X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=block%2Fblktrace.c;h=85049a7e7a179a97c283eb4ebe6c1fe7285f80cf;hb=e436fdae70a31102d2be32969b80fe8545edebd9;hp=8ff33441d8a214259bd2c718ddb4735874dd2fc5;hpb=6c5c934153513dc72e2d6464f39e8ef1f27c0a3e;p=safe%2Fjmp%2Flinux-2.6 diff --git a/block/blktrace.c b/block/blktrace.c index 8ff3344..85049a7 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2006 Jens Axboe + * Copyright (C) 2006 Jens Axboe * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as @@ -22,32 +22,77 @@ #include #include #include +#include #include -static DEFINE_PER_CPU(unsigned long long, blk_trace_cpu_offset) = { 0, }; static unsigned int blktrace_seq __read_mostly = 1; /* - * Send out a notify for this process, if we haven't done so since a trace - * started + * Send out a notify message. */ -static void trace_note_tsk(struct blk_trace *bt, struct task_struct *tsk) +static void trace_note(struct blk_trace *bt, pid_t pid, int action, + const void *data, size_t len) { struct blk_io_trace *t; - t = relay_reserve(bt->rchan, sizeof(*t) + sizeof(tsk->comm)); + t = relay_reserve(bt->rchan, sizeof(*t) + len); if (t) { + const int cpu = smp_processor_id(); + t->magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION; + t->time = ktime_to_ns(ktime_get()); t->device = bt->dev; - t->action = BLK_TC_ACT(BLK_TC_NOTIFY); - t->pid = tsk->pid; - t->cpu = smp_processor_id(); - t->pdu_len = sizeof(tsk->comm); - memcpy((void *) t + sizeof(*t), tsk->comm, t->pdu_len); - tsk->btrace_seq = blktrace_seq; + t->action = action; + t->pid = pid; + t->cpu = cpu; + t->pdu_len = len; + memcpy((void *) t + sizeof(*t), data, len); } } +/* + * Send out a notify for this process, if we haven't done so since a trace + * started + */ +static void trace_note_tsk(struct blk_trace *bt, struct task_struct *tsk) +{ + tsk->btrace_seq = blktrace_seq; + trace_note(bt, tsk->pid, BLK_TN_PROCESS, tsk->comm, sizeof(tsk->comm)); +} + +static void trace_note_time(struct blk_trace *bt) +{ + struct timespec now; + unsigned long flags; + u32 words[2]; + + getnstimeofday(&now); + words[0] = now.tv_sec; + words[1] = now.tv_nsec; + + local_irq_save(flags); + trace_note(bt, 0, BLK_TN_TIMESTAMP, words, sizeof(words)); + local_irq_restore(flags); +} + +void __trace_note_message(struct blk_trace *bt, const char *fmt, ...) +{ + int n; + va_list args; + unsigned long flags; + char *buf; + + local_irq_save(flags); + buf = per_cpu_ptr(bt->msg_data, smp_processor_id()); + va_start(args, fmt); + n = vscnprintf(buf, BLK_TN_MAX_MSG, fmt, args); + va_end(args); + + trace_note(bt, 0, BLK_TN_MESSAGE, buf, n); + local_irq_restore(flags); +} +EXPORT_SYMBOL_GPL(__trace_note_message); + static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector, pid_t pid) { @@ -66,21 +111,9 @@ static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector, */ static u32 ddir_act[2] __read_mostly = { BLK_TC_ACT(BLK_TC_READ), BLK_TC_ACT(BLK_TC_WRITE) }; -/* - * Bio action bits of interest - */ -static u32 bio_act[5] __read_mostly = { 0, BLK_TC_ACT(BLK_TC_BARRIER), BLK_TC_ACT(BLK_TC_SYNC), 0, BLK_TC_ACT(BLK_TC_AHEAD) }; - -/* - * More could be added as needed, taking care to increment the decrementer - * to get correct indexing - */ -#define trace_barrier_bit(rw) \ - (((rw) & (1 << BIO_RW_BARRIER)) >> (BIO_RW_BARRIER - 0)) -#define trace_sync_bit(rw) \ - (((rw) & (1 << BIO_RW_SYNC)) >> (BIO_RW_SYNC - 1)) -#define trace_ahead_bit(rw) \ - (((rw) & (1 << BIO_RW_AHEAD)) << (2 - BIO_RW_AHEAD)) +/* The ilog2() calls