* [PATCH v2] [SCSI] scsi_dh_rdac: avoid possible deadlock in queue_mode_select
@ 2011-01-07 19:51 Mike Snitzer
0 siblings, 0 replies; only message in thread
From: Mike Snitzer @ 2011-01-07 19:51 UTC (permalink / raw)
To: James Bottomley; +Cc: linux-scsi, Babu Moger, Mike Snitzer
From: Babu Moger <babu.moger@lsi.com>
Use GFP_NOIO when allocating memory during activation of a new path to
avoid possiblility of triggering recursive writeback to a failed path.
A failed IO could cause the path to go down, then we would call
queue_mode_select() to active a new path. Allocating memory using
GFP_KERNEL is bad to use here because completing the allocation may
require the system to write out pages to the same disk that just had
the failed path.
Signed-off-by: Babu Moger <babu.moger@lsi.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
---
drivers/scsi/device_handler/scsi_dh_rdac.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
v2: just refreshed and tweaked the patch header
- v1 is available here: https://patchwork.kernel.org/patch/71073/
diff --git a/drivers/scsi/device_handler/scsi_dh_rdac.c b/drivers/scsi/device_handler/scsi_dh_rdac.c
index 5be3ae1..566effd 100644
--- a/drivers/scsi/device_handler/scsi_dh_rdac.c
+++ b/drivers/scsi/device_handler/scsi_dh_rdac.c
@@ -625,7 +625,7 @@ static int queue_mode_select(struct scsi_device *sdev,
struct rdac_queue_data *qdata;
struct rdac_controller *ctlr;
- qdata = kzalloc(sizeof(*qdata), GFP_KERNEL);
+ qdata = kzalloc(sizeof(*qdata), GFP_NOIO);
if (!qdata)
return SCSI_DH_RETRY;
^ permalink raw reply related [flat|nested] only message in thread
only message in thread, other threads:[~2011-01-07 19:51 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-01-07 19:51 [PATCH v2] [SCSI] scsi_dh_rdac: avoid possible deadlock in queue_mode_select Mike Snitzer
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).