[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20160418022620.124098598@linuxfoundation.org>
Date: Mon, 18 Apr 2016 11:29:17 +0900
From: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To: linux-kernel@...r.kernel.org
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
stable@...r.kernel.org, Vlastimil Babka <vbabka@...e.cz>,
Neil Brown <neilb@...e.com>, David Disseldorp <ddiss@...e.de>,
Ilya Dryomov <idryomov@...il.com>
Subject: [PATCH 4.5 085/124] rbd: use GFP_NOIO consistently for request allocations
4.5-stable review patch. If anyone has any objections, please let me know.
------------------
From: David Disseldorp <ddiss@...e.de>
commit 2224d879c7c0f85c14183ef82eb48bd875ceb599 upstream.
As of 5a60e87603c4c533492c515b7f62578189b03c9c, RBD object request
allocations are made via rbd_obj_request_create() with GFP_NOIO.
However, subsequent OSD request allocations in rbd_osd_req_create*()
use GFP_ATOMIC.
With heavy page cache usage (e.g. OSDs running on same host as krbd
client), rbd_osd_req_create() order-1 GFP_ATOMIC allocations have been
observed to fail, where direct reclaim would have allowed GFP_NOIO
allocations to succeed.
Suggested-by: Vlastimil Babka <vbabka@...e.cz>
Suggested-by: Neil Brown <neilb@...e.com>
Signed-off-by: David Disseldorp <ddiss@...e.de>
Signed-off-by: Ilya Dryomov <idryomov@...il.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
---
drivers/block/rbd.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -1955,7 +1955,7 @@ static struct ceph_osd_request *rbd_osd_
osdc = &rbd_dev->rbd_client->client->osdc;
osd_req = ceph_osdc_alloc_request(osdc, snapc, num_ops, false,
- GFP_ATOMIC);
+ GFP_NOIO);
if (!osd_req)
return NULL; /* ENOMEM */
@@ -2004,7 +2004,7 @@ rbd_osd_req_create_copyup(struct rbd_obj
rbd_dev = img_request->rbd_dev;
osdc = &rbd_dev->rbd_client->client->osdc;
osd_req = ceph_osdc_alloc_request(osdc, snapc, num_osd_ops,
- false, GFP_ATOMIC);
+ false, GFP_NOIO);
if (!osd_req)
return NULL; /* ENOMEM */
@@ -2506,7 +2506,7 @@ static int rbd_img_request_fill(struct r
bio_chain_clone_range(&bio_list,
&bio_offset,
clone_size,
- GFP_ATOMIC);
+ GFP_NOIO);
if (!obj_request->bio_list)
goto out_unwind;
} else if (type == OBJ_REQUEST_PAGES) {
Powered by blists - more mailing lists