[arch-dev-public] Move the catalyst drivers from extra to AUR/Community
Here it is, Andreas Radke and I have talked about this and have come to a conclusion. The ATI catalyst drivers are in a pitiful state and AMD is doing close to nothing to improve the situation, they just take Linux as a joke. At least, that is the impression one gets when nvidia releases great drivers for Linux and ATI in the other hand is not showing any promises in making their drivers better and well supported. In the other hand the radeonhd driver, which is in extra shows some promise, and so far have been good, they have been able to make a good driver with only some documentation released by AMD, when AMD can't make one while being the makers, go figure out. This situation is worse in the x86_64 system, we have been stuck with catalyst 8.12 while i686 has 9.2 already, because AMD just wanted to break things at will, because their approach is very strange, break some things to improve others. That is why, we hope for a TU who can bring this drivers to community to raise his hand and help us with this task. Please, if you want to give catalyst the love they don't deserve but need, do not hesitate to offer yourself as the new maintainer. In the worse case scenario where no one raises his hand, the package should be moved to AUR. I personally would not want this, so please TUs respond. A TU will be able to do what is required for the best with this drivers, which includes making 9.2 and later work with x86_64. What do the other developers think? Any developer that wants to take care of them? PS: I do use this drivers, so I do care about them, but this is a choice that had to be made. Thanks, Eduardo "kensai" Romero
On Wed, Feb 25, 2009 at 8:05 PM, Eduardo Romero <k3nsai@gmail.com> wrote:
Here it is,
Andreas Radke and I have talked about this and have come to a conclusion. The ATI catalyst drivers are in a pitiful state and AMD is doing close to nothing to improve the situation, they just take Linux as a joke. At least, that is the impression one gets when nvidia releases great drivers for Linux and ATI in the other hand is not showing any promises in making their drivers better and well supported.
In the other hand the radeonhd driver, which is in extra shows some promise, and so far have been good, they have been able to make a good driver with only some documentation released by AMD, when AMD can't make one while being the makers, go figure out.
This situation is worse in the x86_64 system, we have been stuck with catalyst 8.12 while i686 has 9.2 already, because AMD just wanted to break things at will, because their approach is very strange, break some things to improve others.
That is why, we hope for a TU who can bring this drivers to community to raise his hand and help us with this task. Please, if you want to give catalyst the love they don't deserve but need, do not hesitate to offer yourself as the new maintainer. In the worse case scenario where no one raises his hand, the package should be moved to AUR. I personally would not want this, so please TUs respond. A TU will be able to do what is required for the best with this drivers, which includes making 9.2 and later work with x86_64.
What do the other developers think? Any developer that wants to take care of them?
PS: I do use this drivers, so I do care about them, but this is a choice that had to be made.
Hmm. My thinkpad has an ATI card in it, and I stopped using the catalyst drivers a LONG time ago (a year? I know Travis was still packaging them when I stopped). So, I guess my actual usage follows this pattern, so maybe it's the right way to go. Still, I wonder if this is going to anger a lot of people. Probably, but if they're using catalyst right now they're probably angry at it already :) I'm all for removing shitty packages, but I fear that moving it to community would be simply moving the shit around instead of dealing with it. Perhaps, at this juncture, it would make more sense in the AUR (we'd have to kindly ask the TUs not to package it though).
On Thu, 2009-02-26 at 11:01 -0600, Aaron Griffin wrote:
Hmm. My thinkpad has an ATI card in it, and I stopped using the catalyst drivers a LONG time ago (a year? I know Travis was still packaging them when I stopped).
So, I guess my actual usage follows this pattern, so maybe it's the right way to go.
Still, I wonder if this is going to anger a lot of people. Probably, but if they're using catalyst right now they're probably angry at it already :)
I'm all for removing shitty packages, but I fear that moving it to community would be simply moving the shit around instead of dealing with it. Perhaps, at this juncture, it would make more sense in the AUR (we'd have to kindly ask the TUs not to package it though).
My reason for it to go to community is so a TU can dedicate attention to it, Andreas does not want to waste time on fixing this drivers as they are of no particular interest to him. Maybe a TU in community can apply patches to make the x86_64 9.2 release to work. But well, I'm fine with moving them to AUR, there are some distributions that do not provide catalyst at all in any form, comes to mind paldo GNU/Linux.
On Thu, Feb 26, 2009 at 11:17 AM, Eduardo Romero <k3nsai@gmail.com> wrote:
On Thu, 2009-02-26 at 11:01 -0600, Aaron Griffin wrote:
Hmm. My thinkpad has an ATI card in it, and I stopped using the catalyst drivers a LONG time ago (a year? I know Travis was still packaging them when I stopped).
So, I guess my actual usage follows this pattern, so maybe it's the right way to go.
Still, I wonder if this is going to anger a lot of people. Probably, but if they're using catalyst right now they're probably angry at it already :)
I'm all for removing shitty packages, but I fear that moving it to community would be simply moving the shit around instead of dealing with it. Perhaps, at this juncture, it would make more sense in the AUR (we'd have to kindly ask the TUs not to package it though).
My reason for it to go to community is so a TU can dedicate attention to it, Andreas does not want to waste time on fixing this drivers as they are of no particular interest to him. Maybe a TU in community can apply patches to make the x86_64 9.2 release to work. But well, I'm fine with moving them to AUR, there are some distributions that do not provide catalyst at all in any form, comes to mind paldo GNU/Linux.
Well, I guess what I was saying is: *can* we even fix it? I mean, if they release a broken driver, we can't exactly patch some closed source blob
On Thu, 2009-02-26 at 11:38 -0600, Aaron Griffin wrote:
Well, I guess what I was saying is: *can* we even fix it? I mean, if they release a broken driver, we can't exactly patch some closed source blob
I don't think we can fix most problems, but even if we were able to, licensing does not permit us. One example is the use of /usr/lib64 which is plain ugly. We can use sed on the binary files to load it from /usr/lib, but that violates the license because we're not allowed to distribute altered binaries. Whatever happens, I'm preparing xorg-server 1.6 this week for testing. I don't think AMD supports server-1.6 yet, but Nvidia does, so I see no reason to stick this release in testing as long as we did with server-1.5.
On Thu, 2009-02-26 at 18:57 +0100, Jan de Groot wrote:
On Thu, 2009-02-26 at 11:38 -0600, Aaron Griffin wrote:
Well, I guess what I was saying is: *can* we even fix it? I mean, if they release a broken driver, we can't exactly patch some closed source blob
I don't think we can fix most problems, but even if we were able to, licensing does not permit us. One example is the use of /usr/lib64 which is plain ugly. We can use sed on the binary files to load it from /usr/lib, but that violates the license because we're not allowed to distribute altered binaries.
Whatever happens, I'm preparing xorg-server 1.6 this week for testing. I don't think AMD supports server-1.6 yet, but Nvidia does, so I see no reason to stick this release in testing as long as we did with server-1.5.
Aaron, what you say is true, I thought they could fix it, but as Jan pointed out, is not legal to alter a blob. Didn't thought it that way so Aaron you are right about moving this to AUR instead. Jan, No catalyst does not support xorg-server-1.6 yet, another example on how bad ATIs support is. I do vote for you not to hold xorg-server that long waiting for them.
This is what I've just sent to the closed AMD beta project list.
I am actively trying to get the right development team lined up.
Regards,
Matthew
On 7-Feb-09, at 12:03 PM, Andreas Radke <a.radke@arcor.de> wrote:
No hint? Any other place where we can get support?
-Andy
I try to follow the beta release announcements but reading things like this make me angry wasting more time on this driver: "...Required packaging changes There are no driver changes that require updates to packaging scripts. Deadline for packaging contributions Release: NA Deadline: NA Next planned beta release on this release stream TBD" No communication or public documentation where things will move at all happened. Beside ArchLinux all 64bit non-multilib LinuxFromScratch users are affected and probably more distributions. We are also just packaging Xorg 1.6 about we heard simply nothing here. So we are about to remove support for closed catalyst drivers: http://www.archlinux.org/pipermail/arch-dev-public/2009-February/010394.html -Andy
Yeah, I tried to find a solution of how to work this out and keep catalyst at least in community, but it seems, ATI/AMD just aren't doing a thing to improve the situation. Yet, this package is so important for some users of Arch Linux that a TU should be the one in charge of them. - Cheers Eduardo "kensai" Romero
