Summary of changes from v2.5.73 to v2.5.74 ============================================ Use a non-zero rx_copybreak to avoid charging a full MTU to the socket on tiny packets. Dave Miller suggested 256, I used 200 to be more consistent with the other network drivers. [CPUFREQ] speedstep_detect_speed might not reenable interrupts From Samuel Thibault local_irq_save() is called at the beginning of speedstep_detect_speeds, but local_irq_restore() is not called on I/O errors. [SERIAL] Fix 8250 revision width. Patch from wli, edited by rmk. 'rev' needs to be at least 16 bits. [SCTP] Break out sctp_assoc_valid() from sctp_id2assoc(). sctp_id2assoc() was trying to do too many things. Break out, so sctp_assoc_bh_rcv() can just find out whether the assoc is valid. ppc64: implement any_online_cpu PPC32: Implement force_successful_syscall_return. We use a bit in the thread_info flags for this. The bit is cleared on syscall entry and set by force_successful_syscall_return. If the bit is set on syscall exit, the normal check whether the value from the syscall function is an error code is disabled and the value is just returned to userspace. ppc64: Fix sysinfo translation, spotted by Dave Miller [PATCH] ia64: bitshift fix I think there are other 64p issues in other places, but we found this one while doing 2.5 bringup on hardware. [PATCH] ia64: mark_idle() fixes for sn2 Small patch to fixup the new mark_idle() stuff for sn2. [PATCH] ia64: hwgfs fix for sn2 Small fix for hwgfs on sn2. [PATCH] ia64: PCI fixes for sn2 Updates the PCI routines for sn2. [PATCH] ia64: generic kernel support for sn2 I _think_ I got this one right (at least, it booted on hw). It should allow generic kernels to include the sn2 code. ia64: Drop obsolete ACPI SPCR support. ia64: Two more minor cleanups for 2.5.72. [PATCH] ia64: early_printk for SGI SN2 ia64: Minor cleanups; fix non-SMP build. ia64: Simplify the script to use only $CC and $OBJDUMP. Besides being simpler, this also ensure we really do test the linker which will be used when building the gate DSO. Kconfig: Add back IDE configuration. [PATCH] ia64: sn2 updates for 2.5.72 Here's a patch against your bk tree that keeps things building for sn2 (and working with the discontig patch applied). [PATCH] ia64: cleanup xor build rule to take advantage of lib-m ia64: Fix a alternate-signal-stack bug which could corrupt RNaT bits when bspstore happened to point to an RNaT-slot. Bug reported by Matt Chapman. ia64: More perfmon fixes: - removes 2.4 specific macros/inline functions - fix rum vs. rsm bug for psr.up - introduce cleaner PMC_IS_CONTROL() macro - only save the value of psr.up on ctxsw (not whole psr)" ppc64: Convert from K+R to ansi ppc64: small __access_ok optimisation [PATCH] USB: GFDL in the kernel tree I don't care about this enough to argue ... I attach a patch to make it use the same legalese as the other docs. USB: delete cdc-ether driver as it's no longer needed. [PATCH] USB: add module paramater to visor driver for new devices Allows new palm OS 4.x devices to be added by a module paramater instead of being forced to recompile the driver. [PATCH] USB: fix kaweth warnings - fix compiler warning due to merge oversight [PATCH] USB: pl2303: add ability to report CTS and DSR status to userspace [PATCH] USB: fix UHCI debug kmalloc() usage Here's a patch from Tony Luck that fixes a problem with the UHCI debugging code on architectures with big page sizes. We end up allocating more than kmalloc allows. [XFS] Modified to always pass NULL as data parameter to VOP_ATTR_GET(), whenever ATTR_KERNOVAL flag is used. SGI Modid: 2.5.x-xfs:slinx:147567a [XFS] Remove a use after free which was introduced in the debug build. SGI Modid: 2.5.x-xfs:slinx:148158a [XFS] optimize the code used to initialize inodes during their allocation. SGI Modid: 2.5.x-xfs:slinx:148423a [XFS] Ensure that when we unlock a log item when there is someone waiting for log space, we give at least one thread a chance to flush metadata and free more log space. SGI Modid: 2.5.x-xfs:slinx:149321a [XFS] When flushing inodes in to free log space, make the flush async rather than delayed write. SGI Modid: 2.5.x-xfs:slinx:149418a [XFS] Remove test for impossible condition (unsigned 8-bit >= 256) SGI Modid: 2.5.x-xfs:slinx:149574a [XFS] Fix an assert in log init code SGI Modid: 2.5.x-xfs:slinx:149646a [XFS] V2 log update - tighten checks for corrupt log during recovery SGI Modid: 2.5.x-xfs:slinx:150093a [XFS] Version 2 log code fixes in recovery. SGI Modid: 2.5.x-xfs:slinx:150122a [XFS] fix sign of error return in mount update and statfs operations, mount update could have returned either sign, and statfs was just wrong. SGI Modid: 2.5.x-xfs:slinx:150137a [XFS] Update linux directory inode contents after the initial mkdir, right now size shows up as zero which is wrong. SGI Modid: 2.5.x-xfs:slinx:150334a [XFS] if we remount the fs readonly, there is no need to put out dummy log records to clear the log as these actually have the effect of dirtying it in a readonly unmount. SGI Modid: 2.5.x-xfs:slinx:150738a [PATCH] fix vlsi_ir.c compile if !CONFIG_PROC_FS On Mon, 9 Jun 2003, Jean Tourrilhes wrote: > > Yes, Thanks. In fact walking over the proc-stuff in vlsi_ir, there are > > some more places which need cleanup and fixing. > > > > Jean, please apply the patch below - ChangeLog: > > Would you mind sending it also to Jeff ? I want you to be able > to work without depending on me. > But I'll add that on my todo list anyway... Ok - here we go. Slightly improved patch compared to what I've sent first so we handle now failure in create_proc_entry and !CONFIG_PROC_FS identical, i.e. continue without proc entry. It's not needed for the driver to work. Updated Changelog: * make it compile without CONFIG_PROC_FS (problem reported by Adrian Bunk, original patch by Sam Ravnborg) * get rid of a number of unneeded ifdef/endif CONFIG_PROC_FS (also removing an "unused label" warning) * use proc entrys' owner fields to protect against module removal racing with proc entry access. * handle failure in create_proc_entry like !CONFIG_PROC_FS * set driver version 0.4a Jeff, would please forward it for application to 2.5! Martin [XFS] make doing an iread from a bad location less chatty by default as some code paths can actually do this and handle the result. SGI Modid: 2.5.x-xfs:slinx:150747a [XFS] break dependency between CONFIG_PROCFS and CONFIG_SYSCTL in xfs SGI Modid: 2.5.x-xfs:slinx:150782a [XFS] Turn log head/tail check into an assert - this condition should never be true here SGI Modid: 2.5.x-xfs:slinx:151422a [XFS] SYNC_FSDATA and SYNC_REFCACHE are supposed to be distinct values, they ended up the same somehow. SGI Modid: 2.5.x-xfs:slinx:151227a [XFS] Move the pagebuf_runall_queues out from under a spinlock, if it sleeps (which it might), this is not the right place for it to be. There is also no reason for it to be under a spinlock. SGI Modid: 2.5.x-xfs:slinx:151032a [XFS] Fix deadlock between xfs_finish_reclaim and xfs_iget_core. An inode being reclaimed and removed from memory by one thread while another thread is attempting to reuse the inode and bring it back to life. There was a window between the iget starting to reuse the inode and the reclaim starting. Close the window by marking the inode as being reused under the hash lock, and by abandoning the reclaim if this is detected when it obtains the hash lock. SGI Modid: 2.5.x-xfs:slinx:151123a [XFS] really delete kmem.c SGI Modid: 2.5.x-xfs:slinx:148936a [XFS] Fix some remaining unaligned access issues on 64 bit platforms. SGI Modid: 2.5.x-xfs:slinx:143601a [XFS] Remove a dead code path from mount, noticed by Al Viro. SGI Modid: 2.5.x-xfs:slinx:151226a [XFS] Add in benign missing sector flag to SB version macro SGI Modid: 2.5.x-xfs:slinx:151428a [XFS] Fix remount readonly so it really wont write dummy log records SGI Modid: 2.5.x-xfs:slinx:151423a [XFS] Fix deadlock caused by race between xfs_iunpin marking an inode dirty, and the same inode being reallocated and reused by create. The unlock_new_inode and mark_inode_dirty manage to corrupt the i_state flags and end up leaving I_LOCK|I_NEW set. SGI Modid: 2.5.x-xfs:slinx:151120a [XFS] Close window to prevent a file with delayed write extents from being converted to a realtime file. SGI Modid: 2.5.x-xfs:slinx:150953a [PCMCIA] TI1520 cardbus patch for 2.4.x.. Linus Torvalds wrote: > Here's a patch that apparently improves throughput on Texas > Instrument 1520 controllers a lot by enabling memory bursting. > Apparently to the point that without this you can't even do > things like 802.11g without starting to drop data. Done. Also cleaned up the "Override Table" a bit. [PCMCIA] ti113x - fall back to PCI interrupt (from 2.4.