32-bit graphics stack update?

[pq5190362]

PS:

I was suspicious of the address, but given who David Zeuthen and the lack of finding an official source I'll go ahead and update from the address provided.

Quote from X.Org Wiki:

http://www.x.org/wiki/radeonBuildHowTo/

[...]

Missing Extra Firmware

All radeon gfxcards r600+ (r600 and higher) require extra firmware (ucode) files [1] to work properly with acceleration (Thanks agd5f for clarification on IRC). According to license issues [2] and the fact that no new firmware files will be shipped in upstream kernels, you need to activate the following kernel-config parameters:

Code: [Select]

CONFIG_FIRMWARE_IN_KERNEL=y
CONFIG_EXTRA_FIRMWARE_DIR="/lib/firmware"
CONFIG_EXTRA_FIRMWARE="radeon/CEDAR_me.bin radeon/CEDAR_pfp.bin radeon/CEDAR_rlc.bin radeon/CYPRESS_me.bin radeon/CYPRESS_pfp.bin radeon/CYPRESS_rlc.bin radeon/JUNIPER_me.bin radeon/JUNIPER_pfp.bin radeon/JUNIPER_rlc.bin radeon/R600_rlc.bin radeon/R700_rlc.bin radeon/REDWOOD_me.bin radeon/REDWOOD_pfp.bin radeon/REDWOOD_rlc.bin"
You can omit those firmware files for which you do not actually have hardware. Copy *.bin to /lib/firmware/radeon directory.

UPDATE: Extra ucode files are now in linux-firmware GIT repository [3] (Thanks airlied for information on IRC).

[1] http://people.freedesktop.org/~agd5f/radeon_ucode/

[2] http://people.freedesktop.org/~agd5f/radeon_ucode/LICENSE.radeon

[3] See commit d9076a54d74e371a11e1206b4a26e2e428045b9e "radeon: add RLC firmwares from AMD."

[...]

[nick65go]

For me, the new firmware-radeon.tcz has no negative/positive difference. Because the only file I need from tcz is R520_cp.bin which is the same size 2K and same calendar date 2014-01-02. The Xorg "works" with or without the binary blob.

Without firmware, "dmesg" shows:
[drm] Loading R500 Microcode
radeon 0000:01:00.0: Direct firmware load failed with error -2
radeon 0000:01:00.0: Falling back to user helper
radeon_cp: Failed to load firmware "radeon/R520_cp.bin"
[drm:r100_cp_init] *ERROR* Failed to load firmware!
radeon 0000:01:00.0: failed initializing CP (-12).
radeon 0000:01:00.0: Disabling GPU acceleration
[drm] radeon: cp finalized

and from "/var/log/Xorg0.lst":
[    13.149] (--) RADEON(0): Chipset: "ATI Mobility Radeon X1400" (ChipID = 0x7145)
[    13.149] (II) RADEON(0): GPU accel disabled or not working, using shadowfb for KMS
...
[    13.889] (WW) RADEON(0): Direct rendering disabled
[    13.889] (II) RADEON(0): Acceleration disabled
...
[    13.893] (II) AIGLX: Screen 0 is not DRI2 capable
[    13.893] (EE) AIGLX: reverting to software rendering
[    13.919] (EE) AIGLX error: dlopen of /usr/local/lib/dri/swrast_dri.so failed (libelf.so.1: cannot open shared object file: No such file or directory)
[    13.919] (EE) GLX: could not load software renderer

=====After firmware-radeon.tcz listed in /tce/onboot.lst and reboot=====
[    13.259] (--) RADEON(0): Chipset: "ATI Mobility Radeon X1400" (ChipID = 0x7145)
[    13.259] (II) Module "dri2" already built-in
[    14.563] (II) RADEON(0): Render acceleration enabled for R300/R400/R500 type cards.
...
[    14.036] (EE) AIGLX error: dlopen of /usr/local/lib/dri/r300_dri.so failed (libelf.so.1: cannot open shared object file: No such file or directory)
[    14.036] (EE) AIGLX: reverting to software rendering
[    14.036] (EE) AIGLX error: dlopen of /usr/local/lib/dri/swrast_dri.so failed (libelf.so.1: cannot open shared object file: No such file or directory)
[    14.036] (EE) GLX: could not load software renderer
[    14.036] (II) GLX: no usable GL providers found for screen 0

Hmm, the news are that a dependency is missing! Of course I could load it manually with tce-load -i  elfutils.tcz

[Juanito]

elfutils added to Xorg-7.7-3d deps - thanks

[pq5190362]

Hello,

any update regarding firmware-radeon.tcz?

Regards

[coreplayer2]

Hello,

any update regarding firmware-radeon.tcz?

Regards

I believe an updated extension was posted to the repo long time ago, end of February..

[pq5190362]

Ah, thank you very much !

[pq5190362]

Hello,

any update regarding firmware-radeon.tcz?

Regards

I believe an updated extension was posted to the repo long time ago, end of February..

PS:

firmware-radeon.tcz apparently is not up to date.

The following page:

http://people.freedesktop.org/~agd5f/radeon_ucode/

seems to have been updated in the meantime and therefore has some newer files compared to the current version of firmware-radeon.tcz.

So, could you please update firmware-radeon.tcz again?

Regards

[coreplayer2]

I've updated firmware-radeon with latest updates of 20th April 2015 & 11th May 2015

submitting...

[pq5190362]

I've updated firmware-radeon with latest updates of 20th April 2015 & 11th May 2015

submitting...

Thank you. But I noticed something:

First:

The info file has not been updated:

http://repo.tinycorelinux.net/6.x/x86/tcz/firmware-radeon.tcz.info

Second:

On this page:

http://people.freedesktop.org/~agd5f/radeon_ucode/

there are some subfolders containing files.

But in the extension:

http://repo.tinycorelinux.net/6.x/x86/tcz/firmware-radeon.tcz.info

the files from the subfolders have all been put into:

/usr/local/lib/firmware/radeon/

together with the other files.

So, now there seem to be duplicates of some files in that folder.

For example, there is:

BONAIRE_ce.bin (from 2014-01-02)

and there is:

bonaire_ce.bin (from 2015-05-23)

together in the same folder and so on...

Are you sure it's correct like this?

Regards

[coreplayer2]

I'll look into the discrepancies

I'm not sure why the submitted info file was not the updated

I also see an opportunity to make the extension smaller with the duplicates, the extension is getting quite large now

[coreplayer2]

pq5190362

The info file has not been updated

Actually I'm not sure what transpired with the info file because I'm looking at the new info file yet the old info file was somehow packaged with the updated firmware, will get it right this time  thanks for spotting this

So, now there seem to be duplicates of some files in that folder.
For example, there is:
BONAIRE_ce.bin (from 2014-01-02)
and there is:
bonaire_ce.bin (from 2015-05-23)
together in the same folder and so on...
Are you sure it's correct like this?

yes there are duplicates, just as the upstream source has these duplicates.
The question we should be asking is: Are the older BONAIRE_ce.bin firmware blobs superseded by their lower case counterparts as in bonaire_ce.bin ?
If so, then why not replace the old firmware at source??  Both case versions have been included from source, therefore the driver must be searching for case sensitive firmware.   I'm trying to see the logic..  I am also curious why separate newer firmware by case and sub-directory??  surely separation by name would have been appropriate??

I simply updated those files of matching case with their newer version, otherwise it's entirely possible for an original and two duplicates to exist as in the source

I'll inquire..

[curaga]

Official word from agd5f is to use the radeon subdir in the kernel-firmware git tree. His dir contains many old versions, and linux-firmware is always kept up to date.

[coreplayer2]

Ah ha!

I've been reading up on the differences and it becomes quite complicated since they have attempted to support drivers provided in new kernels, additionally firmware currently supports Radeon and amdgpu hardware.   At least AMD firmware is backwards compatible, apparently..

Thanks curaga, I think that's a good plan after all agd5f is the man on AMD.  I'll grab the current firmware from kernel-firmware git tree and update.. 

Ref:  http://www.phoronix.com/forums/forum/linux-graphics-x-org-drivers/open-source-amd-linux/800259-amd-releases-open-source-vce-1-0-support/page2

[curaga]

Yep, that's the post I meant. Nice googling!

[coreplayer2]

Updated from kernel-firmware git tree clone and submitted.