login
Header Space

 
 

Linux: 2.6.11 Released, "Bug Free"

March 2, 2005 - 8:59am
Submitted by Jeremy on March 2, 2005 - 8:59am.
Linux

Linux creator Linus Torvalds released the much anticipated 2.6.11 Linux kernel declaring, "so it's now _officially_ all bug-free." Though bugs are certain to still remain, quite a bit of effort was made to stabalize this release. On February 12'th, Linus uploaded 2.6.11-rc4 [story] containing only fixes, planning to follow it with the official 2.6.11 release. However, instead it was followed on February 23'rd by 2.6.11-rc5 [story] to deal with "some laptop resource conflicts, various ppc TLB flush issues, some possible stack overflows in networking and a number of other details". Today's release contained minimal changes, "mostly some fixes from various code checkers".

The 2.6.11 patch to the 2.6.10 kernel is about 4MB compressed, and it or the entire 2.6.11 kernel can be obtained from your nearest kernel.org mirror. The kernel makefile still contains the name Woozy Numbat, as given for the 2.6.10 release last Christmas eve [story]. Read on for the relatively short changelog since 2.6.11-rc5.


From: Linus Torvalds [email blocked]
To: Kernel Mailing List [email blocked]
Subject: Linux 2.6.11
Date: 	Wed, 2 Mar 2005 00:02:03 -0800 (PST)


Ok,
 there it is. Only small stuff lately  - as promised. Shortlog from -rc5 
appended, nothing exciting there, mostly some fixes from various code 
checkers (like fixed init sections, and some coverity tool finds).

So it's now _officially_ all bug-free.

			Linus

----
Summary of changes from v2.6.11-rc5 to v2.6.11
============================================

<c.lucas:com.rmk.(none)>:
  o [SERIAL] drivers/serial/*: convert to pci_register_driver

<takis:lumumba.luc.ac.be>:
  o prism54 not releasing region

Alex Williamson:
  o [SERIAL] 8250 woraround for buggy uart

Alexander Nyberg:
  o SELinux: Leak in error path
  o SELinux: null dereference in error path

Andrea Arcangeli:
  o Make the new merged pipe writes check for SIGPIPE

Andrew Morton:
  o binfmt_elf build fix
  o [IA64] ia64 audit build fix
  o genhd: NULL checking fix

Andries E. Brouwer:
  o __devinitdata in parport_pc
  o __init in cfq-iosched.c
  o remove __initdata in scsi_devinfo.c
  o __initdata in apic.c
  o more apic.c

Aurelien Jarno:
  o USB: Fix usbfs regression

Bartlomiej Zolnierkiewicz:
  o [ide] fix build for built-in hpt366 and modular ide-disk
  o [ide] fix IRQ masking in ide_do_request()

Ben Dooks:
  o [ARM PATCH] 2498/1: CREDITS - add Ben Dooks
  o [ARM PATCH] 2505/1: Remove FTVPCI from debug code

Bjorn Helgaas:
  o [SERIAL] discover PNP ports before PCI, etc
  o [SERIAL] add TP560 data/fax/modem support

Chris Wright:
  o fix audit inode filter
  o send audit reply to correct socket

David Gibson:
  o ppc64: hugepage hash flushing bugfix

David Howells:
  o Make keyctl(KEYCTL_JOIN_SESSION_KEYRING) use the correct arg

David S. Miller:
  o [IPV4]: Fix lost routes in fn_hash netlink dumps
  o [AF_UNIX]: Fix SIOCINQ for STREAM and SEQPACKET

Dmitry Torokhov:
  o Input: add more PNP IDs to i8042 driver

Greg Kroah-Hartman:
  o sysfs: fix signedness problem
  o fix module paramater permissions in radeon_base.c
  o USB: fix bug in acm's open function

Harald Welte:
  o [NETFILTER]: ipt_hashlimit rule load time race condition

Hideaki Yoshifuji:
  o add sysctl helper functions to provide milliseconds-based
    interfaces
  o [IPV4] Use appropriate sysctl helpers for gc_min_interval_ms
  o [IPV6]: Unregister per-device snmp6 proc entry earlier

Jens Axboe:
  o [PATCH] Fix bounced bio and dm panic

Kenji Kaneshige:
  o [IA64] pci_irq.c: need signed variable to handle error return from
    acpi

Linus Torvalds:
  o Fix possible pty line discipline race
  o Properly limit keyboard keycodes to KEY_MAX
  o Make pipe "poll()" take direction of pipe into account
  o Linux 2.6.11

Nishanth Aravamudan:
  o [PKTGEN]: Replace interruptible_sleep_on_timeout()

Olaf Hering:
  o Fix incorrect __init on 'modedb[]' array

Patrick McHardy:
  o [NETFILTER]: Prevent NAT from seeing fragments

Randy Dunlap:
  o [ide] make 1-bit fields unsigned
  o srat: initdata section references
  o sound/oss/aedsp16: init/exit section cleanups
  o sonicvibes: fix initdata references
  o sound/oss/opl3as2: fix init section reference
  o isdn: use __init for ICCVersion()
  o dc395x: fix section references
  o hp100: fix section references
  o rrunner: fix section references

Robert Olsson:
  o [PKTGEN]: reduce stack usage

Russell King:
  o [ARM] Fix dma_mmap() size argument

Sascha Hauer:
  o [ARM PATCH] 2496/1: i.MX DMA fix
  o [ARM PATCH] 2497/1: i.MX pll decode



Related Links:

AttachmentSize
patch-2.6.11.bz2.sign248 bytes
patch-2.6.11.bz23.99 MB

Linux: 2.6.10-as5, Security Fixes

March 2, 2005 - 9:53am
maceto (not verified)

all that merged into mainline, that is included in 2.6.11 final.
And with all this development going on without any clear guideline, eg Linux: 2.6.10-as5, Security Fixes, have ppl checked debian/fedora/RHEL kernels in terms of these patches...

It`s come to this: If you know what to do, use upstream kernels with as5 fixes, cause not even RedHat can follow this..

It`s like what to use?, what is best etc, what is beeing meerged back into main, does eg Novell messages that this and this is fixed.. cause I don`t think so- SO much is beeing fixed in newer version, but I don`t see all these fixes beeing meerged into the distros them self

Is this English?

March 2, 2005 - 11:56am

I seriously can't understand a sentence (term used loosely) in here. If English is not your first language, I apologize, but... whaaa?

I would guess that English is

March 2, 2005 - 1:58pm
Anonymous (not verified)

I would guess that English isn't his first language, based on the fact that all of the apostrophes are pointing the wrong direction.

all your base are belong to u

March 2, 2005 - 4:41pm
Anonymous (not verified)

all your base are belong to us

Maybe this machine translation will help:

March 2, 2005 - 9:24pm
Anonymous (not verified)

everything, which mixed in Main LINE, is covered in final round 2.6.11. And with this whole development, which continues without any free guideline, e.g. Linux: 2.6.10-as5, security regulates, has ppl examined debian /fedora/RHELkerne in these marks expressed...,

It`s come to this: If you know that which talk-has to do, upward arranged cores of the use with shiftingnesses as5, not even can this follows to cause you.

It`s may which use?, which best etc. is, beeing which, meerged back in main line, i-don`t e.g. announcements Novell, which don`t this and this locally specified the cause I thinks AS MANY beeing regulated in such a way in the newer version is, but sees that all this shiftingnesses the Beeing into distros her meerged

Heh, go to the IvTV mailing l

March 3, 2005 - 4:14am
Anonymous (not verified)

Heh, go to the IvTV mailing list and read the posts by Chris Kennedy (one of the main deveopers). He's a heck of a coder but hist posts are more or less unreadable (to me anyway, sorry Chris!), and the name suggest he's english speaking by native.

- Peder

Aliens on Earth! This is the

March 2, 2005 - 1:40pm

Aliens on Earth! This is the proof! :)

Changelog

March 2, 2005 - 10:24am

Is there anyone who knows his stuff able to post a changelog summary? I'd like to know what big changes there are in this kernel. Ofcourse I can read the official changelog, but I'm unable to filter out important/big changes.

Importance is relative.

March 2, 2005 - 1:53pm

Importance is relative.

2 major drivers - it821x & sn9c10x

March 4, 2005 - 10:51am

2 big fixes for me, been waiting :))

1) it821x support (ATA raid on my motherboard, 8 ATA drives now possible. I have a gigabyte board)
2) sn9c10x (common cheap USB webcams)

^ both of these I own and new support for me, woo xmas is early! :)

NVidia drivers did compile for me on 2.6.10ac12 so I'm hopefully about 2.6.11 when I get home...

other stuff:

wireless cards region fix
usbfs fixes (affects me with ehci; that's USB2)

edit: darn, looks like AlanCox must have had some troubles meeting 821x into 2.6.11; diodn't make it

NVidia drivers didn't compile :( so searching to see if NVidia have responded... aggh the pain of closed src. NVidia linux support will become abandonware in the future. Need a GPL 3D solution to escape the upgrade tax...

Obligitory...

March 2, 2005 - 11:59am
Mr_Z (J Zbiciak) (not verified)

...there goes my uptime! ;-)

Seriously, I'm happy to see a G200 driver in this one. (Odd, seems like that should've been around forever, but make oldconfig listed it as a new option.) Right now I'm limping along on an old Dual Pentium II/III 450MHz machine with a G200 card. :-/

I spilled a drink into the top fan on my Cooler Master case, thereby evenly distributing said drink over my dual Athlon MP 2700 system's internals. :-( The CPUs didn't get fried, but the mobo (after a thorough washing) was flaky. I sold the CPUs rather than revel in my misfortune.

Now that the Opterons have come down in price, I can't wait to build a dual Opteron 246 system. Any word on what's a good Linux-supporting motherboard that uses both Opteron's memory ports? (The early Opteron boards used only one CPU's memory ports.)

Tyan!

March 2, 2005 - 3:49pm
Anonymous (not verified)

Although I'm not sure about the memory issue, we've been using the 2882's with 2.6.10 and they've been great so far.

Tyan S2885ANRF

March 2, 2005 - 5:12pm
Anonymous (not verified)

I've been deploying dual opterons built on a Tyan S2885ANRF for some time now.

Linux supports the sound chip, onboard LAN, firewire, and best of all, USB/USB2 (with USB storage sticks, keyboard, mice, you name it) reliably. My Athlon64 and Athlon boxes do not have reliable USB storage, so this is a definite plus if, like me, you back up your data to a USB stick.

The board is a little pricey, but well worth it IMHO.

Jean.

Eh?

March 2, 2005 - 6:51pm
Anonymous (not verified)

G200 framebuffer console and DRI have existed for a long time. This patch was just a reorganization.

ati module

March 2, 2005 - 1:47pm
Anonymous (not verified)

And of course the ATI kernel module does not compile.
I get sick of this api change stuff.
This has to end! Every new version.

Complain to your graphics car

March 2, 2005 - 2:16pm
Anonymous (not verified)

Complain to your graphics card manufacturer for the lack of open source drivers. If no one complains, they'll never change anything.

complaining may not help either

March 2, 2005 - 2:49pm
Anonymous (not verified)

Just because not complaining is not helping, does not guarantee that complaining will...
I have been begging ati/gatos/whoever for better support for years and it hasn't helped. Having more competition in this sector would probably achieve more.

Gatos?

March 2, 2005 - 2:56pm
Anonymous (not verified)

What do you mean by more competition? You want to compete against the Gatos project? It's basically a one man show, and a one man show that was fighting with XFree86 for years.

You can beg ATi all you want, and I encourage you to. ATi can make better drivers happen, period. But begging Gatos for better support? They can't make hardware specs appear out of thin air, and they don't have enough developers to keep up with X and the kernel, full tilt.

Gatos has been incorporated into the X.org tree, so I expect to see better work out of that direction soon, but please. They need developers, not beggers.

Business model

March 2, 2005 - 3:14pm
Anonymous (not verified)

Make a board that does everything you want, and GPL the drivers. Use the quality/transparency of the product to drive the market in noble directions.

May the source be with you.

nonsense

March 3, 2005 - 8:08am
Anonymous (not verified)

I think it is their right to make software non-open.
People aren't bad because they do so.
Why make ist harder for hardware vendors as it needs to be?

Turnabout

March 3, 2005 - 5:26pm
Anonymous (not verified)

It's also the kernel developers' right not to encourage binary-only drivers which make Linux more difficult to support. You might instead go find a distribution which hacks up their kernel to make it easier on binary-only drivers, and pay them for the privilege of having them support such hackery for you.

get into kernel

March 2, 2005 - 3:17pm

Read http://www.kroah.com/log/linux/stable_api_nonsense.html?seemore=y It's true. The only way ATI could help you is by GPLing driver.
--
:wq

nvidia ditto

March 2, 2005 - 6:43pm
Anonymous (not verified)

Same here. Nvidia drivers didn't compile either. I hope they'll come with a new version soon. I'm pretty sure they won't GPL them.

Is there grsecurity patch?

March 2, 2005 - 8:19pm
Anonymous (not verified)

Ahahm... just 17(or 4?) hours & 5 clicks away

March 3, 2005 - 3:19am
Anonymous (not verified)

http://grsecurity.net/pipermail/grsecurity/2005-March/000271.html

..."going to hold off on an official release until the PaX team is able to fix a bug that occurs in certain scenarios"...

ERROR

March 28, 2005 - 12:13pm

Hello, this mornig I tried some things, Well for the first time, I tried xconfig. Got Error 2, can not connect to the XWindow Server. Should I be logged into KDE / Gnu or some other XWindowing system to run this??

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
speck-geostationary