-ac) Stumbled over this patch within 2.4.21-ac. Depends on the ti113x.h burst mode patch sent earlier. [SCTP] Change sysctl settable timers to ms granularity (rmlayer@us.ibm.com) [SCTP] Fix for incorrect vtag in the stale cookie ERROR chunk. [SCTP] OOTB Cookie-Echo path does not like stale cookie vtag fix. Move fix out of generic path, into the stale cookie path. [PATCH] __devinitdata declarations should not be marked const. There are some bad compiler issues with section confusion ("const" stuff likes going into the rodata section). [PATCH] loop.c cleanups This does the following: - remove trailing spaces - make loop.h independent by including bio.h, blk.h, spinlock.h - replace the lock/unlock functions by module_get/module_put; in struct loop this is the change - void (*lock)(struct loop_device *); - void (*unlock)(struct loop_device *); + struct module *owner; - replace the integer lo_encrypt_type by the pointer lo_encryption; there was a race with loop_unregister_transfer - fixed an off-by-one in loop_register_transfer This is Step 1 of a series of half a dozen or so. Half of the above is from Jari. Anything that is wrong is mine. [PATCH] fix nbd driver for 2.5 block layer This makes NBD work with the new linux 2.5 block layer design. Specifically, it fixes memory corruption that results from module removal and possible memory corruption from sending or receiving disk data from the server. It essentially rolls together the changes from two of the last patchlets that I emailed: the fix for module removal & the fix for incorrect struct bio usage. I believe it's wisest to roll these both together into this one patch since they both deal with making NBD work better with the 2.5 linux block layer design and without either of which, it's possible that NBD will corrupt memory. Other changes I'd like to see introduced (like in the earlier jumbo patch) meanwhile are feature enhancements so they can wait. This patch also should address all the very helpful concerns that have been raised so far. Particularly: 1. that the very first submitted NBD patch was broken down [Andrew] 2. that only 1 spinlock is used for all the NBD request_queue structures used [Jens,Al] 3. that kmap() is used in case of highmem pages [Jens] 4. that the allocation of request_queue is dynamic and seperate from other allocated objects [Al] [SPARC64]: Update struct compat_statfs. [SPARC64]: Update solaris compat layer for vfs_statfs() changes. [SPARC]: Update for show_stack() changes. [SPARC]: Add {f,}statfs64 syscall entries. [PATCH] Workqueue Exit Neatening Jeff Garzik points out the initializing the exit completion at exit time is foolish: we should just initialize it at creation time live everything else in that structure, and avoid the memory barrier. [PCMCIA] Fix ide-cs driver name (for PCMCIA binding) Problem noticed by Eivind Tagseth. We seem to have resurected the "ide_cs" driver name for the ide-cs module. Restore the driver name to "ide-cs". MTD driver cleanups... - Fix AFS partitioning oops when no partitions are found - Add missing spin_unlock, optimise buffer writes in Intel NOR driver - Fix DiskOnChip Millennium Plus register OutputControl register definition - Fix DiskOnChip drivers to indicate correct ECC type - Fix map drivers to use ARRAY_SIZE instead of redefining it. - Make uCLinux map driver depend on !MMU - Fix NAND write verify problem on some chips - Other trivia from Rusty. Replace mtd_blktrans ->ioctl() method with ->getgeo() and ->flush() ... and also fix the embarrassing bug where NFTL and INFTL will barf and exit if the add_mtd_blktrans_dev() function _exists_, rather than actually calling it and barfing if it returns non-zero :) Trivia: use JFFS2 PAD() macro instead of masking manually. Fix jffs2_statfs w.r.t. statfs64 Remove superfluous debugging in mtd_blkdevs.c [PATCH] PCI: fix non-hotplug build Current BK won't build when CONFIG_HOTPLUG is not set due to undefined references to pci_destroy_dev in hotplug.c. I think it makes sense to not compile hotplug.c in this case at all. Also, this allows to get rid of several function which are unused in non-hotplug kernel. Tested on Alpha. [PATCH] PCI: fix alpha for reimplement pci proc name On Fri, Jun 20, 2003 at 02:24:13PM -0700, Greg KH wrote: > Thanks, I've reverted your previous patch, and fixed the one typo in > this patch and applied it all to my bk tree. Hopefully Linus will pull > from it sometime soon :) Argh, where were my eyes... There was another typo which broke Alpha. [PATCH] ia64: fix static initializers No need to explicitly zero out global variables. [PATCH] Fix XFS proc interface initializers [XFS] the intializers for the /proc interface to xfs got out of order, and we are syncing 1000 times faster than we are supposed to! SGI Modid: 2.5.x-xfs:slinx:151712a [PATCH] MIPS merge, generic mips bits. This contains all the generic 32-bit MIPS code, so all arch/mips/ and include/asm-mips/ stuff. [PATCH] MIPS merge, generic mips64 bits. This updates the generic mips64 code. [PATCH] SGI IP22 bits An update for the Indy aka IP22 support. Consolidates the 32-bit and 64-bit copies of the support code into one directory, so in total this patch deletes quite a bit of code. [PATCH] PCI code This is the entire MIPS PCI code which I'm consolidating in arch/mips/pci/. Applying this patch will result in some code duplication; the remaining patches I'm about to send will clean that. [PATCH] 2.7.73 SGI IP27 update An update of the SGI IP27 aka Origin 200/2000/Onyx 2 support. [PATCH] Code for Galileo boards The support code for two more Galileo evaluation boards. (Evil stuff, just to get your tree in sync again. I've got a cleanup in my work tree). [PATCH] Sibyte updates Update the code for the BCM1250 and evaluation platforms. [PATCH] Alchemy update This is an update of the AMD Alchemy support for 2.5. (This is way behind what we have for 2.4 but forward porting is non-trivial and work in progress.) [PATCH] Update for MIPS Inc's eval boards. This is an update for MIPS Inc's evaluation boards in all their ugly versions ... [PATCH] TX49xx update This adds support for Toshiba's TX49xx SOCs and an evaluation board with the nice nae RBTX4927. [PATCH] Baget update This is an update for the Russian Baget industrial controller. Suffers a bit from bitrot, the authors have promised an update to me. [PATCH] DEC update An update of the code for the DECstations. This also adds 64-bit support for the R4000 versions of DEC's good old workstations. [PATCH] NEC DDB update This updates the support for NEC's DDB series of evaluation boards. [PATCH] NEC VR41xx update This one adds support for a bunch of NEC VR41xx-based platforms such as IBM's workpad and a bunch of eval boards. [PATCH] JMR3927 update This updates support for the JMR3927 eval board. [PATCH] Update Cobalt support This adds back support for the MIPS-based Cobalt Raq 1/2 and Qube 1/2 systems. [PATCH] Add support for SGI IP32 This adds support for SGI's O2 workstation aka IP32. [PATCH] Momentum update This adds the board support code for Momentum's Ocelot series of boards. [PATCH] drivers/sgi update This updates drivers/sgi by removing it :-) With all the conceptually wrong code which has was rewritten or should be rewritten or better live elsewhere there just was no point in keeping this directory around any longer. [PATCH] ARC update This updates the ARC firmware support code. Also removes the 64-bit variant of the code; the 64-bit kernel now uses the 32-bit code also. [PATCH] Lasat support This patch adds support for the Lasat 100 and 200 systems. [PATCH] Remove Nino support This patch mostly removes the support for the Phillips Nino at the request of the author. The only remaining bits are directly related to the SOC the Nino is based on which is used by other ports. Remove SGI subdirectory from driver Makefile, since it is now gone. [IPV6]: Clean-up advmss calculation. [ATM]: Move vccs to global sk-based linked list. [NET]: PPPoE cleanup [TRIVIAL]. This is a cleanup patch, no change in functionality. - Get rid of debug macro's that aren't used anywhere in the code. - Make functions and data structures static where possible - C99 initializer for ppoe_chan_ops - fix whitespace typo [NET]: Fix oops on /proc/net/pppoe. [NET]: Convert PPPoE to new style protocol. [PATCH] update the generic saa7146 driver - update the generic saa7146 driver - remove some #if LINUX_VERSION_CODE constructions - sync with the interrupt handler changes in 2.5.69 - add a missing kfree() call which caused the kernel to leak 32kB of kmalloc()ed memory. iieek! - fixed the capture code to handle cards that have swapped field order (odd and even fields) - added and fixed some debug messages - changed from kmalloc() to pci_consistent() - many small changes necessary to fix warnings/problems for ppc64 compilation [PATCH] update dvb subsystem core - update dvb subsystem core - switched from user-land types like __u8 to u8 and uint16_t to u16 this makes the patch rather large. - updated the dvr (digital videorecording) facility - renamed some structures, like "struct dmxdev_s" to "struct dmxdev" - introduced dvb_functions.[ch], where some linux-kernel specific functions are encapsulated. by this, the dvb subsystem stays quite independent from deeper linux kernel functions. - moved dvb_usercopy() to dvb_functions.c -- this is essentially video_usercopy() which should be generic_usercopy() instead... - Made the dvb-core in dvbdev.c work with devfs again. - remove all typedefs from structs - remove all typedefs from enums [PATCH] update the av7110 and budget drivers - update the av7110 and budget drivers - replaced ddelay() wait function with generic dvb_delay() implementation - new DATA_MPEG_VIDEO_EVENT for direct mpeg2 video playback - added support for DVB-C cards with MSP3400 mixer and analog tuner - fixed up the av7110_ir handler and especially the write_proc() function; this fixed the bug the Stanford Checker has found [PATCH] update dvb frontend drivers - update dvb frontend drivers - C99 initializers - fix up some includes - various bugfixes [PATCH] add a new driver for the cx24110 frontend - add a new driver for the cx24110 frontend by Peter Hettkamp [PATCH] add dvb subsystem as a crc32 lib user - add dvb subsystem as a crc32 lib user [PATCH] update analog saa7146 drivers mxb and dpc7146 - update analog saa7146 drivers mxb and dpc7146 - add MODULE_DEVICE_TABLE entries, so that /sbin/hotplug can handle the devices - fixup due to the latest i2c changes [PATCH] correct the i2c address of saa7111 - this patch corrects the i2c address from "34>>1" to 0x24 and 0x25. Believe me -- or look at the data sheet, for example from http://www.gdv.uni-hannover.de/~hunold1/linux/saa7146/specs/saa7111a.pdf Page 41 says: "Slave address read = 49H or 4BH; note 2 write = 48H or 4AH" They use 8-bit addresses here, but i2c addresses are 7-bit, ie. 0x48>>1 == 0x24 and 0x4a>>1 = 0x25 [PATCH] clean up the parts according to the comments on kernel mailing list clean up the parts according to the comments on kernel mailing list (mainly by Christoph