A Fix For AMD Catalyst On Ubuntu 15.10 Is Coming

[DanL]

AMD often provides Ubuntu with a special release of fglrx during the release cycle (one that's certified to work with the kernel and X found in that release). That didn't happen this time. Maybe it's because Ubuntu chose a relatively new kernel and/or because AMD's proprietary driver team was focused on the amdgpu transition.

But no, don't you ever change the holy release date, regardless if there are critical bugs.

You don't hold up a release for a proprietary driver. You should pull it from the repo and document it in the release notes though.

[molecule-eye]

Would be to much for the Ubuntu team to fix their issues before releasing, I guess.

Seriously. Look how long Wily has been in testing, and how long the bug report had been filed and under active discussion. It's not like the driver fails to install on some limited set of cases. It fails IN EVERY SINGLE CASE. This has to be one of the worst fuck ups I've seen from Canonical, ever.

[dungeon]

So is this a bug in gcc or in Catalyst?

I think it is kernel bug, there was big fpu rewrite (and on rewrites like that one regressions usually happens, i had one or two fpu crashes even if i use opensource radeon driver), xsave changes well maybe it is even avx instruction related Probably some of those instruction gets borked with gcc5 (and tranistion to gcc5 by itself is also big one)

https://lwn.net/Articles/643235/

[chithanh]

But no, don't you ever change the holy release date, regardless if there are critical bugs.

I think Canonical once changed the release date in 2006.

blame canonical because amd is a nice joke

Actually you can blame Dell. AMD seemed perfectly content with supporting only RedHat and SUSE Enterprise distros. But then Dell came and decided to sell Ubuntu computers, basically grabbed AMD and Canonical by their necks and said "Now kiss!"

But I guess it wasn't as bad for Canonical, they never had a problem with bending over for Poulsbo, Cedar View, etc.

[varikonniemi]

AMD should be ashamed

[dungeon]

AMD should be ashamed

Why AMD? Because kernel 4.1 was marked longterm and Ingo decided that is just ideal time to break things in 4.2

It is cool but i am sorry for Ubuntu and rolling distros because they want latest and greatest, but not every time latest is greatest.

[_ONH_]

Ubuntu always was a fucked Systemen and still is. Since 2006 never had a single Problem with since then after switching to fedora (GNOME) or open suse (KDE).
Ubuntu is still cancer it seems. Never realy tried sinc back then because there is no reason to consider it.

[Marc1n]

Seriously. Look how long Wily has been in testing, and how long the bug report had been filed and under active discussion. It's not like the driver fails to install on some limited set of cases. It fails IN EVERY SINGLE CASE. This has to be one of the worst fuck ups I've seen from Canonical, ever.

True. They also did similar things with nvidia when they decided to use nouveau for both Nvidia GTX9xx series chips and Optimus systems in their recent distros. Effect? System that is unable to display either xorg or console after installation.

[dungeon]

I actually like how no one commented on this news - that news should have most attention then anything else

http://www.phoronix.com/scan.php?pag...ntu-15.10-GCC5

For example and to translate it to human readable words:

The libstdc++6 transition will be a large one, having a large fun/pain factor depending on your point of view. In either case, please help it by addressing the outstanding issues.

Fun for devs, pain for the user - currently it is on 41%...

https://release.debian.org/transitio...ibstdc++6.html

So enjoy or feel a pain on Ubuntu 15.10

[profoundWHALE]

Fun for devs, pain for the user - currently it is on 41%...
https://release.debian.org/transitio...ibstdc++6.html
So enjoy or feel a pain on Ubuntu 15.10

What amazes me is the amount of checkmarks under mips64el. I had to look it up.