Disable Interrupts from this point on + ; CONFIG_PREEMPT: This is a must for preempt_schedule_irq() + ; !CONFIG_PREEMPT: To ensure restore_regs is intr safe IRQ_DISABLE r9 +#ifdef CONFIG_PREEMPT + ; Can't preempt if preemption disabled GET_CURR_THR_INFO_FROM_SP r10 ld r8, [r10, THREAD_INFO_PREEMPT_COUNT] @@ -676,9 +678,9 @@ not_exception: brne r9, event_IRQ2, 149f ;------------------------------------------------------------------ - ; if L2 IRQ interrupted a L1 ISR, we'd disbaled preemption earlier - ; so that sched doesnt move to new task, causing L1 to be delayed - ; undeterministically. Now that we've achieved that, lets reset + ; if L2 IRQ interrupted an L1 ISR, we'd disabled preemption earlier + ; so that sched doesn't move to new task, causing L1 to be delayed + ; undeterministically. Now that we've achieved that, let's reset ; things to what they were, before returning from L2 context ;---------------------------------------------------------------- @@ -726,15 +728,15 @@ not_level1_interrupt: debug_marker_syscall: rtie -ARC_EXIT ret_from_exception +END(ret_from_exception) -ARC_ENTRY ret_from_fork +ENTRY(ret_from_fork) ; when the forked child comes here from the __switch_to function ; r0 has the last task pointer. ; put last task in scheduler queue bl @schedule_tail - ; If kernel thread, jump to it's entry-point + ; If kernel thread, jump to its entry-point ld r9, [sp, PT_status32] brne r9, 0, 1f @@ -745,11 +747,11 @@ ARC_ENTRY ret_from_fork ; special case of kernel_thread entry point returning back due to ; kernel_execve() - pretend return from syscall to ret to userland b ret_from_exception -ARC_EXIT ret_from_fork +END(ret_from_fork) ;################### Special Sys Call Wrappers ########################## -ARC_ENTRY sys_clone_wrapper +ENTRY(sys_clone_wrapper) SAVE_CALLEE_SAVED_USER bl @sys_clone DISCARD_CALLEE_SAVED_USER @@ -759,7 +761,7 @@ ARC_ENTRY sys_clone_wrapper bnz tracesys_exit b ret_from_system_call -ARC_EXIT sys_clone_wrapper +END(sys_clone_wrapper) #ifdef CONFIG_ARC_DW2_UNWIND ; Workaround for bug 94179 (STAR ): diff --git a/arch/arc/kernel/head.S b/arch/arc/kernel/head.S index 2c878e964a6..4d2481bd8b9 100644 --- a/arch/arc/kernel/head.S +++ b/arch/arc/kernel/head.S @@ -12,10 +12,42 @@ * to skip certain things during boot on simulator */ +#include <linux/linkage.h> #include <asm/asm-offsets.h> #include <asm/entry.h> -#include <linux/linkage.h> #include <asm/arcregs.h> +#include <asm/cache.h> + +.macro CPU_EARLY_SETUP + + ; Setting up Vectror Table (in case exception happens in early boot + sr @_int_vec_base_lds, [AUX_INTR_VEC_BASE] + + ; Disable I-cache/D-cache if kernel so configured + lr r5, [ARC_REG_IC_BCR] + breq r5, 0, 1f ; I$ doesn't exist + lr r5, [ARC_REG_IC_CTRL] +#ifdef CONFIG_ARC_HAS_ICACHE + bclr r5, r5, 0 ; 0 - Enable, 1 is Disable +#else + bset r5, r5, 0 ; I$ exists, but is not used +#endif + sr r5, [ARC_REG_IC_CTRL] + +1: + lr r5, [ARC_REG_DC_BCR] + breq r5, 0, 1f ; D$ doesn't exist + lr r5, [ARC_REG_DC_CTRL] + bclr r5, r5, 6 ; Invalidate (discard w/o wback) +#ifdef CONFIG_ARC_HAS_DCACHE + bclr r5, r5, 0 ; Enable (+Inv) +#else + bset r5, r5, 0 ; Disable (+Inv) +#endif + sr r5, [ARC_REG_DC_CTRL] + +1: +.endm .cpu A7 @@ -24,13 +56,13 @@ .globl stext stext: ;------------------------------------------------------------------- - ; Don't clobber r0-r4 yet. It might have bootloader provided info + ; Don't clobber r0-r2 yet. It might have bootloader provided info ;------------------------------------------------------------------- - sr @_int_vec_base_lds, [AUX_INTR_VEC_BASE] + CPU_EARLY_SETUP #ifdef CONFIG_SMP - ; Only Boot (Master) proceeds. Others wait in platform dependent way + ; Ensure Boot (Master) proceeds. Others wait in platform dependent way ; IDENTITY Reg [ 3 2 1 0 ] ; (cpu-id) ^^^ => Zero for UP ARC700 ; => #Core-ID if SMP (Master 0) @@ -39,35 +71,25 @@ stext: ; need to make sure only boot cpu takes this path. GET_CPU_ID r5 cmp r5, 0 - jnz arc_platform_smp_wait_to_boot + mov.ne r0, r5 + jne arc_platform_smp_wait_to_boot #endif ; Clear BSS before updating any globals ; XXX: use ZOL here mov r5, __bss_start - mov r6, __bss_stop + sub r6, __bss_stop, r5 + lsr.f lp_count, r6, 2 + lpnz 1f + st.ab 0, [r5, 4] 1: - st.ab 0, [r5,4] - brlt r5, r6, 1b -#ifdef CONFIG_CMDLINE_UBOOT - ; support for bootloader provided cmdline - ; If cmdline passed by u-boot, then - ; r0 = 1 (because ATAGS parsing, now retired, used to use 0) - ; r1 = magic number (board identity) - ; r2 = addr of cmdline string (somewhere in memory/flash) - - brne r0, 1, .Lother_bootup_chores ; u-boot didn't pass cmdline - breq r2, 0, .Lother_bootup_chores ; or cmdline is NULL - - mov r5, @command_line -1: - ldb.ab r6, [r2, 1] - breq r6, 0, .Lother_bootup_chores - b.d 1b - stb.ab r6, [r5, 1] -#endif - -.Lother_bootup_chores: + ; Uboot - kernel ABI + ; r0 = [0] No uboot interaction, [1] cmdline in r2, [2] DTB in r2 + ; r1 = magic number (board identity, unused as of now + ; r2 = pointer to uboot provided cmdline or external DTB in mem + ; These are handled later in setup_arch() + st r0, [@uboot_tag] + st r2, [@uboot_arg] ; Identify if running on ISS vs Silicon ; IDENTITY Reg [ 3 2 1 0 ] @@ -101,7 +123,7 @@ stext: first_lines_of_secondary: - sr @_int_vec_base_lds, [AUX_INTR_VEC_BASE] + CPU_EARLY_SETUP ; setup per-cpu idle task as "current" on this CPU ld r0, [@secondary_idle_tsk] diff --git a/arch/arc/kernel/irq.c b/arch/arc/kernel/irq.c index a4b141ee9a6..7d653c0d077 100644 --- a/arch/arc/kernel/irq.c +++ b/arch/arc/kernel/irq.c @@ -150,24 +150,6 @@ void arch_do_IRQ(unsigned int irq, struct pt_regs *regs) set_irq_regs(old_regs); } -int get_hw_config_num_irq(void) -{ - uint32_t val = read_aux_reg(ARC_REG_VECBASE_BCR); - - switch (val & 0x03) { - case 0: - return 16; - case 1: - return 32; - case 2: - return 8; - default: - return 0; - } - - return 0; -} - /* * arch_local_irq_enable - Enable interrupts. * diff --git a/arch/arc/kernel/kprobes.c b/arch/arc/kernel/kprobes.c index eb1c2ee5eaf..42b05046fad 100644 --- a/arch/arc/kernel/kprobes.c +++ b/arch/arc/kernel/kprobes.c @@ -327,7 +327,7 @@ int __kprobes kprobe_fault_handler(struct pt_regs *regs, unsigned long trapnr) */ /* We increment the nmissed count for accounting, - * we can also use npre/npostfault count for accouting + * we can also use npre/npostfault count for accounting * these specific fault cases. */ kprobes_inc_nmissed_count(cur); diff --git a/arch/arc/kernel/perf_event.c b/arch/arc/kernel/perf_event.c new file mode 100644 index 00000000000..63177e4cb66 --- /dev/null +++ b/arch/arc/kernel/perf_event.c @@ -0,0 +1,326 @@ +/* + * Linux performance counter support for ARC700 series + * + * Copyright (C) 2013 Synopsys, Inc. (www.synopsys.com) + * + * This code is inspired by the perf support of various other architectures. + * + * 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. + * + */ +#include <linux/errno.