Linux 2.6.3-rc2

Uhhuh. There was a bit more pending, so here's a -rc2. Now please calm down, I'd like this to have some time to stabilize..

The rc1->rc2 changes are mostly driver side stuff: PnP update, USB, ACPI, IRDA, i2c, hotplug-PCI and netdrivers etc. But there's a NFSv4 update and soem XFS fixes there too.

And some ARM and sparc updates.

Linus

Summary of changes from v2.6.3-rc1 to v2.6.3-rc2

Adam Belay:

• PCI: Remove uneeded resource structures from pci_dev
• [PNP]: Fix Serial PnP driver
• [PNP]: Resource flags update
• [PNP]: Disable resources fix
• [PNP]: Avoid static requests
• [PNP] Move ID declarations
• [PNP]: file2alias support
• [PNP]: Card matching code fix
• [PNP]: Add additonal sysfs entries
• [PNP]: Cleanup Kconfig

Adrian Bunk:

• USB: remove USB_SCANNER from the build

Alan Stern:

• USB gadget: file_storage.c -- remove device_unregister_wait()
• USB gadget: fix usb/gadget/file_storage.c doesn't compile withgcc 2.95
• USB: change uhci maintainer
• USB: fix unneeded SubClass entry in unusual_devs.h

Alexander Viro:

• [netdrvr qeth] use alloc_etherdev instead of hand-allocatingstruct net_device
• [netdrvr meth] use alloc_etherdev; fix leaks on error/cleanup
• [wireless ray_cs] use alloc_etherdev
• [netdrvr ne3210] remove #if 0'd code
• [appletalk ipddp] dynamically allocate struct net_device
• [arcnet com90io] use alloc_netdev
• [arcnet arc-rimi] use alloc_netdev; module params; fix bugs onerror/cleanup
• [arcnet com20020] netdev dynamic alloc; module params; fix bugs
• [arcnet com90xx] netdev dynamic alloc; module params; fix bugs
• [arcnet] create and use alloc_arcdev helper
• [netdrvr ppp] netdev dynamic alloc; convert ppp_net_init toalloc_netdev setup function
• [wan synclink] netdev dynamic alloc
• [wan cosa] netdev dyamic alloc
• net_device and netdev private struct allocation improvements
• [hamradio dmascc] convert embedded net_device to dynamicallocation
• [char synclinkmp] convert net_device to dynamic allocation
• [pcmcia] synclink_cs] convert net_device to dynamic allocation
• [atm clip] convert to using alloc_netdev(), const-offset priv
• [netdrvr] s/kfree/free_netdev/ where appropriate
• [wireless wl3501_cs] remove unused constructor
• [netdrvr] convert most 8390 drivers to using alloc_ei_netdev()
• [netdrvr smc-mca] alloc_ei_netdev(), other fixes
• [netdrvr pcnet_cs] alloc_ei_netdev-associated cleanups
• [netdrvr 8390] convert 8390 lib to use const-offset priv struct
• [NET] s/kfree/free_netdev/ in bridge driver
• [netdrvr axnet_cs] use embedded private struct
• [netdrvr ns83820] Switched to sane net_device allocation
• [netdrvr ns83820] Plugs the races around too earlyregister_netdev()
• [netdrvr ibmlana] ibmlana switched to sane allocation, leaksfixed
• [netdrvr] A bunch of gratitious ->init() killed; several leaksplugged
• [netdrvr 8390] Forgotten return 0; removed
• [netdrvr pcnet32] Added missing check and cleanup forregister_netdev() failure in pcnet32
• [netdrvr ibmtr, ibmtr_cs] cleanup and leak fixes
• [irda sa1100_ir] convert to using standard alloc_irdadev()
• Fix imm/ppa initializing bug in driver updates

Andrew Morton:

