AMD 8.39.4 Driver Released -- 8.35 no longer does it better

[SparkleX]

The watermark problem. AMD is still figuring out how this can be fixed.

Well, I just installed the 8.39.4 driver on Ubuntu using this script, and it's working without any problems like annoying logos. When I ran the script, it said something like "brutally patched to remove the logo", well anyway - it works, and now even videos are played correctly in Ubuntu without inverted colors and such things

>Screenshot<

Maybe it's a solution for some of you

[sundown]

I just love the drama element that Michael and co. add to their ATi articles.
Always involving a mantion on the Nvidia side.

"I need a saga... what's the saga like..."

[yoshi314]

who knows, maybe somebody slipped a beta version instead of a normal one.

seeing a new fglrx in the middle of the month is really odd =]

[Svartalf]

who knows, maybe somebody slipped a beta version instead of a normal one.

seeing a new fglrx in the middle of the month is really odd =]

I believe I called it: "Brown Paper Bag Version"...

<*sigh*> I'm so glad I'm mostly NVidia or old DRI driver based now. I've only got a smattering of the fglrx driven hardware now, and it's more for testing purposes for LGP work, etc.

Keep in mind- this is really due to their NOT throwing much of any resources at this up to this point. As it stands, they've got open req's for about a half dozen or more developers for the Shanghai, Toronto, Orlando AND Marlborough offices for Linux developers for OpenGL ES 1.1/2.X and OpenGL 2.X driver developers. Not enough, in comparison, to what they really need, but it IS a start in the right direction...IF they can find people to fill the slots. I know I'm not going to be interested for some time to come (As I'm QUITE busy with other consulting jobs and I have little desire to relo to any of those locations save maybe Orlando...)

[sundown]

Besides we never had a ".4" release before nor anything less than 4. If that could be an indicatior.

[r1348]

Well now.... and who says it's working on FC7?
i get exact same behavior
....
ABI class: X.Org XInput driver, version 0.6
(II) Primary Device is: PCI 01:00:0
(II) ATI Proprietary Linux Driver Version Identifier:8.39.4
(II) ATI Proprietary Linux Driver Release Identifier: UNSUPPORTED-8.393.1
(II) ATI Proprietary Linux Driver Build Date: Jul 18 2007 14:41:23
....
Backtrace:
0: X(xf86SigHandler+0x6d) [0x48b99d]
1: /lib64/libc.so.6 [0x2aaaab60e630]
2: /usr/lib64/xorg/modules/drivers//fglrx_drv.so(atiddxSave64BitBAR+0x2d) [0x2aaaacd4cb8d]
3: /usr/lib64/xorg/modules/drivers//fglrx_drv.so(atiddxProbeMain+0x37f) [0x2aaaacd57a7f]
4: X(InitOutput+0x6e7) [0x463707]
5: X(main+0x275) [0x434625]
6: /lib64/libc.so.6(__libc_start_main+0xf4) [0x2aaaab5fbab4]
7: X(FontFileCompleteXLFD+0x229) [0x433ad9]

Fatal server error:
Caught signal 11. Server aborting


anyone can help?
ARCH: x86_64
kernel: 2.6.21

I've too the same issues, I think we can state it officially: on Fedora 7, fgrlx 8.39.4 still segfaults on x86_64.

How sad...

[d2kx]

yoshi, I guess AMD uses the installer and their supported systems so they didn't see this mistake in the testing. If all goes well, every driver is releeased at the middle of the month.

sundown, when there's not much to test, the build number is lower. That's it (oh well, betatesting the 8.25.18 must have been fun).

ATI Proprietary Linux Driver Build Date: Jul 18 2007 14:41:23

That is... new.

[[email protected]]

Update:

Works also with Kernel "2.6.23-0.35.rc0.git6.fc8" here on Fedora Rawhide

[gfxdrone]

well, any doubt about this fake monthly release schedule is removed. The lack of testing of course was already obvious, this simply confirms that their methods are flawed. But this type of error is really funny, almost comical.

At some point hopefully they will realize that this nonsense of scheduled monthly releases is just an invitation to further error and inefficient use of resources, and go back to a more sane model, like: release a beta, get feedback, fix problems, then release final when it's solid. I realize this is a radical notion in software development, but why not give it a try, be on the cutting edge of software engineering, pioneer new and untested methods such as the above.... obviously someone in charge at ATI Linux fglrx driver development must be the very epitomy of the pointy headed boss making bad decisions for bad reasons, Dilbert style.

Anyway, I definitely appreciate the tone of these reviews and articles, at least now it's clear right away that problems exist, so I have to give phoronix credit for that, I get the news I need in the way I need it, so thanks.

[aljo666]

same for me: on Fedora 7, fgrlx 8.39.4 still segfaults on x86_64