Requirements
------------
-Currently, the kernel support patch isn't merged yet. The last posted patch
-is:
-
-http://marc.theaimsgroup.com/?l=linux-kernel&m=112487579802074&w=2
-
-and applies against 2.6.13-rc6-mm2. The patch is also included in the git
-repository you pulled to get this file:
+You need to be running a 2.6.13-mm1 kernel with the blk-trace patch
+included in this repository. If you forgot where you got it, the url is:
rsync://rsync.kernel.org/pub/scm/linux/kernel/git/axboe/blktrace.git
-Additionally, you need to apply the relayfs-read update patch also located
-in the kernel/ directory.
-
Usage
-----
--- /dev/null
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/blktrace.c linux-2.6.13-rc6-mm2/drivers/block/blktrace.c
+--- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/blktrace.c 1970-01-01 01:00:00.000000000 +0100
++++ linux-2.6.13-rc6-mm2/drivers/block/blktrace.c 2005-08-30 20:56:43.000000000 +0200
+@@ -0,0 +1,184 @@
++#include <linux/config.h>
++#include <linux/kernel.h>
++#include <linux/blkdev.h>
++#include <linux/blktrace.h>
++#include <linux/percpu.h>
++#include <linux/init.h>
++#include <asm/uaccess.h>
++
++static DEFINE_PER_CPU(unsigned long long, blk_trace_cpu_offset) = { 0, };
++
++static inline unsigned long long blk_trace_cpu_time(void)
++{
++ unsigned long long offset;
++
++ offset = sched_clock() - per_cpu(blk_trace_cpu_offset, get_cpu());
++ put_cpu();
++ return offset;
++}
++
++void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes,
++ int rw, u32 what, int error, int pdu_len, char *pdu_data)
++{
++ struct blk_io_trace t;
++ unsigned long flags;
++
++ if (rw & (1 << BIO_RW_BARRIER))
++ what |= BLK_TC_ACT(BLK_TC_BARRIER);
++ if (rw & (1 << BIO_RW_SYNC))
++ what |= BLK_TC_ACT(BLK_TC_SYNC);
++
++ if (rw & WRITE)
++ what |= BLK_TC_ACT(BLK_TC_WRITE);
++ else
++ what |= BLK_TC_ACT(BLK_TC_READ);
++
++ if (((bt->act_mask << BLK_TC_SHIFT) & what) == 0)
++ return;
++
++ t.magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION;
++ t.sequence = atomic_add_return(1, &bt->sequence);
++ t.time = blk_trace_cpu_time();
++ t.sector = sector;
++ t.bytes = bytes;
++ t.action = what;
++ t.pid = current->pid;
++ t.error = error;
++ t.pdu_len = pdu_len;
++
++ local_irq_save(flags);
++ __relay_write(bt->rchan, &t, sizeof(t));
++ if (pdu_len)
++ __relay_write(bt->rchan, pdu_data, pdu_len);
++ local_irq_restore(flags);
++}
++
++int blk_stop_trace(struct block_device *bdev)
++{
++ request_queue_t *q = bdev_get_queue(bdev);
++ struct blk_trace *bt = NULL;
++ int ret = -EINVAL;
++
++ if (!q)
++ return -ENXIO;
++
++ down(&bdev->bd_sem);
++
++ spin_lock_irq(q->queue_lock);
++ if (q->blk_trace) {
++ bt = q->blk_trace;
++ q->blk_trace = NULL;
++ ret = 0;
++ }
++ spin_unlock_irq(q->queue_lock);
++
++ up(&bdev->bd_sem);
++
++ if (bt) {
++ relay_close(bt->rchan);
++ kfree(bt);
++ }
++
++ return ret;
++}
++
++int blk_start_trace(struct block_device *bdev, char __user *arg)
++{
++ request_queue_t *q = bdev_get_queue(bdev);
++ struct blk_user_trace_setup buts;
++ struct blk_trace *bt;
++ char b[BDEVNAME_SIZE];
++ int ret = 0;
++
++ if (!q)
++ return -ENXIO;
++
++ if (copy_from_user(&buts, arg, sizeof(buts)))
++ return -EFAULT;
++
++ if (!buts.buf_size || !buts.buf_nr)
++ return -EINVAL;
++
++ strcpy(buts.name, bdevname(bdev, b));
++
++ if (copy_to_user(arg, &buts, sizeof(buts)))
++ return -EFAULT;
++
++ down(&bdev->bd_sem);
++ ret = -EBUSY;
++ if (q->blk_trace)
++ goto err;
++
++ ret = -ENOMEM;
++ bt = kmalloc(sizeof(*bt), GFP_KERNEL);
++ if (!bt)
++ goto err;
++
++ atomic_set(&bt->sequence, 0);
++
++ bt->rchan = relay_open(bdevname(bdev, b), NULL, buts.buf_size,
++ buts.buf_nr, NULL);
++ ret = -EIO;
++ if (!bt->rchan)
++ goto err;
++
++ bt->act_mask = buts.act_mask;
++ if (!bt->act_mask)
++ bt->act_mask = (u16) -1;
++
++ spin_lock_irq(q->queue_lock);
++ q->blk_trace = bt;
++ spin_unlock_irq(q->queue_lock);
++ ret = 0;
++err:
++ up(&bdev->bd_sem);
++ return ret;
++}
++
++static void blk_trace_check_cpu_time(void *data)
++{
++ unsigned long long a, b, *t;
++ struct timeval tv;
++ int cpu = get_cpu();
++
++ t = &per_cpu(blk_trace_cpu_offset, cpu);
++
++ a = sched_clock();
++ do_gettimeofday(&tv);
++ b = sched_clock();
++
++ *t = tv.tv_sec * 1000000000 + tv.tv_usec * 1000;
++ *t -= (a + b) / 2;
++ put_cpu();
++}
++
++static int blk_trace_calibrate_offsets(void)
++{
++ int cpus = num_online_cpus();
++ unsigned long flags;
++ long long *time_offset;
++
++ time_offset = kmalloc(cpus * sizeof(long long), GFP_KERNEL);
++ if (!time_offset) {
++ printk(KERN_ERR "blktrace: time offsets will be unreliable\n");
++ return -ENOMEM;
++ }
++
++ memset(time_offset, 0, cpus * sizeof(long long));
++
++ smp_call_function(blk_trace_check_cpu_time, time_offset, 1, 1);
++ local_irq_save(flags);
++ blk_trace_check_cpu_time(time_offset);
++ local_irq_restore(flags);
++
++ kfree(time_offset);
++ return 0;
++}
++
++static __init int blk_trace_init(void)
++{
++ return blk_trace_calibrate_offsets();
++}
++
++module_init(blk_trace_init);
++
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/elevator.c linux-2.6.13-rc6-mm2/drivers/block/elevator.c
+--- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/elevator.c 2005-08-30 20:55:03.000000000 +0200
++++ linux-2.6.13-rc6-mm2/drivers/block/elevator.c 2005-08-26 11:03:38.000000000 +0200
+@@ -34,6 +34,7 @@
+ #include <linux/slab.h>
+ #include <linux/init.h>
+ #include <linux/compiler.h>
++#include <linux/blktrace.h>
+
+ #include <asm/uaccess.h>
+
+@@ -371,6 +372,9 @@ struct request *elv_next_request(request
+ int ret;
+
+ while ((rq = __elv_next_request(q)) != NULL) {
++
++ blk_add_trace_rq(q, rq, BLK_TA_ISSUE);
++
+ /*
+ * just mark as started even if we don't start it, a request
+ * that has been delayed should not be passed by new incoming
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ioctl.c linux-2.6.13-rc6-mm2/drivers/block/ioctl.c
+--- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ioctl.c 2005-08-30 20:55:03.000000000 +0200
++++ linux-2.6.13-rc6-mm2/drivers/block/ioctl.c 2005-08-26 11:03:38.000000000 +0200
+@@ -4,6 +4,7 @@
+ #include <linux/backing-dev.h>
+ #include <linux/buffer_head.h>
+ #include <linux/smp_lock.h>
++#include <linux/blktrace.h>
+ #include <asm/uaccess.h>
+
+ static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user *arg)
+@@ -188,6 +189,10 @@ static int blkdev_locked_ioctl(struct fi
+ return put_ulong(arg, bdev->bd_inode->i_size >> 9);
+ case BLKGETSIZE64:
+ return put_u64(arg, bdev->bd_inode->i_size);
++ case BLKSTARTTRACE:
++ return blk_start_trace(bdev, (char __user *) arg);
++ case BLKSTOPTRACE:
++ return blk_stop_trace(bdev);
+ }
+ return -ENOIOCTLCMD;
+ }
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Kconfig linux-2.6.13-rc6-mm2/drivers/block/Kconfig
+--- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Kconfig 2005-08-30 20:55:03.000000000 +0200
++++ linux-2.6.13-rc6-mm2/drivers/block/Kconfig 2005-08-29 08:00:42.000000000 +0200
+@@ -419,6 +419,14 @@ config LBD
+ your machine, or if you want to have a raid or loopback device
+ bigger than 2TB. Otherwise say N.
+
++config BLK_DEV_IO_TRACE
++ bool "Support for tracing block io actions"
++ select RELAYFS_FS
++ help
++ Say Y here, if you want to be able to trace the block layer actions
++ on a given queue.
++
++
+ config CDROM_PKTCDVD
+ tristate "Packet writing on CD/DVD media"
+ depends on !UML
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c
+--- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c 2005-08-30 20:55:03.000000000 +0200
++++ linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c 2005-08-26 11:03:38.000000000 +0200
+@@ -29,6 +29,7 @@
+ #include <linux/swap.h>
+ #include <linux/writeback.h>
+ #include <linux/blkdev.h>
++#include <linux/blktrace.h>
+
+ /*
+ * for max sense size
+@@ -1625,6 +1626,12 @@ void blk_cleanup_queue(request_queue_t *
+ if (q->queue_tags)
+ __blk_queue_free_tags(q);
+
++ if (q->blk_trace) {
++ relay_close(q->blk_trace->rchan);
++ kfree(q->blk_trace);
++ q->blk_trace = NULL;
++ }
++
+ blk_queue_ordered(q, QUEUE_ORDERED_NONE);
+
+ kmem_cache_free(requestq_cachep, q);
+@@ -1971,6 +1978,8 @@ rq_starved:
+
+ rq_init(q, rq);
+ rq->rl = rl;
++
++ blk_add_trace_generic(q, bio, rw, BLK_TA_GETRQ);
+ out:
+ return rq;
+ }
+@@ -1999,6 +2008,8 @@ static struct request *get_request_wait(
+ if (!rq) {
+ struct io_context *ioc;
+
++ blk_add_trace_generic(q, bio, rw, BLK_TA_SLEEPRQ);
++
+ __generic_unplug_device(q);
+ spin_unlock_irq(q->queue_lock);
+ io_schedule();
+@@ -2052,6 +2063,8 @@ EXPORT_SYMBOL(blk_get_request);
+ */
+ void blk_requeue_request(request_queue_t *q, struct request *rq)
+ {
++ blk_add_trace_rq(q, rq, BLK_TA_REQUEUE);
++
+ if (blk_rq_tagged(rq))
+ blk_queue_end_tag(q, rq);
+
+@@ -2665,6 +2678,8 @@ static int __make_request(request_queue_
+ if (!q->back_merge_fn(q, req, bio))
+ break;
+
++ blk_add_trace_bio(q, bio, BLK_TA_BACKMERGE);
++
+ req->biotail->bi_next = bio;
+ req->biotail = bio;
+ req->nr_sectors = req->hard_nr_sectors += nr_sectors;
+@@ -2680,6 +2695,8 @@ static int __make_request(request_queue_
+ if (!q->front_merge_fn(q, req, bio))
+ break;
+
++ blk_add_trace_bio(q, bio, BLK_TA_FRONTMERGE);
++
+ bio->bi_next = req->bio;
+ req->bio = bio;
+
+@@ -2705,6 +2722,8 @@ static int __make_request(request_queue_
+ }
+
+ get_rq:
++ blk_add_trace_bio(q, bio, BLK_TA_QUEUE);
++
+ /*
+ * Grab a free request. This is might sleep but can not fail.
+ * Returns with the queue unlocked.
+@@ -2981,6 +3000,10 @@ end_io:
+ blk_partition_remap(bio);
+
+ ret = q->make_request_fn(q, bio);
++
++ if (ret)
++ blk_add_trace_bio(q, bio, BLK_TA_QUEUE);
++
+ } while (ret);
+ }
+
+@@ -3099,6 +3122,8 @@ static int __end_that_request_first(stru
+ int total_bytes, bio_nbytes, error, next_idx = 0;
+ struct bio *bio;
+
++ blk_add_trace_rq(req->q, req, BLK_TA_COMPLETE);
++
+ /*
+ * extend uptodate bool to allow < 0 value to be direct io error
+ */
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Makefile linux-2.6.13-rc6-mm2/drivers/block/Makefile
+--- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Makefile 2005-08-30 20:55:03.000000000 +0200
++++ linux-2.6.13-rc6-mm2/drivers/block/Makefile 2005-08-26 11:03:38.000000000 +0200
+@@ -45,3 +45,5 @@ obj-$(CONFIG_VIODASD) += viodasd.o
+ obj-$(CONFIG_BLK_DEV_SX8) += sx8.o
+ obj-$(CONFIG_BLK_DEV_UB) += ub.o
+
++obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o
++
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blkdev.h linux-2.6.13-rc6-mm2/include/linux/blkdev.h
+--- /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blkdev.h 2005-08-30 20:55:03.000000000 +0200
++++ linux-2.6.13-rc6-mm2/include/linux/blkdev.h 2005-08-26 11:03:38.000000000 +0200
+@@ -22,6 +22,7 @@ typedef struct request_queue request_que
+ struct elevator_queue;
+ typedef struct elevator_queue elevator_t;
+ struct request_pm_state;
++struct blk_trace;
+
+ #define BLKDEV_MIN_RQ 4
+ #define BLKDEV_MAX_RQ 128 /* Default maximum */
+@@ -412,6 +413,8 @@ struct request_queue
+ */
+ struct request *flush_rq;
+ unsigned char ordered;
++
++ struct blk_trace *blk_trace;
+ };
+
+ enum {
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blktrace.h linux-2.6.13-rc6-mm2/include/linux/blktrace.h
+--- /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blktrace.h 1970-01-01 01:00:00.000000000 +0100
++++ linux-2.6.13-rc6-mm2/include/linux/blktrace.h 2005-08-30 09:57:59.000000000 +0200
+@@ -0,0 +1,145 @@
++#ifndef BLKTRACE_H
++#define BLKTRACE_H
++
++#include <linux/config.h>
++#include <linux/blkdev.h>
++#include <linux/relayfs_fs.h>
++
++/*
++ * Trace categories
++ */
++enum {
++ BLK_TC_READ = 1 << 0, /* reads */
++ BLK_TC_WRITE = 1 << 1, /* writes */
++ BLK_TC_BARRIER = 1 << 2, /* barrier */
++ BLK_TC_SYNC = 1 << 3, /* barrier */
++ BLK_TC_QUEUE = 1 << 4, /* queueing/merging */
++ BLK_TC_REQUEUE = 1 << 5, /* requeueing */
++ BLK_TC_ISSUE = 1 << 6, /* issue */
++ BLK_TC_COMPLETE = 1 << 7, /* completions */
++ BLK_TC_FS = 1 << 8, /* fs requests */
++ BLK_TC_PC = 1 << 9, /* pc requests */
++
++ BLK_TC_END = 1 << 15, /* only 16-bits, reminder */
++};
++
++#define BLK_TC_SHIFT (16)
++#define BLK_TC_ACT(act) ((act) << BLK_TC_SHIFT)
++
++/*
++ * Basic trace actions
++ */
++enum {
++ __BLK_TA_QUEUE = 1, /* queued */
++ __BLK_TA_BACKMERGE, /* back merged to existing rq */
++ __BLK_TA_FRONTMERGE, /* front merge to existing rq */
++ __BLK_TA_GETRQ, /* allocated new request */
++ __BLK_TA_SLEEPRQ, /* sleeping on rq allocation */
++ __BLK_TA_REQUEUE, /* request requeued */
++ __BLK_TA_ISSUE, /* sent to driver */
++ __BLK_TA_COMPLETE, /* completed by driver */
++};
++
++/*
++ * Trace actions in full. Additionally, read or write is masked
++ */
++#define BLK_TA_QUEUE (__BLK_TA_QUEUE | BLK_TC_ACT(BLK_TC_QUEUE))
++#define BLK_TA_BACKMERGE (__BLK_TA_BACKMERGE | BLK_TC_ACT(BLK_TC_QUEUE))
++#define BLK_TA_FRONTMERGE (__BLK_TA_FRONTMERGE | BLK_TC_ACT(BLK_TC_QUEUE))
++#define BLK_TA_GETRQ (__BLK_TA_GETRQ | BLK_TC_ACT(BLK_TC_QUEUE))
++#define BLK_TA_SLEEPRQ (__BLK_TA_SLEEPRQ | BLK_TC_ACT(BLK_TC_QUEUE))
++#define BLK_TA_REQUEUE (__BLK_TA_REQUEUE | BLK_TC_ACT(BLK_TC_REQUEUE))
++#define BLK_TA_ISSUE (__BLK_TA_ISSUE | BLK_TC_ACT(BLK_TC_ISSUE))
++#define BLK_TA_COMPLETE (__BLK_TA_COMPLETE| BLK_TC_ACT(BLK_TC_COMPLETE))
++
++#define BLK_IO_TRACE_MAGIC 0x65617400
++#define BLK_IO_TRACE_VERSION 0x02
++
++/*
++ * The trace itself
++ */
++struct blk_io_trace {
++ u32 magic; /* MAGIC << 8 | version */
++ u32 sequence; /* event number */
++ u64 time; /* in microseconds */
++ u64 sector; /* disk offset */
++ u32 bytes; /* transfer length */
++ u32 action; /* what happened */
++ u32 pid; /* who did it */
++ u16 error; /* completion error */
++ u16 pdu_len; /* length of data after this trace */
++};
++
++struct blk_trace {
++ struct rchan *rchan;
++ atomic_t sequence;
++ u16 act_mask;
++};
++
++/*
++ * User setup structure passed with BLKSTARTTRACE
++ */
++struct blk_user_trace_setup {
++ char name[BDEVNAME_SIZE]; /* output */
++ u16 act_mask; /* input */
++ u32 buf_size; /* input */
++ u32 buf_nr; /* input */
++};
++
++#if defined(CONFIG_BLK_DEV_IO_TRACE)
++extern int blk_start_trace(struct block_device *, char __user *);
++extern int blk_stop_trace(struct block_device *);
++extern void __blk_add_trace(struct blk_trace *, sector_t, int, int, u32, int, int, char *);
++
++static inline void blk_add_trace_rq(struct request_queue *q, struct request *rq,
++ u32 what)
++{
++ struct blk_trace *bt = q->blk_trace;
++ int rw = rq->flags & 0x07;
++
++ if (likely(!bt))
++ return;
++
++ if (blk_pc_request(rq)) {
++ what |= BLK_TC_ACT(BLK_TC_PC);
++ __blk_add_trace(bt, 0, rq->data_len, rw, what, rq->errors, sizeof(rq->cmd), rq->cmd);
++ } else {
++ what |= BLK_TC_ACT(BLK_TC_FS);
++ __blk_add_trace(bt, rq->hard_sector, rq->hard_nr_sectors << 9, rw, what, rq->errors, 0, NULL);
++ }
++}
++
++static inline void blk_add_trace_bio(struct request_queue *q, struct bio *bio,
++ u32 what)
++{
++ struct blk_trace *bt = q->blk_trace;
++
++ if (likely(!bt))
++ return;
++
++ __blk_add_trace(bt, bio->bi_sector, bio->bi_size, bio->bi_rw, what, !bio_flagged(bio, BIO_UPTODATE), 0, NULL);
++}
++
++static inline void blk_add_trace_generic(struct request_queue *q,
++ struct bio *bio, int rw, u32 what)
++{
++ struct blk_trace *bt = q->blk_trace;
++
++ if (likely(!bt))
++ return;
++
++ if (bio)
++ blk_add_trace_bio(q, bio, what);
++ else
++ __blk_add_trace(bt, 0, 0, rw, what, 0, 0, NULL);
++}
++
++#else /* !CONFIG_BLK_DEV_IO_TRACE */
++#define blk_start_trace(bdev, arg) (-EINVAL)
++#define blk_stop_trace(bdev) (-EINVAL)
++#define blk_add_trace_rq(q, rq, what) do { } while (0)
++#define blk_add_trace_bio(q, rq, what) do { } while (0)
++#define blk_add_trace_generic(q, rq, rw, what) do { } while (0)
++#endif /* CONFIG_BLK_DEV_IO_TRACE */
++
++#endif
+diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/fs.h linux-2.6.13-rc6-mm2/include/linux/fs.h
+--- /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/fs.h 2005-08-30 20:55:03.000000000 +0200
++++ linux-2.6.13-rc6-mm2/include/linux/fs.h 2005-08-26 11:03:38.000000000 +0200
+@@ -196,6 +196,8 @@ extern int dir_notify_enable;
+ #define BLKBSZGET _IOR(0x12,112,size_t)
+ #define BLKBSZSET _IOW(0x12,113,size_t)
+ #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
++#define BLKSTARTTRACE _IOWR(0x12,115,struct blk_user_trace_setup)
++#define BLKSTOPTRACE _IO(0x12,116)
+
+ #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
+ #define FIBMAP _IO(0x00,1) /* bmap access */
+++ /dev/null
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Kconfig linux-2.6.13-rc6-mm2/drivers/block/Kconfig
---- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Kconfig 2005-08-24 13:17:28.000000000 +0200
-+++ linux-2.6.13-rc6-mm2/drivers/block/Kconfig 2005-08-24 11:52:14.000000000 +0200
-@@ -419,6 +419,14 @@ config LBD
- your machine, or if you want to have a raid or loopback device
- bigger than 2TB. Otherwise say N.
