aboutsummaryrefslogtreecommitdiffstats
path: root/fs
Commit message (Collapse)AuthorAgeFilesLines
* [PATCH] fs/dcache.c: update comment of d_validate()Qinghuang Feng2008-10-231-2/+0
| | | | | | | Parameters @hash and @len have been removed since 2.4.3, now just to delete them. Signed-off-by: Qinghuang Feng <qhfeng.kernel@gmail.com>
* [RFC PATCH] touch_mnt_namespace when the mount flags changeDan Williams2008-10-231-1/+6
| | | | | | | | | | Daemons that need to be launched while the rootfs is read-only can now poll /proc/mounts to be notified when their O_RDWR requests may no longer end in EROFS. Cc: Kay Sievers <kay.sievers@vrfy.org> Cc: Neil Brown <neilb@suse.de> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
* [PATCH] reiserfs: add missing llseek methodChristoph Hellwig2008-10-231-0/+1
| | | | | | | | | Reiserfs currently doesn't set a llseek method for regular files, which means it will fall back to default_llseek. This means no one can seek beyond 2 Gigabytes on reiserfs, and that there's not protection vs the i_size updates from writers. Signed-off-by: Christoph Hellwig <hch@lst.de>
* [PATCH] fix ->llseek for more directoriesChristoph Hellwig2008-10-2311-1/+12
| | | | | | | With this patch all directory fops instances that have a readdir that doesn't take the BKL are switched to generic_file_llseek. Signed-off-by: Christoph Hellwig <hch@lst.de>
* [PATCH vfs-2.6 6/6] vfs: add LOOKUP_RENAME_TARGET intentOGAWA Hirofumi2008-10-231-0/+1
| | | | | | | | | | | This adds LOOKUP_RENAME_TARGET intent for lookup of rename destination. LOOKUP_RENAME_TARGET is going to be used like LOOKUP_CREATE. But since the destination of rename() can be existing directory entry, so it has a difference. Although that difference doesn't matter in my usage, this tells it to user of this intent. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
* [PATCH vfs-2.6 5/6] vfs: remove LOOKUP_PARENT from non LOOKUP_PARENT lookupOGAWA Hirofumi2008-10-231-9/+18
| | | | | | | | | | lookup_hash() with LOOKUP_PARENT is bogus. And this prepares to add new intent on those path. The user of LOOKUP_PARENT intent is nfs only, and it checks whether nd->flags has LOOKUP_CREATE or LOOKUP_OPEN, so the result is same. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
* [PATCH vfs-2.6 4/6] vfs: remove unnecessary fsnotify_d_instantiate()OGAWA Hirofumi2008-10-231-1/+0
| | | | | | | This calls d_move(), so fsnotify_d_instantiate() is unnecessary like rename path. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
* [PATCH vfs-2.6 3/6] vfs: add __d_instantiate() helperOGAWA Hirofumi2008-10-231-15/+16
| | | | | | | | | | | | This adds __d_instantiate() for users which is already taking dcache_lock, and replace with it. The part of d_add_ci() isn't equivalent. But it should be needed fsnotify_d_instantiate() actually, because the path is to add the inode to negative dentry. fsnotify_d_instantiate() should be called after change from negative to positive. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
* [PATCH vfs-2.6 2/6] vfs: add d_ancestor()OGAWA Hirofumi2008-10-232-34/+33
| | | | | | | | | | This adds d_ancestor() instead of d_isparent(), then use it. If new_dentry == old_dentry, is_subdir() returns 1, looks strange. "new_dentry == old_dentry" is not subdir obviously. But I'm not checking callers for now, so this keeps current behavior. Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
* [PATCH vfs-2.6 1/6] vfs: replace parent == dentry->d_parent by IS_ROOT()OGAWA Hirofumi2008-10-232-11/+14
| | | | Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
* [PATCH] get rid of on-stack dentry in udfAl Viro2008-10-231-15/+11
| | | | Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH 2/2] anondev: switch to IDAAlexey Dobriyan2008-10-231-5/+5
| | | | Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
* [PATCH 1/2] anondev: init IDR staticallyAlexey Dobriyan2008-10-231-6/+1
| | | | Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
* [JFFS2] Use d_splice_alias() not d_add() in jffs2_lookup()David Woodhouse2008-10-231-3/+1
| | | | | | | Now that JFFS2 can be exported by NFS, we need to get this right. Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] Optimise NFS readdir hack slightly.David Woodhouse2008-10-231-2/+3
