qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Christian Borntraeger <borntraeger@de.ibm.com>
To: Fam Zheng <famz@redhat.com>, Paolo Bonzini <pbonzini@redhat.com>
Cc: Bo Tu <tubo@linux.vnet.ibm.com>, qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH 5/8] virtio-blk: fix "disabled data plane" mode
Date: Thu, 10 Mar 2016 10:40:17 +0100	[thread overview]
Message-ID: <56E14101.4030405@de.ibm.com> (raw)
In-Reply-To: <56E13849.3060409@de.ibm.com>

On 03/10/2016 10:03 AM, Christian Borntraeger wrote:
> On 03/10/2016 02:51 AM, Fam Zheng wrote:
> [...]
>> The aio_poll() inside "blk_set_aio_context(s->conf->conf.blk, s->ctx)" looks
>> suspicious:
>>
>>        main thread                                          iothread
>> ----------------------------------------------------------------------------
>>     virtio_blk_handle_output()
>>      virtio_blk_data_plane_start()
>>       vblk->dataplane_started = true;
>>       blk_set_aio_context()
>>        bdrv_set_aio_context()
>>         bdrv_drain()
>>          aio_poll()
>>           <snip...>
>>            virtio_blk_handle_output()
>>             /* s->dataplane_started is true */
>> !!!   ->    virtio_blk_handle_request()
>>          event_notifier_set(ioeventfd)
>>                                                     aio_poll()
>>                                                      virtio_blk_handle_request()
>>
>> Christian, could you try the followed patch? The aio_poll above is replaced
>> with a "limited aio_poll" that doesn't disptach ioeventfd.
>>
>> (Note: perhaps moving "vblk->dataplane_started = true;" after
>> blk_set_aio_context() also *works around* this.)
>>
>> ---
>>
>> diff --git a/block.c b/block.c
>> index ba24b8e..e37e8f7 100644
>> --- a/block.c
>> +++ b/block.c
>> @@ -4093,7 +4093,9 @@ void bdrv_attach_aio_context(BlockDriverState *bs,
>>
>>  void bdrv_set_aio_context(BlockDriverState *bs, AioContext *new_context)
>>  {
>> -    bdrv_drain(bs); /* ensure there are no in-flight requests */
>> +    /* ensure there are no in-flight requests */
>> +    bdrv_drained_begin(bs);
>> +    bdrv_drained_end(bs);
>>
>>      bdrv_detach_aio_context(bs);
>>
> 
> That seems to do the trick. 

Or not. Crashed again :-(

here is a trace with debugging enabled. The opaque value is zero, which is not good.

#0  0x0000000010329f98 in bdrv_co_do_rw (opaque=0x0) at block/io.c:2170
#1  0x00000000103b33a2 in coroutine_trampoline (i0=1023, i1=1946159824) at qemu/util/coroutine-ucontext.c:79
#2  0x000003ff7d9d150a in __makecontext_ret () from /lib64/libc.so.6

Still no idea why.

  reply	other threads:[~2016-03-10  9:40 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-02-14 17:17 [Qemu-devel] [PATCH 0/8] virtio: allow migration with dataplane Paolo Bonzini
2016-02-14 17:17 ` [Qemu-devel] [PATCH 1/8] block-migration: acquire AioContext as necessary Paolo Bonzini
2016-02-16  7:17   ` Fam Zheng
2016-02-19  7:41     ` Michael S. Tsirkin
2016-02-19 15:02     ` Paolo Bonzini
2016-02-14 17:17 ` [Qemu-devel] [PATCH 2/8] vring: make vring_enable_notification return void Paolo Bonzini
2016-02-15 16:24   ` Cornelia Huck
2016-02-16  7:18   ` Fam Zheng
2016-02-14 17:17 ` [Qemu-devel] [PATCH 3/8] virtio: add AioContext-specific function for host notifiers Paolo Bonzini
2016-02-16  7:20   ` Fam Zheng
2016-02-16  9:42     ` Cornelia Huck
2016-02-14 17:17 ` [Qemu-devel] [PATCH 4/8] virtio: export vring_notify as virtio_should_notify Paolo Bonzini
2016-02-15 16:44   ` Cornelia Huck
2016-02-16  7:21   ` Fam Zheng
2016-02-14 17:17 ` [Qemu-devel] [PATCH 5/8] virtio-blk: fix "disabled data plane" mode Paolo Bonzini
2016-02-15 17:58   ` Cornelia Huck
2016-02-16 15:45     ` Paolo Bonzini
2016-02-16 16:15       ` Cornelia Huck
2016-02-16 16:29         ` Paolo Bonzini
2016-02-16  7:26   ` Fam Zheng
2016-03-09 12:21   ` Christian Borntraeger
2016-03-09 12:55     ` Paolo Bonzini
2016-03-09 13:02       ` Christian Borntraeger
2016-03-09 13:05         ` Christian Borntraeger
2016-03-09 13:12           ` Cornelia Huck
2016-03-09 14:29       ` Christian Borntraeger
2016-03-09 16:17         ` Paolo Bonzini
2016-03-10  1:51           ` Fam Zheng
2016-03-10  9:03             ` Christian Borntraeger
2016-03-10  9:40               ` Christian Borntraeger [this message]
2016-03-11 10:28                 ` Paolo Bonzini
2016-03-14  9:18                   ` tu bo
2016-03-15 12:45                   ` Fam Zheng
2016-03-15 13:18                     ` Cornelia Huck
2016-03-15 14:08                       ` Paolo Bonzini
2016-03-15 23:34                         ` Fam Zheng
2016-03-16  5:21                     ` tu bo
2016-02-14 17:17 ` [Qemu-devel] [PATCH 6/8] virtio-blk: do not use vring in dataplane Paolo Bonzini
2016-02-16  8:15   ` Fam Zheng
2016-02-14 17:17 ` [Qemu-devel] [PATCH 7/8] virtio-scsi: " Paolo Bonzini
2016-02-16  8:28   ` Fam Zheng
2016-02-14 17:17 ` [Qemu-devel] [PATCH 8/8] vring: remove Paolo Bonzini
2016-02-16  8:32   ` Fam Zheng
2016-02-16  8:57 ` [Qemu-devel] [PATCH 0/8] virtio: allow migration with dataplane Christian Borntraeger
2016-02-16 16:25   ` Paolo Bonzini
2016-02-19  7:42 ` Michael S. Tsirkin
2016-02-19 14:59   ` Paolo Bonzini
2016-02-22 16:01 ` Stefan Hajnoczi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=56E14101.4030405@de.ibm.com \
    --to=borntraeger@de.ibm.com \
    --cc=famz@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=tubo@linux.vnet.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).