Radeon Firmware Boot Error

Hello,

Ever since updating to LEAP 42.3, I’ve had an error on boot:

    4.572122] [drm] radeon: dpm initialized
    4.572133] radeon 0000:01:00.0: Direct firmware load for radeon/TAHITI_uvd.bin failed with error -2
    4.572136] radeon 0000:01:00.0: radeon_uvd: Can't load firmware "radeon/TAHITI_uvd.bin"
    4.572137] radeon 0000:01:00.0: failed UVD (-2) init.
    4.572146] radeon 0000:01:00.0: Direct firmware load for radeon/TAHITI_vce.bin failed with error -2
    4.572148] radeon 0000:01:00.0: radeon_vce: Can't load firmware "radeon/TAHITI_vce.bin"
    4.572149] radeon 0000:01:00.0: failed VCE (-2) init.

Is there any way to fix this? Maybe a replacement file for the firmware?

Even after yesterday’s kernel update, the error persists. This is using the built-in driver. Previously on 42.2 I was using fglrx but the new amdgpupro driver
does not work at all on my machine. It causes my screen to flash some horizontal lines constantly. The graphics card is an Radeon R7 250.

Thank you.

Hi
Is the kernel-firmware package installed?

Yes. Version says 20170530-9.1 from vendor openSUSE. No other versions are available.

here is my Firmware:
http://download.opensuse.org/repositories/home:/Sauerland/openSUSE_Leap_42.3/noarch/kernel-firmware-20170904-2.1.noarch.rpm

But it is there:

rpm -qlp http://download.opensuse.org/distribution/leap/42.3/repo/oss/suse/noarch/kernel-firmware-20170530-9.1.noarch.rpm | grep -i tahiti
/lib/firmware/radeon/TAHITI_ce.bin
/lib/firmware/radeon/TAHITI_mc.bin
/lib/firmware/radeon/TAHITI_mc2.bin
/lib/firmware/radeon/TAHITI_me.bin
/lib/firmware/radeon/TAHITI_pfp.bin
/lib/firmware/radeon/TAHITI_rlc.bin
/lib/firmware/radeon/TAHITI_smc.bin
/lib/firmware/radeon/TAHITI_uvd.bin
/lib/firmware/radeon/TAHITI_vce.bin
/lib/firmware/radeon/tahiti_ce.bin
/lib/firmware/radeon/tahiti_k_smc.bin
/lib/firmware/radeon/tahiti_mc.bin
/lib/firmware/radeon/tahiti_me.bin
/lib/firmware/radeon/tahiti_pfp.bin
/lib/firmware/radeon/tahiti_rlc.bin
/lib/firmware/radeon/tahiti_smc.bin

So install kernel-firmware once more as root:

zypper in -f kernel-firmware

Hi Sauerland,

Thank you for having the package for me! Unfortunately, it did not change anything. My firmware files were
exactly the same (did a binary diff).

The error must not be in the firmware then but in the loading of it.

Hi
Is the package xf86-video-amdgpu installed? If so look at removing this so just have the xf86-video-radeon installed…

Hi,

Indeed it was installed. I removed it but the error is exactly the same.

Also, maybe this can be of a hint, but I’ve always had the following error when not using the proprietary drivers:

[drm:si_dpm_set_power_state [radeon]] *ERROR* si_restrict_performance_levels_before_switch failed

With LEAP 42.2, using fglrx removed this error. That is why I tried the amdgpu drivers first, which apparently I had left the package you mentioned.

Any ideas?

Thanks again!

Hi
Well it’s a Southern Island card, which only (amdgpu support) appeared in the later kernel (eg it works for Tumbleweed). Hmmm is there a drm kmp installed, this has some kernel backports, wonder if that’s causing a fight…

Also if you run the following is radeon and amdgpu present?


/sbin/lspci -nnk | grep -A3 VGA

You also removed the amdgpu-pro package, repository etc?

Hi,

Yes, I have drm-kmp-default installed. Version is 4.9.33_k4.79_4-5.2

The lspci code returns the following, so I think the answer is yes:

01:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Cape Verde PRO [Radeon HD 7750/8740 / R7 250E] [1002:683f]
	Subsystem: PC Partner Limited / Sapphire Technology Device [174b:a001]
	Kernel driver in use: radeon
	Kernel modules: radeon

And, also yes, there is no more amdgpu package or repo. I have no idea what you mean by Southern Island.

Thanks for taking a look!

Hi
Cape Verde (Engineering Name) is a Southern island card, which is GCN 1.0, so wonder if it’s the drm kmp that’s causing it to try and load the (amdgpu) firmware, you could try removing that and see what happens…

Hi again,

For some reason I suspected this would be risky and I think I was right. For the first time Yast2 hung while trying to remove a package. It hung for hours and I had
to kill it. System seems intact but I was really worried it could leave things in an inconsistent state. There were no dependencies either, so I do not know what is
going on.

On Wed 20 Sep 2017 02:36:01 PM CDT, idanan wrote:

malcolmlewis;2837985 Wrote:
> Hi
> Cape Verde (Engineering Name) is a Southern island card, which is GCN
> 1.0, so wonder if it’s the drm kmp that’s causing it to try and load
> the (amdgpu) firmware, you could try removing that and see what
> happens…

Hi again,

For some reason I suspected this would be risky and I think I was right.
For the first time Yast2 hung while trying to remove a package. It hung
for hours and I had
to kill it. System seems intact but I was really worried it could leave
things in an inconsistent state. There were no dependencies either, so I
do not know what is
going on.

Hi
Easier to do it from the command line with zypper rm … :wink:

Any delay is because it’s rebuilding initrd which again can see with
zypper.


Cheers Malcolm °¿° SUSE Knowledge Partner (Linux Counter #276890)
openSUSE Leap 42.2|GNOME 3.20.2|4.4.79-18.26-default
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below… Thanks!

Any delay is because it’s rebuilding initrd which again can see with
zypper.

Or with the history:

cat /var/log/zypp/history

Lucky it did not go through!!! Zypper shows it was going to remove the Kernel!!!


The following 6 packages are going to be REMOVED:
  kernel-default-4.4.79-19.1 kernel-default-4.4.85-22.1 virtualbox virtualbox-host-kmp-default virtualbox-host-source virtualbox-qt


6 packages to remove.



No idea what to try next.

Please post such Things with the whole Commandline, because we can not see, what Command you run.

Hi
Yes, for sure, since those ones should be removed as the kernel is up to 4.4.87-25 for Leap 42.3…

Sure, will do next time. It was just a zypper rm but now that I have updated to this week’s update, the module uninstalled easily through Yast2.

Unfortunately, the error persists. So, I have still no idea what is trying to do with the firmware files.

Regards

So, for no other reason than just to amuse me, when you see the Grub boot menu, press “e”, scroll down to the line that begins with “linux” or whatever, hit the End key to get to the end of the command line, add this:

radeon.si_support=0 radeon.cik_support=0 amdgpu.si_support=1 amdgpu.cik_support=1

and see what happens.

Hi
Those options are only in the Tumbleweed kernel (4.13.x)… don’t think it was there for the backported stuff (4.9.x)?

Same errors exactly. Maybe it’s not supported as someone else suggested.

Thanks for trying though!