linux-kernel - Re: [PATCH v2] PCI: Avoid putting some root ports into D3 on some Ryzen chips

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <90625eb1-9e3b-4c07-903f-6d9139c6bf64@amd.com>
Date: Tue, 17 Dec 2024 10:18:44 -0600
From: Mario Limonciello <mario.limonciello@....com>
To: Werner Sembach <wse@...edocomputers.com>,
 Bjorn Helgaas <bhelgaas@...gle.com>,
 "Rafael J. Wysocki" <rafael.j.wysocki@...el.com>,
 Mika Westerberg <mika.westerberg@...ux.intel.com>
Cc: ggo@...edocomputers.com, linux-pci@...r.kernel.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] PCI: Avoid putting some root ports into D3 on some
 Ryzen chips

On 12/17/2024 10:08, Werner Sembach wrote:
> 
> Am 17.12.24 um 16:58 schrieb Werner Sembach:
>>
>> Am 17.12.24 um 15:11 schrieb Mario Limonciello:
>>> On 12/17/2024 08:07, Werner Sembach wrote:
>>>
>>>>> '''
>>>>> Platform may hang resuming.  Upgrade your firmware or add 
>>>>> pcie_port_pm=off to kernel command line if you have problems
>>>>> '''
>>>> Yes, full log attached (kernel 6.13-rc3 one time without sudo one 
>>>> time with sudo)
>>>
>>> Yes; I see it in your log.
>>>
>>>>> "quirk: disabling D3cold for suspend"
>>>> On the fixed BIOS I see that line. On the unfixed BIOS it aborts the 
>>>> functions at "if (pm_suspend_target_state == PM_SUSPEND_ON)". 
>>>> Skipping the check on the unfixed BIOS it still hangs on resume.
>>>>>
>>>>> I'm /suspecting/ you do see it, but you're having problems with 
>>>>> another root port.
>>>>>
>>>>> I mentioned this in my previous iterations of patches that 
>>>>> eventually landed on that quirk, but Windows and Linux handle root 
>>>>> ports differently at suspend time and that could be why it's 
>>>>> exposing your BIOS bug.
>>>>>
>>>>> If you can please narrow down which root ports actually need the 
>>>>> quirk for your side (feel free to do a similar style to 
>>>>> 7d08f21f8c630) I think we could land on something more narrow and 
>>>>> upstreamable.
>>>>>
>>>>> At a minimum what you're doing today is covering both Rembrandt and 
>>>>> Phoenix and it should only apply to Phoenix.
>>>>
>>>> I also try to find out how many devices where actually shipped with 
>>>> this very first BIOS version.
>>>
>>> OK.
>>
>> Ok found out that the initial bios actually works, then there is one 
>> in between bios where it doesn't and the next one it works again.
>>
>> So i need to find out if the the in between bios was actually shipped, 
>> if not, this issue is actually void.
>>
> Dang it: seems like it.
> 
> So should i create a v3 of the patch with the correct pci ids just for 
> this bios version?

Yes; the quirk needs to be as narrow as makes sense for the situation.