patch md-linear-fix-a-division-by-zero-bug-for-very-small-arrays.patch added to 2.6.27-stable tree
A message that is not in the archive (sent off-list or not yet received): [email protected]
Hi
Quoted 51 lines · On Dienstag, 11. November 2008, [email protected] wrote:
This is a note to let you know that we have just queued up the patch titled
Subject: md: linear: Fix a division by zero bug for very small arrays.
Subject: md: linear: Fix a division by zero bug for very small arrays.to the 2.6.27-stable tree. Its filename is
md-linear-fix-a-division-by-zero-bug-for-very-small-arrays.patch
A git repo of this tree can be found at
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summaryFrom [email protected] Tue Nov 11 09:47:32 2008
From: Andre Noll <[email protected]>
Date: Fri, 7 Nov 2008 00:07:46 GMT
Subject: md: linear: Fix a division by zero bug for very small arrays.
To: [email protected]
Message-ID: <[email protected]>From: Andre Noll <[email protected]>
commit
f1cd14ae52985634d0389e934eba25b5ecf24565upstreamDate: Thu, 6 Nov 2008 19:41:24 +1100
Subject: md: linear: Fix a division by zero bug for very small arrays.We currently oops with a divide error on starting a linear software raid array consisting of at least two very small (< 500K) devices.
The bug is caused by the calculation of the hash table size which tries to compute
sector_div(sz, base)with "base" being zero due to the small size of the component devices of the array.Fix this by requiring the hash spacing to be at least one which implies that also "base" is non-zero.
This bug has existed since about 2.6.14.
Signed-off-by Andre Noll <[email protected]>Signed-off-by NeilBrown <[email protected]>Signed-off-by Greg Kroah-Hartman <[email protected]>--- drivers/md/linear.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/md/linear.c static linear_conf_t *linear_conf(mddev_ · line 164
Quoted 8 lines
drivers/md/linear.c +2
-0@@ -157,6 +157,8 @@ static linear_conf_t *linear_conf(mddev_ min_spacing = conf->array_sectors /2 ;sector_div (min_spacing,PAGE_SIZE /sizeof (struct dev_info *)); +if (min_sectors ==0 ) + min_sectors =1 ;/* min_spacing is the minimum spacing that will fit the hash * table in one PAGE. This may be much smaller than needed.
drivers/md/linear.c: In function 'linear_conf':drivers/md/linear.c:160: error: 'min_sectors' undeclared (first use in this function) drivers/md/linear.c:160: error: (Each undeclared identifier is reported only once drivers/md/linear.c:160: error: for each function it appears in.) make[5]: *** [drivers/md/linear.o] Error 1
make[4]: *** [drivers/md] Error 2
make[4]: *** Waiting for unfinished jobs....
This one obviously depends on:
commit 23242fbb470ff4c8c4d41f178832cf1929273d7d
Author: Andre Noll <[email protected]> Date: Mon Oct 13 11:55:12 2008 +1100
md: linear.c: Make two local variables sector-based.
which in turn depends on previous changes.
Regards
Stefan Lippers-Hollmann
Quoted 29 lines · Stefan Lippers-Hollmann
Hi
drivers/md/linear.c static linear_conf_t *linear_conf(mddev_ · line 164
Quoted 12 lines · On Dienstag, 11. November 2008, [email protected] wrote:
drivers/md/linear.c +2
-0@@ -157,6 +157,8 @@ static linear_conf_t *linear_conf(mddev_ min_spacing = conf->array_sectors /2 ;sector_div (min_spacing,PAGE_SIZE /sizeof (struct dev_info *)); +if (min_sectors ==0 ) + min_sectors =1 ;/* min_spacing is the minimum spacing that will fit the hash * table in one PAGE. This may be much smaller than needed.
drivers/md/linear.c: In function 'linear_conf':drivers/md/linear.c:160: error: 'min_sectors' undeclared (first use in this function)drivers/md/linear.c:160: error: (Each undeclared identifier is reported only oncedrivers/md/linear.c:160: error: for each function it appears in.) make[5]: *** [drivers/md/linear.o] Error 1
make[4]: *** [drivers/md] Error 2
make[4]: *** Waiting for unfinished jobs....This one obviously depends on:
Quoted 7 lines
commit
23242fbb470ff4c8c4d41f178832cf1929273d7dAuthor: Andre Noll <[email protected]> Date: Mon Oct 13 11:55:12 2008 +1100md:
linear.c: Make two local variables sector-based.which in turn depends on previous changes.
Yes, but only syntactically, not semantically.
In 2.6.27, it needs to be
if (min_spacing == 0)
min_spacing = 1;
The other previous patches don't need to be backported.
Thanks for reporting this.
NeilBrown
Quoted 4 lines
Regards
Stefan Lippers-Hollmann
Quoted 36 lines · NeilBrown
Quoted 35 lines · Stefan Lippers-Hollmann
Hi
drivers/md/linear.c static linear_conf_t *linear_conf(mddev_ · line 164
Quoted 12 lines · On Dienstag, 11. November 2008, [email protected] wrote:
drivers/md/linear.c +2
-0@@ -157,6 +157,8 @@ static linear_conf_t *linear_conf(mddev_ min_spacing = conf->array_sectors /2 ;sector_div (min_spacing,PAGE_SIZE /sizeof (struct dev_info *)); +if (min_sectors ==0 ) + min_sectors =1 ;/* min_spacing is the minimum spacing that will fit the hash * table in one PAGE. This may be much smaller than needed.
drivers/md/linear.c: In function 'linear_conf':drivers/md/linear.c:160: error: 'min_sectors' undeclared (first use in this function)drivers/md/linear.c:160: error: (Each undeclared identifier is reported only oncedrivers/md/linear.c:160: error: for each function it appears in.) make[5]: *** [drivers/md/linear.o] Error 1
make[4]: *** [drivers/md] Error 2
make[4]: *** Waiting for unfinished jobs....This one obviously depends on:
commit
23242fbb470ff4c8c4d41f178832cf1929273d7dAuthor: Andre Noll <[email protected]> Date: Mon Oct 13 11:55:12 2008 +1100md:
linear.c: Make two local variables sector-based.
Quoted 7 lines
Quoted 1 line
which in turn depends on previous changes.
Yes, but only syntactically, not semantically.
In 2.6.27, it needs to be
if (min_spacing== 0)min_spacing = 1;
Thanks, I've adjusted the patch to look like this instead.
greg k-h