Re: [Qemu-devel] [PATCH] Flush pending AIO on reboot and shutdown.

[Anthony Liguori]

Gleb Natapov wrote:
> On Wed, Aug 13, 2008 at 10:07:06AM -0500, Anthony Liguori wrote:
>   
> Yes, conceptually this is what should be done. But all outstanding block IO
> requests are stored in one global list in block-raw.c and this file also
> provides a global facility to flush all request from this global queue
> that is why I used it. Otherwise each subsisted will have to be touched
> to have desired affect.
>   

What happens on reset to outstanding disk requests is a semantic of the 
controller itself.  Really, I think the right thing to do is cancel any 
request that can be cancelled.  If a request cannot be cancelled then we 
should wait for it to complete.

Now that I think about it, I think your fixing the wrong problem.  The 
issue isn't that the IO requests need to be completed, but that they 
*will* complete which means that the IDE driver will receive a callback 
for a request that it no longer knows about (because it was reset).  So 
what we really need to do is modify the IDE device such that when it is 
reset, it cancels any pending requests.

The fact that this reset happens as a consequence of a system reset is 
really just a coincidence.

Regards,

Anthony Lgiuori

> --
>                         Gleb.
>
>
>