From: Steven Whitehouse <swhiteho@redhat.com>
To: linux-kernel@vger.kernel.org, cluster-devel@redhat.com
Cc: Bob Peterson <rpeterso@redhat.com>,
Steven Whitehouse <swhiteho@redhat.com>
Subject: [PATCH 17/29] GFS2: Move recovery variables to journal structure in memory
Date: Tue, 1 Apr 2014 10:15:29 +0100 [thread overview]
Message-ID: <1396343741-1524-18-git-send-email-swhiteho@redhat.com> (raw)
In-Reply-To: <1396343741-1524-1-git-send-email-swhiteho@redhat.com>
From: Bob Peterson <rpeterso@redhat.com>
If multiple nodes fail and their recovery work runs simultaneously, they
would use the same unprotected variables in the superblock. For example,
they would stomp on each other's revoked blocks lists, which resulted
in file system metadata corruption. This patch moves the necessary
variables so that each journal has its own separate area for tracking
its journal replay.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 456d8fa..ef26ed9 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -503,6 +503,15 @@ struct gfs2_jdesc {
unsigned int jd_jid;
unsigned int jd_blocks;
int jd_recover_error;
+ /* Replay stuff */
+
+ unsigned int jd_found_blocks;
+ unsigned int jd_found_revokes;
+ unsigned int jd_replayed_blocks;
+
+ struct list_head jd_revoke_list;
+ unsigned int jd_replay_tail;
+
};
struct gfs2_statfs_change_host {
@@ -782,15 +791,6 @@ struct gfs2_sbd {
struct list_head sd_ail1_list;
struct list_head sd_ail2_list;
- /* Replay stuff */
-
- struct list_head sd_revoke_list;
- unsigned int sd_replay_tail;
-
- unsigned int sd_found_blocks;
- unsigned int sd_found_revokes;
- unsigned int sd_replayed_blocks;
-
/* For quiescing the filesystem */
struct gfs2_holder sd_freeze_gh;
diff --git a/fs/gfs2/lops.c b/fs/gfs2/lops.c
index ae1d635..a294d8d 100644
--- a/fs/gfs2/lops.c
+++ b/fs/gfs2/lops.c
@@ -520,13 +520,11 @@ static void buf_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr)
static void buf_lo_before_scan(struct gfs2_jdesc *jd,
struct gfs2_log_header_host *head, int pass)
{
- struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
-
if (pass != 0)
return;
- sdp->sd_found_blocks = 0;
- sdp->sd_replayed_blocks = 0;
+ jd->jd_found_blocks = 0;
+ jd->jd_replayed_blocks = 0;
}
static int buf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
@@ -549,9 +547,9 @@ static int buf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
for (; blks; gfs2_replay_incr_blk(sdp, &start), blks--) {
blkno = be64_to_cpu(*ptr++);
- sdp->sd_found_blocks++;
+ jd->jd_found_blocks++;
- if (gfs2_revoke_check(sdp, blkno, start))
+ if (gfs2_revoke_check(jd, blkno, start))
continue;
error = gfs2_replay_read_block(jd, start, &bh_log);
@@ -572,7 +570,7 @@ static int buf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
if (error)
break;
- sdp->sd_replayed_blocks++;
+ jd->jd_replayed_blocks++;
}
return error;
@@ -615,7 +613,7 @@ static void buf_lo_after_scan(struct gfs2_jdesc *jd, int error, int pass)
gfs2_meta_sync(ip->i_gl);
fs_info(sdp, "jid=%u: Replayed %u of %u blocks\n",
- jd->jd_jid, sdp->sd_replayed_blocks, sdp->sd_found_blocks);
+ jd->jd_jid, jd->jd_replayed_blocks, jd->jd_found_blocks);
}
static void revoke_lo_before_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr)
@@ -677,13 +675,11 @@ static void revoke_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr)
static void revoke_lo_before_scan(struct gfs2_jdesc *jd,
struct gfs2_log_header_host *head, int pass)
{
- struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
-
if (pass != 0)
return;
- sdp->sd_found_revokes = 0;
- sdp->sd_replay_tail = head->lh_tail;
+ jd->jd_found_revokes = 0;
+ jd->jd_replay_tail = head->lh_tail;
}
static int revoke_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
@@ -715,13 +711,13 @@ static int revoke_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
while (offset + sizeof(u64) <= sdp->sd_sb.sb_bsize) {
blkno = be64_to_cpu(*(__be64 *)(bh->b_data + offset));
- error = gfs2_revoke_add(sdp, blkno, start);
+ error = gfs2_revoke_add(jd, blkno, start);
if (error < 0) {
brelse(bh);
return error;
}
else if (error)
- sdp->sd_found_revokes++;
+ jd->jd_found_revokes++;
if (!--revokes)
break;
@@ -741,16 +737,16 @@ static void revoke_lo_after_scan(struct gfs2_jdesc *jd, int error, int pass)
struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
if (error) {
- gfs2_revoke_clean(sdp);
+ gfs2_revoke_clean(jd);
return;
}
if (pass != 1)
return;
fs_info(sdp, "jid=%u: Found %u revoke tags\n",
- jd->jd_jid, sdp->sd_found_revokes);
+ jd->jd_jid, jd->jd_found_revokes);
- gfs2_revoke_clean(sdp);
+ gfs2_revoke_clean(jd);
}
/**
@@ -789,9 +785,9 @@ static int databuf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
blkno = be64_to_cpu(*ptr++);
esc = be64_to_cpu(*ptr++);
- sdp->sd_found_blocks++;
+ jd->jd_found_blocks++;
- if (gfs2_revoke_check(sdp, blkno, start))
+ if (gfs2_revoke_check(jd, blkno, start))
continue;
error = gfs2_replay_read_block(jd, start, &bh_log);
@@ -811,7 +807,7 @@ static int databuf_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start,
brelse(bh_log);
brelse(bh_ip);
- sdp->sd_replayed_blocks++;
+ jd->jd_replayed_blocks++;
}
return error;
@@ -835,7 +831,7 @@ static void databuf_lo_after_scan(struct gfs2_jdesc *jd, int error, int pass)
gfs2_meta_sync(ip->i_gl);
fs_info(sdp, "jid=%u: Replayed %u of %u data blocks\n",
- jd->jd_jid, sdp->sd_replayed_blocks, sdp->sd_found_blocks);
+ jd->jd_jid, jd->jd_replayed_blocks, jd->jd_found_blocks);
}
static void databuf_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr)
diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index c3ef844..ea9c35c 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -128,8 +128,6 @@ static struct gfs2_sbd *init_sbd(struct super_block *sb)
atomic_set(&sdp->sd_log_in_flight, 0);
init_waitqueue_head(&sdp->sd_log_flush_wait);
- INIT_LIST_HEAD(&sdp->sd_revoke_list);
-
return sdp;
}
@@ -575,6 +573,8 @@ static int gfs2_jindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ji_gh)
break;
INIT_LIST_HEAD(&jd->extent_list);
+ INIT_LIST_HEAD(&jd->jd_revoke_list);
+
INIT_WORK(&jd->jd_work, gfs2_recover_func);
jd->jd_inode = gfs2_lookupi(sdp->sd_jindex, &name, 1);
if (!jd->jd_inode || IS_ERR(jd->jd_inode)) {
diff --git a/fs/gfs2/recovery.c b/fs/gfs2/recovery.c
index 963b2d7..7ad4094 100644
--- a/fs/gfs2/recovery.c
+++ b/fs/gfs2/recovery.c
@@ -52,9 +52,9 @@ int gfs2_replay_read_block(struct gfs2_jdesc *jd, unsigned int blk,
return error;
}
-int gfs2_revoke_add(struct gfs2_sbd *sdp, u64 blkno, unsigned int where)
+int gfs2_revoke_add(struct gfs2_jdesc *jd, u64 blkno, unsigned int where)
{
- struct list_head *head = &sdp->sd_revoke_list;
+ struct list_head *head = &jd->jd_revoke_list;
struct gfs2_revoke_replay *rr;
int found = 0;
@@ -81,13 +81,13 @@ int gfs2_revoke_add(struct gfs2_sbd *sdp, u64 blkno, unsigned int where)
return 1;
}
-int gfs2_revoke_check(struct gfs2_sbd *sdp, u64 blkno, unsigned int where)
+int gfs2_revoke_check(struct gfs2_jdesc *jd, u64 blkno, unsigned int where)
{
struct gfs2_revoke_replay *rr;
int wrap, a, b, revoke;
int found = 0;
- list_for_each_entry(rr, &sdp->sd_revoke_list, rr_list) {
+ list_for_each_entry(rr, &jd->jd_revoke_list, rr_list) {
if (rr->rr_blkno == blkno) {
found = 1;
break;
@@ -97,17 +97,17 @@ int gfs2_revoke_check(struct gfs2_sbd *sdp, u64 blkno, unsigned int where)
if (!found)
return 0;
- wrap = (rr->rr_where < sdp->sd_replay_tail);
- a = (sdp->sd_replay_tail < where);
+ wrap = (rr->rr_where < jd->jd_replay_tail);
+ a = (jd->jd_replay_tail < where);
b = (where < rr->rr_where);
revoke = (wrap) ? (a || b) : (a && b);
return revoke;
}
-void gfs2_revoke_clean(struct gfs2_sbd *sdp)
+void gfs2_revoke_clean(struct gfs2_jdesc *jd)
{
- struct list_head *head = &sdp->sd_revoke_list;
+ struct list_head *head = &jd->jd_revoke_list;
struct gfs2_revoke_replay *rr;
while (!list_empty(head)) {
diff --git a/fs/gfs2/recovery.h b/fs/gfs2/recovery.h
index 2226136..6142836 100644
--- a/fs/gfs2/recovery.h
