Bug#1066080: nvidia-driver (525.147.05-10) does not build against kernel 6.6.15-amd64 on Debian Sid

2024-03-12 Thread JON Tauri
Replying to myself. Your statement about the ongoing 64 bit time_t
transition got me thinking. What if I have a problem there? So, I tried
installing the libelf t64 bit again:



$ sudo aptitude -f install libelf1t64
The following NEW packages will be installed:
 libelf1t64
The following packages will be REMOVED:
 libelf1{a} libgphoto2-l10n{u}
The following partially installed packages will be configured:
 nvidia-driver nvidia-kernel-dkms
0 packages upgraded, 1 newly installed, 2 to remove and 267 not upgraded.
Need to get 176 kB of archives. After unpacking 2,769 kB will be freed.
The following packages have unmet dependencies:
libdebuginfod1 : Depends: libelf1 (= 0.190-1+b1) but it is not going to be
installed
libdw1 : Depends: libelf1 (= 0.190-1+b1) but it is not going to be
installed
The following actions will resolve these dependencies:

Keep the following packages at their current version:
1) libelf1 [0.190-1+b1 (now, unstable)]
2) libelf1t64 [Not Installed]



Accept this solution? [Y/n/q/?] n
The following actions will resolve these dependencies:

Remove the following packages:
1) libcurl3-gnutls [8.6.0-3 (now, unstable)]
2) libdebuginfod1 [0.190-1+b1 (now, unstable)]
3) libdw1 [0.190-1+b1 (now, unstable)]
4) libnettle8 [3.9.1-2+b1 (now, unstable)]

Install the following packages:
5) libcurl3t64-gnutls [8.6.0-3.2 (unstable)]
6) libdebuginfod1t64 [0.190-1.1+b1 (unstable)]
7) libdw1t64 [0.190-1.1+b1 (unstable)]
8) libnettle8t64 [3.9.1-2.2 (unstable)]

Rather neat, isn't it? Each library being replaced with the presumably
updated version. So, I said yes on a hunch.

Accept this solution? [Y/n/q/?] y
The following NEW packages will be installed:
 libcurl3t64-gnutls{a} libdebuginfod1t64{a} libdw1t64{a} libelf1t64
libnettle8t64{a}
The following packages will be REMOVED:
 libcurl3-gnutls{a} libdebuginfod1{a} libdw1{a} libelf1{a}
libgphoto2-l10n{u} libnettle8{a}
The following partially installed packages will be configured:
 nvidia-driver nvidia-kernel-dkms
0 packages upgraded, 5 newly installed, 6 to remove and 267 not upgraded.
Need to get 1,165 kB of archives. After unpacking 2,767 kB will be freed.
Do you want to continue? [Y/n/?] y
Get: 1 http://deb.debian.org/debian sid/main amd64 libnettle8t64 amd64
3.9.1-2.2 [296 kB]
Get: 2 http://deb.debian.org/debian sid/main amd64 libcurl3t64-gnutls amd64
8.6.0-3.2 [421 kB]
Get: 3 http://deb.debian.org/debian sid/main amd64 libdw1t64 amd64
0.190-1.1+b1 [243 kB]
Get: 4 http://deb.debian.org/debian sid/main amd64 libelf1t64 amd64
0.190-1.1+b1 [176 kB]
Get: 5 http://deb.debian.org/debian sid/main amd64 libdebuginfod1t64 amd64
0.190-1.1+b1 [28.4 kB]
Fetched 1,165 kB in 1s (963 kB/s)
Retrieving bug reports... Done
Parsing Found/Fixed information... Done
dpkg: libnettle8:amd64: dependency problems, but removing anyway as you
requested:
wget depends on libnettle8.
qemu-utils depends on libnettle8.
libsrt1.5-gnutls:amd64 depends on libnettle8.
librtmp1:amd64 depends on libnettle8.
libhogweed6:amd64 depends on libnettle8.
libgnutls30t64:amd64 depends on libnettle8 (>= 3.9~).
libcurl3-gnutls:amd64 depends on libnettle8.
libarchive13:amd64 depends on libnettle8.
gstreamer1.0-plugins-bad:amd64 depends on libnettle8 (>= 3).
dnsmasq-base depends on libnettle8 (>= 2.4-3).

