PWN-Hunter/android_kernel_xiaomi_sdm845
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

drbd: don't BUG_ON, if bio_add_page of a single page to an empty bio fails

Browse Source 
Just deal with it more gracefully, if we fail to add even a single page
to an empty bio. We used to BUG_ON() there, but it has been observed in
some Xen deployment, so we need to handle that case more robustly now.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
This commit is contained in:
Lars Ellenberg 2011-01-24 14:47:09 +01:00 committed by Philipp Reisner
parent 039312b648
commit 10f6d9926c
2 changed files with 34 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
drivers/block/drbd/drbd_receiver.c
View File

@ -1073,6 +1073,16 @@ void drbd_bump_write_ordering(struct drbd_conf *mdev, enum write_ordering_e wo)
* @mdev: DRBD device.
* @e: epoch entry
* @rw: flag field, see bio->bi_rw
*
* May spread the pages to multiple bios,
* depending on bio_add_page restrictions.
*
* Returns 0 if all bios have been submitted,
* -ENOMEM if we could not allocate enough bios,
* -ENOSPC (any better suggestion?) if we have not been able to bio_add_page a
* single page to an empty bio (which should never happen and likely indicates
* that the lower level IO stack is in some way broken). This has been observed
* on certain Xen deployments.
*/
/* TODO allocate from our own bio_set. */
int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e,
@ -1085,6 +1095,7 @@ int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e,
unsigned ds = e->size;
unsigned n_bios = 0;
unsigned nr_pages = (ds + PAGE_SIZE -1) >> PAGE_SHIFT;
int err = -ENOMEM;
/* In most cases, we will only need one bio. But in case the lower
* level restrictions happen to be different at this offset on this
@ -1110,8 +1121,17 @@ int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e,
page_chain_for_each(page) {
unsigned len = min_t(unsigned, ds, PAGE_SIZE);
if (!bio_add_page(bio, page, len, 0)) {
/* a single page must always be possible! */
BUG_ON(bio->bi_vcnt == 0);
/* A single page must always be possible!
* But in case it fails anyways,
* we deal with it, and complain (below). */
if (bio->bi_vcnt == 0) {
dev_err(DEV,
"bio_add_page failed for len=%u, "
"bi_vcnt=0 (bi_sector=%llu)\n",
len, (unsigned long long)bio->bi_sector);
err = -ENOSPC;
goto fail;
}
goto next_bio;
}
ds -= len;
@ -1137,7 +1157,7 @@ int drbd_submit_ee(struct drbd_conf *mdev, struct drbd_epoch_entry *e,
bios = bios->bi_next;
bio_put(bio);
}
return -ENOMEM;
return err;
}
static int receive_Barrier(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned int data_size)
@ -1436,9 +1456,8 @@ static int recv_resync_read(struct drbd_conf *mdev, sector_t sector, int data_si
if (drbd_submit_ee(mdev, e, WRITE, DRBD_FAULT_RS_WR) == 0)
return true;
/* drbd_submit_ee currently fails for one reason only:
* not being able to allocate enough bios.
* Is dropping the connection going to help? */
/* don't care for the reason here */
dev_err(DEV, "submit failed, triggering re-connect\n");
spin_lock_irq(&mdev->req_lock);
list_del(&e->w.list);
spin_unlock_irq(&mdev->req_lock);
@ -1837,9 +1856,8 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
if (drbd_submit_ee(mdev, e, rw, DRBD_FAULT_DT_WR) == 0)
return true;
/* drbd_submit_ee currently fails for one reason only:
* not being able to allocate enough bios.
* Is dropping the connection going to help? */
/* don't care for the reason here */
dev_err(DEV, "submit failed, triggering re-connect\n");
spin_lock_irq(&mdev->req_lock);
list_del(&e->w.list);
hlist_del_init(&e->colision);
@ -1848,9 +1866,7 @@ static int receive_Data(struct drbd_conf *mdev, enum drbd_packets cmd, unsigned
drbd_al_complete_io(mdev, e->sector);
out_interrupted:
/* yes, the epoch_size now is imbalanced.
* but we drop the connection anyways, so we don't have a chance to
* receive a barrier... atomic_inc(&mdev->epoch_size); */
drbd_may_finish_epoch(mdev, e->epoch, EV_PUT + EV_CLEANUP);
put_ldev(mdev);
drbd_free_ee(mdev, e);
return false;
@ -2096,9 +2112,8 @@ static int receive_DataRequest(struct drbd_conf *mdev, enum drbd_packets cmd, un
if (drbd_submit_ee(mdev, e, READ, fault_type) == 0)
return true;
/* drbd_submit_ee currently fails for one reason only:
* not being able to allocate enough bios.
* Is dropping the connection going to help? */
/* don't care for the reason here */
dev_err(DEV, "submit failed, triggering re-connect\n");
spin_lock_irq(&mdev->req_lock);
list_del(&e->w.list);
spin_unlock_irq(&mdev->req_lock);

7
drivers/block/drbd/drbd_worker.c
View File

@ -369,9 +369,10 @@ static int read_for_csum(struct drbd_conf *mdev, sector_t sector, int size)
if (drbd_submit_ee(mdev, e, READ, DRBD_FAULT_RS_RD) == 0)
return 0;
/* drbd_submit_ee currently fails for one reason only:
* not being able to allocate enough bios.
* Is dropping the connection going to help? */
/* If it failed because of ENOMEM, retry should help. If it failed
* because bio_add_page failed (probably broken lower level driver),
* retry may or may not help.
* If it does not, you may need to force disconnect. */
spin_lock_irq(&mdev->req_lock);
list_del(&e->w.list);
spin_unlock_irq(&mdev->req_lock);