Re: [Qemu-devel] [PATCH for-2.12 v2] monitor: bind dispatch bh to iohandler context

[Eric Blake]

On 04/10/2018 12:01 AM, Stefan Hajnoczi wrote:
> On Tue, Apr 10, 2018 at 12:49:42PM +0800, Peter Xu wrote:
>> Eric Auger reported the problem days ago that OOB broke ARM when running
>> with libvirt:
>>
>> http://lists.gnu.org/archive/html/qemu-devel/2018-03/msg06231.html
>>
>> The problem was that the monitor dispatcher bottom half was bound to
>> qemu_aio_context now, which could be polled unexpectedly in block code.
> 
> And TPM and 9P code, who all use nested event loops.

I'll tweak the commit message to mention that.

> 
>> We should keep the dispatchers run in iohandler_ctx just like what we
>> did before the Out-Of-Band series (chardev uses qio, and qio binds
>> everything with iohandler_ctx).
>>

>> ---
>> v2:
>> - enhanced commit message

Thanks for that.  As you are learning, the hardest part of writing a
patch is often not the code itself, but figuring out the most efficient
way to get reviewers to understand why the patch is important.  The
better you get at writing GOOD commit messages, the easier it is for
your patches to be accepted!

>> ---
>>  monitor.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/monitor.c b/monitor.c
>> index 51f4cf480f..39f8ee17ba 100644
>> --- a/monitor.c
>> +++ b/monitor.c
>> @@ -4467,7 +4467,7 @@ static void monitor_iothread_init(void)
>>   * have assumption to be run on main loop thread.  It would be
>>   * nice that one day we can remove this assumption in the future.
>>   */
>> -mon_global.qmp_dispatcher_bh = aio_bh_new(qemu_get_aio_context(),
>> +mon_global.qmp_dispatcher_bh = aio_bh_new(iohandler_get_aio_context(),
>>monitor_qmp_bh_dispatcher,
>>NULL);
> 
> Reviewed-by: Stefan Hajnoczi 

Reviewed-by: Eric Blake 

and queuing through my qapi tree, pull request for -rc3 going out shortly

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.   +1-919-301-3266
Virtualization:  qemu.org | libvirt.org



signature.asc
Description: OpenPGP digital signature

Re: [Qemu-devel] [PATCH for-2.12 v2] monitor: bind dispatch bh to iohandler context

[Stefan Hajnoczi]

On Tue, Apr 10, 2018 at 12:49:42PM +0800, Peter Xu wrote:
> Eric Auger reported the problem days ago that OOB broke ARM when running
> with libvirt:
> 
> http://lists.gnu.org/archive/html/qemu-devel/2018-03/msg06231.html
> 
> The problem was that the monitor dispatcher bottom half was bound to
> qemu_aio_context now, which could be polled unexpectedly in block code.

And TPM and 9P code, who all use nested event loops.

> We should keep the dispatchers run in iohandler_ctx just like what we
> did before the Out-Of-Band series (chardev uses qio, and qio binds
> everything with iohandler_ctx).
> 
> If without this change, QMP dispatcher might be run even before reaching
> main loop in block IO path, for example, in a stack like (the ARM case,
> "cont" command handler run even during machine init phase):
> 
> #0  qmp_cont ()
> #1  0x006bd210 in qmp_marshal_cont ()
> #2  0x00ac05c4 in do_qmp_dispatch ()
> #3  0x00ac07a0 in qmp_dispatch ()
> #4  0x00472d60 in monitor_qmp_dispatch_one ()
> #5  0x0047302c in monitor_qmp_bh_dispatcher ()
> #6  0x00acf374 in aio_bh_call ()
> #7  0x00acf428 in aio_bh_poll ()
> #8  0x00ad5110 in aio_poll ()
> #9  0x00a08ab8 in blk_prw ()
> #10 0x00a091c4 in blk_pread ()
> #11 0x00734f94 in pflash_cfi01_realize ()
> #12 0x0075a3a4 in device_set_realized ()
> #13 0x009a26cc in property_set_bool ()
> #14 0x009a0a40 in object_property_set ()
> #15 0x009a3a08 in object_property_set_qobject ()
> #16 0x009a0c8c in object_property_set_bool ()
> #17 0x00758f94 in qdev_init_nofail ()
> #18 0x0058e190 in create_one_flash ()
> #19 0x0058e2f4 in create_flash ()
> #20 0x005902f0 in machvirt_init ()
> #21 0x007635cc in machine_run_board_init ()
> #22 0x006b135c in main ()
> 
> Actually the problem is more severe than that.  After we switched to the
> qemu AIO handler it means the monitor dispatcher code can even be called
> with nested aio_poll(), then it can be an explicit aio_poll() inside
> another main loop aio_poll() which could be racy too.
> 
> Switch to use the iohandler_ctx for monitor dispatchers.
> 
> My sincere thanks to Eric Auger who offered great help during both
> debugging and verifying the problem.  The ARM test was carried out by
> applying this patch upon QEMU 2.12.0-rc0 and problem is gone after the
> patch.
> 
> A quick test of mine shows that after this patch applied we can pass all
> raw iotests even with OOB on by default.
> 
> CC: Eric Blake 
> CC: Markus Armbruster 
> CC: Stefan Hajnoczi 
> CC: Fam Zheng 
> Reported-by: Eric Auger 
> Tested-by: Eric Auger 
> Signed-off-by: Peter Xu 
> ---
> v2:
> - enhanced commit message
> ---
>  monitor.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/monitor.c b/monitor.c
> index 51f4cf480f..39f8ee17ba 100644
> --- a/monitor.c
> +++ b/monitor.c
> @@ -4467,7 +4467,7 @@ static void monitor_iothread_init(void)
>   * have assumption to be run on main loop thread.  It would be
>   * nice that one day we can remove this assumption in the future.
>   */
> -mon_global.qmp_dispatcher_bh = aio_bh_new(qemu_get_aio_context(),
> +mon_global.qmp_dispatcher_bh = aio_bh_new(iohandler_get_aio_context(),
>monitor_qmp_bh_dispatcher,
>NULL);

Reviewed-by: Stefan Hajnoczi 


signature.asc
Description: PGP signature