<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
Well, essentially we shouldn't print the error from the kernel in
the first place and just return the correct error code to the
userspace application.<br>
<br>
The userspace application/driver would then say: Oh, I don't have
enough memory for triple buffering, maybe I should just fallback to
double buffering....<br>
<br>
Retrying the same approach over and over again is not really a good
implementation from userspace.<br>
<br>
But devices with only 64MiB of VRAM are rare today, so....<br>
<br>
Regards,<br>
Christian.<br>
<br>
<div class="moz-cite-prefix">Am 04.12.23 um 17:38 schrieb smesgr:<br>
</div>
<blockquote type="cite"
cite="mid:_ghIPI7KEaEIZNuxoFM2Gui55Qu12bw_3fiiU7mO3rwnEBGwLyYT1fCdXb8o4xej6Pw0ApccKIgXRo2_lbeibCDx3EfM1jGkcTcdop9vSww=@protonmail.com">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<div>Hi Chrisitan,</div>
<div
style="font-family: Arial, sans-serif; font-size: 14px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);"><br>
</div>
<div
style="font-family: Arial, sans-serif; font-size: 14px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">you
are right, after bumping the memory in BIOS the error goes away.
Still would be nice if the error would point to "out of VMem" or
something.<br>
</div>
<div
style="font-family: Arial, sans-serif; font-size: 14px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);"><br>
</div>
<div
style="font-family: Arial, sans-serif; font-size: 14px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">Best
Regards<br>
</div>
<div class="protonmail_quote"> Christian König
<a class="moz-txt-link-rfc2396E" href="mailto:ckoenig.leichtzumerken@gmail.com"><ckoenig.leichtzumerken@gmail.com></a> schrieb am Montag, 4.
Dezember 2023 um 16:18:<br>
<br>
<blockquote class="protonmail_quote" type="cite"> Am 02.12.23 um
20:00 schrieb smesgr:<br>
<blockquote type="cite"> <span
style="background-color: rgb(255, 255, 255);">Hi,</span>
<div style="background-color: rgb(255, 255, 255);"><br>
</div>
<div style="background-color: rgb(255, 255, 255);">I have
installed Debian 12 on my old Fujitsu S920 with AMD
GX-424CC.<br>
</div>
<div style="background-color: rgb(255, 255, 255);">After the
installation completed, my logs are flooded by the
following error messages:</div>
<div style="background-color: rgb(255, 255, 255);"><br>
</div>
<div style="background-color: rgb(255, 255, 255);">dmesg:</div>
<div style="background-color: rgb(255, 255, 255);"><span>[
967.869183] radeon 0000:00:01.0: 000000003ba3c813 pin
failed</span>
<div><span>[ 967.869212] [drm:radeon_crtc_page_flip_target
[radeon]] *ERROR* failed to pin new rbo buffer before
flip</span></div>
<div><span>[ 967.886254] radeon 0000:00:01.0:
0000000051be2216 pin failed</span></div>
<div><span>[ 967.886282] [drm:radeon_crtc_page_flip_target
[radeon]] *ERROR* failed to pin new rbo buffer before
flip</span></div>
<div><span>[ 967.907984] radeon 0000:00:01.0:
000000003ba3c813 pin failed</span></div>
<div><span>[ 967.908014] [drm:radeon_crtc_page_flip_target
[radeon]] *ERROR* failed to pin new rbo buffer before
flip</span></div>
<div><span>[ 967.925551] radeon 0000:00:01.0:
0000000051be2216 pin failed</span></div>
<span>[ 967.925580] [drm:radeon_crtc_page_flip_target
[radeon]] *ERROR* failed to pin new rbo buffer before
flip</span></div>
<div style="background-color: rgb(255, 255, 255);"><br>
</div>
<div style="background-color: rgb(255, 255, 255);">Xorg.log</div>
<div style="background-color: rgb(255, 255, 255);"><span>[
47.777] (WW) RADEON(0): flip queue failed: Invalid
argument</span>
<div><span>[ 47.777] (WW) RADEON(0): Page flip failed:
Invalid argument</span></div>
<span>[ 47.777] (EE) RADEON(0): present flip failed</span></div>
<div style="background-color: rgb(255, 255, 255);"><br>
</div>
<div style="background-color: rgb(255, 255, 255);">lspci:</div>
<div style="background-color: rgb(255, 255, 255);"><span>00:01.0
VGA compatible controller: Advanced Micro Devices, Inc.
[AMD/ATI] Mullins [Radeon R4/R5 Graphics] (rev 01)
(prog-if 00 [VGA controller])</span>
<div><span> Subsystem: Fujitsu Technology Solutions
Mullins [Radeon R4/R5 Graphics]</span></div>
<div><span> Control: I/O+ Mem+ BusMaster+ SpecCycle-
MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
DisINTx+</span></div>
<div><span> Status: Cap+ 66MHz- UDF- FastB2B- ParErr-
DEVSEL=fast >TAbort- <TAbort- <MAbort-
>SERR- <PERR- INTx-</span></div>
<div><span> Latency: 0, Cache Line Size: 64 bytes</span></div>
<div><span> Interrupt: pin A routed to IRQ 35</span></div>
<div><span> Region 0: Memory at f8000000 (64-bit,
prefetchable) [size=64M]</span></div>
<div><span> Region 2: Memory at fc000000 (64-bit,
prefetchable) [size=8M]</span></div>
<div><span> Region 4: I/O ports at f000 [size=256]</span></div>
<div><span> Region 5: Memory at feb00000 (32-bit,
non-prefetchable) [size=256K]</span></div>
<div><span> Expansion ROM at 000c0000 [disabled]
[size=128K]</span></div>
<div><span> Capabilities: [48] Vendor Specific
Information: Len=08 <?></span></div>
<div><span> Capabilities: [50] Power Management version 3</span></div>
<div><span> Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA
PME(D0-,D1+,D2+,D3hot+,D3cold-)</span></div>
<div><span> Status: D0 NoSoftRst- PME-Enable- DSel=0
DScale=0 PME-</span></div>
<div><span> Capabilities: [58] Express (v2) Root Complex
Integrated Endpoint, MSI 00</span></div>
<div><span> DevCap: MaxPayload 256 bytes, PhantFunc 0</span></div>
<div><span> ExtTag+ RBE+ FLReset-</span></div>
<div><span> DevCtl: CorrErr- NonFatalErr- FatalErr-
UnsupReq-</span></div>
<div><span> RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+</span></div>
<div><span> MaxPayload 256 bytes, MaxReadReq 512 bytes</span></div>
<div><span> DevSta: CorrErr- NonFatalErr+ FatalErr-
UnsupReq+ AuxPwr- TransPend-</span></div>
<div><span> DevCap2: Completion Timeout: Not Supported,
TimeoutDis- NROPrPrP- LTR-</span></div>
<div><span> 10BitTagComp- 10BitTagReq- OBFF Not Supported,
ExtFmt+ EETLPPrefix+, MaxEETLPPrefixes 1</span></div>
<div><span> EmergencyPowerReduction Not Supported,
EmergencyPowerReductionInit-</span></div>
<div><span> FRS-</span></div>
<div><span> AtomicOpsCap: 32bit- 64bit- 128bitCAS-</span></div>
<div><span> DevCtl2: Completion Timeout: 50us to 50ms,
TimeoutDis- LTR- 10BitTagReq- OBFF Disabled,</span></div>
<div><span> AtomicOpsCtl: ReqEn-</span></div>
<div><span> Capabilities: [a0] MSI: Enable+ Count=1/1
Maskable- 64bit+</span></div>
<div><span> Address: 00000000fee08004 Data: 0023</span></div>
<div><span> Capabilities: [100 v1] Vendor Specific
Information: ID=0001 Rev=1 Len=010 <?></span></div>
<div><span> Capabilities: [270 v1] Secondary PCI Express</span></div>
<div><span> LnkCtl3: LnkEquIntrruptEn- PerformEqu-</span></div>
<div><span> LaneErrStat: 0</span></div>
<div><span> Capabilities: [2b0 v1] Address Translation
Service (ATS)</span></div>
<div><span> ATSCap: Invalidate Queue Depth: 00</span></div>
<div><span> ATSCtl: Enable-, Smallest Translation Unit: 00</span></div>
<div><span> Capabilities: [2c0 v1] Page Request Interface
(PRI)</span></div>
<div><span> PRICtl: Enable- Reset-</span></div>
<div><span> PRISta: RF- UPRGI- Stopped+</span></div>
<div><span> Page Request Capacity: 00000020, Page Request
Allocation: 00000000</span></div>
<div><span> Capabilities: [2d0 v1] Process Address Space
ID (PASID)</span></div>
<div><span> PASIDCap: Exec+ Priv+, Max PASID Width: 10</span></div>
<div><span> PASIDCtl: Enable- Exec- Priv-</span></div>
<div><span> Kernel driver in use: radeon</span></div>
<span> Kernel modules: radeon, amdgpu</span></div>
<div style="background-color: rgb(255, 255, 255);"><br>
</div>