Hellwig) - ugly WRITE_RPS0 define in saa7146_hlp.c has been replaced by a proper inline (I hope) - use not everywhere - include headers after ones - revert the indentation from "static xxx to "static xxx" [NET]: Fix OOPSes with RSVP. [IPSEC]: Add encap_oa member to struct xfrm_encap_tmpl. [IPSEC]: Close SADB_ADD race and add XFRM_MSG_UPDSA (SADB_UPDATE equivalent). [NET]: Trivial patch to netfilter Kconfig. [IPV4]: Be more verbose about invalid ICMPs sent to broadcast. [NET]: Update teql scheduler to dynamic net device. - dynamically allocate and free the network device. previously, used static network device. - support multiple equalizers (default one) via module parameter (max_equalizers) previously, limited to one. [PATCH] PCI: unconfuse arch/i386/pci/Makefile I was looking in this Makefile for link order when my head began to hurt. Apparently you can't have both NUMAQ and VISWS selected, so getting rid of all the ifdefs/ifndefs like this should work. [PATCH] PCI: pci_raw_ops devfn Combine the dev and func arguments to pci_raw_ops into devfn which is more natural all around. [PATCH] PCI Hotplug: fix buggy comparison in cpqphp_pci.c I don't understand the code good enough to be sure my patch is correct, but the current code is definitely buggy: 0xFF is the maximum value for an u8, so tdevice < 0x100 is _always_ true. [PATCH] I2C: add i2c-prosavage driver Using the MMIO method now, the driver should be able to handle multiple video cards. The driver could potentialy also handle other s3 devices. You can try this by adding more pci id's to the prosavage_pci_tbl. [SERIAL] Missing Kconfig dependencies If one turns off SERIAL_8250, these items shouldn't be selectable. Also gets the indentation right in `make oldconfig'. [IPV{4,6}]: Fix "slow multicast on 2.5.69" bug. [SERIAL] 8250_cs update - remove serial_info_t I2C: add i2c-ali1535 bus driver Ported from the i2c cvs tree. [SERIAL] 8250_cs update - remove work queue Remove the work queue for serial_remove; events always happen in process context now so the work queue gains us nothing. [SERIAL] 8250_cs update - incorporate pcmcia-cs 3.1.34 serial_cs fixes - add buggy_uart parameter - use detected Vcc values, not CIS Vcc values - handle OxSemi OXCF950 ports [PATCH] {PATCH 2.5.72] Use mod_timer in drivers_net_wan_sdla_fr.c From: Vinay K Nallamothu [PATCH] [PATCH 2.5.72] Use mod_timer in drivers_net_wan_sdla_ppp.c From: Vinay K Nallamothu [PATCH] [PATCH 2.5.72] Use mod_timer in drivers_net_wan_sdla_x25.c From: Vinay K Nallamothu [PATCH] [PATCH 2.5.72] Use mod_timer in drivers_net_wan_sdla_chdlc.c From: Vinay K Nallamothu [netdrvr sk98lin] PCI API conversion, and some cleanups - PCI API init style conversion for drivers/net/sk98lin/skge.c; - new helpers: SkGeDev{Init/CleanUp}; - sk_devs_lock moved around as it's needed early. Compiles without error. Untested. [PATCH] USB: highdma support for kaweth this tells the network layer that we can DMA from high memory if the host controller supports it. - enable 64bit DMA on platforms that support it ia64: Rename init_thread_union to init_task_mem to avoid conflicting declration in . Drop pcibios_update_resource() #warning. ia64: Rename irq_desc() to irq_descp() to avoid conflict with variable of same name declared in linux/irq.h. The expectation is that this variable will be removed eventually and then irq_descp() can be renamed to irq_desc() again. But for now, this makes it easier to compile against Linus' source tree. ia64: Update defconfig. Add missing include to . Fix compiler warning in perfmon.c. ppc64: remove memset/memcpy function prototypes in eeh header file ppc64: memset destination if access_ok fails in copy_from_user ppc64: pci domain support [XFS]: Fix build error on big-endian. ppc64: fix some compile warnings PPC32: Implement 2-argument show_stack(). ppc64: add dabr and iabr cpu features ppc64: merge Paul's interrupt-map parsing code [IPV6] use macro for maximum payload length [IPV6] Fix large packet length check There were two errors in length check in the output path. We could not send large packet (65535bytes). This patch fixes the problem. [ALPHA] Don't print all interrupts in /proc/stat. From Jeff Wiedemeier. [SCTP] Add ASSOCINFO and RTOINFO sockopts. (Ryan Layer and Anup Pemmaiah) Submitted by pemmaiah@cc.usu.edu and rmlayer@us.ibm.com. [SCTP] Fix wrong logic in hlist change. Otherwise, we leak bind_buckets. [SCTP] Remove some unused source modules. We use crypto api. Get rid of our own versions of hash code. [SCTP] Minor warning cleanups. [SCTP] More typedef removals. sctp_cookie_t -> sctp_cookie, sctp_signed_cookie_t ->sctp_signed_cookie_t. [SCTP] More typedef & name cleanup. [SCTP] Shorten SACK generation path. 1) Remove sctp_tsnmap_update_pending from the SACK path. The pending_data variable is only used by a socket option, so just calculate it when needed rather than the I/O path. 2) Instead of walking the tsnmap twice, change the interface to allow walking once. 3) Only report a fixed number of gabs and reserve this room in the association, saving us a kmalloc every sack generation. TBD: Still need to kick out of tanmap walking early if we get to max_tsn_seen. [SCTP] Don't search gap ack blocks past max_tsn_seen. There is no need to keep walking the tsnmap past the max_tsn_seen, so don't. [PATCH] Fix bug in rpc cache_clean introduced by previous patch Having a 'continue' at the end of this look it pointless... it should be a break. [PATCH] Always flush rpc caches after an update ..as the update might have set an expiry date in the past and we want it to be removed instantly. [PATCH] Include update mode in declaration of RPC information caches. The sunrpc/cache.c caches which store nfsd export information can be updated either "inplace" or by replacing the entry. replacement is needed when an entry hold a reference to some other object, so the reference counts work properly. "inplace" can be used when no such references are held, and should be used when the object could be refered to by another cache (as otherwise the other cache would have to be updated whenever this one is). Previously the type of update (inplace or replace) was specified as an argument to the *_lookup operation. This too easily lead to inconsistancies. With this patch, the update mode is specified when the cache is declared. [PATCH] Add some tracing when showing the content of an RPC cache. The /proc/net/rpc/*/content files now will show entries that are still in the cache, but are either expired or negative, as comment. ip_map_show is enhance to work if called with a negative or incomplete entry. Also if cache debugging is enabled, the expiry time and refcount of each entry will be included in a comment. [PATCH] Define cache_show methods for export and filehandle cache in nfsd. [PATCH] Use schedule_work to regular cache cleaning Cleaning of the export caches is currently done by idle nfsd threads which isn't very reliable. This patch makes use of work_queues to do it all inside cache.c [PATCH] Fix byte counting for NFSv3 readdir replies The amount of free space is calculated wrongly so nfsv3 readdir replies are shorted than they could be. [PATCH] Get buf size and page count right for RPC services. The old calculations BUGed on 64k PAGESIZE machines. [PATCH] do not take ide_setting_sem under ide_lock From -mm tree, fixes oops when removing ide drivers. [PATCH] fix use-after-free in ide_unregister() [PATCH] build fix for pdc4030 without taskfile IO Noticed by Arkadiusz Miskiewicz. [PATCH] fix compilation of NS SC1x00 driver without procfs Noticed by Adrian Bunk. [PATCH] call setup_driver_defaults() only once for each driver [PATCH] enable local APIC on P4 The current local APIC code refuses to enable the local APIC on a P4 if the BIOS booted us with the local APIC disabled. This patch removes this unnecessary restriction. Please apply. Most P4 machines do boot with the local APIC enabled, but Keith Owens reported that the P4 based Compaq Evo N800v disables the local APIC, even though the machine actually works if Linux enables it. It is possible that some P4 machines with broken BIOSen were saved by our refusal to enable the local APIC. We can handle them via the DMI blacklist rules instead. [PNP] pnp_init_resource_table compile fix In the last release, this api was accidently changed and therefore affected some drivers. This patch corrects the issue by renaming the api back to pnp_init_resource_table. [PNP] Locking Fixes The semaphore in pnp_init_resource_table is not needed and, in some cases, can cause resource management lockups. This patch removes the improperly placed semaphore. [SEQ_FILE]: Export seq_path() to modules. [PATCH] USB: add support for 50 baud to io_edgeport.c [PATCH] USB: Desknote/ECS UCR-61S2B card reader (2.5.72 patched) This is for 2.5.72 with the US_PR_DEVICE / US_SC_DEVICE patch in place. Tested and working. [PATCH] USB: make kaweth deal with ENOMEM this fixes the logic kaweth uses to deal with ENOMEM from usb_submit_urb(). Using the interrupt endpoint's completion handller is longer an option because automagic resubmission has been removed. Thus we use workqueues which can be delayed so that kswapd can do its job and can use GFP_NOIO which is much likelier to succeed. [IPV4]: Trim the includes used in util.c [PATCH] Fix compilation of ip2main The following patch fix compilation of drivers/char/ip2main.c. It was broken by the removal of pci_present(). It just adds open and closing braces around the code that declares the pci_dev_i variable. The rest of the patch just change the indentation. [PATCH] PCI: [PATCH] pcibios_scan_acpi() On Mon, Jun 23, 2003 at 02:39:05PM -0700, Greg KH wrote: > > How about acpi_scan_pci_bus_root()? > > I agree, that sounds better. I think it's too long ... so unless anyone has a better idea, I'm going with pci_acpi_scan_root(). Here's the patch, presented in patch -p1 format to make greg's scripts happy ;-) ia64 needs to be passed the pci domain and the acpi handle corresponding to each root bus. Rather than change pcibios_scan_root to take additional arguments, this patch introduces pci_acpi_scan_root(). `cat msg` [NETFILTER]: Forward port cosmetic fixes from 2.4.x [NETFILTER]: Enhancement for ip{,6}_tables, add new /proc files. [NETFILTER]: Fix conntrack master_ct refcounting. [PATCH] PCI documentation Update the PCI Documentation to reflect some of the functions which have recently been added and removed. Index: Documentation/pci.txt =================================================================== RCS file: /var/cvs/linux-2.5/Documentation/pci.txt,v retrieving revision 1.4 [NETFILTER]: Add "recent" iptables facility. [NET]: Need sys_socket cond_syscall() entry. [PATCH] loop.c - part 2 of N This does the following: - IV value is current 512-byte sector relative to start of loop container file. This is what all cryptoloop people have done, if I am not mistaken. Andi or others - if you can demonstrate the need for a more flexible setup an additional ioctl field may be needed. I hope we can do without. - made some things static - made lo_offset a loff_t - added lo_sizelimit If one wanted a (crypto)loop somewhere inside a container file, the old code allowed a starting offset, but no size, so that the cryptoloop always extended to the end of the container file. This field allows one to select an arbitrary interval. Note that this changes struct loop_info64. - improve error handling of loop_init() - removed the unused typedef transfer_proc_t. - added a define for LO_CRYPT_CRYPTOAPI [PATCH] unexpected IO-APIC update Recently there has been a rash of Unexpected IO APIC reports on the linux-smp mailing list. Most of the most recent ones are due to some newer Intel chipsets (865, 875). The IO APIC Version register doesn't indicate the differences in these IO APICs. I have an patch that addresses these chipsets. It has been tested by a few people with good results and has been blessed by Maciej Rozycki. Other than conditionally decoding IO APIC registers 2 and 3, we could alternately ignore them since Linux doesn't use the values for anything other than printing them. This patch ignores IO APIC register 2 if it's the same value as IO APIC register 1. It also reads IO APIC register 3 if the IO APIC version is >= 0x20, but some chipsets don't support this register, so it is also ignored if its value if the same as IO APIC register 1 or 2. Another possible(?) alternative is to read the PID/VID of the device to determine which registers it supports. However, PCI devices have not been scanned at this point in init, so it would require scanning PCI config space directly and I don't yet see the point of doing that. Oh, and the UNEXPECTED_IO_APIC() function doesn't print anything in 2.5.current and I didn't change that. [PATCH] alloc_etherdev for 3c574_cs net_device is no longer allocated as part of the driver's private structure, instead it's allocated via alloc_netdev. compile tested only since no hardware against 2.5.73-bk -daniel ===== drivers/net/pcmcia/3c574_cs.c 1.17 vs edited ===== [PATCH] alloc_etherdev for 3c589_cs net_device is no longer allocated as part of the driver's private structure, instead it's allocated via alloc_netdev. compile tested only since no hardware against 2.5.73-bk -daniel ===== drivers/net/pcmcia/3c589_cs.c 1.17 vs edited ===== [PATCH] alloc_etherdev for fmvj18x_cs net_device is no longer allocated as part of the driver's private structure, instead it's allocated via alloc_netdev. compile tested only since no hardware against 2.5.73-bk -daniel ===== fmvj18x_cs.c 1.21 vs edited ===== [PATCH] alloc_etherdev for nmclan_cs net_device is no longer allocated as part of the driver's private structure, instead it's allocated via alloc_netdev. compile tested only since no hardware against 2.5.73-bk -daniel ===== nmclan_cs.c 1.14 vs edited ===== [PATCH] alloc_etherdev for smc91c92_cs net_device is no longer allocated as part of the driver's private structure, instead it's allocated via alloc_netdev. compile tested only since no hardware against 2.5.73-bk -daniel ===== smc91c92_cs.c 1.18 vs edited ===== [PATCH] Remove CAP_NET_ADMIN check for SIOCETHTOOL's dev_ioctl already checks capable(CAP_NET_ADMIN), so no need to do so in drivers. [PATCH] 2.5.70 - eepro100 - use alloc_etherdev Ignore earlier patch -- this one locks and free's as appropriate. Tested on 2.5.72 with SMP. Of course, it begs the question why have two (now three) versions of drivers for the same hardware... [USB]: Use linux/dma-mapping.h not asm/dma-mapping.h in kaweth.c [PCI]: Export pci_enable_device_bars to modules. [SPARC64]: Update defconfig. [SERIAL]: Sanitize sparc serial console configuration. [SPARC64]: Update defconfig. PPC32: remove check for ERESTARTNOHAND in syscall return path. This error is handled in the signal delivery code and should never be returned from a syscall unless a signal is pending. Grepping seems to indicate that that is in fact the case (but not for ERESTARTSYS, but that is another problem). [SCTP] Peeled off/accepted sockets not in the right bind_bucket. hlist changes caused the peeloff testcase to fail. Investigation shows that the peeloff sockets is not bound into bind_bucket, so the bucket has gone away (original socket closed). Fixing this, shows a problem that inet->num wasn't set on peeled off sockets, so autobind kicks in creating a new bind_bucket. Ugh. One bug had been hiding the other one all this time. Fix 1) peeledoff/accepted sockets need to have their own socket woven into the bind_bucket->owner list. 2) Set inet->num, so autobind doesn't think it needs to kick in. [NET] remove skb_linearize from igmp.c [PATCH] Unionize IO-APIC registers [ Registers of the world, unite! ] This makes the IO-APIC data structures use unions, so that we can cleanly access them both as flat "raw" values, and as the bitmap sub-entries. SCSI: abstract mode_sense 6 and 10 out completely Move the mode_sense request routines to a central location and make all block device consumers use it. Also abstract the header as part of the return to hide the 6/10 differences. scsi_mid_low_api.txt update From Douglas Gilbert [PATCH] remove an unused variable from fd_mcs.c The patch belowremoves an unused variable from drivers/scsi/fd_mcs.c . I've tested the compilation with 2.5.72-mm2. [PATCH] remove an unused function declaration from sym53c416.h The patch below removes a declaration for a function that is no longer present. I've tested the compilation with 2.5.72-mm2. [PATCH] ibmmca cleanup The patch below does the following: - remove an unused static function - removes the declaration of a function that is no longer present - removes a variable declaration that shadows a function parameter I've tested the compilation with 2.5.72-mm2. [PATCH] remove an unused function from nsp32.c The patch below removes an unused function from nsp32.c . I've tested the compilation with 2.5.72-mm2. [PATCH] seagate cleanup The patch below does the following cleanups on drivers/scsi/seagate.{c,h}: - remove two unused functions - remove a function declaration for a function that is no longer present I've tested the compilation with 2.5.72-mm2. [PATCH] Trivial patch for scsi_error.c From: Alan Stern [PATCH] check for can_queue != 0 in scsi_host_alloc This is to give a proper warning if someone tries to load an unconverted old-style driver. [PATCH] REQUEST SENSE 254->252 byte response SPC-3 (rev 13) says that 252 bytes in the maximum (and recommended) length for a REQUEST SENSE reponse. Linux asks for 254 bytes in scsi_error.c . That number was not specified in SPC-2 (although the allocation length field is 1 byte thus limiting it to 255). Seems as though some numbers are being rounded down to be multiples of 4. [PATCH] NCR53C9x compile fix From: Geert Uytterhoeven [PATCH] integer constants in sym53c8xx_2/sym_glue.c too big for int The patch below postfixes a constant in sym53c8xx_2/sym_glue.c with ULL, on 32 bit archs this constant is too big for an int. [PATCH] postfix two constants in ips.c with ULL The patch below postfixes two constants in ips.c with ULL, on 32 bit archs this constant is too big for an int. The cast doesn't do the right thing, 0xffffffffffffffff is in C an int and the cast casts 0xffffffffffffffff interpreted as an int to an u64. [PATCH] fix return value of scsi_device rescan attribute When writing to the rescan attribute, return count as result, otherwise some user apps might retry the write forever. And remove the read capability of the attribute. [PATCH] clean up aacraid use of SCSI constants [PATCH] small patch for sym53c8xx_2 I thought we had an agreement for a long time already that the use of dma64_addr_t in the sym53c8xx_2 driver was wrong, but it's still there. [PATCH] Fix scsi host attributes The shost_attrs stuff looks fine, expect for two points. 1. The scsi_sysfs_modify_shost_attribute() and scsi_sysfs_modify_sdev_attribute() functions appear to be written a bit carelessly. Below is a patch that: permits modification of the first attribute in the list, allocates a new list with entries having the correct size, copies the correct number of entries from the old list, and wraps excessively long source lines. 2. More importantly, the current organization of the code has a serious problem. The SCSI core does not modify the host driver when the reference count for either shost->class_dev or shost->host_gendev drops to 0. Without knowing that, it is unsafe for the driver ever to deallocate a private host data structure, since a user process may continue to hold a reference to an open attribute file indefinitely, even after scsi_unregister() has returned. [PATCH] add scsi_driver.h include/scsi/scsi_driver.h contains æverything related to upper level drivers. Unlike the other header moves there's no compatiblity this time as it's easy to fix up the few users. [PATCH] make sym53c8xx_2 not reject autosense IWR This patch makes sym53c8xx_2 silently ignore the Ignore Wide Residue message on autosense commands rather than rejecting it. This makes the SCSI communications cleaner for targets that return an odd number of sense bytes. [irda] add driver for mips Alchemy Au1000 SIR/FIR Submitted by Ralf Baechle [netdrvr] add driver "meth", for SGI O2 MACE fast eth [netdrvr] sgiseeq update [netdrvr] update ioc3_eth [netdrvr] update declance [netdrvr] au1000_eth update [netdrvr] update sb1250-mac [netdrvr tulip] add mips cobalt support [netdrvr] misc small mips updates Add missing CONFIG_TC35805 entry to Kconfig. Update CONFIG_NET_SB1250_MAC Kconfig entry. Minor cosmetic updates to gt96100eth. ppc64: remove ERESTARTNOHAND code in syscall exit, from ppc32 [PATCH] PCI: more PCI gubbins I noticed we have a couple of redundancies in drivers/pci/Makefile, have a patch... [PATCH] PCI: fixes for pci/probe.c - Combine pci_alloc_primary_bus_parented into pci_scan_bus_parented. - Move the EXPORT_SYMBOL for pci_root_buses up to its definition. - Don't EXPORT_SYMBOL pci_scan_bus since it's a static inline. - Add the pci_domain_nr() to the sysfs name for this bus. [PATCH] PCI Hotplug: fix core problem with kobject lifespans. Added needed release function, now all pci hotplug drivers need to implement it... [PATCH] IBM PCI Hotplug: fixes found by sparse PCI Hotplug: add fake PCI hotplug driver. Useful for testing hotplug issues with pci drivers. [CPUFREQ] speedstep cleanups. From Dominik Brodowski. - separate functions which might be useful for speedstep-piix4, speedstep-via or speedstep-bios into a new speedstep-lib.c. - clean up speedstep-ich as it will only be used for ICH-based chipsets in future. - remove now-obsolete speedstep_coppermine parameter [CPUFREQ] Speedstep support for P4M/533 From Dominik Brodowski. Add support for newest "Mobile Pentium 4-M" and the new "Mobile Pentium 4 with 533 MHz FSB" processors. [CPUFREQ] Silence debug output on centrino speedstep driver. [AGPGART] Add missing pte masking in NVIDIA nForce driver. Noticed by Marcelo Penna Guerra. [NET][IPMR] ipmr fixes These patches fix ip multicast route (ipmr) on 2.5.73. 1 - Trivial C99 initialization 2 - Change functions/variables to static 3 - Drop and reacquire RTNL in error path 4 - Use time_after() 5 - Use alloc_netdev 6 - Fix OOPS on dropped packets 7 - Get rid of skb_linearize Tested on 8-way SMP by bringing up pimd. [AGPGART] SiS 655 support. Needs more testing, especially in x8 mode. [PATCH] GCC 2.95.4 needs the spinlock workaround From: Mikael Pettersson 2.5.73 removed the workaround needed to prevent gcc-2.95.x from miscompiling spinlocks on UP. It turns out that some versions of gcc-2.95 still do have problems with empty structs, so re-introduce the workaround. [PATCH] show_stack fix sysrq-T currently displays the same stack trace for every process. Teach show_stack() to look in the passed task_struct first if the caller did not pass in a stack address. [PATCH] setscheduler needs to force a reschedule From: Robert Love Basically, the problem is that setscheduler() does not set need_resched when needed. There are two basic cases where this is needed: - the task is running, but now it is no longer the highest priority task on the rq - the task is not running, but now it is the highest priority task on the rq In either case, we need to set need_resched to invoke the scheduler. [PATCH] misc fixes - unused var warning in pnp_add_irq_resource() - unused var in fd_mcs.c (Adrian Bunk) - remove dead code in i810_main.c (a u8 cannot exceed 255) (Adrian Bunk) - don't truncate dma_addr_t's in gdth.c (Jes Sorensen) - vgastate.c needs vmalloc.h (Geert Uytterhoeven ) - revert bogus test from rd_open() - it was a 2.4 forward-port, and 2.5 doesn't need it. (The gendisks aren't registered, we cannot get there anyway). [PATCH] AT_SECURE auxv entry From: Stephen Smalley This patch adds an AT_SECURE auxv entry to pass a boolean flag indicating whether "secure mode" should be enabled (i.e. sanitize the environment, initial descriptors, etc) and allows each security module to specify the flag value via a new hook. New userland can then simply obey this flag when present rather than applying other methods of deciding (sample patch for glibc-2.3.2 can be found at http://www.cs.utah.edu/~sds/glibc-secureexec.patch). This change enables security modules like SELinux to request secure mode upon changes to other security attributes (e.g. capabilities, roles/domains, etc) in addition to uid/gid changes or even to completely override the legacy logic. The legacy decision algorithm is preserved in the default hook functions for the dummy and capability security modules. Credit for the idea of adding an AT_SECURE auxv entry goes to Roland McGrath. [PATCH] common name for the kernel DSO From: David Mosberger Give the ia32 vsyscall DSO the same name as ia64's. [PATCH] get_unmapped_area() speedup From: Ingo Molnar Apparently our thread-creation performance has gone down the tubes again, because the mm.free_area_cache search heuristic broke. The initial, more naive hole-cache patch helped the testcode in the beginning. Then after some point glibc started creating a 'small hole' in the vmas, which hole was _below_ the thread stacks, and which hole thus prevented the intended operation of the cache. The new code solves the problem by relaxing the 'smallest address hole cache' rule a bit, the cache is now not re-set at every get_unmapped_area() time, it's only re-set during unmaps. It's also re-set if there are no allocatable mappings at all - ie. correctness is not affected. [PATCH] dentry->d_count fixes: d_invalidate From: Maneesh Soni - d_invalidate() can incorrectly return success instead of returning -EBUSY as we can have situations where lockless d_lookup has found a dentry successfully before d_invalidate drops it [PATCH] dentry->d_count fixes: nfs_unlink From: Maneesh Soni - nfs_unlink() can race with lockless d_lookup() as d_lookup() can successfully lookup a dentry for which nfs_unlink() can assume that no one else is using and can go ahead and do nfs_safe_remove() on it. By using per dentry lock, it is solved as we d_lookup() will fail the lookup for unhashed dentries. [PATCH] dentry->d_count fixes: hpfs From: Maneesh Soni - hpfs_unlink() can race with lockless d_lookup(), as we can have situations where d_lookup() has successfully looked-up a dentry and at the sametime hpfs_unlink()--->d_drop() has dropped it. Taking the per dentry lock before checking the d_count in hpfs_unlink() solves this race condition. [IPV6] DAD has to be destined to solicited node mulitcast address. Check if DAD is destined for solicited node multicast address as RFC2461 required. [PATCH] workaround for smb_proc_getattr oops From: Zwane Mwaikambo smbfs tends to oops over a null server->ops->getattr in smb_proc_getattr(). Urban says: In 2.5 the server->ops is initialized when smbfs gets a connection, to match whatever the server is capable of or not. This happens after the mount syscall so smbfs is then mounted but not usable. Not sure if smb_lookup is always called before any other operation that uses server->ops. If it is then it would be enough to have a test there. Otherwise I will just change all users of server->ops to verify the pointer first and return -EIO or something. Al Viro says (paraphrasing) ugh, we need mount2(). It doesn't look to me like we'll be getting mount2() in the 2.6 timeframe. Zwane's patch implement's Urban's workaround. [PATCH] Column counting fix in n_tty.c From: Chris Heath Here's a column counting bug that was lurking in a corner of n_tty.c. [PATCH] normalise node load for NUMA From: Andrew Theurer This patch ensures that when node loads are compared, the load value is normalised. Without this, load balance across nodes of dissimilar cpu counts can cause unfairness and sometimes lower overall performance. For example, a 2 node system with 4 cpus in the first node and 2 cpus in the second. A workload with 6 running tasks would have 3 tasks running on one node and 3 on the other, leaving one cpu idle in the first node and two tasks sharing a cpu in the second node. The patch would ensure that 4 tasks run in the first node and 2 in the second. I ran some kernel compiles comparing this patch on a 2 node 4 cpu/2 cpu system to show the benefits. Without the patch I got 140 second elapsed time. With the patch I get 132 seconds (6% better). Although it is not very common to have nodes with dissimilar cpu counts, it is already happening. PPC64 systems with partitioning have this happen, and I expect it to be more common on ia32 as partitioning becomes more common. [PATCH] compat_sys_old_getrlimit() depends on From: David Mosberger compat_sys_old_getrlimit() depends on sys_old_getrlimit() and the patch below updates the guarding #ifdef accordingly. [PATCH] ext3: fix page lock vs journal_start ranking bug ext3_block_truncate_page() is calling grab_cache_page() inside a JBD transaction. This is wrong, because transactions nest inside lock_page(). The deadlock is against shrink_list->ext3_journalled_writepage->journal_start. This was not noticed before because we never used to journal writepage() data in journalled-data mode. And because the deadlock against generic_file_write() is covered up by i_sem. Rework things so that we lock the page prior to starting a transaction. [PATCH] ext3: fix memory leak We need to unconditionally brelse() the buffer in there, because journal_remove_journal_head() leaves a ref behind. release_buffer_page() does that. Call it all the time because we can usually strip the buffers and free the page even if it was not marked buffer_freed(). Mainly affects data=journal mode [PATCH] correct mail addresses for visws support From: Andrey Panin this trivial patch changes mailing list address for visws subarch support along with some occurences of my old email addresses. [IPV6] DAD must not have source link-layer option Check if DAD does not have source link-layer address option; RFC2461 7.1.1. [PATCH] NCR53C9x compile fix From: Geert Uytterhoeven NCR53C9x SCSI: Fix compilation after breakage in 2.5.71 [PATCH] htree: set the dir index bit in the right place From: Alex Tomas Don't set the directory's index flag until we know that we're doing ahead with the directory modification. [PATCH] export flush_tlb_all for drm modules From: Jan Dittmer This adds an export for flush_tlb_all to i386_ksyms.c. The drm modules miss this, when compiling for SMP. I changed Jan's patch to use EXPORT_SYMBOL_GPL. [PATCH] Typo after 8250_cs update (SERIAL) From: Francois Romieu s/GetConfiguration/GetConfigurationInfo/ [PATCH] fix type in compat_sys_fcntl64 [Pointed out by Bjorn Helgaas via Arun Sharma] This fixes an obvious cut and paste error in my original patch. SCSI 53c700: add module_exit routine otherwise module isn't removable Add remove method to lasi700.c [NET]: Define LL_RESERVED_SPACE in terms of HH_DATA_MOD. [NETFILTER]: Fix leaks in error paths of ip_recent_ctrl. [IPV6]: Inappropriate static variable in net/ipv6/ndisc.c [NETFILTER]: Missing return in arp_packet_match(). [NETFILTER]: Add arptables mangle module. [SPARC]: Fix LIBS_Y. [SPARC]: Do not use __builtin_trap() on sparc until gcc is fixed. [TCP]: Sanitize initcwnd calculation, add new metrics. 