Re: [Qemu-devel] [PATCH v3] qemu-coroutine: segfault when restarting co_queue

[Kevin Wolf <kwolf@redhat.com>]

Am 10.02.2015 um 06:16 hat Bin Wu geschrieben:
> From: Bin Wu <wu.wubin@huawei.com>
> 
> We tested VMs migration with their disk images by drive_mirror. With
> migration, two VMs copyed large files between each other. During the
> test, a segfault occured. The stack was as follow:
> 
> 00)  0x00007fa5a0c63fc5 in qemu_co_queue_run_restart (co=0x7fa5a1798648) at
> qemu-coroutine-lock.c:66
> 01)  0x00007fa5a0c63bed in coroutine_swap (from=0x7fa5a178f160,
> to=0x7fa5a1798648) at qemu-coroutine.c:97
> 02)  0x00007fa5a0c63dbf in qemu_coroutine_yield () at qemu-coroutine.c:140
> 03)  0x00007fa5a0c9e474 in nbd_co_receive_reply (s=0x7fa5a1a3cfd0,
> request=0x7fa28c2ffa10, reply=0x7fa28c2ffa30, qiov=0x0, offset=0) at
> block/nbd-client.c:165
> 04)  0x00007fa5a0c9e8b5 in nbd_co_writev_1 (client=0x7fa5a1a3cfd0,
> sector_num=8552704, nb_sectors=2040, qiov=0x7fa5a1757468, offset=0) at
> block/nbd-client.c:262
> 05)  0x00007fa5a0c9e9dd in nbd_client_session_co_writev (client=0x7fa5a1a3cfd0,
> sector_num=8552704, nb_sectors=2048, qiov=0x7fa5a1757468) at
> block/nbd-client.c:296
> 06)  0x00007fa5a0c9dda1 in nbd_co_writev (bs=0x7fa5a198fcb0, sector_num=8552704,
> nb_sectors=2048, qiov=0x7fa5a1757468) at block/nbd.c:291
> 07)  0x00007fa5a0c509a4 in bdrv_aligned_pwritev (bs=0x7fa5a198fcb0,
> req=0x7fa28c2ffbb0, offset=4378984448, bytes=1048576, qiov=0x7fa5a1757468,
> flags=0) at block.c:3321
> 08)  0x00007fa5a0c50f3f in bdrv_co_do_pwritev (bs=0x7fa5a198fcb0,
> offset=4378984448, bytes=1048576, qiov=0x7fa5a1757468, flags=(unknown: 0)) at
> block.c:3447
> 09)  0x00007fa5a0c51007 in bdrv_co_do_writev (bs=0x7fa5a198fcb0,
> sector_num=8552704, nb_sectors=2048, qiov=0x7fa5a1757468, flags=(unknown: 0)) at
> block.c:3471
> 10) 0x00007fa5a0c51074 in bdrv_co_writev (bs=0x7fa5a198fcb0, sector_num=8552704,
> nb_sectors=2048, qiov=0x7fa5a1757468) at block.c:3480
> 11) 0x00007fa5a0c652ec in raw_co_writev (bs=0x7fa5a198c110, sector_num=8552704,
> nb_sectors=2048, qiov=0x7fa5a1757468) at block/raw_bsd.c:62
> 12) 0x00007fa5a0c509a4 in bdrv_aligned_pwritev (bs=0x7fa5a198c110,
> req=0x7fa28c2ffe30, offset=4378984448, bytes=1048576, qiov=0x7fa5a1757468,
> flags=0) at block.c:3321
> 13) 0x00007fa5a0c50f3f in bdrv_co_do_pwritev (bs=0x7fa5a198c110,
> offset=4378984448, bytes=1048576, qiov=0x7fa5a1757468, flags=(unknown: 0)) at
> block.c:3447
> 14) 0x00007fa5a0c51007 in bdrv_co_do_writev (bs=0x7fa5a198c110,
> sector_num=8552704, nb_sectors=2048, qiov=0x7fa5a1757468, flags=(unknown: 0)) at
> block.c:3471
> 15) 0x00007fa5a0c542b3 in bdrv_co_do_rw (opaque=0x7fa5a17a0000) at block.c:4706
> 16) 0x00007fa5a0c64e6e in coroutine_trampoline (i0=-1585909408, i1=32677) at
> coroutine-ucontext.c:121
> 17) 0x00007fa59dc5aa50 in __correctly_grouped_prefixwc () from /lib64/libc.so.6
> 18) 0x0000000000000000 in ?? ()
> 
> After analyzing the stack and reviewing the code, we find the
> qemu_co_queue_run_restart should not be put in the coroutine_swap function which
> can be invoked by qemu_coroutine_enter or qemu_coroutine_yield. Only
> qemu_coroutine_enter needs to restart the co_queue.
> 
> The error scenario is as follow: coroutine C1 enters C2, C2 yields
> back to C1, then C1 ternimates and the related coroutine memory
> becomes invalid. After a while, the C2 coroutine is entered again.
> At this point, C1 is used as a parameter passed to
> qemu_co_queue_run_restart. Therefore, qemu_co_queue_run_restart
> accesses an invalid memory and a segfault error ocurrs.
> 
> The qemu_co_queue_run_restart function re-enters coroutines waiting
> in the co_queue. However, this function should be only used int the
> qemu_coroutine_enter context. Only in this context, when the current
> coroutine gets execution control again(after the execution of
> qemu_coroutine_switch), we can restart the target coutine because the
> target coutine has yielded back to the current coroutine or it has
> terminated.
> 
> First we want to put qemu_co_queue_run_restart in qemu_coroutine_enter,
> but we find we can not access the target coroutine if it terminates.
> 
> Signed-off-by: Bin Wu <wu.wubin@huawei.com>

As I replied for v2, I'll send a series with a simpler fix and some
cleanup that should result in a nicer design in the end.

Kevin