Linus Torvalds: Linux 2.6.0-test5

Kernel.org Mirrors

Full Changelog

Lots of small stuff, as usual. I think the biggest "core" change is the Futex changes by Jamie and Hugh, and the dev_t preparations by Al Viro.

But there are ARM and ppc updates here too, and a few drivers have bigger fixes (tg3 driver and the USB gadget interface stand out on diffstat). Watchdog driver updates etc. And Russell King fixed more PCMCIA issues.

Linus

---

Summary of changes from v2.6.0-test4 to v2.6.0-test5

Adrian Bunk:

• [netdrvr sis190] fix build with older gcc
• [NET]: Fix bpqether build with procfs disabled
• Mark more drivers BROKEN{,ON_SMP}
• [wireless airo] fix build with gcc 2.95
• COSA is no longer BROKEN

Alan Stern:

• USB: root hub polling stops after suspend
• USB: Another unusual_devs.h entry update
• USB: storage: Revised update to isd200 I/O buffer patch

Alex Williamson:

• ia64: no discontig w/o NUMA

Alexander Viro:

• dev_t handling cleanups (1-12)
• large dev_t - second series (1-15)

Andi Kleen:

• IOMMU overflow handling fix for MPT fusion
• Make ACPI_SLEEP select SOFTWARE_SUSPEND
• Do 32bit addresses in /proc/self/maps if possible
• x86-64 update
• x86-64 update

Andrew Morton:

• .config checks updated
• random: SMP locking
• random: accounting and sleeping fixes
• disable prefetch on athlons
• fix /proc/pid/fd ownership across setuid()
• Call security hook from pid*_revalidate
• move DAC960 GAM IOCTLs into a new device
• Add the kernel janitors to MAINTAINERS
• Update ide.txt documentation to current ide.c
• v4l use-after-free fix
• ikconfig - Makefile update
• Fix ftape warning
• jffs aops return type fix
• Add 3GB personality
• zeromap_pmd_range bugfix
• don't report async write errors on close() after all
• remove add_wait_queue_cond()
• spin_lock_irqrestore() typo fixes
• zoran: memleak fixes
• zoran: debug->zr_debug
• zoran: add release callback
• zoran: add pci_disable_device() call
• zoran: cleanups
• zoran: more cleanups
• zoran: correct name field breakage
• airo CONFIG_PCI=n build fix
• drivers/char/pcxx.c warning fix
• pcnet32 needs unregister_pci
• c99 struct initialiser conversions
• Fix 'pci=noacpi' with buggy ACPI BIOSes
• /proc/kallsym caching fix
• Fix permissions on /proc/kallsyms
• Kobject doc addition
• vm_enough_memory microoptimisation
• abi doc update
• ni5010.c: remove cli/sti
• do_no_page() fix
• parport_pc rmmod oops fix
• reiserfs writepage-versus-truncate fix
• visws: fix 2.6.0-test4 breakage
• Fix ext3 htree corruption on big-endian platforms
• Fix selinux_file_fcntl
• Fix SELinux avtab
• Fix SELinux format specifiers
• Rework SELinux binprm hooks
• Fix typo in #ifdef for ext2 xattr support
• Add more bad_inode operations
• Fix build with CONFIG_KCORE_AOUT
• kill CONFIG_KCORE_AOUT
• knfsd nfs4 warning fixes
• Fix bluetooth compile warnings
• do_no_page() rss accounting fix
• jbd: remove uninformative printk
• acpi pci_link fix
• add context switch counters
• remove size_t-based printk warnings
• large dev_t 12/12 oops fix
• evdev_ioctl does not report EV_MSC capabilities
• AS: don't anticipate against a task's initial I/O
• hch has moved
• Cyclades ISA serial driver fix
• kbuild: warn if the user has old modutils
• fix arcnet printk parameter types
• floppy driver cleanup
• Use tgid rather than pid in dnotify
• Fix a few declarations
• make voyager work again after the cpumask_t changes
• mtrr cleanups
• compat ioctl_table fix
• raw driver oops fix
• ipc_init() uses vmalloc too early
• vmscan: zone pressure calculation fix
• vmscan: zone pressure simplification and fix
• Remove SSE2 bugs.h check
• HPET 1/6: Support for HPET based timer
• HPET 3/6: makefile and config changes
• HPET 4/6: Core
• HPET 5/6: timer services
• HPET 6/6: rtc emulation
• HPET 2/6: boot parsing
• fix advansys.c if !CONFIG_PROC_FS
• handle setup_swap_extents() error in swapon
• scsi_unregister() oops fix
• tty oops fix
• ext3_setxattr() oops fix
• Add documentation for /proc/stat
• [NET]: Fix 64-bit warnings in af_netlink.c
• misc fixes
• Fix odd code in bio_add_page
• convert /proc/stat to seq_file
• Fix rtc symbol clash and HPET config problems
• add config option for qla1280 SCSI MMIO/ioport
• elevator insertion fixes
• 8250_acpi taints kernel
• proc_misc.c needs irq.h
• more slab page checking
• might_sleep() improvements
• MODULE_ALIAS() in block devices
• MODULE_ALIAS() in char devices
• Remove percpufication of in_flight counter in
• Enable SELinux via boot parameter
• devfs pty fix
• i8042 free_irq() aliasing fix
• Remove Documentation/kmod.txt
• drivers/scsi/imm.c build fix
• hermes.h fails with outw_p() in :?
• cciss error handling cleanup
• MODULE_ALIAS for tty ldisc
• fix /proc/stat handler for ARM, SPARC64, others
• Fix /proc/stat off-by-one

Andries E. Brouwer:

• more keyboard stuff

Andy Molloy:

• USB: Aten 4 Port USB 2.0 KVM C (ACS-1724)

Anton Blanchard:

• sym2 hotplug fix

Arjan van de Ven:

• incomplete asm constraints in arch/i386/pci/pcbios.c

Arnaldo Carvalho de Melo:

• cyc2x: sanitize ioremap usage & more
• scsi: remove include procfs_h from hosts.h
• scsi cleanups

Arnd Bergmann:

• Verify proper usage of ioctl macros

Arun Sharma:

• ia64: fix ia32 execve memory leak

Bart De Schuymer:

• [BRIDGE]: Add arpreply EBTABLES target
• [BRIDGING]: Update Kconfig files for bridging firewall
• [BRIDGE]: Add 802.3 filtering support
• [BRIDGE]: Create CONFIG_BRIDGE_NETFILTER and use it instead of messy tests
• [NETFILTER]: Use CONFIG_BRIDGE_NETFILTER in ipt_REJECT.c

Bartlomiej Zolnierkiewicz:

• cable detection fixes for HPT37x controllers
• fix PowerMac driver breakage caused by recent dynamic queue change
• fix ide.c warning when compiling IDE for non-PCI systems
• fix ide-lib.c warning when compiling IDE without DMA support
• allow drivers (ie. mediabay) to set hwif->gendev.parent
• kill ide_modes.h
• do not set drive->dn twice in probe_hwif()
• kill ide_init_drive() in ide-probe.c
• remove unused exports from ide-probe.c
• remove unused ide_chipsets and IDE_CHIPSET_MODULE
• kill ide_module_t
• kill ide_register()
• ide: fix ide_unregister() vs. driver model
• ide: forward-port siimage driver changes from 2.4.22
• ide: allow LBA48 on Promise 20265
• ide: add very basic support for VIA 8237 SATA controller
• ide: enable LED support for PowerMac
• ide: fix PM with ide-default driver
• ide: remove supports_dma field from ide_driver_t
• ide: fix ide_cs oops with TCQ
• ide: fix imbalance preempt count with taskfile PIO

Ben Collins:

• Update IEEE1394 (r1047)
• Fix compile for raw1394

Benjamin Herrenschmidt:

• cputable.c
• Add new OF tree walking APIs
• Update OF platform & macio driver cores to adapt to device model changes. Fix refcounting
• Update openpic to expose a sys_dev for power management, make it more robust vs. concurrent calls by the PM system and cpufreq
• Update pmac PIC driver to register a sysdev for Power Management
• Major update via-pmu driver, hopefully last before we split it & do major cleanup
• Update PowerMac IDE driver. Adapt to new driver model, add proper support for Kauai ATA/100 and add activity led code.
• Adapt PowerMac i2c-keywest driver to new driver model
• Fix PowerMac ALSA build with device model "name" field change
• Update PowerMac mediabay driver to new model, fix an old bug that could prevent one of the timeouts from working, fix access to MMIO based interface
• Adapt PowerMac "airport" driver to new driver model
• Fix build of controlfb driver
• Adapt PowerMac "platinum" video driver to new driver model
• fixup xmon ADB polling so that it works before ADB core is loaded
• Add back missing fb_set_var to PowerMac platinum driver
• Update PowerMac cpufreq driver to adapt it to some core changes and fix a race with the PMU driver
• For keeping interface ordering consistent between previous kernels and the new driver model probing mecanism, drivers/macintosh has to be linked before ide and scsi
• Don't care about driver registration results for i2c-keywest so failing one don't break the other
• Fix drivers/video Makefile so control & platinum drivers gets proper depedencies on the cfb* files
• Add new pmac_zilog serial driver, obsolete old macserial
• Update "coff" zImage wrapper so it works with larger kernel images
• Fix missing bit in the new .coff wrapper
• some whitespace & tab fixes
• Fix a bug where an ide-pmac hwif returned to the system because it's empty would still be probed thus causing a crash on some machines. Also fix some whitespace/tabs.
• Add & export some routines to access the i2c busses that hang off the PMU, not yet linked to the linux i2c subsystem though. Fix some whitespace/tabs too.
• C99 initializer fixes
• Remove useless junk at beginning of MachineCheck exception handler, this actually is causing problems on some CPUs
• PowerMac: Fix build of via-pmu driver with some .config's

CaT:

• USB: C99: 2.6.0-t3-bk7/Documentation

Chas Williams:

• [ATM]: Clean up the code making use of sti/cli (from vinay-rc@naturesoft.net)
• [ATM]: In ambassador driver, use del_timer_sync instead
• [ATM]: In atm_getaddr() do not copy_to_user() with locks held
• [ATM]: Convert the /proc/net/atm/br2684 to seq_file interface (from shemminger@osdl.org)
• [ATM]: pvc/svc missing .owner for proto_ops/family (from levon@movementarian.org)
• [ATM]: fix atm_dev module refcount bug (from levon@movementarian.org)
• [ATM]: reduce CONFIG_PROC_FS #ifdef clutter in .c code (from levon@movementarian.org)

Christoph Hellwig:

• make scsi_priv.h includable standalone
• make scsi logging level a sysctl
• make /proc/scsi/scsi/ support optional
• don't export proc_scsi
• add a missing extern to scsi_priv.h
• serialize bus scanning
• fixup some tagged queuing mess
• give scsi_allocate_request a gfp_mask
• kill an unused variable in sym2
• kill some dead code in sym2
• check whether a disk got writeable in sd_open
• [IPV6]: Use per-cpu data for icmp sockets
• [NET]: Convert netdev_rx_stat to per-cpu data
• [NET]: Remove reference to CONFIG_IA64_SGI_SN1, it is gone
• fix the scsi_logging_level fix

Daniel Ritz:

• [PCMCIA] Add ToPIC97 and ToPIC100 support

Daniele Bellucci:

• USB: Audit usb_register/usb_serial_register under divers/usb/serial
• USB: CREDITS file update
• Audit and minor cleanups in drivers/usb/*
• ...more usb audit
• Another bad usb_register audit: dvb-ttusb-budget
• Another bad audit in drivers/usb/*: usblp
• Another bad audit in drivers/usb/*: cdc-acm
• Another bad audit in drivers/usb/*: usbskeleton
• Audit and minor cleanups in usbnet
• Audit and minor cleanups in usbstorage

Dave Jones:

• [AGPGART] Fix up ATI's identity crisis
• [DRM] Update MAINTAINERS
• [AGPGART] Update VIA PCI IDs
• [AGPGART] Numerous AMD64 gart driver cleanups
• [AGPGART] Fix indentation
• [AGPGART] Use generic AGP_APBASE define instead of per vendor _APBASE
• [AGPGART] move NVIDIA registers to agp.h
• [AGPGART] Indentation fixes
• [AGPGART] Fix missed AGP_APBASE conversion in VIA AGP driver
• [AGPGART] Remove unneeded 8151 defines
• [AGPGART] Make AMD64 GART driver marchitecture compliant
• [AGPGART] Various renames for AMD64 GART driver
• [CPUFREQ] Move kernel/cpufreq.c to drivers/cpufreq/cpufreq.c Also remove $Id$ tag.
• [CPUFREQ] fix up two typos
• [CPUFREQ] Move drivers/cpufreq/userspace.c to drivers/cpufreq/cpufreq_userspace.c Module names of 'userspace' aren't very informative.
• [CPUFREQ] Fix various oddities in the userspace governer
• [CPUFREQ] Completely separate governors from policies
• [CPUFREQ] Add the "performance" and "powersave" governors as modules
• [CPUFREQ][PPC] Small fixes necessary to separate the governors from policies
• [CPUFREQ][SH] Small fixes necessary to separate the governors from policies
• [CPUFREQ][SPARC64] Small fixes necessary to separate the governors from policies
• [CPUFREQ][X86] Small fixes necessary to separate the governors from policies
• [CPUFREQ][ARM] allow for easier Kconfig usage on ARM, and more features for SA11x0 users
• [CPUFREQ] Fix incorrect entry in Kconfig
• [CPUFREQ] new cpufreq_driver->resume callback, needed (at least) for speedstep-smi
• [CPUFREQ] fix speedstep-ich's .name to reflect its real name
• [CPUFREQ] Inform user about broken powernow-k7 PST tables
• [AGPGART] Fix up compile for i460-agp Missed conversion when killing off the APBASE defines.
• [CPUFREQ] Cache FSB in longhaul driver
• [CPUFREQ] Fix use of fsb before initialisation in longhaul
• [CPUFREQ] Fix longhaul's mult,fsb -> MHz conversions
• [CPUFREQ] Clean up clock_ratio calculation
• [CPUFREQ] Inform user of status of Ezra-T/Nehemiah longhaul support
• [CPUFREQ] Remove the voltage scaling from longhaul driver
• [CPUFREQ] Move longhaul scale/ratio tables to longhaul header file
• [CPUFREQ] Print out FSB in longhaul debug info
• [CPUFREQ] Print out CPU name in debug info
• [AGPGART] Remove unneeded string from AMD64 GART driver
• [AGPGART] Fix up missing brackets on defines

Dave Kleikamp:

• New version of jfsutils needed

David Brownell:

• USB: usb_sg_cancel() + disconnect, fewer messages
• USB: Add Kconfig option for building ax8817x support in usbnet
• USB: usb_new_device() shouldn't be exported
• USB: <linux/usb_gadget.h> minor doc updates
• USB: usbnet minor cleanup
• USB: net2280 fixes: ep halt, sysfs
• USB: usbnet, cdc ethernet descriptor parsing fixes
• USB: ohci -- reset, fault recovery
• USB: uhci-hcd, add uhci_reset()
• USB: net2280, patch dma chains
• USB: net2280 one-liner
• USB: usb hcd states
• USB: usb "gadgetfs" (1/2)
• USB: usb "gadgetfs" (2/2)
• USB: usb_epnumtoep_desc only look

David Jeffery:

• ips: remove arch limitations

David Mosberger:

• ia64: Manual merge with Alex's "UP cmc/cpe polling fix" patch
• ia64: Use offset_in_page() instead of equivalent open code
• ia64: Hook up fadvise64_64() system call
• ia64: The second chunk of the "UP cmc/cpe polling fix" seems to have gotten lost. Please apply the attached for the cpe side of the fix.
• ia64: Fix usage ("corrected" machine checks and platform errors, not "correctable").

David S. Miller:

• [TG3]: Initial implementation of 5705 support
• [TG3]: Fix statistics on 5705
• [TG3]: Do not reset the RX_MAC unless PHY is Serdes
• [TG3]: More missing PCI IDs
• [TG3]: Reset PHY more reliably on 570{3,4,5} chips
• [TG3]: Fix 5788/5901, update TSO code
• [TG3]: Differentiate between TSO capable and TSO enabled
• [ETHTOOL]: Add {G,S}TSO support to ethtool_ops
• [TG3]: Add {get,set}_tso ethtool_ops support
• [TG3]: Bump version/reldate
• [TG3]: Fix tg3_phyreset_57034_5 chip rev test
• [TG3]: Bump version/reldate
• [ETHTOOL]: Add ethtool_op_{set,get}_tso helpers
• [TG3]: More fixes and enhancements
• [SPARC64]: Add some missing PCI error reporting
• [SPARC]: Update ethtool support in Sun net drivers
• [NETFILTER]: Use correct printf format for size_t in ipt_CLASSIFY.c
• [NET]: net/core/ethtool.c needs asm/uaccess.h
• [TG3]: Fix ethtool_ops/sun_5704 changes collision
• [TG3]: Protect get/set TSO support with proper ifdefs
• [SPARC]: Add missing timer_create syscall entries
• [SPARC64]: Make sure init_irqwork_curcpu() is called with PSTATE_IE off
• [IPV6]: Do not mistakedly use ndisc route for normal ipv6 output
• [POSIX_TIMERS]: Do not assume timeval/timespec layout is identical
• [SPARC64]: In sysv IPC translation, mask out IPC_64 as appropriate
• [SPARC]: Mark get_rtc_time() static in SBUS rtc driver
• [IPV6]: Do not BUG() on icmp6 socket contention, just drop
• [IPV6]: Fix typo in icmp BUG() fix
• [IPV6]: Fix types in fl6_renew()
• [IPV6]: linger member of ip6_flowlabel needs to be a long
• [IPV6]: Fix printf format in ip6fl_fl_seq_show
• [BLUETOOTH]: Fix typo in module alias changes
• [IPV4]: Do not BUG() on icmp_xmit_lock() contention, just drop
• [NET]: Kill NET_PROFILE, has not built for years
• [USB]: hiddev_exit() can no longer be __exit, called from init code now
• [SPARC]: Add MODULE_ALIAS_LDISC() defines
• [NET]: Do not ifdef declarations in Space.c
• [NET]: Remove all the silly 'NET4.x' init messages
• [NET]: Print a KERN_INFO msg when protocol families are {un,}registered
• [NET]: Kill more verbose init msgs and unused RTNL_DEBUG define

David T. Hollis:

• USB: Add ax8817x support to usbnet
• USB: Fix building of ax8817x if CONFIG_USB_AX8817X_STANDALONE

Deepak Saxena:

• [ARM PATCH] 1611/1: Add big-endian support to AFLAGS
• [ARM PATCH] 1615/1: Fix IOP3xx timer interrupts
• [ARM PATCH] 1613/1: arch/arm/boot/Makefile fixups for IOP3xx and ADIFCC
• [ARM PATCH] 1616/1: Add PFN_TO_NID to IOP3xx
• [ARM PATCH] 1621/1: IOP3xx CPU detection (cleaned up)
• [ARM PATCH] 1623/1: Updated def-configs for IQ80310/321
• [ARM PATCH] 1620/1: dma_map_single/unmap_single support for ARM
• [ARM PATCH] 1559/1: updated include/asm-arm/checksum.h big-endian support

Douglas Gilbert:

• GFDL issue in Documentation/DocBook/scsidrivers.tmpl

Duncan Sands:

• USB: fix uhci "host controller process error"

Erik Andersen:

• Fix cdrom error handling in 2.6

Ernst Persson:

• [netdrvr] list CONFIG_BMAC in drivers/net/Makefile.lib, as it uses the crc32 library.

Felipe Damasio:

• [NETFILTER]: Remove unneeded version.h inclusion
• [SUNRPC]: Remove unneeded version.h inclusion
• [RXRPC]: Remove unneeded version.h inclusion
• [IPV6]: Remove unnecessary linux/version.h include

Frank Becker:

• [ARM PATCH] 1563/1: Update pxa-regs.h with correct gpio number for 48 MHz clock output

François Romieu:

• [netdrvr sis190] pass irq argument to synchronize_irq()
• [netdrvr sis190] remove unneeded alignment code, other small fixes
• [netdrvr sis190] use PCI DMA API for RX buffers
• sis190 driver fix
• [NET]: Balance alloc_netdev() with free_netdev() in ethertap
• [NET]: Use free_netdev() even in error paths

Gary Algier:

• USB: new ids for io_ti driver

Geert Uytterhoeven:

• vmlinux-*.lds (was: Re: Linux 2.6.0-test4)
• macide (was: Re: Linux 2.6.0-test4)
• m68k asm/sections.h
• m68k asm/local.h
• Amiga z2ram
• Amiga floppy
• M68k switch_to
• Atari floppy
• dmasound core fixes
• dmasound kill MOD_{IN,DE}C_USE_COUNT

Greg Kroah-Hartman:

• USB: remove #include <linux/miscdevice.h> from some usb drivers
• USB: change pci host drivers to use PCI_DEVICE_CLASS() macro
• USB: remove proc code from stv680 driver as it's no longer needed
• USB: convert stv680 driver to handle the driver core changes in the v4l layer
• USB: add sysfs files for stv680 driver, replacing the lost proc functionality
• V4L: fix use after free bug in v4l core
• V4L: add video_device_remove_file() to match video_device_create_file()
• [netdrvr sis900] don't call pci_find_device from irq context
• USB: fix compiler warning in mdc800 driver
• USB: fix up a bunch of copyrights that were incorrectly declared
• USB: hook up the USB driver core to the power management calls of the driver model
• USB: rip out old proc code from the usbvideo driver
• USB: removed the proc code from the se401.c driver
• USB: add support for 2 new devices to the visor driver
• USB: fix usbnet for older versions of gcc
• USB: fix oops in keyspan and whiteheat devices when plugged in
• USB: remove usage of DEVICE_ID_SIZE from usb core as it should not be used
• USB: fix data toggle problem for pl2303 driver
• USB: fix up B0 support in the pl2303 driver

Greg Ungerer:

• use irqreturn_t in m68knommu/5206 config.c
• use irqreturn_t in m68knommu/5206e config.c
• use irqreturn_t in m68knommu/5249 config.c
• use irqreturn_t in m68knommu/5272 config.c
• irqreturn_t fixes for m68knommu irq.h
• create an m68knommu local.h
• create an m68knommu sections.h

Guillaume Morin:

• fix cu3088 group write

Harald Welte:

