Kernel release status

[Posted May 16, 2007 by corbet]

The current 2.6 prepatch is 2.6.22-rc1, released by Linus on May 12. Features added to 2.6.22 since last week's Kernel Page include the eventfd system calls, a new IEEE 1394 ("Firewire") stack "designed for robustness and simplicity," drivers for KingSun DS-620 USB-IrDA dongles, Native Instruments USB audio devices, and WM8753 audio codecs (as found in the OpenMoko phone), a large set of fixes to the "libertas" wireless driver, and support for a number of new ARM processors. See the short-form changelog for the details, or the log-format log for vast amounts of detail.

As of this writing, about 100 changesets (almost all fixes) have been added to the mainline repository since 2.6.22-rc1.

The current -mm tree is 2.6.22-rc1-mm1. Recent changes to -mm include a number of filesystem writeback fixes, a CRC7 implementation, an improved version of the swap prefetch code, and an early startup development tree for the i386 architecture.

to post comments

Kernel release status

Posted May 17, 2007 4:42 UTC (Thu) by jwb (guest, #15467) [Link] (2 responses)

I'd love to see a future LWN feature on the 1394 stack. 1394 has always been rickety on Linux, and not at all like the bulletproof USB stack. Better 1394 support would be of interest to many desktop users.

Kernel release status

Posted May 17, 2007 14:03 UTC (Thu) by i3839 (guest, #31386) [Link]

Which bulletproof USB stack? ;-)

It was rewritten one or two times...

Kernel release status

Posted May 18, 2007 2:40 UTC (Fri) by aleXXX (subscriber, #2742) [Link]

I didn't have problems with the "old" firewire stack, using it for iso
traffic. It's by far easier to work with it than with the high level
firewire interface you get on OSX. And it's also harder to crash the
Linux kernel from user space with firewire while it happened several
times on OSX (I think the reason where these "DMA-programs").

But getting a simpler and more robust stack is a good thing, since it
seemed quite hard to port the old one to other firewire chipsets.
Hopefully this will be easier with the new one.

Alex

Audio codec in the kernel?

Posted May 17, 2007 6:45 UTC (Thu) by Felix.Braun (guest, #3032) [Link] (2 responses)

I've been quite surprised to learn, that this development cycle has seen the addition of an audio codec to the kernel. This seems like the kind of thing that would rather belong into user space. Anybody here care to explain, why this particular codec is different?

Audio codec in the kernel?

Posted May 17, 2007 7:16 UTC (Thu) by jimparis (guest, #38647) [Link]

The WM8753 is a sound chip, and a driver for this chip was added. The term "codec" may be confusing you, this is not something like MP3.

Audio codec in the kernel?

Posted May 17, 2007 19:28 UTC (Thu) by farnz (subscriber, #17727) [Link]

Amplifying on jimparis's reply, there's two usages of the word codec in the audio industry.

The more well-known use is codec as in Ogg or MP3, referring to compression and decompression methods.

The less well-known use is codec as in combination of a digital to analogue and analogue to digital convertor. This is the use referred to in the changelog; the WM8753 chip referenced is a combination of ADC, DAC and analogue I/O (including appropriate amplifiers).