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/perf_event.h> +#include <linux/platform_device.h> +#include <asm/arcregs.h> + +struct arc_pmu { + struct pmu pmu; + int counter_size; /* in bits */ + int n_counters; + unsigned long used_mask[BITS_TO_LONGS(ARC_PMU_MAX_HWEVENTS)]; + int ev_hw_idx[PERF_COUNT_ARC_HW_MAX]; +}; + +/* read counter #idx; note that counter# != event# on ARC! */ +static uint64_t arc_pmu_read_counter(int idx) +{ + uint32_t tmp; + uint64_t result; + + /* + * ARC supports making 'snapshots' of the counters, so we don't + * need to care about counters wrapping to 0 underneath our feet + */ + write_aux_reg(ARC_REG_PCT_INDEX, idx); + tmp = read_aux_reg(ARC_REG_PCT_CONTROL); + write_aux_reg(ARC_REG_PCT_CONTROL, tmp | ARC_REG_PCT_CONTROL_SN); + result = (uint64_t) (read_aux_reg(ARC_REG_PCT_SNAPH)) << 32; + result |= read_aux_reg(ARC_REG_PCT_SNAPL); + + return result; +} + +static void arc_perf_event_update(struct perf_event *event, + struct hw_perf_event *hwc, int idx) +{ + struct arc_pmu *arc_pmu = container_of(event->pmu, struct arc_pmu, pmu); + uint64_t prev_raw_count, new_raw_count; + int64_t delta; + + do { + prev_raw_count = local64_read(&hwc->prev_count); + new_raw_count = arc_pmu_read_counter(idx); + } while (local64_cmpxchg(&hwc->prev_count, prev_raw_count, + new_raw_count) != prev_raw_count); + + delta = (new_raw_count - prev_raw_count) & + ((1ULL << arc_pmu->counter_size) - 1ULL); + + local64_add(delta, &event->count); + local64_sub(delta, &hwc->period_left); +} + +static void arc_pmu_read(struct perf_event *event) +{ + arc_perf_event_update(event, &event->hw, event->hw.idx); +} + +static int arc_pmu_cache_event(u64 config) +{ + unsigned int cache_type, cache_op, cache_result; + int ret; + + cache_type = (config >> 0) & 0xff; + cache_op = (config >> 8) & 0xff; + cache_result = (config >> 16) & 0xff; + if (cache_type >= PERF_COUNT_HW_CACHE_MAX) + return -EINVAL; + if (cache_op >= PERF_COUNT_HW_CACHE_OP_MAX) + return -EINVAL; + if (cache_result >= PERF_COUNT_HW_CACHE_RESULT_MAX) + return -EINVAL; + + ret = arc_pmu_cache_map[cache_type][cache_op][cache_result]; + + if (ret == CACHE_OP_UNSUPPORTED) + return -ENOENT; + + return ret; +} + +/* initializes hw_perf_event structure if event is supported */ +static int arc_pmu_event_init(struct perf_event *event) +{ + struct arc_pmu *arc_pmu = container_of(event->pmu, struct arc_pmu, pmu); + struct hw_perf_event *hwc = &event->hw; + int ret; + + /* ARC 700 PMU does not support sampling events */ + if (is_sampling_event(event)) + return -ENOENT; + + switch (event->attr.type) { + case PERF_TYPE_HARDWARE: + if (event->attr.config >= PERF_COUNT_HW_MAX) + return -ENOENT; + if (arc_pmu->ev_hw_idx[event->attr.config] < 0) + return -ENOENT; + hwc->config = arc_pmu->ev_hw_idx[event->attr.config]; + pr_debug("initializing event %d with cfg %d\n", + (int) event->attr.config, (int) hwc->config); + return 0; + case PERF_TYPE_HW_CACHE: + ret = arc_pmu_cache_event(event->attr.config); + if (ret < 0) + return ret; + hwc->config = arc_pmu->ev_hw_idx[ret]; + return 0; + default: + return -ENOENT; + } +} + +/* starts all counters */ +static void arc_pmu_enable(struct pmu *pmu) +{ + uint32_t tmp; + tmp = read_aux_reg(ARC_REG_PCT_CONTROL); + write_aux_reg(ARC_REG_PCT_CONTROL, (tmp & 0xffff0000) | 0x1); +} + +/* stops all counters */ +static void arc_pmu_disable(struct pmu *pmu) +{ + uint32_t tmp; + tmp = read_aux_reg(ARC_REG_PCT_CONTROL); + write_aux_reg(ARC_REG_PCT_CONTROL, (tmp & 0xffff0000) | 0x0); +} + +/* + * Assigns hardware counter to hardware condition. + * Note that there is no separate start/stop mechanism; + * stopping is achieved by assigning the 'never' condition + */ +static void arc_pmu_start(struct perf_event *event, int flags) +{ + struct hw_perf_event *hwc = &event->hw; + int idx = hwc->idx; + + if (WARN_ON_ONCE(idx == -1)) + return; + + if (flags & PERF_EF_RELOAD) + WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE)); + + event->hw.state = 0; + + /* enable ARC pmu here */ + write_aux_reg(ARC_REG_PCT_INDEX, idx); + write_aux_reg(ARC_REG_PCT_CONFIG, hwc->config); +} + +static void arc_pmu_stop(struct perf_event *event, int flags) +{ + struct hw_perf_event *hwc = &event->hw; + int idx = hwc->idx; + + if (!(event->hw.state & PERF_HES_STOPPED)) { + /* stop ARC pmu here */ + write_aux_reg(ARC_REG_PCT_INDEX, idx); + + /* condition code #0 is always "never" */ + write_aux_reg(ARC_REG_PCT_CONFIG, 0); + + event->hw.state |= PERF_HES_STOPPED; + } + + if ((flags & PERF_EF_UPDATE) && + !(event->hw.state & PERF_HES_UPTODATE)) { + arc_perf_event_update(event, &event->hw, idx); + event->hw.state |= PERF_HES_UPTODATE; + } +} + +static void arc_pmu_del(struct perf_event *event, int flags) +{ + struct arc_pmu *arc_pmu = container_of(event->pmu, struct arc_pmu, pmu); + + arc_pmu_stop(event, PERF_EF_UPDATE); + __clear_bit(event->hw.idx, arc_pmu->used_mask); + + perf_event_update_userpage(event); +} + +/* allocate hardware counter and optionally start counting */ +static int arc_pmu_add(struct perf_event *event, int flags) +{ + struct arc_pmu *arc_pmu = container_of(event->pmu, struct arc_pmu, pmu); + struct hw_perf_event *hwc = &event->hw; + int idx = hwc->idx; + + if (__test_and_set_bit(idx, arc_pmu->used_mask)) { + idx = find_first_zero_bit(arc_pmu->used_mask, + arc_pmu->n_counters); + if (idx == arc_pmu->n_counters) + return -EAGAIN; + + __set_bit(idx, arc_pmu->used_mask); + hwc->idx = idx; + } + + write_aux_reg(ARC_REG_PCT_INDEX, idx); + write_aux_reg(ARC_REG_PCT_CONFIG, 0); + write_aux_reg(ARC_REG_PCT_COUNTL, 0); + write_aux_reg(ARC_REG_PCT_COUNTH, 0); + local64_set(&hwc->prev_count, 0); + + hwc->state = PERF_HES_UPTODATE | PERF_HES_STOPPED; + if (flags & PERF_EF_START) + arc_pmu_start(event, PERF_EF_RELOAD); + + perf_event_update_userpage(event); + + return 0; +} + +static int arc_pmu_device_probe(struct platform_device *pdev) +{ + struct arc_pmu *arc_pmu; + struct arc_reg_pct_build pct_bcr; + struct arc_reg_cc_build cc_bcr; + int i, j, ret; + + union cc_name { + struct { + uint32_t word0, word1; + char sentinel; + } indiv; + char str[9]; + } cc_name; + + + READ_BCR(ARC_REG_PCT_BUILD, pct_bcr); + if (!pct_bcr.v) { + pr_err("This core does not have performance counters!\n"); + return -ENODEV; + } + + arc_pmu = devm_kzalloc(&pdev->dev, sizeof(struct arc_pmu), + GFP_KERNEL); + if (!arc_pmu) + return -ENOMEM; + + arc_pmu->n_counters = pct_bcr.c; + BUG_ON(arc_pmu->n_counters > ARC_PMU_MAX_HWEVENTS); + + arc_pmu->counter_size = 32 + (pct_bcr.s << 4); + pr_info("ARC PMU found with %d counters of size %d bits\n", + arc_pmu->n_counters, arc_pmu->counter_size); + + READ_BCR(ARC_REG_CC_BUILD, cc_bcr); + + if (!cc_bcr.v) + pr_err("Strange! Performance counters exist, but no countable conditions?