Linus Torvalds: Linux 2.6.0-test8

Kernel.org
Mirrors

Full Changelog

More changes than I would have liked, but most of them are
fairly small. The most noticeable changes:

• fix the /proc/PID/stat oops that multiple people reported
• workaround for Athlon prefetch bug (occasional spurious page
  faults)
• fix serverworks PIO autotuning
• fix some cpufrequency calculations
• make NFS O_DIRECT work

The rest are some architecture and driver updates, mostly stuff
that people had queued up and convinced me I wanted to merge before
freezing down totally.

I’ve flamed a number of people who flaunted the freeze (I cursed
a lot more than I usually do ;), and they won’t have any excuse to
do so for test9. So expect the patches to shrink considerably in
the coming weeks.

Linus

Summary of changes from v2.6.0-test7 to v2.6.0-test8

---

<davis_g:com.rmk.(none)>:

• [ARM PATCH] 1684/1: Add uImage build target rule
• [ARM PATCH] 1685/1: Fix bootpImage build target
• [ARM PATCH] 1686/1: Fix params-phys typos
• [ARM PATCH] 1687/1: Add {params,initrd}_phys vars to enable
  bootpImage support on SA1100

<shep:alum.mit.edu>:

• [TCP]: Fix missing connections in /proc/net/tcp
• [TCP]: Fix numbering of lines in /proc/net/tcp

Albert Cahalan:

• number of threads in /proc

Alexander Malysh:

• I2C: i2c-sis630 driver fixes

Alexander Viro:

• reiserfs/procfs.c fix

Alexey Dobriyan:

• USB: Correct module names in drivers/usb/*/Kconfig
• USB: Fix two typos in drivers/usb/README

Andi Kleen:

• APIC fixes for x86-64
• Fix x86-64 build with separate object tree
• Disable non 64bit clean drivers for x86-64
• Fix 64bit bug in trident frame buffer driver
• Prefetch workaround for Athlon/Opteron
• [ATM]: Mark unclean drivers as not-64BIT
• Fix software suspend compilation on x86-64
• Fixing mlockall & PROT_NONE
• Mark non 64bit clean network drivers
• Fix warnings in hp100
• Fix warnings in defxx.c
• Enable interrupts in faults from VM86 mode

Andreas Schwab:

• ia64: fix misnamed syscalls

Andrew Morton:

• drivers/block/initrd.c removal
• Add missing sjcd uaccess checks
• Fix early __might_sleep() calls
• SELINUX: add policyvers to selinuxfs
• Correct case sensitivity in make mandocs
• Don’t swap to files which do not implement readpage
• reiserfs documentation URL fixes
• silence smp_read_mpc_oem() declared static but never defined
  warning
• ext3: i_disksize locking fix
• applicom: fix LEAK, unwind on errors
• ext2/ext3 acl signedness fixes
• saa7134-core.c compile fix for old gcc
• current_is_kswapd() is a function

Anton Blanchard:

• ppc64: siginfo fix from Stephen Rothwell
• ppc64: remove some old irq balance related code
• ppc64: Fix for empty memory nodes when NUMA is enabled
• ppc64: IDE fixes from Linas Vepstas
• ppc64: Add paca export, from Sridhar Samudrala
• ppc64: slabify ppc64 pagetables, from Bill Irwin
• ppc64: add in hugetlb quota fixes
• ppc64: add missing include
• ppc64: merge some changes from the ameslab tree
• ppc64: remove last bits of traceback tables
• ppc64: disable_irq fix from x86
• ppc64: clear IRQ_INPROGRESS in setup_irq and enable_irq
• merge some irq.c fixes from x86
• ppc64: Fix for sym2 problem (first place that checked
  pci_set_mwi return code)
• ppc64: nuke a bunch of warnings, from akpm
• ppc64: Use kmalloc for kernel stacks, like x86 (makes usage
  easy to track via slabinfo)
• ppc64: remove old comment, copy_tofrom_user zeros destination
  on a load fault
• ppc64: arguments to mlockall and msync dont need sign
  extending

Armin Schindler:

• Eicon ISDN driver: use work_struct instead of tasklet for
  dpc

Arnaldo Carvalho de Melo:

• [TCP/IPV4]: Fix tcp_tw_bucket accessed as a sock in
  tcp_bind_conflict()

Arun Sharma:

• ia64: fix ia32 epoll emulation warnings
• ia64: Fix unaligned faults in IA-32 putstat64

Bart De Schuymer:

• [EBTABLES]: Add ebt_among match

Bartlomiej Zolnierkiewicz:

• fix support for PIO modes w/o IORDY flow control
• fix ServerWorks PIO auto-tuning
• remove “biostimings” boot options

Benjamin Herrenschmidt:

• add insert_resource() helper function
• Fix call to Open Firmware “map” call, parameters were flipped
  causing the coff image to randomly fail
• coff images bootloader reserves more memory for the kernel,
  allow booting
• Export mol_trampoline and mmu_hash_lock for MOL (missing from a
  previous cset)
• Export init_mm, some modules need that
• The POWER4 support merged previously was incomplete, add the
  missing bits so that the kernel boots at least on POWER4 and G5
  CPUs
• Fix RAMDAC detection on some IMS TwinTurbo video cards (“Mac”
  cards only), without this, you get no display on machines with
  those cards
• Move definition of TimeBase registers from reg_booke.h to
  reg.h, those registers exist on common CPUs and without those
  definitions, SMP won’t build
• Fix timebase hardware sync on SMP PowerMacs (both oldworld dual
  604s and core99 dual G4s).
• Fix VIA-based TB/Decrementer calibration , previously it
  wouldn’t work properly with HZ != 100, causing tb_to_us to be wrong
  and gettimeofday() to return strangely “off” results
• Fix a crash on sleep in the OF platform core where we would use
  a NULL “driver” pointer and actually try to call it after casting
  it!
• Fix repeat delays in adbhit, they didn’t make the transition to
  jiffies based values to ms. This fix crazy key repeat on ADB based
  PowerMacs
• USB: Conflicting definitions in keyspan driver

Bjorn Helgaas:

• [SERIAL] ACPI serial fix
• ia64: fix serial port naming

Chas Williams:

• [ATM]: sdh should be off by default
• [ATM]: Change ifdef in lanai_ioctl
• [ATM]: Remove vcc reference on outstanding SKBs
• [ATM]: Minor cleanup for vcc_hash conversion

Corey Minyard:

• IPMI fixes for 2.6.0-test7

Dave Jones:

• [CPUFREQ] Make VIA longhaul work on Samuel2 & Ezra
  again
• [CPUFREQ] refix EBLCR FSB only works on Samuel1
• [CPUFREQ] Fix longhauls speed calculations
• [CPUFREQ] Kill longhaul warnings (Blah about unused
  variables).

Dave Kleikamp:

• JFS: Make sure journal records get flushed to disk
• JFS: Improved error handing

David Brownell:

• USB: minor net2280 cleanup
• USB: make more driver names match module names
• USB: ehci-hcd, misc bugfixes
• USB: ohci-hcd PM fixes

David Gibson:

• [NET]: Fix initialization sequence in SunGEM driver

David Mosberger:

• ia64: Patch by Jesse Barnes
• ia64: Fix __delay() to do The Right Thing. In practice, this
  may cause BogoMIPS to drop to half the clock-speed with current
  versions of GCC, but this just shows that GCC doesn’t generate very
  good code for single-cycle loops. Perhaps it will motivate someone
  to improve GCC in this area (though it’s hardly of practical value,
  other than for producing large BogoMIPS values).
• ia64: Fix __delay() even more: it’s just not safe to inline the
  delay- loop because, e.g., the compiler might decide to unroll the
  loop when passed a constant, etc.
• ia64: Important security fix for fsyscalls. Without this patch,
  the McKinley E9 workaround caused fsyscalls to return with the
  wrong privilege level.

David S. Miller:

• [VLAN]: kfree(skb) –> kfree_skb(skb)
• [SPARC64]: Squelch bogus gcc warning in unaligned.c due to bad
  flow analysis
• [NET]: Delete skb_shared() checks from loopback driver
  transmit
• [SPARC64]: Update defconfig
• [SPARC64]: Fix __bzero_noasi declaration
• [SPARC64]: Fix mixed declarations and code in
  flush_dcache_page_all
• [ATM]: Kill PROC_FS ifdef around includes
• [EBTABLES]: Print size_t type properly in ebt_among.c
• [SPARC]: Audit inline asm
• [SPARC64]: Update defconfig
• [SPARC]: Fix types in ext2/minix bitops
• [SPARC]: Provide foo_p aliases for I/O port ops
• [SPARC]: Fix IRQ op build problems
• [SPARC]: Make io.h macros more resiliant
• [IPV6]: Fix typo in addrconf.c
• [NET]: Convert sunlance away from init_etherdev
• [NET]: Convert sunbmac away from init_etherdev
• [NET]: Delete bogus unregister_netdev() call in sunbmac
• [NET]: Convert myri_sbus away from init_etherdev
• [NET]: Delete comment reference to init_etherdev which the
  driver does not even use
• [NET]: Convert sk98lin away from init_etherdev

