diff options
Diffstat (limited to 'drivers/char/ipmi/ipmi_si_intf.c')
| -rw-r--r-- | drivers/char/ipmi/ipmi_si_intf.c | 3556 |
1 files changed, 2452 insertions, 1104 deletions
diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c index a44b97304e9..5d665680ae3 100644 --- a/drivers/char/ipmi/ipmi_si_intf.c +++ b/drivers/char/ipmi/ipmi_si_intf.c @@ -9,6 +9,7 @@ * source@mvista.com * * Copyright 2002 MontaVista Software Inc. + * Copyright 2006 IBM Corp., 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 @@ -38,11 +39,10 @@ * and drives the real SMI state machine. */ -#include <linux/config.h> #include <linux/module.h> #include <linux/moduleparam.h> -#include <asm/system.h> #include <linux/sched.h> +#include <linux/seq_file.h> #include <linux/timer.h> #include <linux/errno.h> #include <linux/spinlock.h> @@ -51,32 +51,31 @@ #include <linux/list.h> #include <linux/pci.h> #include <linux/ioport.h> +#include <linux/notifier.h> +#include <linux/mutex.h> +#include <linux/kthread.h> #include <asm/irq.h> -#ifdef CONFIG_HIGH_RES_TIMERS -#include <linux/hrtime.h> -# if defined(schedule_next_int) -/* Old high-res timer code, do translations. */ -# define get_arch_cycles(a) quick_update_jiffies_sub(a) -# define arch_cycles_per_jiffy cycles_per_jiffies -# endif -static inline void add_usec_to_timer(struct timer_list *t, long v) -{ - t->sub_expires += nsec_to_arch_cycle(v * 1000); - while (t->sub_expires >= arch_cycles_per_jiffy) - { - t->expires++; - t->sub_expires -= arch_cycles_per_jiffy; - } -} -#endif #include <linux/interrupt.h> #include <linux/rcupdate.h> +#include <linux/ipmi.h> #include <linux/ipmi_smi.h> #include <asm/io.h> #include "ipmi_si_sm.h" -#include <linux/init.h> +#include <linux/dmi.h> +#include <linux/string.h> +#include <linux/ctype.h> +#include <linux/pnp.h> +#include <linux/of_device.h> +#include <linux/of_platform.h> +#include <linux/of_address.h> +#include <linux/of_irq.h> + +#ifdef CONFIG_PARISC +#include <asm/hardware.h> /* for register_parisc_driver() stuff */ +#include <asm/parisc-device.h> +#endif -#define IPMI_SI_VERSION "v33" +#define PFX "ipmi_si: " /* Measure times between events in the driver. */ #undef DEBUG_TIMING @@ -86,7 +85,7 @@ static inline void add_usec_to_timer(struct timer_list *t, long v) #define SI_USEC_PER_JIFFY (1000000/HZ) #define SI_TIMEOUT_JIFFIES (SI_TIMEOUT_TIME_USEC/SI_USEC_PER_JIFFY) #define SI_SHORT_TIMEOUT_USEC 250 /* .25ms when the SM request a - short timeout */ + short timeout */ enum si_intf_state { SI_NORMAL, @@ -96,7 +95,9 @@ enum si_intf_state { SI_CLEARING_FLAGS_THEN_SET_IRQ, SI_GETTING_MESSAGES, SI_ENABLE_INTERRUPTS1, - SI_ENABLE_INTERRUPTS2 + SI_ENABLE_INTERRUPTS2, + SI_DISABLE_INTERRUPTS1, + SI_DISABLE_INTERRUPTS2 /* FIXME - add watchdog stuff. */ }; @@ -108,52 +109,137 @@ enum si_intf_state { enum si_type { SI_KCS, SI_SMIC, SI_BT }; +static char *si_to_str[] = { "kcs", "smic", "bt" }; -struct smi_info -{ +static char *ipmi_addr_src_to_str[] = { NULL, "hotmod", "hardcoded", "SPMI", + "ACPI", "SMBIOS", "PCI", + "device-tree", "default" }; + +#define DEVICE_NAME "ipmi_si" + +static struct platform_driver ipmi_driver; + +/* + * Indexes into stats[] in smi_info below. + */ +enum si_stat_indexes { + /* + * Number of times the driver requested a timer while an operation + * was in progress. + */ + SI_STAT_short_timeouts = 0, + + /* + * Number of times the driver requested a timer while nothing was in + * progress. + */ + SI_STAT_long_timeouts, + + /* Number of times the interface was idle while being polled. */ + SI_STAT_idles, + + /* Number of interrupts the driver handled. */ + SI_STAT_interrupts, + + /* Number of time the driver got an ATTN from the hardware. */ + SI_STAT_attentions, + + /* Number of times the driver requested flags from the hardware. */ + SI_STAT_flag_fetches, + + /* Number of times the hardware didn't follow the state machine. */ + SI_STAT_hosed_count, + + /* Number of completed messages. */ + SI_STAT_complete_transactions, + + /* Number of IPMI events received from the hardware. */ + SI_STAT_events, + + /* Number of watchdog pretimeouts. */ + SI_STAT_watchdog_pretimeouts, + + /* Number of asynchronous messages received. */ + SI_STAT_incoming_messages, + + + /* This *must* remain last, add new values above this. */ + SI_NUM_STATS +}; + +struct smi_info { + int intf_num; ipmi_smi_t intf; struct si_sm_data *si_sm; struct si_sm_handlers *handlers; enum si_type si_type; spinlock_t si_lock; - spinlock_t msg_lock; struct list_head xmit_msgs; struct list_head hp_xmit_msgs; struct ipmi_smi_msg *curr_msg; enum si_intf_state si_state; - /* Used to handle the various types of I/O that can occur with - IPMI */ + /* + * Used to handle the various types of I/O that can occur with + * IPMI + */ struct si_sm_io io; int (*io_setup)(struct smi_info *info); void (*io_cleanup)(struct smi_info *info); int (*irq_setup)(struct smi_info *info); void (*irq_cleanup)(struct smi_info *info); unsigned int io_size; + enum ipmi_addr_src addr_source; /* ACPI, PCI, SMBIOS, hardcode, etc. */ + void (*addr_source_cleanup)(struct smi_info *info); + void *addr_source_data; - /* Flags from the last GET_MSG_FLAGS command, used when an ATTN - is set to hold the flags until we are done handling everything - from the flags. */ + /* + * Per-OEM handler, called from handle_flags(). Returns 1 + * when handle_flags() needs to be re-run or 0 indicating it + * set si_state itself. + */ + int (*oem_data_avail_handler)(struct smi_info *smi_info); + + /* + * Flags from the last GET_MSG_FLAGS command, used when an ATTN + * is set to hold the flags until we are done handling everything + * from the flags. + */ #define RECEIVE_MSG_AVAIL 0x01 #define EVENT_MSG_BUFFER_FULL 0x02 #define WDT_PRE_TIMEOUT_INT 0x08 +#define OEM0_DATA_AVAIL 0x20 +#define OEM1_DATA_AVAIL 0x40 +#define OEM2_DATA_AVAIL 0x80 +#define OEM_DATA_AVAIL (OEM0_DATA_AVAIL | \ + OEM1_DATA_AVAIL | \ + OEM2_DATA_AVAIL) unsigned char msg_flags; - /* If set to true, this will request events the next time the - state machine is idle. */ + /* Does the BMC have an event buffer? */ + bool has_event_buffer; + + /* + * If set to true, this will request events the next time the + * state machine is idle. + */ atomic_t req_events; - /* If true, run the state machine to completion on every send - call. Generally used after a panic to make sure stuff goes - out. */ - int run_to_completion; + /* + * If true, run the state machine to completion on every send + * call. Generally used after a panic to make sure stuff goes + * out. + */ + bool run_to_completion; /* The I/O port of an SI interface. */ int port; - /* The space between start addresses of the two ports. For - instance, if the first port is 0xca2 and the spacing is 4, then - the second port is 0xca6. */ + /* + * The space between start addresses of the two ports. For + * instance, if the first port is 0xca2 and the spacing is 4, then + * the second port is 0xca6. + */ unsigned int spacing; /* zero if no irq; */ @@ -162,64 +248,105 @@ struct smi_info /* The timer for this si. */ struct timer_list si_timer; + /* This flag is set, if the timer is running (timer_pending() isn't enough) */ + bool timer_running; + /* The time (in jiffies) the last timeout occurred at. */ unsigned long last_timeout_jiffies; /* Used to gracefully stop the timer without race conditions. */ - volatile int stop_operation; - volatile int timer_stopped; + atomic_t stop_operation; + + /* Are we waiting for the events, pretimeouts, received msgs? */ + atomic_t need_watch; + + /* + * The driver will disable interrupts when it gets into a + * situation where it cannot handle messages due to lack of + * memory. Once that situation clears up, it will re-enable + * interrupts. + */ + bool interrupt_disabled; + + /* From the get device id response... */ + struct ipmi_device_id device_id; - /* The driver will disable interrupts when it gets into a - situation where it cannot handle messages due to lack of - memory. Once that situation clears up, it will re-enable - interrupts. */ - int interrupt_disabled; + /* Driver model stuff. */ + struct device *dev; + struct platform_device *pdev; - unsigned char ipmi_si_dev_rev; - unsigned char ipmi_si_fw_rev_major; - unsigned char ipmi_si_fw_rev_minor; - unsigned char ipmi_version_major; - unsigned char ipmi_version_minor; + /* + * True if we allocated the device, false if it came from + * someplace else (like PCI). + */ + bool dev_registered; /* Slave address, could be reported from DMI. */ unsigned char slave_addr; /* Counters and things for the proc filesystem. */ - spinlock_t count_lock; - unsigned long short_timeouts; - unsigned long long_timeouts; - unsigned long timeout_restarts; - unsigned long idles; - unsigned long interrupts; - unsigned long attentions; - unsigned long flag_fetches; - unsigned long hosed_count; - unsigned long complete_transactions; - unsigned long events; - unsigned long watchdog_pretimeouts; - unsigned long incoming_messages; + atomic_t stats[SI_NUM_STATS]; + + struct task_struct *thread; + + struct list_head link; + union ipmi_smi_info_union addr_info; }; -static void si_restart_short_timer(struct smi_info *smi_info); +#define smi_inc_stat(smi, stat) \ + atomic_inc(&(smi)->stats[SI_STAT_ ## stat]) +#define smi_get_stat(smi, stat) \ + ((unsigned int) atomic_read(&(smi)->stats[SI_STAT_ ## stat])) + +#define SI_MAX_PARMS 4 + +static int force_kipmid[SI_MAX_PARMS]; +static int num_force_kipmid; +#ifdef CONFIG_PCI +static bool pci_registered; +#endif +#ifdef CONFIG_ACPI +static bool pnp_registered; +#endif +#ifdef CONFIG_PARISC +static bool parisc_registered; +#endif + +static unsigned int kipmid_max_busy_us[SI_MAX_PARMS]; +static int num_max_busy_us; + +static bool unload_when_empty = true; + +static int add_smi(struct smi_info *smi); +static int try_smi_init(struct smi_info *smi); +static void cleanup_one_si(struct smi_info *to_clean); +static void cleanup_ipmi_si(void); + +static ATOMIC_NOTIFIER_HEAD(xaction_notifier_list); +static int register_xaction_notifier(struct notifier_block *nb) +{ + return atomic_notifier_chain_register(&xaction_notifier_list, nb); +} static void deliver_recv_msg(struct smi_info *smi_info, struct ipmi_smi_msg *msg) { - /* Deliver the message to the upper layer with the lock - released. */ - spin_unlock(&(smi_info->si_lock)); + /* Deliver the message to the upper layer. */ ipmi_smi_msg_received(smi_info->intf, msg); - spin_lock(&(smi_info->si_lock)); } -static void return_hosed_msg(struct smi_info *smi_info) +static void return_hosed_msg(struct smi_info *smi_info, int cCode) { struct ipmi_smi_msg *msg = smi_info->curr_msg; - /* Make it a reponse */ + if (cCode < 0 || cCode > IPMI_ERR_UNSPECIFIED) + cCode = IPMI_ERR_UNSPECIFIED; + /* else use it as is */ + + /* Make it a response */ msg->rsp[0] = msg->data[0] | 4; msg->rsp[1] = msg->data[1]; - msg->rsp[2] = 0xFF; /* Unknown error. */ + msg->rsp[2] = cCode; msg->rsp_size = 3; smi_info->curr_msg = NULL; @@ -234,14 +361,10 @@ static enum si_sm_result start_next_msg(struct smi_info *smi_info) struct timeval t; #endif - /* No need to save flags, we aleady have interrupts off and we - already hold the SMI lock. */ - spin_lock(&(smi_info->msg_lock)); - /* Pick the high priority queue first. */ - if (! list_empty(&(smi_info->hp_xmit_msgs))) { + if (!list_empty(&(smi_info->hp_xmit_msgs))) { entry = smi_info->hp_xmit_msgs.next; - } else if (! list_empty(&(smi_info->xmit_msgs))) { + } else if (!list_empty(&(smi_info->xmit_msgs))) { entry = smi_info->xmit_msgs.next; } @@ -257,20 +380,24 @@ static enum si_sm_result start_next_msg(struct smi_info *smi_info) link); #ifdef DEBUG_TIMING do_gettimeofday(&t); - printk("**Start2: %d.%9.9d\n", t.tv_sec, t.tv_usec); + printk(KERN_DEBUG "**Start2: %d.%9.9d\n", t.tv_sec, t.tv_usec); #endif + err = atomic_notifier_call_chain(&xaction_notifier_list, + 0, smi_info); + if (err & NOTIFY_STOP_MASK) { + rv = SI_SM_CALL_WITHOUT_DELAY; + goto out; + } err = smi_info->handlers->start_transaction( smi_info->si_sm, smi_info->curr_msg->data, smi_info->curr_msg->data_size); - if (err) { - return_hosed_msg(smi_info); - } + if (err) + return_hosed_msg(smi_info, err); rv = SI_SM_CALL_WITHOUT_DELAY; } - spin_unlock(&(smi_info->msg_lock)); - + out: return rv; } @@ -278,8 +405,10 @@ static void start_enable_irq(struct smi_info *smi_info) { unsigned char msg[2]; - /* If we are enabling interrupts, we have to tell the - BMC to use them. */ + /* + * If we are enabling interrupts, we have to tell the + * BMC to use them. + */ msg[0] = (IPMI_NETFN_APP_REQUEST << 2); msg[1] = IPMI_GET_BMC_GLOBAL_ENABLES_CMD; @@ -287,6 +416,17 @@ static void start_enable_irq(struct smi_info *smi_info) smi_info->si_state = SI_ENABLE_INTERRUPTS1; } +static void start_disable_irq(struct smi_info *smi_info) +{ + unsigned char msg[2]; + + msg[0] = (IPMI_NETFN_APP_REQUEST << 2); + msg[1] = IPMI_GET_BMC_GLOBAL_ENABLES_CMD; + + smi_info->handlers->start_transaction(smi_info->si_sm, msg, 2); + smi_info->si_state = SI_DISABLE_INTERRUPTS1; +} + static void start_clear_flags(struct smi_info *smi_info) { unsigned char msg[3]; @@ -300,39 +440,47 @@ static void start_clear_flags(struct smi_info *smi_info) smi_info->si_state = SI_CLEARING_FLAGS; } -/* When we have a situtaion where we run out of memory and cannot - allocate messages, we just leave them in the BMC and run the system - polled until we can allocate some memory. Once we have some - memory, we will re-enable the interrupt. */ +static void smi_mod_timer(struct smi_info *smi_info, unsigned long new_val) +{ + smi_info->last_timeout_jiffies = jiffies; + mod_timer(&smi_info->si_timer, new_val); + smi_info->timer_running = true; +} + +/* + * When we have a situtaion where we run out of memory and cannot + * allocate messages, we just leave them in the BMC and run the system + * polled until we can allocate some memory. Once we have some + * memory, we will re-enable the interrupt. + */ static inline void disable_si_irq(struct smi_info *smi_info) { if ((smi_info->irq) && (!smi_info->interrupt_disabled)) { - disable_irq_nosync(smi_info->irq); - smi_info->interrupt_disabled = 1; + start_disable_irq(smi_info); + smi_info->interrupt_disabled = true; + if (!atomic_read(&smi_info->stop_operation)) + smi_mod_timer(smi_info, jiffies + SI_TIMEOUT_JIFFIES); } } static inline void enable_si_irq(struct smi_info *smi_info) { if ((smi_info->irq) && (smi_info->interrupt_disabled)) { - enable_irq(smi_info->irq); - smi_info->interrupt_disabled = 0; + start_enable_irq(smi_info); + smi_info->interrupt_disabled = false; } } static void handle_flags(struct smi_info *smi_info) { + retry: if (smi_info->msg_flags & WDT_PRE_TIMEOUT_INT) { /* Watchdog pre-timeout */ - spin_lock(&smi_info->count_lock); - smi_info->watchdog_pretimeouts++; - spin_unlock(&smi_info->count_lock); + smi_inc_stat(smi_info, watchdog_pretimeouts); start_clear_flags(smi_info); smi_info->msg_flags &= ~WDT_PRE_TIMEOUT_INT; - spin_unlock(&(smi_info->si_lock)); ipmi_smi_watchdog_pretimeout(smi_info->intf); - spin_lock(&(smi_info->si_lock)); } else if (smi_info->msg_flags & RECEIVE_MSG_AVAIL) { /* Messages available. */ smi_info->curr_msg = ipmi_alloc_smi_msg(); @@ -371,9 +519,12 @@ static void handle_flags(struct smi_info *smi_info) smi_info->curr_msg->data, smi_info->curr_msg->data_size); smi_info->si_state = SI_GETTING_EVENTS; - } else { + } else if (smi_info->msg_flags & OEM_DATA_AVAIL && + smi_info->oem_data_avail_handler) { + if (smi_info->oem_data_avail_handler(smi_info)) + goto retry; + } else smi_info->si_state = SI_NORMAL; - } } static void handle_transaction_done(struct smi_info *smi_info) @@ -383,7 +534,7 @@ static void handle_transaction_done(struct smi_info *smi_info) struct timeval t; do_gettimeofday(&t); - printk("**Done: %d.