1) Remove srtt etc. tests in tcp_init_cwnd, was buggy anyways. 2) Add RTAX_INITCWMD route metric, use this in tcp_init_cwnd() if non-zero. 3) Add RTAX_FEATURES (to enable/disable ECN/SACK/TIMESTAMPS on a per-route basis), currently unused. [IPV6]: Make several ndisc private stuff static. DiskOnChip Millennium Plus translation layer fixes: - Fix geometry reporting. - Avoid endless loop when deleting a Virtual Unit Chain. Add missing prototype in drivers/mtd/chips/gen_probe.c [PATCH] x86-64 merge for 2.5.73 Bring the x86-64 architecture code uptodate for 2.5.73. One problem is that CONFIG_IDE_TASKFILE corrupts file systems on AMD 8111. It may be worth to disable it in the configuration. - Disable 32bit vsyscalls for now until all bugs can be fixed - Fix warnings - Fix NULL pointer reference in 32bit ptrace - Timing fixes from John Stultz - Sync show_stack prototype - Sync nmi and floppy.h code (Mikael P.) - Set proper defines for AGP - Make Simics work again - Scale unit in 32bit sysinfo (originally from ppc64) [PATCH] Fix ACPI compilation for 2.5.73 Without this patch ACPI won't compile on AMD64 because of mismatched prototypes. Andrew stated that this is the way he wants it to be fixed. [PATCH] Change 64bit epoll ABI for AMD64 As discussed earlier. The 64bit epoll ABI on AMD64 is changed to match 32bit. This way we avoid emulation overhead. To catch old binaries I allocate new syscall slots. [PATCH] remove *_segments() dummy functions again Last November Linus applied a patch to remove the now-unused *_segments() functions from all architectures ... but some of the newer architectures escaped and still have them. SCSI: Add missing scsi/scsi_driver.h file [PATCH] Make SCSI selfcontained Move all the SCSI Kconfig bits from arch/*/Kconfig into driver/scsi/Kconfig. Also add notes about FireWire & USB. [PATCH] PCI Hotplug: pcihp_skeleton: fix delete bug and add release() callback [PATCH] PCI Hotplug: acpiphp: add release() callback [PATCH] PCI Hotplug: cpci: fix delete bug and add release() callback [PATCH] PCI Hotplug: cpqphp: add release() callback and other minor cleanups. [PATCH] PCI Hotplug: ibmphp: add release() callback and other minor cleanups [PATCH] aha1740 update The included patch updates the old aha1740 driver to some of the 2.5 APIs : - Generic DMA API - EISA probing API - SCSI hotplug API It's been heavily tested on x86 (UP and SMP) as well as alpha (UP) for the past two months. [PATCH] PCI: i386/pci/direct.c fixes - Request resources before using them - Don't allocate GFP_KERNEL memory with interrupts disabled - Split pci_direct_init() into three functions to prevent it from getting too long. [PATCH] PCI: create pci_name() This patch introduces pci_name() and converts slot_name into a pointer to dev.bus_id. [PATCH] define KCORE_ELF in m68knommu/Kconfnig Define missing type for CONFIG_KCORE_ELF in m68knommu/Kconfig. [PATCH] fix do_settimeofday() for 'struct timespec' argument Modify m68knommu do_settimeofday() routine to take "struct timespec" argument, and adjust code to handle nsec size quantities. [PATCH] Update IEEE1394 (r972) IEEE1394 : Kbuildify oui2c. ALL : C construct cleanups, macro namespace cleanups. ETH1394 : Limited multicast support. Minor fixes for IPv4 interop. ETH1394 : Add ethtool support. [IPV6] Fixed fragment check in ip6_output.c:ip6_fragment() MTU / alignment check in ip6_fragment() was wrong; first_len was not correct. [NET] Remove some 0 initializers. These small patches remove a number of '.maxlen = 0,' initializers. [XFRM] Set port/proto in acquire messages. Also append policy spec to acquire message. [PATCH] use generic dma direction bits in scsi Make the *_data_direction flags of type enum dma_data_direction from the generic dma mapping API. The SCSI_DATA_* defines are kept for compatiblity and map to the dma mapping API values now. scsi_to_pci_dma_dir and scsi_to_sbus_dma_dir are kept as stubs - we define them to nothing as both the PCI and SBUS constants have the same values as the generic dma mapping ones. Unfortunately we can't check anymore whether someone messes this up as the values are enumes now and not cpp defines. Fix scsi drivers needing to include With the simplification of the DMA direction/SCSI data direction linux/pci.h is no longer pulled in by scsi.h Remove linux/pci.h dependency in SCSI 53c700 The consolidation of the SCSI data direction flags eliminate the need for it. Fix SCSI data direction issues in aha1740 merger and also remove no longer used .command entry Fix up data direction in SCSI abstracted mode sense PCI Hotplug: acpiphp: stupid typo fixes. SCSI: Move can_queue == 0 check Some HBA drivers don't set this until their ->detect routines, so don't check for it until scsi_add_host Fix misreporting of card type and spurious "already scheduled" messages. [NET]: Kill skb_linearize() and bogus feature flag settings in eth1394.c [TCP]: Handle lack of cached dst in tcp_init_cwnd(). [TCP]: If we have a lot of time-wait sockets to kill, do it via workqueue. [PATCH] Fix aacraid status returns aacraid was using the old and erroneously unshifted status return codes. Switch it to use the new SAM_STAT_ codes that don't need shifting. [PATCH] remove a bogus check in sym2 driver The sym2 driver has a check for a very broken bios where it seems devices appear twice. This test is broken on a machine with PCI domains since busnumber:devfn is not a unique identifier. When we move to the scsi hotplug interface all this code will go away but in the short term we need the following fix. [PATCH] Use new svc_export_show to implement e_show for /proc/fs/nfs/exports Also remove the path_buf that was passed around for /proc/fs/nfs/exports as the existance of seq_path removes the need for this. [PATCH] Remove path buffer passed around by cache_show routines this was need for paths, but now we have seq_path... [PATCH] NFSv4 server OPEN_CONFIRM From: "William A.(Andy) Adamson" Add OPEN_CONFIRM functionality to the nfsv4 server, including a preprocess_seqid_op function that will be used in subsequent patches. [PATCH] NFSv4 server Close state From: "William A.(Andy) Adamson" Add share state processing to nfsd4_close. it includes some more debug counters. [PATCH] NFSv4 server - open-downgrade From: "William A.(Andy) Adamson" Add the opendowngrade call with share state processing. it includes nfs4_preprocess_stateid_op() which will be used in read and write state processing. [PATCH] NFSv4 server - Read "share" state From: "William A.(Andy) Adamson" this patch adds share state processing to nfsd4_read(). [PATCH] NFSv4 server - Write "share" state From: "William A.(Andy) Adamson" this patch adds share state processing to nfsd4_write(). [PATCH] NFSv4 server - setattr share state From: "William A.(Andy) Adamson" this patch adds share state processing to nfsd4_setattr(). [PATCH] NFSv4 server - missing locking From: "William A.(Andy) Adamson" sorry. i forgot to lock the nfsv4 state in nfsd4_read, nfsd4_write, and nfsd4_setattr. [PATCH] NFSv4 server - Removed duplicate #define Already defined in state.h [PATCH] Provoide refrigerator support for nfsd Call refrigerator() in svc_recv, so it is there for nfsd and no longer needed explicitly in lockd. [PATCH] remove IO APIC newline This patch is to 2.5.73-bk4 and is purely cosmetic. Please apply. It removes the blank line after "testing the IO APIC....": [PATCH] Fix compile with !CONFIG_VT Architectures using the generic 32/64-bit ioctl() compatibility shims will get a link error if CONFIG_VT is not defined, since the compatbility ioctl() code calls functions in drivers/char/vt.c which is only included in the build if CONFIG_VT is set. This fixes the compile with a couple of #ifdefs: [PATCH] fix arguments of show_stack() Fix the m68knommu version of show_stack(). It expects the task as an argument. [PATCH] conditional ROMfs copy for M5206eLITE board Make the ROMfs copy in the startup code for Motorola/M5206eLITE board conditional on actually using a ROMfs setup. [PATCH] conditional ROMfs copy for NETtel/5272 board Make the ROMfs copy in the startup code for NETtel/5272 board conditional on actually using a ROMfs setup. [PATCH] Use Local Percpu Macros for Local Percpu Variables In general, it is more better to use get_cpu_var() and __get_cpu_var() to access per-cpu variables on this CPU than to use smp_processor_id() and per_cpu(). In the current default implemention they are equivalent, but on IA64 the former is already