Hi Rafael, On 10/6/23 21:20, Rafael J. Wysocki wrote: > On Fri, Oct 6, 2023 at 5:17 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote: >> >> Hi August, >> >> On 10/6/23 16:18, August Wikerfors wrote: >>> Hi Hans, >>> >>> On 2023-10-06 14:33, Hans de Goede wrote: >>>> The TongFang GM6BGEQ, GM6BG5Q and GM6BG0Q are 3 GPU variants of a TongFang >>>> barebone design which is sold under various brand names. >>>> >>>> The ACPI IRQ override for the keyboard IRQ must be used on these AMD Zen >>>> laptops in order for the IRQ to work. >>>> >>>> Adjust the irq1_edge_low_force_override[] DMI match table for this: >>>> >>>> 1. Drop the sys-vendor match from the existing PCSpecialist Elimina Pro 16 >>>> entry for the GM6BGEQ (RTX3050 GPU) model so that it will also match >>>> the laptop when sold by other vendors such as hyperbook.pl. >>>> >>>> 2. Add board-name matches for the GM6BG5Q (RTX4050) and GM6B0Q (RTX4060) >>>> models. >>>> >>>> Suggested-by: August Wikerfors <git@xxxxxxxxxxxxxxxxxx> >>>> Reported-by: Francesco <f.littarru@xxxxxxxxxxx> >>>> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217394 >>>> Link: https://laptopparts4less.frl/index.php?route=product/search&filter_name=GM6BG >>>> Link: https://hyperbook.pl/en/content/14-hyperbook-drivers >>>> Link: https://linux-hardware.org/?probe=bfa70344e3 >>>> Link: https://bbs.archlinuxcn.org/viewtopic.php?id=13313 >>>> Signed-off-by: Hans de Goede <hdegoede@xxxxxxxxxx> >>> Since this is a regression fix for 453b014e2c29 ("ACPI: resource: Fix IRQ override quirk for PCSpecialist Elimina Pro 16 M") (for PCSpecialist systems like Francesco's with product name "Elimina Pro 16 M" but not board name "GM6BGEQ") and 2d331a6ac481 ("ACPI: resource: revert "Remove "Zen" specific match and quirks"") (for other vendors using the same TongFang design), it should have a "Fixes:" tag for at least one of those. >>> >>> Both of those commits are in 6.5 (and 6.6-rc) so this should go into 6.6-rc and be backported to 6.5, but the patch seems to depend on 424009ab2030 ("ACPI: resource: Drop .ident values from dmi_system_id tables") and maybe also d37273af0e42 ("ACPI: resource: Consolidate IRQ trigger-type override DMI tables") to apply cleanly, which seem to only be queued for linux-next/6.7? I'm not familiar with what the correct process is for such cases. >> >> You are right: >> >> Fixes: 453b014e2c29 ("ACPI: resource: Fix IRQ override quirk for PCSpecialist Elimina Pro 16 M") > > OK > >> Note likewise this commit should really also be send as a fix >> for 6.6 and backported to the stable kernels: >> >> https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git/commit/drivers/acpi/resource.c?h=bleeding-edge&id=c1ed72171ed580fbf159e703b77685aa4b0d0df5 > > That's harder, because it does depend on commit 424009ab2030 ("ACPI: > resource: Drop .ident values from dmi_system_id tables") which is not > 6.6-rc material IMV. > > So I'm going to queue this up with the Fixes tag above and Cc: stable > pointing to commit 424009ab2030 as a dependency, but for 6.7. I realize that I'm the architect of this whole mess, but delaying these new quirks to 6.7 does not seem like a good idea to me. This fixes a serious problem (kbd not working pretty much makes the whole laptop unusable under Linux at least on the road). With a very small chance of regressions since this is DMI match base. Maybe we can just take all the resource.c DMI quirk changes as fixes to 6.6 ? I admit the: "ACPI: resource: Consolidate IRQ trigger-type override DMI tables" "ACPI: resource: Drop .ident values from dmi_system_id tables" patches are not really bugfixes. But they too have a very low chance of causing regressions. Regards, Hans p.s. Note that "ACPI: resource: Skip IRQ override on ASUS ExpertBook B1402CBA" should probably go as a fix to 6.6 regardless of the discussion since that is a pure bugfix which applies cleanly on top of 6.6 And the same goes for this (unrelated) drivers/acpi/ec.c patch: "ACPI: EC: Add quirk for the HP Pavilion Gaming 15-dk1xxx" That too is a pure bugfix which IMHO should go to 6.6 as such.
