Summary of changes from v2.6.6-rc2 to v2.6.6-rc3 ============================================ Can not mount from cifs vfs client built with gcc 3.3.1 due to compiler optimization of unsafe global variable. Remove unsafe global variable. Fix problem reconnecting additional mounts to the same server after session failure. Fix invalid dentry when race in mkdir between two clients fix oops in send_sig on unmount of cifs vfs due to sending signal to demultiplex thread after it has exited. Do not treat invalid handle warning in response to oplock break (of file that is now closed) as an error. Fix EIO caused by network timeouts on changing file size. To avoid spurious oplock breaks from server, in the case of inodes that we already have open, avoid doing path based setting of file size if we can do it by handle. This keeps our caching token (oplock) and avoids timeouts when the local oplock break takes longer to flush writebehind data than the SMB timeout for the SetPathInfo request would allow fix to not retime out the same session twice since it can invalidate the newly reestablished session unnecessarily Do not return buffer if request has already timed out. move bad smb session retry to correct location, up one level in cifs vfs code fix endian bug in lockingX and add retry on EAGAIN have to reconnect open files safely, one at a time, as needed finish off move from reopening all files on reconnection (which takes too long under heavy stress) to reopen file as needed after reconnection to server. correct retry on remaining handles based calls Fix compile error Missing soft vs. hard retry mount option Do not grab i_sem (already taken in filemap.c across commit write calls) during reopen of invalidated file handle. Fix oops in mount error path when unload_nls called with bad pointer. Avoid smb data corruption under heavy stress missing message on timed out requests rcvtimeout set improperly for some cifs servers invalidate locally cached pages when server breaks oplock. Do not loop reconnecting for servers that drop tcp session rather than sending smb negprot response Oops on reopen files when dentry already freed invalidate cached pages when last local instance closed so we do not use stale data while someone may be modifying the file on the server. fix double incrementing of transaction counter Fix check of filldir return code during readdir to avoid incomplete search results displayed on very large directories. Fix cleanup of proc entries. Add config parm to allow disabling negotiating Linux extensions allow disabling cifs Linux extensions via proc Fix an incorrect mapping of open flags to cifs open disposition. Fix blocking byte range locks. These fix breakages that were notice running lock tests 1 and 7 of the connectathon posix file api tests set byte range locktimeouts properly fix cifs readme gracefully exit on failed mounts to win98 (which closes tcp session rather than erroring on smb protocol negotiation) fix failed mounts to win98 part II Fix global kernel name space pollution Check return on failed dentry allocation. Suggested by Randy Dunlap Allow null password string pointer and passwords longer than 16 bytes finish handling commas in passwords finish off mount parm sep override Fix caching problem with multiply open files from different clients Relax requested CIFS permissions on open to simply request GENERIC_READ and GENERIC_WRITE (instead of GENERIC_ALL which can unnecessarily conflict with share permissions by asking implicitly for take ownership and other unneeded flags) fix remoting caching part 2 remove spurious debug messages fix problem not connecting to server when port not specified explicitly and port field unitialized reset searches properly when filldir fails ipv6 enablement for cifs vfs fixes Spurious white space and duplicated line cleanup NTFS: Set i_generation in VFS inode from seq_no in NTFS inode. NTFS: Make ntfs_lookup() NFS export safe, i.e. use d_splice_alias(), etc. NTFS: Make it compile... improve resume key resetting logic when filldir returns error and filename is in unicode allow nosuid mounts NTFS: Release 2.1.7 - Enable NFS exporting of mounted NTFS volumes. - Implement ntfs_get_parent() and ntfs_get_dentry() as the NTFS specific export operations ->get_parent() and ->get_dentry() respectively. NTFS: Add missing return -EOPNOTSUPP; in fs/ntfs/aops.c::ntfs_commit_nonresident_write(). fix problem with inode revalidation and cache page invalidation Fix the exec, suid, dev mount parms to not log warnings when specified fix caching data integrity problem use safer i_size_write mechanism to update i_size NTFS: Fix off by one error in ntfs_get_parent(). NTFS: Enforce no atime and no dir atime updates at mount/remount time as they are not implemented yet anyway. PPC32: Add CONFIG_MPC10X_BRIDGE. From: Adrian Cox . PPC32: In some cases we need to make all pages _PAGE_COHERENT. Controll this via the cpu_table flags since it's a requirement in some CPU and host bridge combinations. From: Adrian Cox fixes for fsx truncate/readahead/writebehind bug clean up compiler warnings PPC32: Remove an unneeded include in arch/ppc/boot/ PPC32: Add more useful information to the oops output. PPC32: Fix two typos in arch/ppc/boot/ From: Arthur Othieno Renaming pplus_common.c to hawk_common.c to match gt64260_common.c, etc, plus minor cleanups. Add missing description about how to specify credentials file Invalidate readahead data properly when file closed, but other client changed it on server Updating mcpn765 for 2.6 [SPARC32]: Fix wraparound bug in bitmap allocator Merge bk://linux.bkbits.net/linux-2.5 into linuxbox.(none):/src/linux/ppc/linux-2.5/linux Send NTCreateX with ATTR_POSIX if Linux/Unix extensions negotiated with server. This allows files that differ only in case and improves performance of file creation and file open to such servers Fix 20 second hang on some deletes of reopened file due to semaphore conflict with vfs_delete on i_sem PPC32: Change all #if FOO to #ifdef FOO PPC32: Two minor Carolina PReP fixes. From Hollis Blanchard . PPC32: Assign an interrupt for the VME chip on PReP MVME* boards. From Lutz Rossa . Updating Force PCore to 2.6 fix merge problem with 2.6.5 (rename of page struct field list to lru) ia64: Quiet another compiler-warning. ia64: Drop pci_sal_ext_{read,write}() and instead simply switch to extended config-space addresses when needed. This avoids the fragile SAL version testing. ia64: When delivering a signal, force byte-order to little-endian. [CPUFREQ] Correcting SGTC. Timer is based upon FSB From: Bruno Ducrot I think this patch is needed especially if the FSB is around 166MHz or 200MHz, or else I believe we get instabilities on some K7's motherboard powernow capable (it's called Cool'n Quiet IIRC). * Deduce fsb from cpu_khz and the max multiplier. It will be given as kHz now, so that frequency associated to a multiplier will be computate more accurately. Also, we need it for SGTC (see below). * Fix how cpuid is computed in powernow_decode_bios(). * Be more restrictive for PST. It may be possible (on desktop shipped with low power Athlon models) that FSB can be changed by dip switchs on motherboard for example. * Fix computation for SGTC. It use the bus timer (and then the bus frequency given by fsb). [CPUFREQ] don't use speedstep-centrino on unsupported CPUs From: Dominik Brodowski Don't use the ACPI data on CPUs we don't know nothing about. [CPUFREQ] powernow-k7 ACPI integration. More from Bruno Ducrot. Warning: it will only half work on the ACER Aspire, though: there is no pstate in the DSDT corresponding to the max frequency... I'm looking how to handle that correctly, probably at the init stage, if the max frequency is not given, then add it to powernow_table. * Integrate acpi perflib from Dominik. * Use acpi if the PST tables are known to be broken (ASUS Aspire match one PST, but give brain