fall out because they're constant */ +#define MASK_TC_BIT(rw, __name) ( (rw & (1 << BIO_RW_ ## __name)) << \ + (ilog2(BLK_TC_ ## __name) + BLK_TC_SHIFT - BIO_RW_ ## __name) ) /* * The worker for the various blk_add_trace*() types. Fills out a @@ -100,9 +133,11 @@ void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, return; what |= ddir_act[rw & WRITE]; - what |= bio_act[trace_barrier_bit(rw)]; - what |= bio_act[trace_sync_bit(rw)]; - what |= bio_act[trace_ahead_bit(rw)]; + what |= MASK_TC_BIT(rw, BARRIER); + what |= MASK_TC_BIT(rw, SYNC); + what |= MASK_TC_BIT(rw, AHEAD); + what |= MASK_TC_BIT(rw, META); + what |= MASK_TC_BIT(rw, DISCARD); pid = tsk->pid; if (unlikely(act_log_check(bt, what, sector, pid))) @@ -111,10 +146,7 @@ void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, /* * A word about the locking here - we disable interrupts to reserve * some space in the relay per-cpu buffer, to prevent an irq - * from coming in and stepping on our toes. Once reserved, it's - * enough to get preemption disabled to prevent read of this data - * before we are through filling it. get_cpu()/put_cpu() does this - * for us + * from coming in and stepping on our toes. */ local_irq_save(flags); @@ -128,7 +160,7 @@ void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, t->magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION; t->sequence = ++(*sequence); - t->time = sched_clock() - per_cpu(blk_trace_cpu_offset, cpu); + t->time = ktime_to_ns(ktime_get()); t->sector = sector; t->bytes = bytes; t->action = what; @@ -148,7 +180,7 @@ void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, EXPORT_SYMBOL_GPL(__blk_add_trace); static struct dentry *blk_tree_root; -static struct mutex blk_tree_mutex; +static DEFINE_MUTEX(blk_tree_mutex); static unsigned int root_users; static inline void blk_remove_root(void) @@ -171,6 +203,7 @@ static void blk_remove_tree(struct dentry *dir) static struct dentry *blk_create_tree(const char *blk_name) { struct dentry *dir = NULL; + int created = 0; mutex_lock(&blk_tree_mutex); @@ -178,13 +211,17 @@ static struct dentry *blk_create_tree(const char *blk_name) blk_tree_root = debugfs_create_dir("block", NULL); if (!blk_tree_root) goto err; + created = 1; } dir = debugfs_create_dir(blk_name, blk_tree_root); if (dir) root_users++; - else - blk_remove_root(); + else { + /* Delete root only if we created it */ + if (created) + blk_remove_root(); + } err: mutex_unlock(&blk_tree_mutex); @@ -194,13 +231,15 @@ err: static void blk_trace_cleanup(struct blk_trace *bt) { relay_close(bt->rchan); + debugfs_remove(bt->msg_file); debugfs_remove(bt->dropped_file); blk_remove_tree(bt->dir); free_percpu(bt->sequence); + free_percpu(bt->msg_data); kfree(bt); } -static int blk_trace_remove(request_queue_t *q) +int blk_trace_remove(struct request_queue *q) { struct blk_trace *bt; @@ -214,6 +253,7 @@ static int blk_trace_remove(request_queue_t *q) return 0; } +EXPORT_SYMBOL_GPL(blk_trace_remove); static int blk_dropped_open(struct inode *inode, struct file *filp) { @@ -233,12 +273,50 @@ static ssize_t blk_dropped_read(struct file *filp, char __user *buffer, return simple_read_from_buffer(buffer, count, ppos, buf, strlen(buf)); } -static struct file_operations blk_dropped_fops = { +static const struct file_operations blk_dropped_fops = { .owner = THIS_MODULE, .open = blk_dropped_open, .read = blk_dropped_read, }; +static int blk_msg_open(struct inode *inode, struct file *filp) +{ + filp->private_data = inode->i_private; + + return 