David Woodhouse:

• JFFS2 update; completed support for NAND flash
• Fix jffs2 memory leak on mount error
• Kernel thread signal handling

Dax Kelson:

• USB: Handspring Treo 600 id

Geert Uytterhoeven:

• M68k export csum_partial
• Sun-3 compile fix

Greg Kroah-Hartman:

• I2C: fix i2c-dev class release function bug
• PCI: fix up probe functions for agp drivers
• PCI: fix up probe functions for synclink drivers
• PCI: fix up probe functions for OSS drivers
• I2C: remove unneeded MOD_INC and MOD_DEC calls
• CHAR: Remove unneeded MOD_INC and MOD_DEC calls
• OSS: fix up MOD_INC and MOD_DEC usages to remove compiler
  warnings
• USB: fix visor driver to work with Palm OS 4+ devices
• I2C: fix more define problems in w83781d driver

Herbert Xu:

• [IPIP]: Decapsulate after checking policy
• [IPV4/IPV6]: Clear security path for tunnel packets

Ian Wienand:

• ia64: drop bogus “now < last_tick” message
• ia64: a little more documentation in fsys.txt

Jamie Lokier:

• set sigio target to current->pid and only if not already
  set
• futex bug fixes

Janet Morgan:

• ia64: sys_ia32.c fix for epoll

Jaroslav Kysela:

• ALSA CVS update D:2003/09/26 15:29:05 C:USB generic driver
  A:Takashi Iwai <tiwai@suse.de>
  F:usb/usbmixer.c:1.23->1.24 L:probe units even under a selector
  unit which is marked as ignored or has L:only a single
  selector.
• ALSA CVS update D:2003/09/29 08:31:25 C:PCI drivers A:Jaroslav
  Kysela <perex@suse.cz>
  F:pci/Kconfig:1.8->1.9 L:Removed GAMEPORT dependency (already
  handled in drivers)
• ALSA CVS update D:2003/09/29 19:16:18 C:ALSA<-OSS emulation
  A:Jaroslav Kysela <perex@suse.cz>
  F:core/oss/pcm_oss.c:1.50->1.51 F:include/pcm_oss.h:1.7->1.8
  L:Fixed oops in oss_sync() routine.
• ALSA CVS update D:2003/09/30 08:54:19 C:ALSA<-OSS emulation
  A:Jaroslav Kysela <perex@suse.cz>
  F:core/oss/pcm_oss.c:1.51->1.52 L:Fixed compilation for the sync
  code commited by mistake.
• ALSA CVS update D:2003/09/30 09:58:53 C:Generic drivers
  A:Jaroslav Kysela <perex@suse.cz>
  F:drivers/dummy.c:1.25->1.26 L:Added emu10k1 emulation by
  Takashi
• ALSA CVS update D:2003/09/30 10:28:26 C:Control Midlevel,HWDEP
  Midlevel,ALSA Core,PCM Midlevel,RawMidi Midlevel C:Timer
  Midlevel,ALSA<-OSS emulation,ALSA sequencer,EMU8000 driver
  C:Digigram VX222 driver,USB generic driver A:Jaroslav Kysela
  <perex@suse.cz>
  F:core/control.c:1.37->1.38 F:core/hwdep.c:1.21->1.22
  F:core/init.c:1.38->1.39 F:core/pcm_lib.c:1.43->1.44
  F:core/pcm_native.c:1.81->1.82 F:core/rawmidi.c:1.37->1.38
  F:core/timer.c:1.46->1.47 F:core/oss/pcm_oss.c:1.52->1.53
  F:core/seq/seq_lock.c:1.7->1.8 F:include/rawmidi.h:1.10->1.11
  F:isa/sb/emu8000_patch.c:1.6->1.7
  F:isa/sb/emu8000_pcm.c:1.10->1.11
  F:pci/vx222/vx222_ops.c:1.2->1.3 F:usb/usbaudio.c:1.65->1.66
  L:Revised schedule() and set_current_state() calls.
• ALSA CVS update D:2003/09/30 11:12:09 C:VIA82xx driver
  A:Takashi Iwai <tiwai@suse.de>
  F:pci/via82xx.c:1.52->1.53 L:- fixed the detection of VIA8233A
  (it was overridden by dxs_support L: option).
• ALSA CVS update D:2003/09/30 11:15:44 C:RawMidi Midlevel
  A:Takashi Iwai <tiwai@suse.de>
  F:core/rawmidi.c:1.38->1.39 L:fixed typos (open_lock ->
  open_mutex).

