Bug 118701 - x86_64 GMA500 reports /dev/dri/card0: failed to set DRM interface version 1.4: Inappropriate ioctl for device
Summary: x86_64 GMA500 reports /dev/dri/card0: failed to set DRM interface version 1.4...
Status: RESOLVED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel) (show other bugs)
Hardware: Intel Linux
: P1 normal
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-05-22 10:08 UTC by Stuart Foster
Modified: 2017-02-14 08:34 UTC (History)
1 user (show)

See Also:
Kernel Version: 4.6
Subsystem:
Regression: No
Bisected commit-id:


Attachments
Xorg.0.log, lspci and dmesg outputs. (52.31 KB, application/octet-stream)
2016-05-22 10:08 UTC, Stuart Foster
Details
Add compat ioctl (1.01 KB, patch)
2016-05-24 22:44 UTC, Patrik Jakobsson
Details | Diff

Description Stuart Foster 2016-05-22 10:08:23 UTC
Created attachment 217091 [details]
Xorg.0.log, lspci and dmesg outputs.

MSI U180 atom netbook using gma500 driver works fine if the kernel is build as 32bit however if I build the kernel as 64bit the Xserver fails to start and reports:

[    27.260] (EE) /dev/dri/card0: failed to set DRM interface version 1.4: Inappropriate ioctl for device

Please find attached Xorg.0.log, lspci and dmesg outputs.

Any suggestions as to what is wrong ?
Comment 1 Patrik Jakobsson 2016-05-23 20:07:50 UTC
Hi Stuart

Did you try with a 64-bit userspace? I haven't tried gma500 on a 64-bit system so not sure what the status is. GPU can only address 32-bit so there could be interesting problems.
Comment 2 Stuart Foster 2016-05-23 21:39:00 UTC
(In reply to Patrik Jakobsson from comment #1)
> Hi Stuart
> 
> Did you try with a 64-bit userspace? I haven't tried gma500 on a 64-bit
> system so not sure what the status is. GPU can only address 32-bit so there
> could be interesting problems.

Hi Patrik,

No I have not tried a 64-bit userspace. I only planned on using a 32-bit user space. I use LFS/BLFS for my systems so I could try and build a cut down 64-bit user space if you think it will help. These are not the quickest machines so it may take me some time.
Comment 3 Patrik Jakobsson 2016-05-24 22:44:16 UTC
Created attachment 217321 [details]
Add compat ioctl

Turns out we didn't have compat_ioctl hooked up. Does the patch I attached fix your problem?
Comment 4 Stuart Foster 2016-05-25 07:40:15 UTC
(In reply to Patrik Jakobsson from comment #3)
> Created attachment 217321 [details]
> Add compat ioctl
> 
> Turns out we didn't have compat_ioctl hooked up. Does the patch I attached
> fix your problem?

Hi Patrik,

Just tested your patch and I can report that it solved my problem the xserver starts correctly.

Thank you.
Comment 5 Patrik Jakobsson 2016-05-25 21:36:55 UTC
Thanks for testing, queued for next and stable
Comment 6 Stuart Foster 2016-08-08 07:28:29 UTC
Hi Patrik,

I have checked both linux-4.7 and linux-4.8-rc1 and this fix has not made it through, when do you think it will be included ?

Thanks.
Comment 7 Patrik Jakobsson 2017-02-13 13:45:25 UTC
Hi Stuart,

The fix is available now in most stable kernels. Can you confirm and close the bug if everything is fine?
Comment 8 Stuart Foster 2017-02-14 08:34:31 UTC
Hi Patrik,

Yes everything is ok now.

Thanks.

Note You need to log in before you can comment on or make changes to this bug.