From: Max Krummenacher <max.oss.09@gmail.com>
To: reatmon@ti.com
Cc: Francesco Dolcini <francesco@dolcini.it>,
Randolph Sapp <rs@ti.com>,
meta-ti@lists.yoctoproject.org, max.krummenacher@toradex.com,
Nishanth Menon <nm@ti.com>
Subject: Re: [meta-ti][kirkstone] regression with rogue gpu
Date: Thu, 16 Nov 2023 10:33:33 +0100 [thread overview]
Message-ID: <CAEHkU3WGDxnz9w+Fz1D_kvysQn8+7AgzzRbDetHpH_EqKnwiYA@mail.gmail.com> (raw)
In-Reply-To: <62161039-d848-4e58-a829-b866689754df@ti.com>
Hi Ryan
On Tue, Nov 14, 2023 at 6:37 PM Ryan Eatmon via lists.yoctoproject.org
<reatmon=ti.com@lists.yoctoproject.org> wrote:
>
> On 11/14/2023 10:17 AM, Francesco Dolcini wrote:
> > Hello Randolph,
> >
> > On Mon, Oct 23, 2023 at 01:15:38PM -0300, Max Krummenacher wrote:
> >> On our weekly CI build we see glmark segfaulting on our AM625 based SoM.
> >>
> >> Reverting commit 9b6bce2c ("mesa-pvr: bump srcrev for sgx support") on
> >> top of meta-ti @ 59edb4e8 ("linux-ti-staging_6.1: CI/CD Auto-Merger: cicd.kirkstone.202310191725")
> >> fixes the segfault and glmark shows similar performance as before.
> >>
> >> It looks like the addition of the SGX GPU introduced a regression with
> >> the Rogue GPU.
> >
> > any feedback on this report? We would need to update meta-ti to get a
> > fix on the DM firmware, however we are blocked because of this
> > regression on the GPU.
> >
> > Can you help figure out what's going on? Is there any dependency on the
> > Linux kernel from your downstream branch we need to cherry-pick? Is this
> > working on TI SK AM62?
>
> I believe that this issue was fixed by this commit:
>
> https://git.ti.com/cgit/arago-project/meta-ti/commit/?h=kirkstone&id=00de85a06e252fa17ce004635ba30d055166aa5d
>
> This is on the main kirkstone branch. Please test it to make sure it
> does fix the issue.
>
Actually the issue persists.
I tested our setup with meta-ti in a version containing commit
00de85a06 with the same result as before that commit.
When using the GPU, i.e. by starting glmark-es2-wayland or
weston-simple-egl the process segfaults.
I did setup arago and built the tisdk-default-image for
am62xx-evm to exclude that our setup somehow messes with the
GPU or we need to merge stuff into our U-Boot / Kernel trees.
Running that image on SK-AM62 shows the same segfault when
using the GPU.
