Re: [PATCH v8 2/3] PCI: Add device-specific reset for Qualcomm devices

[Alex Williamson <alex@shazbot.org>]

On Wed, 10 Jun 2026 12:00:59 +0200
Jose Ignacio Tornos Martinez <jtornosm@redhat.com> wrote:

> Hi Alex,
> 
> > My only complaint is that this is positioned a bit generically...  
> Ok, I'll make the comment of the function more specific for these Qualcomm
> devices.
>   
> > Has each of these devices been tested on a platform where D3cold is
> > actually achieved through this method?  
> No, I don't have access to platforms with _PR3 support. All testing was on
> M.2 adapters (D3hot fallback only). Since D3hot works reliably and D3cold
> is a stronger reset mechanism, the approach should work - but I cannot
> confirm the D3cold path through testing.

This seems to suggest that the D3cold reset method is also untested...
I'd suggest dropping the D3cold throughout, unless you can get someone
to volunteer to test it.  It seems like you really just want this quirk
to do a D3hot regardless of NoSoftRst+ with a select set of devices
where it appears to do more than nothing, where doing nothing is
effectively what's left after the other reset methods are being quirked
off for being worse than nothing.  Thanks,

Alex