Re: Problems with Arrow Lake-P GPU

Mikhaylov Aleksandr <[email protected]> Tue, 2 Jun 2026 19:57:57 +0300
Newsgroups	gmane.os.openbsd.bugs
Message-ID	<[email protected]>
Yes, I tried this patch, but unfortunately, 
I'm now getting different errors in dmesg:

root on sd1a (11e3a23e5a8a69ae.a) swap on sd1b dump on sd1b
drm:pid0:guc_wait_ucode *WARNING* [drm] GT0: GUC: excessive init time: 292ms! [status = 0x8002F034, count = 0, ret = 0]
drm:pid0:guc_wait_ucode *WARNING* [drm] GT0: GUC: excessive init time: [freq = 200MHz -> 200MHz vs 2250MHz, perf_limit_reasons = 0x08000000]
drm:pid34291:gsc_fw_load *ERROR* [drm] *ERROR* GT1: Request submission for GSC load failed 0xffffffffffffffc4e
drm:pid34291:intel_uc_fw_mark_load_failed *ERROR* [drm] *ERROR* GT1: Failed to load GSC firmware i915/mtl_gsc_1.bin 0xffffffffffffffc4e
inteldrm0: 2880x1800, 32bpp
wsdisplay0 at inteldrm0 mux 1: console (std, vt100 emulation), using wskbd0
wsdisplay0: screen 1-5 added (std, vt100 emulation)
i915_capture_error_state: stub
drm:pid78279:__intel_engine_reset_bh *NOTICE* [drm] Resetting other0 for stopped heartbeat on other0
drm:pid78279:gen8_engine_reset_prepare *ERROR* [drm] *ERROR* GT1: other0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
drm:pid78279:intel_gt_reset *NOTICE* [drm] GT1: Resetting chip for stopped heartbeat on other0
drm:pid78279:gen8_engine_reset_prepare *ERROR* [drm] *ERROR* GT1: other0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
drm:pid78279:gen8_engine_reset_prepare *ERROR* [drm] *ERROR* GT1: other0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}

and secondly - they appeared Artifacts:

https://downloader.disk.yandex.ru/preview/475d6f8589e3703d986591dbf69c5ee99f1f7342480241edfa82260a1527a7e4/6a1f4203/G3ek20yVbIy82IAgd1xEf5vy40V6jdb2oRP-aafJbHEInvumG_zgk1AEgJhziitZf1f-qBge5Wq_SenjxjckbQ%3D%3D?uid=0&filename=artifacts.png&disposition=inline&hash=&limit=0&content_type=image%2Fpng&owner_uid=0&tknv=v3&is_direct_zip_experiment=1&size=2048x2048

On 02/06/2026, Jonathan Gray wrote:
> On Tue, Jun 02, 2026 at 02:22:48PM +0300, Mikhaylov Aleksandr wrote:
> > It seems like it doesn't always have time to initialize the firmware. I
> > tried disabling all power-saving features in the BIOS, but the problem
> > still occurs occasionally.
> > 
> > I also found a dmesg log from another user with a Lenovo ThinkPad X1
> > Carbon Gen13 with the same error.
> > https://dmesgd.nycbug.org/index.cgi?do=view&id=8765
> > So this isn't the only issue.
> > 
> > Is there a way to increase some timeouts or force the firmware to load?
> 
> The only thing I'm aware of is using less of the firmware.  But then
> low power states aren't possible, and different warnings will appear.
> 
> Index: sys/dev/pci/drm/i915/i915_driver.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/pci/drm/i915/i915_driver.c,v
> diff -u -p -r1.30 i915_driver.c
> --- sys/dev/pci/drm/i915/i915_driver.c	30 Apr 2026 04:33:06 -0000	1.30
> +++ sys/dev/pci/drm/i915/i915_driver.c	22 May 2026 14:27:18 -0000
> @@ -2322,12 +2322,10 @@ inteldrm_attach(struct device *parent, s
>  
>  	dev_priv->display = display;
>  
> -	/*
> -	 * with GuC submission, init sometimes fails on Alder Lake-P
> -	 * and Raptor Lake-S, too early for IS_ALDERLAKE
> -	 */
> -	if (info->platform == INTEL_ALDERLAKE_P ||
> -	    info->platform == INTEL_ALDERLAKE_S)
> +	/* uc_expand_default_options() with no GuC submission */
> +	if (GRAPHICS_VER(dev_priv) >= 12 &&
> +	    (INTEL_INFO(dev_priv)->platform != INTEL_TIGERLAKE) &&
> +	    (INTEL_INFO(dev_priv)->platform != INTEL_ROCKETLAKE))
>  		dev_priv->params.enable_guc = ENABLE_GUC_LOAD_HUC;
>  
>  	mmio_bar = (GRAPHICS_VER(dev_priv) == 2) ? 0x14 : 0x10;
> Index: sys/dev/pci/drm/i915/i915_pci.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/pci/drm/i915/i915_pci.c,v
> diff -u -p -r1.25 i915_pci.c
> --- sys/dev/pci/drm/i915/i915_pci.c	9 Mar 2026 23:58:01 -0000	1.25
> +++ sys/dev/pci/drm/i915/i915_pci.c	22 May 2026 14:21:40 -0000
> @@ -771,7 +771,7 @@ static const struct intel_device_info mt
>  	.has_flat_ccs = 0,
>  	.has_gmd_id = 1,
>  	.has_guc_deprivilege = 1,
> -	.has_guc_tlb_invalidation = 1,
> +	.has_guc_tlb_invalidation = 0,
>  	.has_llc = 0,
>  	.has_mslice_steering = 0,
>  	.has_snoop = 1,

-- 
С уважением, Михайлов Александр.
signature.asc (application/pgp-signature, 870 B)