-
-+config BLK_DEV_IO_TRACE
-+ bool "Support for tracing block io actions"
-+ select RELAYFS_FS
-+ help
-+ Say Y here, if you want to be able to trace the block layer actions
-+ on a given queue.
-+
-+
- config CDROM_PKTCDVD
- tristate "Packet writing on CD/DVD media"
- depends on !UML
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Makefile linux-2.6.13-rc6-mm2/drivers/block/Makefile
---- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/Makefile 2005-08-07 20:18:56.000000000 +0200
-+++ linux-2.6.13-rc6-mm2/drivers/block/Makefile 2005-08-24 11:52:14.000000000 +0200
-@@ -45,3 +45,5 @@ obj-$(CONFIG_VIODASD) += viodasd.o
- obj-$(CONFIG_BLK_DEV_SX8) += sx8.o
- obj-$(CONFIG_BLK_DEV_UB) += ub.o
-
-+obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o
-+
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/blktrace.c linux-2.6.13-rc6-mm2/drivers/block/blktrace.c
---- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/blktrace.c 1970-01-01 01:00:00.000000000 +0100
-+++ linux-2.6.13-rc6-mm2/drivers/block/blktrace.c 2005-08-24 13:22:11.000000000 +0200
-@@ -0,0 +1,124 @@
-+#include <linux/config.h>
-+#include <linux/kernel.h>
-+#include <linux/blkdev.h>
-+#include <linux/blktrace.h>
-+#include <asm/uaccess.h>
-+
-+void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes,
-+ int rw, u32 what, int error, int pdu_len, char *pdu_data)
-+{
-+ struct blk_io_trace t;
-+ unsigned long flags;
-+
-+ if (rw & (1 << BIO_RW_BARRIER))
-+ what |= BLK_TC_ACT(BLK_TC_BARRIER);
-+ if (rw & (1 << BIO_RW_SYNC))
-+ what |= BLK_TC_ACT(BLK_TC_SYNC);
-+
-+ if (rw & WRITE)
-+ what |= BLK_TC_ACT(BLK_TC_WRITE);
-+ else
-+ what |= BLK_TC_ACT(BLK_TC_READ);
-+
-+ if (((bt->act_mask << BLK_TC_SHIFT) & what) == 0)
-+ return;
-+
-+ t.magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION;
-+ t.sequence = atomic_add_return(1, &bt->sequence);
-+ t.time = sched_clock();
-+ t.sector = sector;
-+ t.bytes = bytes;
-+ t.action = what;
-+ t.pid = current->pid;
-+ t.error = error;
-+ t.pdu_len = pdu_len;
-+
-+ local_irq_save(flags);
-+ __relay_write(bt->rchan, &t, sizeof(t));
-+ if (pdu_len)
-+ __relay_write(bt->rchan, pdu_data, pdu_len);
-+ local_irq_restore(flags);
-+}
-+
-+int blk_stop_trace(struct block_device *bdev)
-+{
-+ request_queue_t *q = bdev_get_queue(bdev);
-+ struct blk_trace *bt = NULL;
-+ int ret = -EINVAL;
-+
-+ if (!q)
-+ return -ENXIO;
-+
-+ down(&bdev->bd_sem);
-+
-+ spin_lock_irq(q->queue_lock);
-+ if (q->blk_trace) {
-+ bt = q->blk_trace;
-+ q->blk_trace = NULL;
-+ ret = 0;
-+ }
-+ spin_unlock_irq(q->queue_lock);
-+
-+ up(&bdev->bd_sem);
-+
-+ if (bt) {
-+ relay_close(bt->rchan);
-+ kfree(bt);
-+ }
-+
-+ return ret;
-+}
-+
-+int blk_start_trace(struct block_device *bdev, char __user *arg)
-+{
-+ request_queue_t *q = bdev_get_queue(bdev);
-+ struct blk_user_trace_setup buts;
-+ struct blk_trace *bt;
-+ char b[BDEVNAME_SIZE];
-+ int ret = 0;
-+
-+ if (!q)
-+ return -ENXIO;
-+
-+ if (copy_from_user(&buts, arg, sizeof(buts)))
-+ return -EFAULT;
-+
-+ if (!buts.buf_size || !buts.buf_nr)
-+ return -EINVAL;
-+
-+ strcpy(buts.name, bdevname(bdev, b));
-+
-+ if (copy_to_user(arg, &buts, sizeof(buts)))
-+ return -EFAULT;
-+
-+ down(&bdev->bd_sem);
-+ ret = -EBUSY;
-+ if (q->blk_trace)
-+ goto err;
-+
-+ ret = -ENOMEM;
-+ bt = kmalloc(sizeof(*bt), GFP_KERNEL);
-+ if (!bt)
-+ goto err;
-+
-+ atomic_set(&bt->sequence, 0);
-+
-+ bt->rchan = relay_open(bdevname(bdev, b), NULL, buts.buf_size,
-+ buts.buf_nr, NULL);
-+ ret = -EIO;
-+ if (!bt->rchan)
-+ goto err;
-+
-+ bt->act_mask = buts.act_mask;
-+ if (!bt->act_mask)
-+ bt->act_mask = (u16) -1;
-+
-+ spin_lock_irq(q->queue_lock);
-+ q->blk_trace = bt;
-+ spin_unlock_irq(q->queue_lock);
-+ ret = 0;
-+err:
-+ up(&bdev->bd_sem);
-+ return ret;
-+}
-+
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/elevator.c linux-2.6.13-rc6-mm2/drivers/block/elevator.c
---- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/elevator.c 2005-08-07 20:18:56.000000000 +0200
-+++ linux-2.6.13-rc6-mm2/drivers/block/elevator.c 2005-08-24 11:52:14.000000000 +0200
-@@ -34,6 +34,7 @@
- #include <linux/slab.h>
- #include <linux/init.h>
- #include <linux/compiler.h>
-+#include <linux/blktrace.h>
-
- #include <asm/uaccess.h>
-
-@@ -371,6 +372,9 @@ struct request *elv_next_request(request
- int ret;
-
- while ((rq = __elv_next_request(q)) != NULL) {
-+
-+ blk_add_trace_rq(q, rq, BLK_TA_ISSUE);
-+
- /*
- * just mark as started even if we don't start it, a request
- * that has been delayed should not be passed by new incoming
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ioctl.c linux-2.6.13-rc6-mm2/drivers/block/ioctl.c
---- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ioctl.c 2005-08-07 20:18:56.000000000 +0200
-+++ linux-2.6.13-rc6-mm2/drivers/block/ioctl.c 2005-08-24 11:52:14.000000000 +0200
-@@ -4,6 +4,7 @@
- #include <linux/backing-dev.h>
- #include <linux/buffer_head.h>
- #include <linux/smp_lock.h>
-+#include <linux/blktrace.h>
- #include <asm/uaccess.h>
-
- static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user *arg)
-@@ -188,6 +189,10 @@ static int blkdev_locked_ioctl(struct fi
- return put_ulong(arg, bdev->bd_inode->i_size >> 9);
- case BLKGETSIZE64:
- return put_u64(arg, bdev->bd_inode->i_size);
-+ case BLKSTARTTRACE:
-+ return blk_start_trace(bdev, (char __user *) arg);
-+ case BLKSTOPTRACE:
-+ return blk_stop_trace(bdev);
- }
- return -ENOIOCTLCMD;
- }
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c
---- /opt/kernel/linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c 2005-08-24 13:17:28.000000000 +0200
-+++ linux-2.6.13-rc6-mm2/drivers/block/ll_rw_blk.c 2005-08-24 11:52:14.000000000 +0200
-@@ -29,6 +29,7 @@
- #include <linux/swap.h>
- #include <linux/writeback.h>
- #include <linux/blkdev.h>
-+#include <linux/blktrace.h>
-
- /*
- * for max sense size
-@@ -1625,6 +1626,12 @@ void blk_cleanup_queue(request_queue_t *
- if (q->queue_tags)
- __blk_queue_free_tags(q);
-
-+ if (q->blk_trace) {
-+ relay_close(q->blk_trace->rchan);
-+ kfree(q->blk_trace);
-+ q->blk_trace = NULL;
-+ }
-+
- blk_queue_ordered(q, QUEUE_ORDERED_NONE);
-
- kmem_cache_free(requestq_cachep, q);
-@@ -1971,6 +1978,8 @@ rq_starved:
-
- rq_init(q, rq);
- rq->rl = rl;
-+
-+ blk_add_trace_generic(q, bio, rw, BLK_TA_GETRQ);
- out:
- return rq;
- }
-@@ -1999,6 +2008,8 @@ static struct request *get_request_wait(
- if (!rq) {
- struct io_context *ioc;
-
-+ blk_add_trace_generic(q, bio, rw, BLK_TA_SLEEPRQ);
-+
- __generic_unplug_device(q);
- spin_unlock_irq(q->queue_lock);
- io_schedule();
-@@ -2052,6 +2063,8 @@ EXPORT_SYMBOL(blk_get_request);
- */
- void blk_requeue_request(request_queue_t *q, struct request *rq)
- {
-+ blk_add_trace_rq(q, rq, BLK_TA_REQUEUE);
-+
- if (blk_rq_tagged(rq))
- blk_queue_end_tag(q, rq);
-
-@@ -2665,6 +2678,8 @@ static int __make_request(request_queue_
- if (!q->back_merge_fn(q, req, bio))
- break;
-
-+ blk_add_trace_bio(q, bio, BLK_TA_BACKMERGE);
-+
- req->biotail->bi_next = bio;
- req->biotail = bio;
- req->nr_sectors = req->hard_nr_sectors += nr_sectors;
-@@ -2680,6 +2695,8 @@ static int __make_request(request_queue_
- if (!q->front_merge_fn(q, req, bio))
- break;
-
-+ blk_add_trace_bio(q, bio, BLK_TA_FRONTMERGE);
-+
- bio->bi_next = req->bio;
- req->bio = bio;
-
-@@ -2705,6 +2722,8 @@ static int __make_request(request_queue_
- }
-
- get_rq:
-+ blk_add_trace_bio(q, bio, BLK_TA_QUEUE);
-+
- /*
- * Grab a free request. This is might sleep but can not fail.
- * Returns with the queue unlocked.
-@@ -2981,6 +3000,10 @@ end_io:
- blk_partition_remap(bio);
-
- ret = q->make_request_fn(q, bio);
-+
-+ if (ret)
-+ blk_add_trace_bio(q, bio, BLK_TA_QUEUE);
-+
- } while (ret);
- }
-
-@@ -3099,6 +3122,8 @@ static int __end_that_request_first(stru
- int total_bytes, bio_nbytes, error, next_idx = 0;
- struct bio *bio;
-
-+ blk_add_trace_rq(req->q, req, BLK_TA_COMPLETE);
-+
- /*
- * extend uptodate bool to allow < 0 value to be direct io error
- */
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blkdev.h linux-2.6.13-rc6-mm2/include/linux/blkdev.h
---- /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blkdev.h 2005-08-24 13:17:35.000000000 +0200
-+++ linux-2.6.13-rc6-mm2/include/linux/blkdev.h 2005-08-24 11:52:14.000000000 +0200
-@@ -22,6 +22,7 @@ typedef struct request_queue request_que
- struct elevator_queue;
- typedef struct elevator_queue elevator_t;
- struct request_pm_state;
-+struct blk_trace;
-
- #define BLKDEV_MIN_RQ 4
- #define BLKDEV_MAX_RQ 128 /* Default maximum */
-@@ -412,6 +413,8 @@ struct request_queue
- */
- struct request *flush_rq;
- unsigned char ordered;
-+
-+ struct blk_trace *blk_trace;
- };
-
- enum {
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blktrace.h linux-2.6.13-rc6-mm2/include/linux/blktrace.h
---- /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/blktrace.h 1970-01-01 01:00:00.000000000 +0100
-+++ linux-2.6.13-rc6-mm2/include/linux/blktrace.h 2005-08-24 13:22:24.000000000 +0200
-@@ -0,0 +1,145 @@
-+#ifndef BLKTRACE_H
-+#define BLKTRACE_H
-+
-+#include <linux/config.h>
-+#include <linux/blkdev.h>
-+#include <linux/relayfs_fs.h>
-+
-+/*
-+ * Trace categories
-+ */
-+enum {
-+ BLK_TC_READ = 1 << 0, /* reads */
-+ BLK_TC_WRITE = 1 << 1, /* writes */
-+ BLK_TC_BARRIER = 1 << 2, /* barrier */
-+ BLK_TC_SYNC = 1 << 3, /* barrier */
-+ BLK_TC_QUEUE = 1 << 4, /* queueing/merging */
-+ BLK_TC_REQUEUE = 1 << 5, /* requeueing */
-+ BLK_TC_ISSUE = 1 << 6, /* issue */
-+ BLK_TC_COMPLETE = 1 << 7, /* completions */
-+ BLK_TC_FS = 1 << 8, /* fs requests */
-+ BLK_TC_PC = 1 << 9, /* pc requests */
-+
-+ BLK_TC_END = 1 << 15, /* only 16-bits, reminder */
-+};
-+
-+#define BLK_TC_SHIFT (16)
-+#define BLK_TC_ACT(act) ((act) << BLK_TC_SHIFT)
-+
-+/*
-+ * Basic trace actions
-+ */
-+enum {
-+ __BLK_TA_QUEUE = 1, /* queued */
-+ __BLK_TA_BACKMERGE, /* back merged to existing rq */
-+ __BLK_TA_FRONTMERGE, /* front merge to existing rq */
-+ __BLK_TA_GETRQ, /* allocated new request */
-+ __BLK_TA_SLEEPRQ, /* sleeping on rq allocation */
-+ __BLK_TA_REQUEUE, /* request requeued */
-+ __BLK_TA_ISSUE, /* sent to driver */
-+ __BLK_TA_COMPLETE, /* completed by driver */
-+};
-+
-+/*
-+ * Trace actions in full. Additionally, read or write is masked
-+ */
-+#define BLK_TA_QUEUE (__BLK_TA_QUEUE | BLK_TC_ACT(BLK_TC_QUEUE))
-+#define BLK_TA_BACKMERGE (__BLK_TA_BACKMERGE | BLK_TC_ACT(BLK_TC_QUEUE))
-+#define BLK_TA_FRONTMERGE (__BLK_TA_FRONTMERGE | BLK_TC_ACT(BLK_TC_QUEUE))
-+#define BLK_TA_GETRQ (__BLK_TA_GETRQ | BLK_TC_ACT(BLK_TC_QUEUE))
-+#define BLK_TA_SLEEPRQ (__BLK_TA_SLEEPRQ | BLK_TC_ACT(BLK_TC_QUEUE))
-+#define BLK_TA_REQUEUE (__BLK_TA_REQUEUE | BLK_TC_ACT(BLK_TC_REQUEUE))
-+#define BLK_TA_ISSUE (__BLK_TA_ISSUE | BLK_TC_ACT(BLK_TC_ISSUE))
-+#define BLK_TA_COMPLETE (__BLK_TA_COMPLETE| BLK_TC_ACT(BLK_TC_COMPLETE))
-+
-+#define BLK_IO_TRACE_MAGIC 0x65617400
-+#define BLK_IO_TRACE_VERSION 0x02
-+
-+/*
-+ * The trace itself
-+ */
-+struct blk_io_trace {
-+ u32 magic; /* MAGIC << 8 | version */
-+ u32 sequence; /* event number */
-+ u64 time; /* in microseconds */
-+ u64 sector; /* disk offset */
-+ u32 bytes; /* transfer length */
-+ u32 action; /* what happened */
-+ u32 pid; /* who did it */