faster, and other archs will follow. [PATCH] s390: base fixes - Remove unused variables from smp.c. - Reserve system call number 110 for sys_lookup_dcache. - Fix show_trace. - Remove superfluous asm include file. - Add statfs64 structure. [PATCH] s390: 31 bit compat. - Add missing includes to compat_ioctl.c. - Fix 32 bit emulation of sys_settimeofday. [PATCH] s390: common i/o layer. - Make ccwgroup online attribute consistent with ccw online attribute. - Add link incident record handling to channel subsystem code. - Do path grouping only if the device driver explicitly requests it. - Fix multicast or broadcast flood ping hand on HiperSockets. [PATCH] s390: dasd driver. - Simplify long busy condition handling, add quiesce/resume ioctl. - Add sense data area to reserve/release/steal_lock ccw-chains. [PATCH] s390: set module owner. Pass correct argument to SET_MODULE_OWNER. [PATCH] s390: typos. Fix typos. [PATCH] s390: tty3215_init. Add return statement to tty init function of 3215 driver. [PATCH] [KCONFIG] Make cdrom Kconfig selfcontained This patch moves all the duplicated cdrom Kconfig bits from arch/*/Kconfig to drivers/cdrom/Kconfig [ALPHA] Define memset via symbol copy for ev6 as well. Cset exclude: cweidema@indiana.edu|ChangeSet|20030621162021|08529 [PATCH] ide: TCQ initialization fixes - do not enable TCQ in ide_init_drive(), its too early - enable TCQ in __ide_dma_on() only if CONFIG_BLK_DEV_IDE_TCQ_DEFAULT=y - try to enable TCQ only on disk drives - correct check for two drives on one channel [PATCH] ide: fix IRQ handler returns Make the IDE driver return the proper status return for unhandled interrupts. [PATCH] ide: fix drive->unmask handling for taskfile PIO [PATCH] ide: proper allocation of hwif->io_ports resources [PATCH] ide: tiny cleanup of "ideX=ata66" parameter handling in ide_setup() hwif->udma_four is always preset to 0 [PATCH] ide: remove dead code from ide_raw_build_sglist() [PATCH] ide: remove dead and broken ide_diag_taskfile() variant [PATCH] ide: tiny cleanup of ide_init(), it is called only _once_ [ALPHA] No, really fix memset. Really. [PATCH] module ref counting for airo.c clean up airo.c: remove MOD_(INC|DEC)_USE_COUNT, set the owner field instead. compile tested only. against 2.5.73-bk [PATCH] strip.c: don't allocate net_device as part of private struct hi jeff cleans up strip.c not to allocate struct net_device as part of the private structure. use a separate kmalloc (and kfree). compile tested only. against 2.5.73-bk -daniel [PATCH] alloc_etherdev for netwave_cs.c erm...i didn't actually compile it... sorry. corrected patch below. -daniel On Fri June 27 2003 00:45, Daniel Ritz wrote: > cleans up netwave_cs.c to use alloc_etherdev instead of allocating the > device out of the private data structure. compile tested only. > against 2.5.73-bk [PATCH] fix use-after-free in e100 [NET]: Scale DST/ipv6 intervals like we did for ipv4. [NET]: Fix PPP async regression. Fix PPP over async regression that the PPPoE changes caused. Basically, PPP puts a zero length skbuff in the receive queue as an error token, and the last change caused that to get flushed as bad data. Thanks to Diego Calleja Garcia, Matthew Harrell for validating this. [PATCH] fix an aha1740 merge error it should use scsi_host_alloc/scsi_host_put now. [ALPHA] Set wall_to_monotonic on timer_init and settimeofday. [ALPHA] Eliminate struct declared in prototype warnings in asm/elf.h. Eliminate really old ncr53c8xx driver - Remove direct configuration of this driver - Only remaining use is in the parisc zalon driver - Split the driver up to make it more use to 720 based drivers - Update zalon to new module/SCSI methods [PATCH] avoid Oops in net/core/dev.c This fixes the kernel crash in the case when we do an SIOCSIFNAME ioctl on /proc/net/dev to rename a network interface, and we supply a string such as "foo%sbar". [PATCH] check for presence of readpage() in the readahead code Make sure that the address_space is capable of performing the readahead before going in and allocating the pages. [PATCH] Fix reiserfs BUG From: Oleg Drokin Fix BUG_ON(!buffer_mapped(bh)) in submit_bh() Reiserfs is submitting unmapped buffers into submit_bh(), which trips a BUG. [PATCH] Fix syslog(2) EFAULT reporting From: Andi Kleen Add proper EFAULT reporting to sys_syslog. This fixes some silly LTP test in the 32bit emulation of an AMD64 kernel. [PATCH] Fix ide-cd rw mounts From: Kouichi ONO in 2.5.73, I can't read/write mount DVD-RAM via ide-scsi (only r/o mount). Without ide-scsi, I can mount DVD-RAM read/write mode and works fine. (acked by Jens). [PATCH] Remove check_region and MOD_*_USE_COUNT from From: Bob Miller - Replace the call to MOD_INC_USE_COUNT with a __module_get() when forcing the module to not be unloadable. - Remove the check_region() calls and restructured things to only use request_region(). [PATCH] Remove racy check_mem_region() call from arc-rimi.c From: Bob Miller Removed the check_mem_region() call and replaced with request_mem_region(). Because of the way the driver is structured the first request_mem_region() call gets the default memory area. After probing the complete memory area that is needed to communicate with the device is known, so the first memory area is released and the complete area is requested. [PATCH] kmem_cache_destroy() forgets to drain all objects From: Philippe Elie kmem_cache_destroy() can fail with the following error: slab error in kmem_cache_destroy(): cache `xxx': Can't free all objects but the cache user really free'd all objects This is because drain_array_locked() only frees 80% of thge objects. Fix it by adding a parameter to drain_array_locked() telling it to drain 100% of the objects. [PATCH] ext3: remove the version number The ext3 version number hasn't been updated since ext3 was merged. We track ext3 via the kernel release ID. Remove the ext3 version number. [PATCH] cdrom eject scribbles on the request flags From: Nick Piggin It converts writes to reads and we end up returning the request to the wrong queue. [PATCH] nmclan_cs compile fix nmclan_cs doesn't compile with older compilers [SPARC64]: Fix OBP version parsing on newer systems. [SPARC64]: Fix build error from OBP parsing patch. Add NCR Quad 720 SCSI driver This adds support for the Q720 MCA SCSI driver. It's only really useful on the voyager platform. The driver is preliminary and currently makes no use of the Q720 on-board memory for storing scripts and ccbs. Fix logic reversal in scsi_host_alloc [IPV6] Don't set M flag in last fragment. [IPV6] Use macro for M-Flag and clean-up. Use macro IP6_MF for the M-Flag. Clean-up for readability (commented by ). [IPV6] Convert /proc/net/ip6_flowlabel to seq_file. [XFRM] Fix typo. [XFRM] Set SA saddr correctly The source address of an SA is not necessarily equal to the source address in the selector. This patch addresses this problem. [IPSEC] split xfrm_state_replace + fixes Split xfrm_state_replace into xfrm_state_add and xfrm_state_replace. Fixes: 1. Only update update lifetime and encap options if the state is valid. 2. Disallow updates to states that do not exist. 3. Bail if afinfo cannot be found. This brings SADB_UPDATE in line with what is required by RFC2367. It is also needed by SFS NAT-T support as it needs to update valid states when the encap ports move. I've tweaked the logic slightly so that SADB_UPDATE will fail on a larval state that hasn't undergone SADB_GETSPI. This is what RFC2367 calls for and it simplifies the code in that we don't have to call find_acq for SADB_UPDATE. This doesn't affect any of the three KMs as they either don't use SADB_UPDATE or call SADB_GETSPI before doing an update. [PATCH] Update IEEE1394 (r986) ETH1394 : - Minor cleanups. - Fix incorrect error handling with no broadcast channel. - Set max/default MTU to 1500. - Fix broken spinlock handling on device open. IEEE1394 : Make bus-nodeid printk's of a consistent format. OHCI1394 : Change one printk to a verbose/debug only message. [NET]: Add EDP2 ethernet protocol ID. Fix scsi device starvation handling Doing some debugging after the SDET failure, I found that if we reject a command at zero depth, we can get into the infinite loop where we refuse a command to a LUN because it is starved, but there's no returning command to run the SCSI queues and fix the starvation. The solution is to accept the command and remove the LUN from the starved list. [PATCH] scsi_ioctl_send_command fix The SCSI_IOCTL_SEND_COMMAND ioctl does not work in 2.5.73-bk7. The following patch (possibly mangled by my mail client) does fix the problem (introduced in 2.5.73): [IPV6] put ipv6_rcv_saddr_equal() common place [NET] fixed /proc/net/raw{,6} seq_file support [IPV6] Fix bug in /proc/net/ip6_flowlabel seq_file conversion [ALPHA] Set HAE-4 for SABLE/LYNX From Jay Estabrook: > Here's a real short one that fixes a GENERIC build problem introduced > in pre2, and finalizes the SABLE/LYNX support setting of HAE-4, making > X servers finally work on those platforms. [PCMCIA] Move "owner" field to pcmcia_socket. Move ->owner field from socket operations to pcmcia_socket. (This change is mainly for the SA11xx drivers, which use a core driver for the chip, and a separate module for all the machine specific bits.) ppc64: any_online_cpu returns NR_CPUS on fail [PATCH] USB storage: unusual_devs.h cleanups This patch re-sorts unusual_devs.h and removes some duplicate entries. [PATCH] USB storage: create associate_dev(), more US_*_DEVICE printout This patch creates an associate_dev(), which is the mirror to disassociate_dev(). It also adds some more printout when identifiying devices that qualify for US_*_DEVICE codes so we need less interaction with the reporting user. [PATCH] USB storage: avoid sending URBs when disconnecting This patch avoids sending URBs which are part of the reset system when the device is being disconnected. It also makes the control thread avoid