[elrepo] how to get the amd driver working on RHEL 7 ?

Fri Apr 1 17:13:06 EDT 2016

On 04/01/2016 07:43 PM, jacques klein wrote:
> Ok, did it, took some time ...
>
> http://paste.fedoraproject.org/348323/59528854/ -> 
> http://paste.fedoraproject.org/348323/59528854
>
> What I noticed, is that the radeon module is still loaded ....
>
1. First of all, as far as I notice, you are still using kernel 
3.10.0-229.el7.x86_64, i.e. RHEL ( or a clone of  RHEL ) 7.1, with no 
updates. This is a very bad idea, you miss 8 months of security updates. 
And sometimes (it's not the case now, though ) the packages from elrepo 
are built against and require the kernel release available in the most 
recent version of RHEL

.
>
> 1.
>     # yum search fglrx
> 2.
>     Loaded plugins: fastestmirror, langpacks
> 3.
>     elrepo                                                         |
>     2.9 kB  00:00:00
> 4.
>     elrepo/primary_db                                                
>      | 308 kB  00:00:01
> 5.
>     Loading mirror speeds from cached hostfile
> 6.
>      * base: mirror.euserv.net
> 7.
>      * elrepo: ftp.colocall.net
> 8.
>      * epel: ftp.linux.org.tr
> 9.
>      * extras: centos.schlundtech.de
>10.
>      * updates: mirror.softaculous.com
>11.
>     Warning: No matches found for: fglrx
>12.
>     No matches found
>13.
>14.
>     ===> strange, saw in a mailing list that the fglrx stuf has been
>     moved from elrepo-testing to elrepo
>15.
>16.
>     # yum --enablerepo=elrepo-testing search fglrx
>
So far I have pushed the fglrx  related packages (version 15.9, 15.11 
and 15.12) only via my personal website or ( more recent) via 
elrepo-testing because I am still hunting problems which can be fixed 
via packaging approaches.  So far this was a good idea, with the help of 
several testers who reported problems via the list of the bug tracker we 
isolated ( and fixed ) a couple of issues.

In the output of your lspci we can see:
>
> 1.
>     01:00.0 VGA compatible controller: Advanced Micro Devices, Inc.
>     [AMD/ATI] Saturn XT [FirePro M6100]
>
At https://support.amd.com/en-us/download/desktop?os=Linux+x86_64 => 
Supported Products I do not see your chipset listed as such. I suspect 
that the closest hit is "AMD Radeon™ HD 6000M Series" and we can only 
hope that M6100 really fits in there

>
> 1.
>     # rpm -qa '*fglrx*'
> 2.
>     kmod-fglrx-15.12-3.el7.elrepo.x86_64
> 3.
>     fglrx-x11-drv-15.12-3.el7.elrepo.x86_64
> 4.
>

Good....

> 1.
>     # lsmod | grep fglrx
> 2.
>     fglrx         15517027  0
> 3.
>     amd_iommu_v2           19054  1 fglrx
>
Good....
>
> 1.
>     # lsmod | grep radeon
> 2.
>     radeon           1421005  1
> 3.
>     i2c_algo_bit           13413  1 radeon
> 4.
>     drm_kms_helper         98226  1 radeon
> 5.
>     ttm            93488  1 radeon
> 6.
>     drm           311588  4 ttm,drm_kms_helper,radeon
> 7.
>     i2c_core             40325  7
>     drm,i2c_i801,i2c_hid,drm_kms_helper,i2c_algo_bit,radeon,videodev
> 8.
> 9.
>     ===> strange, radeon seems to be loaded,
>10.
>     ===> and there is no /etc/modprobe.d/blacklist-fglrx.conf file
>     like after amd's installation.
>
We blacklist the video driver via the kernel command line but .... ( see 
below )
>
> 1.
>
>
> 2.
>      ##### /var/log/Xorg.0.log after fglrx install and reboot
> 3.
> 4.
>     X.Org X Server 1.17.2
> 5.
>     Release Date: 2015-06-16
> 6.
>     [  80.251] X Protocol Version 11, Revision 0
> 7.
>     [  80.251] Build Operating System:  2.6.32-220.17.1.el6.x86_64
> 8.
>     [  80.252] Current Operating System: Linux jkn0
>     3.10.0-229.el7.x86_64 #1 SMP Fri Mar 6 11:36:42 UTC 2015 x86_64
> 9.
>     [  80.252] Kernel command line:
>     BOOT_IMAGE=/boot/vmlinuz-0-rescue-b54ecc01e0614077a1954d4047f47202
>     root=/dev/sda3
>
.... this line indicates that you seem to boot in RESCUE mode which does 
not invoke the magic added by the fglrx packages to the regular kernel 
command line.

So, first of all please reboot and start the system using the regular 
menu entry, not in rescue mode. If you still have issues after that, 
please pastebin again /var/log/Xorg.0.log  and also include 
/etc/grub2.cfg and /etc/default/grub

And last but not least, please use the list for communication purposes. 
Others may be interested ( or help ) as well.
Thank you

>
> On Fri, 2016-04-01 at 18:01 +0300, Manuel Wolfshant wrote:
>> On 04/01/2016 05:55 PM, jacques klein wrote:
>> > Finally, I found the elrepo site, and installed kmod-fglrx.x86_64 and
>> > fglrx-x11-drv.x86_64 from elrepo-testing.
>> > The rpms are "better", no "Failed dependencies", but after
>> > installation and reboot, same as with amd's rpm, X server doesn't work.
>> >
>> > What can I do ?, is there a "step-by-step" procedure to debug this
>> > issue ?.
>> once installed and after the system fails to start the X server, push
>> the content of /var/log/Xorg.0.log to a pastebin (fpaste.org for
>> instance --- you could yum install fpaste from EPEL and then just run
>> "fpaste < /var/log/Xorg.0.log" ) and share the URL of the paste over here.
>> please also include the output of the following commands:
>> lspci
>> uname -a
>> rpm -qa \*fglrx\*
>>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.elrepo.org/pipermail/elrepo/attachments/20160402/0cbc7951/attachment.html>