damaged values), or use ACPI if no PST found. [CPUFREQ] Drop unneeded part of last patch. acpi fallback is handled already in the init function. [CPUFREQ] powernow-k7 needs to init later. Due to the possible dependancy on ACPI. [CPUFREQ] Remove bogus newline in powernow-k7 driver. [CPUFREQ] Add a module parameter to force ACPI to be used. Fix misc. minor memory leaks in error paths [CPUFREQ] Make powernow-k7 acpi debug output a little less verbose. [CPUFREQ] powernow-k7 ACPI->PST values were a factor of 10 off. As much as I like the idea of a 13GHz laptop, setting it to 1.3GHz is probably for the best for the time being. [CPUFREQ] clear defaults before powernow-k7 acpi fallback Decoding the legacy tables may have set these values. free cifs read buffer on retry Fix major page leak in read code caused by extra page_cache_get call check permission locally for servers that do not support the CIFS Unix Extensions (allowing file_mode and dir_mode to augment the server permission check, by doing local vfs_permission check) [CPUFREQ] Not all powernow-K7 BIOS's put the frequency at MAX at POST. Remove 64 bit compiler warning [PATCH] ia64: fpswa_interface needs to be exported efivars can be built as a module, but it depends on 'fpswa_interface' which is not exported by fpswa.c. Patch below fixes this problem. [CPUFREQ] Fix debug build of powernow-k8 From Paul Devriendt [CPUFREQ] Fix up missing CONFIG_X86_POWERNOW_K8_ACPI We don't need this, we can infer from CONFIG_ACPI_PROCESSOR [CPUFREQ] Fix broken cast. This breaks on x86-64 with the following warning. drivers/cpufreq/cpufreq_userspace.c: In function `cpufreq_procctl': drivers/cpufreq/cpufreq_userspace.c:170: warning: cast from pointer to integer of different size drivers/cpufreq/cpufreq_userspace.c: In function `cpufreq_sysctl': drivers/cpufreq/cpufreq_userspace.c:208: warning: cast from pointer to integer of different size Remove "badness in remove_proc_entry" warning logged on module unload of cifs [SCTP] Avoid the use of constant SCTP_IP_OVERHEAD to determine the max data size in a SCTP packet. Calculate the overhead based on the socket's protocol family header length. [SCTP] Cleanup sctp_packet and sctp_outq infrastructure. [SCTP] Partial Reliability Extension support. [TG3]: Fix typo in TG3_TSO_FW_RODATA_ADDR definition. [SCTP] Propagate error from sctp_proc_init. (Olaf Kirch) Add in cifs fcntl handling to fix remote dnotify problem [ARM] Add find_first_bit and find_next_bit. [SPARC64]: Fix 32-bit posix timers. Do not cache inode metadata when cache time set to 0 (fix hardlink count caching) [SPARC64]: Missing part of posix timers fix. [PATCH] Add missing __initdata One of the stack size optimizations introduced a new static variable in a function marked with __init. But the variable is not marked appropriately and so 1k of data is never freed. [PATCH] Fix nfsroot option handling From: Trond Myklebust The following patch fixes up a number of bugs in the NFSroot parser rewrite from patchset trond.myklebust@fys.uio.no|ChangeSet|20040411182341|00938 It also ensures that NFSroot mount options are consistent with the userland "mount" program. [PATCH] compute_creds race From: Andy Lutomirski Fixes from me, Olaf Dietsche In fs/exec.c, compute_creds does: task_lock(current); if (bprm->e_uid != current->uid || bprm->e_gid != current->gid) { current->mm->dumpable = 0; if (must_not_trace_exec(current) || atomic_read(¤t->fs->count) > 1 || atomic_read(¤t->files->count) > 1 || atomic_read(¤t->sighand->count) > 1) { if(!capable(CAP_SETUID)) { bprm->e_uid = current->uid; bprm->e_gid = current->gid; } } } current->suid = current->euid = current->fsuid = bprm->e_uid; current->sgid = current->egid = current->fsgid = bprm->e_gid; task_unlock(current); security_bprm_compute_creds(bprm); I assume the task_lock is to prevent another process (on SMP or preempt) from ptracing the execing process between the check and the assignment. If that's the concern then the fact that the lock is dropped before the call to security_brpm_compute_creds means that, if security_bprm_compute_creds does anything interesting, there's a race. For my (nearly complete) caps patch, I obviously need to fix this. But I think it may be exploitable now. Suppose there are two processes, A (the malicious code) and B (which uses exec). B starts out unprivileged (A and B have, e.g., uid and euid = 500). 1. A ptraces B. 2. B calls exec on some setuid-root program. 3. in cap_bprm_set_security, B sets bprm->cap_permitted to the full set. 4. B gets to compute_creds in exec.c, calls task_lock, and does not change its uid. 5. B calls task_unlock. 6. A detaches from B (on preempt or SMP). 7. B gets to task_lock in cap_bprm_compute_creds, changes its capabilities, and returns from compute_creds into load_elf_binary. 8. load_elf_binary calls create_elf_tables (line 852 in 2.6.5-mm1), which calls cap_bprm_secureexec (through LSM), which returns false (!). 9. exec finishes. The setuid program is now running with uid=euid=500 but full permitted capabilities. There are two (or three) ways to effectively get local root now: 1. IIRC, linux 2.4 doesn't check capabilities in ptrace, so A could just ptrace B again. 2. LD_PRELOAD. 3. There are probably programs that will misbehave on their own under these circumstances. Is there some reason why this is not doable? The patch renames bprm_compute_creds to bprm_apply_creds and moves all uid logic into the hook, where the test and the resulting modification can both happen under task_lock(). This way, out-of-tree LSMs will fail to compile instead of malfunctioning. It should also make life easier for LSMs and will certainly make it easier for me to finish the cap patch. [PATCH] ext3 avoid writing kernel memory to disk From: Marc-Christian Petersen Solar Designer discovered an information leak in the ext3 code of Linux. In a worst case an attacker could read sensitive data such as cryptographic keys which would otherwise never hit disk media. Theodore Ts'o developed a correction for this. [PATCH] Call populate_rootfs later in boot populate_rootfs() is called rather early - before we've called init_idle(). But populate_rootfs() does file I/O, which involves calls to cond_resched(), and downing of semaphores, etc. If it scheules, the scheduler emits scheduling-while-atomic warnings and sometimes oopses. So run populate_rootfs() later, after the scheduler is all set up. [PATCH] remove amd7xx_tco From: Zwane Mwaikambo We've had trouble with this driver, it appears to work but the hardware never does the final reboot. I have yet to come across someone with a board which works and don't have personal access to one. So how about scrapping the whole thing. [PATCH] lockfs - vfs bits From: Christoph Hellwig These are the generic lockfs bits. Basically it takes the XFS freezing statemachine into the VFS. It's all behind the kernel-doc documented freeze_bdev and thaw_bdev interfaces. Based on an older patch from Chris Mason. [PATCH] lockfs: reiserfs fix From: Chris Mason reiserfs_write_super_lockfs() is supposed to wait for the transaction to commit. [PATCH] lockfs - xfs bits From: Christoph Hellwig Remove all the code now in the VFS, make XFS's freeze ioctls use the new infastructure and reorganize some code. This code needs some work so the source files shared with 2.4 aren't exposed to the new VFS interfaces directly. You'll get an update once this has been discussed with the other XFS developers and is implemented. Note that the current patch works fine and I wouldn't complain if it gets into Linus' tree as-is. [PATCH] lockfs - dm bits From: Christoph Hellwig This patch makes the device mapper use the new freeze_bdev/thaw_bdev interface. Extracted from Chris Mason's patch. [PATCH] i4l: add compat ioctl's for CAPI From: Marcel Holtmann This patch adds the needed compat ioctl's for the CAPI on 64bit platforms. [PATCH] selinux: change context_to_sid