%9.9d\n", t.tv_sec, t.tv_usec); + printk(KERN_DEBUG "**Done: %d.%9.9d\n", t.tv_sec, t.tv_usec); #endif switch (smi_info->si_state) { case SI_NORMAL: @@ -396,9 +547,11 @@ static void handle_transaction_done(struct smi_info *smi_info) smi_info->curr_msg->rsp, IPMI_MAX_MSG_LENGTH); - /* Do this here becase deliver_recv_msg() releases the - lock, and a new message can be put in during the - time the lock is released. */ + /* + * Do this here becase deliver_recv_msg() releases the + * lock, and a new message can be put in during the + * time the lock is released. + */ msg = smi_info->curr_msg; smi_info->curr_msg = NULL; deliver_recv_msg(smi_info, msg); @@ -412,12 +565,13 @@ static void handle_transaction_done(struct smi_info *smi_info) /* We got the flags from the SMI, now handle them. */ len = smi_info->handlers->get_result(smi_info->si_sm, msg, 4); if (msg[2] != 0) { - /* Error fetching flags, just give up for - now. */ + /* Error fetching flags, just give up for now. */ smi_info->si_state = SI_NORMAL; } else if (len < 4) { - /* Hmm, no flags. That's technically illegal, but - don't use uninitialized data. */ + /* + * Hmm, no flags. That's technically illegal, but + * don't use uninitialized data. + */ smi_info->si_state = SI_NORMAL; } else { smi_info->msg_flags = msg[3]; @@ -435,9 +589,8 @@ static void handle_transaction_done(struct smi_info *smi_info) smi_info->handlers->get_result(smi_info->si_sm, msg, 3); if (msg[2] != 0) { /* Error clearing flags */ - printk(KERN_WARNING - "ipmi_si: Error clearing flags: %2.2x\n", - msg[2]); + dev_warn(smi_info->dev, + "Error clearing flags: %2.2x\n", msg[2]); } if (smi_info->si_state == SI_CLEARING_FLAGS_THEN_SET_IRQ) start_enable_irq(smi_info); @@ -454,9 +607,11 @@ static void handle_transaction_done(struct smi_info *smi_info) smi_info->curr_msg->rsp, IPMI_MAX_MSG_LENGTH); - /* Do this here becase deliver_recv_msg() releases the - lock, and a new message can be put in during the - time the lock is released. */ + /* + * Do this here becase deliver_recv_msg() releases the + * lock, and a new message can be put in during the + * time the lock is released. + */ msg = smi_info->curr_msg; smi_info->curr_msg = NULL; if (msg->rsp[2] != 0) { @@ -467,14 +622,14 @@ static void handle_transaction_done(struct smi_info *smi_info) smi_info->msg_flags &= ~EVENT_MSG_BUFFER_FULL; handle_flags(smi_info); } else { - spin_lock(&smi_info->count_lock); - smi_info->events++; - spin_unlock(&smi_info->count_lock); - - /* Do this before we deliver the message - because delivering the message releases the - lock and something else can mess with the - state. */ + smi_inc_stat(smi_info, events); + + /* + * Do this before we deliver the message + * because delivering the message releases the + * lock and something else can mess with the + * state. + */ handle_flags(smi_info); deliver_recv_msg(smi_info, msg); @@ -490,9 +645,11 @@ static void handle_transaction_done(struct smi_info *smi_info) smi_info->curr_msg->rsp, IPMI_MAX_MSG_LENGTH); - /* Do this here becase deliver_recv_msg() releases the - lock, and a new message can be put in during the - time the lock is released. */ + /* + * Do this here becase deliver_recv_msg() releases the + * lock, and a new message can be put in during the + * time the lock is released. + */ msg = smi_info->curr_msg; smi_info->curr_msg = NULL; if (msg->rsp[2] != 0) { @@ -503,14 +660,14 @@ static void handle_transaction_done(struct smi_info *smi_info) smi_info->msg_flags &= ~RECEIVE_MSG_AVAIL; handle_flags(smi_info); } else { - spin_lock(&smi_info->count_lock); - smi_info->incoming_messages++; - spin_unlock(&smi_info->count_lock); - - /* Do this before we deliver the message - because delivering the message releases the - lock and something else can mess with the - state. */ + smi_inc_stat(smi_info, incoming_messages); + + /* + * Do this before we deliver the message + * because delivering the message releases the + * lock and something else can mess with the + * state. + */ handle_flags(smi_info); deliver_recv_msg(smi_info, msg); @@ -525,14 +682,17 @@ static void handle_transaction_done(struct smi_info *smi_info) /* We got the flags from the SMI, now handle them. */ smi_info->handlers->get_result(smi_info->si_sm, msg, 4); if (msg[2] != 0) { - printk(KERN_WARNING - "ipmi_si: Could not enable interrupts" - ", failed get, using polled mode.\n"); + dev_warn(smi_info->dev, + "Couldn't get irq info: %x.\n", msg[2]); + dev_warn(smi_info->dev, + "Maybe ok, but ipmi might run very slowly.\n"); smi_info->si_state = SI_NORMAL; } else { msg[0] = (IPMI_NETFN_APP_REQUEST << 2); msg[1] = IPMI_SET_BMC_GLOBAL_ENABLES_CMD; - msg[2] = msg[3] | 1; /* enable msg queue int */ + msg[2] = (msg[3] | + IPMI_BMC_RCV_MSG_INTR | + IPMI_BMC_EVT_MSG_INTR); smi_info->handlers->start_transaction( smi_info->si_sm, msg, 3); smi_info->si_state = SI_ENABLE_INTERRUPTS2; @@ -547,9 +707,48 @@ static void handle_transaction_done(struct smi_info *smi_info) /* We got the flags from the SMI, now handle them. */ smi_info->handlers->get_result(smi_info->si_sm, msg, 4); if (msg[2] != 0) { - printk(KERN_WARNING - "ipmi_si: Could not enable interrupts" - ", failed set, using polled mode.\n"); + dev_warn(smi_info->dev, + "Couldn't set irq info: %x.\n", msg[2]); + dev_warn(smi_info->dev, + "Maybe ok, but ipmi might run very slowly.\n"); + } else + smi_info->interrupt_disabled = false; + smi_info->si_state = SI_NORMAL; + break; + } + + case SI_DISABLE_INTERRUPTS1: + { + unsigned char msg[4]; + + /* We got the flags from the SMI, now handle them. */ + smi_info->handlers->get_result(smi_info->si_sm, msg, 4); + if (msg[2] != 0) { + dev_warn(smi_info->dev, "Could not disable interrupts" + ", failed get.\n"); + smi_info->si_state = SI_NORMAL; + } else { + msg[0] = (IPMI_NETFN_APP_REQUEST << 2); + msg[1] = IPMI_SET_BMC_GLOBAL_ENABLES_CMD; + msg[2] = (msg[3] & + ~(IPMI_BMC_RCV_MSG_INTR | + IPMI_BMC_EVT_MSG_INTR)); + smi_info->handlers->start_transaction( + smi_info->si_sm, msg, 3); + smi_info->si_state = SI_DISABLE_INTERRUPTS2; + } + break; + } + + case SI_DISABLE_INTERRUPTS2: + { + unsigned char msg[4]; + + /* We got the flags from the SMI, now handle them. */ + smi_info->handlers->get_result(smi_info->si_sm, msg, 4); + if (msg[2] != 0) { + dev_warn(smi_info->dev, "Could not disable interrupts" + ", failed set.\n"); } smi_info->si_state = SI_NORMAL; break; @@ -557,68 +756,70 @@ static void handle_transaction_done(struct smi_info *smi_info) } } -/* Called on timeouts and events. Timeouts should pass the elapsed - time, interrupts should pass in zero. */ +/* + * Called on timeouts and events. Timeouts should pass the elapsed + * time, interrupts should pass in zero. Must be called with + * si_lock held and interrupts disabled. + */ static enum si_sm_result smi_event_handler(struct smi_info *smi_info, int time) { enum si_sm_result si_sm_result; restart: - /* There used to be a loop here that waited a little while - (around 25us) before giving up. That turned out to be - pointless, the minimum delays I was seeing were in the 300us - range, which is far too long to wait in an interrupt. So - we just run until the state machine tells us something - happened or it needs a delay. */ + /* + * There used to be a loop here that waited a little while + * (around 25us) before giving up. That turned out to be + * pointless, the minimum delays I was seeing were in the 300us + * range, which is far too long to wait in an interrupt. So + * we just run until the state machine tells us something + * happened or it needs a delay. + */ si_sm_result = smi_info->handlers->event(smi_info->si_sm, time); time = 0; while (si_sm_result == SI_SM_CALL_WITHOUT_DELAY) - { si_sm_result = smi_info->handlers->event(smi_info->si_sm, 0); - } - if (si_sm_result == SI_SM_TRANSACTION_COMPLETE) - { - spin_lock(&smi_info->count_lock); - smi_info->complete_transactions++; - spin_unlock(&smi_info->count_lock); + if (si_sm_result == SI_SM_TRANSACTION_COMPLETE) { + smi_inc_stat(smi_info, complete_transactions); handle_transaction_done(smi_info); si_sm_result = smi_info->handlers->event(smi_info->si_sm, 0); - } - else if (si_sm_result == SI_SM_HOSED) - { - spin_lock(&smi_info->count_lock); - smi_info->hosed_count++; - spin_unlock(&smi_info->count_lock); + } else if (si_sm_result == SI_SM_HOSED) { + smi_inc_stat(smi_info, hosed_count); - /* Do the before return_hosed_msg, because that - releases the lock. */ + /* + * Do the before return_hosed_msg, because that + * releases the lock. + */ smi_info->si_state = SI_NORMAL; if (smi_info->curr_msg != NULL) { - /* If we were handling a user message, format - a response to send to the upper layer to - tell it about the error. */ - return_hosed_msg(smi_info); + /* + * If we were handling a user message, format + * a response to send to the upper layer to + * tell it about the error. + */ + return_hosed_msg(smi_info, IPMI_ERR_UNSPECIFIED); } si_sm_result = smi_info->handlers->event(smi_info->si_sm, 0); } - /* We prefer handling attn over new messages. */ - if (si_sm_result == SI_SM_ATTN) - { + /* + * We prefer handling attn over new messages. But don't do + * this if there is not yet an upper layer to handle anything. + */ + if (likely(smi_info->intf) && si_sm_result == SI_SM_ATTN) { unsigned char msg[2]; - spin_lock(&smi_info->count_lock); - smi_info->attentions++; - spin_unlock(&smi_info->count_lock); + smi_inc_stat(smi_info, attentions); - /* Got a attn, send down a get message flags to see - what's causing it. It would be better to handle - this in the upper layer, but due to the way - interrupts work with the SMI, that's not really - possible. */ + /* + * Got a attn, send down a get message flags to see + * what's causing it. It would be better to handle + * this in the upper layer, but due to the way + * interrupts work with the SMI, that's not really + * possible. + */ msg[0] = (IPMI_NETFN_APP_REQUEST << 2); msg[1] = IPMI_GET_MSG_FLAGS_CMD; @@ -630,39 +831,53 @@ static enum si_sm_result smi_event_handler(struct smi_info *smi_info, /* If we are currently idle, try to start the next message. */ if (si_sm_result == SI_SM_IDLE) { - spin_lock(&smi_info->count_lock); - smi_info->idles++; - spin_unlock(&smi_info->count_lock); + smi_inc_stat(smi_info, idles); si_sm_result = start_next_msg(smi_info); if (si_sm_result != SI_SM_IDLE) goto restart; - } + } if ((si_sm_result == SI_SM_IDLE) - && (atomic_read(&smi_info->req_events))) - { - /* We are idle and the upper layer requested that I fetch - events, so do so. */ - unsigned char msg[2]; + && (atomic_read(&smi_info->req_events))) { + /* + * We are idle and the upper layer requested that I fetch + * events, so do so. + */ + atomic_set(&smi_info->req_events, 0); - spin_lock(&smi_info->count_lock); - smi_info->flag_fetches++; - spin_unlock(&smi_info->count_lock); + smi_info->curr_msg = ipmi_alloc_smi_msg(); + if (!smi_info->curr_msg) + goto out; - atomic_set(&smi_info->req_events, 0); - msg[0] = (IPMI_NETFN_APP_REQUEST << 2); - msg[1] = IPMI_GET_MSG_FLAGS_CMD; + smi_info->curr_msg->data[0] = (IPMI_NETFN_APP_REQUEST << 2); + smi_info->curr_msg->data[1] = IPMI_READ_EVENT_MSG_BUFFER_CMD; + smi_info->curr_msg->data_size = 2; smi_info->handlers->start_transaction( - smi_info->si_sm, msg, 2); - smi_info->si_state = SI_GETTING_FLAGS; + smi_info->si_sm, + smi_info->curr_msg->data, + smi_info->curr_msg->data_size); + smi_info->si_state = SI_GETTING_EVENTS; goto restart; } - + out: return si_sm_result; } +static void check_start_timer_thread(struct smi_info *smi_info) +{ + if (smi_info->si_state == SI_NORMAL && smi_info->curr_msg == NULL) { + smi_mod_timer(smi_info, jiffies + SI_TIMEOUT_JIFFIES); + + if (smi_info->thread) + wake_up_process(smi_info->thread); + + start_next_msg(smi_info); + smi_event_handler(smi_info, 0); + } +} + static void sender(void *send_info, struct ipmi_smi_msg *msg, int priority) @@ -674,57 +889,56 @@ static void sender(void *send_info, struct timeval t; #endif - spin_lock_irqsave(&(smi_info->msg_lock), flags); + if (atomic_read(&smi_info->stop_operation)) { + msg->rsp[0] = msg->data[0] | 4; + msg->rsp[1] = msg->data[1]; + msg->rsp[2] = IPMI_ERR_UNSPECIFIED; + msg->rsp_size = 3; + deliver_recv_msg(smi_info, msg); + return; + } + #ifdef DEBUG_TIMING do_gettimeofday(&t); printk("**Enqueue: %d.%9.9d\n", t.tv_sec, t.tv_usec); #endif if (smi_info->run_to_completion) { - /* If we are running to completion, then throw it in - the list and run transactions until everything is - clear. Priority doesn't matter here. */ + /* + * If we are running to completion, then throw it in + * the list and run transactions until everything is + * clear. Priority doesn't matter here. + */ + + /* + * Run to completion means we are single-threaded, no + * need for locks. + */ list_add_tail(&(msg->link), &(smi_info->xmit_msgs)); - /* We have to release the msg lock and claim the smi - lock in this case, because of race conditions. */ - spin_unlock_irqrestore(&(smi_info->msg_lock), flags); - - spin_lock_irqsave(&(smi_info->si_lock), flags); result = smi_event_handler(smi_info, 0); while (result != SI_SM_IDLE) { udelay(SI_SHORT_TIMEOUT_USEC); result = smi_event_handler(smi_info, SI_SHORT_TIMEOUT_USEC); } - spin_unlock_irqrestore(&(smi_info->si_lock), flags); return; - } else { - if (priority > 0) { - list_add_tail(&(msg->link), &(smi_info->hp_xmit_msgs)); - } else { - list_add_tail(&(msg->link), &(smi_info->xmit_msgs)); - } } - spin_unlock_irqrestore(&(smi_info->msg_lock), flags); - spin_lock_irqsave(&(smi_info->si_lock), flags); - if ((smi_info->si_state == SI_NORMAL) - && (smi_info->curr_msg == NULL)) - { - start_next_msg(smi_info); - si_restart_short_timer(smi_info); - } - spin_unlock_irqrestore(&(smi_info->si_lock), flags); + spin_lock_irqsave(&smi_info->si_lock, flags); + if (priority > 0) + list_add_tail(&msg->link, &smi_info->hp_xmit_msgs); + else + list_add_tail(&msg->link, &smi_info->xmit_msgs); + + check_start_timer_thread(smi_info); + spin_unlock_irqrestore(&smi_info->si_lock, flags); } -static void set_run_to_completion(void *send_info, int i_run_to_completion) +static void set_run_to_completion(void *send_info, bool i_run_to_completion) { struct smi_info *smi_info = send_info; enum si_sm_result result; - unsigned long flags; - - spin_lock_irqsave(&(smi_info->si_lock), flags); smi_info->run_to_completion = i_run_to_completion; if (i_run_to_completion) { @@ -735,125 +949,195 @@ static void set_run_to_completion(void *send_info, int i_run_to_completion) SI_SHORT_TIMEOUT_USEC); } } +} - spin_unlock_irqrestore(&(smi_info->si_lock), flags); +/* + * Use -1 in the nsec value of the busy waiting timespec to tell that + * we are spinning in kipmid looking for something and not delaying + * between checks + */ +static inline void ipmi_si_set_not_busy(struct timespec *ts) +{ + ts->tv_nsec = -1; +} +static inline int ipmi_si_is_busy(struct timespec *ts) +{ + return ts->tv_nsec != -1; +} + +static int ipmi_thread_busy_wait(enum si_sm_result smi_result, + const struct smi_info *smi_info, + struct timespec *busy_until) +{ + unsigned int max_busy_us = 0; + + if (smi_info->intf_num < num_max_busy_us) + max_busy_us = kipmid_max_busy_us[smi_info->intf_num]; + if (max_busy_us == 0 || smi_result != SI_SM_CALL_WITH_DELAY) + ipmi_si_set_not_busy(busy_until); + else if (!ipmi_si_is_busy(busy_until)) { + getnstimeofday(busy_until); + timespec_add_ns(busy_until, max_busy_us*NSEC_PER_USEC); + } else { + struct timespec now; + getnstimeofday(&now); + if (unlikely(timespec_compare(&now, busy_until) > 0)) { + ipmi_si_set_not_busy(busy_until); + return 0; + } + } + return 1; } + +/* + * A busy-waiting loop for speeding up IPMI operation. + * + * Lousy hardware makes this hard. This is only enabled for systems + * that are not BT and do not have interrupts. It starts spinning + * when an operation is complete or until max_busy tells it to stop + * (if that is enabled). See the paragraph on kimid_max_busy_us in + * Documentation/IPMI.txt for details. + */ +static int ipmi_thread(void *data) +{ + struct smi_info *smi_info = data; + unsigned long flags; + enum si_sm_result smi_result; + struct timespec busy_until; + + ipmi_si_set_not_busy(&busy_until); + set_user_nice(current, MAX_NICE); + while (!kthread_should_stop()) { + int busy_wait; + + spin_lock_irqsave(&(smi_info->si_lock), flags); + smi_result = smi_event_handler(smi_info, 0); + + /* + * If the driver is doing something, there is a possible + * race with the timer. If the timer handler see idle, + * and the thread here sees something else, the timer + * handler won't restart the timer even though it is + * required. So start it here if necessary. + */ + if (smi_result != SI_SM_IDLE && !smi_info->timer_running) + smi_mod_timer(smi_info, jiffies + SI_TIMEOUT_JIFFIES); + + spin_unlock_irqrestore(&(smi_info->si_lock), flags); + busy_wait = ipmi_thread_busy_wait(smi_result, smi_info, + &busy_until); + if (smi_result == SI_SM_CALL_WITHOUT_DELAY) + ; /* do nothing */ + else if (smi_result == SI_SM_CALL_WITH_DELAY && busy_wait) + schedule(); + else if (smi_result == SI_SM_IDLE) { + if (atomic_read(&smi_info->need_watch)) { + schedule_timeout_interruptible(100); + } else { + /* Wait to be woken up when we are needed. */ + __set_current_state(TASK_INTERRUPTIBLE); + schedule(); + } + } else + schedule_timeout_interruptible(1); + } + return 0; +} + + static void poll(void *send_info) { struct smi_info *smi_info = send_info; + unsigned long flags = 0; + bool run_to_completion = smi_info->run_to_completion; - smi_event_handler(smi_info, 0); + /* + * Make sure there is some delay in the poll loop so we can + * drive time forward and timeout things. + */ + udelay(10); + if (!run_to_completion) + spin_lock_irqsave(&smi_info->si_lock, flags); + smi_event_handler(smi_info, 10); + if (!run_to_completion) + spin_unlock_irqrestore(&smi_info->si_lock, flags); } static void request_events(void *send_info) { struct smi_info *smi_info = send_info; + if (atomic_read(&smi_info->stop_operation) || + !smi_info->has_event_buffer) + return; + atomic_set(&smi_info->req_events, 1); } -static int initialized = 0; - -/* Must be called with interrupts off and with the si_lock held. */ -static void si_restart_short_timer(struct smi_info *smi_info) +static void set_need_watch(void *send_info, bool enable) { -#if defined(CONFIG_HIGH_RES_TIMERS) + struct smi_info *smi_info = send_info; unsigned long flags; - unsigned long jiffies_now; - - if (del_timer(&(smi_info->si_timer))) { - /* If we don't delete the timer, then it will go off - immediately, anyway. So we only process if we - actually delete the timer. */ - /* We already have irqsave on, so no need for it - here. */ - read_lock(&xtime_lock); - jiffies_now = jiffies; - smi_info->si_timer.expires = jiffies_now; - smi_info->si_timer.sub_expires = get_arch_cycles(jiffies_now); - - add_usec_to_timer(&smi_info->si_timer, SI_SHORT_TIMEOUT_USEC); - - add_timer(&(smi_info->si_timer)); - spin_lock_irqsave(&smi_info->count_lock, flags); - smi_info->timeout_restarts++; - spin_unlock_irqrestore(&smi_info->count_lock, flags); - } -#endif + atomic_set(&smi_info->need_watch, enable); + spin_lock_irqsave(&smi_info->si_lock, flags); + check_start_timer_thread(smi_info); + spin_unlock_irqrestore(&smi_info->si_lock, flags); } +static int initialized; + static void smi_timeout(unsigned long data) { struct smi_info *smi_info = (struct smi_info *) data; enum si_sm_result smi_result; unsigned long flags; unsigned long jiffies_now; - unsigned long time_diff; + long time_diff; + long timeout; #ifdef DEBUG_TIMING struct timeval t; #endif - if (smi_info->stop_operation) { - smi_info->timer_stopped = 1; - return; - } - spin_lock_irqsave(&(smi_info->si_lock), flags); #ifdef DEBUG_TIMING do_gettimeofday(&t); - printk("**Timer: %d.%9.9d\n", t.tv_sec, t.tv_usec); + printk(KERN_DEBUG "**Timer: %d.%9.9d\n", t.tv_sec, t.tv_usec); #endif jiffies_now = jiffies; - time_diff = ((jiffies_now - smi_info->last_timeout_jiffies) + time_diff = (((long)jiffies_now - (long)smi_info->last_timeout_jiffies) * SI_USEC_PER_JIFFY); smi_result = smi_event_handler(smi_info, time_diff); - spin_unlock_irqrestore(&(smi_info->si_lock), flags); - - smi_info->last_timeout_jiffies = jiffies_now; - - if ((smi_info->irq) && (! smi_info->interrupt_disabled)) { + if ((smi_info->irq) && (!smi_info->interrupt_disabled)) { /* Running with interrupts, only do long timeouts. */ - smi_info->si_timer.expires = jiffies + SI_TIMEOUT_JIFFIES; - spin_lock_irqsave(&smi_info->count_lock, flags); - smi_info->long_timeouts++; - spin_unlock_irqrestore(&smi_info->count_lock, flags); - goto do_add_timer; + timeout = jiffies + SI_TIMEOUT_JIFFIES; + smi_inc_stat(smi_info, long_timeouts); + goto do_mod_timer; } - /* If the state machine asks for a short delay, then shorten - the timer timeout. */ + /* + * If the state machine asks for a short delay, then shorten + * the timer timeout. + */ if (smi_result == SI_SM_CALL_WITH_DELAY) { - spin_lock_irqsave(&smi_info->count_lock, flags); - smi_info->short_timeouts++; - spin_unlock_irqrestore(&smi_info->count_lock, flags); -#if defined(CONFIG_HIGH_RES_TIMERS) - read_lock(&xtime_lock); - smi_info->si_timer.expires = jiffies; - smi_info->si_timer.sub_expires - = get_arch_cycles(smi_info->si_timer.expires); - read_unlock(&xtime_lock); - add_usec_to_timer(&smi_info->si_timer, SI_SHORT_TIMEOUT_USEC); -#else - smi_info->si_timer.expires = jiffies + 1; -#endif + smi_inc_stat(smi_info, short_timeouts); + timeout = jiffies + 1; } else { - spin_lock_irqsave(&smi_info->count_lock, flags); - smi_info->long_timeouts++; - spin_unlock_irqrestore(&smi_info->count_lock, flags); - smi_info->si_timer.expires = jiffies + SI_TIMEOUT_JIFFIES; -#if defined(CONFIG_HIGH_RES_TIMERS) - smi_info->si_timer.sub_expires = 0; -#endif + smi_inc_stat(smi_info, long_timeouts); + timeout = jiffies + SI_TIMEOUT_JIFFIES; } - do_add_timer: - add_timer(&(smi_info->si_timer)); + do_mod_timer: + if (smi_result != SI_SM_IDLE) + smi_mod_timer(smi_info, timeout); + else + smi_info->timer_running = false; + spin_unlock_irqrestore(&(smi_info->si_lock), flags); } -static irqreturn_t si_irq_handler(int irq, void *data, struct pt_regs *regs) +static irqreturn_t si_irq_handler(int irq, void *data) { struct smi_info *smi_info = data; unsigned long flags; @@ -863,78 +1147,173 @@ static irqreturn_t si_irq_handler(int irq, void *data, struct pt_regs *regs) spin_lock_irqsave(&(smi_info->si_lock), flags); - spin_lock(&smi_info->count_lock); - smi_info->interrupts++; - spin_unlock(&smi_info->count_lock); - - if (smi_info->stop_operation) - goto out; + smi_inc_stat(smi_info, interrupts); #ifdef DEBUG_TIMING do_gettimeofday(&t); - printk("**Interrupt: %d.%9.9d\n", t.tv_sec, t.tv_usec); + printk(KERN_DEBUG "**Interrupt: %d.%9.9d\n", t.tv_sec, t.tv_usec); #endif smi_event_handler(smi_info, 0); - out: spin_unlock_irqrestore(&(smi_info->si_lock), flags); return IRQ_HANDLED; } -static irqreturn_t si_bt_irq_handler(int irq, void *data, struct pt_regs *regs) +static irqreturn_t si_bt_irq_handler(int irq, void *data) { struct smi_info *smi_info = data; /* We need to clear the IRQ flag for the BT interface. */ smi_info->io.outputb(&smi_info->io, IPMI_BT_INTMASK_REG, IPMI_BT_INTMASK_CLEAR_IRQ_BIT | IPMI_BT_INTMASK_ENABLE_IRQ_BIT); - return si_irq_handler(irq, data, regs); + return si_irq_handler(irq, data); +} + +static int smi_start_processing(void *send_info, + ipmi_smi_t intf) +{ + struct smi_info *new_smi = send_info; + int enable = 0; + + new_smi->intf = intf; + + /* Try to claim any interrupts. */ + if (new_smi->irq_setup) + new_smi->irq_setup(new_smi); + + /* Set up the timer that drives the interface. */ + setup_timer(&new_smi->si_timer, smi_timeout, (long)new_smi); + smi_mod_timer(new_smi, jiffies + SI_TIMEOUT_JIFFIES); + + /* + * Check if the user forcefully enabled the daemon. + */ + if (new_smi->intf_num < num_force_kipmid) + enable = force_kipmid[new_smi->intf_num]; + /* + * The BT interface is efficient enough to not need a thread, + * and there is no need for a thread if we have interrupts. + */ + else if ((new_smi->si_type != SI_BT) && (!new_smi->irq)) + enable = 1; + + if (enable) { + new_smi->thread = kthread_run(ipmi_thread, new_smi, + "kipmi%d", new_smi->intf_num); + if (IS_ERR(new_smi->thread)) { + dev_notice(new_smi->dev, "Could not start" + " kernel thread due to error %ld, only using" + " timers to drive the interface\n", + PTR_ERR(new_smi->thread)); + new_smi->thread = NULL; + } + } + + return 0; } +static int get_smi_info(void *send_info, struct ipmi_smi_info *data) +{ + struct smi_info *smi = send_info; + + data->addr_src = smi->addr_source; + data->dev = smi->dev; + data->addr_info = smi->addr_info; + get_device(smi->dev); -static struct ipmi_smi_handlers handlers = + return 0; +} + +static void set_maintenance_mode(void *send_info, bool enable) { + struct smi_info *smi_info = send_info; + + if (!enable) + atomic_set(&smi_info->req_events, 0); +} + +static struct ipmi_smi_handlers handlers = { .owner = THIS_MODULE, + .start_processing = smi_start_processing, + .get_smi_info = get_smi_info, .sender = sender, .request_events = request_events, + .set_need_watch = set_need_watch, + .set_maintenance_mode = set_maintenance_mode, .set_run_to_completion = set_run_to_completion, .poll = poll, }; -/* There can be 4 IO ports passed in (with or without IRQs), 4 addresses, - a default IO port, and 1 ACPI/SPMI address. That sets SI_MAX_DRIVERS */ - -#define SI_MAX_PARMS 4 -#define SI_MAX_DRIVERS ((SI_MAX_PARMS * 2) + 2) -static struct smi_info *smi_infos[SI_MAX_DRIVERS] = -{ NULL, NULL, NULL, NULL }; +/* + * There can be 4 IO ports passed in (with or without IRQs), 4 addresses, + * a default IO port, and 1 ACPI/SPMI address. That sets SI_MAX_DRIVERS. + */ -#define DEVICE_NAME "ipmi_si" +static LIST_HEAD(smi_infos); +static DEFINE_MUTEX(smi_infos_lock); +static int smi_num; /* Used to sequence the SMIs */ -#define DEFAULT_KCS_IO_PORT 0xca2 -#define DEFAULT_SMIC_IO_PORT 0xca9 -#define DEFAULT_BT_IO_PORT 0xe4 #define DEFAULT_REGSPACING 1 +#define DEFAULT_REGSIZE 1 -static int si_trydefaults = 1; +#ifdef CONFIG_ACPI +static bool si_tryacpi = 1; +#endif +#ifdef CONFIG_DMI +static bool si_trydmi = 1; +#endif +static bool si_tryplatform = 1; +#ifdef CONFIG_PCI +static bool si_trypci = 1; +#endif +static bool si_trydefaults = IS_ENABLED(CONFIG_IPMI_SI_PROBE_DEFAULTS); static char *si_type[SI_MAX_PARMS]; #define MAX_SI_TYPE_STR 30 static char si_type_str[MAX_SI_TYPE_STR]; static unsigned long addrs[SI_MAX_PARMS]; -static int num_addrs; +static unsigned int num_addrs; static unsigned int ports[SI_MAX_PARMS]; -static int num_ports; +static unsigned int num_ports; static int irqs[SI_MAX_PARMS]; -static int num_irqs; +static unsigned int num_irqs; static int regspacings[SI_MAX_PARMS]; -static int num_regspacings = 0; +static unsigned int num_regspacings; static int regsizes[SI_MAX_PARMS]; -static int num_regsizes = 0; +static unsigned int num_regsizes; static int regshifts[SI_MAX_PARMS]; -static int num_regshifts = 0; -static int slave_addrs[SI_MAX_PARMS]; -static int num_slave_addrs = 0; +static unsigned int num_regshifts; +static int slave_addrs[SI_MAX_PARMS]; /* Leaving 0 chooses the default value */ +static unsigned int num_slave_addrs; + +#define IPMI_IO_ADDR_SPACE 0 +#define IPMI_MEM_ADDR_SPACE 1 +static char *addr_space_to_str[] = { "i/o", "mem" }; + +static int hotmod_handler(const char *val, struct kernel_param *kp); +module_param_call(hotmod, hotmod_handler, NULL, NULL, 0200); +MODULE_PARM_DESC(hotmod, "Add and remove interfaces. See" + " Documentation/IPMI.txt in the kernel sources for the" + " gory details."); +#ifdef CONFIG_ACPI +module_param_named(tryacpi, si_tryacpi, bool, 0); +MODULE_PARM_DESC(tryacpi, "Setting this to zero will disable the" + " default scan of the interfaces identified via ACPI"); +#endif +#ifdef CONFIG_DMI +module_param_named(trydmi, si_trydmi, bool, 0); +MODULE_PARM_DESC(trydmi, "Setting this to zero will disable the" + " default scan of the interfaces identified via DMI"); +#endif +module_param_named(tryplatform, si_tryplatform, bool, 0); +MODULE_PARM_DESC(tryacpi, "Setting this to zero will disable the" + " default scan of the interfaces identified via platform" + " interfaces like openfirmware"); +#ifdef CONFIG_PCI +module_param_named(trypci, si_trypci, bool, 0); +MODULE_PARM_DESC(tryacpi, "Setting this to zero will disable the" + " default scan of the interfaces identified via pci"); +#endif module_param_named(trydefaults, si_trydefaults, bool, 0); MODULE_PARM_DESC(trydefaults, "Setting this to 'false' will disable the" " default scan of the KCS and SMIC interface at the standard" @@ -944,12 +1323,12 @@ MODULE_PARM_DESC(type, "Defines the type of each interface, each" " interface separated by commas. The types are 'kcs'," " 'smic', and 'bt'. For example si_type=kcs,bt will set" " the first interface to kcs and the second to bt"); -module_param_array(addrs, long, &num_addrs, 0); +module_param_array(addrs, ulong, &num_addrs, 0); MODULE_PARM_DESC(addrs, "Sets the memory address of each interface, the" " addresses separated by commas. Only use if an interface" " is in memory. Otherwise, set it to zero or leave" " it blank."); -module_param_array(ports, int, &num_ports, 0); +module_param_array(ports, uint, &num_ports, 0); MODULE_PARM_DESC(ports, "Sets the port address of each interface, the" " addresses separated by commas. Only use if an interface" " is a port. Otherwise, set it to zero or leave" @@ -981,34 +1360,28 @@ MODULE_PARM_DESC(slave_addrs, "Set the default IPMB slave address for" " the controller. Normally this is 0x20, but can be" " overridden by this parm. This is an array indexed" " by interface number."); +module_param_array(force_kipmid, int, &num_force_kipmid, 0); +MODULE_PARM_DESC(force_kipmid, "Force the kipmi daemon to be enabled (1) or" + " disabled(0). Normally the IPMI driver auto-detects" + " this, but the value may be overridden by this parm."); +module_param(unload_when_empty, bool, 0); +MODULE_PARM_DESC(unload_when_empty, "Unload the module if no interfaces are" + " specified or found, default is 1. Setting to 0" + " is useful for hot add of devices using hotmod."); +module_param_array(kipmid_max_busy_us, uint, &num_max_busy_us, 0644); +MODULE_PARM_DESC(kipmid_max_busy_us, + "Max time (in microseconds) to busy-wait for IPMI data before" + " sleeping. 