Download Linux Kernel 2.4.31

Linux Kernel 2.4.31 has been released to the general public by Marcelo Tosatti and, as usual, contains a whole bunch of tweaks and improvements. The full included changelog looks like this:

2.4.31-rc2 was released as 2.4.31 with no changes.

Summary of changes from v2.4.31-rc1 to v2.4.31-rc2

• Marcelo Tosatti:
  • Change VERSION to 2.4.31-rc2
• Willy Tarreau:
  • off-by-one in mtrr.c found by Brad Spengler and reported by Julien Tinnes
  • IPVS: Replace several unchecked strcpy() with strncpy() (PaX team)

Summary of changes from v2.4.31-pre2 to v2.4.31-rc1

• aleksey_gorelov:phoenix.com:
  • Fix bug in VIA 82C586B PCI IRQ routing
• julien.tinnes:francetelecom.com:
  • Off-by-one in loop.c
• mkasick:club.cc.cmu.edu:
  • JFS oops fix
• vvs:sw.ru:
  • random pool size sysctl fix
• wtarreau:exosec.fr:
  • fix compilation error introduced by moxa correction
• Christoph Hellwig:
  • XFS: fix compilation error
• David S Miller:
  • Add basic bcm5752 support
  • Add bcm5752 to tg3_pci_tbl
  • Add bcm5752 entry to pci_ids.h
  • use TG3_FLG2_5705_PLUS instead of multi-way if’s
  • define TG3_FLG2_5750_PLUS flag
  • use new TG3_FLG2_5750_PLUS flag
  • more use of TG3_FLG2_5705_PLUS flag
  • use TG3_FLG2_57″ rel=”external”>05.50_PLUS flags in tg3_get_invariants
  • check TG3_FLG2_5750_PLUS flag to set TG3_FLG2_5705_PLUS flag
  • add support for bcm5752 rev a1
  • Minor 5752 fixes
  • Split tg3_phy_probe into 2
  • Setup proper GPIO settings
  • Fix tg3_set_power_state()
  • Workaround 5752 A0 chip ID
  • Add GPIO3 for 5752
  • Add nvram detection for 5752
  • Add nvram lockout support for
  • Fix bug in tg3_set_eeprom()
  • Add msi support
  • Add msi test
  • Update driver version and release date
  • Set SA_SAMPLE_RANDOM in request_irq() calls
  • Elide tg3_stop_block messages when such events are normal
  • Ignore tg3_stop_block() errors
  • Add tagged support status
  • Set minimal hw interrupt mitigation
  • Refine DMA boundary setting
  • In tg3_poll(), resample status_tag after doing work
  • Fix stretch ACK performance killer when doing ucopy
  • Fix off-by-one error in rtnetlink.c
• Luca Tettamanti:
  • fbmem: previous radeonfb fix limits the amount of mmap()’able VRAM to the same amount reserved for the framebuffer
• Marcelo Tosatti:
  • USB: fix oops in io_edgeport.c driver (2.6 backport)
  • Change VERSION to 2.4.31-rc1
• Michael Pettersson:
  • x86_64 linkage error on UP_IOAPIC
• Mike Miller:
  • cciss: fix for passthru ioctls
• Suresh B. Siddha:
  • x86_64: fix pci config space access race condition
  • x86_64: Try harder to allocate memory in pci_alloc_consistent()
• Willy Tarreau:
  • floppy typo fix
• Xose Vazquez Perez:
  • Add 5752M device ID

Summary of changes from v2.4.31-pre1 to v2.4.31-pre2

• carlos.pardo:siliconimage.com:
  • sata_sil: Fix FIFO PCI Bus Arbitration
• david.monniaux:ens.fr:
  • fix moxa crash with more than one 1 board
• jason.d.gaston:intel.com:
  • SATA AHCI correction Intel ICH7R
• Brett Russ:
  • AHCI: fix fatal error int handling
  • libata: support descriptor sense in ctrl page
• Chris Wright:
  • backport v2.6 elf_core_dump() flaw fix (CAN-2005-1263)
• Eugene Surovegin:
  • ppc32: backport Book-E decrementer handling fix from 2.6
• Jean Delvare:
  • I2C updates: Fix typo in a comment in i2c.h
  • I2C updates: Fix I2C_FUNC_* defines in i2c.h
  • I2C updates: Fix an iteration bug in the handling of i2c client module parameters
• Jeff Garzik:
  • [libata ahci] support ->tf_read hook
  • [libata sata_sil] Don’t presume PCI cache-line-size reg is > 0
• Marcelo Tosatti:
  • Change EXTRAVERSION to 2.4.31-pre2
• Michael Pettersson:
  • rwsem spinlock linkage error
• Pete Zaitcev:
  • USB: Add HX type pl2303
• Rolf Eike Bear:
  • typo fix in drivers/scsi/sata_svw.c comment
• Steven HARDY:
  • pcnet32: 79C975 fiber fix
• Willy Tarreau:
  • bonding fix

Summary of changes from v2.4.30 to v2.4.31-pre1

• Adrian Bunk:
  • MAINTAINERS: remove obsolete ACP/MWAVE MODEM entry
• Andi Kleen:
  • x86_64: Handle MM going away in context switch
  • x86_64: Backport 2.6 MTRR algorithms
  • x86_64: noexec=off never worked correctly for the kernel direct mappings
  • x86_64: Fix idle=poll
  • x86_64: Fix reference counting bug in change_page_attr on i386/x86-64
  • x86_64: Resend lost APIC IRQs on Uniprocessor too
  • x86_64: Avoid an obnoxious warning during build
  • x86_64: Flush correctly when more than one page is getting flushed
  • x86_64: Reload init_mm on leaving lazy mm
• Andrew Morton:
  • rwsem: Make rwsems use interrupt disabling spinlocks
• Herbert Xu:
  • [NETLINK]: Fix sk_rmem_alloc assertion failure in af_netlink.c
• Hugh Dickins:
  • madvise_willneed -EIO beyond EOF
• Marcelo Tosatti:
  • Change VERSION to 2.4.31-pre1
• Pete Zaitcev:
  • visor: Add Zire 31 device ID