Bug 261440
| Summary: | vt newcons breaks suspend/resume for all graphics cards that do not use KMS drivers | ||
|---|---|---|---|
| Product: | Base System | Reporter: | Stefan B. <sblachmann> |
| Component: | kern | Assignee: | freebsd-bugs (Nobody) <bugs> |
| Status: | New --- | ||
| Severity: | Affects Many People | CC: | emaste, grahamperrin |
| Priority: | --- | Keywords: | vt |
| Version: | Unspecified | ||
| Hardware: | Any | ||
| OS: | Any | ||
|
Description
Stefan B.
2022-01-24 19:20:41 UTC
Just to mention, vt newcons fails no matter whether vesa.ko is loaded or not. Setting hw.vga.textmode="YES" or not also has no effect. OS used for testing is 13.0-RELEASE-p6 on amd64. Note that vesa.ko has no relationship with vt, so it does not matter if it is loaded or not Well, when using vt in text mode, it does make a difference if vesa.ko is loaded or not. Because, if it is loaded, then the console display reappears showing the shutdown proceeding when pressing the power button after resuming with a dark screen instead of X Window. If it is not loaded, there is no displaying of the shutdown proceeds, screen stays dark until power off. If vt is in graphics mode, screen always stays dark until power off, then there is no difference of having vesa loaded or not. (Just tested this using SiS and S3 graphics cards. More tests with other graphics cards follow.) BTW, for sc, vesa has to be loaded for successful resume, *except* for Nvidia cards. As said, I'll be doing more tests later to understand the patterns, will report back. (In reply to Stefan B. from comment #3) Any update? (In reply to Ed Maste from comment #4) Not yet, had no time yet, but testing again with 14 and current is on my todo list, as I have read some reports that suspend/resume now works with Nvidia+vt. As I would like to drop SC from my scripts to not have to maintain different handling for two consoles SC (for Nvidia+Matrox) and vt (Intel, AMD). So I need to also test whether the proposed fix for Matrox (see https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=270509#c29) works, and whether it is already committed/incorporated, as Matrox is a very common onboard graphics chip for Supermicro and other server boards. I will set up a dedicated test PC for these graphics card tests and report back. (In reply to Stefan B. from comment #0) > vt newcons breaks resuming from S3 suspend for all graphics cards > except those being supported by i915kms, radeonkms and amdgpu. I use NVIDIA without breakage. KMS. nvidia-modeset loaded, nvidia-driver-470. Sorry, ignore that. I didn't see KMS in the title bar. (In reply to Graham Perrin from comment #7) Some FreeBSD dev decided to make some changes which render the Freedesktop drivers dysfunctional. Then there was pressure from nv users to incorporate a patch that at least restored the functionality of the nv driver. Patches supplied to restore functionality of the Matrox driver were ignored. The FreeBSD dev said that Matrox is obsolete crap not worth supporting. After being taught that Matrox is _the_ server graphics chip, second following the AST (aspeed) chip, and wide expression of bewilderment that the FreeBSD (supposedly targeted at servers) devs don't know this, they hurried to restore the functionality of both these drivers, too. But only one and a half year after disabling these drivers... And then the FreeBSD dev apparently announced to abandon X Window support in favor of Wayland. This was one of the last straws that made me retire FreeBSD. Anyway, back to the topic: I was referring to the Freedesktop graphics drivers, which flawlessly support suspend/resume. But only when using the SC console. With vt newcons, resuming fails. The graphics cards drivers affected are all of xf86-video-* except the KMS-supporting -ati/amdgpu/intel. |
