You are not logged in.

#51 2019-12-18 00:41:28

loqs
Member
Registered: 2014-03-06
Posts: 10,083

Re: i915 Skylake GPU hangs with kernel 5.3.11

Did linux 5.4.4.arch1-1 resolve the issue?

Offline

#52 2019-12-18 07:44:26

wioo
Member
Registered: 2017-05-18
Posts: 24

Re: i915 Skylake GPU hangs with kernel 5.3.11

Is there something special about 5.4.4? I cant see any difference in code from 5.4.3.

Offline

#53 2019-12-18 09:58:48

ultimaweapon
Member
Registered: 2019-12-18
Posts: 2

Re: i915 Skylake GPU hangs with kernel 5.3.11

The patch for fixing is already in 5.5-rc2. I just installed via https://aur.archlinux.org/packages/linux-mainline/ let see if it works.

Offline

#54 2019-12-18 13:00:28

loqs
Member
Registered: 2014-03-06
Posts: 10,083

Re: i915 Skylake GPU hangs with kernel 5.3.11

wioo wrote:

Is there something special about 5.4.4? I cant see any difference in code from 5.4.3.

I was wondering if the following affected the issue:
https://git.archlinux.org/linux.git/com … e06b3d3856

Offline

#55 2019-12-18 14:06:28

wioo
Member
Registered: 2017-05-18
Posts: 24

Re: i915 Skylake GPU hangs with kernel 5.3.11

loqs wrote:
wioo wrote:

Is there something special about 5.4.4? I cant see any difference in code from 5.4.3.

I was wondering if the following affected the issue:
https://git.archlinux.org/linux.git/com … e06b3d3856

5.4.4-arch1
No issues for now..

Edit:
Unfortunately not fixed

10589.736950] i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0
[10589.736957] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[10589.737974] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

Last edited by wioo (2019-12-18 17:35:26)

Offline

#56 2019-12-18 19:58:56

loqs
Member
Registered: 2014-03-06
Posts: 10,083

Re: i915 Skylake GPU hangs with kernel 5.3.11

wioo could ask for a backported to 5.4 version of the patch on https://gitlab.freedesktop.org/drm/intel/issues/673
Edit:
joanbrugueram is this the patch you used?

diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
index 06a506c29463..b70a59cdcdf2 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.c
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -471,12 +471,6 @@ lrc_descriptor(struct intel_context *ce, struct intel_engine_cs *engine)
 	return desc;
 }
 
-static void unwind_wa_tail(struct i915_request *rq)
-{
-	rq->tail = intel_ring_wrap(rq->ring, rq->wa_tail - WA_TAIL_BYTES);
-	assert_ring_tail_valid(rq->ring, rq->tail);
-}
-
 static struct i915_request *
 __unwind_incomplete_requests(struct intel_engine_cs *engine)
 {
@@ -495,7 +489,6 @@ __unwind_incomplete_requests(struct intel_engine_cs *engine)
 			continue; /* XXX */
 
 		__i915_request_unsubmit(rq);
-		unwind_wa_tail(rq);
 
 		/*
 		 * Push the request back into the queue for later resubmission.
@@ -649,13 +642,29 @@ execlists_schedule_out(struct i915_request *rq)
 	i915_request_put(rq);
 }
 
-static u64 execlists_update_context(const struct i915_request *rq)
+static u64 execlists_update_context(struct i915_request *rq)
 {
 	struct intel_context *ce = rq->hw_context;
-	u64 desc;
+	u64 desc = ce->lrc_desc;
+	u32 tail;
 
-	ce->lrc_reg_state[CTX_RING_TAIL + 1] =
-		intel_ring_set_tail(rq->ring, rq->tail);
+	/*
+	 * WaIdleLiteRestore:bdw,skl
+	 *
+	 * We should never submit the context with the same RING_TAIL twice
+	 * just in case we submit an empty ring, which confuses the HW.
+	 *
+	 * We append a couple of NOOPs (gen8_emit_wa_tail) after the end of
+	 * the normal request to be able to always advance the RING_TAIL on
+	 * subsequent resubmissions (for lite restore). Should that fail us,
+	 * and we try and submit the same tail again, force the context
+	 * reload.
+	 */
+	tail = intel_ring_set_tail(rq->ring, rq->tail);
+	if (unlikely(ce->lrc_reg_state[CTX_RING_TAIL+1] == tail))
+		desc |= CTX_DESC_FORCE_RESTORE;
+	ce->lrc_reg_state[CTX_RING_TAIL+1] = tail;
+	rq->tail = rq->wa_tail;
 
 	/*
 	 * Make sure the context image is complete before we submit it to HW.
@@ -674,9 +683,7 @@ static u64 execlists_update_context(const struct i915_request *rq)
 	 */
 	mb();
 
-	desc = ce->lrc_desc;
 	ce->lrc_desc &= ~CTX_DESC_FORCE_RESTORE;
-
 	return desc;
 }
 
@@ -1149,16 +1156,6 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 			if (!list_is_last(&last->sched.link,
 					  &engine->active.requests))
 				return;
-
-			/*
-			 * WaIdleLiteRestore:bdw,skl
-			 * Apply the wa NOOPs to prevent
-			 * ring:HEAD == rq:TAIL as we resubmit the
-			 * request. See gen8_emit_fini_breadcrumb() for
-			 * where we prepare the padding after the
-			 * end of the request.
-			 */
-			last->tail = last->wa_tail;
 		}
 	}
 

Last edited by loqs (2019-12-18 20:00:33)

Offline

#57 2019-12-18 21:17:10

joanbrugueram
Member
Registered: 2018-11-12
Posts: 13

Re: i915 Skylake GPU hangs with kernel 5.3.11

loqs wrote:

wioo could ask for a backported to 5.4 version of the patch on https://gitlab.freedesktop.org/drm/intel/issues/673
Edit:
joanbrugueram is this the patch you used?

diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
index 06a506c29463..b70a59cdcdf2 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.c
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -471,12 +471,6 @@ lrc_descriptor(struct intel_context *ce, struct intel_engine_cs *engine)
 	return desc;
 }
 
-static void unwind_wa_tail(struct i915_request *rq)
-{
-	rq->tail = intel_ring_wrap(rq->ring, rq->wa_tail - WA_TAIL_BYTES);
-	assert_ring_tail_valid(rq->ring, rq->tail);
-}
-
 static struct i915_request *
 __unwind_incomplete_requests(struct intel_engine_cs *engine)
 {
@@ -495,7 +489,6 @@ __unwind_incomplete_requests(struct intel_engine_cs *engine)
 			continue; /* XXX */
 
 		__i915_request_unsubmit(rq);
-		unwind_wa_tail(rq);
 
 		/*
 		 * Push the request back into the queue for later resubmission.
@@ -649,13 +642,29 @@ execlists_schedule_out(struct i915_request *rq)
 	i915_request_put(rq);
 }
 
-static u64 execlists_update_context(const struct i915_request *rq)
+static u64 execlists_update_context(struct i915_request *rq)
 {
 	struct intel_context *ce = rq->hw_context;
-	u64 desc;
+	u64 desc = ce->lrc_desc;
+	u32 tail;
 
-	ce->lrc_reg_state[CTX_RING_TAIL + 1] =
-		intel_ring_set_tail(rq->ring, rq->tail);
+	/*
+	 * WaIdleLiteRestore:bdw,skl
+	 *
+	 * We should never submit the context with the same RING_TAIL twice
+	 * just in case we submit an empty ring, which confuses the HW.
+	 *
+	 * We append a couple of NOOPs (gen8_emit_wa_tail) after the end of
+	 * the normal request to be able to always advance the RING_TAIL on
+	 * subsequent resubmissions (for lite restore). Should that fail us,
+	 * and we try and submit the same tail again, force the context
+	 * reload.
+	 */
+	tail = intel_ring_set_tail(rq->ring, rq->tail);
+	if (unlikely(ce->lrc_reg_state[CTX_RING_TAIL+1] == tail))
+		desc |= CTX_DESC_FORCE_RESTORE;
+	ce->lrc_reg_state[CTX_RING_TAIL+1] = tail;
+	rq->tail = rq->wa_tail;
 
 	/*
 	 * Make sure the context image is complete before we submit it to HW.
@@ -674,9 +683,7 @@ static u64 execlists_update_context(const struct i915_request *rq)
 	 */
 	mb();
 