+++ b/fs/gfs2/recovery.h
@@ -23,9 +23,9 @@ static inline void gfs2_replay_incr_blk(struct gfs2_sbd *sdp, unsigned int *blk)
extern int gfs2_replay_read_block(struct gfs2_jdesc *jd, unsigned int blk,
struct buffer_head **bh);
-extern int gfs2_revoke_add(struct gfs2_sbd *sdp, u64 blkno, unsigned int where);
-extern int gfs2_revoke_check(struct gfs2_sbd *sdp, u64 blkno, unsigned int where);
-extern void gfs2_revoke_clean(struct gfs2_sbd *sdp);
+extern int gfs2_revoke_add(struct gfs2_jdesc *jd, u64 blkno, unsigned int where);
+extern int gfs2_revoke_check(struct gfs2_jdesc *jd, u64 blkno, unsigned int where);
+extern void gfs2_revoke_clean(struct gfs2_jdesc *jd);
extern int gfs2_find_jhead(struct gfs2_jdesc *jd,
struct gfs2_log_header_host *head);
--
1.8.3.1
next prev parent reply other threads:[~2014-04-01 9:21 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-04-01 9:15 GFS2: Pre-pull patch posting (merge window) Steven Whitehouse
2014-04-01 9:15 ` [PATCH 01/29] GFS2: Plug on AIL flush Steven Whitehouse
2014-04-01 9:15 ` [PATCH 02/29] GFS2: Allocate block for xattr at inode alloc time, if required Steven Whitehouse
2014-04-01 9:15 ` [PATCH 03/29] GFS2: journal data writepages update Steven Whitehouse
2014-04-01 9:15 ` [PATCH 04/29] GFS2: Lock i_mutex and use a local gfs2_holder for fallocate Steven Whitehouse
2014-04-01 9:15 ` [PATCH 05/29] GFS2: Add meta readahead field in directory entries Steven Whitehouse
2014-04-01 9:15 ` [PATCH 06/29] GFS2: Mark functions as static in gfs2/rgrp.c Steven Whitehouse
2014-04-01 9:15 ` [PATCH 07/29] GFS2: add missing newline Steven Whitehouse
2014-04-01 9:15 ` [PATCH 08/29] GFS2: Reduce struct gfs2_trans in size Steven Whitehouse
2014-04-01 9:15 ` [PATCH 09/29] GFS2: Move log buffer lists into transaction Steven Whitehouse
2014-04-01 9:15 ` [PATCH 10/29] GFS2: Move log buffer accounting to transaction Steven Whitehouse
2014-04-01 9:15 ` [PATCH 11/29] fs: NULL dereference in posix_acl_to_xattr() Steven Whitehouse
2014-04-01 9:15 ` [PATCH 12/29] GFS2: Remove extra "if" in gfs2_log_flush() Steven Whitehouse
2014-04-01 9:15 ` [PATCH 13/29] GFS2: replace kmalloc - __vmalloc / memset 0 Steven Whitehouse
2014-04-01 9:15 ` [PATCH 14/29] GFS2: Clean up journal extent mapping Steven Whitehouse
2014-04-01 9:15 ` [PATCH 15/29] GFS2: return -E2BIG if hit the maximum limits of ACLs Steven Whitehouse
2014-04-01 9:15 ` [PATCH 16/29] GFS2: global conversion to pr_foo() Steven Whitehouse
2014-04-01 9:15 ` Steven Whitehouse [this message]
2014-04-01 9:15 ` [PATCH 18/29] GFS2: Use pr_<level> more consistently Steven Whitehouse
2014-04-01 9:15 ` [PATCH 19/29] GFS2: Use fs_<level> more often Steven Whitehouse
2014-04-01 9:15 ` [PATCH 20/29] GFS2: Convert gfs2_lm_withdraw to use fs_err Steven Whitehouse
2014-04-01 9:15 ` [PATCH 21/29] GFS2: check NULL return value in gfs2_ok_to_move Steven Whitehouse
2014-04-01 9:15 ` [PATCH 22/29] GFS2: Ensure workqueue is scheduled after noexp request Steven Whitehouse
2014-04-01 9:15 ` [PATCH 23/29] GFS2: Re-add a call to log_flush_wait when flushing the journal Steven Whitehouse
2014-04-01 9:15 ` [PATCH 24/29] GFS2: Increase the max number of ACLs Steven Whitehouse
2014-04-01 9:15 ` [PATCH 25/29] GFS2: Remove extraneous function gfs2_security_init Steven Whitehouse
2014-04-01 9:15 ` [PATCH 26/29] GFS2: inline function gfs2_set_mode Steven Whitehouse
2014-04-01 9:15 ` [PATCH 27/29] GFS2: Fix return value in slot_get() Steven Whitehouse
2014-04-01 9:15 ` [PATCH 28/29] GFS2: Fix uninitialized VFS inode in gfs2_create_inode Steven Whitehouse
2014-04-01 9:15 ` [PATCH 29/29] GFS2: Fix address space from page function Steven Whitehouse
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=1396343741-1524-18-git-send-email-swhiteho@redhat.com \
--to=swhiteho@redhat.com \
--cc=cluster-devel@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=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
all inboxes | Powered by JetHome®