Jean Delvare:

• I2C: Chip driver initialization fixes
• I2C: correct some errors in i2c/chips/Kconfig

Jeff Garzik:

• [netdrvr 8139too] add pci id
• [netdrvr 8139too] another new PCI ID
• [netdrvr tulip] add pci id

Jens Axboe:

• ide-floppy IOMEGA ZIP fix
• fix ide-floppy IOMEGA logic

Jeroen Vreeken:

• [hamradio scc] fix probe function
• [hamradio bpqether] fix ancient debug line

Jesse Barnes:

• ia64: fix NUMA page table allocation to get memory from the
  right node
• ia64: fix SN2 code for GENERIC builds
• ia64: fix is_headless_node() for systems w/<64p
• ia64: early SN setup fix
• ia64: deal with lack of SRAT in GENERIC kernels
• ia64: fix SN2 interrupt allocation
• ia64: 2nd step to fix GENERIC builds: split contig and
  discontig paging_init functions
• ia64: fix NUMA boot
• ia64: nable SN2 in generic builds
• ia64: SN2 module fix
• ia64: force on appropriate generic options

John Levon:

• USB: default input core support to y

Julian Anastasov:

• [IPV4]: ip_copy_metadata must copy the nfcache field
• [IPV4]: ip_fragment should not reassemble if all frags do not
  have owner

Kenneth W. Chen:

• ia64: fix NUMA hugetlb support

Krzysztof Halasa:

• [WAN]: Fix transmitted headers with generic HDLC + Cisco
  encap

Len Brown:

• [ACPI] Summary of changes for ACPICA version 20031002
• [ACPI] fix acpi_asus module build (Stephen Hemminger)

Linus Torvalds:

• disable_irq() should synchronize with irq handler only if one
  exists
• Revert floppy driver dependence on CONFIG_ISA. The thing
  exists
• Revert the process group accessor functions. They are buggy,
  and cause NULL pointer references in /proc.
• Make kernel stack dumps handle unaligned kernel stacks more
  gracefully
• Clear IRQ_INPROGRESS as part of “enable_irq()”
• Revert recent SMP timer changes – they cause deadlocks
• Mark timer as running in the timer base _before_ we clear the
  pending flag (and order it on SMP), so that del_timer_sync() always
  sees the timer either pending or running if it is active.
• Avoid unnecessary floating point usage in av7110 DVB
  driver
• Avoid unnecessary floating point calculations in amd8111e net
  driver
• Make x86 do_div() macro evaluate its “base” argument only
  once
• Fix APIC address handling at bootup and restore
• Allow Turtle Beach sound drivers to be compiled modular when
  STANDALONE is set, since the firmware is only required for the
  built-in case.

Luca Tettamanti:

• I2C: sensors/w83781d.c creates useless sysfs entries

Madarasz Gergely:

• [NET]: Fix get_random_bytes() call in
  sunhme.c:get_hme_mac_nonsparc()

Manfred Spraul:

• avoid crashes due to unaligned stacks
• [netdrvr natsemi] fix ring clean

Matt Domsch:

• EDD: reads to raw_data return binary data, not hexdump

Matthew Dharm:

• problems with USB memory pen

Matthew Wilcox:

• Build fixes for zoran on PA-RISC

Michael Hunold:

• New DVB frontend driver ves1x93 (obsoletes alps_bsrv2)
• Fix vbi handling in saa7146 core driver
• Add private data pointer to DVB frontends
• Fix DVB network device handling
• Misc. fixes for ALPS TDLB7 DVB frontend driver
• Misc. fixes for AT76C651 DVB frontend driver
• Update the AV7110 DVB driver
• Fix dvb_net network subsystem usage

Mikael Pettersson:

• ftape linkage error

Nathan Scott:

• ia64: correct ino_t to be 64 bits wide

Noah J. Misch:

• USB: Make ISD-200 USB/ATA Bridge depend on BLK_DEV_IDE
• USB: Make Ethernet Gadget depend on CONFIG_NET
• [IPX]: ipx_proc.c needs linux/init.h even when PROC_FS is not
  enabled

Oliver Neukum:

• USB: remove stupid check for NULL in devio.c

Patrick Mochel:

