Page 8 of 21
Re: Kernel 4.9
Posted: Thu Nov 24, 2016 10:38 pm
by xeno74
zappa2009 wrote:Sure here it is.
Xorg.logs_Debian9_Lubuntu.zip
Lubuntu: I habe the Xorg.0.log only because it hangs bevor i get a console.
Debian9: I have dmesg output and Xorg.0.log
Its all in the zip attached.
Boot with initrd-4.9-rc6 bootargs /dev/sda4(or sda6) and kernel4.9-amdgpu with R7-250X 1GB DDR5 Club3D
Many thanks. I have found the following graphics card errors:
R7-250X with the AMDGPU driver
dmesg
Code: Select all
[ 1.572925] [drm:.gfx_v6_0_ring_test_ib] *ERROR* amdgpu: ib test failed (scratch(0x2140)=0xCAFEDEAD)
[ 1.573588] [drm:.amdgpu_ib_ring_tests] *ERROR* amdgpu: failed testing IB on GFX ring (-22).
[ 1.574189] [drm:.amdgpu_device_init] *ERROR* ib ring test failed (-22).
Xorg.0.log
Code: Select all
[ 35.130] (II) [KMS] Kernel modesetting enabled.
[ 35.132] (EE) AMDGPU(0): amdgpu_device_initialize failed
[ 35.132] (II) [KMS] Kernel modesetting enabled.
[ 35.133] (EE) AMDGPU(1): amdgpu_device_initialize failed
[ 35.134] (II) [KMS] Kernel modesetting enabled.
[ 35.135] (EE) AMDGPU(G0): amdgpu_device_initialize failed
[ 35.136] (II) [KMS] Kernel modesetting enabled.
[ 35.137] (EE) AMDGPU(2): amdgpu_device_initialize failed
[ 35.137] (II) [KMS] Kernel modesetting enabled.
[ 35.138] (EE) AMDGPU(3): amdgpu_device_initialize failed
Xorg.0.log
Code: Select all
[ 54.152] (EE) RADEON(0): failed to initialise surface manager
[ 54.152] (EE) RADEON(0): radeon_setup_kernel_mem failed
[ 54.152] (EE)
Fatal server error:
[ 54.152] (EE) AddScreen/ScreenInit failed for driver 0
I will discuss these error messages in the A-EON Linux team.
Re: Kernel 4.9
Posted: Fri Nov 25, 2016 2:06 pm
by xeno74
Hi All,
I think the AMDGPU firmwares are little endian, so we'll probably need to byte swap it before interpreting it in the driver.
Byte swap source code for the RADEON firmwares is available in the kernel source code [1] but I haven't found any byte swap code for the AMDGPU firmwares.
I compiled the RC6 without the AMDGPU support today. Maybe the missing byte swap code for AMDGPU firmwares is the problem.
Download:
vmlinux-4.9-without-AMDGPU.tar.gz
Please test this kernel with your SI graphics cards.
Thanks in advance,
Christian
[1]
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/diff/drivers/gpu/drm/radeon/vce_v1_0.c?id=55ce625fc5e3a87e46a14c2cb062a3579f79b7e0
Re: Kernel 4.9
Posted: Fri Nov 25, 2016 4:45 pm
by zappa2009
Yep it boots in the Login and later in the Desktop with Mesa as SoftwareRenderer.
Re: Kernel 4.9
Posted: Fri Nov 25, 2016 4:49 pm
by xeno74
zappa2009 wrote:
Hi
Yep it boots in the Login and later in the Desktop with Mesa as SoftwareRenderer.
Hi zappa2009,
Many many thanks for testing. OK, we know the AMDGPU driver is the problem.
@Srtest
Could you please test the kernel 4.9 without the AMDGPU driver too?
Thanks,
Christian
Re: Kernel 4.9
Posted: Sun Nov 27, 2016 12:20 am
by Srtest
Tested and no.
Re: Kernel 4.9
Posted: Sun Nov 27, 2016 5:38 am
by xeno74
Srtest wrote:Tested and no.
Sorry because of the test result.
Maybe you have more luck with the new stable kernel
4.8.11.
Re: Kernel 4.9
Posted: Mon Nov 28, 2016 9:52 pm
by xeno74
Hi All,
Here is the
RC7 for testing.
New:
Download:
vmlinux-4.9-rc7-AmigaONE_X1000.tar.gz
In this package are two kernels.
vmlinux-4.9-radeon: Kernel 4.9-rc7 with the RADEON driver
vmlinux-4.9-amdgpu: Kernel 4.9-rc7 with the AMDGPU driver
Screenshot of SeaMonkey 2.40 on openSUSE Tumbleweed PPC64 with the kernel 4.9-rc7:
Please test the
RC7.
