NetBSD Problem Report #51675
From dholland@netbsd.org Tue Nov 29 23:53:32 2016
Return-Path: <dholland@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
(Client CN "mail.netbsd.org", Issuer "Postmaster NetBSD.org" (verified OK))
by mollari.NetBSD.org (Postfix) with ESMTPS id 2DAFF7A2CC
for <gnats-bugs@gnats.NetBSD.org>; Tue, 29 Nov 2016 23:53:32 +0000 (UTC)
Message-Id: <20161129235326.53AB6855E0@mail.netbsd.org>
Date: Tue, 29 Nov 2016 23:53:26 +0000 (UTC)
From: dholland@netbsd.org
Reply-To: dholland@netbsd.org
To: gnats-bugs@NetBSD.org
Subject: radeondrmkms wedge with supertuxkart
X-Send-Pr-Version: 3.95
>Number: 51675
>Category: kern
>Synopsis: radeondrmkms wedge with supertuxkart
>Confidential: no
>Severity: critical
>Priority: medium
>Responsible: riastradh
>State: closed
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Tue Nov 29 23:55:00 +0000 2016
>Closed-Date: Tue Apr 16 14:34:51 +0000 2019
>Last-Modified: Tue Apr 16 14:34:51 +0000 2019
>Originator: David A. Holland
>Release: NetBSD 7.99.42 (20161125)
>Organization:
>Environment:
System: NetBSD valkyrie 7.99.42 NetBSD 7.99.42 (VALKYRIE) #20: Fri Nov 25 18:33:19 EST 2016 dholland@valkyrie:/usr/src/sys/arch/amd64/compile/VALKYRIE amd64
Architecture: x86_64
Machine: amd64
>Description:
Now that we think 50349 is fixed I updated another machine with the
same radeon card to -current, and discovered that supertuxkart wedges
the X server and/or console.
In particular, when you open supertuxkart and start a race, the X
server hangs cold. It doesn't seem to wedge the rest of the system,
although my options for investigating the wedge state are somewhat
limited at the moment.
Once when I did this supertuxkart cored instead, which is maybe/maybe
not a different issue.
While poking around I found that running teapot from MesaDemos dumps
core, as does glxgears; running with env RADEON_THREAD=FALSE (as
suggested in 49838) to disable the thread in the gallium driver makes
glxgears run, but not teapot. (This much requires a fix to the gallium
driver build, or it doesn't load; but that's been committed.)
I haven't tried the thread thing with supertuxkart yet, but I will
next go.
>How-To-Repeat:
>Fix:
>Release-Note:
>Audit-Trail:
Responsible-Changed-From-To: kern-bug-people->riastradh
Responsible-Changed-By: dholland@NetBSD.org
Responsible-Changed-When: Tue, 29 Nov 2016 23:56:05 +0000
Responsible-Changed-Why:
fyi
(maybe we should have a drmkms-bug-people)
State-Changed-From-To: open->closed
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Fri, 18 Jan 2019 02:53:38 +0000
State-Changed-Why:
The system crashes are fixed by sys/external/bsd/drm2/pci/drm_pci.c -r1.32.
I made a new PR for the X server segfaults: 53888.
State-Changed-From-To: closed->pending-pullups
State-Changed-By: mrg@NetBSD.org
State-Changed-When: Fri, 18 Jan 2019 06:47:39 +0000
State-Changed-Why:
this is likely a problem on both -7 and -8. let's do pullups.
State-Changed-From-To: pending-pullups->needs-pullups
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Fri, 18 Jan 2019 07:55:55 +0000
State-Changed-Why:
concur, but they haven't been filed yet
From: David Holland <dholland@netbsd.org>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: kern/51675 (radeondrmkms wedge with supertuxkart)
Date: Fri, 18 Jan 2019 09:07:33 +0000
On Fri, Jan 18, 2019 at 07:55:56AM +0000, dholland@NetBSD.org wrote:
> concur, but they haven't been filed yet
The merge fails on both -8 and -7 because the comment following the
change is different; the following patch applies (to both), but
someone who knows more about the code should eyeball the results to
make sure it's still valid.
Index: drm_pci.c
===================================================================
RCS file: /cvsroot/src/sys/external/bsd/drm2/pci/drm_pci.c,v
retrieving revision 1.17.2.1
diff -u -r1.17.2.1 drm_pci.c
--- drm_pci.c 1 Aug 2017 23:12:06 -0000 1.17.2.1
+++ drm_pci.c 18 Jan 2019 08:19:17 -0000
@@ -145,6 +145,14 @@
continue;
}
+ /*
+ * If it's a 64-bit mapping, don't interpret the second
+ * half of it as another BAR in the next iteration of
+ * the loop -- move on to the next unit.
+ */
+ if (PCI_MAPREG_MEM_TYPE(type) == PCI_MAPREG_MEM_TYPE_64BIT)
+ unit++;
+
/* Inquire about it. We'll map it in drm_core_ioremap. */
if (pci_mapreg_info(pa->pa_pc, pa->pa_tag, reg, type,
&bm->bm_base, &bm->bm_size, &bm->bm_flags) != 0) {
--
David A. Holland
dholland@netbsd.org
From: matthew green <mrg@eterna.com.au>
To: gnats-bugs@NetBSD.org
Cc: riastradh@NetBSD.org, gnats-admin@netbsd.org, netbsd-bugs@netbsd.org,
dholland@netbsd.org
Subject: re: kern/51675 (radeondrmkms wedge with supertuxkart)
Date: Sat, 19 Jan 2019 07:20:55 +1100
> The merge fails on both -8 and -7 because the comment following the
> change is different; the following patch applies (to both), but
> someone who knows more about the code should eyeball the results to
> make sure it's still valid.
i've been running with this change on -8 for 11 hours now.
.mrg.
State-Changed-From-To: needs-pullups->pending-pullups
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Sun, 20 Jan 2019 06:43:23 +0000
State-Changed-Why:
pullup-8 #1165
pullup-7 #1673
State-Changed-From-To: pending-pullups->closed
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Tue, 16 Apr 2019 14:34:51 +0000
State-Changed-Why:
Pulled up.
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.43 2018/01/16 07:36:43 maya Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2017
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.