Re: [Qemu-block] [PATCH v2] iscsi: Fix divide-by-zero regression on raw SG devices

[Eric Blake]

On 09/08/2016 09:27 AM, Holger Schranz wrote:
> Hi Eric,
> 
> Thanks a lot, it seems the patch works. The VM starting.

Should I take that as a Tested-by: line?

> Unfortunately we run into the next issue. By the accessing the
> megasas controller we got a SIGSEGV.

Seems unrelated, let's figure out who's responsible.


> 
> /home/kvm/SOURCES/qemu.git/qemu/hw/scsi> diff megasas.c
> /home/kvm/SOURCES/qemu-2.6.1/hw/scsi/megasas.c

'diff -u' is much easier to read than plain 'diff' (ed script diffs have
no context); in fact, 'git diff' uses -u format by default.  Also, it's
traditional to 'diff old new', but you flipped the argument, so the diff
looks backwards.

> 32c32
> < #include "qapi/error.h"
> ---
>>
> 51c51,55
>     :
> 413a422,423
>>
>>     return bcd_time;
> 1984c1994,1998
> <         cmd->frame->header.cmd_status = frame_status;
> ---
>>         if (cmd->frame) {
>>             cmd->frame->header.cmd_status = frame_status;
>>         } else {
>>             megasas_frame_set_cmd_status(s, frame_addr, frame_status);
>>         }

Here's a nicer way to get the above diff
$ git diff v2.6.1 v2.7.0 hw/scsi/megasas.c

at which point 'git gui blame hw/scsi/megasas.c' shows that it was
commit 8cc46787, by Paolo, that claimed that cmd->frame was non-NULL.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature