X11 unusably slow with DRM 5.15 and 6.1's amdgpu on a RX 800 #302
Comments
|
I have the same issue on my system with an RX570 card, even with my environment limited to WindowMaker and XScreenSaver. DRM version is 5.15 here: From From |
|
I've tested leaving my amd machine (RX580 / polaris 11) all night running some vkcube/glxgears and some webgpu stuff running on firefox to see if I could trigger this but no, everything was smooth in the morning and also later during the day. |
|
Hey manu,
On 28 May 2024, at 07:20, Emmanuel Vadot ***@***.***> wrote:
I've tested leaving my amd machine (RX580 / polaris 11) all night running some vkcube/glxgears and some webgpu stuff running on firefox to see if I could trigger this but no, everything was smooth in the morning and also later during the day.
But I was running x11-wm/awesome, I'll add xfce to my local package list and try again with it.
On my side the symptoms are particularly annoying when I switch virtual desktop. (WindowMaker, no compositing, no desktop window except when testing gbsddialog)
I haven’t tried anything 3D but I have many firefox windows and tabs opened, on different virtual desktops, together with thunderbird, gnome-terminal, deforaos-panel, hexchat, and sometimes libreoffice, qemu, or gbsddialog.
Thanks for your help!
Cheers,
--
khorben
|
|
Yet another person encountering freezing with DRM 5.15 on 14p6. DRM 5.10 works flawlessly. I have 4 monitors and 40 virtual desktops using awesome. Many instances of Chrome (different profiles), remmina RDP to various Windows systems, and other programs. I can do some testing but since this is my daily driver at work results might take a bit to report. EDIT: forgot to include system specs Radeon Pro WX5100, 8GB (5820T) (Polaris 10) |
|
I see this on -CURRENT with both Green Sardine 5600G and Polaris 20 RX 580, single DP-1 3840x2160 output and x11-wm/sway after several poudriere tmpfs runs (fragmented VM?). Delay is significantly worse when all cores are used (not the case on 5.10) and/or larger surfaces are being damaged, and cpuset -c -l 0-(ncores-2) during poudriere run also helps (no difference on 5.10), which suggests that additional work is done $somewhere past 5.10. I only glanced and didn't bisect/investigate, but I speculate past 302b3a8 (see ("ttm_pool.c: use_dma_alloc not implemented") cache-coherent physically contiguous DMA allocations like on Linux are not done which results in additional copying/VM work which makes lag more apparent on larger resolutions. |
|
Thanks for some confirmation. I forgot to include that my 4 monitors are running at 2560x1440.
Message ID: ***@***.***>
|
|
Reproducer (VM behavior not deterministic so could not come up with automated reproducer):
|
|
Found a duplicate FreeBSD PR: PR 277476. As I've just added there, I'd really like to get to the bottom of this. However, I don't plan to have the time to do so before end of September at the very least. |
Title: X11 unusably slow with DRM 5.15 and 6.1's amdgpu on a RX 800
Description and reproduction
With DRM 5.15 running with an AMD RX 800 card, after a few minutes to hours in a X11 session, just clicking on a program in the task bar to switch to it or using Alt-Tab can freeze the whole display during seconds. Generally speaking, any kind of desktop effect (such as application thumbnail display when hovering over the task bar) is slow. As the uptime progresses, freezes tend to last longer (I've measured that a few of them lasted for almost 10 minutes).
DRM 6.1 has the same problem but in a slightly lighter form: It takes more uptime for the problem to start manifesting, and freezes are initially shorter. However, they increase over time to the point that the desktop eventually becomes almost unusable, as for 5.15.
DRM 5.10 works correctly. This problem also doesn't show up on some laptop using Intel Gen10 integrated graphics (driver i915) with DRM 5.15 and DRM 6.1 (although another problem shows up, to be reported separately).
Tested mostly with KDE/KWin, but Xfce has similar problems. Turning off composition in KWin essentially doesn't solve the problem (barely makes an almost imperceptible difference).
System Information
FreeBSD version
FreeBSD 14.1-STABLE #0 n267671-9a8a26aefb36: Mon May 13 13:39:56 CEST 2024 MYCONFIG 1401500 1401500
Kernel MYCONFIG is a stripped-down version of GENERIC close to MINIMAL.
Same problem on an older version:
FreeBSD 14.0-STABLE #1 n266865-245844372d7e: Thu Feb 22 11:11:45 CET 2024
PCI Info
vgapci0@pci0:8:0:0: class=0x030000 rev=0xe7 hdr=0x00 vendor=0x1002 device=0x67df subvendor=0x1043 subdevice=0x0525
vendor = 'Advanced Micro Devices, Inc. [AMD/ATI]'
device = 'Ellesmere [Radeon RX 470/480/570/570X/580/580X/590]'
class = display
subclass = VGA
DRM KMOD version
Problem reproduced with:
drm-515-kmod 5.15.118_4
drm-61-kmod 6.1.69_2
No problem with:
drm-510-kmod 5.10.163_9
Preliminary Investigation
Before clear, long freezes, it is common to observe, after some uptime, Xorg
using ~5% CPU for several seconds or more. Some captured kernel stacks:
The frequency of the stacks with
ttm_pool_free()at the bottom increases as small glitches and freezes appear (not reproduced below). Other stacks that were captured much more rarely:Finally, outputting kernel stack traces every 0.1s during freezes seems to indicate that the process is stuck in (or repeatedly calling):
This stack also appears, although less frequently (it's in fact a sub-stack of the previous):
(Part of the time spent on this report was sponsored by the FreeBSD Foundation.)
The text was updated successfully, but these errors were encountered: