[SOLVED] Linux 3.10/3.11 incompatibility: In function ‘nv_i2c_del_adapter’ error
  1 / 2    
Currently all NVIDIA drivers are incompatible with Linux kernel 3.10. [code]/usr/src/nvidia-313.30/nv-i2c.c: In function ‘nv_i2c_del_adapter’: /usr/src/nvidia-313.30/nv-i2c.c:327:14: error: void value not ignored as it ought to be make[2]: *** [/usr/src/nvidia-313.30/nv-i2c.o] Error 1 make[1]: *** [_module_/usr/src/nvidia-313.30] Error 2 make[1]: Leaving directory `/usr/src/kernels/linux-3.10' NVIDIA: left KBUILD. nvidia.ko failed to build! make: *** [module] Error 1[/code] Please, update them. I'm talking about: [b]313.30 319.32 325.08 beta[/b] LTS releases: [b]304.88 173.14.37[/b] 96.43.23 & 71.86.15 (not really necessary - the supported hardware is pretty ancient) [b]Update: drivers 325.15 are fully compatible with kernel 3.10. There's a little incompatibility with kernel 3.11, which can be fixed by applying [url=https://devtalk.nvidia.com/default/topic/549532/linux/linux-3-10-incompatibility-in-function-lsquo-nv_i2c_del_adapter-rsquo-error-void-value-not/post/3884890/#3884890]this patch[/url].[/b]
Currently all NVIDIA drivers are incompatible with Linux kernel 3.10.

/usr/src/nvidia-313.30/nv-i2c.c: In function ‘nv_i2c_del_adapter’:
/usr/src/nvidia-313.30/nv-i2c.c:327:14: error: void value not ignored as it ought to be
make[2]: *** [/usr/src/nvidia-313.30/nv-i2c.o] Error 1
make[1]: *** [_module_/usr/src/nvidia-313.30] Error 2
make[1]: Leaving directory `/usr/src/kernels/linux-3.10'
NVIDIA: left KBUILD.
nvidia.ko failed to build!
make: *** [module] Error 1


Please, update them.

I'm talking about:

313.30
319.32
325.08 beta


LTS releases:
304.88
173.14.37

96.43.23 & 71.86.15 (not really necessary - the supported hardware is pretty ancient)

Update: drivers 325.15 are fully compatible with kernel 3.10.

There's a little incompatibility with kernel 3.11, which can be fixed by applying this patch.

#1
Posted 07/04/2013 12:10 PM   
I Agree
I Agree

#2
Posted 07/05/2013 03:58 PM   
Really, and something wrong about procfs
Really, and something wrong about procfs

#3
Posted 07/06/2013 12:20 PM   
My GCC didn't go as far as to whine about procfs ;-)
My GCC didn't go as far as to whine about procfs ;-)

#4
Posted 07/06/2013 01:06 PM   
The patch below installs cleanly up to 319.32. slipstream:/usr/src/NVIDIA-Linux-x86_64-325.08 # patch -p1 < pastie-7942599.diff patching file kernel/nv-i2c.c Hunk #1 succeeded at 243 with fuzz 2 (offset -68 lines). Hunk #2 succeeded at 247 (offset -75 lines). patching file kernel/nv-procfs.c Hunk #4 succeeded at 161 with fuzz 1. Hunk #9 FAILED at 424. Hunk #10 succeeded at 546 (offset -1 lines). Hunk #11 succeeded at 575 (offset -1 lines). Hunk #12 succeeded at 600 (offset -1 lines). Hunk #13 succeeded at 617 (offset -1 lines). Hunk #14 succeeded at 629 (offset -1 lines). Hunk #15 succeeded at 637 (offset -1 lines). 1 out of 15 hunks FAILED -- saving rejects to file kernel/nv-procfs.c.rej
The patch below installs cleanly up to 319.32.
slipstream:/usr/src/NVIDIA-Linux-x86_64-325.08 # patch -p1 < pastie-7942599.diff
patching file kernel/nv-i2c.c
Hunk #1 succeeded at 243 with fuzz 2 (offset -68 lines).
Hunk #2 succeeded at 247 (offset -75 lines).
patching file kernel/nv-procfs.c
Hunk #4 succeeded at 161 with fuzz 1.
Hunk #9 FAILED at 424.
Hunk #10 succeeded at 546 (offset -1 lines).
Hunk #11 succeeded at 575 (offset -1 lines).
Hunk #12 succeeded at 600 (offset -1 lines).
Hunk #13 succeeded at 617 (offset -1 lines).
Hunk #14 succeeded at 629 (offset -1 lines).
Hunk #15 succeeded at 637 (offset -1 lines).
1 out of 15 hunks FAILED -- saving rejects to file kernel/nv-procfs.c.rej
Attachments

nv-procfs.c.rej.gz

#5
Posted 07/07/2013 10:10 PM   
I played a bit around with the patches. The nv_i2c file is easy to correct. But the proc* file is tricky. I can just say, that with 3.10 stable kernel release compiling even with patch is not possible with 325 and 319 driver series.
I played a bit around with the patches.
The nv_i2c file is easy to correct. But the proc* file is tricky. I can just say, that with 3.10 stable kernel release compiling even with patch is not possible with 325 and 319 driver series.

#6
Posted 07/08/2013 04:28 PM   
The patch works with 319.32. tindog:~ # uname -r 3.10.0-smp+ tindog:~ # cat /proc/driver/nvidia/version NVRM version: NVIDIA UNIX x86_64 Kernel Module 319.32 Wed Jun 19 15:51:20 PDT 2013 GCC version: gcc version 4.7.3 (SUSE Linux)
The patch works with 319.32.
tindog:~ # uname -r
3.10.0-smp+
tindog:~ # cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module 319.32 Wed Jun 19 15:51:20 PDT 2013
GCC version: gcc version 4.7.3 (SUSE Linux)

#7
Posted 07/08/2013 06:58 PM   
Here's the patch I use with 325.08 and linux-3.10-git: http://pastie.org/8123499
Here's the patch I use with 325.08 and linux-3.10-git:


http://pastie.org/8123499

#8
Posted 07/09/2013 06:01 AM   
I think this patch could compile, but I am not sure what deletion of NV_CREATE_PROC_FILE will cause?! Is the driver still stable and working?
I think this patch could compile, but I am not sure what deletion of NV_CREATE_PROC_FILE will cause?!
Is the driver still stable and working?

#9
Posted 07/09/2013 07:37 AM   
I just built Nvidia 325.08 against linux kernel 3.10.0-1-06005-gd2b4a64-dirty using mlau's patch. It seems to work fine.
I just built Nvidia 325.08 against linux kernel 3.10.0-1-06005-gd2b4a64-dirty using mlau's patch. It seems to work fine.

#10
Posted 07/09/2013 01:29 PM   
dgrat, Aside from /proc/driver/nvidia/warnings/* NVIDIA drivers don't put any crucial information into /proc so I guess it's not really that important.
dgrat,

Aside from /proc/driver/nvidia/warnings/* NVIDIA drivers don't put any crucial information into /proc so I guess it's not really that important.

#11
Posted 07/09/2013 06:39 PM   
The following patch provides a simple fix: [url]http://pastebin.com/N0a5KMZa[/url] It disables the proc fs, but it seems not to be needed anyway (nvidia-settings works fine). I tested it with the nvidia version 295.20.
The following patch provides a simple fix:
http://pastebin.com/N0a5KMZa
It disables the proc fs, but it seems not to be needed anyway (nvidia-settings works fine). I tested it with the nvidia version 295.20.

#12
Posted 07/10/2013 05:43 PM   
I have it working with 304.88 (current Debian-testing), after hacking siddly's patch a little (removed more references to NV_CREATE_PROC_FILE). Seems to work ok (no compile errors, graphics look ok, no bad log entries). Patch is at [url]http://pastebin.com/3nstAMLE[/url]. sean64's patch of 7/10 might have worked (it's at least very simple, as he says), but patch reports it as incomplete ("stops in the middle of a line"). Don't know why.
I have it working with 304.88 (current Debian-testing), after hacking siddly's patch a little (removed more references to NV_CREATE_PROC_FILE). Seems to work ok (no compile errors, graphics look ok, no bad log entries).

Patch is at http://pastebin.com/3nstAMLE.

sean64's patch of 7/10 might have worked (it's at least very simple, as he says), but patch reports it as incomplete ("stops in the middle of a line"). Don't know why.

#13
Posted 07/13/2013 07:03 PM   
I made a couple of modifications to the above patches in order to really clean up the /proc/driver/nvidia directory, which remained on the way after rmmoding the nvidia module, thus causing oopses at the next modprobe when the module tried to create again its proc directory. The patch, for v319.32, is available at [url]http://pastebin.com/JDpkR3kt[/url].
I made a couple of modifications to the above patches in order to really clean up the /proc/driver/nvidia directory, which remained on the way after rmmoding the nvidia module, thus causing oopses at the next modprobe when the module tried to create again its proc directory.

The patch, for v319.32, is available at http://pastebin.com/JDpkR3kt.

#14
Posted 07/19/2013 02:22 PM   
Hey, NVIDIA, when will you fix this incompatibility? 3.10 was released two months ago and people cannot use your drivers normally.
Hey, NVIDIA, when will you fix this incompatibility?

3.10 was released two months ago and people cannot use your drivers normally.

#15
Posted 07/30/2013 04:54 AM   
  1 / 2    
Scroll To Top