Re: [Qemu-devel] [PATCH 03/29] display: enable DIRTY_MEMORY_VGA tracking explicitly

[Mark Cave-Ayland]

On 26/05/15 12:56, Paolo Bonzini wrote:

> On 26/05/2015 13:52, Peter Maydell wrote:
>> On 26 May 2015 at 12:24, Paolo Bonzini <address@hidden> wrote:
>>>> Peter, Mark, Aurelien, can you review and ack this patch?
>> Could you provide some explanation/documentation of when a
>> display device needs to set DIRTY_MEMORY_VGA (and when it
>> doesn't)?
> 
> It needs to set it if it uses memory_region_get/set/clear_dirty with
> DIRTY_MEMORY_VGA as the last argument.
> 
>> If we get it wrong is there any way to make qemu
>> assert or otherwise catch the error?
> 
> It may be possible to check against mr->dirty_log_mask in
> memory_region_get/set/clear_dirty.  However, it is just as likely to
> have some corner case that is correct but triggers the assertion.
> 
> I haven't thought much about it, because a simple grep for
> DIRTY_MEMORY_VGA will catch the device models that need care.

Hi Paolo,

Sorry for a taking a while to get around to this. I've just tested the
v3 patchset on qemu-system-sparc for both CG3 and TCX and I don't see
any display issues (or at least if this were broken I'd expect to see
missing video updates/strange video artifacts).

So while I can't comment on the specifics, it looks reasonable based
upon the patchset and doesn't break anything so:

Acked-by: Mark Cave-Ayland <address@hidden>


ATB,

Mark.