lkml.org 
[lkml]   [2019]   [Feb]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH 4.14 43/46] md/raid5: fix out of memory during raid cache recovery
Date
4.14-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Alexei Naberezhnov <anaberezhnov@fb.com>

commit 483cbbeddd5fe2c80fd4141ff0748fa06c4ff146 upstream.

This fixes the case when md array assembly fails because of raid cache recovery
unable to allocate a stripe, despite attempts to replay stripes and increase
cache size. This happens because stripes released by r5c_recovery_replay_stripes
and raid5_set_cache_size don't become available for allocation immediately.
Released stripes first are placed on conf->released_stripes list and require
md thread to merge them on conf->inactive_list before they can be allocated.

Patch allows final allocation attempt during cache recovery to wait for
new stripes to become availabe for allocation.

Cc: linux-raid@vger.kernel.org
Cc: Shaohua Li <shli@kernel.org>
Cc: linux-stable <stable@vger.kernel.org> # 4.10+
Fixes: b4c625c67362 ("md/r5cache: r5cache recovery: part 1")
Signed-off-by: Alexei Naberezhnov <anaberezhnov@fb.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
drivers/md/raid5-cache.c | 33 ++++++++++++++++++++++-----------
drivers/md/raid5.c | 8 ++++++--
2 files changed, 28 insertions(+), 13 deletions(-)

--- a/drivers/md/raid5-cache.c
+++ b/drivers/md/raid5-cache.c
@@ -1942,12 +1942,14 @@ out:
}

static struct stripe_head *
-r5c_recovery_alloc_stripe(struct r5conf *conf,
- sector_t stripe_sect)
+r5c_recovery_alloc_stripe(
+ struct r5conf *conf,
+ sector_t stripe_sect,
+ int noblock)
{
struct stripe_head *sh;

- sh = raid5_get_active_stripe(conf, stripe_sect, 0, 1, 0);
+ sh = raid5_get_active_stripe(conf, stripe_sect, 0, noblock, 0);
if (!sh)
return NULL; /* no more stripe available */

@@ -2157,7 +2159,7 @@ r5c_recovery_analyze_meta_block(struct r
stripe_sect);

if (!sh) {
- sh = r5c_recovery_alloc_stripe(conf, stripe_sect);
+ sh = r5c_recovery_alloc_stripe(conf, stripe_sect, 1);
/*
* cannot get stripe from raid5_get_active_stripe
* try replay some stripes
@@ -2166,20 +2168,29 @@ r5c_recovery_analyze_meta_block(struct r
r5c_recovery_replay_stripes(
cached_stripe_list, ctx);
sh = r5c_recovery_alloc_stripe(
- conf, stripe_sect);
+ conf, stripe_sect, 1);
}
if (!sh) {
+ int new_size = conf->min_nr_stripes * 2;
pr_debug("md/raid:%s: Increasing stripe cache size to %d to recovery data on journal.\n",
mdname(mddev),
- conf->min_nr_stripes * 2);
- raid5_set_cache_size(mddev,
- conf->min_nr_stripes * 2);
- sh = r5c_recovery_alloc_stripe(conf,
- stripe_sect);
+ new_size);
+ ret = raid5_set_cache_size(mddev, new_size);
+ if (conf->min_nr_stripes <= new_size / 2) {
+ pr_err("md/raid:%s: Cannot increase cache size, ret=%d, new_size=%d, min_nr_stripes=%d, max_nr_stripes=%d\n",
+ mdname(mddev),
+ ret,
+ new_size,
+ conf->min_nr_stripes,
+ conf->max_nr_stripes);
+ return -ENOMEM;
+ }
+ sh = r5c_recovery_alloc_stripe(
+ conf, stripe_sect, 0);
}
if (!sh) {
pr_err("md/raid:%s: Cannot get enough stripes due to memory pressure. Recovery failed.\n",
- mdname(mddev));
+ mdname(mddev));
return -ENOMEM;
}
list_add_tail(&sh->lru, cached_stripe_list);
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -6336,6 +6336,7 @@ raid5_show_stripe_cache_size(struct mdde
int
raid5_set_cache_size(struct mddev *mddev, int size)
{
+ int result = 0;
struct r5conf *conf = mddev->private;

if (size <= 16 || size > 32768)
@@ -6352,11 +6353,14 @@ raid5_set_cache_size(struct mddev *mddev

mutex_lock(&conf->cache_size_mutex);
while (size > conf->max_nr_stripes)
- if (!grow_one_stripe(conf, GFP_KERNEL))
+ if (!grow_one_stripe(conf, GFP_KERNEL)) {
+ conf->min_nr_stripes = conf->max_nr_stripes;
+ result = -ENOMEM;
break;
+ }
mutex_unlock(&conf->cache_size_mutex);

- return 0;
+ return result;
}
EXPORT_SYMBOL(raid5_set_cache_size);


\
 
 \ /
  Last update: 2019-02-04 11:46    [W:0.159 / U:5.100 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site