| | | | | | | | | | | Avoid calling the underlying ->readdir() again when we reached the end already; keep going round the loop only if we stopped due to our own buffer being full. [AV: tidy the things up a bit, while we are there] Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] prepare vfs_readdir() callers to returning filldir resultAl Viro2008-10-234-33/+29
| | | | | | | It's not the final state, but it allows moving ->readdir() instances to passing filldir return value to caller of vfs_readdir(). Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] get rid of on-stack dentry in ext2_get_parent()Al Viro2008-10-233-20/+15
| | | | Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] get rid of on-stack fake dentry in ext3_get_parent()Al Viro2008-10-231-37/+33
| | | | | | | | Better pass parent and qstr to ext3_find_entry() explicitly than use such kludges, especially since the stack footprint is nasty enough and we have every chance to be deep in call chain. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [JFFS2] Reinstate NFS exportabilityDavid Woodhouse2008-10-231-0/+48
| | | | | | | | | | | Now that the readdir/lookup deadlock issues have been dealt with, we can export JFFS2 file systems again. (For now, you have to specify fsid manually; we should add a method to the export_ops to handle that too.) Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] Remove XFS buffered readdir hackDavid Woodhouse2008-10-231-128/+0
| | | | | | | | Now that we've moved the readdir hack to the nfsd code, we can remove the local version from the XFS code. Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] Copy XFS readdir hack into nfsd code.David Woodhouse2008-10-231-15/+93
| | | | | | | | | | | | | | | | | | | | | Some file systems with their own internal locking have problems with the way that nfsd calls the ->lookup() method from within a filldir function called from their ->readdir() method. The recursion back into the file system code can cause deadlock. XFS has a fairly hackish solution to this which involves doing the readdir() into a locally-allocated buffer, then going back through it calling the filldir function afterwards. It's not ideal, but it works. It's particularly suboptimal because XFS does this for local file systems too, where it's completely unnecessary. Copy this hack into the NFS code where it can be used only for NFS export. In response to feedback, use it unconditionally rather than only for the affected file systems. Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] Factor out nfsd_do_readdir() into its own functionDavid Woodhouse2008-10-231-16/+24
| | | | | Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] reduce the stack footprint of exportfs_decode_fh()Al Viro2008-10-231-5/+4
| | | | | | no need to have _two_ 256-byte arrays on stack... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] kill d_alloc_anonChristoph Hellwig2008-10-231-71/+37
| | | | | | | Remove d_alloc_anon now that no users are left. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] switch all filesystems over to d_obtain_aliasChristoph Hellwig2008-10-2319-305/+78
| | | | | | | Switch all users of d_alloc_anon to d_obtain_alias. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] new helper: d_obtain_aliasChristoph Hellwig2008-10-232-24/+37
| | | | | | | | | | | | | | | | The calling conventions of d_alloc_anon are rather unfortunate for all users, and it's name is not very descriptive either. Add d_obtain_alias as a new exported helper that drops the inode reference in the failure case, too and allows to pass-through NULL pointers and inodes to allow for tail-calls in the export operations. Incidentally this helper already existed as a private function in libfs.c as exportfs_d_alloc so kill that one and switch the callers to d_obtain_alias. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] generic_file_llseek tidyupsChristoph Hellwig2008-10-231-18/+40
| | | | | | | | Add kerneldoc for generic_file_llseek and generic_file_llseek_unlocked, use sane variable names and unclutter the code. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] tidy up chrdev_openChristoph Hellwig2008-10-231-7/+14
| | | | | | | Use a single goto label for chrdev_put + return error cases. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] hpfs: cleanup ->setattrChristoph Hellwig2008-10-234-13/+22
| | | | | | | | Reformat hpfs_notify_change to standard kernel style to make it readable and rename it to hpfs_setattr as that's what the method is called. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] make O_EXCL in nd->intent.flags visible in nd->flagsAl Viro2008-10-233-6/+6
| | | | | | | New flag: LOOKUP_EXCL. Set before doing the final step of pathname resolution on the paths that have LOOKUP_CREATE and O_EXCL. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] get rid of path_lookup_create()Al Viro2008-10-231-39/+22
| | | | | | | | ... and don't pass bogus flags when we are just looking for parent. Fold __path_lookup_intent_open() into path_lookup_open() while we are at it; that's the only remaining caller. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] assorted path_lookup() -> kern_path() conversionsAl Viro2008-10-233-29/+24
| | | | | | more nameidata eviction Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] switch nfsd to kern_path()Al Viro2008-10-234-58/+56
| | | | Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] sanitize svc_export_parse()Al Viro2008-10-231-42/+35
| | | | | | clean up the exit paths, get rid of nameidata Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] switch quota_on-related stuff to kern_path()Al Viro2008-10-234-37/+37
| | | | Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] no need for noinline stuff in fs/namespace.c anymoreAl Viro2008-10-231-12/+5
| | | | | | | Stack footprint from hell had been due to many struct nameidata in there. No more. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] finally get rid of nameidata in namespace.cAl Viro2008-10-231-60/+59
| | | | Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* [PATCH] new helper - kern_path()Al Viro2008-10-231-0/+10
| | | | | | Analog of lookup_path(), takes struct path *. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* binfmt_elf_fdpic: Update for cputime changes.Paul Mundt2008-10-201-12/+7
| | | | | | | | | | | | | Commit f06febc96ba8e0af80bcc3eaec0a109e88275fac ("timers: fix itimer/ many thread hang") introduced a new task_cputime interface and subsequently only converted binfmt_elf over to it. This results in the build for binfmt_elf_fdpic blowing up given that p->signal->{u,s}time have disappeared from underneath us. Apply the same trivial fix from binfmt_elf to binfmt_elf_fdpic. Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* Merge branch 'genirq-v28-for-linus' of ↵Linus Torvalds2008-10-201-21/+19
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip This merges branches irq/genirq, irq/sparseirq-v4, timers/hpet-percpu and x86/uv. The sparseirq branch is just preliminary groundwork: no sparse IRQs are actually implemented by this tree anymore - just the new APIs are added while keeping the old way intact as well (the new APIs map 1:1 to irq_desc[]). The 'real' sparse IRQ support will then be a relatively small patch ontop of this - with a v2.6.29 merge target. * 'genirq-v28-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (178 commits) genirq: improve include files intr_remapping: fix typo io_apic: make irq_mis_count available on 64-bit too genirq: fix name space collisions of nr_irqs in arch/* genirq: fix name space collision of nr_irqs in autoprobe.c genirq: use iterators for irq_desc loops proc: fixup irq iterator genirq: add reverse iterator for irq_desc x86: move ack_bad_irq() to irq.c x86: unify show_interrupts() and proc helpers x86: cleanup show_interrupts genirq: cleanup the sparseirq modifications genirq: remove artifacts from sparseirq removal genirq: revert dynarray genirq: remove irq_to_desc_alloc genirq: remove sparse irq code genirq: use inline function for irq_to_desc genirq: consolidate nr_irqs and for_each_irq_desc() x86: remove sparse irq from Kconfig genirq: define nr_irqs for architectures with GENERIC_HARDIRQS=n ...
