[dm-devel] Re: [RFC][PATCH] dm-mirror: fix data corruption

[malahal at us.ibm.com]

Mikulas Patocka [mpatocka at redhat.com] wrote:
> 
> > How do bios queued in ms->failures are processed later? It seems that
> > the bios stay in ms->failures forever, and the upper layer can not
> > receive "success" for those bios. Don't we need a mechanism to block/unblock
> > write bios to fix this issue?
> 
> They are resubmitted with DM_ENDIO_REQUEUE on noflush suspend. My patch 
> has a bug that they aren't --- but I will provide a better patch, also 
> without this periodic polling of ms->failures queue.

Trying to verify this patch. Mikulas, did you provide a better patch
yet? Does this patch work at all?

I would like to verify if this patch works with devices that fail
temporarily. I will plan on using dm-flakey devices for testing purposes.

Thanks, Malahal.