Matthew Garrett ([info]mjg59) wrote,
@ 2009-04-17 00:28:00
Previous Entry  Add to memories!  Tell a Friend  Next Entry
Entry tags:advogato, fedora

Intel's commitment to open source driver support
Update: It turns out that the original PSB code is still on the moblin site - I just failed to notice that there's two pages of git repositories. I fail. Last update is from March 2008, though.

Edited again: An up-to-date version of the drm code is available as part of the Moblin kernel SRPM - repo.moblin.com isn't obviously linked off the Moblin website, but I should have thought to find and check their source repository.

By and large, working with Intel is a pleasure. In fact, a while ago I said so in a fairly obvious way. So it's a shame that the support for Poulsbo is still an absolute fucking mess with absolutely no obvious end in sight. Intel's Moblin group (the people who seem to be closest to having responsibility for providing any public Linux support for that hardware, though it's really not clear if they've got any more say in this than I do) have a kernel team that don't want or need a git tree despite having someone who's working on forward porting the driver to modern kernels - probably a job that requires more than one person given what a godawful fucking mess the last public release was, but it'd at least be nice to see what the current state is. Especially since you don't even seem to be able to get the last public release - moblin's git repositories have been rearranged and I can't find the psb-kmod one on their gitweb any more.

The most terribly depressing thing about this is that development is clearly still happening internally. Ubuntu is still being sent tarball releases by Intel, not that the code's getting any better judging by the type of diff involved. This is made even more entertaining by the bug in question being private, leaving no indication whatsoever about what bugs this code is meant to fix or why.

Closed development, random private tarball drops to partners without any public releases or changelogs, no indication of any upstream release, kernel developers entirely unrelated to the development of the driver trying to get code merged so people can actually use the hardware they bought? I'd expect this of some random far-east vendor with no experience in working with Linux, but not a company that's consistently one of the top ten contributors to the kernel and has frequently touted their level of support. There's no meaningful support for Poulsbo. There's just a thin cardboard cutout that's been carefully placed in front of a hole filled with users' hopes and money.

Seriously Intel. Sort it the fuck out.




(26 comments) - (Post a new comment)


(Anonymous)
2009-04-17 12:24 am UTC (link)
Last Message:

The ball is in your court now.

(Reply to this)


[info]vadi.myopenid.com
2009-04-17 02:44 am UTC (link)
Hm, if Intel has as big of an ego as you, I don't see this getting anywhere.

(Reply to this) (Thread)


[info]realmotk
2009-04-17 03:53 am UTC (link)
Do you actually have a point, or are you just trying to start up a tall-poppy pile-on?

(Reply to this) (Parent)(Thread)


(Anonymous)
2009-04-18 04:49 am UTC (link)
I misread that as "tall-puppy pile-on", and then I was sad.

(Reply to this) (Parent)(Thread)


[info]realmotk
2009-04-18 02:31 pm UTC (link)
YAY PUPPIES

(Reply to this) (Parent)


(Anonymous)
2009-04-17 03:59 am UTC (link)
So, how have you engaged the responsible parties at Intel to get this "sorted the fuck out"? What have YOU done to fix a problem that YOU see? Open Source Development isn't just about code. It's about policy and process and procedure. If all you're going to do is bitch about it in a blog, than you're no better than some random luser who bitches that Linux can't hibernate their laptop as well as Windows can.

(Reply to this) (Thread)


[info]mjg59
2009-04-17 04:10 am UTC (link)
I have undergone all the straightforward and normal procedures. I've spoken to a number of people in Intel about this. I've discussed the situation with people who've worked with Imagination. I've even talked to Intel customers who are shipping Poulsbo hardware. They are all aware of the issue but are currently unable to do anything about it, despite several avenues having been explored. And this is where the analogy with a user's technical issues falls down. I can't file a bug against Intel's management policies. I can't volunteer to fix their issues for them. I can't provide useful descriptions of the precise failure mode I see. If there's no direct way for me to interact with the people who set these policies then the only thing I can do is to attempt to influence them indirectly - that is, by increasing the amount of obvious public unhappiness.

But if you've got other suggestions for things I could do that might improve the situation then please do let me know.

(Reply to this) (Parent)(Thread)


[info]realmotk
2009-04-17 04:14 am UTC (link)
I suspect that anonymous thinks you're just sinking the boot into Ubuntu. I sense the gnashing of gums.

(Reply to this) (Parent)


[info]printf.net
2009-04-17 04:12 am UTC (link)
> you're no better than some random luser who bitches that Linux can't hibernate their laptop as well as Windows can.

Amusing, given that mjg59 has done more to make suspend work with laptops under Linux than, ooh, just about anyone else on the planet.

(Reply to this) (Parent)


[info]izard
2009-04-17 04:59 am UTC (link)
Sorry I don't have access to Poulsbo's issues tracker (only working on bigger iron myself). But I've ppl around who do. So if you have a bug description I can ask someone to open a bug.

(Reply to this) (Thread)


[info]mjg59
2009-04-17 05:02 am UTC (link)
I don't think there's really anything more than "No released open-source Poulsbo driver for Linux".

(Reply to this) (Parent)(Thread)


[info]izard
2009-04-17 05:49 am UTC (link)
too bad. Right now I am reading an internal memo that rephrases your post, only related to bigger servers used for web serving...

(Reply to this) (Parent)


(Anonymous)
2009-04-17 05:21 am UTC (link)
So...

I want to make sure that I have this straight. I specifically choose to purchase a Mini-9 instead of a Mini-12 because of the graphics driver.