| * proc: fixup irq iteratorThomas Gleixner2008-10-161-5/+2
| | | | | | | | | | | | | | There is no need for irq_desc here. Even for sparse_irq we can handle this clever in for_each_irq_nr(). Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
| * genirq: remove sparse irq codeThomas Gleixner2008-10-161-38/+5
| | | | | | | | | | | | | | This code is not ready, but we need to rip it out instead of rebasing as we would lose the APIC/IO_APIC unification otherwise. Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
| * x86: use 28 bits irq NR for pci msi/msix and htYinghai Lu2008-10-161-1/+1
| | | | | | | | | | | | | | | | also print out irq no in /proc/interrups and /proc/stat in hex, so could tell bus/dev/func. Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| * x86_64: make /proc/interrupts work with dyn irq_descYinghai Lu2008-10-161-4/+24
| | | | | | | | | | | | | | loop with irq_desc list Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| * irq, fs/proc: replace loop with nr_irqs for proc/statYinghai Lu2008-10-161-14/+28
| | | | | | | | | | | | | | | | | | | | Replace another nr_irqs loop to avoid the allocation of all sparse irq entries - use for_each_irq_desc instead. v2: make sure arch without GENERIC_HARDIRQS works too Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| * x86: move kstat_irqs from kstat to irq_descYinghai Lu2008-10-161-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | based on Eric's patch ... together mold it with dyn_array for irq_desc, will allcate kstat_irqs for nr_irq_desc alltogether if needed. -- at that point nr_cpus is known already. v2: make sure system without generic_hardirqs works they don't have irq_desc v3: fix merging v4: [mingo@elte.hu] fix typo [ mingo@elte.hu ] irq: build fix fix: arch/x86/xen/spinlock.c: In function 'xen_spin_lock_slow': arch/x86/xen/spinlock.c:90: error: 'struct kernel_stat' has no member named 'irqs' Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| * fs/proc: use nr_irqsYinghai Lu2008-10-161-5/+5
| | | | | | | | | | Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
* | Merge branch 'v28-timers-for-linus' of ↵Linus Torvalds2008-10-202-16/+11
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'v28-timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (36 commits) fix documentation of sysrq-q really Fix documentation of sysrq-q timer_list: add base address to clock base timer_list: print cpu number of clockevents device timer_list: print real timer address NOHZ: restart tick device from irq_enter() NOHZ: split tick_nohz_restart_sched_tick() NOHZ: unify the nohz function calls in irq_enter() timers: fix itimer/many thread hang, fix timers: fix itimer/many thread hang, v3 ntp: improve adjtimex frequency rounding timekeeping: fix rounding problem during clock update ntp: let update_persistent_clock() sleep hrtimer: reorder struct hrtimer to save 8 bytes on 64bit builds posix-timers: lock_timer: make it readable posix-timers: lock_timer: kill the bogus ->it_id check posix-timers: kill ->it_sigev_signo and ->it_sigev_value posix-timers: sys_timer_create: cleanup the error handling posix-timers: move the initialization of timer->sigq from send to create path posix-timers: sys_timer_create: simplify and s/tasklist/rcu/ ... Fix trivial conflicts due to sysrq-q description clahes in Documentation/sysrq.txt and drivers/char/sysrq.c
| | \
| | \
| | \
| | \
| | \
| | \
| *-----. \ Merge branches 'timers/clocksource', 'timers/hrtimers', 'timers/nohz', ↵Thomas Gleixner2008-10-20298-7690/+15600
| |\ \ \ \ \ | | | | | | | | | | | | | | | | | | | | | 'timers/ntp', 'timers/posixtimers' and 'timers/debug' into v28-timers-for-linus
| | | | * | | timers: fix itimer/many thread hangFrank Mayhar2008-09-142-16/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Overview This patch reworks the handling of POSIX CPU timers, including the ITIMER_PROF, ITIMER_VIRT timers and rlimit handling. It was put together with the help of Roland McGrath, the owner and original writer of this code. The problem we ran into, and the reason for this rework, has to do with using a profiling timer in a process with a large number of threads. It appears that the performance of the old implementation of run_posix_cpu_timers() was at least O(n*3) (where "n" is the number of threads in a process) or worse. Everything is fine with an increasing number of threads until the time taken for that routine to run becomes the same as or