starting anything for a device in that state. [PATCH] USB storage: create private I/O buffer This patch creates a private I/O buffer for the 'core' subdrivers. This allows us to fix all cases of DMAing to/from the stacks, statics, or other bad places. [PATCH] USB storage: Cosmetic cleanups This is some cosmetic cleanups. Things like changing dr to cr (they used to stand for usb_devrequest, which is now usb_ctrlrequest), using strlcpy instead of strncpy, and modifying some comments. [PATCH] USB storage: General purpose I/O buffer allocation and management This patch makes our private I/O buffer allocated such that it's pre-mapped for DMA. We then add some logic to make sure that we don't try to re-map it. We also make the size of the buffer large enough for other sub-drivers, which will be converted shortly. [PATCH] USB storage: logic cleanup This patch cleans up the logic involved in setting URB flags a bit so it's more evident that some flags are always set and other flags are only set under certain conditons. [PATCH] USB: zaurus SL-C700 This is a patch for usbnet for working with Zaurus SL-C700. The productid is different from other Zaurus, so i add an entry for it :) [PCMCIA] Ref-count the socket driver module on card insertion/removal. Get/Put module when we insert and remove a card. This avoids a potential deadlock when socket drivers are unloaded, and we have a cardbus card known to the system. [PCMCIA] Remove original module use accounting in register_callback. [PCMCIA] Add work-around for bouncy card detect signals. Add work-around for i82365-based socket drivers to the core PCMCIA code. Since insert processing is not a time critical event, we can afford to delay (by sleeping) these for everyone. [PCMCIA] Allow socket drivers call pcmcia_parse_events directly. Remove register_callback methods. Instead, allow socket drivers pass their events directly into cs.c via pcmcia_parse_events(). [PCMCIA] Remove now obsolete work queues, spinlocks and variables. Since the cs.c pcmcia core has a per-socket thread, we don't need each socket driver having a work queue and associated complexity. Therefore, remove this complexity and allow the socket drivers to call pcmcia_parse_events directly from their interrupt or timer handlers. [ALPHA] Fix SETTLS -- read TLS value to install before clobbering it. [PATCH] USB: turn down some debugging messages in uhci-hcd. [PATCH] move sg_dma_ macros out of asm-i386/pci.h These macros belong in asm-i386/scatterlist.h instead. As the headers are disentangled this has shown up as a problem with my MCA SCSI drivers since they no-longer include asm/pci.h in any form but need to traverse the scatterlist. [PATCH] update show_stack() in voyager for new prototype When show_stack() was changed to take two arguments, the use in the voyager code was not converted. This patch makes the correct conversion. [PATCH] pcmciamtd update Resync with CVS. Minor update including better handling of device removal. [PATCH] conditional ROMfs copy for Motorola M5282EVB board Make the ROMfs copy in the startup code for Motorola M5282EVB board conditional on actually using a ROMfs setup. [PATCH] fix compile warnings ColdFire PIT timer Include interrupt.h in the Motorola ColdFire PIT timer code. Cleans up compiler warning. [PATCH] selection of boot parameters at configure time for Motorola 5307 targets Allow selection of boot arguments at configure time for the Motorola ColdFire 5307 targets. [PATCH] fix return value after hugetlb mmap failure At the moment no one uses is_aligned_hugepage_range or is_hugepage_only_range, but it is reasonable to assume they return true or false. On error we want to return -EINVAL back to userspace. [IPV6] Make ipsec tunnel work with ext hdrs I noticed that using extensions headers along with IPsecv6 tunnel mode did not work in 2.5.73 + patch-2.5.73-bk3. The following patch checks "nexthdr" instead of "iph->nexthdr" which could be an extension header. I tested this with tunnel mode and transport mode with and without extension headers and it worked ok. Fix the code that checks for PCI IDE controller "native" vs "legacy" modes. It used to think that IDE controllers that weren't marked as "IDE storage" PCI class (eg CMD IDE RAID controllers mark themselves as "RAID storage") were always using legacy mode interrupt routing, which is bogus. In fact, the whole native vs legacy thing only makes sense for the IDE storage class, so fix the test to match. [BRIDGE]: Ethernet bridge fixes. 1. STP protocol has no security, so malcontents can fuck with the bridge's topology. The fixes are to ship with STP turned off to protect the ignorant, and run STP packets through ebtables netfilter for the smart. Got this one via a russian hacker "Oleg K. Artemjev" before he published the paper. Bridge netfilter still needs work to give a nice face on this but this patch gives the hooks to filter. 2. STP input processing was lax in it's length checking so I bet you could make up a bomb packet. My inspection while doing #1. 3. Forwarding table could be abused by sending forged packets with bogus source address same as the local host. This came via Lennart from Jerry Kreuscher who ran into it by mistake. [NET] convert /proc/net/igmp to seq_file [NET] convert /proc/net/igmp6 to seq_file [NET] convert /proc/net/mcfilter to seq_file [NET] convert /proc/net/mfilter6 to seq_file [NET] convert /proc/net/anycast6 to seq_file [IPV6]: Fix two ipv6_addr_addr failure checks. [IPV6]: One too many ipv6_addr_type() calls in ndisc_recv_ns(). [PATCH] conditional ROMfs copy for Arnewsh 5307 board Make the ROMfs copy in the startup code for ARNEWSH 5307 board conditional on actually using a ROMfs setup. [PATCH] rework 68360 interrupt handling code Rework the interrupt handling code for the m68knommu 68360 sub-architecture. This is inline with the interrupt changes to the other m68knommu CPU's. Makes this code be the primary interrupt setup for a 68360 based kernel, not just a set of support functions. [PATCH] clean up m68knommu bitops.h Rework the bitops functions for the m68knommu arhcitecture. Change all addr pointers to be "unsigned long", cleans up many compiler warnings. [PATCH] fix ColdFire 5249 dma support The m68knommu ColdFire 5249 CPU has the same DMA engine as the 5307 and 5407 ColdFire CPU's. Modify m68knommu dma.h to reflect this. [PATCH] fix broken MARK parity define for ColdFire UART Fix broken definition of MARK parity for ColdFire UART. [PATCH] Identify Code Section Of Modules for kallsyms Remember the size of the SHF_EXECINSTR sections, which are conveniently at the start of the modules, and use that to more reliably implement module_text_address(). [PATCH] Make runqueues a per-cpu variable Makes scheduler use per-cpu variables for the runqueues. [PATCH] ide-cd: capability flag for MO drives From: Pascal Schmidt Add capability flag for MO drives. Since ATAPI MOs are now recognized by the ide-cd driver, it is useful to have a capability flag to be able to tell MO drives from other drives (needed for later write support). [PATCH] x86-64 merge for 2.5.73-bk9 - Update defconfig - Merge with i386 - Make oprofile compile again - Warning fixes - Minor changes for 32bit SYSCALL path (still disabled) - Remove obsolete FS/GS symbols in ptrace.h [netdrvr de2104x] quiet link timer (can be enabled by ethtool) [PATCH] USB: make struct usb_bus a struct class_device This creates sys/class/usb_host/ which lists all usb hosts in the system. [PATCH] USB: move ehci's sysfs files to the class device instead of the pci device. [PATCH] USB: move ohci's sysfs files to the class device instead of the pci device. [PATCH] USB: usb-skeleton.c usb_buffer_free() not called dev->udev was set to NULL to indicate a device disconnect but we need this value for usb_buffer_free() when device is still opened and cleanup is delayed until skel_release(). I've added a dev->present for preventing device read, write and ioctl. Also URB_NO_DMA_MAP is changed to URB_NO_TRANSFER_DMA_MAP and a unused var is removed. [PATCH] USB: Reconcile unusual_devs.h in 2.4 and 2.5 This patch for 2.5 copies into unusual_devs.h an entry that is present in 2.4 but missing here. It also adds back the IGNORE_SER flag with a dummy value, just for backward compatibility. [PATCH] s390: semtimedop. Fix SEMTIMEDOP operation in sys_ipc. Patch by Ernie Petrides. [PATCH] s390: ptrace. Fix ptrace system call number replacement code. [PATCH] s390: online attribute. Fix online attribute. "echo 1 > online" should enable a device and "echo 0 > online" should disable a device, not the other way round. [PATCH] s390: processor type. Add configuration option to select the different processor types. Add new path group algorith and new relocation types introduces with z990. [PATCH] s390: thin interrupts. Add thin interrupt support to qdio. [PATCH] s390: qeth network driver. The qeth network device driver. Supports the OSA Express Gigabit cards, HiperSockes and VM GuestLAN interfaces. [PATCH] Spelling fix Spelling fix [PATCH] Use KALLSYMS for scripts/kallsyms Since KALLSYMS is defined, might aswell use it somewhere. [PATCH] Fix IPC ABI for AMD64 AMD64 like IA64 needs to force IPC_64 in the IPC functions. This makes 2.5 compatible with 2.4 again. [PATCH] Disable a.out for AMD64 AMD64 does not support a.out, so don't display it in the configuration. [PATCH] Mark taskfile EXPERIMENTAL again Okay, since now corruption happens and there are some other issues to be resolved ("bad: scheduling while atomic" and "/proc/ide/hdX/identify") better set it by default to n for 2.5.74, also mark it EXPERIMENTAL. [PATCH] SH Merge This includes the remainder of arch-specific part of the SH merge. This patch only effects arch/sh and include/asm-sh, against current BK. [PATCH] dm: fix memory leak [PATCH] dm: remove bogus yields Replace a couple of bogus yields() with schedule() and io_schedule() respectively. Linux 2.5.74