login
Search
"After 6 months of careful integration and testing, I'm happy to announce availability of Linux 2.4.35," 2.4 maintainer Willy Tarreau announced on the lkml. This is the second stable 2.4 kernel released since Willy became the 2.4 kernel maintainer nearly a year ago in August of 2006. Source level changes can be viewed through the linux-2.4 gitweb interface. Willy added:
"I'm very conscious that 2.4 has mostly left desktop PCs and notebooks, but it's still commonly found on servers, route reflectors or firewalls. For this reason, I'm open to merge the small updates required to maintain such systems running (eg: PCI IDs and such), but I will generally refuse all patches which add support for new desktop or notebook-specific hardware, unless the people present very convincing arguments. Those people generally would better upgrade their systems to 2.6."
From: Willy Tarreau [email blocked]
To: linux-kernel
Subject: Linux 2.4.35 released
Date: Thu, 26 Jul 2007 22:43:33 +0000
Hi,
after 6 months of careful integration and testing, I'm happy to announce
availability of Linux 2.4.35.
This one contains the same fixes as 2.4.34.6, plus a small set of add-ons,
among which some new PCI IDs, more usb-storage unusual devs, support for
high-speed USB HID, updated e1000 driver, a few watchdog updates, support
for systems with no keyboard controller (mainly blades), backport of the
skge and sky2 drivers from 2.6, support for the "notsc" boot option for
some broken dual-core x86_64 systems with no HPET, and a the latest fixes
from the LVM package.
Note that I'm very conscious that 2.4 has mostly left desktop PCs and
notebooks, but it's still commonly found on servers, route reflectors or
firewalls. For this reason, I'm open to merge the small updates required to
maintain such systems running (eg: PCI IDs and such), but I will generally
refuse all patches which add support for new desktop or notebook-specific
hardware, unless the people present very convincing arguments. Those people
generally would better upgrade their systems to 2.6.
The 2.4.34.X stable branch is now closed and a new 2.4.35.X branch will open
with the first next fixes. This model has proven very efficient to provide
riskless fixes, especially in a situation where several weeks may pass before
a patch gets tested on a production environment.
This version has been tested on x86 SMP, sparc64 SMP and alpha. Please report
any unexpected problem. As usual, all feedback is welcome.
The patch and changelog will appear soon at the following locations:
ftp://ftp.all.kernel.org/pub/linux/kernel/v2.4/
ftp://ftp.all.kernel.org/pub/linux/kernel/v2.4/patch-2.4.35.bz2
ftp://ftp.all.kernel.org/pub/linux/kernel/v2.4/ChangeLog-2.4.35
Enjoy,
Willy
--
Summary of changes from v2.4.35-rc1 to v2.4.35
============================================
final:
- v2.4.34 was released as 2.4.35 with no changes.
Summary of changes from v2.4.35-pre5 to v2.4.35-rc1
============================================
Filippo Carletti (2):
2.4.34 - VIA VT8237A support
Add some AHCI PCI IDs
Krzysztof Strasburger (1):
notsc support for x86_64
Lior Dotan (1):
Fix divide by 0 in vegas_cong_avoid()
PaX Team (1):
random device reseed bugfix, possibly security problem
Vincent Bernat (1):
SATA update: add ICH8 PCI IDs
Willy Tarreau (2):
Documentations/SubmittingPatches was outdated
Change VERSION to 2.4.35-rc1
Summary of changes from v2.4.35-pre4 to v2.4.35-pre5
============================================
Brian Maly (1):
fix 'pc_keyb: controller jammed (0xA7)' error on systems with KVM
Willy Tarreau (7):
do not mark init_idle() __init
Bluetooth: correct fix for CVE-2007-1353
[BACKPORT] Bluetooth: Fix NULL pointer dereference in HCI line discipline
[BACKPORT] Bluetooth: Fix unintentional fall-through in HCI line discipline
lvm: update to latest fixes from the LVM package
lvm: do not update extent count if snapshot allocation fails
Change VERSION to 2.4.35-pre5
Summary of changes from v2.4.35-pre3 to v2.4.35-pre4
============================================
Urs Thermann (1):
recent patch to fib_semantics broke build
Willy Tarreau (1):
Change VERSION to 2.4.35-pre4
Summary of changes from v2.4.35-pre2 to v2.4.35-pre3
============================================
Adrian Bunk (1):
init/main.c: remove comment that gcc 4 was not supported
Alexey Korolev (1):
Linux 2.4.x MTD CFI P30/P33 support
Tal Kelrich (2):
2.6 backport of Watchdog wdt83627 (Winbond W83627HF/F/HG/G) driver
Watchdog w83977ef (Winbond W83977EF) driver
Willy Tarreau (6):
[DECNet] fib: Fix out of bound access of dn_fib_props[]
[IPv4] fib: Fix out of bound access of fib_props[]
[PPP]: Don't leak an sk_buff on interface destruction.
Fix TCP receiver side SWS handling
[Bluetooth] Fix L2CAP and HCI setsockopt() information leaks (CVE-2007-1353)
Change VERSION to v2.4.35-pre3
Summary of changes from v2.4.35-pre1 to v2.4.35-pre2
============================================
Andrzej Stypula (2):
fix channel balance on TV cards with LG head
fix build of serial with DEBUG enabled
Gilles Espinasse (2):
usb-storage: backport unusual_devs from 2.6.20
usb-storage: HP-FDC-GOLD need US_FL_SINGLE_LUN
Krzysztof Strasburger (3):
make it possible to compile x86_64 without VT support
add nforce-MCP55 chipset ID
restore ability to set disk geometry on command line
Mariusz Kozlowski (24):
arm: dma-arc add missing parenthesis
add missing parenthesis in pci-v320usc
amijoy joystick parenthesis fix
rio parenthesis fix
serial_amba parenthesis fix
ide legacy hd parenthesis fix
ide serverworks parenthesis fix
message fusion parenthesis fix
net wavelan parenthesis fix
scsi 53c7,8xx parenthesis fix
scsi 53c7xx parenthesis fix
scsi aic7xxx parenthesis fix
video pm3fb parenthesis fix
video sis parenthesis fix
fs ufs macro parenthesis fix
arm sa1100 parenthesis fix
mips/mips64 mv64340 parenthesis fixes
mips64 klconfig parenthesis fix
parisc pdc parenthesis fix
ppc m48t35 parenthesis fix
ppc ppc4xx_dma parenthesis fix
sparc64 dma parenthesis fixes
byteorder swab parenthesis fix
reiserfs parenthesis fix
Pete Zaitcev (1):
usb 2.4: Support high-speed HID
Willy Tarreau (6):
ia64 kernel entry fix
add missing parenthesis in video/cyberfb
fix missing parenthesis in unused macro in ip_nat_standalone
remove excess parenthesis in video/cyberfb
IPV6: ipv6_fl_socklist is inadvertently shared.
Change VERSION to v2.4.35-pre2
Summary of changes from v2.4.34 to v2.4.35-pre1
============================================
Christian Praehauser (1):
[NET] ethernet: Fix first packet goes out with MAC 00:00:00:00:00:00
dann frazier (1):
smbfs: fix problems introduced by last security backport
Jesse Brandeburg (3):
e1000: integrate latest 2.4 linux driver
[MAINTAINERS]: update e1000 maintainers
e1000: update README
Oliver Neukum (1):
proper locking on disconnect for mdc800
Willy Tarreau (2):
merge 2.6 backport of skge/sky2 network drivers
Change VERSION to 2.4.35-pre1
While I applaud Willy for
While I applaud Willy for his work and his willingness to maintain an ancient kernel for those who still use it, I cannot agree with his reasoning.
Certainly, he's right when he says that people would be better off upgrading their systems to 2.6; however, the same thing can reasonably be said about servers as well. It may be just me, but I honestly do not see the difference: upgrading a server is work, things may break, and it may lead to downtime or other problems, but the same is true for desktop systems as well. And yes, people depend on servers; but certainly, they depend on desktop systems just as much.
Willy is clearly giving servers a preferential treatment here, and for something as trivial as PCI IDs, too. I could understand it if he said he was opposed to larger changes (although even in that case, I'd expect him to be agnostic about the server/desktop distinction when he makes a decision about what does or doesn't go in), but PCI IDs? That's just asinine.
Huh?
Moron. People doesn't depend on servers; companies do. Comparing breaking a server with breaking a desktop is just insane. Willy is spot on.
Name-calling doesn't get you
Name-calling doesn't get you anywhere, and I'm talking to both you and the top poster.
Anyway, it depends on the desktop, the application running on it, and how it's used. I actually will continue to run Linux 2.4 for certain boxes, because it works and there's no need to upgrade. My new stuff indeed does get kernel 2.6.
Server downtime = millions
Server downtime = millions of dollars lost
Desktop downtime = a few minutes lost
Agreed. What's the worst
Agreed. What's the worst case scenario for a desktop? It was a $10,000 workstation where some 6 figure a year tech does critical things each day and it needs to go down for a day so the admin can do an upgrade?
In this scenario you're using networked data, so the highest cost you could really create would be the emergency purchase of a second $10,000 workstation for these contingencies.
Server downtime goes over $10,000 very quickly.
So true, wouldn't put it
So true, wouldn't put it better myself :-)
expiration date
which leads us to an interesting question :
how long should/could an "old" kernel such as 2.4 be maintained and how do you
qualify it as outdated if it keeps being maintained/updated forever.
while not really slowing down 2.6 development it still keeps some developers back.