)]}' { "log": [ { "commit": "7978672c4d9a1e6a6081de3a9d9ba5e5b24904a0", "tree": "a7deb3a6e339e691078c770d4b9cbc05d99a2243", "parents": [ "ff60a5dc4fa584d47022d2533bc5c53b80096fb5" ], "author": { "name": "George Anzinger", "email": "george@wildturkeyranch.net", "time": "Wed Feb 01 03:05:11 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Wed Feb 01 08:53:13 2006 -0800" }, "message": "[PATCH] hrtimers: cleanups and simplifications\n\nClean up the interface to hrtimers by changing the init code to pass the mode\nas well as the clock. This allow the init code to select the correct base and\neliminates extra timer re-init code in posix-timers. We also simplify the\nrestart interface nanosleep use.\n\nSigned-off-by: George Anzinger \u003cgeorge@mvista.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "5fd63b308569060ffa40af52ed122d9734111bff", "tree": "746c2701ec42374b1ec389a6af0d438c9105a9d7", "parents": [ "99f7b77d3c62045bb47dfc4228a2341ba71bff1d" ], "author": { "name": "Ravikiran G Thirumalai", "email": "kiran@scalex86.org", "time": "Wed Jan 11 22:46:15 2006 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Wed Jan 11 19:05:01 2006 -0800" }, "message": "[PATCH] x86_64: Inclusion of ScaleMP vSMP architecture patches - vsmp_align\n\nvSMP specific alignment patch to\n1. Define INTERNODE_CACHE_SHIFT for vSMP\n2. Use this for alignment of critical structures\n3. Use INTERNODE_CACHE_SHIFT for ARCH_MIN_TASKALIGN,\n and let the slab align task_struct allocations to the internode cacheline size\n4. Introduce and use ARCH_MIN_MMSTRUCT_ALIGN for mm_struct slab allocations.\n\nSigned-off-by: Ravikiran Thirumalai \u003ckiran@scalemp.com\u003e\nSigned-off-by: Shai Fultheim \u003cshai@scalemp.com\u003e\nSigned-off-by: Andi Kleen \u003cak@suse.de\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "c59ede7b78db329949d9cdcd7064e22d357560ef", "tree": "f9dc9d464fdad5bfd464d983e77c1af031389dda", "parents": [ "e16885c5ad624a6efe1b1bf764e075d75f65a788" ], "author": { "name": "Randy.Dunlap", "email": "rdunlap@xenotime.net", "time": "Wed Jan 11 12:17:46 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Wed Jan 11 18:42:13 2006 -0800" }, "message": "[PATCH] move capable() to capability.h\n\n- Move capable() from sched.h to capability.h;\n\n- Use \u003clinux/capability.h\u003e where capable() is used\n\t(in include/, block/, ipc/, kernel/, a few drivers/,\n\tmm/, security/, \u0026 sound/;\n\tmany more drivers/ to go)\n\nSigned-off-by: Randy Dunlap \u003crdunlap@xenotime.net\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "2ff678b8da6478d861c1b0ecb3ac14575760e906", "tree": "0ca983ce820ac8bb9f6e8b193926e0804116a7e0", "parents": [ "df78488de7befd387e9d060da6e18bb5d1cb882c" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Jan 09 20:52:34 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Tue Jan 10 08:01:38 2006 -0800" }, "message": "[PATCH] hrtimer: switch itimers to hrtimer\n\nswitch itimers to a hrtimers-based implementation\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "408894ee4dd4debfdedd472eb4d8414892fc90f6", "tree": "b324c1086b804cc05c2839f9d9675eef2bc7c949", "parents": [ "f3f54ffa703c6298240ffd69616451d645bae4d5" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Jan 09 15:59:20 2006 -0800" }, "committer": { "name": "Ingo Molnar", "email": "mingo@hera.kernel.org", "time": "Mon Jan 09 15:59:20 2006 -0800" }, "message": "[PATCH] mutex subsystem, debugging code\n\nmutex implementation - add debugging code.\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Arjan van de Ven \u003carjan@infradead.org\u003e\n" }, { "commit": "fe7d37d1fbf8ffe78abd72728b24fb0c64f7af55", "tree": "274d82b06d2a711e84381b1357c6332077181d27", "parents": [ "ddc0f846aa7621940b74cee0c91cd26405058a4d" ], "author": { "name": "Oleg Nesterov", "email": "oleg@tv-sign.ru", "time": "Sun Jan 08 01:04:02 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Jan 08 20:14:01 2006 -0800" }, "message": "[PATCH] copy_process: error path cleanup\n\nThis patch moves \u0027fork_out:\u0027 under \u0027bad_fork_free:\u0027, and removes now\nunneeded \u0027if (retval)\u0027 check.\n\nSigned-off-by: Oleg Nesterov \u003coleg@tv-sign.ru\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "9a5d3023e626a0baf86ac6b892c983b3db13f22b", "tree": "41d1be6513fc3a5367414654eeaf11bbb12d4814", "parents": [ "fb86a35b9ded8a7e53a432cbf28df603cdd4849c" ], "author": { "name": "Oren Laadan", "email": "orenl@cs.columbia.edu", "time": "Sun Jan 08 01:03:51 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Jan 08 20:14:00 2006 -0800" }, "message": "[PATCH] fork: fix race in setting child\u0027s pgrp and tty\n\nIn fork, child should recopy parent\u0027s pgrp/tty after it has tasklist_lock.\nOtherwise following a setpgid() on the parent, *after* copy_signal(), the\nchild will own a stale pgrp (which may be reused); (eg. if copy_mm()\nsleeps a long while due to memory pressure). Similar issue for the tty.