(Reading database ... 415473 files and directories currently installed.)
Removing libnettle8:amd64 (3.9.1-2+b1) ...
Selecting previously unselected package libnettle8t64:amd64.
(Reading database ... 415465 files and directories currently installed.)
Preparing to unpack .../libnettle8t64_3.9.1-2.2_amd64.deb ...
Unpacking libnettle8t64:amd64 (3.9.1-2.2) ...
Setting up libnettle8t64:amd64 (3.9.1-2.2) ...
dpkg: libcurl3-gnutls:amd64: dependency problems, but removing anyway as
you requested:
qemu-block-extra depends on libcurl3-gnutls (>= 7.16.3).
python3-pycurl depends on libcurl3-gnutls (>= 8.6.0).
octave depends on libcurl3-gnutls (>= 7.16.2).
network-manager depends on libcurl3-gnutls (>= 7.24.0).
libxerces-c3.2:amd64 depends on libcurl3-gnutls (>= 7.16.2).
libvirt0:amd64 depends on libcurl3-gnutls (>= 7.28.0).
libsane1:amd64 depends on libcurl3-gnutls (>= 7.16.2).
libraptor2-0:amd64 depends on libcurl3-gnutls (>= 7.16.2).
libqalculate22:amd64 depends on libcurl3-gnutls (>= 7.16.2).
libproxy1v5:amd64 depends on libcurl3-gnutls (>= 7.16.2).
libkolabxml1v5:amd64 depends on libcurl3-gnutls (>= 7.16.2).
libdebuginfod1:amd64 depends on libcurl3-gnutls (>= 7.28.0).
libappstream5:amd64 depends on libcurl3-gnutls (>= 7.63.0).
gstreamer1.0-plugins-bad:amd64 depends on libcurl3-gnutls (>= 7.55.0).
git depends on libcurl3-gnutls (>= 7.56.1).

(Reading database ... 415473 files and directories currently installed.)
Removing libcurl3-gnutls:amd64 (8.6.0-3) ...
Selecting previously unselected package libcurl3t64-gnutls:amd64.
(Reading database ... 415466 files and directories currently installed.)
Preparing to unpack .../libcurl3t64-gnutls_8.6.0-3.2_amd64.deb ...

Bug#1066080: nvidia-driver (525.147.05-10) does not build against kernel 6.6.15-amd64 on Debian Sid

2024-03-12 Thread JON Tauri
Thanks.

I have run updates multiple times a day for the past week hoping what you
are suggesting would transpire, but things are not getting fixed through a
simple update.

Here is the information requested:

$ dpkg -l libelf1 libelf1t64
dpkg-query: no packages found matching libelf1t64
Desired=Unknown/Install/Remove/Purge/Hold
|
Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend

|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name   Version  Architecture Description
+++-==---===

ri  libelf1:amd64  0.190-1+b1   amd64library to read and write ELF
files

An attempt to install the 64 bit version of the library fails:

$ sudo apt-get install libelf1t64
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
plasma-workspace : Depends: gdb-minimal but it is not going to be installed
or
gdb
   Recommends: qml-module-org-kde-pipewire but it is not
going to be installed
E: Error, pkgProblemResolver::Resolve generated breaks, this may be caused
by held packages.

Incidentally, I get the same error when I try to install the kernel headers
for kernel 6.7.7-amd64 (so I am not upgrading the kernel either). I was
trying that as a possible solution hoping that the newer kernel may have
fixed any problems of compatibility between nvidia-kernel and
linux-headers. No joy:

$ sudo apt-get install linux-headers-6.7.7-amd64
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
plasma-workspace : Depends: gdb-minimal but it is not going to be installed
or
gdb
   Recommends: qml-module-org-kde-pipewire but it is not
going to be installed
E: Error, pkgProblemResolver::Resolve generated breaks, this may be caused
by held packages.

The above error has also been persistent for the past 4-5 days.

Use of aptitude -f install for installation of kernel headers suggests that
I uninstall libelf:

The following actions will resolve these dependencies:

Keep the following packages at their current version:
1) linux-headers-6.7.7-amd64 [Not Installed]


Accept this solution? [Y/n/q/?] n
The following actions will resolve these dependencies:

 Remove the following packages:
1)  libcurl3-gnutls [8.6.0-3 (now, unstable)]
2)  libdebuginfod1 [0.190-1+b1 (now, unstable)]
3)  libdw1 [0.190-1+b1 (now, unstable)]
4)  libelf1 [0.190-1+b1 (now, unstable)]
5)  libnettle8 [3.9.1-2+b1 (now, unstable)]

 Install the following packages:
6)  libcurl3t64-gnutls [8.6.0-3.2 (unstable)]
7)  libdebuginfod1t64 [0.190-1.1+b1 (unstable)]
8)  libdw1t64 [0.190-1.1+b1 (unstable)]
9)  libelf1t64 [0.190-1.1+b1 (unstable)]
10) libnettle8t64 [3.9.1-2.2 (unstable)]
11) linux-kbuild-6.7.7 [6.7.7-1 (unstable)]

Accept this solution? [Y/n/q/?]

So, I have a cluster of broken packages that are keeping me from fixing up
the situation through an update. From what you mention, maybe I have a
broken installation, but I have never seen a Debian installation break like
this through regular upgrades (I used to use Testing, but moved to Sid 3
years ago).



On Tue, Mar 12, 2024 at 6:31 PM Andreas Beckmann  wrote:

> That's the actual culprit:
>
> On 12/03/2024 13.33, JON Tauri wrote:
> > # LD [M]  /var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.o
> >ld -m elf_x86_64 -z noexecstack --no-warn-rwx-segments   -r -o
> > /var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.o
> > @/var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.mod  ;
> > ./tools/objtool/objtool --hacks=jump_label --hacks=noinstr
> --hacks=skylake
> > --ibt --orc --retpoline --rethunk --sls --static-call --uaccess
> --prefix=16
> >   --link  --module
> /var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.o
> > ./tools/objtool/objtool: error while loading shared libraries:
> libelf.so.1:
> > cannot open shared object file: No such file or directory
> > make[4]: ***
> > [/usr/src/linux-headers-6.6.15-common/scripts/Makefile.build:443:
> > /var/lib/dkms/nvidia-current/525.147.05/build/nvidia.o] Error 127
> > make[4]: *** Deleting file
> > 

Bug#1066080: nvidia-driver (525.147.05-10) does not build against kernel 6.6.15-amd64 on Debian Sid

2024-03-12 Thread Andreas Beckmann

That's the actual culprit:

On 12/03/2024 13.33, JON Tauri wrote:

# LD [M]  /var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.o
   ld -m elf_x86_64 -z noexecstack --no-warn-rwx-segments   -r -o
/var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.o
@/var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.mod  ;
./tools/objtool/objtool --hacks=jump_label --hacks=noinstr --hacks=skylake
--ibt --orc --retpoline --rethunk --sls --static-call --uaccess --prefix=16
  --link  --module /var/lib/dkms/nvidia-current/525.147.05/build/nvidia-uvm.o
./tools/objtool/objtool: error while loading shared libraries: libelf.so.1:
cannot open shared object file: No such file or directory
make[4]: ***
[/usr/src/linux-headers-6.6.15-common/scripts/Makefile.build:443:
/var/lib/dkms/nvidia-current/525.147.05/build/nvidia.o] Error 127
make[4]: *** Deleting file
'/var/lib/dkms/nvidia-current/525.147.05/build/nvidia.o'
make[4]: *** Waiting for unfinished jobs


This is not a bug in nvidia-kernel-dkms.
I assume this is a temporary breakage due to the ongoing 64-bit time_t 
transition which involves a huge amount of package renames.


I cannot reproduce it in an up-to-date sid chroot with 
linux-headers-6.6.15-amd64 installed (which has been superseded by 6.7.* 
btw).


Updating your system again should probably fix the issue.

In case it persists: Which variant and version of the libelf.so.1 
library do you have installed?


dpkg -l libelf1 libelf1t64


Andreas



Bug#1066080: nvidia-driver (525.147.05-10) does not build against kernel 6.6.15-amd64 on Debian Sid

2024-03-12 Thread Andreas Beckmann

On 12/03/2024 08.53, JON Tauri wrote:

Contents of the make.log:


Please send the complete make.log to the bug. The part you pasted only 
contained the secondary errors.



Andreas



Bug#1066080: nvidia-driver (525.147.05-10) does not build against kernel 6.6.15-amd64 on Debian Sid

2024-03-12 Thread JON Tauri
Package: nvidia-driver
Version: 525.147.05-10

An attempt to upgrade nvidia-driver to current version (have retried this
after a purge remove in an attempt to restart from a clean slate) fails. I
did not have any issues with the previous version of the driver in Sid
(don't know the old version number), so this is not a hardware problem
(lspci output is included) but a driver problem.

However, this is not the first time nvidia-driver has broken with Sid
(which is fine - it is called unstable for a reason), but it has been 5-6
days already since this happened, I am not seeing any movement on the
package tracker. I was hoping that someone had reported this showstopper
already and a fix was on the way.

Details (following https://www.debian.org/Bugs/Reporting suggestions) are
below:

$ sudo apt-get install nvidia-driver firmware-misc-nonfree
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
firmware-misc-nonfree is already the newest version (20230625-2).

...

The following additional packages will be installed:
 curl firmware-nvidia-gsp glx-alternative-mesa glx-alternative-nvidia
glx-diversions libcuda1 libcurl4t64 libegl-nvidia0 libgl1-nvidia-glvnd-glx
libgles-nvidia1 libgles-nvidia2 libgles1 libglx-nvidia0 libnss-mymachines
libnss-systemd libnvcuvid1 libnvidia-allocator1 libnvidia-cfg1
libnvidia-egl-gbm1 libnvidia-egl-wayland1 libnvidia-eglcore
libnvidia-encode1 libnvidia-glcore libnvidia-glvkspirv libnvidia-ml1
libnvidia-ptxjitcompiler1 libnvidia-rtcore libpam-systemd libpsl5t64
libssh2-1t64 libsystemd-shared libsystemd0 nvidia-alternative
nvidia-driver-bin nvidia-driver-libs nvidia-egl-common nvidia-egl-icd
nvidia-installer-cleanup nvidia-kernel-common nvidia-kernel-dkms
nvidia-kernel-support nvidia-legacy-check nvidia-modprobe
nvidia-persistenced nvidia-settings nvidia-smi nvidia-support
nvidia-suspend-common nvidia-vdpau-driver nvidia-vulkan-common
nvidia-vulkan-icd systemd systemd-container systemd-coredump
systemd-timesyncd update-glx xserver-xorg-video-nvidia
Suggested packages:
 nvidia-cuda-mps vulkan-tools systemd-homed systemd-userdbd systemd-boot
systemd-resolved libtss2-mu-4.0.1-0 libtss2-rc0
Recommended packages:
 libcuda1:i386 nvidia-driver-libs:i386
The following packages will be REMOVED:
 libcurl4 libpsl5 libssh2-1
The following NEW packages will be installed:
 firmware-nvidia-gsp glx-alternative-mesa glx-alternative-nvidia
glx-diversions libcuda1 libcurl4t64 libegl-nvidia0 libgl1-nvidia-glvnd-glx
libgles-nvidia1 libgles-nvidia2 libgles1 libglx-nvidia0 libnvcuvid1
libnvidia-allocator1 libnvidia-cfg1 libnvidia-egl-gbm1
libnvidia-egl-wayland1 libnvidia-eglcore libnvidia-encode1 libnvidia-glcore
libnvidia-glvkspirv libnvidia-ml1 libnvidia-ptxjitcompiler1
libnvidia-rtcore libpsl5t64 libssh2-1t64 nvidia-alternative nvidia-driver
nvidia-driver-bin nvidia-driver-libs nvidia-egl-common nvidia-egl-icd
nvidia-installer-cleanup nvidia-kernel-common nvidia-kernel-dkms
nvidia-kernel-support nvidia-legacy-check nvidia-modprobe
nvidia-persistenced nvidia-settings nvidia-smi nvidia-support
nvidia-suspend-common nvidia-vdpau-driver nvidia-vulkan-common
nvidia-vulkan-icd update-glx xserver-xorg-video-nvidia
The following packages will be upgraded:
 curl libnss-mymachines libnss-systemd libpam-systemd libsystemd-shared
libsystemd0 systemd systemd-container systemd-coredump systemd-timesyncd
10 upgraded, 48 newly installed, 3 to remove and 264 not upgraded.

...
...

Setting up nvidia-kernel-dkms (525.147.05-10) ...
Loading new nvidia-current-525.147.05 DKMS files...
Building for 6.6.15-amd64
Building initial module for 6.6.15-amd64
Error! Bad return status for module build on kernel: 6.6.15-amd64 (x86_64)
Consult /var/lib/dkms/nvidia-current/525.147.05/build/make.log for more
information.
dpkg: error processing package nvidia-kernel-dkms (--configure):
installed nvidia-kernel-dkms package post-installation script subprocess
returned error exit status 10
dpkg: dependency problems prevent configuration of nvidia-driver:
nvidia-driver depends on nvidia-kernel-dkms (= 525.147.05-10) |
nvidia-kernel-525.147.05 | nvidia-open-kernel-525.1
47.05; however:
 Package nvidia-kernel-dkms is not configured yet.
 Package nvidia-kernel-525.147.05 is not installed.
 Package nvidia-kernel-dkms which provides nvidia-kernel-525.147.05 is not
configured yet.
 Package nvidia-open-kernel-525.147.05 is not installed.

dpkg: error processing package nvidia-driver (--configure):
dependency problems - leaving unconfigured
...

Contents of the make.log:
DKMS make.log for nvidia-current-525.147.05 for kernel 6.6.15-amd64
(x86_64)
Tue Mar 12 12:26:40 PM IST 2024
make KBUILD_OUTPUT=/lib/modules/6.6.15-amd64/build V=1 -C
/lib/modules/6.6.15-amd64/source M=/var/lib/dkms/nvidia-cu
rrent/525.147.05/build ARCH=x86_64
NV_KERNEL_SOURCES=/lib/modules/6.6.15-amd64/source
NV_KERNEL_OUTPUT=/lib/modules/
6.6.15-amd64/build NV_KERNEL_MODULES="nvidia nvidia-uvm nvidia-modeset
nvidia-drm nvidia-peermem"