-+ u16 error; /* completion error */
-+ u16 pdu_len; /* length of data after this trace */
-+};
-+
-+struct blk_trace {
-+ struct rchan *rchan;
-+ atomic_t sequence;
-+ u16 act_mask;
-+};
-+
-+/*
-+ * User setup structure passed with BLKSTARTTRACE
-+ */
-+struct blk_user_trace_setup {
-+ char name[BDEVNAME_SIZE]; /* output */
-+ u16 act_mask; /* input */
-+ u32 buf_size; /* input */
-+ u32 buf_nr; /* input */
-+};
-+
-+#if defined(CONFIG_BLK_DEV_IO_TRACE)
-+extern int blk_start_trace(struct block_device *, char __user *);
-+extern int blk_stop_trace(struct block_device *);
-+extern void __blk_add_trace(struct blk_trace *, sector_t, int, int, u32, int, int, char *);
-+
-+static inline void blk_add_trace_rq(struct request_queue *q, struct request *rq,
-+ u32 what)
-+{
-+ struct blk_trace *bt = q->blk_trace;
-+ int rw = rq->flags & 0x07;
-+
-+ if (likely(!bt))
-+ return;
-+
-+ if (blk_pc_request(rq)) {
-+ what |= BLK_TC_ACT(BLK_TC_PC);
-+ __blk_add_trace(bt, 0, rq->data_len, rw, what, rq->errors, sizeof(rq->cmd), rq->cmd);
-+ } else {
-+ what |= BLK_TC_ACT(BLK_TC_FS);
-+ __blk_add_trace(bt, rq->hard_sector, rq->hard_nr_sectors << 9, rw, what, rq->errors, 0, NULL);
-+ }
-+}
-+
-+static inline void blk_add_trace_bio(struct request_queue *q, struct bio *bio,
-+ u32 what)
-+{
-+ struct blk_trace *bt = q->blk_trace;
-+
-+ if (likely(!bt))
-+ return;
-+
-+ __blk_add_trace(bt, bio->bi_sector, bio->bi_size, bio->bi_rw, what, !bio_flagged(bio, BIO_UPTODATE), 0, NULL);
-+}
-+
-+static inline void blk_add_trace_generic(struct request_queue *q,
-+ struct bio *bio, int rw, u32 what)
-+{
-+ struct blk_trace *bt = q->blk_trace;
-+
-+ if (likely(!bt))
-+ return;
-+
-+ if (bio)
-+ blk_add_trace_bio(q, bio, what);
-+ else
-+ __blk_add_trace(bt, 0, 0, rw, what, 0, 0, NULL);
-+}
-+
-+#else /* !CONFIG_BLK_DEV_IO_TRACE */
-+#define blk_start_trace(bdev, arg) (-EINVAL)
-+#define blk_stop_trace(bdev) (-EINVAL)
-+#define blk_add_trace_rq(q, rq, what) do { } while (0)
-+#define blk_add_trace_bio(q, rq, what) do { } while (0)
-+#define blk_add_trace_generic(q, rq, rw, what) do { } while (0)
-+#endif /* CONFIG_BLK_DEV_IO_TRACE */
-+
-+#endif
-diff -urpN -X linux-2.6.13-rc6-mm2/Documentation/dontdiff /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/fs.h linux-2.6.13-rc6-mm2/include/linux/fs.h
---- /opt/kernel/linux-2.6.13-rc6-mm2/include/linux/fs.h 2005-08-24 13:17:35.000000000 +0200
-+++ linux-2.6.13-rc6-mm2/include/linux/fs.h 2005-08-24 11:52:14.000000000 +0200
-@@ -196,6 +196,8 @@ extern int dir_notify_enable;
- #define BLKBSZGET _IOR(0x12,112,size_t)
- #define BLKBSZSET _IOW(0x12,113,size_t)
- #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
-+#define BLKSTARTTRACE _IOWR(0x12,115,struct blk_user_trace_setup)
-+#define BLKSTOPTRACE _IO(0x12,116)
-
- #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
- #define FIBMAP _IO(0x00,1) /* bmap access */
+++ /dev/null
-Hi,
-
-The current relayfs read implementation works fine, but was designed
-to be used mainly for 'draining' the buffer after a tracing run. It
-turns out that people really want to be able to read from the buffer
-during a live trace, for example the blktrace application submitted
-recently:
-
-http://marc.theaimsgroup.com/?l=linux-kernel&m=112480046405961&w=2
-
-Here's an improved read implementation for relayfs which allows for
-that.
-
-This version has been tested pretty thoroughly, using both the
-blktrace application and a new example I added to the relay-apps
-tarball called 'readtest' which is basically a unit test for the read
-functionality. All the tests I've come up with have passed and it
-looks pretty solid at this point. Here's a link to the test code:
-
-http://prdownloads.sourceforge.net/relayfs/relay-apps-0.8.tar.gz?download
-
-Andrew, please apply.
-
-Thanks,
-
-Tom
-
-
-Signed-off-by: Tom Zanussi <zanussi@us.ibm.com>
-
-diff -urpN -X dontdiff linux-2.6.13-rc6-mm2/Documentation/filesystems/relayfs.txt linux-2.6.13-rc6-mm2-cur/Documentation/filesystems/relayfs.txt
---- linux-2.6.13-rc6-mm2/Documentation/filesystems/relayfs.txt 2005-08-25 19:28:59.000000000 -0500
-+++ linux-2.6.13-rc6-mm2-cur/Documentation/filesystems/relayfs.txt 2005-08-25 17:07:48.000000000 -0500
-@@ -82,10 +82,15 @@ mmap() results in channel buffer being
- memory space. Note that you can't do a partial mmap - you must
- map the entire file, which is NRBUF * SUBBUFSIZE.
-
--read() read the contents of a channel buffer. If there are active
-- channel writers, results may be unpredictable - users should
-- make sure that all logging to the channel has ended before
-- using read().
-+read() read the contents of a channel buffer. The bytes read are
-+ 'consumed' by the reader i.e. they won't be available again
-+ to subsequent reads. If the channel is being used in
-+ no-overwrite mode (the default), it can be read at any time
-+ even if there's an active kernel writer. If the channel is
-+ being used in overwrite mode and there are active channel
-+ writers, results may be unpredictable - users should make
-+ sure that all logging to the channel has ended before using
-+ read() with overwrite mode.
-
- poll() POLLIN/POLLRDNORM/POLLERR supported. User applications are
- notified when sub-buffer boundaries are crossed.
-@@ -256,8 +261,8 @@ consulted.
-
- The default subbuf_start() implementation, used if the client doesn't
- define any callbacks, or doesn't define the subbuf_start() callback,
--implements the simplest possible 'overwrite' mode i.e. it does nothing
--but return 1.
-+implements the simplest possible 'no-overwrite' mode i.e. it does
-+nothing but return 0.