0; +} + +static ssize_t blk_msg_write(struct file *filp, const char __user *buffer, + size_t count, loff_t *ppos) +{ + char *msg; + struct blk_trace *bt; + + if (count > BLK_TN_MAX_MSG) + return -EINVAL; + + msg = kmalloc(count, GFP_KERNEL); + if (msg == NULL) + return -ENOMEM; + + if (copy_from_user(msg, buffer, count)) { + kfree(msg); + return -EFAULT; + } + + bt = filp->private_data; + __trace_note_message(bt, "%s", msg); + kfree(msg); + + return count; +} + +static const struct file_operations blk_msg_fops = { + .owner = THIS_MODULE, + .open = blk_msg_open, + .write = blk_msg_write, +}; + /* * Keep track of how many times we encountered a full subbuffer, to aid * the user space app in telling how many lost events there were. @@ -281,33 +359,26 @@ static struct rchan_callbacks blk_relay_callbacks = { /* * Setup everything required to start tracing */ -static int blk_trace_setup(request_queue_t *q, struct block_device *bdev, - char __user *arg) +int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, + struct blk_user_trace_setup *buts) { - struct blk_user_trace_setup buts; struct blk_trace *old_bt, *bt = NULL; struct dentry *dir = NULL; - char b[BDEVNAME_SIZE]; int ret, i; - if (copy_from_user(&buts, arg, sizeof(buts))) - return -EFAULT; - - if (!buts.buf_size || !buts.buf_nr) + if (!buts->buf_size || !buts->buf_nr) return -EINVAL; - strcpy(buts.name, bdevname(bdev, b)); + strncpy(buts->name, name, BLKTRACE_BDEV_SIZE); + buts->name[BLKTRACE_BDEV_SIZE - 1] = '\0'; /* * some device names have larger paths - convert the slashes * to underscores for this to work as expected */ - for (i = 0; i < strlen(buts.name); i++) - if (buts.name[i] == '/') - buts.name[i] = '_'; - - if (copy_to_user(arg, &buts, sizeof(buts))) - return -EFAULT; + for (i = 0; i < strlen(buts->name); i++) + if (buts->name[i] == '/') + buts->name[i] = '_'; ret = -ENOMEM; bt = kzalloc(sizeof(*bt), GFP_KERNEL); @@ -318,13 +389,17 @@ static int blk_trace_setup(request_queue_t *q, struct block_device *bdev, if (!bt->sequence) goto err; + bt->msg_data = __alloc_percpu(BLK_TN_MAX_MSG); + if (!bt->msg_data) + goto err; + ret = -ENOENT; - dir = blk_create_tree(buts.name); + dir = blk_create_tree(buts->name); if (!dir) goto err; bt->dir = dir; - bt->dev = bdev->bd_dev; + bt->dev = dev; atomic_set(&bt->dropped, 0); ret = -EIO; @@ -332,21 +407,25 @@ static int blk_trace_setup(request_queue_t *q, struct block_device *bdev, if (!bt->dropped_file) goto err; - bt->rchan = relay_open("trace", dir, buts.buf_size, buts.buf_nr, &blk_relay_callbacks); + bt->msg_file = debugfs_create_file("msg", 0222, dir, bt, &blk_msg_fops); + if (!bt->msg_file) + goto err; + + bt->rchan = relay_open("trace", dir, buts->buf_size, + buts->buf_nr, &blk_relay_callbacks, bt); if (!bt->rchan) goto err; - bt->rchan->private_data = bt; - bt->act_mask = buts.act_mask; + bt->act_mask = buts->act_mask; if (!bt->act_mask) bt->act_mask = (u16) -1; - bt->start_lba = buts.start_lba; - bt->end_lba = buts.end_lba; + bt->start_lba = buts->start_lba; + bt->end_lba = buts->end_lba; if (!bt->end_lba) bt->end_lba = -1ULL; - bt->pid = buts.pid; + bt->pid = buts->pid; bt->trace_state = Blktrace_setup; ret = -EBUSY; @@ -361,10 +440,12 @@ err: if (dir) blk_remove_tree(dir); if (bt) { + if (bt->msg_file) + debugfs_remove(bt->msg_file); if (bt->dropped_file) debugfs_remove(bt->dropped_file); - if (bt->sequence) - free_percpu(bt->sequence); + free_percpu(bt->sequence); + free_percpu(bt->msg_data); if (bt->rchan) relay_close(bt->rchan); kfree(bt); @@ -372,7 +453,28 @@ err: return ret; } -static int blk_trace_startstop(request_queue_t *q, int start) +int blk_trace_setup(struct request_queue *q, char *name, dev_t dev, + char __user *arg) +{ + struct blk_user_trace_setup buts; + int ret; + + ret = copy_from_user(&buts, arg, sizeof(buts)); + if (ret) + return -EFAULT; + + ret = do_blk_trace_setup(q, name, dev, &buts); + if (ret) + return ret; + + if (copy_to_user(arg, &buts, sizeof(buts))) + return -EFAULT; + + return 0; +} +EXPORT_SYMBOL_GPL(blk_trace_setup); + +int blk_trace_startstop(struct request_queue *q, int start) { struct blk_trace *bt; int ret; @@ -391,6 +493,8 @@ static int blk_trace_startstop(request_queue_t *q, int start) blktrace_seq++; smp_mb(); bt->trace_state = Blktrace_running; + + trace_note_time(bt); ret = 0; } } else { @@ -403,6 +507,7 @@ static int blk_trace_startstop(request_queue_t *q, int start) return ret; } +EXPORT_SYMBOL_GPL(blk_trace_startstop); /** * blk_trace_ioctl: - handle the ioctls associated with tracing @@ -413,8 +518,9 @@ static int blk_trace_startstop(request_queue_t *q, int start) **/ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg) { - request_queue_t *q; + struct request_queue *q; int ret, start = 0; + char b[BDEVNAME_SIZE]; q = bdev_get_queue(bdev); if (!q) @@ -424,7 +530,8 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg) switch (cmd) { case BLKTRACESETUP: - ret = blk_trace_setup(q, bdev, arg); + bdevname(bdev, b); + ret = blk_trace_setup(q, b, bdev->bd_dev, arg); break; case BLKTRACESTART: start = 1; @@ -448,95 +555,10 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg) * @q: the request queue associated with the device * **/ -void blk_trace_shutdown(request_queue_t *q) +void blk_trace_shutdown(struct request_queue *q) { if (q->blk_trace) { blk_trace_startstop(q, 0); blk_trace_remove(q); } } - -/* - * Average offset over two calls to sched_clock() with a gettimeofday() - * in the middle - */ -static void blk_check_time(unsigned long long *t) -{ - unsigned long long a, b; - struct timeval tv; - - a = sched_clock(); - do_gettimeofday(&tv); - b = sched_clock(); - - *t = tv.tv_sec * 1000000000 + tv.tv_usec * 1000; - *t -= (a + b) / 2; -} - -static void blk_trace_check_cpu_time(void *data) -{ - unsigned long long *t; - int cpu = get_cpu(); - - t = &per_cpu(blk_trace_cpu_offset, cpu); - - /* - * Just call it twice, hopefully the second call will be cache hot - * and a little more precise - */ - blk_check_time(t); - blk_check_time(t); - - put_cpu(); -} - -/* - * Call blk_trace_check_cpu_time() on each CPU to calibrate our inter-CPU - * timings - */ -static void blk_trace_calibrate_offsets(void) -{ - unsigned long flags; - - smp_call_function(blk_trace_check_cpu_time, NULL, 1, 1); - local_irq_save(flags); - blk_trace_check_cpu_time(NULL); - local_irq_restore(flags); -} - -static void blk_trace_set_ht_offsets(void) -{ -#if defined(CONFIG_SCHED_SMT) - int cpu, i; - - /* - * now make sure HT siblings have the same time offset - */ - preempt_disable(); - for_each_online_cpu(cpu) { - unsigned long long *cpu_off, *sibling_off; - - for_each_cpu_mask(i, cpu_sibling_map[cpu]) { - if (i == cpu) - continue; - - cpu_off = &per_cpu(blk_trace_cpu_offset, cpu); - sibling_off = &per_cpu(blk_trace_cpu_offset, i); - *sibling_off = *cpu_off; - } - } - preempt_enable(); -#endif -} - -static __init int blk_trace_init(void) -{ - mutex_init(&blk_tree_mutex); - blk_trace_calibrate_offsets(); - blk_trace_set_ht_offsets(); - - return 0; -} - -module_init(blk_trace_init); -