)]}' { "log": [ { "commit": "6d08a2567c0b9103c3ff946df17ad4be9a917e2f", "tree": "6228064bb45946616dd2f0a6f6b2f9e88e9d8e52", "parents": [ "187841a800f328e93529086ca58145e7a44bff3f" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Apr 07 15:39:18 2014 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Apr 07 16:36:11 2014 -0700" }, "message": "ipc: use device_initcall\n\n... since __initcall is now deprecated.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "187841a800f328e93529086ca58145e7a44bff3f", "tree": "8d5111595735e5b3cd563c54fa27f9714d72d632", "parents": [ "6aa7a29aa8378fa868fabef6a84b0a40d5d9d677" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Apr 07 15:39:17 2014 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Apr 07 16:36:11 2014 -0700" }, "message": "ipc/compat.c: remove sc_semopm macro\n\nThis macro appears to have been introduced back in the 2.5 era for\nsemtimedop32 backward compatibility on ia32:\n\n https://lkml.org/lkml/2003/4/28/78\n\nNowadays, this syscall in compat just defaults back to the code found in\nsem.c, so it is no longer used and can thus be removed:\n\nlong compat_sys_semtimedop(int semid, struct sembuf __user *tsems,\n\t\tunsigned nsops, const struct compat_timespec __user *timeout)\n{\n\tstruct timespec __user *ts64;\n\tif (compat_convert_timespec(\u0026ts64, timeout))\n\t\treturn -EFAULT;\n\treturn sys_semtimedop(semid, tsems, nsops, ts64);\n}\n\nFurthermore, there are no users in compat.c. After this change, kernel\nbuilds just fine with both CONFIG_SYSVIPC_COMPAT and CONFIG_SYSVIPC.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7125764c5d1a5c72d522f1011b6cc8b8100b48fe", "tree": "678f3355ac872a4379b28dbe36f5beddd0e284d8", "parents": [ "c6f21243ce1e8d81ad8361da4d2eaa5947b667c4", "dce44e03b0a3448ad11ac6c6e0cbe299e0400791" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 02 12:51:41 2014 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 02 12:51:41 2014 -0700" }, "message": "Merge branch \u0027x86-x32-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull compat time conversion changes from Peter Anvin:\n \"Despite the branch name this is really neither an x86 nor an\n x32-specific patchset, although it the implementation of the\n discussions that followed the x32 security hole a few months ago.\n\n This removes get/put_compat_timespec/val() and replaces them with\n compat_get/put_timespec/val() which are savvy as to the current status\n of COMPAT_USE_64BIT_TIME.\n\n It removes several unused and/or incorrect/misleading functions (like\n compat_put_timeval_convert which doesn\u0027t in fact do any conversion)\n and also replaces several open-coded implementations what is now\n called compat_convert_timespec() with that function\"\n\n* \u0027x86-x32-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n compat: Fix sparse address space warnings\n compat: Get rid of (get|put)_compat_time(val|spec)\n" }, { "commit": "190f918660a69d1c56fd05dc8c6cbb8336a8a0af", "tree": "2320b4aab7f048d97c35ac6ade5b7dedc71da89d", "parents": [ "176ab02d4916f09d5d8cb63372d142df4378cdea", "1e4ec6217dcf4b26cf959b70298a3b990479c955" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Mar 31 14:32:17 2014 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Mar 31 14:32:17 2014 -0700" }, "message": "Merge branch \u0027compat\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux\n\nPull s390 compat wrapper rework from Heiko Carstens:\n \"S390 compat system call wrapper simplification work.\n\n The intention of this work is to get rid of all hand written assembly\n compat system call wrappers on s390, which perform proper sign or zero\n extension, or pointer conversion of compat system call parameters.\n Instead all of this should be done with C code eg by using Al\u0027s\n COMPAT_SYSCALL_DEFINEx() macro.\n\n Therefore all common code and s390 specific compat system calls have\n been converted to the COMPAT_SYSCALL_DEFINEx() macro.\n\n In order to generate correct code all compat system calls may only\n have eg compat_ulong_t parameters, but no unsigned long parameters.\n Those patches which change parameter types from unsigned long to\n compat_ulong_t parameters are separate in this series, but shouldn\u0027t\n cause any harm.\n\n The only compat system calls which intentionally have 64 bit\n parameters (preadv64 and pwritev64) in support of the x86/32 ABI\n haven\u0027t been changed, but are now only available if an architecture\n defines __ARCH_WANT_COMPAT_SYS_PREADV64/PWRITEV64.\n\n System calls which do not have a compat variant but still need proper\n zero extension on s390, like eg \"long sys_brk(unsigned long brk)\" will\n get a proper wrapper function with the new s390 specific\n COMPAT_SYSCALL_WRAPx() macro:\n\n COMPAT_SYSCALL_WRAP1(brk, unsigned long, brk);\n\n which generates the following code (simplified):\n\n asmlinkage long sys_brk(unsigned long brk);\n asmlinkage long compat_sys_brk(long brk)\n {\n return sys_brk((u32)brk);\n }\n\n Given that the C file which contains all the COMPAT_SYSCALL_WRAP lines\n includes both linux/syscall.h and linux/compat.h, it will generate\n build errors, if the declaration of sys_brk() doesn\u0027t match, or if\n there exists a non-matching compat_sys_brk() declaration.\n\n In addition this will intentionally result in a link error if\n somewhere else a compat_sys_brk() function exists, which probably\n should have been used instead. Two more BUILD_BUG_ONs make sure the\n size and type of each compat syscall parameter can be handled\n correctly with the s390 specific macros.\n\n I converted the compat system calls step by step to verify the\n generated code is correct and matches the previous code. In fact it\n did not always match, however that was always a bug in the hand\n written asm code.\n\n In result we get less code, less bugs, and much more sanity checking\"\n\n* \u0027compat\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (44 commits)\n s390/compat: add copyright statement\n compat: include linux/unistd.h within linux/compat.h\n s390/compat: get rid of compat wrapper assembly code\n s390/compat: build error for large compat syscall args\n mm/compat: convert to COMPAT_SYSCALL_DEFINE with changing parameter types\n kexec/compat: convert to COMPAT_SYSCALL_DEFINE with changing parameter types\n net/compat: convert to COMPAT_SYSCALL_DEFINE with changing parameter types\n ipc/compat: convert to COMPAT_SYSCALL_DEFINE with changing parameter types\n fs/compat: convert to COMPAT_SYSCALL_DEFINE with changing parameter types\n ipc/compat: convert to COMPAT_SYSCALL_DEFINE\n fs/compat: convert to COMPAT_SYSCALL_DEFINE\n security/compat: convert to COMPAT_SYSCALL_DEFINE\n mm/compat: convert to COMPAT_SYSCALL_DEFINE\n net/compat: convert to COMPAT_SYSCALL_DEFINE\n kernel/compat: convert to COMPAT_SYSCALL_DEFINE\n fs/compat: optional preadv64/pwrite64 compat system calls\n ipc/compat_sys_msgrcv: change msgtyp type from long to compat_long_t\n s390/compat: partial parameter conversion within syscall wrappers\n s390/compat: automatic zero, sign and pointer conversion of syscalls\n s390/compat: add sync_file_range and fallocate compat syscalls\n ...\n" }, { "commit": "4f87dac386cc43d5525da7a939d4b4e7edbea22c", "tree": "80f2b7381fe213302d1903b32f9617d1de3501ce", "parents": [ "3b4df68d0697129c172e66fd925cf32192044f9a" ], "author": { "name": "Michael Kerrisk", "email": "mtk.manpages@gmail.com", "time": "Mon Mar 10 14:46:07 2014 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Mar 16 10:41:04 2014 -0700" }, "message": "ipc: Fix 2 bugs in msgrcv() MSG_COPY implementation\n\nWhile testing and documenting the msgrcv() MSG_COPY flag that Stanislav\nKinsbursky added in commit 4a674f34ba04 (\"ipc: introduce message queue\ncopy feature\" \u003d\u003e kernel 3.8), I discovered a couple of bugs in the\nimplementation. The two bugs concern MSG_COPY interactions with other\nmsgrcv() flags, namely:\n\n (A) MSG_COPY + MSG_EXCEPT\n (B) MSG_COPY + !IPC_NOWAIT\n\nThe bugs are distinct (and the fix for the first one is obvious),\nhowever my fix for both is a single-line patch, which is why I\u0027m\ncombining them in a single mail, rather than writing two mails+patches.\n\n \u003d\u003d\u003d\u003d\u003d (A) MSG_COPY + MSG_EXCEPT \u003d\u003d\u003d\u003d\u003d\n\nWith the addition of the MSG_COPY flag, there are now two msgrcv()\nflags--MSG_COPY and MSG_EXCEPT--that modify the meaning of the \u0027msgtyp\u0027\nargument in unrelated ways. Specifying both in the same call is a\nlogical error that is currently permitted, with the effect that MSG_COPY\nhas priority and MSG_EXCEPT is ignored. The call should give an error\nif both flags are specified. The patch below implements that behavior.\n\n \u003d\u003d\u003d\u003d\u003d (B) (B) MSG_COPY + !IPC_NOWAIT \u003d\u003d\u003d\u003d\u003d\n\nThe test code that was submitted in commit 3a665531a3b7 (\"selftests: IPC\nmessage queue copy feature test\") shows MSG_COPY being used in\nconjunction with IPC_NOWAIT. In other words, if there is no message at\nthe position \u0027msgtyp\u0027. return immediately with the error in ENOMSG.