blktrace: support per-partition tracing

Though one can specify '-d /dev/sda1' when using blktrace, it still
traces the whole sda.

To support per-partition tracing, when we start tracing, we initialize
bt->start_lba and bt->end_lba to the start and end sector of that
partition.

Note some actions are per device, thus we don't filter 0-sector events.

The original patch and discussion can be found here:
	http://marc.info/?l=linux-btrace&m=122949374214540&w=2

Signed-off-by: Shawn Du <duyuyang@gmail.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jens Axboe <jens.axboe@oracle.com>
LKML-Reference: <49E42620.4050701@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

block/compat_ioctl.c

@@ -568,7 +568,7 @@ static int compat_blk_trace_setup(struct block_device *bdev, char __user *arg)
 	memcpy(&buts.name, &cbuts.name, 32);
 
 	mutex_lock(&bdev->bd_mutex);
-	ret = do_blk_trace_setup(q, b, bdev->bd_dev, &buts);
+	ret = do_blk_trace_setup(q, b, bdev->bd_dev, bdev, &buts);
 	mutex_unlock(&bdev->bd_mutex);
 	if (ret)
 		return ret;

drivers/scsi/sg.c

@@ -1065,6 +1065,7 @@ sg_ioctl(struct inode *inode, struct file *filp,
 		return blk_trace_setup(sdp->device->request_queue,
 				       sdp->disk->disk_name,
 				       MKDEV(SCSI_GENERIC_MAJOR, sdp->index),
+				       NULL,
 				       (char *)arg);
 	case BLKTRACESTART:
 		return blk_trace_startstop(sdp->device->request_queue, 1);

include/linux/blktrace_api.h

@@ -165,8 +165,9 @@ struct blk_trace {
 
 extern int blk_trace_ioctl(struct block_device *, unsigned, char __user *);
 extern void blk_trace_shutdown(struct request_queue *);
-extern int do_blk_trace_setup(struct request_queue *q,
-	char *name, dev_t dev, struct blk_user_trace_setup *buts);
+extern int do_blk_trace_setup(struct request_queue *q, char *name,
+			      dev_t dev, struct block_device *bdev,
+			      struct blk_user_trace_setup *buts);
 extern void __trace_note_message(struct blk_trace *, const char *fmt, ...);
 
 /**
@@ -193,6 +194,7 @@ extern void __trace_note_message(struct blk_trace *, const char *fmt, ...);
 extern void blk_add_driver_data(struct request_queue *q, struct request *rq,
 				void *data, size_t len);
 extern int blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
+			   struct block_device *bdev,
 			   char __user *arg);
 extern int blk_trace_startstop(struct request_queue *q, int start);
 extern int blk_trace_remove(struct request_queue *q);
@@ -200,15 +202,15 @@ extern int blk_trace_remove(struct request_queue *q);
 extern struct attribute_group blk_trace_attr_group;
 
 #else /* !CONFIG_BLK_DEV_IO_TRACE */
-#define blk_trace_ioctl(bdev, cmd, arg)		(-ENOTTY)
-#define blk_trace_shutdown(q)			do { } while (0)
-#define do_blk_trace_setup(q, name, dev, buts)	(-ENOTTY)
-#define blk_add_driver_data(q, rq, data, len)	do {} while (0)
-#define blk_trace_setup(q, name, dev, arg)	(-ENOTTY)
-#define blk_trace_startstop(q, start)		(-ENOTTY)
-#define blk_trace_remove(q)			(-ENOTTY)
-#define blk_add_trace_msg(q, fmt, ...)		do { } while (0)
-
+# define blk_trace_ioctl(bdev, cmd, arg)		(-ENOTTY)
+# define blk_trace_shutdown(q)				do { } while (0)
+# define do_blk_trace_setup(q, name, dev, bdev, buts)	(-ENOTTY)
+# define blk_add_driver_data(q, rq, data, len)		do {} while (0)
+# define blk_trace_setup(q, name, dev, bdev, arg)	(-ENOTTY)
+# define blk_trace_startstop(q, start)			(-ENOTTY)
+# define blk_trace_remove(q)				(-ENOTTY)
+# define blk_add_trace_msg(q, fmt, ...)			do { } while (0)
 #endif /* CONFIG_BLK_DEV_IO_TRACE */
+
 #endif /* __KERNEL__ */
 #endif

kernel/trace/blktrace.c

@@ -147,7 +147,7 @@ static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector,
 {
 	if (((bt->act_mask << BLK_TC_SHIFT) & what) == 0)
 		return 1;
-	if (sector < bt->start_lba || sector > bt->end_lba)
+	if (sector && (sector < bt->start_lba || sector > bt->end_lba))
 		return 1;
 	if (bt->pid && pid != bt->pid)
 		return 1;
@@ -192,7 +192,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes,
 	what |= MASK_TC_BIT(rw, DISCARD);
 
 	pid = tsk->pid;
-	if (unlikely(act_log_check(bt, what, sector, pid)))
+	if (act_log_check(bt, what, sector, pid))
 		return;
 	cpu = raw_smp_processor_id();
 
@@ -407,11 +407,13 @@ static struct rchan_callbacks blk_relay_callbacks = {
  * Setup everything required to start tracing
  */
 int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
-			struct blk_user_trace_setup *buts)
+		       struct block_device *bdev,
+		       struct blk_user_trace_setup *buts)
 {
 	struct blk_trace *old_bt, *bt = NULL;
 	struct dentry *dir = NULL;
 	int ret, i;
+	struct hd_struct *part = NULL;
 
 	if (!buts->buf_size || !buts->buf_nr)
 		return -EINVAL;
@@ -480,11 +482,21 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
 	if (!bt->act_mask)
 		bt->act_mask = (u16) -1;
 
-	bt->start_lba = buts->start_lba;
-	bt->end_lba = buts->end_lba;
-	if (!bt->end_lba)
+	if (bdev)
+		part = bdev->bd_part;
+
+	if (part) {
+		bt->start_lba = part->start_sect;
+		bt->end_lba = part->start_sect + part->nr_sects;
+	} else
 		bt->end_lba = -1ULL;
 
+	/* overwrite with user settings */
+	if (buts->start_lba)
+		bt->start_lba = buts->start_lba;
+	if (buts->end_lba)
+		bt->end_lba = buts->end_lba;
+
 	bt->pid = buts->pid;
 	bt->trace_state = Blktrace_setup;
 
@@ -505,6 +517,7 @@ err:
 }
 
 int blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
+		    struct block_device *bdev,
 		    char __user *arg)
 {
 	struct blk_user_trace_setup buts;
@@ -514,7 +527,7 @@ int blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
 	if (ret)
 		return -EFAULT;
 
-	ret = do_blk_trace_setup(q, name, dev, &buts);
+	ret = do_blk_trace_setup(q, name, dev, bdev, &buts);
 	if (ret)
 		return ret;
 
@@ -582,7 +595,7 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg)
 	switch (cmd) {
 	case BLKTRACESETUP:
 		bdevname(bdev, b);
-		ret = blk_trace_setup(q, b, bdev->bd_dev, arg);
+		ret = blk_trace_setup(q, b, bdev->bd_dev, bdev, arg);
 		break;
 	case BLKTRACESTART:
 		start = 1;