linux-kernel - Re: [PATCH v8 2/2] PCI/ACPI: Use device constraints instead of dates to opt devices into D3

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

Date:   Thu, 3 Aug 2023 21:32:35 -0500
From:   Mario Limonciello <mario.limonciello@....com>
To:     Mika Westerberg <mika.westerberg@...ux.intel.com>
Cc:     "Rafael J . Wysocki" <rafael@...nel.org>,
        Bjorn Helgaas <helgaas@...nel.org>, linux-pci@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-acpi@...r.kernel.org,
        Iain Lane <iain@...ngesquash.org.uk>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Kuppuswamy Sathyanarayanan 
        <sathyanarayanan.kuppuswamy@...ux.intel.com>
Subject: Re: [PATCH v8 2/2] PCI/ACPI: Use device constraints instead of dates
 to opt devices into D3

On 8/3/23 06:38, Mario Limonciello wrote:
> On 8/3/23 00:01, Mika Westerberg wrote:
>> Hi,
>>
>> On Wed, Aug 02, 2023 at 03:10:13PM -0500, Mario Limonciello wrote:
>>> @@ -3036,11 +3044,8 @@ bool pci_bridge_d3_possible(struct pci_dev 
>>> *bridge)
>>>           if (dmi_check_system(bridge_d3_blacklist))
>>>               return false;
>>> -        /*
>>> -         * It should be safe to put PCIe ports from 2015 or newer
>>> -         * to D3.
>>> -         */
>>> -        if (dmi_get_bios_year() >= 2015)
>>> +        /* the platform indicates in a device constraint that D3 is 
>>> needed */
>>> +        if (platform_constraint_d3(bridge))
>>
>> This for sure causes some sort of power regression on the Intel
>> platforms made after 2015. Why not check for the constraint and:
>>
> Are you sure?  I saw it as an explanation of how Windows could put the 
> systems into D3 when there is no other PM related ACPI objects.
> 
>> - If present and enabled, use the desired D-state
>> - If present and disabled, leave the device in D0
>> - If not present use the existing cutoff date
>>
>> ?
> 
> Thanks! That sounds very reasonable to me.  I'll double check it in my 
> case.

I've played with this a bit, and I found that I can make it work by 
moving the constraints check into pci_target_state() in the non-ACPI 
power manageable case.

To me this works pretty well to reflect spec policy ambiguity but should 
avoid regressions dropping the 2015 check.  I'll send out a v9 with this 
approach.