Who wrote 3.5 [LWN.net]

Who wrote 3.5

Posted Jul 26, 2012 14:06 UTC (Thu) by intgr (subscriber, #39733)
Parent article: Who wrote 3.5

It's really sad to see that AMD -- who used to be a regular appearance before 3.x or so -- has been absent for many recent releases. Does anyone know why is that so?

to post comments

Who wrote 3.5

Posted Jul 26, 2012 16:13 UTC (Thu) by iabervon (subscriber, #722) [Link] (2 responses)

I think their hardware requires relatively little variation in the kernel portion of drivers. At least, I'm not aware of anything that AMD ought to work on in the kernel. Of course, they could arbitrarily employ kernel developers working on things other than support for their hardware, but I'd much rather they put their resources toward support for their hardware in whatever part of the system needs it, which is currently userspace graphics drivers.

Who wrote 3.5

Posted Jul 28, 2012 15:31 UTC (Sat) by jengelh (subscriber, #33263) [Link]

>I think [AMD's] hardware requires relatively little variation in the kernel portion of drivers.

Requiring less code is indeed a wonderful thing.

AHCI is one of these masterpieces: a single driver driving many consumer-level disk chipsets at once. The sound club tried something similar with AC97 and HDA. Why can't we have a unified HW API like that for further types of silicon as well, like, network cards?

Who wrote 3.5

Posted Aug 6, 2012 14:48 UTC (Mon) by nhippi (subscriber, #34640) [Link]

> I think their hardware requires relatively little variation in the kernel portion of drivers. At least, I'm not aware of anything that AMD ought to work on in the kernel.

The AMD GPU's would need more work. OTOH much that is on the MESA/Xorg side rather than in the kernel.

Who wrote 3.5

Posted Aug 23, 2012 13:37 UTC (Thu) by daenzer (subscriber, #7050) [Link] (1 responses)

Eh? AMD and in particular Alex Deucher were among the most active employers / developers of the 3.4 cycle, both by changesets and number of lines changed. Similarly for many previous cycles. I'm afraid you mistook an exception for the rule.

Who wrote 3.5

Posted Aug 23, 2012 15:17 UTC (Thu) by intgr (subscriber, #39733) [Link]

> I'm afraid you mistook an exception for the rule.

Ok let's have a look at the numbers. Here are the stats for AMD in particular...

3.5 : missing
3.4 : changesets 1.2% #19, lines 1.5% #15
3.3 : missing
3.2 : missing
3.1 : missing
3.0 : changesets 1.8% #12
2.6.39: changesets 1.3% #16
2.6.38: changesets 2.0% #9 , lines 1.1% #19
2.6.37: changesets 1.5% #14
2.6.36: changesets 1.4% #14
2.6.35: changesets 1.7% #10
2.6.34: changesets 1.3% #17, lines 1.5% #19
2.6.33: changesets 1.6% #13, lines 0.8% #18
2.6.32: changesets 2.3% #9
2.6.31: changesets 1.2% #13
2.6.30: changesets N/A, lines 2.1% #10

I'm afraid you mistook an exception for the rule. AMD had significant presence in all releases between 2.6.30 - 3.0, but then suddenly disappeared (apart from one exception that is 3.4).

(I created an index of LWN development statistics articles for easier reference here:
http://kernelnewbies.org/DevelopmentStatistics)