<div style="background-color: rgb(255, 255, 255);">I'm a bit
confused. Is this an kernel issue, Xorg issue or something
else?</div>
</blockquote>
<br>
Well, what is connected to the GPU? E.g. what monitor,
projector or other display device do you use?<br>
<br>
What happens is that only 64MiB is assigned to the device and
for some reason a process (maybe the compositor?) Is trying to
do double or triple buffering and doesn't has enough memory
for that.<br>
<br>
So it falls back to single buffering, that might work but will
probably not look as good and you see tons of error messages.<br>
<br>
You most likely can go into the BIOS and reserve more memory
to the GPU (it will just be stolen from system memory).<br>
<br>
Regards,<br>
Christian.<br>
<br>
<blockquote type="cite">
<div style="background-color: rgb(255, 255, 255);"><br>
</div>
<div style="background-color: rgb(255, 255, 255);">kernel:</div>
<div style="background-color: rgb(255, 255, 255);"><span>Linux
debian 6.1.0-13-amd64 #1 SMP PREEMPT_DYNAMIC Debian
6.1.55-1 (2023-09-29) x86_64 GNU/Linux</span><br>
<span></span></div>
<div style="background-color: rgb(255, 255, 255);"><br>
</div>
<div style="background-color: rgb(255, 255, 255);">Xorg:<br>
</div>
<div style="background-color: rgb(255, 255, 255);"><span>[
33.144] </span>
<div><span><a target="_blank"
rel="noreferrer nofollow noopener"
href="http://X.Org" moz-do-not-send="true">X.Org</a>
X Server 1.21.1.7</span></div>
<div><span>X Protocol Version 11, Revision 0</span></div>
<div><span>[ 33.144] Current Operating System: Linux
debian 6.1.0-13-amd64 #1 SMP PREEMPT_DYNAMIC Debian
6.1.55-1 (2023-09-29) x86_64</span></div>
<div><span>[ 33.144] Kernel command line:
BOOT_IMAGE=/boot/vmlinuz-6.1.0-13-amd64
root=UUID=4101c568-e073-42f4-8dfe-08373fceb13c ro
quiet</span></div>
<div><span>[ 33.144] xorg-server 2:21.1.7-3 (<a
class="moz-txt-link-freetext" target="_blank"
rel="noreferrer nofollow noopener"
href="https://www.debian.org/support"
moz-do-not-send="true">https://www.debian.org/support</a>)
</span></div>
<div><span>[ 33.144] Current version of pixman: 0.42.2</span></div>
<div><span>[ 33.144] Before reporting problems, check <a
class="moz-txt-link-freetext" target="_blank"
rel="noreferrer nofollow noopener"
href="http://wiki.x.org" moz-do-not-send="true">http://wiki.x.org</a></span></div>
<span> to make sure that you have the latest version.</span></div>
<div style="background-color: rgb(255, 255, 255);"><span><br>
</span></div>
<div style="background-color: rgb(255, 255, 255);"><span>I
searched bug tracker and could only find this issue: <span><a
class="moz-txt-link-freetext" target="_blank"
rel="noreferrer nofollow noopener"
href="https://gitlab.freedesktop.org/xorg/driver/xf86-video-ati/-/issues/180"
moz-do-not-send="true">https://gitlab.freedesktop.org/xorg/driver/xf86-video-ati/-/issues/180</a></span>
but I don't have a black screen. The system is rather
sluggish but seems to work.</span></div>
<div style="background-color: rgb(255, 255, 255);"><span>I
did try older version of Debian back to 10.13. All have
this issue. I'm very sure the eLux which came with the
box did not have this.<br>
</span></div>
<div style="background-color: rgb(255, 255, 255);"><span><br>
</span></div>
<div style="background-color: rgb(255, 255, 255);"><span>Any
hints where to go next is very <span>appreciated</span>.</span></div>
<div style="background-color: rgb(255, 255, 255);"><span><br>
</span></div>
<span style="background-color: rgb(255, 255, 255);"><span>Best
Regards</span></span>
<div
style="font-family: Arial, sans-serif; font-size: 14px;"
class="protonmail_signature_block"> </div>
</blockquote>
<br>
</blockquote>
<br>
</div>
</blockquote>
<br>
</body>
</html>