[SeaBIOS] Confirm the status about BZ#1377575

[Shouta.Uehara at jp.yokogawa.com]

Hello Laszlo,

Thank you for your advice.
At last, I've been able to get a crash dump of WS2016.

I changed following configuration;

(1) Disabled cirrus, and use std graphic card.
(2) To explicit invoke recent bios from qemu.

My testbed is installed CentOS 7.4 which has seabios-bin-1.10.2-3.el7 RPM package,
but my modified qemu is ported from CentOS 7.1.
Therefore, qemu invoked the obsolete bios which is included in its source.

Finally, I have realized to get a crash dump on WS2016.

Thank you!

Shota

-----Original Message-----
Hello Shouta-san,

On 02/14/18 05:48, Shouta.Uehara at jp.yokogawa.com wrote:
> Dear SeaBIOS development members,
>
> I'm Shota Uehara work for Yokogawa Electric Corporation.
> I would like to confirm the status about a bug fixing of BZ#1377575.
>
> I use Windows Server 2016 virtual machine on qemu-kvm with CentOS 7.
> And now I attempt to add a dedicated virtual device to qemu-kvm, and develop its device driver for WS2016.
>
> For debugging the windows driver I want to get a crash dump on WS2016, however it doesn't be created.
> According to RedHat knowledge base, that is a Windows incompatibility issue to SeaBIOS, as known BZ#1377575.
>  <https://access.redhat.com/articles/2470791#known-issues-14>
>
> Therefore, I have 2 questions as follows;
> (1) Could you tell me the current status about a bug fixing of BZ#1377575 ?
> (2) If possible, could you tell me the work-around to avoid such bug and realize to get a crash dump ?

Please refer to the following blog post by my (then-)colleague Ladi Prosek, who worked around the Windows issue (BZ#1377575) in the RHEL7 SeaBIOS package:

  https://ladipro.wordpress.com/2016/10/21/windows-vga-bug/

As stated in the blog post, both abandoning Cirrus video, and patching SeaBIOS, are necessary for working around the Windows issue. The SeaBIOS patch was sent to the SeaBIOS mailing list; please see the discussion thread at:

  [SeaBIOS] [PATCH] vgabios: Reorder video modes to work around a
                    Windows bug
  https://www.coreboot.org/pipermail/seabios/2016-October/010963.html

The thread seems to have petered out back then. For using an upstream SeaBIOS build, you might have to apply the patch manually. However, when using SeaBIOS from CentOS 7, the issue should already be fixed:

- in 7.3.z, starting with seabios-1.9.1-5.el7_3.1
  (BZ#1392028, <https://rhn.redhat.com/errata/RHBA-2017-0079.html>),
- in 7.4, starting with seabios-1.10.1-1.el7
  (BZ#1377575, <https://access.redhat.com/errata/RHBA-2017:1855>).

Can you check if you are using a CentOS 7 SeaBIOS package that is at least this recent?

... In fact, I see a private (Red Hat only) reminder under the Knowledge Base article that we should update the "Known Issues" text, regarding BZ#1377575, once the related updates are out. Given that both BZs above are in CLOSED ERRATA status, I'm going to ping our docs guys now. :)

... Additionally, I've found a note in BZ#1377575 that Windows 10 RS2, version 1703+ (possibly earlier), should no longer have this bug. Not sure how (or if) that maps to Windows Server 2016.

Thank you!
Laszlo
-----
CONFIDENTIAL: This e-mail may contain information that is confidential or otherwise protected from disclosure and intended only for the party to whom it is addressed. If you are not the intended recipient, please notify the sender by return and delete this e-mail. You are hereby formally advised that any unauthorized use, disclosure or copying of this email is strictly prohibited and may be unlawful.