greater than the tick time, at which point things degrade rather quickly. This patch fixes bug 9906, "Weird hang with NPTL and SIGPROF." Code Changes This rework corrects the implementation of run_posix_cpu_timers() to make it run in constant time for a particular machine. (Performance may vary between one machine and another depending upon whether the kernel is built as single- or multiprocessor and, in the latter case, depending upon the number of running processors.) To do this, at each tick we now update fields in signal_struct as well as task_struct. The run_posix_cpu_timers() function uses those fields to make its decisions. We define a new structure, "task_cputime," to contain user, system and scheduler times and use these in appropriate places: struct task_cputime { cputime_t utime; cputime_t stime; unsigned long long sum_exec_runtime; }; This is included in the structure "thread_group_cputime," which is a new substructure of signal_struct and which varies for uniprocessor versus multiprocessor kernels. For uniprocessor kernels, it uses "task_cputime" as a simple substructure, while for multiprocessor kernels it is a pointer: struct thread_group_cputime { struct task_cputime totals; }; struct thread_group_cputime { struct task_cputime *totals; }; We also add a new task_cputime substructure directly to signal_struct, to cache the earliest expiration of process-wide timers, and task_cputime also replaces the it_*_expires fields of task_struct (used for earliest expiration of thread timers). The "thread_group_cputime" structure contains process-wide timers that are updated via account_user_time() and friends. In the non-SMP case the structure is a simple aggregator; unfortunately in the SMP case that simplicity was not achievable due to cache-line contention between CPUs (in one measured case performance was actually _worse_ on a 16-cpu system than the same test on a 4-cpu system, due to this contention). For SMP, the thread_group_cputime counters are maintained as a per-cpu structure allocated using alloc_percpu(). The timer functions update only the timer field in the structure corresponding to the running CPU, obtained using per_cpu_ptr(). We define a set of inline functions in sched.h that we use to maintain the thread_group_cputime structure and hide the differences between UP and SMP implementations from the rest of the kernel. The thread_group_cputime_init() function initializes the thread_group_cputime structure for the given task. The thread_group_cputime_alloc() is a no-op for UP; for SMP it calls the out-of-line function thread_group_cputime_alloc_smp() to allocate and fill in the per-cpu structures and fields. The thread_group_cputime_free() function, also a no-op for UP, in SMP frees the per-cpu structures. The thread_group_cputime_clone_thread() function (also a UP no-op) for SMP calls thread_group_cputime_alloc() if the per-cpu structures haven't yet been allocated. The thread_group_cputime() function fills the task_cputime structure it is passed with the contents of the thread_group_cputime fields; in UP it's that simple but in SMP it must also safely check that tsk->signal is non-NULL (if it is it just uses the appropriate fields of task_struct) and, if so, sums the per-cpu values for each online CPU. Finally, the three functions account_group_user_time(), account_group_system_time() and account_group_exec_runtime() are used by timer functions to update the respective fields of the thread_group_cputime structure. Non-SMP operation is trivial and will not be mentioned further. The per-cpu structure is always allocated when a task creates its first new thread, via a call to thread_group_cputime_clone_thread() from copy_signal(). It is freed at process exit via a call to thread_group_cputime_free() from cleanup_signal(). All functions that formerly summed utime/stime/sum_sched_runtime values from from all threads in the thread group now use thread_group_cputime() to snapshot the values in the thread_group_cputime structure or the values in the task structure itself if the per-cpu structure hasn't been allocated. Finally, the code in kernel/posix-cpu-timers.c has changed quite a bit. The run_posix_cpu_timers() function has been split into a fast path and a slow path; the former safely checks whether there are any expired thread timers and, if not, just returns, while the slow path does the heavy lifting. With the dedicated thread group fields, timers