-
- Header information can be reserved at the beginning of each sub-buffer
- by calling the subbuf_start_reserve() helper function from within the
-diff -urpN -X dontdiff linux-2.6.13-rc6-mm2/fs/relayfs/inode.c linux-2.6.13-rc6-mm2-cur/fs/relayfs/inode.c
---- linux-2.6.13-rc6-mm2/fs/relayfs/inode.c 2005-08-25 19:29:02.000000000 -0500
-+++ linux-2.6.13-rc6-mm2-cur/fs/relayfs/inode.c 2005-08-25 18:21:31.000000000 -0500
-@@ -295,101 +295,143 @@ static int relayfs_release(struct inode
- }
-
- /**
-- * relayfs_read_start - find the first available byte to read
-- *
-- * If the read_pos is in the middle of padding, return the
-- * position of the first actually available byte, otherwise
-- * return the original value.
-+ * relayfs_read_consume - update the consumed count for the buffer
- */
--static inline size_t relayfs_read_start(size_t read_pos,
-- size_t avail,
-- size_t start_subbuf,
-- struct rchan_buf *buf)
-+static void relayfs_read_consume(struct rchan_buf *buf,
-+ size_t read_pos,
-+ size_t bytes_consumed)
- {
-- size_t read_subbuf, adj_read_subbuf;
-- size_t padding, padding_start, padding_end;
- size_t subbuf_size = buf->chan->subbuf_size;
- size_t n_subbufs = buf->chan->n_subbufs;
-+ size_t read_subbuf;
-
-- read_subbuf = read_pos / subbuf_size;
-- adj_read_subbuf = (read_subbuf + start_subbuf) % n_subbufs;
-+ if (buf->bytes_consumed + bytes_consumed > subbuf_size) {
-+ relay_subbufs_consumed(buf->chan, buf->cpu, 1);
-+ buf->bytes_consumed = 0;
-+ }
-
-- if ((read_subbuf + 1) * subbuf_size <= avail) {
-- padding = buf->padding[adj_read_subbuf];
-- padding_start = (read_subbuf + 1) * subbuf_size - padding;
-- padding_end = (read_subbuf + 1) * subbuf_size;
-- if (read_pos >= padding_start && read_pos < padding_end) {
-- read_subbuf = (read_subbuf + 1) % n_subbufs;
-- read_pos = read_subbuf * subbuf_size;
-+ buf->bytes_consumed += bytes_consumed;
-+ read_subbuf = read_pos / buf->chan->subbuf_size;
-+ if (buf->bytes_consumed + buf->padding[read_subbuf] == subbuf_size) {
-+ if ((read_subbuf == buf->subbufs_produced % n_subbufs) &&
-+ (buf->offset == subbuf_size))
-+ return;
-+ relay_subbufs_consumed(buf->chan, buf->cpu, 1);
-+ buf->bytes_consumed = 0;
-+ }
-+}
-+
-+/**
-+ * relayfs_read_avail - boolean, are there unconsumed bytes available?
-+ */
-+static int relayfs_read_avail(struct rchan_buf *buf, size_t read_pos)
-+{
-+ size_t bytes_produced, bytes_consumed, write_offset;
-+ size_t subbuf_size = buf->chan->subbuf_size;
-+ size_t n_subbufs = buf->chan->n_subbufs;
-+ size_t produced = buf->subbufs_produced % n_subbufs;
-+ size_t consumed = buf->subbufs_consumed % n_subbufs;
-+
-+ write_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
-+
-+ if (consumed > produced) {
-+ if ((produced > n_subbufs) &&
-+ (produced + n_subbufs - consumed <= n_subbufs))
-+ produced += n_subbufs;
-+ } else if (consumed == produced) {
-+ if (buf->offset > subbuf_size) {
-+ produced += n_subbufs;
-+ if (buf->subbufs_produced == buf->subbufs_consumed)
-+ consumed += n_subbufs;
- }
- }
-
-- return read_pos;
-+ if (buf->offset > subbuf_size)
-+ bytes_produced = (produced - 1) * subbuf_size + write_offset;
-+ else
-+ bytes_produced = produced * subbuf_size + write_offset;
-+ bytes_consumed = consumed * subbuf_size + buf->bytes_consumed;
-+
-+ if (bytes_produced == bytes_consumed)
-+ return 0;
-+
-+ relayfs_read_consume(buf, read_pos, 0);
-+
-+ return 1;
- }
-
- /**
-- * relayfs_read_end - return the end of available bytes to read
-- *
-- * If the read_pos is in the middle of a full sub-buffer, return
-- * the padding-adjusted end of that sub-buffer, otherwise return
-- * the position after the last byte written to the buffer. At
-- * most, 1 sub-buffer can be read at a time.
-- *
-+ * relayfs_read_subbuf_avail - return bytes available in sub-buffer
- */
--static inline size_t relayfs_read_end(size_t read_pos,
-- size_t avail,
-- size_t start_subbuf,
-- struct rchan_buf *buf)
-+static size_t relayfs_read_subbuf_avail(size_t read_pos,
-+ struct rchan_buf *buf)
- {
-- size_t padding, read_endpos, buf_offset;
-- size_t read_subbuf, adj_read_subbuf;
-+ size_t padding, avail = 0;
-+ size_t read_subbuf, read_offset, write_subbuf, write_offset;
- size_t subbuf_size = buf->chan->subbuf_size;
-- size_t n_subbufs = buf->chan->n_subbufs;
-
-- buf_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
-+ write_subbuf = (buf->data - buf->start) / subbuf_size;
-+ write_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
- read_subbuf = read_pos / subbuf_size;
-- adj_read_subbuf = (read_subbuf + start_subbuf) % n_subbufs;
-+ read_offset = read_pos % subbuf_size;
-+ padding = buf->padding[read_subbuf];
-
-- if ((read_subbuf + 1) * subbuf_size <= avail) {
-- padding = buf->padding[adj_read_subbuf];
-- read_endpos = (read_subbuf + 1) * subbuf_size - padding;
-+ if (read_subbuf == write_subbuf) {
-+ if (read_offset + padding < write_offset)
-+ avail = write_offset - (read_offset + padding);
- } else
-- read_endpos = read_subbuf * subbuf_size + buf_offset;
-+ avail = (subbuf_size - padding) - read_offset;
-
-- return read_endpos;
-+ return avail;
- }
-
- /**
-- * relayfs_read_avail - return total available along with buffer start
-- *
-- * Because buffers are circular, the 'beginning' of the buffer
-- * depends on where the buffer was last written. If the writer
-- * has cycled around the buffer, the beginning is defined to be
-- * the beginning of the sub-buffer following the last sub-buffer
-- * written to, otherwise it's the beginning of sub-buffer 0.
-+ * relayfs_read_start_pos - find the first available byte to read
- *
-+ * If the read_pos is in the middle of padding, return the
-+ * position of the first actually available byte, otherwise
-+ * return the original value.