• [kobject] export kset syms for module use
• [driver model] Add device_unregister_wait()
• [kobject] Fix memory leak in kobject_set_name()
• [power] Clean up ACPI STR assembly
• [driver model] Remove unneeded error check
• [power] Fix some typos
• [power] swsusp Update
• [swsusp] Don’t free buffer when going to sleep

Paul Mackerras:

• PPC32: Export a couple of symbols (csum_partial and
  init_task)
• Fix compile on PPC
• USB: Fix USB suspend in 2.6.0-test6

Paul Mundt:

• net/sunrpc/clnt.c compile fix

Pekka Pietikäinen:

• b44 enable interrupts after tx timeout (2.6-test version)

Peter Chubb:

• ia64: drop stale check for GAS_HAS_LOCAL_TAGS

Petr Vandrovec:

• [NET]: Fix socket test in dev_queue_xmit_nit()

Randy Dunlap:

• [NET]: Grab RTNL semaphore in init_netdev() if we call
  dev_alloc_name()

Russell Cattelan:

• [XFS] Fix remount,ro path

Russell King:

• [ARM] Update LART default configuration
• [ARM] Clean generated files in “clean” rather than
  “mrproper”
• [ARM] Add ARM specific ignores to BK ignore file
• [ARM] Remove no_action function from CLPS7500 code
• [ARM] Add missing exports for Integrator logic module
  drivers
• [ARM] Add missing csum_partial export
• Fix pcnet_cs network hotplug
• [ARM] Fix div64 implementation

Rusty Lynch:

• ia64: fix perfmon typo that broke Itanium build

Rusty Russell:

• Fix pcmcia_tcic.c if PCMCIA_DEBUG is defined
• ipt_limit fix for HZ=1000
• More barriers in module code

Scott Feldman:

• ethtool_ops eeprom stuff
• hang on ZEROCOPY/TSO when hitting no-Tx-resources

Shirley Ma:

• [IPV6]: MIB fix, provide timestamps in ifa_cacheinfo

Simon Kelley:

• atmel wireless driver

Stephen Hemminger:

• [IRDA]: Fix memory leak in sa1100_ir.c
• [WAN]: Incorrect comparison for register_netdev in cosa.c
• [WAN]: Fix register_netdev() return value check in
  hostess_sv11.c
• [NET]: Fix register_netdev() return value check in
  synclink_cs.c
• [NET]: Fix syncppp wan device regression
• [NET]: Fix sealevel regression
• [NET]: Resolve device probe difference
• [netdrvr xircom_cb] fix race in statistics pointer setting by
  converting to use alloc_etherdev.
• [NET]: Mark dev_alloc as deprecated
• [NET]: Mark init_etherdev() as deprecated
• [NET]: Catch buggy net drivers changing getstats op after
  registry

Stephen Lord:

• [XFS] Implement deletion of inode clusters in XFS
• [XFS] Document ikeep mount option
• [XFS] switch xfs to use linux imode flags internally
• [XFS] Match up minor formatting changes between SGI and Linus
  trees
• Revert from using __kernelfsidt to
  fsid_t since ia64 now defines this correctly
• Change the way XFS implements the invisible I/O mechanism used
  by the online backup tools.
• More uio changes, add code attribution

Stéphane Eranian:

• ia64: perfmon-2 fixes

Tigran Aivazian:

• update to microcode update driver

Tom Rini:

• PPC32: Always print the processor number in /proc/cpuinfo
• PPC32: Remove CONFIG_DCACHE_DISABLE, as it’s not all that
  useful
• PPC32: Fix a typo
• PPC: Change how we export some Openfirmware device nodes
• PPC32: Fix the ns1655x uart driver in the bootwrapper
• PPC32: Update ‘make help’

Tony Luck:

• ia64: fix register numbers in MCA save/restore
• ia64: infinite loop in ia64_mcawakeupipi_wait
• ia64: cannot convert region 5 address to physical by clearing
  bits 63:61
• ia64: Another MCA fix
• ia64: MCA min_state area must be uncacheable

Trond Myklebust:

• NFS: fix the synchronous READ/WRITE bugs
• NFS: remove broken O_DIRECT wait code
• NFS: synchronous NFSv3/v4 COMMIT
• NFS: Fix O_DIRECT code
• NFS: Enable NFS_DIRECTIO support

Venkatesh Pallipadi:

• Bug in timer_tsc cpufreq callback

Wensong Zhang:

• [IPVS]: Fix IPVS sync state check
• [IPVS]: Fix to set the statistics of dest zero when bound to a
  new service

Yoshinori Sato:

• H8/300 support update

Yves Rutschle:

• [ARM] Fix LART build