On Thu, Oct 30, 2014 at 10:38 PM, Jens Axboe <axboe at kernel.dk>
wrote:> Forgot to CC you...
>
>
> -------- Forwarded Message --------
> Subject: Re: kernel BUG at drivers/block/virtio_blk.c:172
> Date: Thu, 30 Oct 2014 08:38:08 -0600
> From: Jens Axboe <axboe at kernel.dk>
> To: Luk?? Czerner <lczerner at redhat.com>,
> virtualization at lists.linux-foundation.org
> CC: hch at lst.de
>
> On 2014-10-30 08:33, Luk?? Czerner wrote:
>>
>> Hi,
>>
>> I've just hit this BUG at drivers/block/virtio_blk.c when updated
to
>> the kernel from the top of the Linus git tree.
>>
>> commit a7ca10f263d7e673c74d8e0946d6b9993405cc9c
>>
>> This is my virtual machine running on RHEL7 guest
>>
>> qemu-kvm-1.5.3-60.el7.x86_64
>>
>> The last upstream kernel (3.17.0-rc4) worked well. I'll try to
>> bisect, but meanwhile this is a backtrace I got very early in the boot.
>> The root fs on that guest is xfs and I am using raw disk image and
virtio
>> driver.
>>
>> Let me know if you need more information.
>
>
> Ming, looks like this still isn't really fixed. The above upstream
> commit has the latest fixup as well for the segments being wrong, so
> nothing else should be pending.
That looks weird, and I can't reproduce with mkfs.xfs & mount
in my environment.
Luk??, could you reproduce the issue with attached debug patch
and post the result? BTW, do you pass 'scsi=off' in the qemu
command line for the virito-blk device?
Thanks,
--
Ming Lei
-------------- next part --------------
A non-text attachment was scrubbed...
Name: blk-seg.patch
Type: text/x-patch
Size: 583 bytes
Desc: not available
URL:
<http://lists.linuxfoundation.org/pipermail/virtualization/attachments/20141031/6995a123/attachment.bin>