\n\nSigned-off-by: Oren Laadan \u003corenl@cs.columbia.edu\u003e\nCc: Oleg Nesterov \u003coleg@tv-sign.ru\u003e\nCc: Roland McGrath \u003croland@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "b4b2641843db124637fa3d2cb2101982035dcc82", "tree": "fe8ed223f2e1828a6d14090711bad1c864aded09", "parents": [ "59dac16fb95f09253b8086134443abeb439703cd" ], "author": { "name": "Paul Jackson", "email": "pj@sgi.com", "time": "Sun Jan 08 01:01:53 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Jan 08 20:13:43 2006 -0800" }, "message": "[PATCH] cpuset: fork hook fix\n\nFix obscure, never seen in real life, cpuset fork race. The cpuset_fork()\ncall in fork.c was setting up the correct task-\u003ecpuset pointer after the\ntasklist_lock was dropped, which briefly exposed the newly forked process with\nan unsafe (copied from parent without locks or usage counter increment) cpuset\npointer.\n\nIn theory, that exposed cpuset pointer could have been pointing at a cpuset\nthat was already freed and removed, and in theory another task that had been\nsitting on the tasklist_lock waiting to scan the task list could have raced\ndown the entire tasklist, found our new child at the far end, and dereferenced\nthat bogus cpuset pointer.\n\nTo fix, setup up the correct cpuset pointer in the new child by calling\ncpuset_fork() before the new task is linked into the tasklist, and with that,\nadd a fork failure case, to dereference that cpuset, if the fork fails along\nthe way, after cpuset_fork() was called.\n\nHad to remove a BUG_ON() from cpuset_exit(), because it was no longer valid -\nthe call to cpuset_exit() from a failed fork would not have PF_EXITING set.\n\nSigned-off-by: Paul Jackson \u003cpj@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "e56d090310d7625ecb43a1eeebd479f04affb48b", "tree": "2f479215dff4a2d8f3a9ed85200a5bc4f51534be", "parents": [ "4369ef3c3e9d3bd9b879580678778f558d481e90" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Sun Jan 08 01:01:37 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Jan 08 20:13:40 2006 -0800" }, "message": "[PATCH] RCU signal handling\n\nRCU tasklist_lock and RCU signal handling: send signals RCU-read-locked\ninstead of tasklist_lock read-locked. This is a scalability improvement on\nSMP and a preemption-latency improvement under PREEMPT_RCU.\n\nSigned-off-by: Paul E. McKenney \u003cpaulmck@us.ibm.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nAcked-by: William Irwin \u003cwli@holomorphy.com\u003e\nCc: Roland McGrath \u003croland@redhat.com\u003e\nCc: Oleg Nesterov \u003coleg@tv-sign.ru\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "8c4b8add83c93306b07d78469fd351dc462e4b66", "tree": "e99523b961cb84f0462057d329de27e1d85546e6", "parents": [ "3148890bfa4f36c9949871264e06ef4d449eeff9" ], "author": { "name": "Paul Jackson", "email": "pj@sgi.com", "time": "Mon Nov 28 13:44:05 2005 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Nov 28 14:42:24 2005 -0800" }, "message": "[PATCH] cpuset fork locking fix\n\nMove the cpuset_fork() call below the write_unlock_irq call in\nkernel/fork.c copy_process().\n\nSince the cpuset-dual-semaphore-locking-overhaul.patch, the cpuset_fork()\nroutine acquires task_lock(), so cannot be called while holding the\ntasklist_lock for write.\n\nSigned-off-by: Paul Jackson \u003cpj@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "c13cf856cbe16aec3007604dc013cbf3a16c6686", "tree": "0d05f46a20de1abf59c02af4e819ccfae0a8d009", "parents": [ "ff88a3b2f56ae4f3296ea957ea38f99f8bd0e5a8" ], "author": { "name": "Andrew Morton", "email": "akpm@osdl.org", "time": "Mon Nov 28 13:43:48 2005 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Nov 28 14:42:23 2005 -0800" }, "message": "[PATCH] fork.c: proc_fork_connector() called under write_lock()\n\nDon\u0027t do that - it does GFP_KERNEL allocations, for a start.\n\n(Reported by Guillaume Thouvenin \u003cguillaume.thouvenin@bull.net\u003e)\n\nAcked-by: Matt Helsley \u003cmatthltc@us.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "0b0db14c536debd92328819fe6c51a49717e8440", "tree": "9fb2df9bc06a30bbc44d980614e441974be55660", "parents": [ "689bcebfda16d7bace742740bfb3137fff30b529" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Mon Nov 21 21:32:20 2005 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Tue Nov 22 09:13:43 2005 -0800" }, "message": "[PATCH] unpaged: copy_page_range vma\n\nFor copy_one_pte\u0027s print_bad_pte to show the task correctly (instead of\n\"???\"), dup_mmap must pass down parent vma rather than child vma.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "20dcae32439384b6863c626bb3b2a09bed65b33e", "tree": "9750c39119447fb32963448bf1935e1ba22b2f9d", "parents": [ "4557398f8cbaf9f254cff747534b4724c7f75c4f" ], "author": { "name": "Zach Brown", "email": "zach.brown@oracle.com", "time": "Sun Nov 13 16:07:33 2005 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Nov 13 18:14:16 2005 -0800" }, "message": "[PATCH] aio: remove kioctx from mm_struct\n\nSync iocbs have a life cycle that don\u0027t need a kioctx. Their retrying, if\nany, is done in the context of their owner who has allocated them on the\nstack.\n\nThe sole user of a sync iocb\u0027s ctx reference was aio_complete() checking for\nan elevated iocb ref count that could never happen. No path which grabs an\niocb ref has access to sync iocbs.\n\nIf we were to implement sync iocb cancelation it would be done by the owner of\nthe iocb using its on-stack reference.\n\nRemoving this chunk from aio_complete allows us to remove the entire kioctx\ninstance from mm_struct, reducing its size by a third. On a i386 testing box\nthe slab size went from 768 to 504 bytes and from 5 to 8 per page.\n\nSigned-off-by: Zach Brown \u003czach.brown@oracle.com\u003e\nAcked-by: Benjamin LaHaise \u003cbcrl@kvack.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "10ebffde3d3916026974352b7900e44afe2b243f", "tree": "ac88fa33694f5bea1b6958e1e56bcd4c9b95b002", "parents": [ "a1261f54611ec4ad6a7ab7080f86747e3ac3685b" ], "author": { "name": "Al Viro", "email": "viro@parcelfarce.linux.theplanet.co.uk", "time": "Sun Nov 13 16:06:56 2005 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Nov 13 18:14:13 2005 -0800" }, "message": "[PATCH] m68k: introduce setup_thread_stack() and end_of_stack()\n\nencapsulates the rest of arch-dependent operations with thread_info access.\nTwo new helpers - setup_thread_stack() and end_of_stack(). For normal case\nthe former consists of copying thread_info of parent to new thread_info and\nthe latter returns pointer immediately past the end of thread_info.\n\nSigned-off-by: Al Viro \u003cviro@parcelfarce.linux.theplanet.co.uk\u003e\nSigned-off-by: Roman Zippel \u003czippel@linux-m68k.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "a1261f54611ec4ad6a7ab7080f86747e3ac3685b", "tree": "8a65c419da590e3712543f69284fb5f8cd613a37", "parents": [ "7feacd53347c04aee789ba5d632eda0c3fc421c4" ], "author": { "name": "Al Viro", "email": "viro@parcelfarce.linux.theplanet.co.uk", "time": "Sun Nov 13 16:06:55 2005 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Nov 13 18:14:13 2005 -0800" }, "message": "[PATCH] m68k: introduce task_thread_info\n\nnew helper - task_thread_info(task). On platforms that have thread_info\nallocated separately (i.e. in default case) it simply returns\ntask-\u003ethread_info. m68k wants (and for good reasons) to embed its thread_info\ninto task_struct. So it will (in later patch) have task_thread_info() of its\nown. For now we just add a macro for generic case and convert existing\ninstances of its body in core kernel to uses of new macro. Obviously safe -\nall normal architectures get the same preprocessor output they used to get.\n\nSigned-off-by: Al Viro \u003cviro@parcelfarce.linux.theplanet.co.uk\u003e\nSigned-off-by: Roman Zippel \u003czippel@linux-m68k.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "dedeb0029b9c83420fc1337d4ee53daa7b2a0ad4", "tree": "d87e66e1d6240cd412c20ecbc12f5b810c9807e4", "parents": [ "b8ae48656db860d4c83a29aa7b0588fc89361935" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Mon Nov 07 14:09:01 2005 -0800" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Nov 07 14:09:01 2005 -0800" }, "message": "[SPARC64] mm: context switch ptlock\n\nsparc64 is unique among architectures in taking the page_table_lock in\nits context switch (well, cris does too, but erroneously, and it\u0027s not\nyet SMP anyway).\n\nThis seems to be a private affair between switch_mm and activate_mm,\nusing page_table_lock as a per-mm lock, without any relation to its uses\nelsewhere. That\u0027s fine, but comment it as such; and unlock sooner in\nswitch_mm, more like in activate_mm (preemption is disabled here).\n\nThere is a block of \"if (0)\"ed code in smp_flush_tlb_pending which would\nhave liked to rely on the page_table_lock, in switch_mm and elsewhere;\nbut its comment explains how dup_mmap\u0027s flush_tlb_mm defeated it. And\nthough that could have been changed at any time over the past few years,\nnow the chance vanishes as we push the page_table_lock downwards, and\nperhaps split it per page table page. Just delete that block of code.\n\nWhich leaves the mysterious spin_unlock_wait(\u0026oldmm-\u003epage_table_lock)\nin kernel/fork.c copy_mm. Textual analysis (supported by Nick Piggin)\nsuggests that the comment was written by DaveM, and that it relates to\nthe defeated approach in the sparc64 smp_flush_tlb_pending. Just delete\nthis block too.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "9f46080c41d5f3f7c00b4e169ba4b0b2865258bf", "tree": "e2c029ef7f0cd5fb8ea9b78db3f7be5badaf59b1", "parents": [ "49364ce2534418462d681ad99e52e79a00b0f40b" ], "author": { "name": "Matt Helsley", "email": "matthltc@us.ibm.com", "time": "Mon Nov 07 00:59:16 2005 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Nov 07 07:53:35 2005 -0800" }, "message": "[PATCH] Process Events Connector\n\nThis patch adds a connector that reports fork, exec, id change, and exit\nevents for all processes to userspace. It replaces the fork_advisor patch\nthat ELSA is currently using. Applications that may find these events\nuseful include accounting/auditing (e.g. ELSA), system activity monitoring\n(e.g. top), security, and resource management (e.g. CKRM).\n\nSigned-off-by: Matt Helsley \u003cmatthltc@us.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "c74df32c724a1652ad8399b4891bb02c9d43743a", "tree": "5a79d56fdcf7dc2053a277dbf6db7c3b339e9659", "parents": [ "1bb3630e89cb8a7b3d3807629c20c5bad88290ff" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Sat Oct 29 18:16:23 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sat Oct 29 21:40:40 2005 -0700" }, "message": "[PATCH] mm: ptd_alloc take ptlock\n\nSecond step in pushing down the page_table_lock. Remove the temporary\nbridging hack from __pud_alloc, __pmd_alloc, __pte_alloc: expect callers not\nto hold page_table_lock, whether it\u0027s on init_mm or a user mm; take\npage_table_lock internally to check if a racing task already allocated.\n\nConvert their callers from common code. But avoid coming back to change them\nagain later: instead of moving the spin_lock(\u0026mm-\u003epage_table_lock) down,\nswitch over to new macros pte_alloc_map_lock and pte_unmap_unlock, which\nencapsulate the mapping+locking and unlocking+unmapping together, and in the\nend may use alternatives to the mm page_table_lock itself.\n\nThese callers all hold mmap_sem (some exclusively, some not), so at no level\ncan a page table be whipped away from beneath them; and pte_alloc uses the\n\"atomic\" pmd_present to test whether it needs to allocate. It appears that on\nall arches we can safely descend without page_table_lock.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "7ee78232501ea9de2b6c8f10d32c9a0fee541357", "tree": "2041a36a13bdd8b096dfbf52b63a87739ea97d6b", "parents": [ "fd3e42fcc888a773572282575d2fdbf5cfd6216e" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Sat Oct 29 18:16:08 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sat Oct 29 21:40:38 2005 -0700" }, "message": "[PATCH] mm: dup_mmap down new mmap_sem\n\nOne anomaly remains from when Andrea rationalized the responsibilities of\nmmap_sem and page_table_lock: in dup_mmap we add vmas to the child holding its\npage_table_lock, but not the mmap_sem which normally guards the vma list and\nrbtree. Which could be an issue for unuse_mm: though since it just walks down\nthe list (today with page_table_lock, tomorrow not), it\u0027s probably okay. Will\nneed a memory barrier? Oh, keep it simple, Nick and I agreed, no harm in\ntaking child\u0027s mmap_sem here.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "fd3e42fcc888a773572282575d2fdbf5cfd6216e", "tree": "ffe2c223b9200185eb67881582f784d534958883", "parents": [ "ae859762332f19bfc06f4c4a1b1fefb41e9e1084" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Sat Oct 29 18:16:06 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sat Oct 29 21:40:38 2005 -0700" }, "message": "[PATCH] mm: dup_mmap use oldmm more\n\nUse the parent\u0027s oldmm throughout dup_mmap, instead of perversely going back\nto current-\u003emm. (Can you hear the sigh of relief from those mpnts? Usually I\nsquash them, but not today.)\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "4294621f41a85497019fae64341aa5351a1921b7", "tree": "fdeb7eb44384a99d0679ffa6de5019bab0ea2166", "parents": [ "404351e67a9facb475abf1492245374a28d13e90" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Sat Oct 29 18:16:05 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sat Oct 29 21:40:38 2005 -0700" }, "message": "[PATCH] mm: rss \u003d file_rss + anon_rss\n\nI was lazy when we added anon_rss, and chose to change as few places as\npossible. So currently each anonymous page has to be counted twice, in rss\nand in anon_rss. Which won\u0027t be so good if those are atomic counts in some\nconfigurations.\n\nChange that around: keep file_rss and anon_rss separately, and add them\ntogether (with get_mm_rss macro) when the total is needed - reading two\natomics is much cheaper than updating two atomics. And update anon_rss\nupfront, typically in memory.c, not tucked away in page_add_anon_rmap.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "404351e67a9facb475abf1492245374a28d13e90", "tree": "5ef4e78b399b36a46eda339ad0cd27556fc5b9a2", "parents": [ "fc2acab31be8e869b2d5f6de12f557f6f054f19c" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Sat Oct 29 18:16:04 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sat Oct 29 21:40:38 2005 -0700" }, "message": "[PATCH] mm: mm_init set_mm_counters\n\nHow is anon_rss initialized? In dup_mmap, and by mm_alloc\u0027s memset; but\nthat\u0027s not so good if an mm_counter_t is a special type. And how is rss\ninitialized? By set_mm_counter, all over the place. Come on, we just need to\ninitialize them both at once by set_mm_counter in mm_init (which follows the\nmemcpy when forking).\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "ab50b8ed818016cfecd747d6d4bb9139986bc029", "tree": "33c666578c14dccce05b3f7a5538405098eebcc4", "parents": [ "72866f6f277ec0ddd6df7a3b6ecdcf59a28de115" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Sat Oct 29 18:15:56 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sat Oct 29 21:40:37 2005 -0700" }, "message": "[PATCH] mm: vm_stat_account unshackled\n\nThe original vm_stat_account has fallen into disuse, with only one user, and\nonly one user of vm_stat_unaccount. It\u0027s easier to keep track if we convert\nthem all to __vm_stat_account, then free it from its __shackles.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "d1209d049bbc3df66650f8417637be4f7b57b604", "tree": "0712622c596cd3ca93234e8eca661240b5fa042b", "parents": [ "f9b25fabfddf142b5af8268692701d386cf37e4f" ], "author": { "name": "Alan Stern", "email": "stern@rowland.harvard.edu", "time": "Wed Oct 19 21:23:51 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Wed Oct 19 23:04:31 2005 -0700" }, "message": "[PATCH] Threads shouldn\u0027t inherit PF_NOFREEZE\n\nThe PF_NOFREEZE process flag should not be inherited when a thread is\nforked. This patch (as585) removes the flag from the child.\n\nThis problem is starting to show up more and more as drivers turn to the\nkthread API instead of using kernel_thread(). As a result, their kernel\nthreads are now children of the kthread worker instead of modprobe, and\nthey inherit the PF_NOFREEZE flag. This can cause problems during system\nsuspend; the kernel threads are not getting frozen as they ought to be.\n\nSigned-off-by: Alan Stern \u003cstern@rowland.harvard.edu\u003e\nAcked-by: Pavel Machek \u003cpavel@ucw.cz\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "26ff6ad9786abf6f40a6d3cbb89753b4fa50cb00", "tree": "8abf9b85bc5f081460f05c3e8f76328dc690f167", "parents": [ "fa72e26a640e551637f91122d782d5a85d514f5b" ], "author": { "name": "Srivatsa Vaddagiri", "email": "vatsa@in.ibm.com", "time": "Fri Sep 16 19:27:40 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sat Sep 17 11:49:59 2005 -0700" }, "message": "[PATCH] CPU hotplug breaks wake_up_new_task\n\nFix a problem wherein a new-born task is added to a dead CPU.\n\nSigned-off-by: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nAcked-by: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nAcked-by: Shaohua Li \u003cshaohua.li@intel.com\u003e\nAcked-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "ab2af1f5005069321c5d130f09cce577b03f43ef", "tree": "73a70ba486f522cd9eeeef376ede2b5a1c1b473b", "parents": [ "6e72ad2c581de121cc7e772469e2a8f6b1fd4379" ], "author": { "name": "Dipankar Sarma", "email": "dipankar@in.ibm.com", "time": "Fri Sep 09 13:04:13 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Sep 09 13:57:55 2005 -0700" }, "message": "[PATCH] files: files struct with RCU\n\nPatch to eliminate struct files_struct.file_lock spinlock on the reader side\nand use rcu refcounting rcuref_xxx api for the f_count refcounter. The\nupdates to the fdtable are done by allocating a new fdtable structure and\nsetting files-\u003efdt to point to the new structure. The fdtable structure is\nprotected by RCU thereby allowing lock-free lookup. For fd arrays/sets that\nare vmalloced, we use keventd to free them since RCU callbacks can\u0027t sleep. A\nglobal list of fdtable to be freed is not scalable, so we use a per-cpu list.\nIf keventd is already handling the current cpu\u0027s work, we use a timer to defer\nqueueing of that work.\n\nSince the last publication, this patch has been re-written to avoid using\nexplicit memory barriers and use rcu_assign_pointer(), rcu_dereference()\npremitives instead. This required that the fd information is kept in a\nseparate structure (fdtable) and updated atomically.\n\nSigned-off-by: Dipankar Sarma \u003cdipankar@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "badf16621c1f9d1ac753be056fce11b43d6e0be5", "tree": "3fdf833fdf2e3d3a439090743539680449ec3428", "parents": [ "c0dfb2905126e9e94edebbce8d3e05001301f52d" ], "author": { "name": "Dipankar Sarma", "email": "dipankar@in.ibm.com", "time": "Fri Sep 09 13:04:10 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Sep 09 13:57:55 2005 -0700" }, "message": "[PATCH] files: break up files struct\n\nIn order for the RCU to work, the file table array, sets and their sizes must\nbe updated atomically. Instead of ensuring this through too many memory\nbarriers, we put the arrays and their sizes in a separate structure. This\npatch takes the first step of putting the file table elements in a separate\nstructure fdtable that is embedded withing files_struct. It also changes all\nthe users to refer to the file table using files_fdtable() macro. Subsequent\napplciation of RCU becomes easier after this.\n\nSigned-off-by: Dipankar Sarma \u003cdipankar@in.ibm.com\u003e\nSigned-Off-By: David Howells \u003cdhowells@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "b0d62e6d5b3318b6b722121d945afa295f7201b5", "tree": "9dfed36c97bcdbeaf1576b62752bfcafbfae808a", "parents": [ "f76baf9365bd66216bf0e0ebfc083e22eda6215b" ], "author": { "name": "Jason Baron", "email": "jbaron@redhat.com", "time": "Fri Sep 09 13:02:01 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Sep 09 13:57:31 2005 -0700" }, "message": "[PATCH] fix disassociate_ctty vs. fork race\n\nRace is as follows. Process A forks process B, both being part of the same\nsession. Then, A calls disassociate_ctty while B forks C:\n\nA\t\t\t\tB\n\u003d\u003d\u003d\u003d\t\t\t\t\u003d\u003d\u003d\u003d\n\t\t\t\tfork()\n\t\t\t\t copy_signal()\ndissasociate_ctty()\t\t....\n\t\t\t\t attach_pid(p, PIDTYPE_SID, p-\u003esignal-\u003esession);\n\nNow, C can have current-\u003esignal-\u003etty pointing to a freed tty structure, as\nit hasn\u0027t yet been added to the session group (to have its controlling tty\ncleared on the diassociate_ctty() call).\n\nThis has shown up as an oops but could be even more serious. I haven\u0027t\ntried to create a test case, but a customer has verified that the patch\nbelow resolves the issue, which was occuring quite frequently. I\u0027ll try\nand post the test case if i can.\n\nThe patch simply checks for a NULL tty *after* it has been attached to the\nproper session group and clears it as necessary. Alternatively, we could\nsimply do the tty assignment after the the process is added to the proper\nsession group.\n\nSigned-off-by: Jason Baron \u003cjbaron@redhat.com\u003e\nCc: Roland McGrath \u003croland@redhat.com\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "4b5d37ac02954572e80e09255bb5737277aaee8e", "tree": "6072a602430af9a95d2bcadf149bd41c8045ef67", "parents": [ "1299232b5743da454c73853b90b3d2d83dce1737" ], "author": { "name": "Giancarlo Formicuccia", "email": "giancarlo.formicuccia@gmail.com", "time": "Fri Sep 09 13:01:22 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Sep 09 13:56:43 2005 -0700" }, "message": "[PATCH] Clear task_struct-\u003efs_excl on fork()\n\nAn oversight. We don\u0027t want to carry the IO scheduler\u0027s \"we hold exclusive fs\nresources\" hint over to the child across fork().\n\nAcked-by: Jens Axboe \u003caxboe@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "ed75e8d58010fdc06e2c3a81bfbebae92314c7e3", "tree": "3f6f8dc5a34c9e03f613d4b907e02802ab075a9e", "parents": [ "94c80b2598dbd2b8a6fe5f5c2c3af1beb37f66c7" ], "author": { "name": "Laurent Vivier", "email": "LaurentVivier@wanadoo.fr", "time": "Sat Sep 03 15:57:18 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@evo.osdl.org", "time": "Mon Sep 05 00:06:20 2005 -0700" }, "message": "[PATCH] UML Support - Ptrace: adds the host SYSEMU support, for UML and general usage\n\n Jeff Dike \u003cjdike@addtoit.com\u003e,\n Paolo \u0027Blaisorblade\u0027 Giarrusso \u003cblaisorblade_spam@yahoo.it\u003e,\n Bodo Stroesser \u003cbstroesser@fujitsu-siemens.com\u003e\n\nAdds a new ptrace(2) mode, called PTRACE_SYSEMU, resembling PTRACE_SYSCALL\nexcept that the kernel does not execute the requested syscall; this is useful\nto improve performance for virtual environments, like UML, which want to run\nthe syscall on their own.\n\nIn fact, using PTRACE_SYSCALL means stopping child execution twice, on entry\nand on exit, and each time you also have two context switches; with SYSEMU you\navoid the 2nd stop and so save two context switches per syscall.\n\nAlso, some architectures don\u0027t have support in the host for changing the\nsyscall number via ptrace(), which is currently needed to skip syscall\nexecution (UML turns any syscall into getpid() to avoid it being executed on\nthe host). Fixing that is hard, while SYSEMU is easier to implement.\n\n* This version of the patch includes some suggestions of Jeff Dike to avoid\n adding any instructions to the syscall fast path, plus some other little\n changes, by myself, to make it work even when the syscall is executed with\n SYSENTER (but I\u0027m unsure about them). It has been widely tested for quite a\n lot of time.\n\n* Various fixed were included to handle the various switches between\n various states, i.e. when for instance a syscall entry is traced with one of\n PT_SYSCALL / _SYSEMU / _SINGLESTEP and another one is used on exit.\n Basically, this is done by remembering which one of them was used even after\n the call to ptrace_notify().\n\n* We\u0027re combining TIF_SYSCALL_EMU with TIF_SYSCALL_TRACE or TIF_SINGLESTEP\n to make do_syscall_trace() notice that the current syscall was started with\n SYSEMU on entry, so that no notification ought to be done in the exit path;\n this is a bit of a hack, so this problem is solved in another way in next\n patches.\n\n* Also, the effects of the patch:\n\"Ptrace - i386: fix Syscall Audit interaction with singlestep\"\nare cancelled; they are restored back in the last patch of this series.\n\nDetailed descriptions of the patches doing this kind of processing follow (but\nI\u0027ve already summed everything up).\n\n* Fix behaviour when changing interception kind #1.\n\n In do_syscall_trace(), we check the status of the TIF_SYSCALL_EMU flag\n only after doing the debugger notification; but the debugger might have\n changed the status of this flag because he continued execution with\n PTRACE_SYSCALL, so this is wrong. This patch fixes it by saving the flag\n status before calling ptrace_notify().\n\n* Fix behaviour when changing interception kind #2:\n avoid intercepting syscall on return when using SYSCALL again.\n\n A guest process switching from using PTRACE_SYSEMU to PTRACE_SYSCALL\n crashes.\n\n The problem is in arch/i386/kernel/entry.S. The current SYSEMU patch\n inhibits the syscall-handler to be called, but does not prevent\n do_syscall_trace() to be called after this for syscall completion\n interception.\n\n The appended patch fixes this. It reuses the flag TIF_SYSCALL_EMU to\n remember \"we come from PTRACE_SYSEMU and now are in PTRACE_SYSCALL\", since\n the flag is unused in the depicted situation.\n\n* Fix behaviour when changing interception kind #3:\n avoid intercepting syscall on return when using SINGLESTEP.\n\n When testing 2.6.9 and the skas3.v6 patch, with my latest patch and had\n problems with singlestepping on UML in SKAS with SYSEMU. It looped\n receiving SIGTRAPs without moving forward. EIP of the traced process was\n the same for all SIGTRAPs.\n\nWhat\u0027s missing is to handle switching from PTRACE_SYSCALL_EMU to\nPTRACE_SINGLESTEP in a way very similar to what is done for the change from\nPTRACE_SYSCALL_EMU to PTRACE_SYSCALL_TRACE.\n\nI.e., after calling ptrace(PTRACE_SYSEMU), on the return path, the debugger is\nnotified and then wake ups the process; the syscall is executed (or skipped,\nwhen do_syscall_trace() returns 0, i.e. when using PTRACE_SYSEMU), and\ndo_syscall_trace() is called again. Since we are on the return path of a\nSYSEMU\u0027d syscall, if the wake up is performed through ptrace(PTRACE_SYSCALL),\nwe must still avoid notifying the parent of the syscall exit. Now, this\nbehaviour is extended even to resuming with PTRACE_SINGLESTEP.\n\nSigned-off-by: Paolo \u0027Blaisorblade\u0027 Giarrusso \u003cblaisorblade@yahoo.it\u003e\nCc: Jeff Dike \u003cjdike@addtoit.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "3b6bfcdb116f2cc2cab921fcac6d39d4022952d2", "tree": "b9bdc0f85d16bac6de043883880c203627bab9c5", "parents": [ "70d1d47c47c4643af357cb44d0d891c1b765f2ab" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jul 12 13:58:09 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Tue Jul 12 16:00:58 2005 -0700" }, "message": "[PATCH] lower VM_DONTCOPY total_vm\n\ndup_mmap of a VM_DONTCOPY vma forgot to lower the child\u0027s total_vm. (But\nno way does this account for the recent report of total_vm seen too low.)\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "22e2c507c301c3dbbcf91b4948b88f78842ee6c9", "tree": "9a97c91d1362e69703aa286021daffb8a5456f4c", "parents": [ "020f46a39eb7b99a575b9f4d105fce2b142acdf1" ], "author": { "name": "Jens Axboe", "email": "axboe@suse.de", "time": "Mon Jun 27 10:55:12 2005 +0200" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Mon Jun 27 14:33:29 2005 -0700" }, "message": "[PATCH] Update cfq io scheduler to time sliced design\n\nThis updates the CFQ io scheduler to the new time sliced design (cfq\nv3). It provides full process fairness, while giving excellent\naggregate system throughput even for many competing processes. It\nsupports io priorities, either inherited from the cpu nice value or set\ndirectly with the ioprio_get/set syscalls. The latter closely mimic\nset/getpriority.\n\nThis import is based on my latest from -mm.\n\nSigned-off-by: Jens Axboe \u003caxboe@suse.de\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "476d139c218e44e045e4bc6d4cc02b010b343939", "tree": "82a6537b829b2b35156fba5a312f4e44273a4356", "parents": [ "674311d5b411e9042df4fdf7aef0b3c8217b6240" ], "author": { "name": "Nick Piggin", "email": "nickpiggin@yahoo.com.au", "time": "Sat Jun 25 14:57:29 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Jun 25 16:24:44 2005 -0700" }, "message": "[PATCH] sched: consolidate sbe sbf\n\nConsolidate balance-on-exec with balance-on-fork. This is made easy by the\nsched-domains RCU patches.\n\nAs well as the general goodness of code reduction, this allows the runqueues\nto be unlocked during balance-on-fork.