Regards
Max
root@am62xx-evm:~# glmark2-es2-wayland
=======================================================
glmark2 2021.12
=======================================================
OpenGL Information
GL_VENDOR: Imagination Technologies
GL_RENDERER: PowerVR A-Series AXE-1-16M
GL_VERSION: OpenGL ES 3.1 build 23.2@6460340
=======================================================
[build] use-vbo=false:[ 26.004288] kauditd_printk_skb: 1 callbacks suppressed
[ 26.004317] audit: type=1701 audit(1651169516.600:15):
auid=4294967295 uid=0 gid=0 ses=4294967295 pid=726
comm="glmark2-es2-way"
exe="/usr/bin/glmark2-es2-wayland" sig=11 res=1
[ 26.071116] audit: type=1334 audit(1651169516.664:16): prog-id=13 op=LOAD
[ 26.078085] audit: type=1334 audit(1651169516.672:17): prog-id=14 op=LOAD
Segmentation fault (core dumped)
root@am62xx-evm:~# [ 27.111672] audit: type=1334
audit(1651169517.708:18): prog-id=14 op=UNLOAD
[ 27.119012] audit: type=1334 audit(1651169517.708:19): prog-id=13 op=UNLOAD
root@am62xx-evm:~# weston-simple-egl
has EGL_EXT_buffer_age and EGL_EXT_swap_buffers_with_damage
[ 38.378006] audit: type=1701 audit(1651169528.972:20):
auid=4294967295 uid=0 gid=0 ses=4294967295 pid=733
comm="weston-simple-e"
exe="/usr/bin/weston-simple-egl" sig=11 res=1
[ 38.406617] audit: type=1334 audit(1651169529.000:21): prog-id=15 op=LOAD
[ 38.413536] audit: type=1334 audit(1651169529.008:22): prog-id=16 op=LOAD
Segmentation fault (core dumped)
root@am62xx-evm:~# [ 39.179515] audit: type=1334
audit(1651169529.776:23): prog-id=16 op=UNLOAD
[ 39.186626] audit: type=1334 audit(1651169529.776:24): prog-id=15 op=UNLOAD
root@am62xx-evm:~#
Used layer hashes:
Build Configuration (mc:default):
BB_VERSION = "2.0.0"
BUILD_SYS = "x86_64-linux"
NATIVELSBSTRING = "fedora-38"
TARGET_SYS = "aarch64-oe-linux"
MACHINE = "am62xx-evm"
DISTRO = "arago"
DISTRO_VERSION = "2023.10"
TUNE_FEATURES = "aarch64"
TARGET_FPU = ""
meta-arago-distro
meta-arago-extras
meta-arago-demos
meta-arago-test = "kirkstone:1bbf7db62c15977004217f83fb72fdf165597bad"
meta-qt5 = "kirkstone:ae8a97f79364bed1abc297636f7933d0e35f22be"
meta-virtualization = "kirkstone:478a91800c4a95c7410dc8ef52e229a81be24b4e"
meta-networking
meta-python
meta-oe
meta-gnome
meta-filesystems = "kirkstone:7da6cb848bc42b3e6bd5d2b37b52ba75510a6ca0"
meta-ti-extras
meta-ti-bsp = "kirkstone:9efb0c6fc4f5089f4f2a305c2823dba311505487"
meta-arm
meta-arm-toolchain = "kirkstone:b187fb9232ca0a6b5f8f90b4715958546fc41d73"
meta = "kirkstone:4bb6373e5f4a1330a063d1afe855d6c24d5461e7"
>
>
> > Thanks for your support,
> > Francesco
> >
>
> --
> Ryan Eatmon reatmon@ti.com
> -----------------------------------------
> Texas Instruments, Inc. - LCPD - MGTS
>
> -=-=-=-=-=-=-=-=-=-=-=-
> Links: You receive all messages sent to this group.
> View/Reply Online (#17297): https://lists.yoctoproject.org/g/meta-ti/message/17297
> Mute This Topic: https://lists.yoctoproject.org/mt/102138831/3617484
> Group Owner: meta-ti+owner@lists.yoctoproject.org
> Unsubscribe: https://lists.yoctoproject.org/g/meta-ti/leave/11972926/3617484/839770832/xyzzy [max.oss.09@gmail.com]
> -=-=-=-=-=-=-=-=-=-=-=-
>
prev parent reply other threads:[~2023-11-16 9:33 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-23 16:15 [meta-ti][kirkstone] regression with rogue gpu Max Krummenacher
[not found] ` <ZVOdgoK/6uaA76cH@francesco-nb.int.toradex.com>
2023-11-14 17:36 ` Ryan Eatmon
2023-11-16 9:33 ` Max Krummenacher [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAEHkU3WGDxnz9w+Fz1D_kvysQn8+7AgzzRbDetHpH_EqKnwiYA@mail.gmail.com \
--to=max.oss.09@gmail.com \
--cc=francesco@dolcini.it \
--cc=max.krummenacher@toradex.com \
--cc=meta-ti@lists.yoctoproject.org \
--cc=nm@ti.com \
--cc=reatmon@ti.com \
--cc=rs@ti.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).