1/ Introduce "mddev->resync_max_sectors" so that an md personality
can ask for resync to cover a different address range than that of a
single drive. raid10 will use this.
2/ fix is_mddev_idle so that if there seem to be a negative number
of events, it doesn't immediately assume activity.
3/ make "sync_io" (the count of IO sectors used for array resync)
an atomic_t to avoid SMP races.
4/ Pass md_sync_acct a "block_device" rather than the containing "rdev",
as the whole rdev isn't needed. Also make this an inline function.
5/ Make sure recovery gets interrupted on any error.
diff ./drivers/md/md.c~current~ ./drivers/md/md.c
Signed-off-by: Neil Brown <neilb@xxxxxxxxxxxxxxx>
### Diffstat output
./drivers/md/md.c | 32 ++++++++++++++++++++++----------
./drivers/md/raid1.c | 4 ++--
./drivers/md/raid5.c | 5 +++--
./drivers/md/raid6main.c | 5 +++--
./include/linux/genhd.h | 2 +-
./include/linux/raid/md.h | 1 -
./include/linux/raid/md_k.h | 6 ++++++
7 files changed, 37 insertions(+), 18 deletions(-)
diff ./drivers/md/md.c~current~ ./drivers/md/md.c
--- ./drivers/md/md.c~current~ 2004-08-23 12:07:44.000000000 +1000
+++ ./drivers/md/md.c 2004-08-23 12:23:14.000000000 +1000
@@ -1684,6 +1684,8 @@ static int do_md_run(mddev_t * mddev)
mddev->pers = pers[pnum];
spin_unlock(&pers_lock);
+ mddev->resync_max_sectors = mddev->size << 1; /* may be over-ridden by personality */
+
err = mddev->pers->run(mddev);
if (err) {
printk(KERN_ERR "md: pers->run() failed ...\n");
@@ -2989,6 +2991,7 @@ void md_error(mddev_t *mddev, mdk_rdev_t
if (!mddev->pers->error_handler)
return;
mddev->pers->error_handler(mddev,rdev);
+ set_bit(MD_RECOVERY_INTR, &mddev->recovery);
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
md_wakeup_thread(mddev->thread);
}
@@ -3021,7 +3024,11 @@ static void status_resync(struct seq_fil
unsigned long max_blocks, resync, res, dt, db, rt;
resync = (mddev->curr_resync - atomic_read(&mddev->recovery_active))/2;
- max_blocks = mddev->size;
+
+ if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
+ max_blocks = mddev->resync_max_sectors >> 1;
+ else
+ max_blocks = mddev->size;
/*
* Should not happen.
@@ -3257,11 +3264,6 @@ int unregister_md_personality(int pnum)
return 0;
}
-void md_sync_acct(mdk_rdev_t *rdev, unsigned long nr_sectors)
-{
- rdev->bdev->bd_contains->bd_disk->sync_io += nr_sectors;
-}
-
static int is_mddev_idle(mddev_t *mddev)
{
mdk_rdev_t * rdev;
@@ -3274,8 +3276,12 @@ static int is_mddev_idle(mddev_t *mddev)
struct gendisk *disk = rdev->bdev->bd_contains->bd_disk;
curr_events = disk_stat_read(disk, read_sectors) +
disk_stat_read(disk, write_sectors) -
- disk->sync_io;
- if ((curr_events - rdev->last_events) > 32) {
+ atomic_read(&disk->sync_io);
+ /* Allow some slack between valud of curr_events and last_events,
+ * as there are some uninteresting races.
+ * Note: the following is an unsigned comparison.
+ */
+ if ((curr_events - rdev->last_events + 32) > 64) {
rdev->last_events = curr_events;
idle = 0;
}
@@ -3409,7 +3415,14 @@ static void md_do_sync(mddev_t *mddev)
}
} while (mddev->curr_resync < 2);
- max_sectors = mddev->size << 1;
+ if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
+ /* resync follows the size requested by the personality,
+ * which default to physical size, but can be virtual size
+ */
+ max_sectors = mddev->resync_max_sectors;
+ else
+ /* recovery follows the physical size of devices */
+ max_sectors = mddev->size << 1;
printk(KERN_INFO "md: syncing RAID array %s\n", mdname(mddev));
printk(KERN_INFO "md: minimum _guaranteed_ reconstruction speed:"
@@ -3832,7 +3845,6 @@ module_exit(md_exit)
EXPORT_SYMBOL(register_md_personality);
EXPORT_SYMBOL(unregister_md_personality);
EXPORT_SYMBOL(md_error);
-EXPORT_SYMBOL(md_sync_acct);
EXPORT_SYMBOL(md_done_sync);
EXPORT_SYMBOL(md_write_start);
EXPORT_SYMBOL(md_write_end);
diff ./drivers/md/raid1.c~current~ ./drivers/md/raid1.c
--- ./drivers/md/raid1.c~current~ 2004-08-23 12:07:44.000000000 +1000
+++ ./drivers/md/raid1.c 2004-08-23 12:20:59.000000000 +1000
@@ -903,7 +903,7 @@ static void sync_request_write(mddev_t *
atomic_inc(&conf->mirrors[i].rdev->nr_pending);
atomic_inc(&r1_bio->remaining);
- md_sync_acct(conf->mirrors[i].rdev, wbio->bi_size >> 9);
+ md_sync_acct(conf->mirrors[i].rdev->bdev, wbio->bi_size >> 9);
generic_make_request(wbio);
}
@@ -1143,7 +1143,7 @@ static int sync_request(mddev_t *mddev,
bio = r1_bio->bios[disk];
r1_bio->sectors = nr_sectors;
- md_sync_acct(mirror->rdev, nr_sectors);
+ md_sync_acct(mirror->rdev->bdev, nr_sectors);
generic_make_request(bio);
diff ./drivers/md/raid5.c~current~ ./drivers/md/raid5.c
--- ./drivers/md/raid5.c~current~ 2004-08-23 12:07:44.000000000 +1000
+++ ./drivers/md/raid5.c 2004-08-23 12:20:59.000000000 +1000
@@ -1071,7 +1071,8 @@ static void handle_stripe(struct stripe_
PRINTK("Reading block %d (sync=%d)\n",
i, syncing);
if (syncing)
- md_sync_acct(conf->disks[i].rdev, STRIPE_SECTORS);
+ md_sync_acct(conf->disks[i].rdev->bdev,
+ STRIPE_SECTORS);
}
}
}
@@ -1256,7 +1257,7 @@ static void handle_stripe(struct stripe_
if (rdev) {
if (test_bit(R5_Syncio, &sh->dev[i].flags))
- md_sync_acct(rdev, STRIPE_SECTORS);
+ md_sync_acct(rdev->bdev, STRIPE_SECTORS);
bi->bi_bdev = rdev->bdev;
PRINTK("for %llu schedule op %ld on disc %d\n",
diff ./drivers/md/raid6main.c~current~ ./drivers/md/raid6main.c
--- ./drivers/md/raid6main.c~current~ 2004-08-23 12:07:44.000000000 +1000
+++ ./drivers/md/raid6main.c 2004-08-23 12:20:59.000000000 +1000
@@ -1208,7 +1208,8 @@ static void handle_stripe(struct stripe_
PRINTK("Reading block %d (sync=%d)\n",
i, syncing);
if (syncing)
- md_sync_acct(conf->disks[i].rdev, STRIPE_SECTORS);
+ md_sync_acct(conf->disks[i].rdev->bdev,
+ STRIPE_SECTORS);
}
}
}
@@ -1418,7 +1419,7 @@ static void handle_stripe(struct stripe_
if (rdev) {
if (test_bit(R5_Syncio, &sh->dev[i].flags))
- md_sync_acct(rdev, STRIPE_SECTORS);
+ md_sync_acct(rdev->bdev, STRIPE_SECTORS);
bi->bi_bdev = rdev->bdev;
PRINTK("for %llu schedule op %ld on disc %d\n",
diff ./include/linux/genhd.h~current~ ./include/linux/genhd.h
--- ./include/linux/genhd.h~current~ 2004-08-23 12:07:44.000000000 +1000
+++ ./include/linux/genhd.h 2004-08-23 12:20:59.000000000 +1000
@@ -100,7 +100,7 @@ struct gendisk {
struct timer_rand_state *random;
int policy;
- unsigned sync_io; /* RAID */
+ atomic_t sync_io; /* RAID */
unsigned long stamp, stamp_idle;
int in_flight;
#ifdef CONFIG_SMP
diff ./include/linux/raid/md.h~current~ ./include/linux/raid/md.h
--- ./include/linux/raid/md.h~current~ 2004-08-23 12:07:44.000000000 +1000
+++ ./include/linux/raid/md.h 2004-08-23 12:20:59.000000000 +1000
@@ -74,7 +74,6 @@ extern void md_write_start(mddev_t *mdde
extern void md_write_end(mddev_t *mddev);
extern void md_handle_safemode(mddev_t *mddev);
extern void md_done_sync(mddev_t *mddev, int blocks, int ok);
-extern void md_sync_acct(mdk_rdev_t *rdev, unsigned long nr_sectors);
extern void md_error (mddev_t *mddev, mdk_rdev_t *rdev);
extern void md_unplug_mddev(mddev_t *mddev);
diff ./include/linux/raid/md_k.h~current~ ./include/linux/raid/md_k.h
--- ./include/linux/raid/md_k.h~current~ 2004-08-23 12:07:44.000000000 +1000
+++ ./include/linux/raid/md_k.h 2004-08-23 12:20:59.000000000 +1000
@@ -216,6 +216,7 @@ struct mddev_s
unsigned long resync_mark; /* a recent timestamp */
sector_t resync_mark_cnt;/* blocks written at resync_mark */
+ sector_t resync_max_sectors; /* may be set by personality */
/* recovery/resync flags
* NEEDED: we might need to start a resync/recover
* RUNNING: a thread is running, or about to be started
@@ -263,6 +264,11 @@ static inline void rdev_dec_pending(mdk_
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
}
+static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
+{
+ atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
+}
+
struct mdk_personality_s
{
char *name;
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