-	desc = ce->lrc_desc;
 	ce->lrc_desc &= ~CTX_DESC_FORCE_RESTORE;
-
 	return desc;
 }
 
@@ -1149,16 +1156,6 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 			if (!list_is_last(&last->sched.link,
 					  &engine->active.requests))
 				return;
-
-			/*
-			 * WaIdleLiteRestore:bdw,skl
-			 * Apply the wa NOOPs to prevent
-			 * ring:HEAD == rq:TAIL as we resubmit the
-			 * request. See gen8_emit_fini_breadcrumb() for
-			 * where we prepare the padding after the
-			 * end of the request.
-			 */
-			last->tail = last->wa_tail;
 		}
 	}
 

Yes, this is exactly the patch I have been using those last few days on 5.4.2 and 5.4.3, and so far I haven't  experienced any of those mini-hangs. But for me they were already infrequent, so it's possible none would have happened anyway without the patch. So it would be great if more people could confirm it fixes the issue for them.

Last edited by joanbrugueram (2019-12-18 21:18:31)

Offline

#58 2019-12-18 22:58:24

wioo
Member
Registered: 2017-05-18
Posts: 24

Re: i915 Skylake GPU hangs with kernel 5.3.11

loqs wrote:

wioo could ask for a backported to 5.4 version of the patch on https://gitlab.freedesktop.org/drm/intel/issues/673
Edit:
joanbrugueram is this the patch you used?

diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
index 06a506c29463..b70a59cdcdf2 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.c
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -471,12 +471,6 @@ lrc_descriptor(struct intel_context *ce, struct intel_engine_cs *engine)
 	return desc;
 }
 
-static void unwind_wa_tail(struct i915_request *rq)
-{
-	rq->tail = intel_ring_wrap(rq->ring, rq->wa_tail - WA_TAIL_BYTES);
-	assert_ring_tail_valid(rq->ring, rq->tail);
-}
-
 static struct i915_request *
 __unwind_incomplete_requests(struct intel_engine_cs *engine)
 {
@@ -495,7 +489,6 @@ __unwind_incomplete_requests(struct intel_engine_cs *engine)
 			continue; /* XXX */
 
 		__i915_request_unsubmit(rq);
-		unwind_wa_tail(rq);
 
 		/*
 		 * Push the request back into the queue for later resubmission.
@@ -649,13 +642,29 @@ execlists_schedule_out(struct i915_request *rq)
 	i915_request_put(rq);
 }
 
-static u64 execlists_update_context(const struct i915_request *rq)
+static u64 execlists_update_context(struct i915_request *rq)
 {
 	struct intel_context *ce = rq->hw_context;
-	u64 desc;
+	u64 desc = ce->lrc_desc;
+	u32 tail;
 
-	ce->lrc_reg_state[CTX_RING_TAIL + 1] =
-		intel_ring_set_tail(rq->ring, rq->tail);
+	/*
+	 * WaIdleLiteRestore:bdw,skl
+	 *
+	 * We should never submit the context with the same RING_TAIL twice
+	 * just in case we submit an empty ring, which confuses the HW.
+	 *
+	 * We append a couple of NOOPs (gen8_emit_wa_tail) after the end of
+	 * the normal request to be able to always advance the RING_TAIL on
+	 * subsequent resubmissions (for lite restore). Should that fail us,
+	 * and we try and submit the same tail again, force the context
+	 * reload.
+	 */
+	tail = intel_ring_set_tail(rq->ring, rq->tail);
+	if (unlikely(ce->lrc_reg_state[CTX_RING_TAIL+1] == tail))
+		desc |= CTX_DESC_FORCE_RESTORE;
+	ce->lrc_reg_state[CTX_RING_TAIL+1] = tail;
+	rq->tail = rq->wa_tail;
 
 	/*
 	 * Make sure the context image is complete before we submit it to HW.
@@ -674,9 +683,7 @@ static u64 execlists_update_context(const struct i915_request *rq)
 	 */
 	mb();
 
-	desc = ce->lrc_desc;
 	ce->lrc_desc &= ~CTX_DESC_FORCE_RESTORE;
-
 	return desc;
 }
 
@@ -1149,16 +1156,6 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 			if (!list_is_last(&last->sched.link,
 					  &engine->active.requests))
 				return;
-
-			/*
-			 * WaIdleLiteRestore:bdw,skl
-			 * Apply the wa NOOPs to prevent
-			 * ring:HEAD == rq:TAIL as we resubmit the
-			 * request. See gen8_emit_fini_breadcrumb() for
-			 * where we prepare the padding after the
-			 * end of the request.
-			 */
-			last->tail = last->wa_tail;
 		}
 	}
 

I applied this patch on top of 5.4.5 (now in testing).

Dmesg look ok

[    1.541492] i915 0000:00:02.0: vgaarb: deactivate vga console
[    1.542163] i915 0000:00:02.0: Direct firmware load for i915/gvt/vid_0x8086_did_0x191b_rid_0x06.golden_hw_state failed with error -2
[    1.558187] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
[    1.558547] [drm] Finished loading DMC firmware i915/skl_dmc_ver1_27.bin (v1.27)
[    1.590839] [drm] Initialized i915 1.6.0 20190822 for 0000:00:02.0 on minor 0
[    1.598511] fbcon: i915drmfb (fb0) is primary device
[    1.631115] i915 0000:00:02.0: fb0: i915drmfb frame buffer device
[    1.673463] i915 0000:00:02.0: MDEV: Registered

Will report if problem appears again.

Offline

#59 2019-12-19 08:15:42

Archanfel80HUN
Member
Registered: 2017-04-27
Posts: 6

Re: i915 Skylake GPU hangs with kernel 5.3.11

installed 5.4.5-arch1-1 today. the issue is not fixed. my system hang after around half hour. the same like before.

Offline

#60 2019-12-19 14:04:39

wioo
Member
Registered: 2017-05-18
Posts: 24

Re: i915 Skylake GPU hangs with kernel 5.3.11

Unfortunately, another gpu hang here

i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0

ultimaweapon, did you try mainline kernel?

Last edited by wioo (2019-12-19 14:06:23)

Offline

#61 2019-12-19 14:08:45

loqs
Member
Registered: 2014-03-06
Posts: 10,083

Re: i915 Skylake GPU hangs with kernel 5.3.11

Offline

#62 2019-12-19 15:54:12

venereo
Member
Registered: 2019-12-19
Posts: 2

Re: i915 Skylake GPU hangs with kernel 5.3.11

Linux manjaro 5.4.2-1-MANJARO #1 SMP PREEMPT Thu Dec 5 09:55:57 UTC 2019 x86_64 GNU/Linux
warning: xf86-video-intel: ignoring package upgrade (1:2.99.917+893+gbff5eca4-1 => 1:2.99.917+897+g0867eea6-1)

Downgrading xf86-video-intel fixed the issue for me

Offline

#63 2019-12-20 01:23:53

Fandekasp
Member
From: Japan
Registered: 2012-02-12
Posts: 22
Website

Re: i915 Skylake GPU hangs with kernel 5.3.11

@venereo did you downgrade on December 5th and never got any issue since?

Will try this, but it's weird that my version was prior to yours (892 vs 893) yet I had many freezes.

:: xf86-video-intel and xf86-video-intel-git are in conflict (xf86-video-intel-sna). Remove xf86-video-intel-git? [y/N] y

Packages (2) xf86-video-intel-git-1:2.99.917+892+gc6cb1b19-2 [removal]
             xf86-video-intel-1:2.99.917+893+gbff5eca4-1

Offline

#64 2019-12-20 02:57:30

venereo
Member
Registered: 2019-12-19
Posts: 2

Re: i915 Skylake GPU hangs with kernel 5.3.11

@Fandekasp 10/12 it's when problematic update happened, been trying to downgrade some of them, xorg didn't worked, this one 3 days no freeze
but It can be just a coincidence

Fandekasp wrote:

@venereo did you downgrade on December 5th and never got any issue since?

Will try this, but it's weird that my version was prior to yours (892 vs 893) yet I had many freezes.