(and much to my dismay I ended up with a broadcom wifi.. which boggled my mind considering the numerous other vendors that provide good wifi for Linux. Within a day it was replaced with a Atheros and end up with superior results)

So... We have:
1. Open Source DRM stuff trying to make it's way into the kernel.
2. Open Source xf86 driver for 2D
3. Closed source OpenGL driver.

Is this right?

I've been looking forward to Intel's Larrabee, but this sort of thing has me questioning Intel's commitment when it comes to hardware that is actually competitive with the rest of the industry in terms of performance...

(Reply to this) (Thread)


[info]thristian
2009-04-17 06:28 am UTC (link)
To be fair, I believe all the graphics hardware Intel has built themselves work pretty well under Linux, it's just this this one specific chipset licensed from another manufacturer that's causing issues. Larrabee is very definitely an all-Intel effort, I expect it'll work just as well as all of the other non-Poulsbo graphics hardware.

(Reply to this) (Parent)(Thread)


[info]ajaxxx
2009-04-17 06:38 am UTC (link)
While Larrabee is indeed all-Intel, that's not a guarantee that the people currently doing the Intel 965 and 915 drivers are in any way working on the Larrabee drivers, or that there's indeed anyone tasked with making Larrabee work on Linux at launch. Remember, Intel is a huge company. It's entirely too easy for the left hand to not know what the right is doing. That's how we got the Poulsbo problem in the first place.

(Reply to this) (Parent)(Thread)

Larabee
(Anonymous)
2009-04-17 08:10 pm UTC (link)
While I don't know how well Larabee will perform, I do know a few of the engineers personally, and Linux is top on their list.

(Reply to this) (Parent)


[info]ajaxxx
2009-04-17 06:30 am UTC (link)
If by "open source' for those two first two features you are okay with meaning "open source with some APIs you can't ever fix because it'll break feature #3", then yes, that's accurate.

(Reply to this) (Parent)


[info]reddragdiva
2009-04-17 07:17 pm UTC (link)
Intel is a rather large company. A lack of joined-up thinking is industry standard best practice in such situations. The good thing is that there is anyone at all there in the first place who (a) gets it (b) is allowed to put getting it into practice.

(Reply to this) (Parent)

dont want or need a git tree?
(Anonymous)
2009-04-17 05:30 pm UTC (link)
I agree that the poulsbo driver situation sucks.. but I don't think you can hold against them their inability to understand how glorious git is.

After all, it is quite hated by many in the community. I like git, but I did use way too much time learning it and it's internals. More worryingly, I can see how git is going to become the biggest barrier for entry for newcomer contributors.

(Reply to this) (Thread)

Re: dont want or need a git tree?
[info]reddragdiva
2009-04-17 07:20 pm UTC (link)
whuh?

What's hard about git? It's worked fantastically well for Wine, for example. We get Windows refugees who've just installed Ubuntu coming to the wine-users list, getting pointed at RegressionTesting and, if their heads don't explode at a command line, we get usable bug reports with the guilty commit identified. This suggests it's not that hard for newcomers to at least regression-test.

The "many in the community" who are claimed to hate it are not newcomers. I'm interested in the evidence that git itself puts off teh n00bs.

(Reply to this) (Parent)(Thread)

Re: dont want or need a git tree?
[info]reddragdiva
2009-04-17 07:22 pm UTC (link)
And when I say wine-users "list", I actually mean forum (which is gatewayed to and from a mailing list). So they don't even require familiarity with mailing lists to get that far.

(Reply to this) (Parent)

Re: dont want or need a git tree?
(Anonymous)
2009-04-17 08:13 pm UTC (link)
Actually, they do have an internal GIT tree. What they can post externally has to go through an approval process by the vendor that supplied the IP (which makes life miserable). The external git tree is very low priority, and with their recent cutbacks and holds on overtime, it is one of the first things neglected.

(Reply to this) (Parent)

Re: dont want or need a git tree?
[info]fooishbar
2009-04-17 09:56 pm UTC (link)
This is not about git vs. svn for some random small project. If you do not understand git, then you do not understand kernel development.

(Reply to this) (Parent)

i'd like to see..
(Anonymous)
2009-04-17 08:26 pm UTC (link)
I'd like to see them fix the driver for the goddamn GMA950/X3*00 chipset that's in the majority of laptops from 2006 to 2008 first.
Freedesktop's bug tracker is filled with high priority bugs that crash X and they aren't getting solved, and the "new" Xorg 7.4/intel's 2.6.x driver is utter crap, worse than the one from two years ago (hard crash when going to suspend, with GL apps, when switching to VTs, etc)

(Reply to this)

the updated patches, git etc
(Anonymous)
2009-04-18 03:04 pm UTC (link)
So the "kernel rpm" that is in the edit just has the patches that were posted to lkml and the dri devel lists.... nothing special, just patches that were sent upstream (and then rejected, but that's a different story) and then put in a distro kernel rpm.

Not using git for maintaining a kernel rpm is not very special, even though mjg59 seems to pretend it is. Even Fedora does not use git for doing this, they use CVS....


Arjan -- who works for Intel on the Moblin distribution

(Reply to this)

Fix Poulsbo!!!
(Anonymous)
2009-08-03 08:08 pm UTC (link)
C'mon Intel, How many more linux users will have to bitch before you dedicate some resources to fixing the problem with Poulsbo!? Sure, I know there's one binary blob floating out there.. and I'm fine with a binary only driver, as long as it works as well as it does in Windows. I'd like working 3d with all the video acceleration features of the GMA500. The current binary blob barely works at all, and crashes like crazy (if you can even get it to install) on modern distros.

(Reply to this)


(26 comments) - (Post a new comment)

Create an Account
Forgot your login or password?
Login w/ OpenID
English • Español • Deutsch • Русский…