\n\nschedstats is a problem. Maybe just have balance-on-event instead of\ndistinguishing fork and exec?\n\nSigned-off-by: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nAcked-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "45918e1a8bfcabc1cb4570b8df276655020eac45", "tree": "4caf6a8b634de389e758d2b4c037208f098b3141", "parents": [ "91612e0df20a52f61db3cac280c153311b36df7a" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jun 21 17:15:08 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Tue Jun 21 18:46:19 2005 -0700" }, "message": "[PATCH] dup_mmap: update comment on new vma\n\nRemove part of comment on linking new vma in dup_mmap: since anon_vma rmap\ncame in, try_to_unmap_one knows the vma without needing find_vma. But add\na comment to note that here vma is inserted without mmap_sem.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "1363c3cd8603a913a27e2995dccbd70d5312d8e6", "tree": "405e7fc1ef44678f3ca0a54c536d0457e6e80f45", "parents": [ "e7c8d5c9955a4d2e88e36b640563f5d6d5aba48a" ], "author": { "name": "Wolfgang Wander", "email": "wwc@rentec.com", "time": "Tue Jun 21 17:14:49 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Tue Jun 21 18:46:16 2005 -0700" }, "message": "[PATCH] Avoiding mmap fragmentation\n\nIngo recently introduced a great speedup for allocating new mmaps using the\nfree_area_cache pointer which boosts the specweb SSL benchmark by 4-5% and\ncauses huge performance increases in thread creation.\n\nThe downside of this patch is that it does lead to fragmentation in the\nmmap-ed areas (visible via /proc/self/maps), such that some applications\nthat work fine under 2.4 kernels quickly run out of memory on any 2.6\nkernel.\n\nThe problem is twofold:\n\n 1) the free_area_cache is used to continue a search for memory where\n the last search ended. Before the change new areas were always\n searched from the base address on.\n\n So now new small areas are cluttering holes of all sizes\n throughout the whole mmap-able region whereas before small holes\n tended to close holes near the base leaving holes far from the base\n large and available for larger requests.\n\n 2) the free_area_cache also is set to the location of the last\n munmap-ed area so in scenarios where we allocate e.g. five regions of\n 1K each, then free regions 4 2 3 in this order the next request for 1K\n will be placed in the position of the old region 3, whereas before we\n appended it to the still active region 1, placing it at the location\n of the old region 2. Before we had 1 free region of 2K, now we only\n get two free regions of 1K -\u003e fragmentation.\n\nThe patch addresses thes issues by introducing yet another cache descriptor\ncached_hole_size that contains the largest known hole size below the\ncurrent free_area_cache. If a new request comes in the size is compared\nagainst the cached_hole_size and if the request can be filled with a hole\nbelow free_area_cache the search is started from the base instead.\n\nThe results look promising: Whereas 2.6.12-rc4 fragments quickly and my\n(earlier posted) leakme.c test program terminates after 50000+ iterations\nwith 96 distinct and fragmented maps in /proc/self/maps it performs nicely\n(as expected) with thread creation, Ingo\u0027s test_str02 with 20000 threads\nrequires 0.7s system time.\n\nTaking out Ingo\u0027s patch (un-patch available per request) by basically\ndeleting all mentions of free_area_cache from the kernel and starting the\nsearch for new memory always at the respective bases we observe: leakme\nterminates successfully with 11 distinctive hardly fragmented areas in\n/proc/self/maps but thread creating is gringdingly slow: 30+s(!) system\ntime for Ingo\u0027s test_str02 with 20000 threads.\n\nNow - drumroll ;-) the appended patch works fine with leakme: it ends with\nonly 7 distinct areas in /proc/self/maps and also thread creation seems\nsufficiently fast with 0.71s for 20000 threads.\n\nSigned-off-by: Wolfgang Wander \u003cwwc@rentec.com\u003e\nCredit-to: \"Richard Purdie\" \u003crpurdie@rpsys.net\u003e\nSigned-off-by: Ken Chen \u003ckenneth.w.chen@intel.com\u003e\nAcked-by: Ingo Molnar \u003cmingo@elte.hu\u003e (partly)\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "1da177e4c3f41524e886b7f1b8a0c1fc7321cac2", "tree": "0bba044c4ce775e45a88a51686b5d9f90697ea9d", "parents": [], "author": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "message": "Linux-2.6.12-rc2\n\nInitial git repository build. I\u0027m not bothering with the full history,\neven though we have it. We can create a separate \"historical\" git\narchive of that later if we want to, and in the meantime it\u0027s about\n3.2GB when imported into git - space that would just make the early\ngit days unnecessarily complicated, when we don\u0027t have a lot of good\ninfrastructure for it.\n\nLet it rip!\n" } ] }