Thanks,
Christian
Re: Kernel 4.9
Posted: Sat Dec 03, 2016 7:57 am
by xeno74
Hi All,
Tommysammy has successfully tested the RC7 (AMDGPU version) with his Saphire Radeon 7850. Additionally he was able to boot and install the ubuntu MATE 16.04.1 LTS Live DVD with his Saphire Radeon 7850.
Link:
a1k.org
Please test the
RC7.
Have a nice weekend!
Cheers,
Christian
Re: Kernel 4.9
Posted: Sun Dec 04, 2016 4:47 am
by Srtest
Srtest wrote:Srtest wrote:I don't know how to say it one more time: no kernel works for me since 4.7. That darren patch that fixed things for all of you NI guys didn't do so for me with a R9-270 GCN 1.0. It goes to a very early boot stage when it says - "boots something something ........" and then nothing. It is similiar to the way the microcode didn't work a few kernels back until it was reverted.
I have finally made progress and what I mean by progress is getting closer to know what is this situation. Xeno's split kernels between RADEON and AMDGPU provided more information as there is a difference for me between the 2:
When I boot with the "normal" modules of RADEON I get to that early message and it gets stuck. However, when booting with the AMDGPU modules I get past that message yet get a black screen (of death?) that is exactly the same screen and the same behaviour I got last time when the microcodes didn't work for SI users. I guess now it is the same situation that doesn't affect all users of SI (from what I read the users who reported back didn't have a R9 card) or at least most importantly - a graphics card situation. Debian have recently updated their AMD microcode I guess it is time to check a different microcode with my card via the early boot microcode/modules and continue from there. My guess it is exactly the same situation as last time.
I guess I need to quote myself because I got the same result.
Did you made sure you disabled everything amdgpu related in the config for the radeon kernel? because even on 4.7 xorg tells me it is looking for amdgpu in the kernel level and not radeon and if xorg looks for that then the kernel tells it to look for that. Maybe on my card amdgpu always takes precedent. My only chance is if you put my card's firmware as a radeon module and disable everything amdgpu including the basic amdgpu support (not just the experimental SI).
Re: Kernel 4.9
Posted: Sun Dec 04, 2016 11:28 am
by xeno74
Srtest wrote:
I guess I need to quote myself because I got the same result.
Did you made sure you disabled everything amdgpu related in the config for the radeon kernel? because even on 4.7 xorg tells me it is looking for amdgpu in the kernel level and not radeon and if xorg looks for that then the kernel tells it to look for that. Maybe on my card amdgpu always takes precedent. My only chance is if you put my card's firmware as a radeon module and disable everything amdgpu including the basic amdgpu support (not just the experimental SI).
Yes, I disabled all AMDGPU options but the AMDGPU's firmwares are still in the kernel config file.
Code: Select all
amdgpu/carrizo_ce.bin amdgpu/carrizo_me.bin amdgpu/carrizo_mec2.bin amdgpu/carrizo_mec.bin amdgpu/carrizo_pfp.bin amdgpu/carrizo_rlc.bin amdgpu/carrizo_sdma1.bin amdgpu/carrizo_sdma.bin amdgpu/carrizo_uvd.bin amdgpu/carrizo_vce.bin amdgpu/fiji_ce.bin amdgpu/fiji_mc.bin amdgpu/fiji_me.bin amdgpu/fiji_mec2.bin amdgpu/fiji_mec.bin amdgpu/fiji_pfp.bin amdgpu/fiji_rlc.bin amdgpu/fiji_sdma1.bin amdgpu/fiji_sdma.bin amdgpu/fiji_smc.bin amdgpu/fiji_uvd.bin amdgpu/fiji_vce.bin amdgpu/tonga_ce.bin amdgpu/tonga_mc.bin amdgpu/tonga_me.bin amdgpu/tonga_mec2.bin amdgpu/tonga_mec.bin amdgpu/tonga_pfp.bin amdgpu/tonga_rlc.bin amdgpu/tonga_sdma1.bin amdgpu/tonga_sdma.bin amdgpu/tonga_smc.bin amdgpu/tonga_uvd.bin amdgpu/tonga_vce.bin amdgpu/topaz_ce.bin amdgpu/topaz_mc.bin amdgpu/topaz_me.bin amdgpu/topaz_mec2.bin amdgpu/topaz_mec.bin amdgpu/topaz_pfp.bin amdgpu/topaz_rlc.bin amdgpu/topaz_sdma1.bin amdgpu/topaz_sdma.bin amdgpu/topaz_smc.bin
I will delete the AMDGPU firmwares and compile a new RC7.