Hi Vijay, I no longer actively maintain the error recovery stack. However, your patch and explanation for it looks reasonable to me --- it was certainly never the intent that a non-capable driver report a successful recovery. I'm guessing that the bug that you are fixing was introduced when the original pci error recovery code was integrated with AER (PCI-AER was standardized after the original error recovery system was created, and I did not really pay very much attention to AER at the time). FWIW, I guess I could add a Signed-off-by: Linas Vepstas <linasvepstas@xxxxxxxxx> but I'm not sure that really means much any more :-) -- Linas On 7 November 2012 21:50, Pandarathil, Vijaymohan R <vijaymohan.pandarathil@xxxxxx> wrote: > Hi Bjorn/Linas, > > I got your names from the PCI/PCI-error-recovery maintainers list. I had > posted the following request last week. > http://article.gmane.org/gmane.linux.kernel.pci/18465 > > Being a newbie, it would be extremely helpful to get your guidance on > these changes, specifically on any side effects it has. I am also working > on another set of patches to improve the PCI error recovery/containment > capabilities of KVM. But if the above change cannot be made first, I may > have to figure out alternate approach for the second set of patches. > > Thanks > > Vijay
Attachment:
signature.asc
Description: Digital signature