are no longer "rebalanced" and the process_timer_rebalance() function and related code has gone away. All summing loops are gone and all code that used them now uses the thread_group_cputime() inline. When process-wide timers are set, the new task_cputime structure in signal_struct is used to cache the earliest expiration; this is checked in the fast path. Performance The fix appears not to add significant overhead to existing operations. It generally performs the same as the current code except in two cases, one in which it performs slightly worse (Case 5 below) and one in which it performs very significantly better (Case 2 below). Overall it's a wash except in those two cases. I've since done somewhat more involved testing on a dual-core Opteron system. Case 1: With no itimer running, for a test with 100,000 threads, the fixed kernel took 1428.5 seconds, 513 seconds more than the unfixed system, all of which was spent in the system. There were twice as many voluntary context switches with the fix as without it. Case 2: With an itimer running at .01 second ticks and 4000 threads (the most an unmodified kernel can handle), the fixed kernel ran the test in eight percent of the time (5.8 seconds as opposed to 70 seconds) and had better tick accuracy (.012 seconds per tick as opposed to .023 seconds per tick). Case 3: A 4000-thread test with an initial timer tick of .01 second and an interval of 10,000 seconds (i.e. a timer that ticks only once) had very nearly the same performance in both cases: 6.3 seconds elapsed for the fixed kernel versus 5.5 seconds for the unfixed kernel. With fewer threads (eight in these tests), the Case 1 test ran in essentially the same time on both the modified and unmodified kernels (5.2 seconds versus 5.8 seconds). The Case 2 test ran in about the same time as well, 5.9 seconds versus 5.4 seconds but again with much better tick accuracy, .013 seconds per tick versus .025 seconds per tick for the unmodified kernel. Since the fix affected the rlimit code, I also tested soft and hard CPU limits. Case 4: With a hard CPU limit of 20 seconds and eight threads (and an itimer running), the modified kernel was very slightly favored in that while it killed the process in 19.997 seconds of CPU time (5.002 seconds of wall time), only .003 seconds of that was system time, the rest was user time. The unmodified kernel killed the process in 20.001 seconds of CPU (5.014 seconds of wall time) of which .016 seconds was system time. Really, though, the results were too close to call. The results were essentially the same with no itimer running. Case 5: With a soft limit of 20 seconds and a hard limit of 2000 seconds (where the hard limit would never be reached) and an itimer running, the modified kernel exhibited worse tick accuracy than the unmodified kernel: .050 seconds/tick versus .028 seconds/tick. Otherwise, performance was almost indistinguishable. With no itimer running this test exhibited virtually identical behavior and times in both cases. In times past I did some limited performance testing. those results are below. On a four-cpu Opteron system without this fix, a sixteen-thread test executed in 3569.991 seconds, of which user was 3568.435s and system was 1.556s. On the same system with the fix, user and elapsed time were about the same, but system time dropped to 0.007 seconds. Performance with eight, four and one thread were comparable. Interestingly, the timer ticks with the fix seemed more accurate: The sixteen-thread test with the fix received 149543 ticks for 0.024 seconds per tick, while the same test without the fix received 58720 for 0.061 seconds per tick. Both cases were configured for an interval of 0.01 seconds. Again, the other tests were comparable. Each thread in this test computed the primes up to 25,000,000. I also did a test with a large number of threads, 100,000 threads, which is impossible without the fix. In this case each thread computed the primes only up to 10,000 (to make the runtime manageable). System time dominated, at 1546.968 seconds out of a total 2176.906 seconds (giving a user time of 629.938s). It received 147651 ticks for 0.015 seconds per tick, still quite accurate. There is obviously no comparable test without the fix. Signed-off-by: Frank Mayhar <fmayhar@google.com> Cc: Roland McGrath <roland@redhat.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
generated by cgit v1.1 at 2024-09-14 00:59:15 +0000