Summary: | vesa.ko: Invalid BIOS call when resuming from S3 suspend/sleep causes nvidia driver hang | ||||||
---|---|---|---|---|---|---|---|
Product: | Base System | Reporter: | Stefan B. <sblachmann> | ||||
Component: | kern | Assignee: | freebsd-bugs (Nobody) <bugs> | ||||
Status: | In Progress --- | ||||||
Severity: | Affects Many People | CC: | emaste, grahamperrin, jkim, linimon, pi, sblachmann | ||||
Priority: | --- | ||||||
Version: | 12.2-RELEASE | ||||||
Hardware: | amd64 | ||||||
OS: | Any | ||||||
See Also: | https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=224069 | ||||||
Attachments: |
|
Description
Stefan B.
2021-02-21 00:00:48 UTC
Created attachment 223218 [details]
Disable POST and save/restore states on NVIDIA cards
Please try the attached patch. I cannot test it because I don't use syscons(4) any more.
Thank you very much, Jung-uk! I am going to test the patch on my computers using different nvidia cards/drivers. To make sure it works reliably, I want to test long enough, to accumulate sufficient uptime and suspend/resume cycles. So it might take a few days until I report back. Sorry for late update. I was busy with other things. Tested the patch. It does not work. Still hangs in text mode when resuming. Not sure for which exact cause yet. Apparently I didn't express myself clearly enough; it is *only* the LOAD_STATE call which breaks resume and needs to be omitted in case of Nvidia card/chip. I verified this is still valid by commenting out the x86bios_intr() call in the case STATE_LOAD: of vesa_bios_save_restore(). So I believe the other VESA calls, including POST, do *not* have a negative impact on suspend/resume. I didn't test yet whether vesa_find_pci_device() actually finds the card which responds to the VESA BIOS call (but will do soon using some debug printfs). So I can not rule out yet that a problem there could be the potential cause for the patch not working. Another issue I am not yet clear of whether it matters: There are some OEMs who had in some cases their onboard video BIOS at other locations than C000. I remember some cases I personally encountered, where video BIOS was at E000. For this reason I am not really sure whether the approach of checking for a C000 BIOS start address is 100% safe. I am now thinking about scanning the OEM string which gets returned by function 4F00 for "nvidia" (case-independent), eg the string the VESA 1.2 OEMStringPtr points to. This approach would be independent of the Option BIOS memory address. I have about ten different Nvidia cards and onboard chips, NV4 and higher, and will read out their OEMString via debug printfs, to find out whether this alternative approach could be viable. As I am currently moving my hardware lab, it will take about 1-3 weeks until I report back, maybe with an updated patch. Since VESA is not supported by the default vt(4) console anyway I propose removing it from GENERIC: https://reviews.freebsd.org/D33141 sc(4) users who want to use VESA modes can still load it as a module. A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/src/commit/?id=b8cf1c5c30a5e6da4e2c9702ffd607a90453fb33 commit b8cf1c5c30a5e6da4e2c9702ffd607a90453fb33 Author: Ed Maste <emaste@FreeBSD.org> AuthorDate: 2021-11-27 20:27:45 +0000 Commit: Ed Maste <emaste@FreeBSD.org> CommitDate: 2021-11-28 16:29:17 +0000 Remove options VESA from x86 GENERIC options VESA / vesa.ko provides VESA Bios Extensions (VBE) support for the legacy sc(4) console. It is not used by the default console, vt(4). There is a report[1] of an incompatibility between VESA and the Nvidia driver breaking suspend/resume. Since VESA is not used by the default configuration anyway, just remove options VESA from GENERIC. The kernel module is still available and may be loaded by sc(4) users who want to select a VBE mode. (Note that vt(4) does not support selecting a VBE mode. The loader can set a VBE mode and vt(4) will use it via the vt_vbefb driver.) [1] https://lists.freebsd.org/archives/freebsd-hackers/2021-November/000469.html PR: 253733 Reported by: Stefan Blachmann [1] Reviewed by: imp, manu, tsoome Relnotes: Yes Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D33141 sys/amd64/conf/GENERIC | 1 - sys/i386/conf/GENERIC | 1 - 2 files changed, 2 deletions(-) A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/src/commit/?id=777526ed83822e1af2b7f7ea4186dbf7d3d3d60a commit 777526ed83822e1af2b7f7ea4186dbf7d3d3d60a Author: Ed Maste <emaste@FreeBSD.org> AuthorDate: 2021-11-28 19:10:28 +0000 Commit: Ed Maste <emaste@FreeBSD.org> CommitDate: 2021-11-28 19:37:46 +0000 Remove options VESA from x86 MINIMAL Followup to b8cf1c5c30a5, remove from MINIMAL in addition to GENERIC. options VESA / vesa.ko provides VESA Bios Extensions (VBE) support for the legacy sc(4) console. It is not used by the default console, vt(4). PR: 253733 Fixes: b8cf1c5c30a5 ("Remove options VESA from x86 GENERIC") Relnotes: Yes Sponsored by: The FreeBSD Foundation sys/amd64/conf/MINIMAL | 1 - sys/i386/conf/MINIMAL | 1 - 2 files changed, 2 deletions(-) Hmm, just noticed that the commits refer to x86?!? Does this mean i386 and? or? amd64? My systems are all amd64, so I can only say for amd64 that making suspend/resume work by removing vesa.ko from the kernel. I haven't done any tests on i386 yet. (In reply to Stefan B. from comment #7) Both i386 and amd64. ^Triage: assign to committer. To committer: does this need MFC to 13? Unassign: I did not commit a fix for this issue, just removed VESA from 14 and later. The removal won't be merged back to stable/13. Someone will need to investigate the underlying issue and produce a real fix. |