mdadm-3.1.1 segfaults when mdadm -G -Z

mdadm-3.1.1 segfaults when mdadm -G -Z

[Marcin Mirosław]

Hello,
I tried to grow array size (i resized /dev/sdc5-> resync->remove
/dev/sda6 -> resize /dev/sda6 -> resync-> curent situation )
and i got segfault:

# cat /proc/mdstat
md6 : active raid1 sdc5[1](W) sda7[0]
      125001664 blocks [2/2] [UU]
      bitmap: 16/239 pages [64KB], 256KB chunk

# gdb /sbin/mdadm
[...]
<http://bugs.gentoo.org/>...
Reading symbols from
/var/tmp/portage/sys-fs/mdadm-3.1.1/work/mdadm-3.1.1/mdadm...done.
(gdb) run  -G -Z /dev/md6
Starting program:
/var/tmp/portage/sys-fs/mdadm-3.1.1/work/mdadm-3.1.1/mdadm -G -Z /dev/md6

Program received signal SIGSEGV, Segmentation fault.
0xb7f067f8 in strcmp () from /lib/libc.so.6
(gdb) bt
#0  0xb7f067f8 in strcmp () from /lib/libc.so.6
#1  0x0804b70b in main (argc=4, argv=0xbffff3f4) at mdadm.c:408
(gdb)

$  uname -r
2.6.32-gentoo-r1

May i get any suggestions?
Regards

Re: mdadm-3.1.1 segfaults when mdadm -G -Z

[Zdenek Behan]

On 01/10/2010 01:13 PM, Marcin Mirosław wrote:
> Hello,
> I tried to grow array size (i resized /dev/sdc5-> resync->remove
> /dev/sda6 -> resize /dev/sda6 -> resync-> curent situation )
> and i got segfault:
>
> # cat /proc/mdstat
> md6 : active raid1 sdc5[1](W) sda7[0]
>       125001664 blocks [2/2] [UU]
>       bitmap: 16/239 pages [64KB], 256KB chunk
>
> # gdb /sbin/mdadm
> [...]
> <http://bugs.gentoo.org/>...
> Reading symbols from
> /var/tmp/portage/sys-fs/mdadm-3.1.1/work/mdadm-3.1.1/mdadm...done.
> (gdb) run  -G -Z /dev/md6
> Starting program:
> /var/tmp/portage/sys-fs/mdadm-3.1.1/work/mdadm-3.1.1/mdadm -G -Z /dev/md6
>
> Program received signal SIGSEGV, Segmentation fault.
> 0xb7f067f8 in strcmp () from /lib/libc.so.6
> (gdb) bt
> #0  0xb7f067f8 in strcmp () from /lib/libc.so.6
> #1  0x0804b70b in main (argc=4, argv=0xbffff3f4) at mdadm.c:408
> (gdb)
>
> $  uname -r
> 2.6.32-gentoo-r1
>
> May i get any suggestions?
> Regards
>   
       -Z, --array-size=
              This is only meaningful with --grow and its effect is 
not  per‐
              sistent:  when  the  array  is  stopped an restarted the
default
              array size will be restored.

You didn't supply any argument to -Z at all, or to be precise, you
supplied /dev/md6 as the argument, whereas you should have put in the
array size.

Also, another thing I have noticed lately is, that _some_ short options
seem to be somehow oddly broken on my system and not accepting
arguments, even if they are supposed to be. If the segfault persists
after using the correct syntax, try using the long opts instead. I was
planning to dig more into this bug myself, so I might just kick self to
do it now.


Zdenek
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: mdadm-3.1.1 segfaults when mdadm -G -Z

[Marcin Mirosław]

W dniu 2010-01-10 14:47, Zdenek Behan pisze:
>        -Z, --array-size=
>               This is only meaningful with --grow and its effect is 
> not  per‐
>               sistent:  when  the  array  is  stopped an restarted the
> default
>               array size will be restored.

Yes, i missed it.

