Re: [PATCH v2 0/1] PCI: pcie_failed_link_retrain() return if dev is not ASM2824

[Matthew W Carlis <mattc@purestorage.com>]

On Mon, 1 Dec 2025 Maciej W. Rozycki wrote:

>> Root port can tell us if PCIe errors are going to the BIOS. IF any of the
>> ErrCorrectable, ErrNon-Fatal, ErrFatal, are set in the RootCtrl then those
>> error types would most likely go to the BIOS even if the OS thinks it took
>> control. Someone will have to correct me if wrong about ARM. If you sent
>> the full lspci -vvv of root port, USP/DSP/USP combo I could figure out
>> whats going on.
>> 
>
>  I've attached a full `lspci -xxxx' dump instead, just in case anything 
> else turns out useful.  I'll appreciate if you have a look.

I looked through the tree and don't see anything obviously wrong in the error
settings. Those CE's on 02:03.0 might be left-over from boot. In any case I
wouldn't expect to see BadTLP or BadDLLP unless there was something wrong...
It would be interesting to clear them and see if a link reset could trigger
them again.

As for the quirk itself I want to re-ping this review since it looks like the
link speed 'quirk' has also been implicated by Alok here in a CR here:
https://lore.kernel.org/all/c296df33-f9c0-42f7-8add-6966d89d00c4@oracle.com/

We know the quirk can be executed on devices that are functioning completely correctly. 
There might be a justification here to have an additional/new 'pci_fixup_pass'
that would execute after link-training failures & let the specific devices
decide how to handle link recovery, but I don't want this specific quirk running
against any of the devices in my systems.

  Matt