Re: [Qemu-devel] qcow crypto changes crash on migration (was: [PULL 10/12] block: convert qcow/qcow2 to use generic cipher API)

[Aurelien Jarno]

On 2015-07-09 12:53, Christian Borntraeger wrote:
> Forgot some CCs (patch author and migration folks)
> 
> 
> Am 09.07.2015 um 12:17 schrieb Christian Borntraeger:
> > Am 07.07.2015 um 16:12 schrieb Paolo Bonzini:
> >> From: "Daniel P. Berrange" <address@hidden>
> >>
> >> Switch the qcow/qcow2 block driver over to use the generic cipher
> >> API, this allows it to use the pluggable AES implementations,
> >> instead of being hardcoded to use QEMU's built-in impl.
> >>
> >> Signed-off-by: Daniel P. Berrange <address@hidden>
> >> Message-Id: <address@hidden>
> >> Signed-off-by: Paolo Bonzini <address@hidden>
> > 
> > For whatever reason this breaks migration(or virsh restore)
> > from guests that were created with an older version of QEMU.
> > 
> > 
> > 
> > Thread 1 (Thread 0x3fffb856bd0 (LWP 32226)):
> > #0  0x0000000080354a54 in qcrypto_cipher_free (cipher=0x0) at 
> > /home/cborntra/REPOS/qemu/crypto/cipher-builtin.c:357
> > #1  0x00000000802ca912 in qcow2_close (bs=0x80a0cea0) at 
> > /home/cborntra/REPOS/qemu/block/qcow2.c:1477
> > #2  0x00000000802caa32 in qcow2_invalidate_cache (bs=0x80a0cea0, 
> > errp=0x81a3fc58) at /home/cborntra/REPOS/qemu/block/qcow2.c:1509
> > #3  0x000000008029bac0 in bdrv_invalidate_cache (bs=0x80a0cea0, 
> > errp=0x81a3fd08) at /home/cborntra/REPOS/qemu/block.c:3135
> > #4  0x000000008029bbe6 in bdrv_invalidate_cache_all (errp=0x81a3fdd0) at 
> > /home/cborntra/REPOS/qemu/block.c:3160
> > #5  0x000000008021de50 in process_incoming_migration_co (opaque=0x80acaae0) 
> > at /home/cborntra/REPOS/qemu/migration/migration.c:160
> > #6  0x00000000802ab96a in coroutine_trampoline (i0=0, i1=-2137149984) at 
> > /home/cborntra/REPOS/qemu/coroutine-ucontext.c:80
> > #7  0x000003fffc463ca2 in __makecontext_ret () from /lib64/libc.so.6
> > Backtrace stopped: previous frame identical to this frame (corrupt stack?)

This is the same kind of backtrace I got on a MIPS host (see my other
mail). The reason is that a NULL pointer is dereferenced before testing
it is non NULL in qcrypto_cipher_free.

-- 
Aurelien Jarno                          GPG: 4096R/1DDD8C9B
address@hidden                 http://www.aurel32.net