Re: [Qemu-devel] Xen patches - status summary

[Anthony Liguori <anthony@codemonkey.ws>]

Gerd Hoffmann wrote:
> Ian Jackson wrote:
>   
>> I don't know about you but I'm starting to lose track of all the
>> patches we're submitting from the Xen tree.  So here's a summary of
>> recent activity:
>>
>> Patches from Stefano Stabellini awaiting commit to qemu:
>>
>>  [1/3] vnc dynamic resolution	} minor discussion, all issues resolved
>>  [2/3] WMVi extension support	}  so should IMO be applied (updated version
>>  [3/3] vga shared buffer	}  in case of vga shared buffer patch)
>>  sdl shared buffer support	  no discussion, should be applied IMO
>>  opengl rendering in the sdl ...  no discussion, should be applied IMO
>>     
>
> I think they still sitting in anthonys inbox waiting for review ...
>   

Yeah, sorry, I'm on vacation this week and my network access is much 
worse than I expected it to be.  I'm queuing it all up and will go 
through it this weekend when I get back.

>> Looking at my git logs here there are quite a few more which will be
>> forthcoming but perhaps it would be good to digest these first ?
>>     
>
> In case you don't know what to do with your time meanwhile:
> /me waits for qemu-xen catching up with upstream, so I can update and
> respin the patch series ;)
>
>   
>> [1] Re: Use fd signal trick to break us out of select
>>     Re: Introduce #define QEMU_ASYNC_EVENTLOOP
>>
>> Anthony Ligouri had a different patch based on a helper thread to
>> provide an emulation of signalfd rather than use of a signal handler.
>> However (as confirmed by Jamie Lokier) some versions of glibc have
>> a bug in the handling of the glibc internal aio helper thread(s)
>> which can make efforts to block signals ineffective.  This means
>> that Anthony's patch will not work properly on those libcs and as
>> I wrote already this means that my patch should be applied.
>>     
>
> I also tend to dislike os-specific stuff due to the portability issues
> it brings.  Sometimes there are good reasons to use it nevertheless.
>   

I haven't made up my mind yet about which approach is better.  I'll 
commit something this weekend though.

>> The QEMU_ASYNC_EVENTLOOP change is a tidying up of the NBD feature to
>> make qemu-nbd share more commonality with qemu-img.  As discussed
>> there are perhaps even more cleanups that could be done to improve
>> this but as I say this change is a good start and should be applied.
>>     
>
> Cleanups is, well, not the correct word IMHO.  The complete block device
> handling needs a major redesign.  That this ifdef is needed in the first
> place is a blatant layering violation.  Also we should be able to
> support async I/O in some form (be it libaio, threads or whatever)
> without hacking support for it into each and every file format handler.
>   

I have something queued up to help us get closer to this.

>> [3] Re: IDE error checking
>>     
>
> Yep, this certainly needs to be fixed.
> /me votes for applying them, although I don't know IDE good enougth to
> comment on the patches themself.
>   

I think I explained what I didn't like about this patch--that all IO 
errors were being reported to the guest as ECC errors.  I don't feel 
this is correct, in particular for ENOSPC.  When ENOSPC occurs, the 
guest is likely to really mess itself up.  Since ENOSPC is almost 
certainly going to be the most common error, it needs to be handled 
specially.

Regards,

Anthony Liguori

> cheers,
>   Gerd
>
>
>
>
>