linux-kernel - Re: [External] [PATCH 2/2] platform/x86: thinkpad_acpi: Add a s2idle resume quirk for a number of laptops

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:   Fri, 29 Apr 2022 08:21:43 -0400
From:   Mark Pearson <markpearson@...ovo.com>
To:     Mario Limonciello <mario.limonciello@....com>,
        Hans de Goede <hdegoede@...hat.com>
CC:     Henrique de Moraes Holschuh <hmh@....eng.br>,
        Mark Gross <mgross@...ux.intel.com>,
        "open list:THINKPAD ACPI EXTRAS DRIVER" 
        <ibm-acpi-devel@...ts.sourceforge.net>,
        "open list:THINKPAD ACPI EXTRAS DRIVER" 
        <platform-driver-x86@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [External] [PATCH 2/2] platform/x86: thinkpad_acpi: Add a s2idle
 resume quirk for a number of laptops


On 4/28/22 23:05, Mario Limonciello wrote:
> Lenovo laptops that contain NVME SSDs across a variety of generations have
> trouble resuming from suspend to idle when the IOMMU translation layer is
> active for the NVME storage device.
> 
> This generally manifests as a large resume delay or page faults. These
> delays and page faults occur as a result of a Lenovo BIOS specific SMI
> that runs during the D3->D0 transition on NVME devices.
> 
> This SMI occurs because of a flag that is set during resume by Lenovo
> firmware:
> 
> ```
> OperationRegion (PM80, SystemMemory, 0xFED80380, 0x10)
> Field (PM80, AnyAcc, NoLock, Preserve)
> {
> 	SI3R,   1
> }
> 
> Method (_ON, 0, NotSerialized)  // _ON_: Power On
> {
> 	TPST (0x60D0)
> 	If ((DAS3 == 0x00))
> 	{
> 		If (SI3R)
> 		{
> 			TPST (0x60E0)
> 			M020 (NBRI, 0x00, 0x00, 0x04, (NCMD | 0x06))
> 			M020 (NBRI, 0x00, 0x00, 0x10, NBAR)
> 			APMC = HDSI /* \HDSI */
> 			SLPS = 0x01
> 			SI3R = 0x00
> 			TPST (0x60E1)
> 		}
> 		D0NV = 0x01
> 	}
> }
> ```
> 
> Create a quirk that will run early in the resume process to prevent this
> SMI from running. As any of these machines are fixed, they can be peeled
> back from this quirk or narrowed down to individual firmware versions.
> 
> Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1910>> Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1689>> Signed-off-by: Mario Limonciello <mario.limonciello@....com>

Thanks Mario,

Tested the series on the T14 G1 AMD and it fixes the long delay when
resuming from S0ix suspend nicely.

Patches look good to me. I am working with the Lenovo FW teams for the
various platforms, but as S0ix support wasn't in the original scope I
don't know yet when (or if in some cases) they'll get FW fixes released.
I believe these changes are benign even once the FW is fixed so think
this is a great idea.

I'll push updates removing platform quirks once FW is released though.

Tested-by: Mark Pearson <markpearson@...vo.com>