Re: [Qemu-devel] Fwd: qemu drive mirror assert fault

[Fam Zheng <famz@redhat.com>]

On Thu, 04/30 17:50, Paolo Bonzini wrote:
> John, Fam,
> 
> I got this report offlist.  This happens if a bit in the hbitmap is
> cleared and the HBitmap has _not_ yet reached the bit.  See this comment
> in include/qemu/hbitmap.h:
> 
>  * Resetting bits before the current
>  * position of the iterator is also okay.  However, concurrent
>  * resetting of bits can lead to unexpected behavior if the iterator
>  * has not yet reached those bits.
> 
> Can you please take a look?

Since the gdb output is suggesting 1.5.3, it's worth to trying 2.3 which has
this:

    commit c4237dfa635900e4d1cdc6038d5efe3507f45f0c
    Author: Vladimir Sementsov-Ogievskiy <vsementsov@parallels.com>
    Date:   Thu Nov 27 12:40:46 2014 +0300

        block: fix spoiling all dirty bitmaps by mirror and migration

        Mirror and migration use dirty bitmaps for their purposes, and since
        commit [block: per caller dirty bitmap] they use their own bitmaps, not
        the global one. But they use old functions bdrv_set_dirty and
        bdrv_reset_dirty, which change all dirty bitmaps.

        Named dirty bitmaps series by Fam and Snow are affected: mirroring and
        migration will spoil all (not related to this mirroring or migration)
        named dirty bitmaps.

        This patch fixes this by adding bdrv_set_dirty_bitmap and
        bdrv_reset_dirty_bitmap, which change concrete bitmap. Also, to prevent
        such mistakes in future, old functions bdrv_(set,reset)_dirty are made
        static, for internal block usage.

        Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@parallels.com>
        CC: John Snow <jsnow@redhat.com>
        CC: Fam Zheng <famz@redhat.com>
        CC: Denis V. Lunev <den@openvz.org>
        CC: Stefan Hajnoczi <stefanha@redhat.com>
        CC: Kevin Wolf <kwolf@redhat.com>
        Reviewed-by: John Snow <jsnow@redhat.com>
        Reviewed-by: Fam Zheng <famz@redhat.com>
        Message-id: 1417081246-3593-1-git-send-email-vsementsov@parallels.com
        Signed-off-by: Max Reitz <mreitz@redhat.com>

Fam

> 
> Thanks,
> 
> Paolo
> 
> -------- Forwarded Message --------
> Subject: 	qemu drive mirror assert fault
> Date: 	Wed, 29 Apr 2015 10:50:28 +0800
> From: 	wangxiaolong <wangxiaolong@ucloud.cn>
> To: 	pbonzini <pbonzini@redhat.com>
> 
> hello,
> 
> I used drive mirror to do live migration, and I run into such an assert
> fault:
> 
> (gdb) bt
> 
> #0  0x00007fd2c6e678a5 in raise (sig=6) at
> ../nptl/sysdeps/unix/sysv/linux/raise.c:64
> 
> #1  0x00007fd2c6e69085 in abort () at abort.c:92
> 
> #2  0x00007fd2c6e60a1e in __assert_fail_base (fmt=<value optimized out>,
> assertion=0x7fd2ca215aa0 "cur", file=0x7fd2ca215a78 "util/hbitmap.c",
> line=<value optimized out>,
> 
>     function=<value optimized out>) at assert.c:96
> 
> #3  0x00007fd2c6e60ae0 in __assert_fail (assertion=0x7fd2ca215aa0 "cur",
> file=0x7fd2ca215a78 "util/hbitmap.c", line=129, function=0x7fd2ca215bf0
> "hbitmap_iter_skip_words")
> 
>     at assert.c:105
> 
> #4  0x00007fd2ca1b3bb8 in hbitmap_iter_skip_words (hbi=<value optimized
> out>) at util/hbitmap.c:129
> 
> #5  0x00007fd2c9f8f8e0 in hbitmap_iter_next (opaque=0x7fd2cc59c730) at
> /usr/src/debug/qemu-kvm-1.5.3/include/qemu/hbitmap.h:166
> 
> #6  mirror_iteration (opaque=0x7fd2cc59c730) at block/mirror.c:163
> 
> #7  mirror_run (opaque=0x7fd2cc59c730) at block/mirror.c:407
> 
> #8  0x00007fd2c9fc45bb in coroutine_trampoline (i0=<value optimized
> out>, i1=<value optimized out>) at coroutine-ucontext.c:118
> 
> #9  0x00007fd2c6e78b70 in ?? () from /lib64/libc-2.12.so
> 
> #10 0x00007fff53eede80 in ?? ()
> 
> #11 0x0000000000000000 in ?? ()
> 
> 
> and I just can’t figure out what is the cause of this situation,
> could you help me figure it out, thanks!
> 
> 
> 
>