handling for no-policy case From: Stephen Smalley This patch changes the behavior of security_context_to_sid in the no-policy case so that it simply accepts all contexts and maps them to the kernel SID rather than rejecting anything other than an initial SID. The change avoids error conditions when using SELinux in permissive/no-policy mode, so that any file contexts left on disk from prior use of SELinux with a policy will not cause an error when they are looked up and userspace attempts to set contexts can succeed. [PATCH] selinux: add runtime disable From: Stephen Smalley This patch adds a kernel configuration option that enables writing to a new selinuxfs node 'disable' that allows SELinux to be disabled at runtime prior to initial policy load. SELinux will then remain disabled until next boot. This option is similar to the selinux=0 boot parameter, but is to support runtime disabling of SELinux, e.g. from /sbin/init, for portability across platforms where boot parameters are difficult to employ (based on feedback by Jeremy Katz). [PATCH] selinux: remove hardcoded policy assumption from get_user_sids() logic From: Stephen Smalley This patch removes a hardcoded policy assumption from the get_user_sids logic in the SELinux module that was preventing it from returning contexts that had the same type as the caller even if the policy allowed such a transition. The assumption is not valid for all policies, and can be handled via policy configuration and userspace rather than hardcoding it in the module logic. [PATCH] i810_dma range check From: Andrea Arcangeli Correctly range-check an incoming-from-userspace argument. Found by the Stanford checker. [PATCH] i386 hugetlb tlb correction From: William Lee Irwin III i386 does hardware interpretation of pagetables, so pte_clear() can't be used on present ptes, as it sets the upper half of the hugepte prior to setting the lower half (which includes the valid bit). i.e. there is a window where having a hugepage mapped at 56GB and doing pte_clear() in unmap_hugepage_range() allows other threads of the process to see a hugepage at 0 in place of the original hugepage at 56GB. This patch corrects the situation by using ptep_get_and_clear(), which clears the lower word of the pte prior to clearing the upper word. There is another nasty where huge_page_release() needs to wait for TLB flushes before returning the hugepages to the free pool, analogous to the issue tlb_remove_page() and tlb_flush_mm() repair. [PATCH] loop_set_fd() sendfile check fix From: Yury Umanets I have found small inconsistency in loop_set_fd(). It checks if ->sendfile() is implemented for passed block device file. But in fact, loop back device driver never calls it. It uses ->sendfile() from backing store file. [CPUFREQ] Fix unbalanced try_get_module/put_module Spotted by Charles Coffing [CPUFREQ] Remove redundant part of powernow-k7 module parm If used as a bootparam, this would've become powernow-k7.powernow_acpi_force which looks silly. [CPUFREQ] Make an educated guess at the current P-state in the ACPI driver. One big limitation of the ACPI specification is that it's impossible to detect the current P-State by reading from ACPI-defined registers. And the CPU isn't always at P0 when the system boots. So, try to "guess" the current P-State by analyzing cpu_khz. From Dominik. [CPUFREQ] Export an array of acpi driver supported frequencies in sysfs From Dominik. [CPUFREQ] Fix security hole in proc handler. Brad Spengler found an exploitable bug in the proc handler of cpufreq, where a user-supplied unsigned int is cast to a signed int and then passed on to copy_[to|from]_user() allowing arbitary amounts of memory to be written (root only thankfully), or read (as any user). The Common Vulnerabilities and Exposures project (cve.mitre.org) has assigned the name CAN-2004-0228 to this issue. Revert fb_ioctl "fix" with extreme prejudice. As Arjan points out, the patch does exactly the opposite of what it was claimed to do. Andrea: tssk tssk. Cset exclude: akpm@osdl.org[torvalds]|ChangeSet|20040421144431|15930 [PATCH] ia64: add mq support for ia64 [ARM] Add support for ARM Versatile platform. This cset adds minimal support for ARM Ltd's ARM926EJ-S "Versatile" platform. [SERIAL] Correct PL011 help text. [PATCH] ide-disk.c: fix for IDE CF card ejection with devfs From: Pavel Roskin If I eject IDE CompactFlash card, I get a stack dump from devfs_remove() because ide/host2/bus0/target0/lun0 doesn't exist. After del_gendisk() is called from idedisk_cleanup() drive->devfs_name refers to a non-existent directory and should be erased, so that ide_unregister() doesn't try to remove that directory again. [PATCH] ide-probe.c: kill duplicate #include From: Arthur Othieno [PATCH] removal of MOD_{INC,DEC}_USE_COUNT in ide-cs.c From: Pavel Roskin The "ide-cs" module cannot be unloaded because it uses obsolete MOD_INC_USE_COUNT and MOD_DEC_USE_COUNT macros. In fact, they are not needed in ide-cs.c in 2.6 kernels. The generic PCMCIA code already increases use count for every device served by the driver, so it's impossible to unload the ide-cs driver while it's in use. I was told that the removal of IDE interfaces may be unsafe in 2.6 kernels. However, MOD_INC_USE_COUNT only prevents removal of the module, not the interface. It's also the first obstacle, albeit a trivial one, for anybody debugging those problems (i.e. loading a modified module requires "rmmod -f" or reboot to unload the old version). [PATCH] PPC64 iSeries virtual ethernet fix This patch is needed due to other patches that were applied in parallel with the inclusion of the iSeries virtual ethernet driver. [PATCH] New set of input patches This updates the vsxxx driver to it's current version. Even DEC tablet support (VSXXX-AB) is now tested - it works:) You can even hotplug between mouse and digitizer... [PATCH] lkkbd: Current version This updates the lkkbd driver to it's current version. It also incorporates two patches suggested on LKML (fixing some leading whitespace and an unneccessary check). [IPV4]: Fix return value on MCAST_MSFILTER error case. [NETFILTER]: Missing ip_rt_put in ipt_MASQUERADE. [TCP]: Better packing of frto fields into tcp_opt. [TCP]: Add sysctl to turn off matrics caching. [PATCH] atmel wireless update Fixes to avoid problems when the driver shares an interrupt. These were caused because the ISR cannot determine the status of the chip without changing its state and the driver was only disabling the interrupts from the card during some critical regions. An interrupt originating from another device at the wrong moment could run the ISR and corrupt chip state. Fixed by blocking interupts in the processor in critical regions and never checking the chip interrupt status before the driver is fully up. Added the ability to override the Regulatory Domain stored in ROM. The following command achieves this iwpriv regdomain vaild domains are USA,Canada,Europe,Spain,France,MKK,MKK1,Israel The current regulatory domain affects which radio channels are available for use. Some minor tidying to calibrate busy-wait loops using udelay and remove inclusion of obsolete include/linux/802_11.h [PATCH] pcmcia netdev ordering fixes From: Russell King This is an *untested* patch (PCMCIA people, please test) to fix initialisation ordering issues in these network device drivers. I don't have the hardware to be able to test these changes, although it passes a modular build without warnings. Andrew - it may be worth sticking these in -mm so they get some visibility via your tree as well. Many of these drivers were calling register_netdev() before they had completed their initialisation. In addition, they were calling register_netdev with the "DEV_CONFIG_PENDING" flag still set, which prevents hotplug scripts from bringing up the interface. Also, we take care to ensure that link->dev is correctly set - this is used to tell the PCMCIA release code if the netdev is currently registered (yes, that's probably racy at present, but lets sort one problem at a time.) I've arranged that all drivers register the netdevice as close as possible to the end of their initialisation, copy the net device's name for cardmgr to pass to it's network scripts, and then print out whatever information the driver wants to. Finally, a note about ibmtr - it seemed to assume that cardmgr wanted "tr0" or "tr1" depending on the base address, and completely ignoring the real device which the netdev layer allocated it. I've assumed that this is wrong (since the netdev name is used in printk messages), and changed it to behave the same as the other drivers. [PATCH] Au1000 IrDA driver update Get to build under 2.6, random updates to debugging code, new eval boards. [PATCH] Remove RCS Id string [PATCH] meth updates More work on the meth driver for SGI IP32 aka O2. [PATCH] BCM1250 network driver updates Resurrect into working order for 2.6. [PATCH] sgiseeq fixes Resurrect into working order for 2.6. [PATCH] IOC3 updates Use generic MII code, compile time configurable rx and tx checksumming. Try to handle IOC3 attached 16552 serials. Fix descriptor setup when crossing a 16kB page. Use generic PCI API. Use generic ethtool code, small cleanups, use a temporary kludge to exploit virtual device functionality for significantly better performance. [PATCH] declance updates Various small changes, DEC still doesn't work under 2.6. Retry 2nd time after failure on correct port [TCP]: Abstract out all settings of tcp_opt->ca_state into a function. [TCP]: Add vegas congestion avoidance support. A forward port of an old 2.3.x kernel hack done years ago. I (DaveM) did the first rough port, Stephen Hemminger actually cleaned it up and made it usable. [PATCH] USB: Don't try to suspend devices that do not support it. Patch originally from luming.yu@intel.com and closes bug #1557 [PATCH] sysfs module unload race fix for bin_attributes - Add module locking to sysfs bin_attribute files. Update all in-tree users to set module owner. Compile tested. booted. stress tests pass: while true; do modprobe mymod; rmmod mymod; done & while true; do hexdump -C /sys/path/to/sysfs/binary/file; done [PATCH] USB: Important bugfix for UHCI list management code A major bug in the UHCI driver turned up recently. Thanks to a lot of help from Simone Gotti it was identified and fixed late last week. It turned out to be entirely my fault -- a previous patch had introduced two (!) errors. (A combination of carelessness and a nasty thinko, and somehow it passed the regression tests...) Anyway, it's entirely possible that many of the problems people have been seeing are caused by that bug. This patch is the solution. [PATCH] USB: fix cdc-acm as it is still (differently) broken [PATCH] USB: fix cdc-acm warnings due to previous patch [PATCH] remove show_trace_task() It no longer has any callers. [PATCH] ppc64: iSeries virtual cdrom module fix From: Stephen Rothwell This patch fixes loading viocd as a module. It would oops because I was passing the address of a static buffer to dma_map_single and when loaded as a module, this address is not valid for that purpose. There are a couple of simple cleanups here as well. [PATCH] s390: core s390. From: Martin Schwidefsky s390 core changes: - Fix race in do_call_softirq in regard to kernel preemption. - Fix typo in compat mq system call wrappers. - Add s390 to Kconfig for AUDITSYSCALL. - Redefine TASK_SIZE to TASK31_SIZE for compilation of binfmt_elf32. - Use correct error value for sys32_ipc when called with an invalid number. - New default configuration. [PATCH] s390: common i/o layer. From: Martin Schwidefsky Common i/o layer changes: - Quiesce active subchannels for lpar reipl. - Delete timer after reception of interrupt for kill on timeout. - Cleanup some comments in qdio. [PATCH] s390: 3270 device driver. From: Martin Schwidefsky 3270 device driver changes: - Add NULL pointer checks. [PATCH] s390: network device drivers. From: Martin Schwidefsky Network driver changes: - qeth: Fix reference counting in regard to sysfs backing store patches. - qeth: Prefix kernel thread names with qeth_. - qeth: Remove inbound and outbound tasklets. Handle buffers directly in the interrupts handlers. - iucv: Add missing kfree in iucv_register_program. - iucv: Add missing return in netiucv_transmit_skb. - iucv: Check for NULL pointer in conn_action_txdone. [PATCH] s390: dasd device driver. From: Martin Schwidefsky dasd device driver changes: - Initialize open_count with -1 to account for blkdev_open in dasd_scan_partitions. - Introduce USE_ERP request flag to selectivly switch off error recovery for reserve, release & unconditional reserve ioctls. [PATCH] s390: zfcp adapter fixes. From: Martin Schwidefsky zfcp host adapter changes: - Fix error recovery stall in case of unavailable nameserver. - Reset host_scribble field to NULL in scsi_cmd. - Remove request debug code. [PATCH] s390: crypto api. From: Martin Schwidefsky Add support for z990 crypto instructions to in-kernel crypto api. [PATCH] EDD: set sysfs attr owner field From: Matt Domsch The patch below from Michael E. Brown properly sets the owner field of a sysfs attribute. Without this patch, it is possible to crash the kernel with a simultaneous insmod/rmmod while reading files exported by the module. [PATCH] dynamic proc cleanups From: Matt Mackall Delete obsolete comment and kill test of obsolete define. [PATCH] fbdev comment fix Fix the incorrect comment which caused the fb_ioctl confusion. [PATCH] MIPS: PCI code is now shared. From: Ralf Baechle All MIPS systems use the same PCI code now. [PATCH] Add Pete Popov to credits From: Ralf Baechle Add Pete to CREDITS for all the time he's invested into supporting the AMD Alchemy of SOCs and eval boards. [PATCH] Merge missing MIPS i8042 bits From: Ralf Baechle - Add HPC3 PS/2 driver bits for SGI IP22 aka Indy - Add Mace PS/2 driver bits for SGI IP32 aka O2 - Add R4030 PS/2 driver bits for Jazz family - Don't register I/O ports where we're using the I/O port memory window to access the i8042 registers [PATCH] MIPS is an a.out free zone From: Ralf Baechle MIPS never uses a.out [PATCH] Update comment in fs/compat.c From: Ralf Baechle MIPS now also uses the generic ioctl compat code. [PATCH] Set ARCH_MIN_TASKALIGN on ppc32 From: Benjamin Herrenschmidt From: David Woodhouse Without this the task struct gets unaligned when using SLAB_DEBUG, causing random problems with FP and Altivec. [PATCH] sunrpc rmmod oops fix From: "J. Bruce Fields" Unregister svcauth_gss caches on exit from gss module; fixes an oops on rmmod. [PATCH] ppc64: add some iSeries proc entries From: Stephen Rothwell This patch just adds some proc entries for the virtual tape and cdrom drivers to allow mapping between linux devices and OS/400 ones. This is expected by existing users and there is no other way to do this translation. [PATCH] ppc64: Set ARCH_MIN_TASKALIGN From: Benjamin Herrenschmidt We need some alignement of those structs for proper operations especially with FP and Altivec, or SLAB_DEBUG can break us. [PATCH] clean up Pentium M quirk code in nmi.c This simplifies the Pentium M quirk code in nmi.c, and eliminates an unnecessary apic_read(). Local APIC accesses are not zero-cycle; let's not inflict more damage than we must. [PATCH] use smp_processor_id() in init_IRQ() This replaces current_thread_info()->cpu in i386' init_IRQ() by the equivalent smp_processor_id(). Reduces overhead on UP, and makes the code cleaner. [PATCH] USB: ehci handles pci misbehavior better Cope better when PCI misbehaves badly and registers misbehave: - terminate some loops before they get to infinity * capability scan * port reset - after init failure, memory may already be cleaned up Some systems have been reporting such problems after ACPI resume. [PATCH] USB: rndis gadget driver updates Various build fixes: 64bit (Andrew Morton), static linking, broken on big-endian, etc. Tighten up the integration with the main "ether" driver, so state transitions and host ethernet addresses are shared too. Add missing spinlock calls around RNDIS command outcall, fix GET_INTERFACE issue, host mustn't clobber netdev flags. Minor code cleanups. [PATCH] USB: silence dpcm warning Warnings aren't terribly important in and of themselves, but there isn't really much the warning tells us to do here, so it would appear that caving in to the compiler is the thing to do for now. [PATCH] USB: tiglusb: wrong timeout value Hi, this patch (cumulative; 2.4 & 2.6) fixes another bug in the tiglusb driver. The formula used to calculate jiffies from timeout is wrong. The new formula is ok and takes care of integer computation/rounding. This is the same kind of bug than in the tipar char driver. [PATCH] USB: ftdi patch fixup I just checked out the latest 2.6.6-rc1-mm1 to see that the ID patch for ftdi has applied cleanly, but apparently someone was faster and in the combined ID table our ID's were missing. Most probably the patch program got confused and applied the hung at wrong spot.. Attached is a patch that places that 2 lines in the correct spot. [PATCH] USB: usbnet and pl2301/2302 reset Handle some PL-2301/2302 devices better. RFC1002 fixup [PATCH] tipar char driver: wrong timeout value this patch (2.4 & 2.6) fixes a bug about the timeout value. The formula used to calculate jiffies from timeout is wrong. The new formula is ok and takes care of integer computation/rounding. There is the same bug in the tiglusb.c module which will be fixed by another patch. [PATCH] symlink doesn't support kobj name > 20 charaters (KOBJ_NAME_LEN) Since symlink.c uses "name" field of a kobj when it calculates the length, it gets a wrong value if the kobj's name has more than 20 charathers. A correct way to do that is to call kobject_name(kobj) instead of using kobj->name directly. exit from waiting on smb response when session dead Update change log for 1.10 cifs vfs ia64: Add message-queue support to copy_siginfo_from_user(). proper rc on host down fix error code mapping on bad host fix timeout on close operation when pending signal NTFS: Move a few assignments after a NULL check in fs/ntfs/attrib.c. do not allow routine user signals to kill SendReceive wait for response (which was damaging performance badly) [PATCH] writeback livelock fix If a filesystem's ->writepage implementation repeatedly refuses to write the page (it keeps on redirtying it instead) (reiserfs seems to do this) then the writeback logic can get stuck repeately trying to write the same page. Fix that up by correctly setting wbc->pages_skipped, to tell the writeback logic that things aren't working out. NTFS: Finally fix NFS exporting of mounted NTFS volumes by checking the return of d_splice_alias() and acting accordingly rather than just ignoring the returned dentry. [SPARC64]: Update defconfig. [ARM PATCH] 1824/1: guard against gcc not respecting local variable register assignment Patch from Nicolas Pitre The recent discovery of a problem with gcc not always respecting register assignment for local variables to be used with inline assembly is worrisome. If unnoticed, such problems can cover bugs that might prove hard to find especially when the code is right but silently ignored. This patch adds a test where needed to have the assembler confirm our register selection and fail the kernel build if the wrong registers are allocated. [ARM PATCH] 1825/1: abort on bad code generation with div64 in some cases Patch from Nicolas Pitre One case was found when the kernel is compiled with -O2 (instead of the default -Os). However, all gcc-3.* versions seem to have a problem with this. Instead of blacklisting them all unconditionally, better just abort when a bad case occurs especially since this bad case doesn't seem to appear with current default compiler flags. Depends on patch #1824. [SPARC64]: Fix zero-extension issues wrt. {pgd,pmd}_val(). [TCP]: Report vegas info via tcp_diag. [CRYPTO]: Add module autoloads for null module. [NET]: linux/if.h needs linux/compiler.h for __user. [ARM] Remove extraneous "volatile" from atomic_t pointers. [ARM PATCH] 1815/1: Generic DMA buffer bouncing support for ARM targets Patch from Deepak Saxena Latest (and hopefully last :) patch for generic DMA buffer bouncing. - Fixed SA1111 dma_needs_bounce - Added check for out of bounds buffers. - Made dmabounce.c directly implement dma_* API Note that I didn't do the following: #ifndef CONFIG_DMABOUNCE inline implementations of all functions #else extern declarations #endif Instead I wrapped it individual in #ifndef/#else blocks to keep the comments in the same area as function declarations. IMHO this makes it easier for someone to go look at the comments if they need to know what a specific API does. [PATCH] e100: ICH 10/H Tx hang fix * Need to carry forward this workaround from old e100 driver to avoid a Tx hang on ICH systems linked at 10/Half. Workaround adds a stall before each Tx command queued by issuing a NOP command followed by 1us delay. Yuck! Otherwise HW locks hard. (Probably needed for eepro100 also ;-). [netdrvr tulip] fix use-after-free The attached patch fixes an "oops" in tulip driver when a live interface is deactivated (i.e. PC Card ejected or module unloaded) without being brought down first. [PATCH] generic PCI IDE support for Toshiba Piccolo chips From: Daniel Drake Adds Toshiba chips to the list supported by the generic PCI IDE driver. The 2.4 tree already contained an entry for 0x1179:0x0102, this patch adds that entry to 2.6 as well as two new ones (0x0103 and 0x0105). This then allows DMA to be enabled on disks. [PATCH] use kernel min/max in IDE code (1/2) From: Randy Dunlap From: Michael Veeck ide-cd.h: remove unnecessary MIN() macro ide-cd.c: change MIN() calls to use kernel.h calls ide-tape.c: use min_t()/max_t() instead of min()/max() [PATCH] use kernel min/max in IDE code (2/2) From: Randy Dunlap From: Michael Veeck Removes unnecessary IDE_MIN()/IDE_MAX() macros and changes calls to use kernel.h macros instead. [PATCH] prevent module unloading for legacy IDE chipset drivers It is unsafe thing to do (no locking, no reference counting etc). Just remove module_exit() as it was done for IDE PCI drivers. [ATM]: [fore200e] 0.3e version by Christophe Lizzi (lizzi@cnam.fr) [ATM]: [fore200e] make use tasklet configurable [PATCH] V4L: Update the saa7146 driver - [DVB] saa7146 driver updates: - remove bogus v_calc and h_calc parameters, which can be easily retrieved from other values - add class parameter to i2c initialization - let resource handling provide more useful informations - sanitize overlay/capture locking [PATCH] DVB: Documentation and Kconfig updazes - [DVB] misc. documentation updates, KConfig help file updates - [DVB] make Twinhan driver depend on bt8xx [PATCH] DVB: Update DVB budget drivers - [DVB] budget-av: patch by Kenneth Aafløy to add support for Typhoon DVB-S budget card - [DVB] budget.c: support for Fujitsu-Siemens Activy Card - [DVB] budget-ci: add preliminary CI support [PATCH] DVB: Add EN50221 cam support to dvb-core - [DVB] add generic functions for EN50221 CAM interfaces [PATCH] DVB: Other DVB core updates - [DVB] remove superflous memset() which caused section data to be overwritten when a) there are two sections in one TS packet, and b) the first section was smaller than 18 bytes; thanks to Jean-Claude Repetto for tracking this down - [DVB] starting a ts filter on a running section filter's pid did break the section filter; fixed. - [DVB] integrate ULE Decapsulation code, thanks to gcs - Global Communication & Services GmbH. and Institute for Computer Sciences Salzburg University. Hilmar Linder and Wolfram Stering - [DVB] fix the module use count bugs, thanks to Hernan A.Perez Masci for his initial work on this problem - [DVB] if dvb_frontend_internal_ioctl() returns an error code, be sure to deliver it to the calling application, don't ignore it (fixes the bug that the frontend0 doesn't respond properly to unknown ioctls...) - [DVB] major frontend code clean up, rewritten core tuning loop. Thanks to Andrew de Quincey. - [DVB] follow changes in dvb-core in skystar2, dvb-bt8xx [PATCH] DVB: AV7110 DVB driver updates - [DVB] av7110 update: - speed up firmware loading - follow internal API changes in saa7146 driver - introduced some symbolic constants for a/v dec cmds - change default for hw_sections to 0 to enable crc checks [PATCH] DVB: Misc. DVB frontend driver updates - [DVB] follow changes in dvb-core for frontend drivers (ves1x93, ves1820, nxt6000, sp887x, tda1004x, stv0299, mt312, alps_tdlb7, alps_tdmb7, at76c651, cx24110, dst, dvb_dummy_fe, grundig_29504-401, grundig_29504-491) - [DVB] tda1004x: updated timeout to 800ms, implemented FE_SLEEP - [DVB] cx24110: add FE_CAN_RECOVER to reduce kdvb-fe CPU load - [DVB] grundig_29504-401: added 200ms delay after first FE_INIT, Implemented FE_GET_FRONTEND - [DVB] alps_tdlb7, alps_tdmb7: upped tuning delays to fix tuning [PATCH] DVB: Misc. DVB USB driver updates - [DVB] ttusb-dec: - Add a parameter to dvb_filter_pes2ts function to specify whether the packet is a payload unit start or not - Use the hotplug firmware loader for 2.6 kernels instead of compiling the firmware into the module. - Correct the USB id of the DEC3000-s, add basic support - [DVB] ttusb-budget: - Remove spurious discontinuity message when starting streaming [PATCH] DVB: Follow saa7146 changes in affected V4L drivers - [V4L] follow changes in saa7146 driver: mxb, dpc7146, hexium_orion, hexium_gemini [PATCH] ISDN CAPI: add ncci list semaphore Fix race conditions of ISDN CAPI's internal ncci list handling by using a per capidev semaphore. [PATCH] don't log drive loading failures This is an old bug that often confuses people, rightfully. For some laptop and server type cdroms, the drives advertise themselves as tray loading even if they are caddy-like or slot-in loaders. This means they cannot insert the media on their own, and this generates a 5/24/00 error to START_STOP_UNIT. This prevents the logging of such a failure, and also fixes the generation of such a bad command in case the mech type was correctly set by the drive. [PATCH] USB: fix up fake usb_interface structure in hiddev This fixes a oops in the current kernel tree. [PATCH] USB: further cleanup of the hiddev driver, fixing another possible oops on disconnect. [PATCH] correct LoEj logic The logic got a little foo-bar'ed in the last patch, we should have keep the old logic. Ie LoEj bit (2nd) must always be set, just clear it for known non-capable of loading drives. [PATCH] ia64: bug w/ shared interrupts I just ran into a bug introduced by the most recent iosapic.c patch. The scenario is a builtin driver is up and running happily. A module loads for a devices that happens to share the same interrupt vector, in this case a network driver. The module calls pci_enable_device() as it should, which eventually lands in iosapic_enable_intr(). We then proceed to mask the interrupt and kill the device that's already running. As a bonus, request_interrupt() doesn't fix the problem because we only call the startup for the interrupt handler on the first action attached to the interrupt. I think the best way out of this is simply to detect when an action is already attached to a vector and leave it alone. This also prevents interrupts from moving to other cpus (on boxes w/o irq redirection) for no good reason. [PATCH] ia64: Remove SN PDA page overflow check There's an obsolete check in sn_init_pdas() left over from the 2.4 days that panics if the cpu PDA and cpu_data area span more than a page. With 2.6, we're not limited to a page, so remove the check entirely. [TCP]: Add vegas sysctl docs. [IPV4]: Spelling fixed for ip-sysctl.txt [IRDA]: Export irda_task_delete. [SUNRPC]: Missing NULL kmalloc check in unix_domain_find(). [PATCH] One more USB fix Here's another fix it'd be good to merge ... No changes to the Linux code at all, but it'll makes the Windows configuration of a Linux device running the Ethernet/RNDIS gadget behave correctly on more versions of Windows. (It might prevent some Bluescreening too.) [TG3]: Fix jimbo frame PHY programming. 1. Added new workaround for 5705 to improve bit error rate. 2. Changed to use read-modify-write to set the extended packet bit in tg3_phy_reset. 3. Removed some phy setup code in tg3_phy_probe. All of that is already in the end of tg3_phy_reset. Writing to the phy during probe time may contend with ASF because ASF is constantly reading phy registers to determine the link. 4. Removed the write to phy register 0x18 in tg3_phy_copper_begin. Not sure what it was for and it will overwrite that regsiter. [TG3]: Fix comment typo. [TG3]: Dump NIC-specific statistics via ethtool. Contributed by Tony Cureington @ HP, updated (from pre-ethtool_ops) and modified by me. [TG3]: Update driver version and reldate. [PATCH] create singlethread_workqueue() From: Rusty Russell Workqueues are a great primitive for running things from user context from a completely clean environment. Unfortunately, they currently insist on creating one thread per CPU, which is overkill for many situations, so the more generic keventd workqueue is used for these. Recently deadlocks using keventd were demonstrated, showing that it is not suitable for all uses. 1) Clean up CPU iterators. Always a nice touch. 2) Add __create_workqueue() and create_singlethread_workqueue(), keeping source compatibility. 3) Put workqueues in workqueue list even if !CONFIG_HOTPLUG_CPU (means we need a lock to protect that list). Now we can tell if a wq is single-threaded using list_empty(&wq->list). 4) For single-threaded workqueues, override CPU in queue_work, delayed_work_timer_fn and flush_workqueue to be 0. flush_workqueue now does redundant passes for single-threaded workqueues, but the code remains simple. 5) Make create_workqueue_thread return the thread, so we can easily kthread_bind for multi-threaded workqueues. akpm fixes: - Fix up is_single_threaded() handling - single-threaded wq thread does not have "/0" appended. [PATCH] Use workqueue for call_usermodehelper From: Rusty Russell call_usermodehelper uses keventd to create a thread, guaranteeing a nice, clean kernel thread. Unfortunately, there is a case where call_usermodehelper is called with &bus->subsys.rwsem held (via bus_add_driver()), but keventd could be running bus_add_device(), which is blocked on the same lock. The result is deadlock, and it comes from using keventd for both. In this case, it can be fixed by using a completely independent thread for call_usermodehelper, or an independent workqueue. Workqueues have the infrastructure we need, so we use one. Move EXPORT_SYMBOL while we're there, too. akpm fixes: Make it compile with !CONFIG_KMOD [PATCH] reiserfs: ignore prepared and locked buffers From: Chris Mason block_write_full_page might see and lock clean metadata buffers, which leads to bogus vs-12339 messages. Change the message to ignore bh locked. [PATCH] credentials locking fix From: Chris Wright Contributions from: Stephen Smalley Andy Lutomirski During exec the LSM bprm_apply_creds() hooks may tranisition the program to a new security context (like setuid binaries). The security context of the new task is dependent on state such as if the task is being ptraced. ptrace_detach() doesn't take the task_lock() when clearing task->ptrace. So there is a race possible where a process starts off being ptraced, the malicious ptracer detaches and if any checks agains task->ptrace are done more than once, the results are indeterminate. This patch ensures task_lock() is held while bprm_apply_creds() hooks are called, keeping it safe against ptrace_attach() races. Additionally, tests against task->ptrace (and ->fs->count, ->files->count and ->sighand->count all of which signify potential unsafe resource sharing during a security context transition) are done only once the results are passed down to hooks, making it safe against ptrace_detach() races. Additionally: - s/must_must_not_trace_exec/unsafe_exec/ - move unsafe_exec() call above security_bprm_apply_creds() call rather than in call for readability. - fix dummy hook to honor the case where root is ptracing - couple minor formatting/spelling fixes [PATCH] ext3 journalled quota locking fix From: Jan Kara I've attached a fix for a problem in ext3 journalled quota patch - the problem is that detecting whether dqput() sleeps was wrong and so we could possibly schedule when holding a spinlock. [PATCH] Bigger quota hashtable From: Jan Kara I found out that quota uses hash table with just 43 entries to hash dquot entries. I guess that we can afford using one page for that (quotactl(Q_GETQUOTA...), got faster like 3x for 4000 users). Attached patch implements that. [PATCH] Per-sb dquot dirty lists From: Jan Kara When there are lots of dirty dquots the vfs_quota_sync() is too slow (it has O(N^2) behaviour). Attached patch implements list of dirty dquots for each superblock and quota type. Using this lists sync is trivially linear. Attached patch is against 2.6.5 with journalled quota and previous patch for hash table size. (Jan had a test which went from 8 minutes to 0.8 seconds...) [PATCH] dquot: remove unneeded test We're testing the nullness of `sb' potentially after rereferencing it (although the compiler will have reordered things to avoid such a bug). Just remove the test - the superblock pointer shouldn't be null in there. [PATCH] Minor fixes for ext3 journalled quotas From: Jan Kara It fixes a memory leak when turning journalled quotas off. [PATCH] ppc64: Split prom.c Into pre-reloc and post-reloc Functions From: Rusty Russell Move all the functions that run normally (ie. with the kernel in the correct address) to the bottom of prom.c, and keep all the functions which use RELOC() to the top, above "prom_init()" which calls them. Also, make prom_initialize_dart_table and prom_initialize_tce_table static. [PATCH] ppc64: Rearrage finish_device_tree() and its functions in C Order From: Rusty Russell Rather than pre-declaring static functions, order them so it's not neccessary. This makes the code easier to navigate and patch. [PATCH] ppc64: Rearrage copy_device_tree() and its functions in C Order From: Rusty Russell Rather than pre-declaring static functions, order them so it's not neccessary. This makes the code easier to navigate and patch. [PATCH] ppc64: Rearrage interpret_funcs in C Order From: Rusty Russell Rather than pre-declaring static functions, order them so it's not neccessary. This makes the code easier to navigate and patch. [PATCH] ppc64: Rearrage Rest of prom.c in C Order From: Rusty Russell Rather than pre-declaring static functions, order them so it's not neccessary. This makes the code easier to navigate and patch. [PATCH] ppc64: Make finish_device_tree use lmb_alloc, not klimit From: Rusty Russell finish_device_tree simply allocates nodes by incrementing klimit (ie. using memory on top of the kernel). Change it to figure out how much memory it needs, then use lmb_alloc to allocate that, then fill it in. This gets rid of the only manipulation of klimit after prom_init. [PATCH] ppc64: make_room macro for ppc64 prom.c From: Rusty Russell copy_device_tree calls inspect_node() which recurses to copy the Open Firmware device tree. It just copied into memory above the kernel for about 8MB, which can overwrite the initrd, and/or get truncated on big machines. The real solution is to allocate memory properly for all the prom_init memory users. However, that's quite a change. This patch introduces a "make_room" macro which inspect_node uses to get memory. This does the same as before, except skips over the initrd if neccessary. [PATCH] ppc64: Fix prom.c to boot on G5 after make_room fix From: Rusty Russell Ben Herrenschmidt made this actually boot. [PATCH] ppc64: Clean up prom functions in prom.c From: Rusty Russell 1) Define PROM_ERROR: checks for call_prom() < 0 are bogus (returns ulong) 2) All OF access functions should be marked __init. 3) prom_strtoul isn't used at all. [PATCH] ppc64: Initrd Cleanup From: Rusty Russell Clean up initrd handling. 1) Expose initrd_start and initrd_end to prom.c (replacing its local initrd_start and initrd_len). 2) Don't hand mem (aka klimit) through functions which don't need it. 3) Add more debugging under DEBUG_PROM in case we broke anything. [PATCH] ppc64: Move Initrd From: Rusty Russell This patch moves the initrd when "make_room" runs out of space, rather than stepping over it. The problem with stepping over it is that it gets copied with the kernel when this happens, which wastes space. Also ensures that the initrd isn't where the kernel wants to be moved to: if it is, it gets moved out the way. [PATCH] ppc64: prom.c fix for CONFIG_BLK_DEV_INITRD=n From: Rusty Russell Doesn't compile with CONFIG_BLK_DEV_INITRD=n. Be more careful with the conditionals. Spotted by Dave Boutcher... [PATCH] hugetlbpage: remove include linux/module.h From: Paul Jackson The #include of linux/module.h in several arch/*/mm/hugetlbpage.c files for EXPORT_SYMBOL(hugetlb_total_pages) is now unneeded. [PATCH] slab alignment fixes From: Manfred Spraul Below is a patch that redefines the kmem_cache_alloc `align' argument: - align not zero: use the specified alignment. I think values smaller than sizeof(void*) will work, even on archs with strict alignment requirement (or at least: slab shouldn't crash. Obviously the user must handle the alignment properly). - align zero: * debug on: align to sizeof(void*) * debug off, SLAB_HWCACHE_ALIGN clear: align to sizeof(void*) * debug off, SLAB_HWCACHE_ALIGN set: align to the smaller of - cache_line_size() - the object size, rounded up to the next power of two. Slab never honored cache align for tiny objects: otherwise the 32-byte kmalloc objects would use 128 byte objects. There is one additional point: right now slab uses ints for the bufctls. Using short would save two bytes for each object. Initially I had used short, but davem objected. IIRC because some archs do not handle short efficiently. Should I allow arch overrides for the bufctls? On i386, saving two bytes might allow a few additional anon_vma objects in each page. [PATCH] prune_dcache comment fix From: Maneesh Soni Clarify some things which were confusing poeple... [PATCH] m68k: Amiga A2065 Ethernet KERN_* From: Geert Uytterhoeven Amiga A2065 Ethernet: Add KERN_* prefixes to printk() messages [PATCH] m68k bitops From: Geert Uytterhoeven M68k bitops updates (from Roman Zippel): - Optimize find_{first,next}_zero_bit() - Add missing implementations of find_{first,next}_bit() [PATCH] efivars: remove from arch/ia64 From: Matt Tolentino I broke up the efivars driver update patch I had sent out quite a while ago into several smaller patches. This includes several fixes and suggestions that were pointed out. The patches are broken down as follows: 1 - remove all traces of efivars from arch/ia64/ 2 - add new sysfs based efivars driver into drivers/firmware with accompanying Kconfig/Makefile changes to make it fully functional for ia64 again. 