Kernel Traffic #223 For 20 Jul 2003

By Zack Brown

Table Of Contents


Today is my birthday. It's also the anniversary of the Apollo 11 Moon-landing in 1969, and the Viking 1 Mars-landing in 1976. Some would say this is fitting.

Mailing List Stats For This Week

We looked at 1421 posts in 7214K.

There were 424 different contributors. 211 posted more than once. 167 posted last week too.

The top posters of the week were:


1. Linux 2.5.73 Released
22 Jun 2003 - 4 Jul 2003 (26 posts) Archive Link: "Linux 2.5.73"
Topics: FS: ext3
People: Linus Torvalds

Linus Torvalds announced 2.5.73 ( , and said, "Various updates all over the map here. Lots of ia64 updates, and Andrew merged the ext3 locking cleanups/fixes that have been in the -mm tree for a while."


2. Virtual Memory Documentation For 2.6
30 Jun 2003 - 4 Jul 2003 (79 posts) Archive Link: "What to expect with the 2.6 VM"
Topics: Virtual Memory
People: Mel GormanDaniel Phillips

Mel Gorman posted a small document ( and explained, "I'm writing a small paper on the 2.6 VM for a conference. It is based on 2.5.73 and is basically an introduction to some of the new features that are up and coming. As it's only an introduction, it's a little light on heavy detail. Below is the parts of the paper which I thought would be of general interest (I left out stuff like the abstract, introduction, conclusion and bibliography)." Daniel Phillips replied, "Nice stuff, and very timely." And Mel said, "I was hoping someone else would write it so I could read it but thats what I said about the 2.4 VM :-) . Yep, once again, my contributions are mainly documenting related, believe it or not, I actually do code a bit from time to time." He went on, "I was going to update the whole document based on this thread and repost it but it's looking like it'll take me a few days for a week before I work through it all (so I'm slow, sue me). This is especially true as there is a lot of old email threads I need to read before I understand 100% of the current discussion (which is also why I'm not replying to most posts in this thread). Instead, I'm going to post up the bits that are changed and hopefully get everything together." Elsewhere, a lot of folks discussed the technical issues Mel had referred to; and Mel worked on collecting it all together.


3. Encryption Over Loopback; API Changes In 2.5
2 Jul 2003 - 8 Jul 2003 (38 posts) Archive Link: "[PATCH] cryptoloop"
Topics: Advanced Encryption Standard
People: Andries BrouwerChristoph HellwigJari RuusuAndrew Morton

Andries Brouwer posted a patch which "adds a module cryptoloop, with source file drivers/block/cryptoloop.c and config option BLK_DEV_CRYPTOLOOP. It is the missing link between the kernel crypto subdirectory and the loop device." He went on, "This means that one no longer needs kernel patches to use encryption via the loop device. This goes together with mount and losetup from util-linux 2.12 which will be released as soon as this goes in."

Andrew Morton pointed out some complication in the code, that he felt was unnecessary. He suggested a solution, but Andries replied that Andrew's suggestion would change the APIs, and require source modifications of all modules currently using the loop device. He said he'd be willing to do it if no one objected, and Christoph Hellwig said, "I don't think that matters. This is 2.5 and APIs tend to change in unstable series :)"

Jari Ruusu protested that this would be a big inconvenience for him personally, as he would have to clean up the loop-AES code to compensate for the API change. The discussion began to revolve around Jari's patch submission practices. Christoph and Andrew said he should send small, self-contained patches that might get into the official tree, instead of the big one he'd sent earlier. Christoph in particular said that if Jari didn't want to take the necessary steps to get his code into the tree where it could be better protected from these kinds of API changes, then he would have to deal with them himself.

The thread petered out around there, with the API change looking more likely than the status quo.


4. Linux 2.5.74 Released
2 Jul 2003 - 3 Jul 2003 (13 posts) Archive Link: "Linux 2.5.74"
Topics: USB
People: Linus Torvalds

Linus Torvalds announced Linux 2.5.74 ( and said:

Updates all over, the patch itself is big largely because of a MIPS/MIPS64 merge (and SH, for that matter).

Network driver updates, USB updates, PnP, SCTP, s390, you name it. See the changelog for more details.


5. SELinux Module Patch For 2.5
3 Jul 2003 - 8 Jul 2003 (13 posts) Archive Link: "[PATCH] Add SELinux module to 2.5.74-bk1"
People: Stephen Smalley

Stephen Smalley announced, "The patch against 2.5.74-bk1 available from adds the SELinux module to the tree and modifies the security/Makefile and security/KConfig files for SELinux. The last dependency for SELinux, the vm_enough_memory security hook, was included in -bk1." Comments were generally favorable, though several folks had suggestions for implementation details.


6. Framebuffer Updates; RadeonFB Long-Term Bugs
3 Jul 2003 - 9 Jul 2003 (10 posts) Archive Link: "New fbdev updates."
Topics: Framebuffer, I2C, Sound: i810, Version Control
People: James SimmonsJoshua Kwan

James Simmons said:

I have updates to the framebuffer layer. Alot of bug fixes accumlated. A couple of driver updates as well. I have more code to go in but haven't had time to add them in. Please test. This is not the final code going in just yet. More needs to be done. The patches are at the usual

or BK



The NVIDIA driver hasn't been updated to the new cursor fixes. I haven't had time to fix it yet. So it doesn't compile. Also the i810 driver might have issues with the cursor changes.