0 (default) means to wait forever. Set to 100-500" + " if kipmid is using up a lot of CPU time."); -#define IPMI_MEM_ADDR_SPACE 1 -#define IPMI_IO_ADDR_SPACE 2 - -#if defined(CONFIG_ACPI_INTERPRETER) || defined(CONFIG_X86) || defined(CONFIG_PCI) -static int is_new_interface(int intf, u8 addr_space, unsigned long base_addr) +static void std_irq_cleanup(struct smi_info *info) { - int i; - - for (i = 0; i < SI_MAX_PARMS; ++i) { - /* Don't check our address. */ - if (i == intf) - continue; - if (si_type[i] != NULL) { - if ((addr_space == IPMI_MEM_ADDR_SPACE && - base_addr == addrs[i]) || - (addr_space == IPMI_IO_ADDR_SPACE && - base_addr == ports[i])) - return 0; - } - else - break; - } - - return 1; + if (info->si_type == SI_BT) + /* Disable the interrupt in the BT interface. */ + info->io.outputb(&info->io, IPMI_BT_INTMASK_REG, 0); + free_irq(info->irq, info); } -#endif static int std_irq_setup(struct smi_info *info) { @@ -1020,7 +1393,7 @@ static int std_irq_setup(struct smi_info *info) if (info->si_type == SI_BT) { rv = request_irq(info->irq, si_bt_irq_handler, - SA_INTERRUPT, + IRQF_SHARED, DEVICE_NAME, info); if (!rv) @@ -1030,104 +1403,93 @@ static int std_irq_setup(struct smi_info *info) } else rv = request_irq(info->irq, si_irq_handler, - SA_INTERRUPT, + IRQF_SHARED, DEVICE_NAME, info); if (rv) { - printk(KERN_WARNING - "ipmi_si: %s unable to claim interrupt %d," - " running polled\n", - DEVICE_NAME, info->irq); + dev_warn(info->dev, "%s unable to claim interrupt %d," + " running polled\n", + DEVICE_NAME, info->irq); info->irq = 0; } else { - printk(" Using irq %d\n", info->irq); + info->irq_cleanup = std_irq_cleanup; + dev_info(info->dev, "Using irq %d\n", info->irq); } return rv; } -static void std_irq_cleanup(struct smi_info *info) -{ - if (!info->irq) - return; - - if (info->si_type == SI_BT) - /* Disable the interrupt in the BT interface. */ - info->io.outputb(&info->io, IPMI_BT_INTMASK_REG, 0); - free_irq(info->irq, info); -} - static unsigned char port_inb(struct si_sm_io *io, unsigned int offset) { - unsigned int *addr = io->info; + unsigned int addr = io->addr_data; - return inb((*addr)+(offset*io->regspacing)); + return inb(addr + (offset * io->regspacing)); } static void port_outb(struct si_sm_io *io, unsigned int offset, unsigned char b) { - unsigned int *addr = io->info; + unsigned int addr = io->addr_data; - outb(b, (*addr)+(offset * io->regspacing)); + outb(b, addr + (offset * io->regspacing)); } static unsigned char port_inw(struct si_sm_io *io, unsigned int offset) { - unsigned int *addr = io->info; + unsigned int addr = io->addr_data; - return (inw((*addr)+(offset * io->regspacing)) >> io->regshift) & 0xff; + return (inw(addr + (offset * io->regspacing)) >> io->regshift) & 0xff; } static void port_outw(struct si_sm_io *io, unsigned int offset, unsigned char b) { - unsigned int *addr = io->info; + unsigned int addr = io->addr_data; - outw(b << io->regshift, (*addr)+(offset * io->regspacing)); + outw(b << io->regshift, addr + (offset * io->regspacing)); } static unsigned char port_inl(struct si_sm_io *io, unsigned int offset) { - unsigned int *addr = io->info; + unsigned int addr = io->addr_data; - return (inl((*addr)+(offset * io->regspacing)) >> io->regshift) & 0xff; + return (inl(addr + (offset * io->regspacing)) >> io->regshift) & 0xff; } static void port_outl(struct si_sm_io *io, unsigned int offset, unsigned char b) { - unsigned int *addr = io->info; + unsigned int addr = io->addr_data; - outl(b << io->regshift, (*addr)+(offset * io->regspacing)); + outl(b << io->regshift, addr+(offset * io->regspacing)); } static void port_cleanup(struct smi_info *info) { - unsigned int *addr = info->io.info; - int mapsize; - - if (addr && (*addr)) { - mapsize = ((info->io_size * info->io.regspacing) - - (info->io.regspacing - info->io.regsize)); + unsigned int addr = info->io.addr_data; + int idx; - release_region (*addr, mapsize); + if (addr) { + for (idx = 0; idx < info->io_size; idx++) + release_region(addr + idx * info->io.regspacing, + info->io.regsize); } - kfree(info); } static int port_setup(struct smi_info *info) { - unsigned int *addr = info->io.info; - int mapsize; + unsigned int addr = info->io.addr_data; + int idx; - if (!addr || (!*addr)) + if (!addr) return -ENODEV; info->io_cleanup = port_cleanup; - /* Figure out the actual inb/inw/inl/etc routine to use based - upon the register size. */ + /* + * Figure out the actual inb/inw/inl/etc routine to use based + * upon the register size. + */ switch (info->io.regsize) { case 1: info->io.inputb = port_inb; @@ -1142,94 +1504,61 @@ static int port_setup(struct smi_info *info) info->io.outputb = port_outl; break; default: - printk("ipmi_si: Invalid register size: %d\n", - info->io.regsize); + dev_warn(info->dev, "Invalid register size: %d\n", + info->io.regsize); return -EINVAL; } - /* Calculate the total amount of memory to claim. This is an - * unusual looking calculation, but it avoids claiming any - * more memory than it has to. It will claim everything - * between the first address to the end of the last full - * register. */ - mapsize = ((info->io_size * info->io.regspacing) - - (info->io.regspacing - info->io.regsize)); - - if (request_region(*addr, mapsize, DEVICE_NAME) == NULL) - return -EIO; - return 0; -} - -static int try_init_port(int intf_num, struct smi_info **new_info) -{ - struct smi_info *info; - - if (!ports[intf_num]) - return -ENODEV; - - if (!is_new_interface(intf_num, IPMI_IO_ADDR_SPACE, - ports[intf_num])) - return -ENODEV; - - info = kmalloc(sizeof(*info), GFP_KERNEL); - if (!info) { - printk(KERN_ERR "ipmi_si: Could not allocate SI data (1)\n"); - return -ENOMEM; + /* + * Some BIOSes reserve disjoint I/O regions in their ACPI + * tables. This causes problems when trying to register the + * entire I/O region. Therefore we must register each I/O + * port separately. + */ + for (idx = 0; idx < info->io_size; idx++) { + if (request_region(addr + idx * info->io.regspacing, + info->io.regsize, DEVICE_NAME) == NULL) { + /* Undo allocations */ + while (idx--) { + release_region(addr + idx * info->io.regspacing, + info->io.regsize); + } + return -EIO; + } } - memset(info, 0, sizeof(*info)); - - info->io_setup = port_setup; - info->io.info = &(ports[intf_num]); - info->io.addr = NULL; - info->io.regspacing = regspacings[intf_num]; - if (!info->io.regspacing) - info->io.regspacing = DEFAULT_REGSPACING; - info->io.regsize = regsizes[intf_num]; - if (!info->io.regsize) - info->io.regsize = DEFAULT_REGSPACING; - info->io.regshift = regshifts[intf_num]; - info->irq = 0; - info->irq_setup = NULL; - *new_info = info; - - if (si_type[intf_num] == NULL) - si_type[intf_num] = "kcs"; - - printk("ipmi_si: Trying \"%s\" at I/O port 0x%x\n", - si_type[intf_num], ports[intf_num]); return 0; } -static unsigned char mem_inb(struct si_sm_io *io, unsigned int offset) +static unsigned char intf_mem_inb(struct si_sm_io *io, unsigned int offset) { return readb((io->addr)+(offset * io->regspacing)); } -static void mem_outb(struct si_sm_io *io, unsigned int offset, +static void intf_mem_outb(struct si_sm_io *io, unsigned int offset, unsigned char b) { writeb(b, (io->addr)+(offset * io->regspacing)); } -static unsigned char mem_inw(struct si_sm_io *io, unsigned int offset) +static unsigned char intf_mem_inw(struct si_sm_io *io, unsigned int offset) { return (readw((io->addr)+(offset * io->regspacing)) >> io->regshift) - && 0xff; + & 0xff; } -static void mem_outw(struct si_sm_io *io, unsigned int offset, +static void intf_mem_outw(struct si_sm_io *io, unsigned int offset, unsigned char b) { writeb(b << io->regshift, (io->addr)+(offset * io->regspacing)); } -static unsigned char mem_inl(struct si_sm_io *io, unsigned int offset) +static unsigned char intf_mem_inl(struct si_sm_io *io, unsigned int offset) { return (readl((io->addr)+(offset * io->regspacing)) >> io->regshift) - && 0xff; + & 0xff; } -static void mem_outl(struct si_sm_io *io, unsigned int offset, +static void intf_mem_outl(struct si_sm_io *io, unsigned int offset, unsigned char b) { writel(b << io->regshift, (io->addr)+(offset * io->regspacing)); @@ -1239,7 +1568,7 @@ static void mem_outl(struct si_sm_io *io, unsigned int offset, static unsigned char mem_inq(struct si_sm_io *io, unsigned int offset) { return (readq((io->addr)+(offset * io->regspacing)) >> io->regshift) - && 0xff; + & 0xff; } static void mem_outq(struct si_sm_io *io, unsigned int offset, @@ -1251,7 +1580,7 @@ static void mem_outq(struct si_sm_io *io, unsigned int offset, static void mem_cleanup(struct smi_info *info) { - unsigned long *addr = info->io.info; + unsigned long addr = info->io.addr_data; int mapsize; if (info->io.addr) { @@ -1260,35 +1589,36 @@ static void mem_cleanup(struct smi_info *info) mapsize = ((info->io_size * info->io.regspacing) - (info->io.regspacing - info->io.regsize)); - release_mem_region(*addr, mapsize); + release_mem_region(addr, mapsize); } - kfree(info); } static int mem_setup(struct smi_info *info) { - unsigned long *addr = info->io.info; + unsigned long addr = info->io.addr_data; int mapsize; - if (!addr || (!*addr)) + if (!addr) return -ENODEV; info->io_cleanup = mem_cleanup; - /* Figure out the actual readb/readw/readl/etc routine to use based - upon the register size. */ + /* + * Figure out the actual readb/readw/readl/etc routine to use based + * upon the register size. + */ switch (info->io.regsize) { case 1: - info->io.inputb = mem_inb; - info->io.outputb = mem_outb; + info->io.inputb = intf_mem_inb; + info->io.outputb = intf_mem_outb; break; case 2: - info->io.inputb = mem_inw; - info->io.outputb = mem_outw; + info->io.inputb = intf_mem_inw; + info->io.outputb = intf_mem_outw; break; case 4: - info->io.inputb = mem_inl; - info->io.outputb = mem_outl; + info->io.inputb = intf_mem_inl; + info->io.outputb = intf_mem_outl; break; #ifdef readq case 8: @@ -1297,82 +1627,382 @@ static int mem_setup(struct smi_info *info) break; #endif default: - printk("ipmi_si: Invalid register size: %d\n", - info->io.regsize); + dev_warn(info->dev, "Invalid register size: %d\n", + info->io.regsize); return -EINVAL; } - /* Calculate the total amount of memory to claim. This is an + /* + * Calculate the total amount of memory to claim. This is an * unusual looking calculation, but it avoids claiming any * more memory than it has to. It will claim everything * between the first address to the end of the last full - * register. */ + * register. + */ mapsize = ((info->io_size * info->io.regspacing) - (info->io.regspacing - info->io.regsize)); - if (request_mem_region(*addr, mapsize, DEVICE_NAME) == NULL) + if (request_mem_region(addr, mapsize, DEVICE_NAME) == NULL) return -EIO; - info->io.addr = ioremap(*addr, mapsize); + info->io.addr = ioremap(addr, mapsize); if (info->io.addr == NULL) { - release_mem_region(*addr, mapsize); + release_mem_region(addr, mapsize); return -EIO; } return 0; } -static int try_init_mem(int intf_num, struct smi_info **new_info) +/* + * Parms come in as <op1>[:op2[:op3...]]. ops are: + * add|remove,kcs|bt|smic,mem|i/o,<address>[,<opt1>[,<opt2>[,...]]] + * Options are: + * rsp=<regspacing> + * rsi=<regsize> + * rsh=<regshift> + * irq=<irq> + * ipmb=<ipmb addr> + */ +enum hotmod_op { HM_ADD, HM_REMOVE }; +struct hotmod_vals { + char *name; + int val; +}; +static struct hotmod_vals hotmod_ops[] = { + { "add", HM_ADD }, + { "remove", HM_REMOVE }, + { NULL } +}; +static struct hotmod_vals hotmod_si[] = { + { "kcs", SI_KCS }, + { "smic", SI_SMIC }, + { "bt", SI_BT }, + { NULL } +}; +static struct hotmod_vals hotmod_as[] = { + { "mem", IPMI_MEM_ADDR_SPACE }, + { "i/o", IPMI_IO_ADDR_SPACE }, + { NULL } +}; + +static int parse_str(struct hotmod_vals *v, int *val, char *name, char **curr) { - struct smi_info *info; + char *s; + int i; - if (!addrs[intf_num]) - return -ENODEV; + s = strchr(*curr, ','); + if (!s) { + printk(KERN_WARNING PFX "No hotmod %s given.\n", name); + return -EINVAL; + } + *s = '\0'; + s++; + for (i = 0; hotmod_ops[i].name; i++) { + if (strcmp(*curr, v[i].name) == 0) { + *val = v[i].val; + *curr = s; + return 0; + } + } - if (!is_new_interface(intf_num, IPMI_MEM_ADDR_SPACE, - addrs[intf_num])) - return -ENODEV; + printk(KERN_WARNING PFX "Invalid hotmod %s '%s'\n", name, *curr); + return -EINVAL; +} - info = kmalloc(sizeof(*info), GFP_KERNEL); - if (!info) { - printk(KERN_ERR "ipmi_si: Could not allocate SI data (2)\n"); - return -ENOMEM; +static int check_hotmod_int_op(const char *curr, const char *option, + const char *name, int *val) +{ + char *n; + + if (strcmp(curr, name) == 0) { + if (!option) { + printk(KERN_WARNING PFX + "No option given for '%s'\n", + curr); + return -EINVAL; + } + *val = simple_strtoul(option, &n, 0); + if ((*n != '\0') || (*option == '\0')) { + printk(KERN_WARNING PFX + "Bad option given for '%s'\n", + curr); + return -EINVAL; + } + return 1; } - memset(info, 0, sizeof(*info)); + return 0; +} - info->io_setup = mem_setup; - info->io.info = &addrs[intf_num]; - info->io.addr = NULL; - info->io.regspacing = regspacings[intf_num]; - if (!info->io.regspacing) - info->io.regspacing = DEFAULT_REGSPACING; - info->io.regsize = regsizes[intf_num]; - if (!info->io.regsize) - info->io.regsize = DEFAULT_REGSPACING; - info->io.regshift = regshifts[intf_num]; - info->irq = 0; - info->irq_setup = NULL; - *new_info = info; +static struct smi_info *smi_info_alloc(void) +{ + struct smi_info *info = kzalloc(sizeof(*info), GFP_KERNEL); - if (si_type[intf_num] == NULL) - si_type[intf_num] = "kcs"; + if (info) + spin_lock_init(&info->si_lock); + return info; +} - printk("ipmi_si: Trying \"%s\" at memory address 0x%lx\n", - si_type[intf_num], addrs[intf_num]); - return 0; +static int hotmod_handler(const char *val, struct kernel_param *kp) +{ + char *str = kstrdup(val, GFP_KERNEL); + int rv; + char *next, *curr, *s, *n, *o; + enum hotmod_op op; + enum si_type si_type; + int addr_space; + unsigned long addr; + int regspacing; + int regsize; + int regshift; + int irq; + int ipmb; + int ival; + int len; + struct smi_info *info; + + if (!str) + return -ENOMEM; + + /* Kill any trailing spaces, as we can get a "\n" from echo. */ + len = strlen(str); + ival = len - 1; + while ((ival >= 0) && isspace(str[ival])) { + str[ival] = '\0'; + ival--; + } + + for (curr = str; curr; curr = next) { + regspacing = 1; + regsize = 1; + regshift = 0; + irq = 0; + ipmb = 0; /* Choose the default if not specified */ + + next = strchr(curr, ':'); + if (next) { + *next = '\0'; + next++; + } + + rv = parse_str(hotmod_ops, &ival, "operation", &curr); + if (rv) + break; + op = ival; + + rv = parse_str(hotmod_si, &ival, "interface type", &curr); + if (rv) + break; + si_type = ival; + + rv = parse_str(hotmod_as, &addr_space, "address space", &curr); + if (rv) + break; + + s = strchr(curr, ','); + if (s) { + *s = '\0'; + s++; + } + addr = simple_strtoul(curr, &n, 0); + if ((*n != '\0') || (*curr == '\0')) { + printk(KERN_WARNING PFX "Invalid hotmod address" + " '%s'\n", curr); + break; + } + + while (s) { + curr = s; + s = strchr(curr, ','); + if (s) { + *s = '\0'; + s++; + } + o = strchr(curr, '='); + if (o) { + *o = '\0'; + o++; + } + rv = check_hotmod_int_op(curr, o, "rsp", ®spacing); + if (rv < 0) + goto out; + else if (rv) + continue; + rv = check_hotmod_int_op(curr, o, "rsi", ®size); + if (rv < 0) + goto out; + else if (rv) + continue; + rv = check_hotmod_int_op(curr, o, "rsh", ®shift); + if (rv < 0) + goto out; + else if (rv) + continue; + rv = check_hotmod_int_op(curr, o, "irq", &irq); + if (rv < 0) + goto out; + else if (rv) + continue; + rv = check_hotmod_int_op(curr, o, "ipmb", &ipmb); + if (rv < 0) + goto out; + else if (rv) + continue; + + rv = -EINVAL; + printk(KERN_WARNING PFX + "Invalid hotmod option '%s'\n", + curr); + goto out; + } + + if (op == HM_ADD) { + info = smi_info_alloc(); + if (!info) { + rv = -ENOMEM; + goto out; + } + + info->addr_source = SI_HOTMOD; + info->si_type = si_type; + info->io.addr_data = addr; + info->io.addr_type = addr_space; + if (addr_space == IPMI_MEM_ADDR_SPACE) + info->io_setup = mem_setup; + else + info->io_setup = port_setup; + + info->io.addr = NULL; + info->io.regspacing = regspacing; + if (!info->io.regspacing) + info->io.regspacing = DEFAULT_REGSPACING; + info->io.regsize = regsize; + if (!info->io.regsize) + info->io.regsize = DEFAULT_REGSPACING; + info->io.regshift = regshift; + info->irq = irq; + if (info->irq) + info->irq_setup = std_irq_setup; + info->slave_addr = ipmb; + + rv = add_smi(info); + if (rv) { + kfree(info); + goto out; + } + rv = try_smi_init(info); + if (rv) { + cleanup_one_si(info); + goto out; + } + } else { + /* remove */ + struct smi_info *e, *tmp_e; + + mutex_lock(&smi_infos_lock); + list_for_each_entry_safe(e, tmp_e, &smi_infos, link) { + if (e->io.addr_type != addr_space) + continue; + if (e->si_type != si_type) + continue; + if (e->io.addr_data == addr) + cleanup_one_si(e); + } + mutex_unlock(&smi_infos_lock); + } + } + rv = len; + out: + kfree(str); + return rv; } +static int hardcode_find_bmc(void) +{ + int ret = -ENODEV; + int i; + struct smi_info *info; + + for (i = 0; i < SI_MAX_PARMS; i++) { + if (!ports[i] && !addrs[i]) + continue; + + info = smi_info_alloc(); + if (!info) + return -ENOMEM; + + info->addr_source = SI_HARDCODED; + printk(KERN_INFO PFX "probing via hardcoded address\n"); + + if (!si_type[i] || strcmp(si_type[i], "kcs") == 0) { + info->si_type = SI_KCS; + } else if (strcmp(si_type[i], "smic") == 0) { + info->si_type = SI_SMIC; + } else if (strcmp(si_type[i], "bt") == 0) { + info->si_type = SI_BT; + } else { + printk(KERN_WARNING PFX "Interface type specified " + "for interface %d, was invalid: %s\n", + i, si_type[i]); + kfree(info); + continue; + } + + if (ports[i]) { + /* An I/O port */ + info->io_setup = port_setup; + info->io.addr_data = ports[i]; + info->io.addr_type = IPMI_IO_ADDR_SPACE; + } else if (addrs[i]) { + /* A memory port */ + info->io_setup = mem_setup; + info->io.addr_data = addrs[i]; + info->io.addr_type = IPMI_MEM_ADDR_SPACE; + } else { + printk(KERN_WARNING PFX "Interface type specified " + "for interface %d, but port and address were " + "not set or set to zero.\n", i); + kfree(info); + continue; + } + + info->io.addr = NULL; + info->io.regspacing = regspacings[i]; + if (!info->io.regspacing) + info->io.regspacing = DEFAULT_REGSPACING; + info->io.regsize = regsizes[i]; + if (!info->io.regsize) + info->io.regsize = DEFAULT_REGSPACING; + info->io.regshift = regshifts[i]; + info->irq = irqs[i]; + if (info->irq) + info->irq_setup = std_irq_setup; + info->slave_addr = slave_addrs[i]; + + if (!add_smi(info)) { + if (try_smi_init(info)) + cleanup_one_si(info); + ret = 0; + } else { + kfree(info); + } + } + return ret; +} -#ifdef CONFIG_ACPI_INTERPRETER +#ifdef CONFIG_ACPI #include <linux/acpi.h> -/* Once we get an ACPI failure, we don't try any more, because we go - through the tables sequentially. Once we don't find a table, there - are no more. */ -static int acpi_failure = 0; +/* + * Once we get an ACPI failure, we don't try any more, because we go + * through the tables sequentially. Once we don't find a table, there + * are no more. + */ +static int acpi_failure; /* For GPE-type interrupts. */ -static u32 ipmi_acpi_gpe(void *context) +static u32 ipmi_acpi_gpe(acpi_handle gpe_device, + u32 gpe_number, void *context) { struct smi_info *smi_info = context; unsigned long flags; @@ -1382,24 +2012,26 @@ static u32 ipmi_acpi_gpe(void *context) spin_lock_irqsave(&(smi_info->si_lock), flags); - spin_lock(&smi_info->count_lock); - smi_info->interrupts++; - spin_unlock(&smi_info->count_lock); - - if (smi_info->stop_operation) - goto out; + smi_inc_stat(smi_info, interrupts); #ifdef DEBUG_TIMING do_gettimeofday(&t); printk("**ACPI_GPE: %d.%9.9d\n", t.tv_sec, t.tv_usec); #endif smi_event_handler(smi_info, 0); - out: spin_unlock_irqrestore(&(smi_info->si_lock), flags); return ACPI_INTERRUPT_HANDLED; } +static void acpi_gpe_irq_cleanup(struct smi_info *info) +{ + if (!info->irq) + return; + + acpi_remove_gpe_handler(NULL, info->irq, &ipmi_acpi_gpe); +} + static int acpi_gpe_irq_setup(struct smi_info *info) { acpi_status status; @@ -1414,29 +2046,20 @@ static int acpi_gpe_irq_setup(struct smi_info *info) &ipmi_acpi_gpe, info); if (status != AE_OK) { - printk(KERN_WARNING - "ipmi_si: %s unable to claim ACPI GPE %d," - " running polled\n", - DEVICE_NAME, info->irq); + dev_warn(info->dev, "%s unable to claim ACPI GPE %d," + " running polled\n", DEVICE_NAME, info->irq); info->irq = 0; return -EINVAL; } else { - printk(" Using ACPI GPE %d\n", info->irq); + info->irq_cleanup = acpi_gpe_irq_cleanup; + dev_info(info->dev, "Using ACPI GPE %d\n", info->irq); return 0; } } -static void acpi_gpe_irq_cleanup(struct smi_info *info) -{ - if (!info->irq) - return; - - acpi_remove_gpe_handler(NULL, info->irq, &ipmi_acpi_gpe); -} - /* * Defined at - * http://h21007.www2.hp.com/dspp/files/unprotected/devresource/Docs/TechPapers/IA64/hpspmi.pdf + * http://h21007.www2.hp.com/portal/download/files/unprot/hpspmi.pdf */ struct SPMITable { s8 Signature[4]; @@ -1458,14 +2081,18 @@ struct SPMITable { */ u8 InterruptType; - /* If bit 0 of InterruptType is set, then this is the SCI - interrupt in the GPEx_STS register. */ + /* + * If bit 0 of InterruptType is set, then this is the SCI + * interrupt in the GPEx_STS register. + */ u8 GPE; s16 Reserved; - /* If bit 1 of InterruptType is set, then this is the I/O - APIC/SAPIC interrupt. */ + /* + * If bit 1 of InterruptType is set, then this is the I/O + * APIC/SAPIC interrupt. + */ u32 GlobalSystemInterrupt; /* The actual register address. */ @@ -1476,185 +2103,284 @@ struct SPMITable { s8 spmi_id[1]; /* A '\0' terminated array starts here. */ }; -static int try_init_acpi(int intf_num, struct smi_info **new_info) +static int try_init_spmi(struct SPMITable *spmi) { struct smi_info *info; - acpi_status status; - struct SPMITable *spmi; - char *io_type; - u8 addr_space; - - if (acpi_failure) - return -ENODEV; + int rv; - status = acpi_get_firmware_table("SPMI", intf_num+1, - ACPI_LOGICAL_ADDRESSING, - (struct acpi_table_header **) &spmi); - if (status != AE_OK) { - acpi_failure = 1; + if (spmi->IPMIlegacy != 1) { + printk(KERN_INFO PFX "Bad SPMI legacy %d\n", spmi->IPMIlegacy); return -ENODEV; } - if (spmi->IPMIlegacy != 1) { - printk(KERN_INFO "IPMI: Bad SPMI legacy %d\n", spmi->IPMIlegacy); - return -ENODEV; + info = smi_info_alloc(); + if (!info) { + printk(KERN_ERR PFX "Could not allocate SI data (3)\n"); + return -ENOMEM; } - if (spmi->addr.address_space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) - addr_space = IPMI_MEM_ADDR_SPACE; - else - addr_space = IPMI_IO_ADDR_SPACE; - if (!is_new_interface(-1, addr_space, spmi->addr.address)) - return -ENODEV; - - if (!spmi->addr.register_bit_width) { - acpi_failure = 1; - return -ENODEV; - } + info->addr_source = SI_SPMI; + printk(KERN_INFO PFX "probing via SPMI\n"); /* Figure out the interface type. */ - switch (spmi->InterfaceType) - { + switch (spmi->InterfaceType) { case 1: /* KCS */ - si_type[intf_num] = "kcs"; + info->si_type = SI_KCS; break; - case 2: /* SMIC */ - si_type[intf_num] = "smic"; + info->si_type = SI_SMIC; break; - case 3: /* BT */ - si_type[intf_num] = "bt"; + info->si_type = SI_BT; break; - default: - printk(KERN_INFO "ipmi_si: Unknown ACPI/SPMI SI type %d\n", - spmi->InterfaceType); + printk(KERN_INFO PFX "Unknown ACPI/SPMI SI type %d\n", + spmi->InterfaceType); + kfree(info); return -EIO; } - info = kmalloc(sizeof(*info), GFP_KERNEL); - if (!info) { - printk(KERN_ERR "ipmi_si: Could not allocate SI data (3)\n"); - return -ENOMEM; - } - memset(info, 0, sizeof(*info)); - if (spmi->InterruptType & 1) { /* We've got a GPE interrupt. */ info->irq = spmi->GPE; info->irq_setup = acpi_gpe_irq_setup; - info->irq_cleanup = acpi_gpe_irq_cleanup; } else if (spmi->InterruptType & 2) { /* We've got an APIC/SAPIC interrupt. */ info->irq = spmi->GlobalSystemInterrupt; info->irq_setup = std_irq_setup; - info->irq_cleanup = std_irq_cleanup; } else { /* Use the default interrupt setting. */ info->irq = 0; info->irq_setup = NULL; } - if (spmi->addr.register_bit_width) { + if (spmi->addr.bit_width) { /* A (hopefully) properly formed register bit width. */ - regspacings[intf_num] = spmi->addr.register_bit_width / 8; - info->io.regspacing = spmi->addr.register_bit_width / 8; + info->io.regspacing = spmi->addr.bit_width / 8; } else { - /* Some broken systems get this wrong and set the value - * to zero. Assume it is the default spacing. If that - * is wrong, too bad, the vendor should fix the tables. */ - regspacings[intf_num] = DEFAULT_REGSPACING; info->io.regspacing = DEFAULT_REGSPACING; } - regsizes[intf_num] = regspacings[intf_num]; - info->io.regsize = regsizes[intf_num]; - regshifts[intf_num] = spmi->addr.register_bit_offset; - info->io.regshift = regshifts[intf_num]; + info->io.regsize = info->io.regspacing; + info->io.regshift = spmi->addr.bit_offset; - if (spmi->addr.address_space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) { - io_type = "memory"; + if (spmi->addr.space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) { info->io_setup = mem_setup; - addrs[intf_num] = spmi->addr.address; - info->io.info = &(addrs[intf_num]); - } else if (spmi->addr.address_space_id == ACPI_ADR_SPACE_SYSTEM_IO) { - io_type = "I/O"; + info->io.addr_type = IPMI_MEM_ADDR_SPACE; + } else if (spmi->addr.space_id == ACPI_ADR_SPACE_SYSTEM_IO) { info->io_setup = port_setup; - ports[intf_num] = spmi->addr.address; - info->io.info = &(ports[intf_num]); + info->io.addr_type = IPMI_IO_ADDR_SPACE; } else { kfree(info); - printk("ipmi_si: Unknown ACPI I/O Address type\n"); + printk(KERN_WARNING PFX "Unknown ACPI I/O Address type\n"); return -EIO; } + info->io.addr_data = spmi->addr.address; - *new_info = info; + pr_info("ipmi_si: SPMI: %s %#lx regsize %d spacing %d irq %d\n", + (info->io.addr_type == IPMI_IO_ADDR_SPACE) ? "io" : "mem", + info->io.addr_data, info->io.regsize, info->io.regspacing, + info->irq); - printk("ipmi_si: ACPI/SPMI specifies \"%s\" %s SI @ 0x%lx\n", - si_type[intf_num], io_type, (unsigned long) spmi->addr.address); - return 0; + rv = add_smi(info); + if (rv) + kfree(info); + + return rv; } -#endif -#ifdef CONFIG_X86 -typedef struct dmi_ipmi_data +static void spmi_find_bmc(void) { + acpi_status status; + struct SPMITable *spmi; + int i; + + if (acpi_disabled) + return; + + if (acpi_failure) + return; + + for (i = 0; ; i++) { + status = acpi_get_table(ACPI_SIG_SPMI, i+1, + (struct acpi_table_header **)&spmi); + if (status != AE_OK) + return; + + try_init_spmi(spmi); + } +} + +static int ipmi_pnp_probe(struct pnp_dev *dev, + const struct pnp_device_id *dev_id) +{ + struct acpi_device *acpi_dev; + struct smi_info *info; + struct resource *res, *res_second; + acpi_handle handle; + acpi_status status; + unsigned long long tmp; + int rv; + + acpi_dev = pnp_acpi_device(dev); + if (!acpi_dev) + return -ENODEV; + + info = smi_info_alloc(); + if (!info) + return -ENOMEM; + + info->addr_source = SI_ACPI; + printk(KERN_INFO PFX "probing via ACPI\n"); + + handle = acpi_dev->handle; + info->addr_info.acpi_info.acpi_handle = handle; + + /* _IFT tells us the interface type: KCS, BT, etc */ + status = acpi_evaluate_integer(handle, "_IFT", NULL, &tmp); + if (ACPI_FAILURE(status)) + goto err_free; + + switch (tmp) { + case 1: + info->si_type = SI_KCS; + break; + case 2: + info->si_type = SI_SMIC; + break; + case 3: + info->si_type = SI_BT; + break; + default: + dev_info(&dev->dev, "unknown IPMI type %lld\n", tmp); + goto err_free; + } + + res = pnp_get_resource(dev, IORESOURCE_IO, 0); + if (res) { + info->io_setup = port_setup; + info->io.addr_type = IPMI_IO_ADDR_SPACE; + } else { + res = pnp_get_resource(dev, IORESOURCE_MEM, 0); + if (res) { + info->io_setup = mem_setup; + info->io.addr_type = IPMI_MEM_ADDR_SPACE; + } + } + if (!res) { + dev_err(&dev->dev, "no I/O or memory address\n"); + goto err_free; + } + info->io.addr_data = res->start; + + info->io.regspacing = DEFAULT_REGSPACING; + res_second = pnp_get_resource(dev, + (info->io.addr_type == IPMI_IO_ADDR_SPACE) ? + IORESOURCE_IO : IORESOURCE_MEM, + 1); + if (res_second) { + if (res_second->start > info->io.addr_data) + info->io.regspacing = res_second->start - info->io.addr_data; + } + info->io.regsize = DEFAULT_REGSPACING; + info->io.regshift = 0; + + /* If _GPE exists, use it; otherwise use standard interrupts */ + status = acpi_evaluate_integer(handle, "_GPE", NULL, &tmp); + if (ACPI_SUCCESS(status)) { + info->irq = tmp; + info->irq_setup = acpi_gpe_irq_setup; + } else if (pnp_irq_valid(dev, 0)) { + info->irq = pnp_irq(dev, 0); + info->irq_setup = std_irq_setup; + } + + info->dev = &dev->dev; + pnp_set_drvdata(dev, info); + + dev_info(info->dev, "%pR regsize %d spacing %d irq %d\n", + res, info->io.regsize, info->io.regspacing, + info->irq); + + rv = add_smi(info); + if (rv) + kfree(info); + + return rv; + +err_free: + kfree(info); + return -EINVAL; +} + +static void ipmi_pnp_remove(struct pnp_dev *dev) +{ + struct smi_info *info = pnp_get_drvdata(dev); + + cleanup_one_si(info); +} + +static const struct pnp_device_id pnp_dev_table[] = { + {"IPI0001", 0}, + {"", 0}, +}; + +static struct pnp_driver ipmi_pnp_driver = { + .name = DEVICE_NAME, + .probe = ipmi_pnp_probe, + .remove = ipmi_pnp_remove, + .id_table = pnp_dev_table, +}; + +MODULE_DEVICE_TABLE(pnp, pnp_dev_table); +#endif + +#ifdef CONFIG_DMI +struct dmi_ipmi_data { u8 type; u8 addr_space; unsigned long base_addr; u8 irq; u8 offset; u8 slave_addr; -} dmi_ipmi_data_t; - -static dmi_ipmi_data_t dmi_data[SI_MAX_DRIVERS]; -static int dmi_data_entries; - -typedef struct dmi_header -{ - u8 type; - u8 length; - u16 handle; -} dmi_header_t; +}; -static int decode_dmi(dmi_header_t __iomem *dm, int intf_num) +static int decode_dmi(const struct dmi_header *dm, + struct dmi_ipmi_data *dmi) { - u8 __iomem *data = (u8 __iomem *)dm; + const u8 *data = (const u8 *)dm; unsigned long base_addr; u8 reg_spacing; - u8 len = readb(&dm->length); - dmi_ipmi_data_t *ipmi_data = dmi_data+intf_num; + u8 len = dm->length; - ipmi_data->type = readb(&data[4]); + dmi->type = data[4]; memcpy(&base_addr, data+8, sizeof(unsigned long)); if (len >= 0x11) { if (base_addr & 1) { /* I/O */ base_addr &= 0xFFFE; - ipmi_data->addr_space = IPMI_IO_ADDR_SPACE; - } - else { + dmi->addr_space = IPMI_IO_ADDR_SPACE; + } else /* Memory */ - ipmi_data->addr_space = IPMI_MEM_ADDR_SPACE; - } + dmi->addr_space = IPMI_MEM_ADDR_SPACE; + /* If bit 4 of byte 0x10 is set, then the lsb for the address is odd. */ - ipmi_data->base_addr = base_addr | ((readb(&data[0x10]) & 0x10) >> 4); + dmi->base_addr = base_addr | ((data[0x10] & 0x10) >> 4); - ipmi_data->irq = readb(&data[0x11]); + dmi->irq = data[0x11]; /* The top two bits of byte 0x10 hold the register spacing. */ - reg_spacing = (readb(&data[0x10]) & 0xC0) >> 6; - switch(reg_spacing){ + reg_spacing = (data[0x10] & 0xC0) >> 6; + switch (reg_spacing) { case 0x00: /* Byte boundaries */ - ipmi_data->offset = 1; + dmi->offset = 1; break; case 0x01: /* 32-bit boundaries */ - ipmi_data->offset = 4; + dmi->offset = 4; break; case 0x02: /* 16-byte boundaries */ - ipmi_data->offset = 16; + dmi->offset = 16; break; default: /* Some other interface, just ignore it. */ @@ -1662,461 +2388,940 @@ static int decode_dmi(dmi_header_t __iomem *dm, int intf_num) } } else { /* Old DMI spec. */ - /* Note that technically, the lower bit of the base + /* + * Note that technically, the lower bit of the base * address should be 1 if the address is I/O and 0 if * the address is in memory. So many systems get that * wrong (and all that I have seen are I/O) so we just * ignore that bit and assume I/O. Systems that use - * memory should use the newer spec, anyway. */ - ipmi_data->base_addr = base_addr & 0xfffe; - ipmi_data->addr_space = IPMI_IO_ADDR_SPACE; - ipmi_data->offset = 1; - } - - ipmi_data->slave_addr = readb(&data[6]); - - if (is_new_interface(-1, ipmi_data->addr_space,ipmi_data->base_addr)) { - dmi_data_entries++; - return 