Re: [PATCH 07/11] PCI: rcar-gen4: Recover the Root Port on link down

[Marek Vasut <marek.vasut@mailbox.org>]

On 9/28/26 6:06 AM, Koichiro Den wrote:

Hello Den-san,

>>>> Out of curiosity, do they trigger SError, and does the firmware (TFA)
>>>> trap/fix those up in EL3?
>>>
>>> I haven't confirmed whether those DBI accesses trigger an SError.
>>> The BL31 on the Spider I'm using should be based on rcar-s4_v2.5 [1] from the
>>> Spider BSP (as I haven't updated it). I just checked that this version also sets
>>> HANDLE_EA_EL3_FIRST=1, plus plat_ea_handler() is empty. So IIUC an SError would
>>> be taken to EL3
>>
>> Yes.
>>
>>> , and the normal path would return to the kernel without fixing
>>> the underlying error.
>>
>> That is correct.
>>
>> R-Car Gen3 does PCIe link error fixup in TFA, it looks this way:
>>
>> https://github.com/ARM-software/arm-trusted-firmware/commit/0969397f295621aa26b3d14b76dd397d22be58bf
> 
> Thanks for the info! It does look similar, so now I'm wondering about (*) below.
> 
>>
>>> And, regardless of whether it's a synchronous abort or an SError:
>>> - There was no crash report (via report_unhandled_exception) on the console.
> 
> (*) I checked that CRASH_REPORTING is enabled in that version, but there might
>      be a separate issue with report_unhandled_exception.

It could be that the DBI access triggers some other type of exception, 
not SError.

>>> - Once that DBI access in the small window hangs, it never recovers,
>>>     whereas adding udelay(300) before the access avoids the hang. So I
>>>     doubt it's just retrying the same access after a transient synchronous
>>>     abort as well (though I can't rule it out).
>>
>> ( It just crossed my mind, this sounds similar to 0056d29f8c1b ("PCI:
>> rcar-gen4: Assure reset occurs before DBI access") )
> 
> Yes, exactly. I'm almost ready to send v2, with an explicit mention of that
> commit in the cover letter.
> 
> A delay-based approach might be an option here, but I suspect it would involve
> about the same level of complexity in the end. The reset request shares
> intreq_pcim_sub with iMSI-RX, while AER arrives on another IRQ, so we'd still
> need to coordinate both paths to avoid unsafe DBI access.

ACK

-- 
Best regards,
Marek Vasut