This discussion have got some press coverage by phoronix: http://www.phoronix.com/scan.php?page=news_item&px=NzEwMg And by distrowatch: http://distrowatch.com/weekly.php?issue=20090302#news Well, some developers did not liked the press coverage we got from phoronix, I have to say, I did liked it, they made it look more dramatic than it was but we got outstanding support from the community. There were close to no one against our decision. And yes the problem in part was the poor quality of this drivers, so why try to deviate the attention from it? All this have lead me to a conclusion, which was later reaffirmed by Pierrre publishing the percentage of use for the graphics driver in Arch Linux. Those results put catalyst below the open source driver xf86-video-ati. Having seen all that, and the positive reaction of the community in the places where this was covered (phoronix), having the blessing of Aaron Griffin, and all other developers. I think it would be for the best to drop support from this drivers and concentrate our efforts in the drivers that really work, and are not but a pain to the developers. I hereby grant the community the privilege to upload this drivers to AUR/Community and take good care of them. Once they are uploaded on AUR/Community I will proceed to delete them from extra and make a front page news announcement. No other objection? -Thanks Eduardo "kensai" Romero
2009/3/2 Eduardo Romero <k3nsai@gmail.com>:
This discussion have got some press coverage by phoronix: http://www.phoronix.com/scan.php?page=news_item&px=NzEwMg And by distrowatch: http://distrowatch.com/weekly.php?issue=20090302#news
Well, some developers did not liked the press coverage we got from phoronix, I have to say, I did liked it, they made it look more dramatic than it was but we got outstanding support from the community
ROFLMAO. It just goes to show you have to be more careful what you say lest you be quoted out of context! ;-) Dusty
On Monday 02 March 2009 10:30:04 am Dusty Phillips wrote:
2009/3/2 Eduardo Romero <k3nsai@gmail.com>:
This discussion have got some press coverage by phoronix: http://www.phoronix.com/scan.php?page=news_item&px=NzEwMg And by distrowatch: http://distrowatch.com/weekly.php?issue=20090302#news
Well, some developers did not liked the press coverage we got from phoronix, I have to say, I did liked it, they made it look more dramatic than it was but we got outstanding support from the community
ROFLMAO. It just goes to show you have to be more careful what you say lest you be quoted out of context! ;-)
Dusty Yeah, you know how the press is they just quote the bad things you say, and leave out whatever good things you said about the subject. That happens when one gets famous. :D
Eduardo Romero schrieb:
This discussion have got some press coverage by phoronix: http://www.phoronix.com/scan.php?page=news_item&px=NzEwMg And by distrowatch: http://distrowatch.com/weekly.php?issue=20090302#news
Well, some developers did not liked the press coverage we got from phoronix, I have to say, I did liked it, they made it look more dramatic than it was but we got outstanding support from the community. There were close to no one against our decision. And yes the problem in part was the poor quality of this drivers, so why try to deviate the attention from it?
I think they missed the point of what we consider "bad quality". The article on phoronix talks about driver quality in terms of features (like OpenGL extensions and such), while we talk about quality in terms of compatibility and packagibility (I want the trademark on that word!). ATI hardcodes paths that should be configurable (and actually are configurable in Xorg - and other drivers respect those paths) and does not ensure compatibility to current Xorg and Linux developments (apparently, only to current Ubuntu developments). That is what we dislike about it, not some nice features it has that you COULD use if the driver WOULD work. The amount of time and uglification we need to make it work in our distribution is just too much and keeps holding us back.
On Mon, Mar 2, 2009 at 9:07 AM, Eduardo Romero <k3nsai@gmail.com> wrote:
I hereby grant the community the privilege to upload this drivers to AUR/Community and take good care of them. Once they are uploaded on AUR/Community I will proceed to delete them from extra and make a front page news announcement.
No other objection?
And hot on the heels of this decision - AMD's dropping support for anything older than a Radeon HD 2000: http://www.phoronix.com/scan.php?page=article&item=amd_r500_legacy&num=1
On Thursday 05 March 2009 03:57:19 pm Travis Willard wrote:
And hot on the heels of this decision - AMD's dropping support for anything older than a Radeon HD 2000:
http://www.phoronix.com/scan.php?page=article&item=amd_r500_legacy&num=1 And even hotter in the heels: ATI/AMT digs its own grave: 9.3 won't have xorg-server 1.6 support, so we will have to wait until 9.4 in April, maybe that will be middle to late april, so I say, as soon as someone takes this to AUR or Community this should be removed without a pity, they just demonstrate little seriousness to Linux and its support. http://www.phoronix.com/scan.php?page=news_item&px=NzExOQ Wow, I'm starting to appreciate more and more my Nvidia card on my desktop, I never have to pray for them to work after an upgrade, for ATI I have to make a whole group of prayers for the sake of seeing the catalyst driver boot up X fine after an upgrade.