\n"); + + pr_info("ARC PMU has %d countable conditions\n", cc_bcr.c); + + cc_name.str[8] = 0; + for (i = 0; i < PERF_COUNT_HW_MAX; i++) + arc_pmu->ev_hw_idx[i] = -1; + + for (j = 0; j < cc_bcr.c; j++) { + write_aux_reg(ARC_REG_CC_INDEX, j); + cc_name.indiv.word0 = read_aux_reg(ARC_REG_CC_NAME0); + cc_name.indiv.word1 = read_aux_reg(ARC_REG_CC_NAME1); + for (i = 0; i < ARRAY_SIZE(arc_pmu_ev_hw_map); i++) { + if (arc_pmu_ev_hw_map[i] && + !strcmp(arc_pmu_ev_hw_map[i], cc_name.str) && + strlen(arc_pmu_ev_hw_map[i])) { + pr_debug("mapping %d to idx %d with name %s\n", + i, j, cc_name.str); + arc_pmu->ev_hw_idx[i] = j; + } + } + } + + arc_pmu->pmu = (struct pmu) { + .pmu_enable = arc_pmu_enable, + .pmu_disable = arc_pmu_disable, + .event_init = arc_pmu_event_init, + .add = arc_pmu_add, + .del = arc_pmu_del, + .start = arc_pmu_start, + .stop = arc_pmu_stop, + .read = arc_pmu_read, + }; + + ret = perf_pmu_register(&arc_pmu->pmu, pdev->name, PERF_TYPE_RAW); + + return ret; +} + +#ifdef CONFIG_OF +static const struct of_device_id arc_pmu_match[] = { + { .compatible = "snps,arc700-pmu" }, + {}, +}; +MODULE_DEVICE_TABLE(of, arc_pmu_match); +#endif + +static struct platform_driver arc_pmu_driver = { + .driver = { + .name = "arc700-pmu", + .of_match_table = of_match_ptr(arc_pmu_match), + }, + .probe = arc_pmu_device_probe, +}; + +module_platform_driver(arc_pmu_driver); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Mischa Jonker <mjonker@synopsys.com>"); +MODULE_DESCRIPTION("ARC PMU driver"); diff --git a/arch/arc/kernel/process.c b/arch/arc/kernel/process.c index 07a3a968fe4..fdd89715d2d 100644 --- a/arch/arc/kernel/process.c +++ b/arch/arc/kernel/process.c @@ -151,6 +151,29 @@ int copy_thread(unsigned long clone_flags, } /* + * Do necessary setup to start up a new user task + */ +void start_thread(struct pt_regs * regs, unsigned long pc, unsigned long usp) +{ + set_fs(USER_DS); /* user space */ + + regs->sp = usp; + regs->ret = pc; + + /* + * [U]ser Mode bit set + * [L] ZOL loop inhibited to begin with - cleared by a LP insn + * Interrupts enabled + */ + regs->status32 = STATUS_U_MASK | STATUS_L_MASK | + STATUS_E1_MASK | STATUS_E2_MASK; + + /* bogus seed values for debugging */ + regs->lp_start = 0x10; + regs->lp_end = 0x80; +} + +/* * Some archs flush debug and FPU info here */ void flush_thread(void) diff --git a/arch/arc/kernel/ptrace.c b/arch/arc/kernel/ptrace.c index 5d76706139d..13b3ffb27a3 100644 --- a/arch/arc/kernel/ptrace.c +++ b/arch/arc/kernel/ptrace.c @@ -146,6 +146,10 @@ long arch_ptrace(struct task_struct *child, long request, pr_debug("REQ=%ld: ADDR =0x%lx, DATA=0x%lx)\n", request, addr, data); switch (request) { + case PTRACE_GET_THREAD_AREA: + ret = put_user(task_thread_info(child)->thr_ptr, + (unsigned long __user *)data); + break; default: ret = ptrace_request(child, request, addr, data); break; diff --git a/arch/arc/kernel/setup.c b/arch/arc/kernel/setup.c index 643eae4436e..119dddb752b 100644 --- a/arch/arc/kernel/setup.c +++ b/arch/arc/kernel/setup.c @@ -29,7 +29,10 @@ int running_on_hw = 1; /* vs. on ISS */ -char __initdata command_line[COMMAND_LINE_SIZE]; +/* Part of U-boot ABI: see head.S */ +int __initdata uboot_tag; +char __initdata *uboot_arg; + const struct machine_desc *machine_desc; struct task_struct *_current_task[NR_CPUS]; /* For stack switching */ @@ -311,19 +314,40 @@ void setup_processor(void) arc_chk_fpu(); } +static inline int is_kernel(unsigned long addr) +{ + if (addr >= (unsigned long)_stext && addr <= (unsigned long)_end) + return 1; + return 0; +} + void __init setup_arch(char **cmdline_p) { - /* This also populates @boot_command_line from /bootargs */ - machine_desc = setup_machine_fdt(__dtb_start); - if (!machine_desc) - panic("Embedded DT invalid\n"); - - /* Append any u-boot provided cmdline */ -#ifdef CONFIG_CMDLINE_UBOOT - /* Add a whitespace seperator between the 2 cmdlines */ - strlcat(boot_command_line, " ", COMMAND_LINE_SIZE); - strlcat(boot_command_line, command_line, COMMAND_LINE_SIZE); -#endif + /* make sure that uboot passed pointer to cmdline/dtb is valid */ + if (uboot_tag && is_kernel((unsigned long)uboot_arg)) + panic("Invalid uboot arg\n"); + + /* See if u-boot passed an external Device Tree blob */ + machine_desc = setup_machine_fdt(uboot_arg); /* uboot_tag == 2 */ + if (!machine_desc) { + /* No, so try the embedded one */ + machine_desc = setup_machine_fdt(__dtb_start); + if (!machine_desc) + panic("Embedded DT invalid\n"); + + /* + * If we are here, it is established that @uboot_arg didn't + * point to DT blob. Instead if u-boot says it is cmdline, + * Appent to embedded DT cmdline. + * setup_machine_fdt() would have populated @boot_command_line + */ + if (uboot_tag == 1) { + /* Ensure a whitespace between the 2 cmdlines */ + strlcat(boot_command_line, " ", COMMAND_LINE_SIZE); + strlcat(boot_command_line, uboot_arg, + COMMAND_LINE_SIZE); + } + } /* Save unparsed command line copy for /proc/cmdline */ *cmdline_p = boot_command_line; diff --git a/arch/arc/kernel/smp.c b/arch/arc/kernel/smp.c index c2f9ebbc38f..c802bb50060 100644 --- a/arch/arc/kernel/smp.c +++ b/arch/arc/kernel/smp.c @@ -138,7 +138,7 @@ void start_kernel_secondary(void) if (machine_desc->init_smp) machine_desc->init_smp(smp_processor_id()); - arc_local_timer_setup(cpu); + arc_local_timer_setup(); local_irq_enable(); preempt_disable(); @@ -197,51 +197,65 @@ int __init setup_profiling_timer(unsigned int multiplier) /* Inter Processor Interrupt Handling */ /*****************************************************************************/ -/* - * structures for inter-processor calls - * A Collection of single bit ipi messages - * - */ - -/* - * TODO_rajesh investigate tlb message types. - * IPI Timer not needed because each ARC has an individual Interrupting Timer - */ enum ipi_msg_type { - IPI_NOP = 0, + IPI_EMPTY = 0, IPI_RESCHEDULE = 1, IPI_CALL_FUNC, - IPI_CPU_STOP + IPI_CPU_STOP, }; -struct ipi_data { - unsigned long bits; -}; +/* + * In arches with IRQ for each msg type (above), receiver can use IRQ-id to + * figure out what msg was sent. For those which don't (ARC has dedicated IPI + * IRQ), the msg-type needs to be conveyed via per-cpu data + */ -static DEFINE_PER_CPU(struct ipi_data, ipi_data); +static DEFINE_PER_CPU(unsigned long, ipi_data); -static void ipi_send_msg(const struct cpumask *callmap, enum ipi_msg_type msg) +static void ipi_send_msg_one(int cpu, enum ipi_msg_type msg) { + unsigned long __percpu *ipi_data_ptr = per_cpu_ptr(&ipi_data, cpu); + unsigned long old, new; unsigned long flags; - unsigned int cpu; + + pr_debug("%d Sending msg [%d] to %d\n", smp_processor_id(), msg, cpu); local_irq_save(flags); - for_each_cpu(cpu, callmap) { - struct ipi_data *ipi = &per_cpu(ipi_data, cpu); - set_bit(msg, &ipi->bits); - } + /* + * Atomically write new msg bit (in case others are writing too), + * and read back old value + */ + do { + new = old = *ipi_data_ptr; + new |= 1U << msg; + } while (cmpxchg(ipi_data_ptr, old, new) != old); - /* Call the platform specific cross-CPU call function */ - if (plat_smp_ops.ipi_send) - plat_smp_ops.ipi_send((void *)callmap); + /* + * Call the platform specific IPI kick function, but avoid if possible: + * Only do so if there's no pending msg from other concurrent sender(s). + * Otherwise, recevier will see this msg as well when it takes the + * IPI corresponding to that msg. This is true, even if it is already in + * IPI handler, because !@old means it has not yet dequeued the msg(s) + * so @new msg can be a free-loader + */ + if (plat_smp_ops.ipi_send && !old) + plat_smp_ops.ipi_send(cpu); local_irq_restore(flags); } +static void ipi_send_msg(const struct cpumask *callmap, enum ipi_msg_type msg) +{ + unsigned int cpu; + + for_each_cpu(cpu, callmap) + ipi_send_msg_one(cpu, msg); +} + void smp_send_reschedule(int cpu) { - ipi_send_msg(cpumask_of(cpu), IPI_RESCHEDULE); + ipi_send_msg_one(cpu, IPI_RESCHEDULE); } void smp_send_stop(void) @@ -254,7 +268,7 @@ void smp_send_stop(void) void arch_send_call_function_single_ipi(int cpu) { - ipi_send_msg(cpumask_of(cpu), IPI_CALL_FUNC); + ipi_send_msg_one(cpu, IPI_CALL_FUNC); } void arch_send_call_function_ipi_mask(const struct cpumask *mask) @@ -265,33 +279,29 @@ void arch_send_call_function_ipi_mask(const struct cpumask *mask) /* * ipi_cpu_stop - handle IPI from smp_send_stop() */ -static void ipi_cpu_stop(unsigned int cpu) +static void ipi_cpu_stop(void) { machine_halt(); } -static inline void __do_IPI(unsigned long *ops, struct ipi_data *ipi, int cpu) +static inline void __do_IPI(unsigned long msg) { - unsigned long msg = 0; - - do { - msg = find_next_bit(ops, BITS_PER_LONG, msg+1); + switch (msg) { + case IPI_RESCHEDULE: + scheduler_ipi(); + break; - switch (msg) { - case IPI_RESCHEDULE: - scheduler_ipi(); - break; + case IPI_CALL_FUNC: + generic_smp_call_function_interrupt(); + break; - case IPI_CALL_FUNC: - generic_smp_call_function_interrupt(); - break; - - case IPI_CPU_STOP: - ipi_cpu_stop(cpu); - break; - } - } while (msg < BITS_PER_LONG); + case IPI_CPU_STOP: + ipi_cpu_stop(); + break; + default: + pr_warn("IPI with unexpected msg %ld\n", msg); + } } /* @@ -300,19 +310,25 @@ static inline void __do_IPI(unsigned long *ops, struct ipi_data *ipi, int cpu) */ irqreturn_t do_IPI(int irq, void *dev_id) { - int cpu = smp_processor_id(); - struct ipi_data *ipi = &per_cpu(ipi_data, cpu); - unsigned long ops; + unsigned long pending; + + pr_debug("IPI [%ld] received on cpu %d\n", + *this_cpu_ptr(&ipi_data), smp_processor_id()); if (plat_smp_ops.ipi_clear) - plat_smp_ops.ipi_clear(cpu, irq); + plat_smp_ops.ipi_clear(irq); /* - * XXX: is this loop really needed - * And do we need to move ipi_clean inside + * "dequeue" the msg corresponding to this IPI (and possibly other + * piggybacked msg from elided IPIs: see ipi_send_msg_one() above) */ - while ((ops = xchg(&ipi->bits, 0)) != 0) - __do_IPI(&ops, ipi, cpu); + pending = xchg(this_cpu_ptr(&ipi_data), 0); + + do { + unsigned long msg = __ffs(pending); + __do_IPI(msg); + pending &= ~(1U << msg); + } while (pending); return IRQ_HANDLED; } @@ -321,8 +337,19 @@ irqreturn_t do_IPI(int irq, void *dev_id) * API called by platform code to hookup arch-common ISR to their IPI IRQ */ static DEFINE_PER_CPU(int, ipi_dev); + +static struct irqaction arc_ipi_irq = { + .name = "IPI Interrupt", + .flags = IRQF_PERCPU, + .handler = do_IPI, +}; + int smp_ipi_irq_setup(int cpu, int irq) { - int *dev_id = &per_cpu(ipi_dev, smp_processor_id()); - return request_percpu_irq(irq, do_IPI, "IPI Interrupt", dev_id); + if (!cpu) + return setup_irq(irq, &arc_ipi_irq); + else + arch_unmask_irq(irq); + + return 0; } diff --git a/arch/arc/kernel/time.c b/arch/arc/kernel/time.c index e5f3a837fb3..36c2aa99436 100644 --- a/arch/arc/kernel/time.c +++ b/arch/arc/kernel/time.c @@ -155,22 +155,6 @@ static void arc_timer_event_setup(unsigned int limit) write_aux_reg(ARC_REG_TIMER0_CTRL, TIMER_CTRL_IE | TIMER_CTRL_NH); } -/* - * Acknowledge the interrupt (oneshot) and optionally re-arm it (periodic) - * -Any write to CTRL Reg will ack the intr (NH bit: Count when not halted) - * -Rearming is done by setting the IE bit - * - * Small optimisation: Normal code would have been - * if (irq_reenable) - * CTRL_REG = (IE | NH); - * else - * CTRL_REG = NH; - * However since IE is BIT0 we can fold the branch - */ -static void arc_timer_event_ack(unsigned int irq_reenable) -{ - write_aux_reg(ARC_REG_TIMER0_CTRL, irq_reenable | TIMER_CTRL_NH); -} static int arc_clkevent_set_next_event(unsigned long delta, struct clock_event_device *dev) @@ -207,10 +191,22 @@ static DEFINE_PER_CPU(struct clock_event_device, arc_clockevent_device) = { static irqreturn_t timer_irq_handler(int irq, void *dev_id) { - struct clock_event_device *clk = this_cpu_ptr(&arc_clockevent_device); + /* + * Note that generic IRQ core could have passed @evt for @dev_id if + * irq_set_chip_and_handler() asked for handle_percpu_devid_irq() + */ + struct clock_event_device *evt = this_cpu_ptr(&arc_clockevent_device); + int irq_reenable = evt->mode == CLOCK_EVT_MODE_PERIODIC; + + /* + * Any write to CTRL reg ACks the interrupt, we rewrite the + * Count when [N]ot [H]alted bit. + * And re-arm it if perioid by [I]nterrupt [E]nable bit + */ + write_aux_reg(ARC_REG_TIMER0_CTRL, irq_reenable | TIMER_CTRL_NH); + + evt->event_handler(evt); - arc_timer_event_ack(clk->mode == CLOCK_EVT_MODE_PERIODIC); - clk->event_handler(clk); return IRQ_HANDLED; } @@ -222,14 +218,14 @@ static struct irqaction arc_timer_irq = { /* * Setup the local event timer for @cpu - * N.B. weak so that some exotic ARC SoCs can completely override it */ -void __weak arc_local_timer_setup(unsigned int cpu) +void arc_local_timer_setup() { - struct clock_event_device *clk = &per_cpu(arc_clockevent_device, cpu); + struct clock_event_device *evt = this_cpu_ptr(&arc_clockevent_device); + int cpu = smp_processor_id(); - clk->cpumask = cpumask_of(cpu); - clockevents_config_and_register(clk, arc_get_core_freq(), + evt->cpumask = cpumask_of(cpu); + clockevents_config_and_register(evt, arc_get_core_freq(), 0, ARC_TIMER_MAX); /* @@ -266,7 +262,7 @@ void __init time_init(void) clocksource_register_hz(&arc_counter, arc_get_core_freq()); /* sets up the periodic event timer */ - arc_local_timer_setup(smp_processor_id()); + arc_local_timer_setup(); if (machine_desc->init_time) machine_desc->init_time(); diff --git a/arch/arc/kernel/troubleshoot.c b/arch/arc/kernel/troubleshoot.c index 73a7450ee62..1badf9b84b5 100644 --- a/arch/arc/kernel/troubleshoot.c +++ b/arch/arc/kernel/troubleshoot.c @@ -86,12 +86,13 @@ static void show_faulting_vma(unsigned long address, char *buf) unsigned long ino = 0; dev_t dev = 0; char *nm = buf; + struct mm_struct *active_mm = current->active_mm; /* can't use print_vma_addr() yet as it doesn't check for * non-inclusive vma */ - - vma = find_vma(current->active_mm, address); + down_read(&active_mm->mmap_sem); + vma = find_vma(active_mm, address); /* check against the find_vma( ) behaviour which returns the next VMA * if the container VMA is not found @@ -110,9 +111,10 @@ static void show_faulting_vma(unsigned long address, char *buf) vma->vm_start < TASK_UNMAPPED_BASE ? address : address - vma->vm_start, nm, vma->vm_start, vma->vm_end); - } else { + } else pr_info(" @No matching VMA found\n"); - } + + up_read(&active_mm->mmap_sem); } static void show_ecr_verbose(struct pt_regs *regs) diff --git a/arch/arc/kernel/vmlinux.lds.S b/arch/arc/kernel/vmlinux.lds.S index 2555f5886af..dd35bde39f6 100644 --- a/arch/arc/kernel/vmlinux.lds.S +++ b/arch/arc/kernel/vmlinux.lds.S @@ -116,7 +116,7 @@ SECTIONS _edata = .; - BSS_SECTION(0, 0, 0) + BSS_SECTION(4, 4, 4) #ifdef CONFIG_ARC_DW2_UNWIND . = ALIGN(PAGE_SIZE); |