:: xf86-video-intel and xf86-video-intel-git are in conflict (xf86-video-intel-sna). Remove xf86-video-intel-git? [y/N] y

Packages (2) xf86-video-intel-git-1:2.99.917+892+gc6cb1b19-2 [removal]
             xf86-video-intel-1:2.99.917+893+gbff5eca4-1

Last edited by venereo (2019-12-20 02:57:50)

Offline

#65 2019-12-20 06:26:46

ultimaweapon
Member
Registered: 2019-12-18
Posts: 2

Re: i915 Skylake GPU hangs with kernel 5.3.11

wioo wrote:

Unfortunately, another gpu hang here

i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0

ultimaweapon, did you try mainline kernel?

Yes. I'm running on 5.5-rc2 for the whole 2 days and still no freezing. So far so good.

Offline

#66 2019-12-20 08:18:46

V1del
Forum Moderator
Registered: 2012-10-16
Posts: 9,062

Re: i915 Skylake GPU hangs with kernel 5.3.11

In general the xf86-video-intel driver is quite unmaintained, if you don't have an actual reason (expanding multi-monitor screens via bumblebee/intel-virtual-output being the only thing I can think of off-hand) for using it you might want to check whether this is reproducible with it removed and just using the modesetting driver built into xorg.

Offline

#67 2019-12-21 11:10:01

Everette88
Member
Registered: 2018-02-17
Posts: 38

Re: i915 Skylake GPU hangs with kernel 5.3.11

Offline

#68 2019-12-21 12:00:48

V1del
Forum Moderator
Registered: 2012-10-16
Posts: 9,062

Re: i915 Skylake GPU hangs with kernel 5.3.11

It's not really, if you look at the actual commits, nearly nothing is an actual functional change, there are a lot of "Let's try changing this, ooh woops we have to revert because it broke something catastrophically"

Chris Wilson is it's primary developer, and I assume he does consider it his "baby" of some sort and yes there are use cases where it indeed can be the preferred option, it shouldn't be what you opt for as a default on anything that can safely use modesetting.

Also "Intel" as in the company doesn't reccommend it, all mainstream distros don't set it up, actual projects have issues only present on that driver, you might want to look/read through the links in one of my earlier posts on the subject: https://bbs.archlinux.org/viewtopic.php … 4#p1874734

Offline

#69 2019-12-21 14:49:28

Everette88
Member
Registered: 2018-02-17
Posts: 38

Re: i915 Skylake GPU hangs with kernel 5.3.11

It's not really, if you look at the actual commits, nearly nothing is an actual functional change, there are a lot of "Let's try changing this, ooh woops we have to revert because it broke something catastrophically"

I looked at the actual commits and all of them seem fixes for more or less severe issues. Some of them are done in response for user reports from bugtracker. If you report bug for xf86-video-intel driver the you can expect it will be fixed when intel devs found the solution. Of course when some fix causes regression is will be reverted the same as it's done for modesetting driver. I don't think this state can be called "unmaintained" and I believe it's unfair for intel developers.

Chris Wilson is it's primary developer, and I assume he does consider it his "baby" of some sort and yes there are use cases where it indeed can be the preferred option, it shouldn't be what you opt for as a default on anything that can safely use modesetting.

Yes, Chris Wilson is lead developer for intel userspace and kernel drivers. Honestly I value his opinion higher than those from random persons from forums or blogs most of which never read or write any relevant code.

Also "Intel" as in the company doesn't reccommend it, all mainstream distros don't set it up, actual projects have issues only present on that driver, you might want to look/read through the links in one of my earlier posts on the subject

Perhaps "as in the company" Intel doesn't recommend neither xf86-video-intel or modesetting driver. The fact is intel developer(s) recommend xf86-video-intel driver unless you can found and quote any of them favoring modesetting one. It's true that various distro maintainers recommend modesetting but they never confronted this with anyone from intel. Everyone is free to decide who's opinion they trust better but the facts shouldn't be skewed for the one side.

Personally I followed closely discussion about this topic and used modesetting driver for years then switched back to intel one as it was more stable for me.

Offline

#70 2019-12-21 15:35:32

V1del
Forum Moderator
Registered: 2012-10-16
Posts: 9,062

Re: i915 Skylake GPU hangs with kernel 5.3.11

It's really not the point of this topic, it was simply a suggestion, as it would help rule out a potential data point as to what causes/and can be used to circumvent the issue in this topic. The "official" intel graphics stack does not include xf86-video-intel anymore. I also think one should use what's best, for one's system. There are bugs that only happen on xf86-video-intel and bugs that only happen on modesetting. And from what I'm seeing, the bugs on xf86-video-intel with actual visible impact far outweigh the ones on the modesetting driver.

