Bug 27703

Summary: Update request to nvidia-current nonfree package with kernel 5.9.1x
Product: Mageia Reporter: Xavier BERTAUX <bertauxx>
Component: RPM PackagesAssignee: Kernel and Drivers maintainers <kernel>
Status: RESOLVED OLD QA Contact:
Severity: normal    
Priority: Normal CC: davidwhodgins, ghibomgx, ouaurelien, sysadmin-bugs
Version: 7   
Target Milestone: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Source RPM: nvidia-current-430.64-11.mga7.nonfree.src.rpm CVE:
Status comment: nvidia-current-450.57-3.mga7.nonfree.src.rpm in backports

Description Xavier BERTAUX 2020-12-02 13:50:49 CET
Description of problem:
after update on kernel 5.9.11 or  5.9.12 dkms nvidia 450.57 don't build

Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.
nvidia-current (450.57-3.mga7.nonfree): Installing module.
..........(bad exit status: 10)
  Build failed.  Installation skipped.
Comment 1 Dave Hodgins 2020-12-02 14:09:06 CET
There was a message from the tmb titled "Headsup: kernel 5.9 series lands in mga7 updates_testing" to the dev and qa mailing lists warning that nvidia,
virtualbox, and some other packages that require dkms are not ready, but the
kernel itself is ready for early testing, on systems that do not require
those modules.

Those fixes are being worked on.

CC: (none) => davidwhodgins

Comment 2 Aurelien Oudelet 2020-12-02 17:43:39 CET
Hi, thanks for reporting.
I added the committers in CC.

Since there is not Bug assigned to QA for nvidia-current package update yet, this can't be passed though us. Meanwhile, Kernel 5.9.1x is not ready yet to be pushed.
It will be ASAP when work done for nvidia-current has been done.

(Please set the status to 'assigned' if you are working on it)

CC: (none) => ghibomgx, ouaurelien
Component: Release (media or process) => RPM Packages

Comment 3 Aurelien Oudelet 2020-12-02 17:48:19 CET
This bug was previously opened against nvidia-current-450.57-3.mga7.nonfree.src.rpm in backports.

Assigning this to nvidia-current-430.64-11.mga7.nonfree.src.rpm in updates.

Summary: with last kernel 5.9.11 & 5.9.12 dkms nvidia 450 don't build => Update request to nvidia-current nonfree package with kernel 5.9.1x
Status comment: (none) => nvidia-current-450.57-3.mga7.nonfree.src.rpm in backports
Source RPM: (none) => nvidia-current-430.64-11.mga7.nonfree.src.rpm

Aurelien Oudelet 2020-12-02 17:48:36 CET

Assignee: bugsquad => kernel

Comment 4 Giuseppe Ghibò 2020-12-02 18:26:13 CET
For mga7 there is nvidia-current is 430.64-11 in updates and 450.57-11 in backports_testing (the sources for 450.57 for mga7 are in in svn:
updates/7/nvidia-current/branches/450). None of them however would work anymore in kernel 5.9.

For kernel 5.9 we have to distinguish module building from module working. Including the kernel extra patches in nvidia packages for kernel 5.9 would let module building flawlessly, but it's not said it would work, due to the GPL-condom changes in kernel 5.9 series, that the public patches are not yet able to fix.

In particular only official nvidia series working with kernel 5.9, is 455.45.01. The other older series, were not yet fixed upstream. AFAIK this is the series that nvidia worked on for kernel 5.9, but AFAIK even that still has problems with CUDA (for uvm IIRC) and the kernel-5.9 condom.

A workaround that we had included in kernel 5.9.x package previously was to revert that kernel restriction, reverting the condom patch to the previous state. This had worked, but then it was preferred to revert to the standard kernel upstream state, because reverting the patch there was giving us the feeling of being "unfair". To avoid misunderstanding an alternative solution, could be to include the old reverting patch in a special kernel (enclosed in a conditional build in the spec file), to be uploaded maybe only in "tainted/updates" section. The only problem doing so, beyond the extra work to upload everytime two kernels in two different places, it's that the patch applies only to kernel 5.9, and not anymore to the upcoming kernel 5.10 LTS, so it would have anyway a pretty short life.

A third alternative solution to provide for mga7, IMHO not much feasible, is to use kernel 5.8 instead of 5.9. kernel 5.8 hasn't the GPL condom patch, however it reached EOL just after 5.8.16. To have it up-to-date for security fixes, we could follow the Ubuntu kernel for which the 5.8 is the basis for their 20.10 distro. This is however a long job, because if you cherrypick their patchset, it's not said that patches would apply flawlessly to our older kernel 5.8, and as long as the kernel 5.8 diverge, backporting patches would become harder and harder.
Comment 5 Aurelien Oudelet 2021-07-06 13:17:57 CEST
Mageia 7 is EOL since July 1st 2021.
There will not have any further bugfix for this release.

You are encouraged to upgrade to Mageia 8 as soon as possible.

@reporter, if this bug still apply with Mageia 8, please let us know it.

@packager, if you work on the Mageia 7 version of your package, please check the Mageia 8 package if issue is also present. In this case, please fix the Mageia 8 version instead.

This bug report will be closed OLD if there is no further notice within 1st September 2021.
Comment 6 Marja Van Waes 2021-09-07 14:08:58 CEST
Hi bug reporter and hi assignee and others involved,

Please reopen this bug report if it is still valid for Mageia 8 or 9(cauldron), and change "Version:" in the upper left of this report accordingly.

This report is being closed as OLD because it was filed against Mageia 7, for which  support ended on June 30th 2021.

Thanks,
Marja

Status: NEW => RESOLVED
Resolution: (none) => OLD