• [NETFILTER]: Fix ipt_REJECT if used on bridge
• [NETFILTER]: Remove ipt_MIRROR target from 2.6.x
• [NETFILTER]: Remove ipt_unclean match from 2.6.x
• [NETFILTER]: Remove EXPERIMENTAL mark from some netfilter stuff
• [NETFILTER]: Cosmetic netfilter patch
• [NETFILTER]: NAT optimization
• [NETFILTER]: Conntrack optimization (LIST_DELETE)
• [NETFILTER]: New iptables modules (iprange, CLASSIFY, SAME, NETMAP)
• [NETFILTER]: Fix ipt_helper build problem wrt. Kconfig
• [NETFILTER]: Fix email address in MODULE_AUTHOR
• [NETFILTER]: NAT range calculation fix

Herbert Xu:

• free_netdev typo
• [NET]: Add MODULE_LICENSE to xfrm_user.c

Hideaki Yoshifuji:

• [NET]: Fix OOPS in multicast procfs usage
• [IPV{4,6}]: Fixing a bug that reading /proc/net/{udp,udp6} may drop some data

Hirofumi Ogawa:

• [netdrvr 8139too] lwake unlock fix
• [netdrvr 8139too] remove unused RxConfigMask
• [netdrvr 8139too] add more h/w revision ids
• [netdrvr 8139too] remove driver-based poisoning of net_device
• [netdrvr 8139too] don't start thread when it's not needed
• vfat_valid_longname() cleanup

Holger Freyther:

• [ARM PATCH] 1595/1: [PATCH] 1/10 Simpad changes
• [ARM PATCH] 1598/1: [PATCH] 4/10 Simpad changes
• [ARM PATCH] 1603/1: [PATCH] 9/10 Simpad changes

Hugh Dickins:

• Fix futex hashing bugs

Ian Abbott:

• USB: ftdi_sio - fix memory leak and tidy up write bulk callback

Ian Molton:

• arm26 updates

Jakub Jelínek:

• [COMPAT]: Add missing set_fs() calls to {clock,timer}_*() handlers
• [SPARC64]: Fix struct sigevent32
• [SPARC64]: sys_timer_create needs 32-bit translation
• [SPARC]: Fix typos

James Bottomley:

• Fix typo introduced into 53c700 by tag fixup patch
• Add extern for scsi_logging_level so scsi_sysctl.c can compile
• Fix up scsi_alloc_request in sr.c to take a GFP_ flag
• fix remap of shared read only mappings

Jamie Lenehan:

• dc395x [1/6] - make functions static
• dc395x [2/6] - cleanup devices
• dc395x [3/6] - cleanup adapter init
• dc395x [4/6] - cleanup adapter uninit
• dc395x [5/6] - check for device
• dc395x [6/6] - use pci resource len

Jamie Lokier:

• Unpinned futexes v2: indexing changes
• Common PROT_xxx -> VM_xxx mapping

Javier Achirica:

• [wireless airo] add support for MIC and latest firmwares
• [wireless airo] build fix when MIC support is disabled

Jean Tourrilhes:

• [irda] IrCOMM mod refcount
• [irda] NSC 3839x probe fixes
• [irda] irtty cleanup
• [irda] LAP close race
• [irda] connect watchdog fixes
• [irda] init failure cleanups
• [irda] Dongle module aliases

Jeff Garzik:

• [TG3]: Bug fixes for 5705 support
• [TG3]: More 5705 updates
• [TG3]: More 5705 fixes
• [TG3]: Another 5705 fix: enable eeprom write prot as needed
• [TG3]: Only write the on-nic sram addr on non-5705
• [TG3]: Add 5782 pci id
• [netdrvr sis900] ethtool_ops support
• [netdrvr 8139cp] ethtool_ops support
• [netdrvr sis190] convert TX path to use PCI DMA API
• [netdrvr sis190] make driver depend on CONFIG_BROKEN
• [netdrvr 3c501] ethtool_ops support
• [netdrvr] ethtool_ops support in 3c503, 3c505, 3c507
• [netdrvr] ethtool_ops support for 3c515, 3c523, 3c527, and dmfe
• [netdrvr pcmcia] ethtool_ops for 3c574, 3c589, axnet
• [netdrvr pcmcia] convert several drivers to ethtool_ops
• [netdrvr xircom_cb] ethtool_ops support
• [wireless ray_cs] ethtool_ops support
• [netdrvr sis190] small bug fixes
• [TG3]: Remove pci-set-dma-mask casts
• [netdrvr 8139cp] build TX checksumming code, but default OFF
• [netdrvr 8139cp] support NAPI on RX path; Ditch RX frag handling
• [netdrvr 8139cp] update todo list in header
• [netdrvr 8139cp] remove mentions of RTL8169 (now handled by "r8169")
• [netdrvr 8139cp] small cleanups
• [netdrvr 8139cp] fix NAPI bug; remove board_type distinction, not needed
• [netdrvr 8139cp] bump version
• [netdrvr 8139cp] stats improvements and fixes
• [netdrvr 8139too] make features more persistent; fix PCI DAC mode
• [netdrvr pcmcia] support SIOC[GS]MII{PHY,REG} ioctls
• [netdrvr 8139too] remove useless board names
• [netdrvr 3c509] dev->name removal build fix
• [netdrvr 8139cp] must call NAPI-specific vlan hook
• [netdrvr ixgb] must call NAPI-specific vlan hook
• [netdrvr sk_mca] somebody typo'd in their cli()-to-spinlock conversion
• [netdrvr sk_mca] remove ancient-kernel compat code; fix bugs
• [netdrvr 8139cp] PCI MWI cleanup; remove unneeded workaround
• [netdrvr de2104x] ethtool_ops support
• [PCI] Remove cases where PCI_CACHE_LINE_SIZE is hardcoded
• [tokenring lanstreamer] clean up MWI / PCI_CACHE_LINE_SIZE usage
• [video planb] don't hardware pci command/cacheline/latency values, use the PCI layer instead to provide those for us.
• [netdrvr 8390] new function alloc_ei_netdev()
• [netdrvr ne2k-pci] allocate netdev+8390 struct using new alloc_ei_netdev()
• [netdrvr ne2k-pci] ethtool_ops support
• [NET] move netif_* helpers from tg3 driver to linux/netdevice.h
• [netdrvr] ethtool_ops for epic100, fealnx, winbond-840, via-rhine
• [netdrvr fealnx] merge typo build fix (non-x86) from 2.4
• [NET] move ethtool_opsettx_csum from 8139cp drvr to net/core/ethtool.c, where it belongs.
• [PCI, ia32] don't assume "c->x86 > 6" applies to non-Intel CPUs when programming PCI cache line size.
• [SUNHME]: Fix non-sbus build
• Fix non-modular compile of 3c515.c
• [NET]: Fix ethtool_ops thinko in sungem.c

Jens Axboe:

• cciss init problem
• software hd led support
• IO scheduler, not elevator
• amiflop error handling
• fix IO hangs
• Fix noop elevator request merging

John Levon:

• OProfile: correct CPU type for x86-64
• [NET]: Kill net/README, obsolete and out-of-date
• [ATM]: Remove bogus UNUSED macro usage in pppoatm.c

Joris Struyve:

• unusual_devs.h entry

Jürgen Quade:

• USB: writing usb driver documentation update
• USB: usb-skeleton bugfix

Karsten Keil:

• ISDN bugfixes part 1
• next fixes

Krishna Kumar:

• [IPV4]: Fix creat_proc_read_entry() args
• Remaining task queue to work queue conversion

Linus Torvalds:

• Input: typo in device matching
• Don't claim exclusive ownership of the device when doing the SG_IO and SCSI_IOCTL_SEND_COMMAND ioctl's. That just screws things up when the drive is mounted.
• Fix del_timer_sync() SMP memory ordering (from Tejun Huh <tejun@aratech.co.kr>)
• Be a lot more careful about TS_USEDFPU and preemption
• Fix keyboard double E0-sequence release case
• Add the isicom serial driver to the list of drivers that are broken on SMP (due to expecting global irq locking).
• Instead of asking for "broken drivers", ask for a "clean compile"
• Avoid a negative in config questions: we don't want to have users forced into double negatives.
• Fix mprotect() to do proper PROT_xxx -> VM_xxx translation
• Fix PCMCIA typo (extra 'R') that broke the compile
• Fix over-eager ioctl number fix. _IOC() does want the sizeof()
• Undo static on ide_probe_for_pdc4030: it can (and will) be called from the IDE init code when compiled-in.
• Mark drivers that can't be built stand-alone in the configuration files.
• Fix various scripts to be marked executable
• Fix more ioctl _IOR/_IOW misusage
• Arnd's new-and-improved _IOx() argument checking showed two sound drivers whose microcode load ioctl's used types with sizes that

Marc Zyngier:

• [netdrvr de4x5] big modernization / cleanup

Marcelo Abreu:

• [NET]: Remove dead comment from dummy.c driver

Martin Diehl:

• [IRDA]: vlsi_ir v0.5 update, 1-7

Martin Hicks:

• ia64: paddr_to_nid fixup

Matthew Dharm:

• sr.c should issue TEST_UNIT_READY

Matthew Wilcox:

• ia64: default to building compressed
• ia64: ia64/lib/Makefile: use call-if-changed
• [netdrvr 3c59x] ethtool_ops support
• [ethtool] fix ethtool_get_strings counting bug
• [netdrvr 8139too] ethtool_ops support
• bio.c: reduce verbosity at boot
• ncr & sym1 patches
• sym2 patchset
• use size_t for the broken ioctl numbers
• CONFIG_64BIT

Matthias Bruestle:

• USB: Cyberjack patch

Maximilian Attems:

• [NET]: Use list_for_each() where applicable

Michel Dänzer:

• USB: linuxppc-2.5 fixlets for usbtest.c

Mike Anderson:

• fix Kernel Panic in scsi_host_dev_release

Miles Bader:

• Give v850 its own version of the vmlinux.lds.h RODATA macro
• Properly export symbols that depend on CONFIG_MMU

Mitchell Blank Jr.:

• [ATM]: Lanai driver updates

Neil Brown:

• Fix module ref counting for md
• Honour the read-ahead for for reads in raid5
• Set max_sectors for raid0 only, not for all raid levels
• Fix md superblock incompatabilities with 2.4 kernels
• Track nfsv4 open files by "struct inode" rather than dev/ino/generation
• fix in NFSv4 server for bad sequence id errors
• Fix compile errors in NFSv4 server

Nicolas Pitre:

• [ARM PATCH] 1565/1: syscall macros clobbering returned error value

Olaf Hering:

• USB: io_edgeport.o differences in 2.4 vs. 2.6

Pat LaVarre:

• USB: storage: cbw/csw trace in order

Patrick Mochel:

• [sysfs] Fix memory leak
• [power] Turn off debugging
• [kobject] Support unlimited name lengths
• [sysfs] Use kobject_name() when creating directories for kobjects
• [driver model] Use kobject_set_name() when registering objects
• [kobject] Don't use kobject->k_name after it's been freed
• [driver model] Add exports for sys devices
• [sysfs/kobject] Update documentation

Paul Fulghum:

• 2.6.0-test4 synclink.c
• 2.6.0-test4 synclinkmp.c
• 2.6.0-test4 synclink_cs.c

Paul Mackerras:

• PPC32: Define MCA_bus_isa_macro. From Christoph Hellwig
• PPC32: Add the fadvise64_64 system call
• PPC32: Declare cpu_onlinemap and cpu_possiblemap as cpumask_t
• PPC32: Update some of the example configs
• PPC32: Add support for the PPC 440 family of embedded processors
• PPC32: Eliminate one use of struct device name field

Paul Mundt:

• [netdrvr 8139too] fix and pci ids needed for SH platform

Paul Thompson:

• [NET]: Fix probing messages in 3c509.c

Pete Zaitcev:

• [SPARC]: Add pci_{map,unmap}_page()

Randy Dunlap:

• imm driver needs scsi_unregister()
• advansys build with ADVANSYS_DEBUG defined
• ia64: fix printk type warning
• [SCTP]: Fix printf format string
• [IPVS]: Fix printf format strings
• [HAMRADIO]: Missing return statement in yam.c driver
• USB: fix printk parameter types
• USB: fix functions to match prototypes
• label needs statement following it
• [NET]: remove duplicate #includes in net/
• [CRYPTO]: remove duplicate #includes in crypto/
• janitor: remove unneeded version.h #includes
• janitor: add static to comx
• janitor: sc520_wdt
• janitor: oss/ali copy*user fixes
• janitor: oss/ite8172 copy*user fixes
• janitor: skfddi copy*user fixes
• janitor: remove __SMP__
• janitor: jffs2 add/delete version.h
• janitor: adfs: add/remove version.h
• janitor: fix oss/harmony copy*user
• janitor: cdrom module owner
• janitor: fix input serport register failure
• janitor: fix blk_init_queue() comments
• janitor: fix oss/swarm copy*user
• janitor: saa7134 pci alloc/free consistent checking
• janitor: coda delete version.h
• janitor: oss/au1000 copy*user fixes
• janitor: fix copy*user in tc/zs
• janitor: more init/exit cleanups
• janitor: oss/forte copy*user fixes
• janitor: paride: better return codes
• janitor: qla1280 pci alloc/free consistent checking
• janitor: x86_64/sys_ia32
• rename make check* targets, add versioncheck
• jffs2: add linux/version.h as needed

Randy Hron:

• USB: version.h cleanup 1-4

Rob Radez:

• [SPARC]: Two build fixes

Russell King:

• [ARM] Noddy indentation fix for arch/arm/boot/Makefile
• [ARM] Fix vmlinux linker script
• [ARM] Remove reference to struct device name element
• [ARM] Fix device suspend/resume calls
• [ARM] Fix ecard.c manufacturer and product files
• [ARM] Tweak the bridge control register for PCI and cardbus bridges
• [ARM] Remove pci_dev->dev.name in favour of pci_name()
• [ARM] Remove old binutils compatibility
• [ARM] Update AMBA suspend/resume model
• [ARM] Update SA1111 suspend/resume model
• [ARM] Fix EBSA285 CLOCK_TICK_RATE
• [PCMCIA] Use #define'd constants in ZV code where possible
• [PCMCIA] Clean up yenta overrides
• [PCMCIA] Move socket initialisation to the quirk table
• [PCMCIA] Add generic and per-controller power management handling
• [PCMCIA] Move PM restore from socket initialisation
• [PCMCIA] Put socket initialisation to where it should be
• [PCMCIA] Move more controllers to the more advanced quirks
• [PCMCIA] Don't add CIS cache entries on failure
• [ARM] Remove more 26-bit ARM support
• [ARM] Remove more reminants of 26-bit ARM support
• Don't #ifdef prototypes
• [ARM] arch/arm/kernel/setup.c needs to include asm/cacheflush.h
• [ARM] Fix wrong cache flush call for ARM1020 CPUs
• Move MODULE_ALIAS_LDISC to tty_ldisc.h
• [ARM] Newer binutils want -mcpu=xscale not -mxscale
• [ARM] Restore preempt count before reporting unbalanced preempt count
• [ARM] Don't read the CPU control reg back - it may be write only
• [ARM] Fix ARM suspend-to-RAM
• [ARM] Fix PXA and SA1100 suspend/resume
• [ARM] Don't sleep in cpufreq code if IRQs are disabled (during resume.)
• [ARM] Kill snprintf formatting warning
• [PCMCIA] Fix cs.c debugging
• [PCMCIA] Use "yenta" instead of pci_name() when allocating irq
• [PCMCIA] Fix race condition causing cards to be incorrectly recognised
• Fixes to allow ARM to build in the standard tree
• [SERIAL] Make SA11x0 serial driver build

Rusty Russell:

• [NETFILTER]: Trivial 2.6 tftp conntrack fix
• [NET]: Use MODULE_ALIAS() in network families
• [CRYPTO]: Use try_then_request_module()
• Futex-fd error return fix
• Modules: Be stricter recognizing init&exit sesections
• modprobe -q: quieter when modules missing

Sam Ravnborg:

• kbuild: Do not duplicate A/CFLAGS
• kbuild: arch/i386/boot*, use kbuild syntax when descending into compressed
• kbuild/ieee1394: Makefile update
• kbuild/isdn: Defer md5sum calculation until needed
• kbuild/eisa: Makefile update
• kbuild: genksyms, add explicit reference to include dir

Sridhar Samudrala:

• [SCTP] Fix bugs in sysctl set/get of sctp rto parameters
• [SCTP] draft07 API changes: sctp_getpaddrs(), sctp_getladdrs() now return a packed array of sockaddr_in/sockaddr_in6 structures instead
• [SCTP] SCTP_SET_PEER_PRIMARY socket option support. (Kevin Gao)
• [SCTP] draft07 API changes: sctp_bindx() now takes a packed array of sockaddr_in/sockaddr_in6 structures instead of an array of sockaddr_storage structures.
• [SCTP] Convert sctp_param2sockaddr() and sockaddr2sctp_addr() to address family specific routines af->from_addr_parm() and af->to_addr_param() respectively.
• [SCTP] Fix a couple of issues with the call to sctp_ssnmap_new() in sctp_process_init().
• [SCTP] draft 07 API changes: Disable listening when backlog is 0
• [SCTP] draft 07 API changes: By default, all the event notifications are turned off even for one-to-many style sockets.
• [SCTP] Move a local variable declaration ahead of the function code

Stefan Rompf:

• [netdrvr 8139too] use mii_check_media lib function, instead of homebrew MII bitbanging.

Stelian Pop:

• sonypi driver update
• meye driver update
• reenable CAPTURE button in sonypi
• meye driver update

Stephen Hemminger:

• [IPV4]: Route cache /proc interface cleanup
• [AX25]: Make sure and hold ref to dev
• [AX25]: Convert to seq_file
• [LLC]: Need to pskb_may_pull() in fix_up_incoming_skb()
• [LLC]: Missing sk_set_owner() in llc_sk_alloc
• [LLC]: Set module owner on /proc/net/llc directory
• [ECONET]: Missing sk_set_owner()
• [IPX]: Missing sk_set_owner()
• [ATM]: Missing sk_set_owner()
• [AX25/NETROM/ROSE]: Missing sk_set_owner()
• [IRDA]: Missing sk_set_owner()
• [DDP]: Missing sk_set_owner()
• [DDP]: Invert logic for clarity
• [ATALK]: Fix whitespace in /proc/net/atalk/interfaces header
• [ATALK]: AARP ->last_sent field never set
• [ATALK]: Purge AARP table on module unload
• [ATALK]: AARP needs to use del_timer_sync()
• [ATALK]: Convert AARP over to seq_file
• [ATALK]: Set owner on /proc/net/atalk directory
• [DDP]: Fix obsolete comment about module handling
• [DDP]: Fix oops in aecho socket handling
• [ATALK]: Move aarp procfs file into atalk subdirectory
• [DDP]: Missing netdev refcounting
• [DDP]: Convert to new protocol interface
• [BLUETOOTH]: Missing sk_set_owner()
• [NET]: Convert af_netlink.c over to seq_file
• [NET]: ethertap fixes
• [NET]: DLCI driver cleanups for 2.6.x
• [NET]: Add probe_old_netdevs() hook
• [NET]: Convert SDLA to new initialization
• [NET]: Convert cops over to new initialization
• [NET]: Convert ether probes to probe_old_netdevs()
• [NET]: Convert tr probes to probe_old_netdevs()
• [NET]: Convert sbni initialization
• [NET]: Loopback device simplification
• [NET]: Convert ltpc to new initialization
• ikconfig - cleanups
• [NET]: Convert /proc/net/unix to seq_file
• [NET]: COSA driver fixes
• [NET]: More SDLS fixes
• sdla non-module build fix

Steve French:

• Fix scheduling while atomic problem in getting attributes of newly created file. Fix truncate of existing file when O_CREAT but not O_TRUNC specified
• Fix oops in reconnection logic when no dentry for file being reconnected
• Match smb pid to current->tgid
• update change log for 0.9.1 cifs vfs
• Return error correctly on revalidate so dentry will be dropped
• fix bad return code mapping when server lacks hard link support

Tom Rini:

• PPC32: Update the Motorola MCP(n) 765 support code
• PPC32: Cleanup arch/ppc/boot/simple/Makefile
• PPC32: Fix a warning in the boot serial code
• PPC32: Allow for hooks into the bootwrapper
• PPC32: Fix udelay in the PPC boot code for non-16.6 MHz timebases
• PPC32: Minor fixups to the Motorola Sandpoint platform
• PPC32: Export flush_tlb_page
• PPC32: Fix a warning in the 'mktree' boot util
• PPC32: A number of minor KGDB fixes and tweaks
• PPC32: Add Magic SysRq support to the MPC8260 platforms
• PPC32: Change the default behavior of a kernel with KGDB
• PPC32: Fix KGDB and userland GDB interactions

Ulrich Drepper:

• [NET]: Check tgid not pid in scm_check_creds()
• More ->pid to ->tgid changes

Ville Nuorvala:

• [IPV6]: Fix two bugs in ip6_tunnel.c ICMP error handling
• [IPV6]: Use free_netdev as ip6_tunnel device destructor
• [IPV6]: Set dev->{dev_addr,broadcast} in ip6_tnls
• [IPV6]: Remove sockets from ip6_tunnel.c

Vinay K. Nallamothu:

• vx_entry.c: remove release timer
• [NET]: Fix 'spin_lock_irqrestore' typos in sk_mca.c
• [NET]: Fix MCA device name handling in 3c509.c
• USB: digi_acceleport.c: typo fix
• pcmciamtd.c: remove release timer
• [IPV6]: Fix timer handling in ip6_flowlabel.c

Wim Van Sebroeck:

• [WATCHDOG] advantechwdt.c - patch
• [WATCHDOG] wafer5823wdt.c - patch
• [WATCHDOG] wafer5823wdt.c - patch2
• [WATCHDOG] wafer5823wdt.c - patch3
• [WATCHDOG] acquirewdt.c - patch
• [WATCHDOG] alim1535_wdt.c
• [WATCHDOG] wafer5823wdt.c - patch4
• [WATCHDOG] Documentation

Xose Vazquez Perez:

• [TG3]: More missing PCI ids
• [TG3]: ICH2 needs MBOX write reorder bug workaround too

Yoshinori Sato:

• h8300 interrupt problem fix
• h8300 include update

Yusuf Wilajati Purna:

• [netdrvr] fix skb_padto bugs introduced when skb_padto was introduced