Skeleton i2c support is added. It hasn't been tested or even compiles in yet.

Joshua Kwan asked if James worked on RadeonFB issues in 2.5, since that driver had been flaking out for a long time. But James replied, "Nope. I don't have this hardware. Once I get my hands on that kind of hardware I will give it a shot." Elsewhere, Scott Robert Ladd also asked about the RadeonFB, and James said he'd accept patches if anyone else wanted to send them to him. Scott said he'd take a look and see if he could tackle it.


7. FAT Filesystem Maintainership
4 Jul 2003 - 8 Jul 2003 (10 posts) Archive Link: "FAT statfs loop abort on read-error"
People: Randy DunlapOGAWA HirofumiJeff GarzikChristoph Hellwig

Sancho Dauskardt had a question about the FAT filesystem, but could not get any reply from the maintainer, Gordon Chaffee. He posted to the list, remarking on Gordon's disappearance, and Randy Dunlap replied, "Try asking OGAWA Hirofumi (cc-ed). He's the de facto FAT maintainter. (I asked him to add a patch to MAINTAINTERS...)" OGAWA Hirofumi replied thanking Randy for the vote of confidence, but adding, "honestly, I may not have skill enough." Jeff Garzik said with a smile, "The people who apply your patches think you do." Christoph Hellwig also said, "Given that you have done a nice job ob fatfs in 2.5 and there's no one coming near that many useful contributions in that timeframe I think it would be a good idea to declare you maintainer. According to MAINTAINERS it currently doesn't have any formal maintainer anyway." Ogawa said OK, he'd give it a try; and posted a patch listing himself as maintainer.


8. BitKeeper Snapshots For The 2.4 Tree
5 Jul 2003 (4 posts) Archive Link: "[ANN] 2.4.x snapshots started"
Topics: Version Control
People: Jeff GarzikWilly Tarreau

Jeff Garzik announced:

Just like 2.5.x, nightly snapshots of Marcelo's latest 2.4.x BK repository are being posted on

ftp://ftp.?? (

I created the first snapshot midday as a test, and the standard cron job created a second one, so the current release is 2.4.21-bk2.

Note that snapshots are NOT based off -pre and -rc releases. 2.4.21-bkN will continue, for ascending values of N, until 2.4.22 is released.

Just like 2.5.x, when a new version is released, the old snapshots are moved into ftp://ftp.?? (

Just like 2.5.x, when a new snapshot appears, another automated job generates the incremental diff between this and the last snapshot: ftp://ftp.?? (

Willy Tarreau was very happy to hear about this, and pointed out a bug, which Jeff fixed.


9. Status Of Anticipatory I/O Scheduler
5 Jul 2003 - 7 Jul 2003 (6 posts) Archive Link: "anticipatory scheduler merged"
People: Andrew Morton

Andrew Morton announced:

Linus has merged the anticipator scheduler. A few things to note on this:

  • For some workloads it is around 10% slower than deadline. Most notably database workloads which seek all over the disk performing reads and synchronous writes.

    Sorry, we just weren't able to get the last little bit back.

    But for other workloads (reading lots of filesystem objects when there's a lot of writeout around, when there's a streaming read, etc) it is up to 1000% faster. I believe it is a better all-round IO scheduler.

    Unless Nick pulls a big rabbit out of his hat, database people will need to boot their kernels with `elevator=deadline' to get the last bit of performance back.

  • These changes have been well tested, but it is five months work and over 100 patches. There's probably a bug or two. If you suspect that something has gone wrong at the block layer (lots of tasks stuck in D state) then please retest with `elevator=deadline'.


10. New libsysfs Library To Access SysFS Data
7 Jul 2003 (3 posts) Archive Link: "[ANNOUNCE] libsysfs v0.1.0"
Topics: FS: devfs, FS: sysfs, Hot-Plugging
People: Daniel StekloffGreg KH

Daniel Stekloff said:

I'd like to announce libsysfs - a small library built over sysfs, the virtual filesystem that exports system devices in the Linux 2.5+ kernels. You can find the initial version of the library in a small package called sysutils at: (

The library grew from the requirements of several applications all needing access to system device information in sysfs. We felt it was better to provide a library of common code rather than having each application create their own access. Greg KH's udev application, a User Space replacement for devfs, is one of the applications needing sysfs access. You can see Greg's original announcement here:

The library doesn't implement any device or bus specifics but simply provides generic bus, class, and device access as represented in sysfs. Included with the library in the sysutils package are two small commands that hopefully make viewing sysfs and system device information easier.

systool is a command that can list devices by bus, by class, or by device root - as represented through sysfs.







Sharon And Joy

Kernel Traffic is grateful to be developed on a computer donated by Professor Greg Benson and Professor Allan Cruse in the Department of Computer Science at the University of San Francisco. This is the same department that invented FlashMob Computing. Kernel Traffic is hosted by the generous folks at All pages on this site are copyright their original authors, and distributed under the terms of the GNU General Public License, version 2.0.