0; + * memory should use the newer spec, anyway. + */ + dmi->base_addr = base_addr & 0xfffe; + dmi->addr_space = IPMI_IO_ADDR_SPACE; + dmi->offset = 1; } - memset(ipmi_data, 0, sizeof(dmi_ipmi_data_t)); + dmi->slave_addr = data[6]; - return -1; + return 0; } -static int dmi_table(u32 base, int len, int num) +static void try_init_dmi(struct dmi_ipmi_data *ipmi_data) { - u8 __iomem *buf; - struct dmi_header __iomem *dm; - u8 __iomem *data; - int i=1; - int status=-1; - int intf_num = 0; + struct smi_info *info; - buf = ioremap(base, len); - if(buf==NULL) - return -1; + info = smi_info_alloc(); + if (!info) { + printk(KERN_ERR PFX "Could not allocate SI data\n"); + return; + } - data = buf; + info->addr_source = SI_SMBIOS; + printk(KERN_INFO PFX "probing via SMBIOS\n"); - while(i<num && (data - buf) < len) - { - dm=(dmi_header_t __iomem *)data; + switch (ipmi_data->type) { + case 0x01: /* KCS */ + info->si_type = SI_KCS; + break; + case 0x02: /* SMIC */ + info->si_type = SI_SMIC; + break; + case 0x03: /* BT */ + info->si_type = SI_BT; + break; + default: + kfree(info); + return; + } - if((data-buf+readb(&dm->length)) >= len) - break; + switch (ipmi_data->addr_space) { + case IPMI_MEM_ADDR_SPACE: + info->io_setup = mem_setup; + info->io.addr_type = IPMI_MEM_ADDR_SPACE; + break; - if (readb(&dm->type) == 38) { - if (decode_dmi(dm, intf_num) == 0) { - intf_num++; - if (intf_num >= SI_MAX_DRIVERS) - break; - } - } + case IPMI_IO_ADDR_SPACE: + info->io_setup = port_setup; + info->io.addr_type = IPMI_IO_ADDR_SPACE; + break; - data+=readb(&dm->length); - while((data-buf) < len && (readb(data)||readb(data+1))) - data++; - data+=2; - i++; + default: + kfree(info); + printk(KERN_WARNING PFX "Unknown SMBIOS I/O Address type: %d\n", + ipmi_data->addr_space); + return; } - iounmap(buf); + info->io.addr_data = ipmi_data->base_addr; + + info->io.regspacing = ipmi_data->offset; + if (!info->io.regspacing) + info->io.regspacing = DEFAULT_REGSPACING; + info->io.regsize = DEFAULT_REGSPACING; + info->io.regshift = 0; + + info->slave_addr = ipmi_data->slave_addr; + + info->irq = ipmi_data->irq; + if (info->irq) + info->irq_setup = std_irq_setup; + + pr_info("ipmi_si: SMBIOS: %s %#lx regsize %d spacing %d irq %d\n", + (info->io.addr_type == IPMI_IO_ADDR_SPACE) ? "io" : "mem", + info->io.addr_data, info->io.regsize, info->io.regspacing, + info->irq); + + if (add_smi(info)) + kfree(info); +} - return status; +static void dmi_find_bmc(void) +{ + const struct dmi_device *dev = NULL; + struct dmi_ipmi_data data; + int rv; + + while ((dev = dmi_find_device(DMI_DEV_TYPE_IPMI, NULL, dev))) { + memset(&data, 0, sizeof(data)); + rv = decode_dmi((const struct dmi_header *) dev->device_data, + &data); + if (!rv) + try_init_dmi(&data); + } } +#endif /* CONFIG_DMI */ + +#ifdef CONFIG_PCI + +#define PCI_ERMC_CLASSCODE 0x0C0700 +#define PCI_ERMC_CLASSCODE_MASK 0xffffff00 +#define PCI_ERMC_CLASSCODE_TYPE_MASK 0xff +#define PCI_ERMC_CLASSCODE_TYPE_SMIC 0x00 +#define PCI_ERMC_CLASSCODE_TYPE_KCS 0x01 +#define PCI_ERMC_CLASSCODE_TYPE_BT 0x02 + +#define PCI_HP_VENDOR_ID 0x103C +#define PCI_MMC_DEVICE_ID 0x121A +#define PCI_MMC_ADDR_CW 0x10 -static inline int dmi_checksum(u8 *buf) +static void ipmi_pci_cleanup(struct smi_info *info) { - u8 sum=0; - int a; + struct pci_dev *pdev = info->addr_source_data; - for(a=0; a<15; a++) - sum+=buf[a]; - return (sum==0); + pci_disable_device(pdev); } -static int dmi_decode(void) +static int ipmi_pci_probe_regspacing(struct smi_info *info) { - u8 buf[15]; - u32 fp=0xF0000; + if (info->si_type == SI_KCS) { + unsigned char status; + int regspacing; -#ifdef CONFIG_SIMNOW - return -1; -#endif + info->io.regsize = DEFAULT_REGSIZE; + info->io.regshift = 0; + info->io_size = 2; + info->handlers = &kcs_smi_handlers; - while(fp < 0xFFFFF) - { - isa_memcpy_fromio(buf, fp, 15); - if(memcmp(buf, "_DMI_", 5)==0 && dmi_checksum(buf)) - { - u16 num=buf[13]<<8|buf[12]; - u16 len=buf[7]<<8|buf[6]; - u32 base=buf[11]<<24|buf[10]<<16|buf[9]<<8|buf[8]; - - if(dmi_table(base, len, num) == 0) - return 0; + /* detect 1, 4, 16byte spacing */ + for (regspacing = DEFAULT_REGSPACING; regspacing <= 16;) { + info->io.regspacing = regspacing; + if (info->io_setup(info)) { + dev_err(info->dev, + "Could not setup I/O space\n"); + return DEFAULT_REGSPACING; + } + /* write invalid cmd */ + info->io.outputb(&info->io, 1, 0x10); + /* read status back */ + status = info->io.inputb(&info->io, 1); + info->io_cleanup(info); + if (status) + return regspacing; + regspacing *= 4; } - fp+=16; } - - return -1; + return DEFAULT_REGSPACING; } -static int try_init_smbios(int intf_num, struct smi_info **new_info) +static int ipmi_pci_probe(struct pci_dev *pdev, + const struct pci_device_id *ent) { - struct smi_info *info; - dmi_ipmi_data_t *ipmi_data = dmi_data+intf_num; - char *io_type; + int rv; + int class_type = pdev->class & PCI_ERMC_CLASSCODE_TYPE_MASK; + struct smi_info *info; - if (intf_num >= dmi_data_entries) - return -ENODEV; + info = smi_info_alloc(); + if (!info) + return -ENOMEM; - switch(ipmi_data->type) { - case 0x01: /* KCS */ - si_type[intf_num] = "kcs"; - break; - case 0x02: /* SMIC */ - si_type[intf_num] = "smic"; - break; - case 0x03: /* BT */ - si_type[intf_num] = "bt"; - break; - default: - return -EIO; - } + info->addr_source = SI_PCI; + dev_info(&pdev->dev, "probing via PCI"); - info = kmalloc(sizeof(*info), GFP_KERNEL); - if (!info) { - printk(KERN_ERR "ipmi_si: Could not allocate SI data (4)\n"); + switch (class_type) { + case PCI_ERMC_CLASSCODE_TYPE_SMIC: + info->si_type = SI_SMIC; + break; + + case PCI_ERMC_CLASSCODE_TYPE_KCS: + info->si_type = SI_KCS; + break; + + case PCI_ERMC_CLASSCODE_TYPE_BT: + info->si_type = SI_BT; + break; + + default: + kfree(info); + dev_info(&pdev->dev, "Unknown IPMI type: %d\n", class_type); return -ENOMEM; } - memset(info, 0, sizeof(*info)); - if (ipmi_data->addr_space == 1) { - io_type = "memory"; - info->io_setup = mem_setup; - addrs[intf_num] = ipmi_data->base_addr; - info->io.info = &(addrs[intf_num]); - } else if (ipmi_data->addr_space == 2) { - io_type = "I/O"; + rv = pci_enable_device(pdev); + if (rv) { + dev_err(&pdev->dev, "couldn't enable PCI device\n"); + kfree(info); + return rv; + } + + info->addr_source_cleanup = ipmi_pci_cleanup; + info->addr_source_data = pdev; + + if (pci_resource_flags(pdev, 0) & IORESOURCE_IO) { info->io_setup = port_setup; - ports[intf_num] = ipmi_data->base_addr; - info->io.info = &(ports[intf_num]); + info->io.addr_type = IPMI_IO_ADDR_SPACE; } else { - kfree(info); - printk("ipmi_si: Unknown SMBIOS I/O Address type.\n"); - return -EIO; + info->io_setup = mem_setup; + info->io.addr_type = IPMI_MEM_ADDR_SPACE; } + info->io.addr_data = pci_resource_start(pdev, 0); - regspacings[intf_num] = ipmi_data->offset; - info->io.regspacing = regspacings[intf_num]; - if (!info->io.regspacing) - info->io.regspacing = DEFAULT_REGSPACING; - info->io.regsize = DEFAULT_REGSPACING; - info->io.regshift = regshifts[intf_num]; + info->io.regspacing = ipmi_pci_probe_regspacing(info); + info->io.regsize = DEFAULT_REGSIZE; + info->io.regshift = 0; - info->slave_addr = ipmi_data->slave_addr; + info->irq = pdev->irq; + if (info->irq) + info->irq_setup = std_irq_setup; - irqs[intf_num] = ipmi_data->irq; + info->dev = &pdev->dev; + pci_set_drvdata(pdev, info); - *new_info = info; + dev_info(&pdev->dev, "%pR regsize %d spacing %d irq %d\n", + &pdev->resource[0], info->io.regsize, info->io.regspacing, + info->irq); - printk("ipmi_si: Found SMBIOS-specified state machine at %s" - " address 0x%lx, slave address 0x%x\n", - io_type, (unsigned long)ipmi_data->base_addr, - ipmi_data->slave_addr); - return 0; + rv = add_smi(info); + if (rv) { + kfree(info); + pci_disable_device(pdev); + } + + return rv; } -#endif /* CONFIG_X86 */ -#ifdef CONFIG_PCI +static void ipmi_pci_remove(struct pci_dev *pdev) +{ + struct smi_info *info = pci_get_drvdata(pdev); + cleanup_one_si(info); + pci_disable_device(pdev); +} -#define PCI_ERMC_CLASSCODE 0x0C0700 -#define PCI_HP_VENDOR_ID 0x103C -#define PCI_MMC_DEVICE_ID 0x121A -#define PCI_MMC_ADDR_CW 0x10 +static struct pci_device_id ipmi_pci_devices[] = { + { PCI_DEVICE(PCI_HP_VENDOR_ID, PCI_MMC_DEVICE_ID) }, + { PCI_DEVICE_CLASS(PCI_ERMC_CLASSCODE, PCI_ERMC_CLASSCODE_MASK) }, + { 0, } +}; +MODULE_DEVICE_TABLE(pci, ipmi_pci_devices); -/* Avoid more than one attempt to probe pci smic. */ -static int pci_smic_checked = 0; +static struct pci_driver ipmi_pci_driver = { + .name = DEVICE_NAME, + .id_table = ipmi_pci_devices, + .probe = ipmi_pci_probe, + .remove = ipmi_pci_remove, +}; +#endif /* CONFIG_PCI */ -static int find_pci_smic(int intf_num, struct smi_info **new_info) +static struct of_device_id ipmi_match[]; +static int ipmi_probe(struct platform_device *dev) { - struct smi_info *info; - int error; - struct pci_dev *pci_dev = NULL; - u16 base_addr; - int fe_rmc = 0; - - if (pci_smic_checked) - return -ENODEV; +#ifdef CONFIG_OF + const struct of_device_id *match; + struct smi_info *info; + struct resource resource; + const __be32 *regsize, *regspacing, *regshift; + struct device_node *np = dev->dev.of_node; + int ret; + int proplen; - pci_smic_checked = 1; + dev_info(&dev->dev, "probing via device tree\n"); - if ((pci_dev = pci_get_device(PCI_HP_VENDOR_ID, PCI_MMC_DEVICE_ID, - NULL))) - ; - else if ((pci_dev = pci_get_class(PCI_ERMC_CLASSCODE, NULL)) && - pci_dev->subsystem_vendor == PCI_HP_VENDOR_ID) - fe_rmc = 1; - else - return -ENODEV; + match = of_match_device(ipmi_match, &dev->dev); + if (!match) + return -EINVAL; - error = pci_read_config_word(pci_dev, PCI_MMC_ADDR_CW, &base_addr); - if (error) - { - pci_dev_put(pci_dev); - printk(KERN_ERR - "ipmi_si: pci_read_config_word() failed (%d).\n", - error); - return -ENODEV; + ret = of_address_to_resource(np, 0, &resource); + if (ret) { + dev_warn(&dev->dev, PFX "invalid address from OF\n"); + return ret; } - /* Bit 0: 1 specifies programmed I/O, 0 specifies memory mapped I/O */ - if (!(base_addr & 0x0001)) - { - pci_dev_put(pci_dev); - printk(KERN_ERR - "ipmi_si: memory mapped I/O not supported for PCI" - " smic.\n"); - return -ENODEV; + regsize = of_get_property(np, "reg-size", &proplen); + if (regsize && proplen != 4) { + dev_warn(&dev->dev, PFX "invalid regsize from OF\n"); + return -EINVAL; } - base_addr &= 0xFFFE; - if (!fe_rmc) - /* Data register starts at base address + 1 in eRMC */ - ++base_addr; + regspacing = of_get_property(np, "reg-spacing", &proplen); + if (regspacing && proplen != 4) { + dev_warn(&dev->dev, PFX "invalid regspacing from OF\n"); + return -EINVAL; + } - if (!is_new_interface(-1, IPMI_IO_ADDR_SPACE, base_addr)) { - pci_dev_put(pci_dev); - return -ENODEV; + regshift = of_get_property(np, "reg-shift", &proplen); + if (regshift && proplen != 4) { + dev_warn(&dev->dev, PFX "invalid regshift from OF\n"); + return -EINVAL; } - info = kmalloc(sizeof(*info), GFP_KERNEL); + info = smi_info_alloc(); + if (!info) { - pci_dev_put(pci_dev); - printk(KERN_ERR "ipmi_si: Could not allocate SI data (5)\n"); + dev_err(&dev->dev, + "could not allocate memory for OF probe\n"); return -ENOMEM; } - memset(info, 0, sizeof(*info)); - info->io_setup = port_setup; - ports[intf_num] = base_addr; - info->io.info = &(ports[intf_num]); - info->io.regspacing = regspacings[intf_num]; - if (!info->io.regspacing) - info->io.regspacing = DEFAULT_REGSPACING; - info->io.regsize = DEFAULT_REGSPACING; - info->io.regshift = regshifts[intf_num]; + info->si_type = (enum si_type) match->data; + info->addr_source = SI_DEVICETREE; + info->irq_setup = std_irq_setup; + + if (resource.flags & IORESOURCE_IO) { + info->io_setup = port_setup; + info->io.addr_type = IPMI_IO_ADDR_SPACE; + } else { + info->io_setup = mem_setup; + info->io.addr_type = IPMI_MEM_ADDR_SPACE; + } + + info->io.addr_data = resource.start; - *new_info = info; + info->io.regsize = regsize ? be32_to_cpup(regsize) : DEFAULT_REGSIZE; + info->io.regspacing = regspacing ? be32_to_cpup(regspacing) : DEFAULT_REGSPACING; + info->io.regshift = regshift ? be32_to_cpup(regshift) : 0; - irqs[intf_num] = pci_dev->irq; - si_type[intf_num] = "smic"; + info->irq = irq_of_parse_and_map(dev->dev.of_node, 0); + info->dev = &dev->dev; - printk("ipmi_si: Found PCI SMIC at I/O address 0x%lx\n", - (long unsigned int) base_addr); + dev_dbg(&dev->dev, "addr 0x%lx regsize %d spacing %d irq %d\n", + info->io.addr_data, info->io.regsize, info->io.regspacing, + info->irq); - pci_dev_put(pci_dev); + dev_set_drvdata(&dev->dev, info); + + ret = add_smi(info); + if (ret) { + kfree(info); + return ret; + } +#endif return 0; } -#endif /* CONFIG_PCI */ -static int try_init_plug_and_play(int intf_num, struct smi_info **new_info) +static int ipmi_remove(struct platform_device *dev) { -#ifdef CONFIG_PCI - if (find_pci_smic(intf_num, new_info)==0) - return 0; +#ifdef CONFIG_OF + cleanup_one_si(dev_get_drvdata(&dev->dev)); #endif - /* Include other methods here. */ + return 0; +} + +static struct of_device_id ipmi_match[] = +{ + { .type = "ipmi", .compatible = "ipmi-kcs", + .data = (void *)(unsigned long) SI_KCS }, + { .type = "ipmi", .compatible = "ipmi-smic", + .data = (void *)(unsigned long) SI_SMIC }, + { .type = "ipmi", .compatible = "ipmi-bt", + .data = (void *)(unsigned long) SI_BT }, + {}, +}; + +static struct platform_driver ipmi_driver = { + .driver = { + .name = DEVICE_NAME, + .owner = THIS_MODULE, + .of_match_table = ipmi_match, + }, + .probe = ipmi_probe, + .remove = ipmi_remove, +}; + +#ifdef CONFIG_PARISC +static int ipmi_parisc_probe(struct parisc_device *dev) +{ + struct smi_info *info; + int rv; + + info = smi_info_alloc(); + + if (!info) { + dev_err(&dev->dev, + "could not allocate memory for PARISC probe\n"); + return -ENOMEM; + } + + info->si_type = SI_KCS; + info->addr_source = SI_DEVICETREE; + info->io_setup = mem_setup; + info->io.addr_type = IPMI_MEM_ADDR_SPACE; + info->io.addr_data = dev->hpa.start; + info->io.regsize = 1; + info->io.regspacing = 1; + info->io.regshift = 0; + info->irq = 0; /* no interrupt */ + info->irq_setup = NULL; + info->dev = &dev->dev; - return -ENODEV; + dev_dbg(&dev->dev, "addr 0x%lx\n", info->io.addr_data); + + dev_set_drvdata(&dev->dev, info); + + rv = add_smi(info); + if (rv) { + kfree(info); + return rv; + } + + return 0; } +static int ipmi_parisc_remove(struct parisc_device *dev) +{ + cleanup_one_si(dev_get_drvdata(&dev->dev)); + return 0; +} + +static struct parisc_device_id ipmi_parisc_tbl[] = { + { HPHW_MC, HVERSION_REV_ANY_ID, 0x004, 0xC0 }, + { 0, } +}; + +static struct parisc_driver ipmi_parisc_driver = { + .name = "ipmi", + .id_table = ipmi_parisc_tbl, + .probe = ipmi_parisc_probe, + .remove = ipmi_parisc_remove, +}; +#endif /* CONFIG_PARISC */ + +static int wait_for_msg_done(struct smi_info *smi_info) +{ + enum si_sm_result smi_result; + + smi_result = smi_info->handlers->event(smi_info->si_sm, 0); + for (;;) { + if (smi_result == SI_SM_CALL_WITH_DELAY || + smi_result == SI_SM_CALL_WITH_TICK_DELAY) { + schedule_timeout_uninterruptible(1); + smi_result = smi_info->handlers->event( + smi_info->si_sm, jiffies_to_usecs(1)); + } else if (smi_result == SI_SM_CALL_WITHOUT_DELAY) { + smi_result = smi_info->handlers->event( + smi_info->si_sm, 0); + } else + break; + } + if (smi_result == SI_SM_HOSED) + /* + * We couldn't get the state machine to run, so whatever's at + * the port is probably not an IPMI SMI interface. + */ + return -ENODEV; + + return 0; +} static int try_get_dev_id(struct smi_info *smi_info) { - unsigned char msg[2]; - unsigned char *resp; - unsigned long resp_len; - enum si_sm_result smi_result; - int rv = 0; + unsigned char msg[2]; + unsigned char *resp; + unsigned long resp_len; + int rv = 0; resp = kmalloc(IPMI_MAX_MSG_LENGTH, GFP_KERNEL); if (!resp) return -ENOMEM; - /* Do a Get Device ID command, since it comes back with some - useful info. */ + /* + * Do a Get Device ID command, since it comes back with some + * useful info. + */ msg[0] = IPMI_NETFN_APP_REQUEST << 2; msg[1] = IPMI_GET_DEVICE_ID_CMD; smi_info->handlers->start_transaction(smi_info->si_sm, msg, 2); - smi_result = smi_info->handlers->event(smi_info->si_sm, 0); - for (;;) - { - if (smi_result == SI_SM_CALL_WITH_DELAY) { - set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(1); - smi_result = smi_info->handlers->event( - smi_info->si_sm, 100); - } - else if (smi_result == SI_SM_CALL_WITHOUT_DELAY) - { - smi_result = smi_info->handlers->event( - smi_info->si_sm, 0); - } - else - break; - } - if (smi_result == SI_SM_HOSED) { - /* We couldn't get the state machine to run, so whatever's at - the port is probably not an IPMI SMI interface. */ - rv = -ENODEV; + rv = wait_for_msg_done(smi_info); + if (rv) + goto out; + + resp_len = smi_info->handlers->get_result(smi_info->si_sm, + resp, IPMI_MAX_MSG_LENGTH); + + /* Check and record info from the get device id, in case we need it. */ + rv = ipmi_demangle_device_id(resp, resp_len, &smi_info->device_id); + + out: + kfree(resp); + return rv; +} + +static int try_enable_event_buffer(struct smi_info *smi_info) +{ + unsigned char msg[3]; + unsigned char *resp; + unsigned long resp_len; + int rv = 0; + + resp = kmalloc(IPMI_MAX_MSG_LENGTH, GFP_KERNEL); + if (!resp) + return -ENOMEM; + + msg[0] = IPMI_NETFN_APP_REQUEST << 2; + msg[1] = IPMI_GET_BMC_GLOBAL_ENABLES_CMD; + smi_info->handlers->start_transaction(smi_info->si_sm, msg, 2); + + rv = wait_for_msg_done(smi_info); + if (rv) { + printk(KERN_WARNING PFX "Error getting response from get" + " global enables command, the event buffer is not" + " enabled.