Re: [PATCH xf86-video-amdgpu 00/19] Removing UMS remnants and assorted patches

2018-04-10 Thread Michel Dänzer

Hi Emil,


thanks for the patches and sorry for the late follow-up, I was on
vacation and backlogged.


> Here is some work that I had for ages - was planning to do a similar
> series for the radeon driver. Since ETIME for the latter I'm getting
> this out.

FWIW, it would be nice to bring the radeon probe code more in line with
amdgpu, which should make porting these changes easy.


> This series has two goals, although since the split it a bit hard I've
> left it one big bunch:
>  - remove most of the UMS remnants
>  - simplify the {pci,platform}_probe duplication
> 
> To top it up there's a TODO as 'Patch' 19. It mostly seeks to gather
> feedback on the items listed.

Patches 1, 15 & 16 are

Reviewed-by: Michel Dänzer 


I'm going to follow up with some comments to individual patches. Some of
them should be dropped, the rest modified or rebased.


-- 
Earthling Michel Dänzer   |   http://www.amd.com
Libre software enthusiast | Mesa and X developer
___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx


[PATCH xf86-video-amdgpu 00/19] Removing UMS remnants and assorted patches

2018-04-04 Thread Emil Velikov
Hi all,

Here is some work that I had for ages - was planning to do a similar
series for the radeon driver. Since ETIME for the latter I'm getting
this out.

This series has two goals, although since the split it a bit hard I've
left it one big bunch:
 - remove most of the UMS remnants
 - simplify the {pci,platform}_probe duplication

To top it up there's a TODO as 'Patch' 19. It mostly seeks to gather
feedback on the items listed.

As always, input is greatly appreciated.

Thanks
Emil

Emil Velikov (19):
  Move amdgpu_bus_id/amgpu_kernel_mode within amdgpu_kernel_open_fd
  Guard ODEV_ATTRIB_FD usage with the correct ifdef
  Use ODEV_ATTRIB_PATH where possible for the device node.
  Remove drmCheckModesettingSupported and kernel module loading
  Kill off drmOpen/Close/drmSetInterfaceVersion in favour of drmDevices
  Introduce amdgpu_device_setup helper
  Factor out common code to amdgpu_probe()
  Simplify fd_ref handling in amdgpu_probe()
  Remove error handling on xnfcalloc()
  Remove unneeded xf86GetEntityInfo() call
  Don't leak a AMDGPUEntRec instance if amdgpu_device_setup fails
  Remove ancient "pointer" macro
  configure: check for XORG_DRIVER_CHECK_EXT prior to using it
  Do not export gAMDGPUEntityIndex
  Do not export the DriverRec AMDGPU
  Remove set but unused amdgpu_dri2::pKernelDRMVersion
  Store device_name as AMDGPUEntRec::master_node
  Keep track of who owns the fd in AMDGPUEnt
  TODO

 configure.ac |   6 +
 src/amdgpu_dri2.c|  12 +-
 src/amdgpu_dri2.h|   2 -
 src/amdgpu_dri3.c|   3 +-
 src/amdgpu_glamor.c  |   2 +-
 src/amdgpu_glamor_wrappers.c |   4 +-
 src/amdgpu_kms.c |  25 +---
 src/amdgpu_misc.c|   2 +-
 src/amdgpu_probe.c   | 304 ---
 src/amdgpu_probe.h   |   3 +-
 src/compat-api.h |   4 +-
 src/drmmode_display.c|   2 +-
 todo |   9 ++
 13 files changed, 146 insertions(+), 232 deletions(-)
 create mode 100644 todo

-- 
2.16.0

___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx