From: tada keisuke <keisuke1.tada@kioxia.com>
To: "song@kernel.org" <song@kernel.org>,
"yukuai3@huawei.com" <yukuai3@huawei.com>
Cc: "linux-raid@vger.kernel.org" <linux-raid@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: [PATCH 03/11] md: workaround for inconsistency of config state in takeover
Date: Tue, 26 Mar 2024 10:29:28 +0000 [thread overview]
Message-ID: <df3a5e85db254f2ab44b93734cac27bb@kioxia.com> (raw)
This patch depends on patch 01.
An inconsistency occurs between config state in takeover and percpu_ref state.
Differentiate percpu_ref state setting for takeover and other to avoid the inconsistency.
Therefore, add percpu_ref state setting to match config state in takeover.
Signed-off-by: Keisuke TADA <keisuke1.tada@kioxia.com>
Signed-off-by: Toshifumi OHTAKE <toshifumi.ootake@kioxia.com>
---
drivers/md/raid5.c | 39 +++++++++++++++++++++++++++++++--------
1 file changed, 31 insertions(+), 8 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 118dfed9695f..c7186ebcd299 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -7405,7 +7405,8 @@ static unsigned long raid5_cache_count(struct shrinker *shrink,
return max_stripes - min_stripes;
}
-static struct r5conf *setup_conf(struct mddev *mddev)
+
+static struct r5conf *setup_conf(struct mddev *mddev, bool quiesce)
{
struct r5conf *conf;
int raid_disk, memory, max_disks;
@@ -7416,6 +7417,7 @@ static struct r5conf *setup_conf(struct mddev *mddev)
int group_cnt;
struct r5worker_group *new_group;
int ret = -ENOMEM;
+ unsigned int percpu_ref_init_flags;
if (mddev->new_level != 5
&& mddev->new_level != 4
@@ -7486,6 +7488,10 @@ static struct r5conf *setup_conf(struct mddev *mddev)
init_llist_head(&conf->released_stripes);
atomic_set(&conf->active_stripes, 0);
atomic_set(&conf->preread_active_stripes, 0);
+ if (quiesce)
+ percpu_ref_init_flags = PERCPU_REF_ALLOW_REINIT | PERCPU_REF_INIT_DEAD;
+ else
+ percpu_ref_init_flags = PERCPU_REF_ALLOW_REINIT;
atomic_set(&conf->active_aligned_reads, 0);
spin_lock_init(&conf->pending_bios_lock);
conf->batch_bio_dispatch = true;
@@ -7666,6 +7672,23 @@ static struct r5conf *setup_conf(struct mddev *mddev)
return ERR_PTR(ret);
}
+static struct r5conf *setup_conf_for_run(struct mddev *mddev)
+{
+ return setup_conf(mddev, false);
+}
+
+static struct r5conf *setup_conf_for_takeover(struct mddev *mddev)
+{
+ struct r5conf *conf;
+ bool quiesce = false;
+
+ if (mddev->level == 4 || mddev->level == 5 || mddev->level == 6) {
+ conf = mddev->private;
+ quiesce = false;
+ }
+ return setup_conf(mddev, quiesce);
+}
+
static int only_parity(int raid_disk, int algo, int raid_disks, int max_degraded)
{
switch (algo) {
@@ -7893,7 +7916,7 @@ static int raid5_run(struct mddev *mddev)
}
if (mddev->private == NULL)
- conf = setup_conf(mddev);
+ conf = setup_conf_for_run(mddev);
else
conf = mddev->private;
@@ -8661,7 +8684,7 @@ static void *raid45_takeover_raid0(struct mddev *mddev, int level)
/* make sure it will be not marked as dirty */
mddev->recovery_cp = MaxSector;
- return setup_conf(mddev);
+ return setup_conf_for_takeover(mddev);
}
static void *raid5_takeover_raid1(struct mddev *mddev)
@@ -8689,7 +8712,7 @@ static void *raid5_takeover_raid1(struct mddev *mddev)
mddev->new_layout = ALGORITHM_LEFT_SYMMETRIC;
mddev->new_chunk_sectors = chunksect;
- ret = setup_conf(mddev);
+ ret = setup_conf_for_takeover(mddev);
if (!IS_ERR(ret))
mddev_clear_unsupported_flags(mddev,
UNSUPPORTED_MDDEV_FLAGS);
@@ -8726,7 +8749,7 @@ static void *raid5_takeover_raid6(struct mddev *mddev)
mddev->new_layout = new_layout;
mddev->delta_disks = -1;
mddev->raid_disks -= 1;
- return setup_conf(mddev);
+ return setup_conf_for_takeover(mddev);
}
static int raid5_check_reshape(struct mddev *mddev)
@@ -8804,7 +8827,7 @@ static void *raid5_takeover(struct mddev *mddev)
if (mddev->level == 4) {
mddev->new_layout = ALGORITHM_PARITY_N;
mddev->new_level = 5;
- return setup_conf(mddev);
+ return setup_conf_for_takeover(mddev);
}
if (mddev->level == 6)
return raid5_takeover_raid6(mddev);
@@ -8824,7 +8847,7 @@ static void *raid4_takeover(struct mddev *mddev)
mddev->layout == ALGORITHM_PARITY_N) {
mddev->new_layout = 0;
mddev->new_level = 4;
- return setup_conf(mddev);
+ return setup_conf_for_takeover(mddev);
}
return ERR_PTR(-EINVAL);
}
@@ -8874,7 +8897,7 @@ static void *raid6_takeover(struct mddev *mddev)
mddev->new_layout = new_layout;
mddev->delta_disks = 1;
mddev->raid_disks += 1;
- return setup_conf(mddev);
+ return setup_conf_for_takeover(mddev);
}
static int raid5_change_consistency_policy(struct mddev *mddev, const char *buf)
--
2.34.1
reply other threads:[~2024-03-26 11:18 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=df3a5e85db254f2ab44b93734cac27bb@kioxia.com \
--to=keisuke1.tada@kioxia.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-raid@vger.kernel.org \
--cc=song@kernel.org \
--cc=yukuai3@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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 read-only IMAP folder(s) and NNTP newsgroup(s).