http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 Michiel Janssens changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |michiel@nexigon.net -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 http://bugzilla.opensuse.org/show_bug.cgi?id=1168776#c4 --- Comment #4 from Achim Gratz --- (In reply to Takashi Iwai from comment #1)

So the message appears repeatedly?

Yes, several times a second actually, in bursts of a few seconds apart.

Interesting, a patch to make the message to DRM_INFO_ONCE() was submitted in more than year ago, but it didn't get merged. https://lore.kernel.org/patchwork/patch/960241/

Yes I'd seen that, but it'

And you've hit this while it didn't happen on 5.5. So the call pattern has changed between 5.5 and 5.6 and starts triggering it...

Or the kernel decides to use the reported BIOS features differently. THe other thing that irks me is that I am apparently unable to switch off the framebuffer compression even though the kernel parameter was recognized. What's up with that? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 http://bugzilla.opensuse.org/show_bug.cgi?id=1168776#c6 Achim Gratz changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(Stromeko@NexGo.DE | |) | --- Comment #6 from Achim Gratz --- (In reply to Achim Gratz from comment #5)

(In reply to Achim Gratz from comment #3)

...

(In reply to Jiri Slaby from comment #2)

...

What do you have in /sys/kernel/debug/dri/0/i915_gem_stolen in both kernels?

That sysfs file does not exist under the 5.5.13 I'm currently running. I'll have to reboot into the 5.6.0 kernel later.

The new kernel doesn't have that either... maybe this is what you were looking for?

# cat /sys/kernel/debug/dri/0/i915_gem_objects 391 shrinkable [0 free] objects, 594219008 bytes system: total:0x00000001eb41d000, available:0x00000001eb41d000 bytes stolen: total:0x0000000002000000, available:0x0000000002000000 bytes

the same output from the 5.5.13 kernel is just: # cat /sys/kernel/debug/dri/0/i915_gem_objects 367 shrinkable [0 free] objects, 591519744 bytes OK, so I dove into /sys/module/i915/parameters and see that fbc_enable is reading "0". So indeed the 5.5.13 kernel did _not_ switch on fb compression. Switching it on manually yields this: [ 617.221643] Setting dangerous option enable_fbc - tainting kernel [ 617.267790] [drm] Reducing the compressed framebuffer size. This may lead to less power savings than a non-reduced-size. Try to increase stolen memory size if available in BIOS. So far that has been the only message. So I start to think that the bug with 5.6.0 is actually that the fb compression gets set to on unconditionally (as evidenced by the fact that the kernel parameter was read as "0" on boot, but the compression stayed on anyway). I might try to manually disable it via sysfs on 5.6.0 later, but I have to shut down for today. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 http://bugzilla.opensuse.org/show_bug.cgi?id=1168776#c8 --- Comment #8 from Achim Gratz --- The plot thickens... back on the original system (Xeon E3-1225v3 and a 4k display); FBC is actually _not_ on: # cat /sys/kernel/debug/dri/0/i915_fbc_status FBC disabled: disabled at runtime per module param That likely explains why trying to switch it off just taints the kernel and doesn't seem to do aynthing else. # lspci -vv -s 00:02.0 00:02.0 VGA compatible controller: Intel Corporation Xeon E3-1200 v3 Processor Integrated Graphics Controller (rev 06) (prog-if 00 [VGA controller]) DeviceName: Onboard IGD Subsystem: Dell Device 0620 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+ Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort- /sys/module/i915/parameters/enable_fbc # cat /sys/kernel/debug/dri/0/i915_fbc_status FBC disabled: pixel rate is too big The messages continue to be spewed out, but the actual error that trigger them is probably unrelated to framebuffer compression and a bug somewhere else in the driver. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 http://bugzilla.opensuse.org/show_bug.cgi?id=1168776#c11 --- Comment #11 from Takashi Iwai --- Well, basically there is no chance to fix such a bug in the distro side for TW. This must be handled in the upstream, and they need the proper bug report. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 http://bugzilla.opensuse.org/show_bug.cgi?id=1168776#c13 --- Comment #13 from Achim Gratz --- (In reply to Takashi Iwai from comment #11)

Well, basically there is no chance to fix such a bug in the distro side for TW. This must be handled in the upstream, and they need the proper bug report.

Well, one would have thought that somebody has figured out a way to "upstream" distro bugreports by now, as that's not exactly an uncommon occurence. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 http://bugzilla.opensuse.org/show_bug.cgi?id=1168776#c15 --- Comment #15 from Achim Gratz --- So upstreams solution it to apply the almost two year old patch that makes the message appear only once per boot. Can that be backported to 5.6? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=1168776 http://bugzilla.opensuse.org/show_bug.cgi?id=1168776#c17 --- Comment #17 from Achim Gratz --- (In reply to Takashi Iwai from comment #16)

Thanks for the update. Now backported to stable branch. It looks merely like a workaround rather than a proper fix, at least the annoyance would be gone by this, so let's close now.

Yes, it seems to be papering over the issue of why we get that message in the first place. FB compression is already off for this display size, so the message must get triggered by some other error. I haven't got the round tuit to go look by which path it snakes there.

Feel free to reopen if you still encounter the problem even with the patch.

I've been running that for two days now (not too thoroughly stressed) and aside from the missing journal entries everything still seems to work as it should. -- You are receiving this mail because: You are on the CC list for the bug.