Andrew,
I haven't done anything about the deadlock yet. This fix took
priority since it can completely destroy the root fs. In fact I
had to reinstall after doing so.
You can import this changeset into BK by piping this whole message to:
'| bk receive [path to repository]' or apply the patch as usual.
===================================================================
ChangeSet@1.550, 2002-05-09 07:40:45-05:00, shaggy@kleikamp.austin.ibm.com
JFS: Flush dirty metadata to disk when remounting from read-write
to read-only. Also fix umount ordering to make sure metadata is
written before journal closed.
With Andrew Morton's recent changes, JFS is no longer writing much
of its dirty metadata when remounting from read-write to read-only.
This causes severe file system corruption. A JFS root file system
will be corrupted on shutdown.
This patch fixes the problem by explicitly writing the dirty metadata
before the journal is closed. It also fixes the ordering so that
the dirty metadata is completely written before the journal is closed
for the normal unmount case as well.
jfs_logmgr.c | 47 ++++++++++++++++++++++++++++++-----------------
jfs_logmgr.h | 1 +
jfs_umount.c | 45 +++++++++++++++++++++++++++++++--------------
3 files changed, 62 insertions(+), 31 deletions(-)
diff -Nru a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c
--- a/fs/jfs/jfs_logmgr.c Thu May 9 08:13:46 2002
+++ b/fs/jfs/jfs_logmgr.c Thu May 9 08:13:46 2002
@@ -1386,6 +1386,35 @@
/*
+ * NAME: lmLogWait()
+ *
+ * FUNCTION: wait for all outstanding log records to be written to disk
+ */
+void lmLogWait(log_t *log)
+{
+ int i;
+
+ jFYI(1, ("lmLogWait: log:0x%p\n", log));
+
+ if (log->cqueue.head || !list_empty(&log->synclist)) {
+ /*
+ * If there was very recent activity, we may need to wait
+ * for the lazycommit thread to catch up
+ */
+
+ for (i = 0; i < 800; i++) { /* Too much? */
+ current->state = TASK_INTERRUPTIBLE;
+ schedule_timeout(HZ / 4);
+ if ((log->cqueue.head == NULL) &&
+ list_empty(&log->synclist))
+ break;
+ }
+ }
+ assert(log->cqueue.head == NULL);
+ assert(list_empty(&log->synclist));
+}
+
+/*
* NAME: lmLogShutdown()
*
* FUNCTION: log shutdown at last LogClose().
@@ -1411,23 +1440,7 @@
jFYI(1, ("lmLogShutdown: log:0x%p\n", log));
- if (log->cqueue.head || !list_empty(&log->synclist)) {
- /*
- * If there was very recent activity, we may need to wait
- * for the lazycommit thread to catch up
- */
- int i;
-
- for (i = 0; i < 800; i++) { /* Too much? */
- current->state = TASK_INTERRUPTIBLE;
- schedule_timeout(HZ / 4);
- if ((log->cqueue.head == NULL) &&
- list_empty(&log->synclist))
- break;
- }
- }
- assert(log->cqueue.head == NULL);
- assert(list_empty(&log->synclist));
+ lmLogWait(log);
/*
* We need to make sure all of the "written" metapages
diff -Nru a/fs/jfs/jfs_logmgr.h b/fs/jfs/jfs_logmgr.h
--- a/fs/jfs/jfs_logmgr.h Thu May 9 08:13:46 2002
+++ b/fs/jfs/jfs_logmgr.h Thu May 9 08:13:46 2002
@@ -489,6 +489,7 @@
}
extern int lmLogOpen(struct super_block *sb, log_t ** log);
+extern void lmLogWait(log_t * log);
extern int lmLogClose(struct super_block *sb, log_t * log);
extern int lmLogSync(log_t * log, int nosyncwait);
extern int lmLogQuiesce(log_t * log);
diff -Nru a/fs/jfs/jfs_umount.c b/fs/jfs/jfs_umount.c
--- a/fs/jfs/jfs_umount.c Thu May 9 08:13:46 2002
+++ b/fs/jfs/jfs_umount.c Thu May 9 08:13:46 2002
@@ -64,15 +64,11 @@
*
* if mounted read-write and log based recovery was enabled
*/
- if ((log = sbi->log)) {
+ if ((log = sbi->log))
/*
- * close log:
- *
- * remove file system from log active file system list.
+ * Wait for outstanding transactions to be written to log:
*/
- log = sbi->log;
- rc = lmLogClose(sb, log);
- }
+ lmLogWait(log);
/*
* close fileset inode allocation map (aka fileset inode)
@@ -113,6 +109,14 @@
sbi->ipimap = NULL;
/*
+ * Make sure all metadata makes it to disk before we mark
+ * the superblock as clean
+ */
+ filemap_fdatawait(sbi->direct_inode->i_mapping);
+ filemap_fdatawrite(sbi->direct_inode->i_mapping);
+ filemap_fdatawait(sbi->direct_inode->i_mapping);
+
+ /*
* ensure all file system file pages are propagated to their
* home blocks on disk (and their in-memory buffer pages are
* invalidated) BEFORE updating file system superblock state
@@ -120,10 +124,16 @@
* consistent state) and log superblock active file system
* list (to signify skip logredo()).
*/
- if (log) /* log = NULL if read-only mount */
+ if (log) { /* log = NULL if read-only mount */
rc = updateSuper(sb, FM_CLEAN);
-
+ /*
+ * close log:
+ *
+ * remove file system from log active file system list.
+ */
+ rc = lmLogClose(sb, log);
+ }
jFYI(0, (" UnMount JFS Complete: %d\n", rc));
return rc;
}
@@ -132,8 +142,9 @@
int jfs_umount_rw(struct super_block *sb)
{
struct jfs_sb_info *sbi = JFS_SBI(sb);
+ log_t *log = sbi->log;
- if (!sbi->log)
+ if (!log)
return 0;
/*
@@ -141,13 +152,19 @@
*
* remove file system from log active file system list.
*/
- lmLogClose(sb, sbi->log);
+ lmLogWait(log);
+ /*
+ * Make sure all metadata makes it to disk
+ */
dbSync(sbi->ipbmap);
diSync(sbi->ipimap);
+ filemap_fdatawait(sbi->direct_inode->i_mapping);
+ filemap_fdatawrite(sbi->direct_inode->i_mapping);
+ filemap_fdatawait(sbi->direct_inode->i_mapping);
- sbi->log = 0;
+ updateSuper(sb, FM_CLEAN);
+ sbi->log = NULL;
- return updateSuper(sb, FM_CLEAN);
-
+ return lmLogClose(sb, log);
}
===================================================================
This BitKeeper patch contains the following changesets:
1.550
## Wrapped with gzip_uu ##
begin 664 bkpatch1793
M'XL(``IVVCP``\U8^W/;-A+^F?PKMNU<:B=Z\*F7SVD<Q[[JZC@9/R9S-YW1
M0"0D,B()%0"MZ*K^[]T%*=F.I2;V76;.EBV+6.P;W[?P#W"MN!Q8*F'3Z=+^
M`7X62@^L6<;3&<OG+58JG1:M=)RW(I'C^H40N-Y.1,[;U:;V>-;^.%%-KQ7:
M*/">Z2B!&R[5P');_N:)7L[YP+HX^<?UV=&%;1\>PG'"BBF_Y!H.#VTMY`W+
M8O6*Z20314M+5JB<:T9V5QO1E><X'GZ';M=WPL[*[3A!=Q6YL>NRP.6QXP6]
M3F!7GKW:$<;GZD*GX_I^WW56@>/W/?L-N*TP=,#QVD[8=OK@=`>!,PC"IA,.
M'`?^6CN\\*'IV*_A?QO2L1W!/T\O!W":E2J!.)5Z":0M9IJA+7RB9K!(>`&2
MYZ(LT*4I3"0Z)#F+FPN9:HXZ4-)\%D6V;`$<94K`)/T$I=D#0L9<TDZ4R]F,
M@RHEO[63*E1!JC3:&?.)P,6/HI0%RR#*A.)Q"P7P]2'5"1P5L>0+>"ND%L6/
M"@U''&U$)G35H'A0(Q0",#]3+HUF,IZ748)*Q`12K3Z/]0LQWH\0M5PE:"/"
M(G$%BF-G<@PXP\B62O,<(B%E.=<I5@C387R2V.1W92CF-,LPX+4TCT$4V`BE
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M./`=SPVZ?G\5]!RO;[C!]1]0@_NUU.!`T^U^$W)XC:`W`U%J6+`*7.CL8C0P
M;+\C4%P?_@HPL_Q,3"]K``-6T&$W@)\BO"!.SIED)%L6$6%CH]KP`54W#,8@
M!!0&$%F6(1IR)K.42X,\N&"$CPE$$+`,+J*86*A*7)([DVT\QB;:*%F[?0>S
M&C#&V&JX(GRB$%&.X*=J_'?0E`OS0CAYOZW#GH!*0]?O]0`K#\_A_.CMR<#:
M)&)O'Q_2\]/K\^.KX;OS@46I-WG',*D62F-F:T>)_A"I%06/B5L#<$W=J*=M
MWX@TODWT'FX::7B.;_OV[[:5(O"F!_:OMO7Q]%_#/;<!>]]OA`=D8N!\^MO\
MU^+[!GW8WS>RZ01(4?-E]%O)2]Y*D!MAM8+OLE3I$<_G>KGWS`BH91'1P_U]
M0&M6^SG^PNB&$V(`3/H"`1_I<[GF<8:=<9/J90-I`'MG"07'3L!X*`O5WC5]
M9.P_2RQDCMG1";&S:4A#E>7<B+;)5XOD]U(X!.<`4O@[]!SZX\4+]`C]@2LA
MS&SP$\E;EA654J(GZ+JF9CV$JZ/+7T;#\ZN3BXOK]U?#UV<G!R2GHH3'9<9'
M.LTYEF7OYW]#&X)]LT@)>I@AA*CSZ[.S?7CVC(0`O_XB821BC>D$DLH_;'HQ
MA:.UWJWYX%9DM^(#^P],#);BC1NX`2!V#%W$)'!MZUZ?H.!V5$V>@JI?20CV
M32K%JQPAK#5790M3O%M7X(5$!'YOY2&254`:/AE'W6\U8A_%=PX@#6]:T,T%
M,::BKB]B3/(4C`GZ+E:4?T+P*V`[",#6(J]GE4<5^7&3TY<N4SL5$^%W<##L
MK;!S'=]4_+]@3BRY&WR3FG^@"T4]V1)6(P;1%87^E.T%5!,OW2K:HK&5MFYG
M=J(W`_>W_$>M8R;'W:VSSML36N=-IXN=,S2_-TB&0*C&:?.E(0$4Z8-O#[O4
M8@:4/ZQ)ZBY!F30RP_5;&(JX!>PW7=]H"HVFS_%GZ+HA]&RR\'9S<R0>W+`[
MS1>*!HSU=;7.FJ$/.3-;B2U4.>=RG(EH1G>,*..LL`U%6'0IR]E\-"%]1#-[
M)E`L!X_T*"U$S)LOTQ&*S#$H@MC[.^B"^,@M7V$$F<L@M.=3+:JW->\2<Q%U
M564AX`=<V5Q0Z^["V'![I]K>@>X=^C57L;H`]*!Z2O??F_L7V6JL0S.&E^^O
M$9FT:IZU+(EX<:=%,;Q_(6SCP!(9>-L/'G78Z0#?]RDWS9J+\S@PH2%"!
M60B";>R$ST-L'A/6UW=(57;<VZ6]_Y_5Q\#[5>!]\&RKG.-&?DDM;!)[^G9T
M?'9R=$ZZUVFL.X'VABYN&B+\4M(DUW@%WUZ:S7_2<):)9JK,#YG7ZSICO-S\
)":.AL&#&$P``
`
end
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/