You don't need to follow developer comments for that, just read the boards.

Offline

#71 2019-12-21 16:14:40

Everette88
Member
Registered: 2018-02-17
Posts: 38

Re: i915 Skylake GPU hangs with kernel 5.3.11

It's really not the point of this topic, it was simply a suggestion

I just wanted to clarify your suggestion with two facts. Nothing you wrote undermines those two and the whole discussion shouldn't have been started.

The "official" intel graphics stack" points to linux kernel 4.16 which nobody supports and nobody uses. Informational value of "official" intel graphics stack 2018Q1"  for today is zero.

Offline

#72 2019-12-21 20:56:55

dcc24
Member
Registered: 2009-10-31
Posts: 731

Re: i915 Skylake GPU hangs with kernel 5.3.11

Can we please keep this thread on topic? A lot of us are subscribed to (and get emails from) it because we want to be notified on any developments related to the actual bug, not random people's opinions on tangentially related topics.


It is better to keep your mouth shut and be thought a fool than to open it and remove all doubt. (Mark Twain)

My AUR packages

Offline

#73 2019-12-22 02:55:13

tpfkanep
Member
From: South Africa
Registered: 2009-12-15
Posts: 103

Re: i915 Skylake GPU hangs with kernel 5.3.11

CarbonChauvinist wrote:

[...]
Edit: -- Also, in my case, all my hangs have been while using Chromium which is one of the few applications using XWayland (though I'm using FF now also under XWayland). Also not sure if that's relevant or not.

Mine's with Opera. Played around with downgrading versions: did not help. Since release of linux 5.4, disabled Use hardware acceleration when available in settings, and so far no issues.

Offline

#74 2019-12-24 14:07:04

Tumi
Member
From: Czech Republic
Registered: 2010-04-07
Posts: 9

Re: i915 Skylake GPU hangs with kernel 5.3.11

Hola folks, issue seems to be still present for me, however much less common.

  1. It happens infrequently on both linux (5.4.6.arch1-1) and linux-zen (5.4.6.zen1-1) kernels.

  2. On a machine with Intel Core i7-7600U CPU (Intel(R) HD Graphics 620) - Kaby Lake.

  3. Using X11 with KDE Plasma and hardware accelerated Chromium - chromium-vaapi-bin (79.0.3945.88-1).

  4. The xf86-video-intel package is not installed.

The actual dmesg output is still the same:

kernel: i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0
...
kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
kernel: [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
kernel: i915 0000:00:02.0: Resetting chip for hang on rcs0
kernel: [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
kernel: [drm:gen8_reset_engines [i915]] *ERROR* rcs0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
kernel: [drm] GuC communication enabled
kernel: i915 0000:00:02.0: GuC firmware i915/kbl_guc_33.0.0.bin version 33.0 submission:disabled
kernel: i915 0000:00:02.0: HuC firmware i915/kbl_huc_ver02_00_1810.bin version 2.0 authenticated:yes
...
kernel: GpuWatchdog[3073]: segfault at 0 ip 000055c965180858 sp 00007f8aa4839470 error 6 in chromium[55c96156b000+7041000]
kernel: Code: 48 c1 c9 03 48 81 f9 af 00 00 00 0f 87 a8 00 00 00 48 8d 15 d2 08 dd fb f6 04 11 01 0f 84 97 00 00 00 be 01 00 00 00 ff 50 30 <c7> 04 25 00 00 00 00 37 13 00 00 c6 05 be 63 b7 03 01 eb 5b 49 8b
kernel: audit: type=1701 audit(1577194933.441:119): auid=1000 uid=1000 gid=1000 ses=3 pid=3050 comm="GpuWatchdog" exe="/usr/lib/chromium/chromium" sig=11 res=1
...

Kernel linux-lts works just fine, as expected.

Offline

#75 2019-12-26 12:34:51

ingolfured
Member
Registered: 2019-08-07
Posts: 2

Re: i915 Skylake GPU hangs with kernel 5.3.11

I had the same problem. I couldn't downgrade to linux-lts so upgrading to linux-mainline did the trick for me.

Offline

Board footer

Powered by FluxBB