> You didn't supply any argument to -Z at all, or to be precise, you
> supplied /dev/md6 as the argument, whereas you should have put in the
> array size.
> 
> Also, another thing I have noticed lately is, that _some_ short options
> seem to be somehow oddly broken on my system and not accepting
> arguments, even if they are supposed to be. If the segfault persists
> after using the correct syntax, try using the long opts instead. I was
> planning to dig more into this bug myself, so I might just kick self to
> do it now.

long option, with suplied array size works correctly, so indeed problem
is in short option.
Thank you!
Regards.
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: mdadm-3.1.1 segfaults when mdadm -G -Z

[Luca Berra]

[-- Attachment #1: Type: text/plain, Size: 1233 bytes --]

On Sun, Jan 10, 2010 at 02:54:02PM +0100, Marcin Mirosław wrote:
>W dniu 2010-01-10 14:47, Zdenek Behan pisze:
>>        -Z, --array-size=
>>               This is only meaningful with --grow and its effect is 
>> not  per‐
>>               sistent:  when  the  array  is  stopped an restarted the
>> default
>>               array size will be restored.
>
>Yes, i missed it.
>
>> You didn't supply any argument to -Z at all, or to be precise, you
>> supplied /dev/md6 as the argument, whereas you should have put in the
>> array size.
>> 
>> Also, another thing I have noticed lately is, that _some_ short options
>> seem to be somehow oddly broken on my system and not accepting
>> arguments, even if they are supposed to be. If the segfault persists
>> after using the correct syntax, try using the long opts instead. I was
>> planning to dig more into this bug myself, so I might just kick self to
>> do it now.
>
>long option, with suplied array size works correctly, so indeed problem
>is in short option.
>Thank you!
>Regards.

attached patch fixes it.

Regards,
L.

-- 
Luca Berra -- bluca@comedia.it
         Communication Media & Services S.r.l.
  /"\
  \ /     ASCII RIBBON CAMPAIGN
   X        AGAINST HTML MAIL
  / \

[-- Attachment #2: 0002-fix-segfault-when-Z-was-given-without-argument-repor.patch --]
[-- Type: text/plain, Size: 1381 bytes --]

From d61bcf0c5f425bd4009f90cc1d467d933fde0a2c Mon Sep 17 00:00:00 2001
From: Luca Berra <bluca@vodka.it>
Date: Sat, 30 Jan 2010 10:46:07 +0100
Subject: [PATCH 2/2] =?UTF-8?q?fix=20segfault=20when=20-Z=20was=20given=20without=20argument=20(reported=20by=20Marcin
=20Miros=C5=82aw)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit


Signed-off-by: Luca Berra <bluca@vodka.it>
---
 ReadMe.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/ReadMe.c b/ReadMe.c
index cd7b685..45619a4 100644
--- a/ReadMe.c
+++ b/ReadMe.c
@@ -86,11 +86,11 @@ char Version[] = Name " - v3.1.1 - 19th November 2009\n";
  *     At the time if writing, there is only minimal support.
  */
 
-char short_options[]="-ABCDEFGIQhVXWZvqbc:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:";
+char short_options[]="-ABCDEFGIQhVXWZ:vqbc:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:";
 char short_bitmap_options[]=
-                   "-ABCDEFGIQhVXWZvqb:c:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:";
+                   "-ABCDEFGIQhVXWZ:vqb:c:i:l:p:m:n:x:u:c:d:z:U:N:sarfRSow1tye:";
 char short_bitmap_auto_options[]=
-                   "-ABCDEFGIQhVXWZvqb:c:i:l:p:m:n:x:u:c:d:z:U:N:sa:rfRSow1tye:";
+                   "-ABCDEFGIQhVXWZ:vqb:c:i:l:p:m:n:x:u:c:d:z:U:N:sa:rfRSow1tye:";
 
 struct option long_options[] = {
     {"manage",    0, 0, '@'},
-- 
1.6.6.1

Re: mdadm-3.1.1 segfaults when mdadm -G -Z

[Neil Brown]

On Sat, 30 Jan 2010 10:56:11 +0100
Luca Berra <bluca@comedia.it> wrote:

> On Sun, Jan 10, 2010 at 02:54:02PM +0100, Marcin Mirosław wrote:
> >W dniu 2010-01-10 14:47, Zdenek Behan pisze:
> >>        -Z, --array-size=
> >>               This is only meaningful with --grow and its effect is 
> >> not  per‐
> >>               sistent:  when  the  array  is  stopped an restarted the
> >> default
> >>               array size will be restored.
> >
> >Yes, i missed it.
> >
> >> You didn't supply any argument to -Z at all, or to be precise, you
> >> supplied /dev/md6 as the argument, whereas you should have put in the
> >> array size.
> >> 
> >> Also, another thing I have noticed lately is, that _some_ short options
> >> seem to be somehow oddly broken on my system and not accepting
> >> arguments, even if they are supposed to be. If the segfault persists
> >> after using the correct syntax, try using the long opts instead. I was
> >> planning to dig more into this bug myself, so I might just kick self to
> >> do it now.
> >
> >long option, with suplied array size works correctly, so indeed problem
> >is in short option.
> >Thank you!
> >Regards.
> 
> attached patch fixes it.

Applied, thanks.

NeilBrown
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html