From: Bob Peterson <rpeterso@redhat.com>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] [GFS2 PATCH 06/28] gfs2: Introduce concept of a pending withdraw
Date: Thu, 20 Feb 2020 13:53:07 -0600 [thread overview]
Message-ID: <20200220195329.952027-7-rpeterso@redhat.com> (raw)
In-Reply-To: <20200220195329.952027-1-rpeterso@redhat.com>
File system withdraws can be delayed when inconsistencies are
discovered when we cannot withdraw immediately, for example, when
critical spin_locks are held. But delaying the withdraw can cause
gfs2 to ignore the error and keep running for a short period of time.
For example, an rgrp glock may be dequeued and demoted while there
are still buffers that haven't been properly revoked, due to io
errors writing to the journal.
This patch introduces a new concept of a pending withdraw, which
means an inconsistency has been discovered and we need to withdraw
at the earliest possible opportunity. In these cases, we aren't
quite withdrawn yet, but we still need to not dequeue glocks and
other critical things. If we dequeue the glocks and the withdraw
results in our journal being replayed, the replay could overwrite
data that's been modified by a different node that acquired the
glock in the meantime.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
---
fs/gfs2/incore.h | 1 +
fs/gfs2/log.c | 21 ++++++++-------------
fs/gfs2/util.c | 14 +++++++-------
fs/gfs2/util.h | 12 +++++++++++-
4 files changed, 27 insertions(+), 21 deletions(-)
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 9fd88ed18807..3cd2de3db40a 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -618,6 +618,7 @@ enum {
SDF_FORCE_AIL_FLUSH = 9,
SDF_AIL1_IO_ERROR = 10,
SDF_FS_FROZEN = 11,
+ SDF_WITHDRAWING = 12, /* Will withdraw eventually */
};
enum gfs2_freeze_state {
diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c
index c4c7c013f7a7..d1ab04135b2f 100644
--- a/fs/gfs2/log.c
+++ b/fs/gfs2/log.c
@@ -88,8 +88,7 @@ static void gfs2_remove_from_ail(struct gfs2_bufdata *bd)
static int gfs2_ail1_start_one(struct gfs2_sbd *sdp,
struct writeback_control *wbc,
- struct gfs2_trans *tr,
- bool *withdraw)
+ struct gfs2_trans *tr)
__releases(&sdp->sd_ail_lock)
__acquires(&sdp->sd_ail_lock)
{
@@ -108,7 +107,7 @@ __acquires(&sdp->sd_ail_lock)
!test_and_set_bit(SDF_AIL1_IO_ERROR,
&sdp->sd_flags)) {
gfs2_io_error_bh(sdp, bh);
- *withdraw = true;
+ gfs2_withdraw_delayed(sdp);
}
list_move(&bd->bd_ail_st_list, &tr->tr_ail2_list);
continue;
@@ -149,7 +148,6 @@ void gfs2_ail1_flush(struct gfs2_sbd *sdp, struct writeback_control *wbc)
struct list_head *head = &sdp->sd_ail1_list;
struct gfs2_trans *tr;
struct blk_plug plug;
- bool withdraw = false;
trace_gfs2_ail_flush(sdp, wbc, 1);
blk_start_plug(&plug);
@@ -158,13 +156,12 @@ void gfs2_ail1_flush(struct gfs2_sbd *sdp, struct writeback_control *wbc)
list_for_each_entry_reverse(tr, head, tr_list) {
if (wbc->nr_to_write <= 0)
break;
- if (gfs2_ail1_start_one(sdp, wbc, tr, &withdraw) &&
- !gfs2_withdrawn(sdp))
+ if (gfs2_ail1_start_one(sdp, wbc, tr) && !gfs2_withdrawn(sdp))
goto restart;
}
spin_unlock(&sdp->sd_ail_lock);
blk_finish_plug(&plug);
- if (withdraw)
+ if (test_bit(SDF_WITHDRAWING, &sdp->sd_flags))
gfs2_withdraw(sdp);
trace_gfs2_ail_flush(sdp, wbc, 0);
}
@@ -193,8 +190,7 @@ static void gfs2_ail1_start(struct gfs2_sbd *sdp)
*
*/
-static void gfs2_ail1_empty_one(struct gfs2_sbd *sdp, struct gfs2_trans *tr,
- bool *withdraw)
+static void gfs2_ail1_empty_one(struct gfs2_sbd *sdp, struct gfs2_trans *tr)
{
struct gfs2_bufdata *bd, *s;
struct buffer_head *bh;
@@ -208,7 +204,7 @@ static void gfs2_ail1_empty_one(struct gfs2_sbd *sdp, struct gfs2_trans *tr,
if (!buffer_uptodate(bh) &&
!test_and_set_bit(SDF_AIL1_IO_ERROR, &sdp->sd_flags)) {
gfs2_io_error_bh(sdp, bh);
- *withdraw = true;
+ gfs2_withdraw_delayed(sdp);
}
list_move(&bd->bd_ail_st_list, &tr->tr_ail2_list);
}
@@ -226,11 +222,10 @@ static int gfs2_ail1_empty(struct gfs2_sbd *sdp)
struct gfs2_trans *tr, *s;
int oldest_tr = 1;
int ret;
- bool withdraw = false;
spin_lock(&sdp->sd_ail_lock);
list_for_each_entry_safe_reverse(tr, s, &sdp->sd_ail1_list, tr_list) {
- gfs2_ail1_empty_one(sdp, tr, &withdraw);
+ gfs2_ail1_empty_one(sdp, tr);
if (list_empty(&tr->tr_ail1_list) && oldest_tr)
list_move(&tr->tr_list, &sdp->sd_ail2_list);
else
@@ -239,7 +234,7 @@ static int gfs2_ail1_empty(struct gfs2_sbd *sdp)
ret = list_empty(&sdp->sd_ail1_list);
spin_unlock(&sdp->sd_ail_lock);
- if (withdraw) {
+ if (test_bit(SDF_WITHDRAWING, &sdp->sd_flags)) {
gfs2_lm(sdp, "fatal: I/O error(s)\n");
gfs2_withdraw(sdp);
}
diff --git a/fs/gfs2/util.c b/fs/gfs2/util.c
index ec8e8c5ce848..47cd40de08b1 100644
--- a/fs/gfs2/util.c
+++ b/fs/gfs2/util.c
@@ -249,13 +249,13 @@ void gfs2_io_error_bh_i(struct gfs2_sbd *sdp, struct buffer_head *bh,