• [netdrvr] new-probe warning fix
• USB: gcc-3.5: drivers/usb/gadget/net2280.c
• USB: gcc-3.5: drivers/usb/input/hid-core.c
• USB: gcc-3.5: drivers/usb/misc/uss720.c
• USB: gcc-3.5: drivers/usb/storage/usb.c
• Fix qla2xxx warnings
• v850: Add some #includes for the v850 to eliminate some compilerwarnings
• v850: Define ARCH_HAS_*_EXTABLE macros for v850
• v850: make __delay function handle a loop count of zero
• add device id to radeonfb
• Fix ppa/imm warnings
• Fix __filemap_fdatawrite() comment
• bitmap_snprintf() optimization
• alpha: fix build error due to __attribute_const__ conversion
• vx222_ops.c warning fix
• cciss: PCI BAR sizing fix
• cciss: Fix freeing of incorrect IO memory address
• cciss: Add support for SA 6i embedded controller
• cciss: IRQ sharing fix
• cciss: disble prefetching in ASIC
• cciss: intialisation oops fix
• cciss: avoid reading PCI config space
• cciss: printk format fix
• cciss: improve /proc presentation
• cciss: use pci_module_init()
• cciss: rmmod oops fix
• NFS: fix for older gcc's
• Remove duplicated HPPA bits in kernel/sysctl.c
• getxattr error checking fix

Andrew Vasquez:

• qla2xxx Kconfig fix

Bartlomiej Zolnierkiewicz:

• remove MOD_INC_USE_COUNT from drivers/ide/
• fix ns87415.c for PA-RISC Super I/O chip
• fix too early probing of default IDE ports
• fix duplication of DMA {black,white}list in icside.c
• fix OOPS for multiple IDE PCI modules and CONFIG_PROC_FS=y

Benjamin Herrenschmidt:

• ide-cd: incorrect use of sector_div

Christoph Hellwig:

• [SUNRPC]: Use completions instead of sleep_on for rpciod
• [XFS] Fix buffer teardown on _pagebuf_lookup_pages failure
• [XFS] Remove the lockable/not lockable buffer distinction
• [XFS] Remove PBF_MAPPABLE
• [XFS] plug a pagebuf leak
• [XFS] Avoid NULL returns from pagebuf_get
• [XFS] Fix gcc 3.5 compilation for real

Dan Scholnik:

• USB: fix Casio digicam entry in unusual_devs.h

David Brownell:

• USB: fix Bug 1821: sleeping function called
• USB: remove pci_unmap_single() calls from usbcore
• PCI: dma_pool fixups
• USB Gadget: ethernet gadget locking tweaks
• USB Gadget: pxa2xx_udc updates
• USB: usbtest updates
• USB: usbnet updates (new devices)
• USB: USB misc OHCI updates
• USB: re-factor enumeration/reset paths

David Martinez Moreno:

• I2C: fix typos

David S. Miller:

• [SPARC]: Fix AIO syscall numbering
• [ECONET]: Use LL_RESERVED_SPACE() where applicable. Noticed byyoshfuji

Deepak Saxena:

• PCI: Replace pci_pool with generic dma_pool
• PCI: Replace pci_pool with generic dma_pool, part 2
• USB: remove reference to usb_hcd.refcnt in ohci-sa111.c

Dely Sy:

• PCI: Patch to get cpqphp working with IOAPIC

Dominik Brodowski:

• [ACPI] update passive cooling algorithm from Dominik Brodowski
• [ACPI] remove unnecessary check in acpi-cpufreq driver fromDominik Brodowski
• [ACPI] update _PPC handling -- from Dominik Brodowski
• [ACPI] acpi-cpufreq fixups from Dominik Brodowski
• [ACPI] more acpi-cpufreq fixups from Dominik Brodowski
• [ACPI] Move the _PSS and _PCT access to drivers/acpi/processor.cso that it can be used by various low-level drivers (centrino, acpi-io,powernow-k{7,8}, ...) from Dominik Brodowski
• [ACPI] Move the /proc/acpi/processor/./performance output todrivers/acpi/processor.c -- it's the same for all lowlevel drivers.
• [ACPI] Abstract the registration method between low-level driversand the ACPI Processor P-States driver.
• [ACPI] add _PDC support
• [ACPI] make # of performance states dynamic

Eric Sandeen:

• [XFS] Make more xfs errors trappable with panic_mask

Eugene Teo:

• Kobject: export some missing symbols

Geert Uytterhoeven:

• m68k-related net driver fixes
• sun3-related net driver fixes
• 2.6.x experimental net driver queue fix
• I2C: add Hydra i2c bus driver

Greg Kroah-Hartman:

• USB: remove unused usb-debug.c file
• PCI: add back some pci.ids entries that got deleted in the lastbig update
• dmapool: use dev_err() whenever we can to get the betterinformation in it
• USB: add support for the Aceeca Meazura device to the visordriver
• I2C: fix oops when CONFIG_I2C_DEBUG_CORE is enabled and the bttvdriver is loaded
• Driver core: remove device_unregister_wait() as it's a very badidea
• dmapool: Remove sentance in documentation that is now false on2.6
• USB: remove scanner driver files
• PCI: remove stupid MSI debugging code that was never used
• PCI: move pci_msi.h out of include/linux to drivers/pci where itbelongs
• USB: fix bug number 1980 about keyspan devices not gettingrecognized
• Driver Core: fix up list_for_each() calls tolist_for_each_entry()
• dmapool: fix up list_for_each() calls to list_for_each_entry()

Hideaki Yoshifuji:

• [IPV6]: Kill broken and unused IPV6_AUTHHDR support
• [WANROUTER]: Use LL_RESERVED_SPACE() where applicable
• [PACKET]: Use LL_RESERVED_SPACE() where applicable
• [IPVS]: Use LL_RESERVED_SPACE() where applicable
• [IPV4/IGMP]: Use LL_RESERVED_SPACE() where applicable
• [NETFILER/ipt_REJECT]: Use LL_RESERVED_SPACE() where applicable
• [IPV6]: Use LL_RESERVED_SPACE() where applicable

Hirofumi Ogawa:

• [AF_UNIX]: Print out paths correctly in /proc/net/unix, matching2.4.x

James Bottomley:

• SCSI: undelete qlogicfc driver

Jean Delvare:

• I2C: Bring w83l785ts in compliance with sysfs naming conventions
• I2C: Handle read errors in w83l785ts
• I2C: add new driver, gl518sm
• I2C: Update I2C maintainers entry
• I2C: add new chip driver: fscher

Jean Tourrilhes:

• [IRDA]: Ultra sendto support
• [IRDA]: IrLAP disconnection pending race
• [IRDA]: Remove net notifier
• [IRDA]: nsc-ircc irq retval
• [IRDA]: ali-ircc irq retval
• [IRDA]: smsc-ircc2 irq retval
• [IRDA]: via-ircc irq retval
• [IRDA]: w83977af_ir irq retval

Jeff Garzik:

• [netdrvr bonding] fix broken build

Jens Axboe:

• ide-cd invalidate toc cache on last close

Jes Sorensen:

• [ACPI] Check for overflow when parsing MADT entries from JesSorensen

John Belmonte:

• toshiba_acpi 0.17 from John Belmonte

Jon Smirl:

• Driver core: add hotplug support for class_simple

Jonathan Corbet:

• Char drivers: cdev_unmap()

Karol Kozimor:

• acpi4asus update from Karol 'sziwan' Kozimor

Kevin Owen:

• USB: ehci - clear TT buffer command patch

Len Brown:

• [ACPI] add Bruno Ducrot to CREDITS
• fix build error from undefined NR_IRQ_VECTORS
• [ACPI] quiet numa boot -- from Jes Sorensen
• [ACPI] proposed fix for AML parameter passing from Bob Moore http://bugzilla.kernel.org/show_bug.cgi?id=1766
• [ACPI] fix IA64 build warning from Martin Hicks

Linus Torvalds:

• Don't read i8042 data if no data exists
• Make SET_INPUT_KEYCODE return the old value, and clean up usersof this that were very confused indeed.
• Add forward-declaration of "struct nfs4_client" to makenfs_idmap.h independent of CONFIG_NFS4 and other header files.
• Clean up dentry pointer validation by moving it into a functionof its own.
• Linux 2.6.3-rc2

Marcelo Tosatti:

• Fix pc300_tty.c -> implement tiocmset/tiocmget

Markus Demleitner:

• USB: DSBR-100 tiny patch

Matthew Dobson:

• PCI: fix "pcibus_class" Device Class, release function

Matthew Wilcox:

• PA-RISC needs IPC64 structs
• PA-RISC Harmony driver update
• New ptrace.h definitions

Nathan Scott:

• [XFS] Use list_move for moving pagebufs between lists, notlist_add/list_del
• [XFS] Fix compile warning, ensure _pagebuf_lookup_pages returnvalue is inited
• [XFS] Add back a missing pflags check in releasepage
• [XFS] Fix data loss when writing into unwritten extents whilememory is being reclaimed
• [XFS] Fix a trivial compiler warning, remove some no-longer-usedmacros
• [XFS] Update XFS config entries - add security entry, update ACLentry

Olaf Hering:

• USB storage: fix sign bug in usb-storage datafab

Oliver Neukum:

• USB: fix DMA to stack in tt-usb
• USB: fix URB leak in belkin driver

Petri Koistinen:

• [SCTP]: Unify URL referencing in Kconfig documentation
• [NET_SCHED]: Unify URL referencing in Kconfig documentation
• [IPV6]: Unify URL referencing in Kconfig documentation
• [IPVS]: Unify URL referencing in Kconfig documentation
• USB: usb-storage Kconfig_URL_update
• USB: drivers/usb/net config URI update and unify
• USB: drivers/usb/misc/Kconfig URI update & unify: modules.txt
• USB: drivers/usb/input/Kconfig URI unify
• USB: drivers/usb/media/Kconfig URL fixups

Rolf Eike Beer:

• PCI Hotplug: coding style for cpqphp_pci.c
• PCI Hotplug: Whitespace fixes for acpiphp
• PCI Hotplug: Kill spaces before n in ibmphp*
• PCI Hotplug: Kill useless instructions from ibmphp_core.c
• PCI: avoid two returns directly after each other in pcidriver.c
• PCI Hotplug: Coding style fixes for drivers/pci/hotplug.c
• PCI Hotplug: very small optimisations for ibmphp_pci.c
• PCI Hotplug: kill hpcrc from several functions in ibmphp_core.c
• PCI Hotplug: make ibm_unconfigure_device void
• PCI Hotplug: mark functions __init/__exit in acpiphp
• PCI Hotplug: Convert error paths in ibmphp to use goto

Russell King:

• [netdrvr pcmcia] fix hot unplugging
• [irda sa1100_ir] "resurrect from bitrot hell"
• [ARM] Fix asm syntax for gcc3
• [ARM] Use scsi_host_{alloc,put}() rather thanscsi_{un,}register()
• [ARM] Add prefix to driver constants to prevent namespace clashes

Shmulik Hen:

• bonding cleanup 2.6 - Simplify ifenslave
• bonding cleanup 2.6 - Consolidate prints
• bonding cleanup 2.6 - death of typedefs
• bonding cleanup 2.6 - remove dead code
• bonding cleanup 2.6 - Consolidate timer handling
• bonding cleanup 2.6 - Fix handling of bond->primary
• bonding cleanup 2.6 - Remove multicast_mode module param
• bonding cleanup 2.6 - Fix slave list iteration
• bonding cleanup 2.6 - Consolidate function declarations
• bonding cleanup 2.6 - consolidate param names of function paramsand variables
• bonding cleanup 2.6 - Re-org struct bonding members (re-send)
• bonding cleanup 2.6 - consolidate return values of functions
• bonding cleanup 2.6 - Consolidate conditions & statements
• bonding cleanup 2.6 - Consolidate error handling in all xmitfunctions
• bonding cleanup 2.6 - Whitespace cleanup
• bonding cleanup 2.6 - empty lines cleanup
• bonding cleanup 2.6 - fix indentations
• bonding cleanup 2.6 - Code re-org
• bonding cleanup 2.6 - Fix rejects from previous patches

Sridhar Samudrala:

• [SCTP] provide valid tos and oif values for ip_route_output_key. (ja@ssi.bg)
• [SCTP] Add sysctl parameters to update socket send/receivebuffers
• [SCTP] Removed the deprecated ADLER32 checksum support
• [SCTP] Removed SCTP specific rmem/wmem sysctl's and use theglobal rmem_default/wmem_default values for SCTP socket buffer sizes.

Stephen Hemminger:

• [netdrvr skfddi] use PCI hotplug API; other cleanups
• USB: fix usb hc and shared irq handling
• USB: uhci - unused urbp element

Tom Rini:

• USB: mark the scanner driver BROKEN

Tony Lindgren:

• USB: Update ohci-omap to compile

Trond Myklebust:

• NFSv4/RPCSEC_GSS: Ensure that RPC userland upcalls time outcorrectly if the corresponding userland daemon is not up and running.
• RPCSEC_GSS: More fixes to the upcall mechanism
• RPCSEC_GSS: Make the upcalls detect if the userland daemon dieswhile processing a request.
• NFSv4: Fix an Oopsable condition if we fail to get the rootdirectory under NFSv4.
• NFSv4: Bugfixes for the NFSv4 client name to uid mapper
• NFSv4: Bugfixes and cleanups for the NFSv4 client name to uidmapper. Includes a fix by Tim Woods to deal with a caching bug in thecase where a user and a group share the same numerical id and/or name.
• RPCSEC_GSS: Make it safe to share crypto tfms among multiplethreads.
• RPCSEC_GSS: Oops. Major memory leak here
• RPCSEC_GSS: Fix two more memory leaks found by the Stanfordchecker.
• RPCSEC_GSS: Fix yet more memory leaks
• RPCSEC_GSS: Miscellaneous cleanups of the krb5 code required forthe integrity checksumming mode.
• RPCSEC_GSS: Instead of having gss_get_mic allocate memory for themic, require the caller to pass an output buffer whose data pointeralready points to preallocated memory.
• RPCSEC_GSS: Client-side only support for rpcsec_gss integrityprotection. Since this requires checksumming an entire request, insteadof just the header, and since the request may include, for example,pages with write data, we modify the gss_api routines to pass xdr_bufsinstead of xdr_netobjs where necessary.
• RPCSEC_GSS: Move the gss sequence number history from the taskstructure to the request structure, where it makes more sense.
• RPC: xdr_encode_pages either leaves the tail iovec pointing tonull or, if padding onthe page data is needed, sets it to point to alittle bit of static data. This is a problem if we're expecting tolater append some data in gss_wrap_req. Modify xdr_encode_pages to maketail point to the end of the data in head, as xdr_inline_pages andxdr_write_pages do.
• RPC: Add support for sharing the same RPC transport andcredential caches between different mountpoints by allowing cloning ofthe rpc_client struct.
• NFSv4/RPCSEC_GSS: Make Frank's server->client_sys feature useRPC cloning in order to avoid duplicating sockets etc. Make NFSv4 sharea single socket for all communication to the same server.
• NFSv4: Convert the RENEW operation from using nfs4_compound, tobeing a standalone RPC call in preparation for the renew daemon
• NFSv4: Convert the lease renewal daemon from being per-mountpointto being per-server. Instead of running it on top of rpciod, convert itto use keventd. This mean we can use the struct nfs4_client semaphoresfor ordering purposes.
• NFSv4: Split out the code for retrieving static serverinformation out of the GETATTR compound.
• NFSv4: Convert SETCLIENTID and SETCLIENTID_CONFIRM to bestandalone operations. Ensure that SETCLIENTID_CONFIRM always returnsthe lease timeout length.
• NFSv4: Don't translate those NFSv4 errors that are needed by thekernel itself into EIO.
• NFSv4: Preparation for the server reboot recovery code
• NFSv4: Basic code for recovering file OPEN state after a serverreboot.
• RPC/NFSv4: Allow lease RENEW calls to be soft (i.e. to time
• RPC: Ensure that we disconnect TCP sockets if there has been noNFS traffic for the last 5 minutes. This code also affects NFSv2/v3.
• NFSv4: Atomic open(). Fixes races w.r.t. opening files
• NFSv4: Share open_owner structs between several differentprocesses. Reduces the load on the server.
• NFSv4: Fix a bug which was causing Oopses if the client wasmounting more than one partition from the same server.
• NFSv4: Add support for POSIX file locking