[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[PATCH 12/24] aoe: don't call set_capacity from irq context



Updating the block device size from irq context can lead to torn
writes of the 64-bit value, and prevents us from using normal
process context locking primitives to serialize access to the 64-bit
nr_sectors value.  Defer the set_capacity to the already existing
workqueue handler, where it can be merged with the update of the
block device size by using set_capacity_and_notify.  As an extra
bonus this also adds proper uevent notifications for the resize.

Signed-off-by: Christoph Hellwig <hch@xxxxxx>
---
 drivers/block/aoe/aoecmd.c | 15 ++++-----------
 1 file changed, 4 insertions(+), 11 deletions(-)

diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c
index 313f0b946fe2b3..ac720bdcd983e7 100644
--- a/drivers/block/aoe/aoecmd.c
+++ b/drivers/block/aoe/aoecmd.c
@@ -890,19 +890,13 @@ void
 aoecmd_sleepwork(struct work_struct *work)
 {
        struct aoedev *d = container_of(work, struct aoedev, work);
-       struct block_device *bd;
-       u64 ssize;
 
        if (d->flags & DEVFL_GDALLOC)
                aoeblk_gdalloc(d);
 
        if (d->flags & DEVFL_NEWSIZE) {
-               ssize = get_capacity(d->gd);
-               bd = bdget_disk(d->gd, 0);
-               if (bd) {
-                       bd_set_nr_sectors(bd, ssize);
-                       bdput(bd);
-               }
+               set_capacity_and_notify(d->gd, d->ssize);
+
                spin_lock_irq(&d->lock);
                d->flags |= DEVFL_UP;
                d->flags &= ~DEVFL_NEWSIZE;
@@ -971,10 +965,9 @@ ataid_complete(struct aoedev *d, struct aoetgt *t, 
unsigned char *id)
        d->geo.start = 0;
        if (d->flags & (DEVFL_GDALLOC|DEVFL_NEWSIZE))
                return;
-       if (d->gd != NULL) {
-               set_capacity(d->gd, ssize);
+       if (d->gd != NULL)
                d->flags |= DEVFL_NEWSIZE;
-       } else
+       else
                d->flags |= DEVFL_GDALLOC;
        schedule_work(&d->work);
 }
-- 
2.28.0




 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.