- */
--static inline size_t relayfs_read_avail(struct rchan_buf *buf,
-- size_t *start_subbuf)
-+static size_t relayfs_read_start_pos(size_t read_pos,
-+ struct rchan_buf *buf)
- {
-- size_t avail, complete_subbufs, cur_subbuf, buf_offset;
-+ size_t read_subbuf, padding, padding_start, padding_end;
- size_t subbuf_size = buf->chan->subbuf_size;
- size_t n_subbufs = buf->chan->n_subbufs;
-+
-+ read_subbuf = read_pos / subbuf_size;
-+ padding = buf->padding[read_subbuf];
-+ padding_start = (read_subbuf + 1) * subbuf_size - padding;
-+ padding_end = (read_subbuf + 1) * subbuf_size;
-+ if (read_pos >= padding_start && read_pos < padding_end) {
-+ read_subbuf = (read_subbuf + 1) % n_subbufs;
-+ read_pos = read_subbuf * subbuf_size;
-+ }
-
-- buf_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
-+ return read_pos;
-+}
-
-- if (buf->subbufs_produced >= n_subbufs) {
-- complete_subbufs = n_subbufs - 1;
-- cur_subbuf = (buf->data - buf->start) / subbuf_size;
-- *start_subbuf = (cur_subbuf + 1) % n_subbufs;
-- } else {
-- complete_subbufs = buf->subbufs_produced;
-- *start_subbuf = 0;
-- }
-+/**
-+ * relayfs_read_end_pos - return the new read position
-+ */
-+static size_t relayfs_read_end_pos(struct rchan_buf *buf,
-+ size_t read_pos,
-+ size_t count)
-+{
-+ size_t read_subbuf, padding, end_pos;
-+ size_t subbuf_size = buf->chan->subbuf_size;
-+ size_t n_subbufs = buf->chan->n_subbufs;
-
-- avail = complete_subbufs * subbuf_size + buf_offset;
-+ read_subbuf = read_pos / subbuf_size;
-+ padding = buf->padding[read_subbuf];
-+ if (read_pos % subbuf_size + count + padding == subbuf_size)
-+ end_pos = (read_subbuf + 1) * subbuf_size;
-+ else
-+ end_pos = read_pos + count;
-+ if (end_pos >= subbuf_size * n_subbufs)
-+ end_pos = 0;
-
-- return avail;
-+ return end_pos;
- }
-
- /**
-@@ -401,13 +443,6 @@ static inline size_t relayfs_read_avail(
- *
- * Reads count bytes or the number of bytes available in the
- * current sub-buffer being read, whichever is smaller.
-- *
-- * NOTE: The results of reading a relayfs file which is currently
-- * being written to are undefined. This is because the buffer is
-- * circular and an active writer in the kernel could be
-- * overwriting the data currently being read. Therefore read()
-- * is mainly useful for reading the contents of a buffer after
-- * logging has completed.
- */
- static ssize_t relayfs_read(struct file *filp,
- char __user *buffer,
-@@ -416,33 +451,30 @@ static ssize_t relayfs_read(struct file
- {
- struct inode *inode = filp->f_dentry->d_inode;
- struct rchan_buf *buf = RELAYFS_I(inode)->buf;
-- size_t read_start, read_end, avail, start_subbuf;
-- size_t buf_size = buf->chan->subbuf_size * buf->chan->n_subbufs;
-+ size_t read_start, avail;
-+ ssize_t ret = 0;
- void *from;
-
-- avail = relayfs_read_avail(buf, &start_subbuf);
-- if (*ppos >= avail)
-- return 0;
--
-- read_start = relayfs_read_start(*ppos, avail, start_subbuf, buf);
-- if (read_start == 0 && *ppos)
-- return 0;
--
-- read_end = relayfs_read_end(read_start, avail, start_subbuf, buf);
-- if (read_end == read_start)
-- return 0;
--
-- from = buf->start + start_subbuf * buf->chan->subbuf_size + read_start;
-- if (from >= buf->start + buf_size)
-- from -= buf_size;
--
-- count = min(count, read_end - read_start);
-- if (copy_to_user(buffer, from, count))
-- return -EFAULT;
--
-- *ppos = read_start + count;
--
-- return count;
-+ down(&inode->i_sem);
-+ if(!relayfs_read_avail(buf, *ppos))
-+ goto out;
-+
-+ read_start = relayfs_read_start_pos(*ppos, buf);
-+ avail = relayfs_read_subbuf_avail(read_start, buf);
-+ if (!avail)
-+ goto out;
-+
-+ from = buf->start + read_start;
-+ ret = count = min(count, avail);
-+ if (copy_to_user(buffer, from, count)) {
-+ ret = -EFAULT;
-+ goto out;
-+ }
-+ relayfs_read_consume(buf, read_start, count);
-+ *ppos = relayfs_read_end_pos(buf, read_start, count);
-+out:
-+ up(&inode->i_sem);
-+ return ret;
- }
-
- /**
-@@ -481,6 +513,7 @@ struct file_operations relayfs_file_oper
- .poll = relayfs_poll,
- .mmap = relayfs_mmap,
- .read = relayfs_read,
-+ .llseek = no_llseek,
- .release = relayfs_release,
- };
-
-diff -urpN -X dontdiff linux-2.6.13-rc6-mm2/fs/relayfs/relay.c linux-2.6.13-rc6-mm2-cur/fs/relayfs/relay.c
---- linux-2.6.13-rc6-mm2/fs/relayfs/relay.c 2005-08-25 19:29:02.000000000 -0500
-+++ linux-2.6.13-rc6-mm2-cur/fs/relayfs/relay.c 2005-08-25 21:12:29.000000000 -0500
-@@ -58,6 +58,9 @@ static int subbuf_start_default_callback
- void *prev_subbuf,
- size_t prev_padding)
- {
-+ if (relay_buf_full(buf))
-+ return 0;
-+
- return 1;
- }
-
-@@ -120,6 +123,7 @@ static inline void __relay_reset(struct
-
- buf->subbufs_produced = 0;
- buf->subbufs_consumed = 0;
-+ buf->bytes_consumed = 0;
- buf->finalized = 0;
- buf->data = buf->start;
- buf->offset = 0;
-@@ -262,6 +266,7 @@ struct rchan *relay_open(const char *bas
- for_each_online_cpu(i) {
- sprintf(tmpname, "%s%d", base_filename, i);
- chan->buf[i] = relay_open_buf(chan, tmpname, parent);
-+ chan->buf[i]->cpu = i;
- if (!chan->buf[i])
- goto free_bufs;
- }
-@@ -328,7 +333,7 @@ size_t relay_switch_subbuf(struct rchan_
- return length;
-
- toobig:
-- printk(KERN_WARNING "relayfs: event too large (%u)\n", length);
-+ printk(KERN_WARNING "relayfs: event too large (%Zd)\n", length);
- WARN_ON(1);
- return 0;
- }
-diff -urpN -X dontdiff linux-2.6.13-rc6-mm2/include/linux/relayfs_fs.h linux-2.6.13-rc6-mm2-cur/include/linux/relayfs_fs.h
---- linux-2.6.13-rc6-mm2/include/linux/relayfs_fs.h 2005-08-25 19:29:03.000000000 -0500
-+++ linux-2.6.13-rc6-mm2-cur/include/linux/relayfs_fs.h 2005-08-24 00:16:37.000000000 -0500
-@@ -22,7 +22,7 @@
- /*
- * Tracks changes to rchan_buf struct
- */
--#define RELAYFS_CHANNEL_VERSION 4
-+#define RELAYFS_CHANNEL_VERSION 5
-
- /*
- * Per-cpu relay channel buffer
-@@ -44,6 +44,8 @@ struct rchan_buf
- unsigned int finalized; /* buffer has been finalized */
- size_t *padding; /* padding counts per sub-buffer */
- size_t prev_padding; /* temporary variable */
-+ size_t bytes_consumed; /* bytes consumed in cur read subbuf */
-+ unsigned int cpu; /* this buf's cpu */
- } ____cacheline_aligned;
-
- /*
-
-
projects / blktrace.git / commitdiff