On Thu, 2009-03-05 at 22:38 -0400, Eduardo Romero wrote:
And even hotter in the heels: ATI/AMT digs its own grave: 9.3 won't have xorg-server 1.6 support, so we will have to wait until 9.4 in April, maybe that will be middle to late april, so I say, as soon as someone takes this to AUR or Community this should be removed without a pity, they just demonstrate little seriousness to Linux and its support.
AMD is digging catalysts grave, not their own. Don't forget that AMD has released specifications of all chips that are no longer supported by their catalyst. Compare this to nvidia and then look at the development speed of nouveau compared to xf86-video-ati. At least we won't have users crying for Catalyst drivers in archlinux anymore, as it's impossible to get these things working on any modern linux distribution.
On Friday 06 March 2009 04:38:53 am Jan de Groot wrote:
AMD is digging catalysts grave, not their own. Don't forget that AMD has released specifications of all chips that are no longer supported by their catalyst.
Yeah sometimes I need to look at my phrasing, in fact I love AMD I have never own another make of CPU, but well, is the catalyst driver to be blamed, so sorry for the phrasing.
It happened again: Ubuntu released a not yet published driver to their unstable repos to prepare the 9.04 isos. See http://www.phoronix.com/scan.php?page=news_item&px=NzE0Nw The driver were not even published to the AMD beta program. Not one word on the list about any progress with Xorg 1.6. That makes sense how they treat the community beside the holy Canonical. Well. Our community picked it already up and pushed a package to AUR and I think this time it's not a license violation because Ubuntu will have received a special ok to publish this driver. So now I'm going to remove the drivers from extra very soon. Any objections? I'm also going to ask to remove my account from the "beta program". It's a pity that our community will loose this last chance to catch the non existent support. If you want fun, read http://www.phoronix.com/scan.php?page=article&item=amd_beta_declassified&num=1 Read the 2nd page and think about "what the program is not"... I'm done with this closed crap now. -Andy
On Wed, Mar 18, 2009 at 13:54, Andreas Radke <a.radke@arcor.de> wrote:
It happened again: Ubuntu released a not yet published driver to their unstable repos to prepare the 9.04 isos. See
http://www.phoronix.com/scan.php?page=news_item&px=NzE0Nw
The driver were not even published to the AMD beta program. Not one word on the list about any progress with Xorg 1.6. That makes sense how they treat the community beside the holy Canonical.
Well. Our community picked it already up and pushed a package to AUR and I think this time it's not a license violation because Ubuntu will have received a special ok to publish this driver.
So now I'm going to remove the drivers from extra very soon. Any objections?
I'm also going to ask to remove my account from the "beta program". It's a pity that our community will loose this last chance to catch the non existent support.
If you want fun, read http://www.phoronix.com/scan.php?page=article&item=amd_beta_declassified&num=1
Read the 2nd page and think about "what the program is not"...
I'm done with this closed crap now.
-Andy
I can just see the headlines now, "Arch Linux calls ATI crap, drops support for graphics"
On Wed, 2009-03-18 at 13:58 -0400, Daenyth Blank wrote:
I can just see the headlines now, "Arch Linux calls ATI crap, drops support for graphics"
What support for graphics? I thought the latest ATI drivers dropped support for a lot of graphics ;)
Andreas Radke I am with you, ATI just doesn't care about Linux nor its users. I say, just remove them officially and leave those on AUR as the ones everyone will use. Don't remove them in any way, just if they break the system.
Just to let you know, I have tested the 9.4 version, it gives me 300 FPS more in fgl_glgears, and it now works withouth the ugly symlink. So yeah, this driver is going to be good. On Wed, Mar 18, 2009 at 8:14 PM, Eduardo Romero <k3nsai@gmail.com> wrote:
Andreas Radke I am with you, ATI just doesn't care about Linux nor its users. I say, just remove them officially and leave those on AUR as the ones everyone will use. Don't remove them in any way, just if they break the system.
Catalyst drivers are now part of the ArchLinux history :D -Andy
Am Donnerstag, 19. März 2009 11:42:49 schrieb Andreas Radke:
Catalyst drivers are now part of the ArchLinux history :D
So, we are ready for xorg-server 1.6 now? -- Pierre Schmitz Clemens-August-Straße 76 53115 Bonn Telefon 0228 9716608 Mobil 0160 95269831 Jabber pierre@jabber.archlinux.de WWW http://www.archlinux.de
participants (9)
-
Aaron Griffin
-
Andreas Radke
-
Daenyth Blank
-
Dusty Phillips
-
Eduardo Romero
-
Jan de Groot
-
Pierre Schmitz
-
Thomas Bächler
-
Travis Willard