3 - cleans up x86 references to the /proc version of the efivars driver. [PATCH] efivars: add to drivers/firmware From: Matt Tolentino Second efivars driver update patch that adds efivars into the drivers/firmware directory. [PATCH] efivars: remove x86 references From: Matt Tolentino Third efivars driver update patch that removes x86 references to the /proc version of the driver. [PATCH] efivars fixes From: Matt Domsch Patch below fixes three small bugs in efivars.c as posted by Matt Tolentino last week and included in the latest -mm. Aside from this small patch, I'm quite pleased with Matt T's work, thanks! - dummy() used for reading write-only sysfs files should return -ENODEV to indicate failure, not 0. - efivar_create() should return the number of bytes written on success, not zero. - efivar_delete() should return the number of bytes written on success, not zero. [PATCH] efibootmgr location change From: Matt Domsch I moved the home of the efibootmgr utility from domsch.com to linux.dell.com. Note the move in drivers/firmware/Kconfig, also note version 0.5.0-test3 or above is necessary. [PATCH] s390: no timer interrupts in idle. From: Martin Schwidefsky This patch add a system control that allows to switch off the jiffies timer interrupts while a cpu sleeps in idle. This is useful for a system running with virtual cpus under z/VM. [PATCH] task_lock() comment update [PATCH] dio_bio_reap() return value fix This function doesn't correctly accumulate -EIO for its caller. [PATCH] fix CONFIG_SYSFS=n compile warning From: Matt Mackall [PATCH] isofs "default NLS charset not used" fix From: Marek Szuba Use the configured default NLS rather than hardwire it. [PATCH] slab: use order 0 for vfs caches We have interesting deadlocks when slab decides to use order-1 allocations for ext3_inode_cache. This is because ext3_alloc_inode() needs to perform a GFP_NOFS 1-order allocation. Sometimes the 1-order allocation needs to free a huge number of pages (tens of megabytes) before a 1-order grouping becomes available. But the GFP_NOFS allocator cannot free dcache (and hence icache) due to the deadlock problems identified in shrink_dcache_memory(). So change slab so that it will force 0-order allocations for shrinkable VFS objects. We can handle those OK. [PATCH] smb_writepage retval fix ->writepage() is supposed to return zero or a negative error code. Returning the number of bytes written will confuse the VFS. [PATCH] simplify put_page() By requiring that compound pages implement destructors we can drop some code from put_page(). [PATCH] hugepage fixes From: William Lee Irwin III mm/hugetlb.c is putting the destructor in head->lru.prev not head[1].mapping; fix below along with nuking huge_page_release(), which simply duplicates put_page(). [PATCH] ISDN Eicon driver: remove call to trap usermode helper From: Armin Schindler This patch removes the function to call a usermode helper if the hardware and its firmware has a problem. This feature is already implemented in the user utility and the driver does not need to care about it. [PATCH] doc: tips for S3 resume on radeon cards From: Pavel Machek Stefan has pretty useful tips for getting S3 to work on radeon notebooks. This brings whole new class of systems to be usable for S3. [PATCH] ppc32: fix head_44x.S copyrights From: Matt Porter A copyright change to properly show the lineage of this file. [PATCH] fs/proc/array.c: workaround for gcc-2.96 From: Alan Stern This patch is needed to work around gcc-2.96's limited ability to cope with long long intermediate expression types. I don't know why the code compiled okay earlier and failed now. [PATCH] blkdev.h: functions no longer inline From: "Randy.Dunlap" These are EXPORTed SYMBOLs; 'inline' was removed from them in ll_rw_blk.c on 2002-11-25. [PATCH] nfs_writepage() retval fix Both ->writepage() and ->writepages() should return 0 or a negative error code. It shouldn't return the number of bytes which were written. From: Trond Myklebust Redirty the page if all of the modified parts of the page weren't written out. [PATCH] ppc32: dma_unmap_page() fix From: (Arthur Othieno) Duplicate definition of dma_unmap_single() should actually be dma_unmap_page(). (acked by Tom Rini) [PATCH] fix fs/proc/task_nommu.c compile From: Christoph Hellwig this file has been broken for ages, but it seems few !CONFIG_MMU users use mainline at all. [PATCH] remove Documentation/DocBook/parportbook.tmpl From: Christoph Hellwig The partportbook is licensed under the GFDL and Linus agreed to remove all GFDL licensed files in http://www.ussg.iu.edu/hypermail/linux/kernel/0306.1/1968.html. I pinged the author the first time on the 2nd of april but still didn't get a reply, then send a patch to Linus to remove it last week but linus ignored it. Here's the patch again: [PATCH] Set module license in mcheck/non-fatal.c From: Herbert Xu This patch sets the module license for mcheck/non-fatal.c. The module doesn't work at all without this as one of the symbols it needs is only exported as GPL. [PATCH] SubmittingPatches diffing update. From: Zwane Mwaikambo A kernel janitor recently got confused by the advice in SubmittingPatches and was sending patches with the wrong strip level, i think just about everyone would prefer standard patches. Also mention various patch management scripts for batching up large deltas. [PATCH] ppc64: remove duplicated mb() and comment from __cpu_up From: Nathan Lynch This seems to have slipped in during a manual merge at some point. [PATCH] SELinux ptrace race fix From: Stephen Smalley Looking again at the SELinux ptrace check, I believe that there is an unrelated race due to the fact that the parent link is only updated after releasing the task lock in ptrace_attach (and this is necessary as task lock doesn't nest with write lock of tasklist_lock). The patch below changes SELinux to save the tracing process' SID upon a successful selinux_ptrace hook call and then use that SID in the ptrace check in apply_creds in order to avoid such races. This allows us to preserve the fine-grained process-to-process ptrace check upon exec (vs. the global CAP_SYS_PTRACE privilege => PT_PTRACE_CAP flag used by the capability module) while still avoiding races. [PATCH] doc: specifiying module parameters From: "Randy.Dunlap" kernel-parameters.txt: add info on how to specify loadable module parameters vs. built-in module parameters [PATCH] kbuild: Improved external module support From: Sam Ravnborg The external module support recently introduced caused a number of problems: - To build an external module the Module.symvers file was needed - To create the Module.symvers file a module was required - If Module.symvers was missing kbuild boiled out with an error - If vmlinux was missing also the stage 2 of module build failed (make -k) - It was not documented what was needed to actually bauild a module The following patch addresses this by adding the following functionality: - Always generate the Module.symvers file - Ignore a missing Module.symvers file - Add a new target modules_prepare, it prepares the kernel for building external modules, and is also usefull with O= - And it adds some more comments to Makefile.modpost, so others may follow it with some luck - .modpost.cmd is no longer generated This should close all reports on issues with respect to building external modules with current kernel - which has been identified as kernel problems. JFS: Fix non-ascii file name problem An unintentional sign extention caused non-ascii characters to be stored incorrectly, leading to inaccessible files. Changing a declaration to unsigned char fixes the problem. [PATCH] fix SG_IO page leak We cannot always rely on ->biotail remaining untouched. Currently we leak all the pinned user pages when doing cdda ripping at least, so I see no way around keeping the bio pointer seperate and passing it back in for unmap. Alternatively, we could invent a struct blk_map_data and put it on the stack for passing to both map and unmap. Include in files that need them. [CRYPTO]: Add module aliases for des and sha512. [TG3]: Undo comment typo fix, it was wrong. [COMPAT]: HDIO_DRIVE_TASK is a compatible ioctl. [PATCH] Fix cpumask iterator over empty cpu set Can't use _ffs() without first checking for zero, and if bits beyond NR_CPUS set it'll give bogus results. Use find_first_bit Linux 2.6.6-rc3