diff options
author | Patrick J Volkerding <volkerdi@slackware.com> | 2023-07-18 19:58:10 +0000 |
---|---|---|
committer | Eric Hameleers <alien@slackware.com> | 2023-07-18 22:30:10 +0200 |
commit | a1b07eafc1e2522790168868d732f16d0c442ff8 (patch) | |
tree | 83fe618f3da1e4ca9cc12de4b74f1719de354018 /ChangeLog.rss | |
parent | 53caf4f3e2e7fff4c43ef1d9e92949049e63a1b0 (diff) | |
download | current-a1b07eafc1e2522790168868d732f16d0c442ff8.tar.gz current-a1b07eafc1e2522790168868d732f16d0c442ff8.tar.xz |
Tue Jul 18 19:58:10 UTC 202320230718195810
a/tar-1.35-x86_64-1.txz: Upgraded.
x/mesa-23.2.0_rc1-x86_64-1.txz: Upgraded.
OK, usually I won't use rc versions even in -current, but in this case I'm
going to. Some time ago my desktop machine with RS880 / Radeon HD 4290
graphics on the motherboard began acting up with the screen going black
for a few seconds before returning. This after an hour or so of uptime,
usually, then becoming more frequent with more uptime. Eventually I'd lose
mouse and/or keyboard too, and have to reboot. Here's a couple of errors
from dmesg:
[ 9942.677675] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait
timed out.
[ 9942.677741] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed
testing IB on GFX ring (-110).
I also noticed that the backtrace started with ttm_bo_release, and seeing
this in recent kernel patches had been chalking this up to a kernel bug.
I *still* think it could be, and there are a bunch of kernel patches coming
soon to -stable from Alex Deucher that could address the underlying causes
(not for 6.1.39 though, unfortunately). Anyway, I'd recently figured out
that reverting Mesa sufficiently made the issue go away. And now it seems
this 23.2.0 release candidate also fixes the issue.
Yes, I could go search for the commits to cherry-pick, but we'll be moving
to mesa-23.2.0 when it's released, so we might as well start testing now.
Diffstat (limited to 'ChangeLog.rss')
-rw-r--r-- | ChangeLog.rss | 36 |
1 files changed, 34 insertions, 2 deletions
diff --git a/ChangeLog.rss b/ChangeLog.rss index c34b46f44..1b47af390 100644 --- a/ChangeLog.rss +++ b/ChangeLog.rss @@ -11,10 +11,42 @@ <description>Tracking Slackware development in git.</description> <language>en-us</language> <id xmlns="http://www.w3.org/2005/Atom">urn:uuid:c964f45e-6732-11e8-bbe5-107b4450212f</id> - <pubDate>Mon, 17 Jul 2023 19:17:19 GMT</pubDate> - <lastBuildDate>Mon, 17 Jul 2023 19:59:01 GMT</lastBuildDate> + <pubDate>Tue, 18 Jul 2023 19:58:10 GMT</pubDate> + <lastBuildDate>Tue, 18 Jul 2023 20:30:06 GMT</lastBuildDate> <generator>maintain_current_git.sh v 1.17</generator> <item> + <title>Tue, 18 Jul 2023 19:58:10 GMT</title> + <pubDate>Tue, 18 Jul 2023 19:58:10 GMT</pubDate> + <link>https://git.slackware.nl/current/tag/?h=20230718195810</link> + <guid isPermaLink="false">20230718195810</guid> + <description> + <![CDATA[<pre> +a/tar-1.35-x86_64-1.txz: Upgraded. +x/mesa-23.2.0_rc1-x86_64-1.txz: Upgraded. + OK, usually I won't use rc versions even in -current, but in this case I'm + going to. Some time ago my desktop machine with RS880 / Radeon HD 4290 + graphics on the motherboard began acting up with the screen going black + for a few seconds before returning. This after an hour or so of uptime, + usually, then becoming more frequent with more uptime. Eventually I'd lose + mouse and/or keyboard too, and have to reboot. Here's a couple of errors + from dmesg: + [ 9942.677675] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait + timed out. + [ 9942.677741] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed + testing IB on GFX ring (-110). + I also noticed that the backtrace started with ttm_bo_release, and seeing + this in recent kernel patches had been chalking this up to a kernel bug. + I *still* think it could be, and there are a bunch of kernel patches coming + soon to -stable from Alex Deucher that could address the underlying causes + (not for 6.1.39 though, unfortunately). Anyway, I'd recently figured out + that reverting Mesa sufficiently made the issue go away. And now it seems + this 23.2.0 release candidate also fixes the issue. + Yes, I could go search for the commits to cherry-pick, but we'll be moving + to mesa-23.2.0 when it's released, so we might as well start testing now. + </pre>]]> + </description> + </item> + <item> <title>Mon, 17 Jul 2023 19:17:19 GMT</title> <pubDate>Mon, 17 Jul 2023 19:17:19 GMT</pubDate> <link>https://git.slackware.nl/current/tag/?h=20230717191719</link> |