\n"); goto out; } - /* Otherwise, we got some data. */ resp_len = smi_info->handlers->get_result(smi_info->si_sm, resp, IPMI_MAX_MSG_LENGTH); - if (resp_len < 6) { - /* That's odd, it should be longer. */ + + if (resp_len < 4 || + resp[0] != (IPMI_NETFN_APP_REQUEST | 1) << 2 || + resp[1] != IPMI_GET_BMC_GLOBAL_ENABLES_CMD || + resp[2] != 0) { + printk(KERN_WARNING PFX "Invalid return from get global" + " enables command, cannot enable the event buffer.\n"); rv = -EINVAL; goto out; } - if ((resp[1] != IPMI_GET_DEVICE_ID_CMD) || (resp[2] != 0)) { - /* That's odd, it shouldn't be able to fail. */ - rv = -EINVAL; + if (resp[3] & IPMI_BMC_EVT_MSG_BUFF) + /* buffer is already enabled, nothing to do. */ + goto out; + + msg[0] = IPMI_NETFN_APP_REQUEST << 2; + msg[1] = IPMI_SET_BMC_GLOBAL_ENABLES_CMD; + msg[2] = resp[3] | IPMI_BMC_EVT_MSG_BUFF; + smi_info->handlers->start_transaction(smi_info->si_sm, msg, 3); + + rv = wait_for_msg_done(smi_info); + if (rv) { + printk(KERN_WARNING PFX "Error getting response from set" + " global, enables command, the event buffer is not" + " enabled.\n"); goto out; } - /* Record info from the get device id, in case we need it. */ - smi_info->ipmi_si_dev_rev = resp[4] & 0xf; - smi_info->ipmi_si_fw_rev_major = resp[5] & 0x7f; - smi_info->ipmi_si_fw_rev_minor = resp[6]; - smi_info->ipmi_version_major = resp[7] & 0xf; - smi_info->ipmi_version_minor = resp[7] >> 4; + resp_len = smi_info->handlers->get_result(smi_info->si_sm, + resp, IPMI_MAX_MSG_LENGTH); + + if (resp_len < 3 || + resp[0] != (IPMI_NETFN_APP_REQUEST | 1) << 2 || + resp[1] != IPMI_SET_BMC_GLOBAL_ENABLES_CMD) { + printk(KERN_WARNING PFX "Invalid return from get global," + "enables command, not enable the event buffer.\n"); + rv = -EINVAL; + goto out; + } + if (resp[2] != 0) + /* + * An error when setting the event buffer bit means + * that the event buffer is not supported. + */ + rv = -ENOENT; out: kfree(resp); return rv; } -static int type_file_read_proc(char *page, char **start, off_t off, - int count, int *eof, void *data) +static int smi_type_proc_show(struct seq_file *m, void *v) { - char *out = (char *) page; - struct smi_info *smi = data; + struct smi_info *smi = m->private; - switch (smi->si_type) { - case SI_KCS: - return sprintf(out, "kcs\n"); - case SI_SMIC: - return sprintf(out, "smic\n"); - case SI_BT: - return sprintf(out, "bt\n"); - default: - return 0; - } + return seq_printf(m, "%s\n", si_to_str[smi->si_type]); +} + +static int smi_type_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, smi_type_proc_show, PDE_DATA(inode)); } -static int stat_file_read_proc(char *page, char **start, off_t off, - int count, int *eof, void *data) +static const struct file_operations smi_type_proc_ops = { + .open = smi_type_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static int smi_si_stats_proc_show(struct seq_file *m, void *v) { - char *out = (char *) page; - struct smi_info *smi = data; + struct smi_info *smi = m->private; - out += sprintf(out, "interrupts_enabled: %d\n", + seq_printf(m, "interrupts_enabled: %d\n", smi->irq && !smi->interrupt_disabled); - out += sprintf(out, "short_timeouts: %ld\n", - smi->short_timeouts); - out += sprintf(out, "long_timeouts: %ld\n", - smi->long_timeouts); - out += sprintf(out, "timeout_restarts: %ld\n", - smi->timeout_restarts); - out += sprintf(out, "idles: %ld\n", - smi->idles); - out += sprintf(out, "interrupts: %ld\n", - smi->interrupts); - out += sprintf(out, "attentions: %ld\n", - smi->attentions); - out += sprintf(out, "flag_fetches: %ld\n", - smi->flag_fetches); - out += sprintf(out, "hosed_count: %ld\n", - smi->hosed_count); - out += sprintf(out, "complete_transactions: %ld\n", - smi->complete_transactions); - out += sprintf(out, "events: %ld\n", - smi->events); - out += sprintf(out, "watchdog_pretimeouts: %ld\n", - smi->watchdog_pretimeouts); - out += sprintf(out, "incoming_messages: %ld\n", - smi->incoming_messages); - - return (out - ((char *) page)); -} - -/* Returns 0 if initialized, or negative on an error. */ -static int init_one_smi(int intf_num, struct smi_info **smi) -{ - int rv; - struct smi_info *new_smi; - - - rv = try_init_mem(intf_num, &new_smi); - if (rv) - rv = try_init_port(intf_num, &new_smi); -#ifdef CONFIG_ACPI_INTERPRETER - if ((rv) && (si_trydefaults)) { - rv = try_init_acpi(intf_num, &new_smi); + seq_printf(m, "short_timeouts: %u\n", + smi_get_stat(smi, short_timeouts)); + seq_printf(m, "long_timeouts: %u\n", + smi_get_stat(smi, long_timeouts)); + seq_printf(m, "idles: %u\n", + smi_get_stat(smi, idles)); + seq_printf(m, "interrupts: %u\n", + smi_get_stat(smi, interrupts)); + seq_printf(m, "attentions: %u\n", + smi_get_stat(smi, attentions)); + seq_printf(m, "flag_fetches: %u\n", + smi_get_stat(smi, flag_fetches)); + seq_printf(m, "hosed_count: %u\n", + smi_get_stat(smi, hosed_count)); + seq_printf(m, "complete_transactions: %u\n", + smi_get_stat(smi, complete_transactions)); + seq_printf(m, "events: %u\n", + smi_get_stat(smi, events)); + seq_printf(m, "watchdog_pretimeouts: %u\n", + smi_get_stat(smi, watchdog_pretimeouts)); + seq_printf(m, "incoming_messages: %u\n", + smi_get_stat(smi, incoming_messages)); + return 0; +} + +static int smi_si_stats_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, smi_si_stats_proc_show, PDE_DATA(inode)); +} + +static const struct file_operations smi_si_stats_proc_ops = { + .open = smi_si_stats_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static int smi_params_proc_show(struct seq_file *m, void *v) +{ + struct smi_info *smi = m->private; + + return seq_printf(m, + "%s,%s,0x%lx,rsp=%d,rsi=%d,rsh=%d,irq=%d,ipmb=%d\n", + si_to_str[smi->si_type], + addr_space_to_str[smi->io.addr_type], + smi->io.addr_data, + smi->io.regspacing, + smi->io.regsize, + smi->io.regshift, + smi->irq, + smi->slave_addr); +} + +static int smi_params_proc_open(struct inode *inode, struct file *file) +{ + return single_open(file, smi_params_proc_show, PDE_DATA(inode)); +} + +static const struct file_operations smi_params_proc_ops = { + .open = smi_params_proc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +/* + * oem_data_avail_to_receive_msg_avail + * @info - smi_info structure with msg_flags set + * + * Converts flags from OEM_DATA_AVAIL to RECEIVE_MSG_AVAIL + * Returns 1 indicating need to re-run handle_flags(). + */ +static int oem_data_avail_to_receive_msg_avail(struct smi_info *smi_info) +{ + smi_info->msg_flags = ((smi_info->msg_flags & ~OEM_DATA_AVAIL) | + RECEIVE_MSG_AVAIL); + return 1; +} + +/* + * setup_dell_poweredge_oem_data_handler + * @info - smi_info.device_id must be populated + * + * Systems that match, but have firmware version < 1.40 may assert + * OEM0_DATA_AVAIL on their own, without being told via Set Flags that + * it's safe to do so. Such systems will de-assert OEM1_DATA_AVAIL + * upon receipt of IPMI_GET_MSG_CMD, so we should treat these flags + * as RECEIVE_MSG_AVAIL instead. + * + * As Dell has no plans to release IPMI 1.5 firmware that *ever* + * assert the OEM[012] bits, and if it did, the driver would have to + * change to handle that properly, we don't actually check for the + * firmware version. + * Device ID = 0x20 BMC on PowerEdge 8G servers + * Device Revision = 0x80 + * Firmware Revision1 = 0x01 BMC version 1.40 + * Firmware Revision2 = 0x40 BCD encoded + * IPMI Version = 0x51 IPMI 1.5 + * Manufacturer ID = A2 02 00 Dell IANA + * + * Additionally, PowerEdge systems with IPMI < 1.5 may also assert + * OEM0_DATA_AVAIL and needs to be treated as RECEIVE_MSG_AVAIL. + * + */ +#define DELL_POWEREDGE_8G_BMC_DEVICE_ID 0x20 +#define DELL_POWEREDGE_8G_BMC_DEVICE_REV 0x80 +#define DELL_POWEREDGE_8G_BMC_IPMI_VERSION 0x51 +#define DELL_IANA_MFR_ID 0x0002a2 +static void setup_dell_poweredge_oem_data_handler(struct smi_info *smi_info) +{ + struct ipmi_device_id *id = &smi_info->device_id; + if (id->manufacturer_id == DELL_IANA_MFR_ID) { + if (id->device_id == DELL_POWEREDGE_8G_BMC_DEVICE_ID && + id->device_revision == DELL_POWEREDGE_8G_BMC_DEVICE_REV && + id->ipmi_version == DELL_POWEREDGE_8G_BMC_IPMI_VERSION) { + smi_info->oem_data_avail_handler = + oem_data_avail_to_receive_msg_avail; + } else if (ipmi_version_major(id) < 1 || + (ipmi_version_major(id) == 1 && + ipmi_version_minor(id) < 5)) { + smi_info->oem_data_avail_handler = + oem_data_avail_to_receive_msg_avail; + } } +} + +#define CANNOT_RETURN_REQUESTED_LENGTH 0xCA +static void return_hosed_msg_badsize(struct smi_info *smi_info) +{ + struct ipmi_smi_msg *msg = smi_info->curr_msg; + + /* Make it a response */ + msg->rsp[0] = msg->data[0] | 4; + msg->rsp[1] = msg->data[1]; + msg->rsp[2] = CANNOT_RETURN_REQUESTED_LENGTH; + msg->rsp_size = 3; + smi_info->curr_msg = NULL; + deliver_recv_msg(smi_info, msg); +} + +/* + * dell_poweredge_bt_xaction_handler + * @info - smi_info.device_id must be populated + * + * Dell PowerEdge servers with the BT interface (x6xx and 1750) will + * not respond to a Get SDR command if the length of the data + * requested is exactly 0x3A, which leads to command timeouts and no + * data returned. This intercepts such commands, and causes userspace + * callers to try again with a different-sized buffer, which succeeds. + */ + +#define STORAGE_NETFN 0x0A +#define STORAGE_CMD_GET_SDR 0x23 +static int dell_poweredge_bt_xaction_handler(struct notifier_block *self, + unsigned long unused, + void *in) +{ + struct smi_info *smi_info = in; + unsigned char *data = smi_info->curr_msg->data; + unsigned int size = smi_info->curr_msg->data_size; + if (size >= 8 && + (data[0]>>2) == STORAGE_NETFN && + data[1] == STORAGE_CMD_GET_SDR && + data[7] == 0x3A) { + return_hosed_msg_badsize(smi_info); + return NOTIFY_STOP; + } + return NOTIFY_DONE; +} + +static struct notifier_block dell_poweredge_bt_xaction_notifier = { + .notifier_call = dell_poweredge_bt_xaction_handler, +}; + +/* + * setup_dell_poweredge_bt_xaction_handler + * @info - smi_info.device_id must be filled in already + * + * Fills in smi_info.device_id.start_transaction_pre_hook + * when we know what function to use there. + */ +static void +setup_dell_poweredge_bt_xaction_handler(struct smi_info *smi_info) +{ + struct ipmi_device_id *id = &smi_info->device_id; + if (id->manufacturer_id == DELL_IANA_MFR_ID && + smi_info->si_type == SI_BT) + register_xaction_notifier(&dell_poweredge_bt_xaction_notifier); +} + +/* + * setup_oem_data_handler + * @info - smi_info.device_id must be filled in already + * + * Fills in smi_info.device_id.oem_data_available_handler + * when we know what function to use there. + */ + +static void setup_oem_data_handler(struct smi_info *smi_info) +{ + setup_dell_poweredge_oem_data_handler(smi_info); +} + +static void setup_xaction_handlers(struct smi_info *smi_info) +{ + setup_dell_poweredge_bt_xaction_handler(smi_info); +} + +static inline void wait_for_timer_and_thread(struct smi_info *smi_info) +{ + if (smi_info->intf) { + /* + * The timer and thread are only running if the + * interface has been started up and registered. + */ + if (smi_info->thread != NULL) + kthread_stop(smi_info->thread); + del_timer_sync(&smi_info->si_timer); + } +} + +static struct ipmi_default_vals +{ + int type; + int port; +} ipmi_defaults[] = +{ + { .type = SI_KCS, .port = 0xca2 }, + { .type = SI_SMIC, .port = 0xca9 }, + { .type = SI_BT, .port = 0xe4 }, + { .port = 0 } +}; + +static void default_find_bmc(void) +{ + struct smi_info *info; + int i; + + for (i = 0; ; i++) { + if (!ipmi_defaults[i].port) + break; +#ifdef CONFIG_PPC + if (check_legacy_ioport(ipmi_defaults[i].port)) + continue; #endif -#ifdef CONFIG_X86 - if ((rv) && (si_trydefaults)) { - rv = try_init_smbios(intf_num, &new_smi); - } -#endif - if ((rv) && (si_trydefaults)) { - rv = try_init_plug_and_play(intf_num, &new_smi); + info = smi_info_alloc(); + if (!info) + return; + + info->addr_source = SI_DEFAULT; + + info->si_type = ipmi_defaults[i].type; + info->io_setup = port_setup; + info->io.addr_data = ipmi_defaults[i].port; + info->io.addr_type = IPMI_IO_ADDR_SPACE; + + info->io.addr = NULL; + info->io.regspacing = DEFAULT_REGSPACING; + info->io.regsize = DEFAULT_REGSPACING; + info->io.regshift = 0; + + if (add_smi(info) == 0) { + if ((try_smi_init(info)) == 0) { + /* Found one... */ + printk(KERN_INFO PFX "Found default %s" + " state machine at %s address 0x%lx\n", + si_to_str[info->si_type], + addr_space_to_str[info->io.addr_type], + info->io.addr_data); + } else + cleanup_one_si(info); + } else { + kfree(info); + } } +} +static int is_new_interface(struct smi_info *info) +{ + struct smi_info *e; - if (rv) - return rv; + list_for_each_entry(e, &smi_infos, link) { + if (e->io.addr_type != info->io.addr_type) + continue; + if (e->io.addr_data == info->io.addr_data) + return 0; + } + + return 1; +} + +static int add_smi(struct smi_info *new_smi) +{ + int rv = 0; + + printk(KERN_INFO PFX "Adding %s-specified %s state machine", + ipmi_addr_src_to_str[new_smi->addr_source], + si_to_str[new_smi->si_type]); + mutex_lock(&smi_infos_lock); + if (!is_new_interface(new_smi)) { + printk(KERN_CONT " duplicate interface\n"); + rv = -EBUSY; + goto out_err; + } + + printk(KERN_CONT "\n"); /* So we know not to free it unless we have allocated one. */ new_smi->intf = NULL; new_smi->si_sm = NULL; new_smi->handlers = NULL; - if (!new_smi->irq_setup) { - new_smi->irq = irqs[intf_num]; - new_smi->irq_setup = std_irq_setup; - new_smi->irq_cleanup = std_irq_cleanup; - } + list_add_tail(&new_smi->link, &smi_infos); - /* Default to KCS if no type is specified. */ - if (si_type[intf_num] == NULL) { - if (si_trydefaults) - si_type[intf_num] = "kcs"; - else { - rv = -EINVAL; - goto out_err; - } - } +out_err: + mutex_unlock(&smi_infos_lock); + return rv; +} + +static int try_smi_init(struct smi_info *new_smi) +{ + int rv = 0; + int i; - /* Set up the state machine to use. */ - if (strcmp(si_type[intf_num], "kcs") == 0) { + printk(KERN_INFO PFX "Trying %s-specified %s state" + " machine at %s address 0x%lx, slave address 0x%x," + " irq %d\n", + ipmi_addr_src_to_str[new_smi->addr_source], + si_to_str[new_smi->si_type], + addr_space_to_str[new_smi->io.addr_type], + new_smi->io.addr_data, + new_smi->slave_addr, new_smi->irq); + + switch (new_smi->si_type) { + case SI_KCS: new_smi->handlers = &kcs_smi_handlers; - new_smi->si_type = SI_KCS; - } else if (strcmp(si_type[intf_num], "smic") == 0) { + break; + + case SI_SMIC: new_smi->handlers = &smic_smi_handlers; - new_smi->si_type = SI_SMIC; - } else if (strcmp(si_type[intf_num], "bt") == 0) { + break; + + case SI_BT: new_smi->handlers = &bt_smi_handlers; - new_smi->si_type = SI_BT; - } else { + break; + + default: /* No support for anything else yet. */ rv = -EIO; goto out_err; @@ -2125,7 +3330,8 @@ static int init_one_smi(int intf_num, struct smi_info **smi) /* Allocate the state machine's data and initialize it. */ new_smi->si_sm = kmalloc(new_smi->handlers->size(), GFP_KERNEL); if (!new_smi->si_sm) { - printk(" Could not allocate state machine memory\n"); + printk(KERN_ERR PFX + "Could not allocate state machine memory\n"); rv = -ENOMEM; goto out_err; } @@ -2135,142 +3341,199 @@ static int init_one_smi(int intf_num, struct smi_info **smi) /* Now that we know the I/O size, we can set up the I/O. */ rv = new_smi->io_setup(new_smi); if (rv) { - printk(" Could not set up I/O space\n"); + printk(KERN_ERR PFX "Could not set up I/O space\n"); goto out_err; } - spin_lock_init(&(new_smi->si_lock)); - spin_lock_init(&(new_smi->msg_lock)); - spin_lock_init(&(new_smi->count_lock)); - /* Do low-level detection first. */ if (new_smi->handlers->detect(new_smi->si_sm)) { + if (new_smi->addr_source) + printk(KERN_INFO PFX "Interface detection failed\n"); rv = -ENODEV; goto out_err; } - /* Attempt a get device id command. If it fails, we probably - don't have a SMI here. */ + /* + * Attempt a get device id command. If it fails, we probably + * don't have a BMC here. + */ rv = try_get_dev_id(new_smi); - if (rv) + if (rv) { + if (new_smi->addr_source) + printk(KERN_INFO PFX "There appears to be no BMC" + " at this location\n"); goto out_err; + } - /* Try to claim any interrupts. */ - new_smi->irq_setup(new_smi); + setup_oem_data_handler(new_smi); + setup_xaction_handlers(new_smi); INIT_LIST_HEAD(&(new_smi->xmit_msgs)); INIT_LIST_HEAD(&(new_smi->hp_xmit_msgs)); new_smi->curr_msg = NULL; atomic_set(&new_smi->req_events, 0); - new_smi->run_to_completion = 0; + new_smi->run_to_completion = false; + for (i = 0; i < SI_NUM_STATS; i++) + atomic_set(&new_smi->stats[i], 0); + + new_smi->interrupt_disabled = true; + atomic_set(&new_smi->stop_operation, 0); + atomic_set(&new_smi->need_watch, 0); + new_smi->intf_num = smi_num; + smi_num++; - new_smi->interrupt_disabled = 0; - new_smi->timer_stopped = 0; - new_smi->stop_operation = 0; + rv = try_enable_event_buffer(new_smi); + if (rv == 0) + new_smi->has_event_buffer = true; - /* Start clearing the flags before we enable interrupts or the - timer to avoid racing with the timer. */ + /* + * Start clearing the flags before we enable interrupts or the + * timer to avoid racing with the timer. + */ start_clear_flags(new_smi); /* IRQ is defined to be set when non-zero. */ if (new_smi->irq) new_smi->si_state = SI_CLEARING_FLAGS_THEN_SET_IRQ; - /* The ipmi_register_smi() code does some operations to - determine the channel information, so we must be ready to - handle operations before it is called. This means we have - to stop the timer if we get an error after this point. */ - init_timer(&(new_smi->si_timer)); - new_smi->si_timer.data = (long) new_smi; - new_smi->si_timer.function = smi_timeout; - new_smi->last_timeout_jiffies = jiffies; - new_smi->si_timer.expires = jiffies + SI_TIMEOUT_JIFFIES; - add_timer(&(new_smi->si_timer)); + if (!new_smi->dev) { + /* + * If we don't already have a device from something + * else (like PCI), then register a new one. + */ + new_smi->pdev = platform_device_alloc("ipmi_si", + new_smi->intf_num); + if (!new_smi->pdev) { + printk(KERN_ERR PFX + "Unable to allocate platform device\n"); + goto out_err; + } + new_smi->dev = &new_smi->pdev->dev; + new_smi->dev->driver = &ipmi_driver.driver; + + rv = platform_device_add(new_smi->pdev); + if (rv) { + printk(KERN_ERR PFX + "Unable to register system interface device:" + " %d\n", + rv); + goto out_err; + } + new_smi->dev_registered = true; + } rv = ipmi_register_smi(&handlers, new_smi, - new_smi->ipmi_version_major, - new_smi->ipmi_version_minor, - new_smi->slave_addr, - &(new_smi->intf)); + &new_smi->device_id, + new_smi->dev, + "bmc", + new_smi->slave_addr); if (rv) { - printk(KERN_ERR - "ipmi_si: Unable to register device: error %d\n", - rv); + dev_err(new_smi->dev, "Unable to register device: error %d\n", + rv); goto out_err_stop_timer; } rv = ipmi_smi_add_proc_entry(new_smi->intf, "type", - type_file_read_proc, NULL, - new_smi, THIS_MODULE); + &smi_type_proc_ops, + new_smi); if (rv) { - printk(KERN_ERR - "ipmi_si: Unable to create proc entry: %d\n", - rv); + dev_err(new_smi->dev, "Unable to create proc entry: %d\n", rv); goto out_err_stop_timer; } rv = ipmi_smi_add_proc_entry(new_smi->intf, "si_stats", - stat_file_read_proc, NULL, - new_smi, THIS_MODULE); + &smi_si_stats_proc_ops, + new_smi); if (rv) { - printk(KERN_ERR - "ipmi_si: Unable to create proc entry: %d\n", - rv); + dev_err(new_smi->dev, "Unable to create proc entry: %d\n", rv); goto out_err_stop_timer; } - *smi = new_smi; + rv = ipmi_smi_add_proc_entry(new_smi->intf, "params", + &smi_params_proc_ops, + new_smi); + if (rv) { + dev_err(new_smi->dev, "Unable to create proc entry: %d\n", rv); + goto out_err_stop_timer; + } - printk(" IPMI %s interface initialized\n", si_type[intf_num]); + dev_info(new_smi->dev, "IPMI %s interface initialized\n", + si_to_str[new_smi->si_type]); return 0; out_err_stop_timer: - new_smi->stop_operation = 1; - - /* Wait for the timer to stop. This avoids problems with race - conditions removing the timer here. */ - while (!new_smi->timer_stopped) { - set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(1); - } + atomic_inc(&new_smi->stop_operation); + wait_for_timer_and_thread(new_smi); out_err: - if (new_smi->intf) + new_smi->interrupt_disabled = true; + + if (new_smi->intf) { ipmi_unregister_smi(new_smi->intf); + new_smi->intf = NULL; + } - new_smi->irq_cleanup(new_smi); + if (new_smi->irq_cleanup) { + new_smi->irq_cleanup(new_smi); + new_smi->irq_cleanup = NULL; + } - /* Wait until we know that we are out of any interrupt - handlers might have been running before we freed the - interrupt. */ + /* + * Wait until we know that we are out of any interrupt + * handlers might have been running before we freed the + * interrupt. + */ synchronize_sched(); if (new_smi->si_sm) { if (new_smi->handlers) new_smi->handlers->cleanup(new_smi->si_sm); kfree(new_smi->si_sm); + new_smi->si_sm = NULL; + } + if (new_smi->addr_source_cleanup) { + new_smi->addr_source_cleanup(new_smi); + new_smi->addr_source_cleanup = NULL; + } + if (new_smi->io_cleanup) { + new_smi->io_cleanup(new_smi); + new_smi->io_cleanup = NULL; + } + + if (new_smi->dev_registered) { + platform_device_unregister(new_smi->pdev); + new_smi->dev_registered = false; } - new_smi->io_cleanup(new_smi); return rv; } -static __init int init_ipmi_si(void) +static int init_ipmi_si(void) { - int rv = 0; - int pos = 0; int i; char *str; + int rv; + struct smi_info *e; + enum ipmi_addr_src type = SI_INVALID; if (initialized) return 0; initialized = 1; + if (si_tryplatform) { + rv = platform_driver_register(&ipmi_driver); + if (rv) { + printk(KERN_ERR PFX "Unable to register " + "driver: %d\n", rv); + return rv; + } + } + /* Parse out the si_type string into its components. */ str = si_type_str; if (*str != '\0') { - for (i=0; (i<SI_MAX_PARMS) && (*str != '\0'); i++) { + for (i = 0; (i < SI_MAX_PARMS) && (*str != '\0'); i++) { si_type[i] = str; str = strchr(str, ','); if (str) { @@ -2282,123 +3545,208 @@ static __init int init_ipmi_si(void) } } - printk(KERN_INFO "IPMI System Interface driver version " - IPMI_SI_VERSION); - if (kcs_smi_handlers.version) - printk(", KCS version %s", kcs_smi_handlers.version); - if (smic_smi_handlers.version) - printk(", SMIC version %s", smic_smi_handlers.version); - if (bt_smi_handlers.version) - printk(", BT version %s", bt_smi_handlers.version); - printk("\n"); + printk(KERN_INFO "IPMI System Interface driver.\n"); + + /* If the user gave us a device, they presumably want us to use it */ + if (!hardcode_find_bmc()) + return 0; + +#ifdef CONFIG_PCI + if (si_trypci) { + rv = pci_register_driver(&ipmi_pci_driver); + if (rv) + printk(KERN_ERR PFX "Unable to register " + "PCI driver: %d\n", rv); + else + pci_registered = true; + } +#endif + +#ifdef CONFIG_ACPI + if (si_tryacpi) { + pnp_register_driver(&ipmi_pnp_driver); + pnp_registered = true; + } +#endif -#ifdef CONFIG_X86 - dmi_decode(); +#ifdef CONFIG_DMI + if (si_trydmi) + dmi_find_bmc(); #endif - rv = init_one_smi(0, &(smi_infos[pos])); - if (rv && !ports[0] && si_trydefaults) { - /* If we are trying defaults and the initial port is - not set, then set it. */ - si_type[0] = "kcs"; - ports[0] = DEFAULT_KCS_IO_PORT; - rv = init_one_smi(0, &(smi_infos[pos])); - if (rv) { - /* No KCS - try SMIC */ - si_type[0] = "smic"; - ports[0] = DEFAULT_SMIC_IO_PORT; - rv = init_one_smi(0, &(smi_infos[pos])); - } - if (rv) { - /* No SMIC - try BT */ - si_type[0] = "bt"; - ports[0] = DEFAULT_BT_IO_PORT; - rv = init_one_smi(0, &(smi_infos[pos])); +#ifdef CONFIG_ACPI + if (si_tryacpi) + spmi_find_bmc(); +#endif + +#ifdef CONFIG_PARISC + register_parisc_driver(&ipmi_parisc_driver); + parisc_registered = true; + /* poking PC IO addresses will crash machine, don't do it */ + si_trydefaults = 0; +#endif + + /* We prefer devices with interrupts, but in the case of a machine + with multiple BMCs we assume that there will be several instances + of a given type so if we succeed in registering a type then also + try to register everything else of the same type */ + + mutex_lock(&smi_infos_lock); + list_for_each_entry(e, &smi_infos, link) { + /* Try to register a device if it has an IRQ and we either + haven't successfully registered a device yet or this + device has the same type as one we successfully registered */ + if (e->irq && (!type || e->addr_source == type)) { + if (!try_smi_init(e)) { + type = e->addr_source; + } } } - if (rv == 0) - pos++; - for (i=1; i < SI_MAX_PARMS; i++) { - rv = init_one_smi(i, &(smi_infos[pos])); - if (rv == 0) - pos++; + /* type will only have been set if we successfully registered an si */ + if (type) { + mutex_unlock(&smi_infos_lock); + return 0; } - if (smi_infos[0] == NULL) { - printk("ipmi_si: Unable to find any System Interface(s)\n"); - return -ENODEV; + /* Fall back to the preferred device */ + + list_for_each_entry(e, &smi_infos, link) { + if (!e->irq && (!type || e->addr_source == type)) { + if (!try_smi_init(e)) { + type = e->addr_source; + } + } } + mutex_unlock(&smi_infos_lock); - return 0; + if (type) + return 0; + + if (si_trydefaults) { + mutex_lock(&smi_infos_lock); + if (list_empty(&smi_infos)) { + /* No BMC was found, try defaults. */ + mutex_unlock(&smi_infos_lock); + default_find_bmc(); + } else + mutex_unlock(&smi_infos_lock); + } + + mutex_lock(&smi_infos_lock); + if (unload_when_empty && list_empty(&smi_infos)) { + mutex_unlock(&smi_infos_lock); + cleanup_ipmi_si(); + printk(KERN_WARNING PFX + "Unable to find any System Interface(s)\n"); + return -ENODEV; + } else { + mutex_unlock(&smi_infos_lock); + return 0; + } } module_init(init_ipmi_si); -static void __exit cleanup_one_si(struct smi_info *to_clean) +static void cleanup_one_si(struct smi_info *to_clean) { - int rv; + int rv = 0; unsigned long flags; - if (! to_clean) + if (!to_clean) return; - /* Tell the timer and interrupt handlers that we are shutting - down. */ - spin_lock_irqsave(&(to_clean->si_lock), flags); - spin_lock(&(to_clean->msg_lock)); - - to_clean->stop_operation = 1; - - to_clean->irq_cleanup(to_clean); + list_del(&to_clean->link); - spin_unlock(&(to_clean->msg_lock)); - spin_unlock_irqrestore(&(to_clean->si_lock), flags); + /* Tell the driver that we are shutting down. */ + atomic_inc(&to_clean->stop_operation); - /* Wait until we know that we are out of any interrupt - handlers might have been running before we freed the - interrupt. */ - synchronize_sched(); + /* + * Make sure the timer and thread are stopped and will not run + * again. + */ + wait_for_timer_and_thread(to_clean); - /* Wait for the timer to stop. This avoids problems with race - conditions removing the timer here. */ - while (!to_clean->timer_stopped) { - set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(1); + /* + * Timeouts are stopped, now make sure the interrupts are off + * for the device. A little tricky with locks to make sure + * there are no races. + */ + spin_lock_irqsave(&to_clean->si_lock, flags); + while (to_clean->curr_msg || (to_clean->si_state != SI_NORMAL)) { + spin_unlock_irqrestore(&to_clean->si_lock, flags); + poll(to_clean); + schedule_timeout_uninterruptible(1); + spin_lock_irqsave(&to_clean->si_lock, flags); + } + disable_si_irq(to_clean); + spin_unlock_irqrestore(&to_clean->si_lock, flags); + while (to_clean->curr_msg || (to_clean->si_state != SI_NORMAL)) { + poll(to_clean); + schedule_timeout_uninterruptible(1); } - /* Interrupts and timeouts are stopped, now make sure the - interface is in a clean state. */ - while ((to_clean->curr_msg) || (to_clean->si_state != SI_NORMAL)) { + /* Clean up interrupts and make sure that everything is done. */ + if (to_clean->irq_cleanup) + to_clean->irq_cleanup(to_clean); + while (to_clean->curr_msg || (to_clean->si_state != SI_NORMAL)) { poll(to_clean); - set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(1); + schedule_timeout_uninterruptible(1); } - rv = ipmi_unregister_smi(to_clean->intf); + if (to_clean->intf) + rv = ipmi_unregister_smi(to_clean->intf); + if (rv) { - printk(KERN_ERR - "ipmi_si: Unable to unregister device: errno=%d\n", + printk(KERN_ERR PFX "Unable to unregister device: errno=%d\n", rv); } - to_clean->handlers->cleanup(to_clean->si_sm); + if (to_clean->handlers) + to_clean->handlers->cleanup(to_clean->si_sm); kfree(to_clean->si_sm); - to_clean->io_cleanup(to_clean); + if (to_clean->addr_source_cleanup) + to_clean->addr_source_cleanup(to_clean); + if (to_clean->io_cleanup) + to_clean->io_cleanup(to_clean); + + if (to_clean->dev_registered) + platform_device_unregister(to_clean->pdev); + + kfree(to_clean); } -static __exit void cleanup_ipmi_si(void) +static void cleanup_ipmi_si(void) { - int i; + struct smi_info *e, *tmp_e; if (!initialized) return; - for (i=0; i<SI_MAX_DRIVERS; i++) { - cleanup_one_si(smi_infos[i]); - } +#ifdef CONFIG_PCI + if (pci_registered) + pci_unregister_driver(&ipmi_pci_driver); +#endif +#ifdef CONFIG_ACPI + if (pnp_registered) + pnp_unregister_driver(&ipmi_pnp_driver); +#endif +#ifdef CONFIG_PARISC + if (parisc_registered) + unregister_parisc_driver(&ipmi_parisc_driver); +#endif + + platform_driver_unregister(&ipmi_driver); + + mutex_lock(&smi_infos_lock); + list_for_each_entry_safe(e, tmp_e, &smi_infos, link) + cleanup_one_si(e); + mutex_unlock(&smi_infos_lock); } module_exit(cleanup_ipmi_si); MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Corey Minyard <minyard@mvista.com>"); +MODULE_DESCRIPTION("Interface to the IPMI driver for the KCS, SMIC, and BT" + " system interfaces."); |