const char *function, char *file, unsigned int line,
bool withdraw)
{
- if (!gfs2_withdrawn(sdp))
- fs_err(sdp,
- "fatal: I/O error\n"
- " block = %llu\n"
- " function = %s, file = %s, line = %u\n",
- (unsigned long long)bh->b_blocknr,
- function, file, line);
+ if (gfs2_withdrawn(sdp))
+ return;
+
+ fs_err(sdp, "fatal: I/O error\n"
+ " block = %llu\n"
+ " function = %s, file = %s, line = %u\n",
+ (unsigned long long)bh->b_blocknr, function, file, line);
if (withdraw)
gfs2_withdraw(sdp);
}
diff --git a/fs/gfs2/util.h b/fs/gfs2/util.h
index 572399e75ce6..16b2cc6c4560 100644
--- a/fs/gfs2/util.h
+++ b/fs/gfs2/util.h
@@ -172,13 +172,23 @@ static inline unsigned int gfs2_tune_get_i(struct gfs2_tune *gt,
return x;
}
+/**
+ * gfs2_withdraw_delayed - withdraw as soon as possible without deadlocks
+ * @sdp: the superblock
+ */
+static inline void gfs2_withdraw_delayed(struct gfs2_sbd *sdp)
+{
+ set_bit(SDF_WITHDRAWING, &sdp->sd_flags);
+}
+
/**
* gfs2_withdrawn - test whether the file system is withdrawing or withdrawn
* @sdp: the superblock
*/
static inline bool gfs2_withdrawn(struct gfs2_sbd *sdp)
{
- return test_bit(SDF_WITHDRAWN, &sdp->sd_flags);
+ return test_bit(SDF_WITHDRAWN, &sdp->sd_flags) ||
+ test_bit(SDF_WITHDRAWING, &sdp->sd_flags);
}
#define gfs2_tune_get(sdp, field) \
--
2.24.1
next prev parent reply other threads:[~2020-02-20 19:53 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-02-20 19:53 [Cluster-devel] [GFS2 PATCH 00/28] GFS2 recovery patches v10 Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 01/28] gfs2: Split gfs2_lm_withdraw into two functions Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 02/28] gfs2: Report errors before withdraw Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 03/28] gfs2: Remove usused cluster_wide arguments of gfs2_consist functions Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 04/28] gfs2: Turn gfs2_consist into void functions Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 05/28] gfs2: Return bool from gfs2_assert functions Bob Peterson
2020-02-20 19:53 ` Bob Peterson [this message]
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 07/28] gfs2: clear ail1 list when gfs2 withdraws Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 08/28] gfs2: Rework how rgrp buffer_heads are managed Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 09/28] gfs2: log error reform Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 10/28] gfs2: Only complain the first time an io error occurs in quota or log Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 11/28] gfs2: Ignore dlm recovery requests if gfs2 is withdrawn Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 12/28] gfs2: move check_journal_clean to util.c for future use Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 13/28] gfs2: Allow some glocks to be used during withdraw Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 14/28] gfs2: Force withdraw to replay journals and wait for it to finish Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 15/28] gfs2: fix infinite loop when checking ail item count before go_inval Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 16/28] gfs2: Add verbose option to check_journal_clean Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 17/28] gfs2: Issue revokes more intelligently Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 18/28] gfs2: Prepare to withdraw as soon as an IO error occurs in log write Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 19/28] gfs2: Check for log write errors before telling dlm to unlock Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 20/28] gfs2: Do log_flush in gfs2_ail_empty_gl even if ail list is empty Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 21/28] gfs2: Withdraw in gfs2_ail1_flush if write_cache_pages fails Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 22/28] gfs2: drain the ail2 list after io errors Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 23/28] gfs2: Don't demote a glock until its revokes are written Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 24/28] gfs2: Do proper error checking for go_sync family of glops functions Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 25/28] gfs2: flesh out delayed withdraw for gfs2_log_flush Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 26/28] fs: clean up __block_commit_write Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 27/28] gfs2: don't allow releasepage to free bd still used for revokes Bob Peterson
2020-02-20 19:53 ` [Cluster-devel] [GFS2 PATCH 28/28] gfs2: allow journal replay to hold sd_log_flush_lock Bob Peterson
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=20200220195329.952027-7-rpeterso@redhat.com \
--to=rpeterso@redhat.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 NNTP newsgroup(s).