[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1029602: vmwgfx: kernel oops when using fbterm in vmware



Continued testing and found that this bug:

- Not reproducible in current Linux 6.2-rcX mainline
- Reproducible in Linux 6.1.7-1 (bookworm kernel package)

The git history of drivers/gpu/drm/vmwgfx shows that the offending function `vmw_fb_dirty_flush()` in file vmwgfx_fb.c has been removed by commit df42523c12f8d58a41f547f471b46deffd18c203. It seems that vmwgfx will use "drm fb helpers" instead of its own fb implementation in Linux 6.2.x so this bug is gone in mainline.

I'm going to send bug report to upstream after a working 6.1.x vanilla kernel compiled and tested, as it seems certain that this issue exists in all stable & LTS kernels before current mainline 6.2-rcX.

On 1/29/23 20:56, Salvatore Bonaccorso wrote:
Hi,

On Sun, Jan 29, 2023 at 04:46:38PM +0800, taoky wrote:
Hi Bonaccorso,

I have tested the vanilla 5.10.165 and the 6.0.12-1 in bullseye-backports,
and they both have this bug.

I'm going to compile a mainline kernel and then report this bug. BTW, I'm
not sure... does it mean to Cc 1029602@bugs.debian.org when reporting to
upstream to "keep in the loop"?

Yes, Cc the Debian downstream bug as well so we keep informed. It's
not a must but helps noticing if there is upstream change/fix for the
issue. Alternatively let us know when the report has happened upstream
and set the bug forwarded to the upstream report.

Regards,
Salvatore


Reply to: