Re: [PATCH] thermal: intel: intel_pch_thermal: Add 200 series PCH-H device ID

From: Rafael J. Wysocki (Intel)

Date: Fri Nov 20 2026 - 14:24:34 EST


On Wed, Nov 18, 2026 at 4:53 PM Magnus Leatti
<magnusleatti@xxxxxxxxxxxxxx> wrote:
>
> The PCH thermal sensor on 200 series boards (Z270, Z370, X299, ...) is
> 8086:a2b1 and it's not in the ID table, so these boards, including mine,
> get no PCH temperature at all. The 100 series (Skylake) and 300 series
> (Cannon Lake) are there, Kaby belongs perfectly in between, I see no
> reason to not have it there.
>
> It's the same sensor. Intel's own datasheet (335193, chapter 21.2) has
> the register layout the driver already uses, so just add the ID and
> point it at PCH_BOARD_SKL like a131.
>
> Tested on an ASRock X299 Taichi, Kernel @ 7.4-rc3 with this patch. The
> driver loads by itself and the BIOS already turns the sensor on, so it
> just reads it. The zone shows 66°C / 115°C Crit, the readings look
> correct.
>
> Adding the ID has one more effect. I am not going to call it a side
> effect but more, an intended effect? AFAICS these boards were just
> missing the s2idle cooling delay that the 100 and 300 series already
> have, and now they get it. X299 usually runs above the 50 C threshold,
> it would wait the full 60s. Meaning that the wait *WILL* be present.
>
> The reason I am saying it *WILL* be present, is that I could not put
> this specific box to suspend, due to the GPU not allowing it under Linux
> and X299 not having onboard GPU on the CPU. But I manually followed the
> path (intel_pch_thermal_suspend_noirq()) and it seems to be precisely
> what will happen.
>
> I am not the first to notice this. Henry Bent put a patch in bugzilla
> back in 2024, and Jean Delvare asked about it on linux-pm last year.
>
> Link: https://cdrdv2-public.intel.com/335193/200-series-chipset-pch-datasheet-vol-2.pdf
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=219197
> Link: https://lore.kernel.org/linux-pm/20251222154444.496284d9@endymion/
> Signed-off-by: Magnus Leatti <magnusleatti@xxxxxxxxxxxxxx>
> ---
>
> Thanks,
> Magnus
>
> drivers/thermal/intel/intel_pch_thermal.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/thermal/intel/intel_pch_thermal.c b/drivers/thermal/intel/intel_pch_thermal.c
> index 074bc8ca5f0f..7d8803608486 100644
> --- a/drivers/thermal/intel/intel_pch_thermal.c
> +++ b/drivers/thermal/intel/intel_pch_thermal.c
> @@ -24,6 +24,7 @@
> #define PCH_THERMAL_DID_WPT 0x9CA4 /* Wildcat Point */
> #define PCH_THERMAL_DID_SKL 0x9D31 /* Skylake PCH */
> #define PCH_THERMAL_DID_SKL_H 0xA131 /* Skylake PCH 100 series */
> +#define PCH_THERMAL_DID_KBL_H 0xA2B1 /* Kaby Lake PCH 200 series */
> #define PCH_THERMAL_DID_CNL 0x9Df9 /* CNL PCH */
> #define PCH_THERMAL_DID_CNL_H 0xA379 /* CNL-H PCH */
> #define PCH_THERMAL_DID_CNL_LP 0x02F9 /* CNL-LP PCH */
> @@ -372,6 +373,8 @@ static const struct pci_device_id intel_pch_thermal_id[] = {
> .driver_data = PCH_BOARD_SKL, },
> { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_SKL_H),
> .driver_data = PCH_BOARD_SKL, },
> + { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_KBL_H),
> + .driver_data = PCH_BOARD_SKL, },
> { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_CNL),
> .driver_data = PCH_BOARD_CNL, },
> { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCH_THERMAL_DID_CNL_H),
>

Applied as 7.5 material, thanks!