\n\nWhat was not (fully) tested is the behavior if MSG_COPY is specified\n*without* IPC_NOWAIT, and there is an odd behavior. If the queue\ncontains less than \u0027msgtyp\u0027 messages, then the call blocks until the\nnext message is written to the queue. At that point, the msgrcv() call\nreturns a copy of the newly added message, regardless of whether that\nmessage is at the ordinal position \u0027msgtyp\u0027. This is clearly bogus, and\nproblematic for applications that might want to make use of the MSG_COPY\nflag.\n\nI considered the following possible solutions to this problem:\n\n (1) Force the call to block until a message *does* appear at the\n position \u0027msgtyp\u0027.\n\n (2) If the MSG_COPY flag is specified, the kernel should implicitly add\n IPC_NOWAIT, so that the call fails with ENOMSG for this case.\n\n (3) If the MSG_COPY flag is specified, but IPC_NOWAIT is not, generate\n an error (probably, EINVAL is the right one).\n\nI do not know if any application would really want to have the\nfunctionality of solution (1), especially since an application can\ndetermine in advance the number of messages in the queue using msgctl()\nIPC_STAT. Obviously, this solution would be the most work to implement.\n\nSolution (2) would have the effect of silently fixing any applications\nthat tried to employ broken behavior. However, it would mean that if we\nlater decided to implement solution (1), then user-space could not\neasily detect what the kernel supports (but, since I\u0027m somewhat doubtful\nthat solution (1) is needed, I\u0027m not sure that this is much of a\nproblem).\n\nSolution (3) would have the effect of informing broken applications that\nthey are doing something broken. The downside is that this would cause\na ABI breakage for any applications that are currently employing the\nbroken behavior. However:\n\na) Those applications are almost certainly not getting the results they\n expect.\nb) Possibly, those applications don\u0027t even exist, because MSG_COPY is\n currently hidden behind CONFIG_CHECKPOINT_RESTORE.\n\nThe upside of solution (3) is that if we later decided to implement\nsolution (1), user-space could determine what the kernel supports, via\nthe error return.\n\nIn my view, solution (3) is mildly preferable to solution (2), and\nsolution (1) could still be done later if anyone really cares. The\npatch below implements solution (3).\n\nPS. For anyone out there still listening, it\u0027s the usual story:\ndocumenting an API (and the thinking about, and the testing of the API,\nthat documentation entails) is the one of the single best ways of\nfinding bugs in the API, as I\u0027ve learned from a lot of experience. Best\nto do that documentation before releasing the API.\n\nSigned-off-by: Michael Kerrisk \u003cmtk.manpages@gmail.com\u003e\nAcked-by: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nCc: stable@vger.kernel.org\nCc: Serge Hallyn \u003cserge.hallyn@canonical.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Pavel Emelyanov \u003cxemul@parallels.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8eee9093cdbeb2aa89d67dc1a3fd118acabaea52", "tree": "ad181d1ee2fab9db22b51676cb259083fe890423", "parents": [ "932602e238329da99f8482c1b721549531fbfe7f" ], "author": { "name": "Heiko Carstens", "email": "heiko.carstens@de.ibm.com", "time": "Tue Mar 04 16:19:16 2014 +0100" }, "committer": { "name": "Heiko Carstens", "email": "heiko.carstens@de.ibm.com", "time": "Thu Mar 06 16:30:45 2014 +0100" }, "message": "ipc/compat: convert to COMPAT_SYSCALL_DEFINE with changing parameter types\n\nIn order to allow the COMPAT_SYSCALL_DEFINE macro generate code that\nperforms proper zero and sign extension convert all 64 bit parameters\nto their corresponding 32 bit compat counterparts.\n\nSigned-off-by: Heiko Carstens \u003cheiko.carstens@de.ibm.com\u003e\n" }, { "commit": "5d70a59637911e84687b421afeb4c111a579fb2b", "tree": "645055f75e5a47120128f17430f82da40fa00eff", "parents": [ "625b1d7e812d55df8d42253a134002c006de7468" ], "author": { "name": "Heiko Carstens", "email": "heiko.carstens@de.ibm.com", "time": "Tue Mar 04 11:17:50 2014 +0100" }, "committer": { "name": "Heiko Carstens", "email": "heiko.carstens@de.ibm.com", "time": "Thu Mar 06 16:30:44 2014 +0100" }, "message": "ipc/compat: convert to COMPAT_SYSCALL_DEFINE\n\nConvert all compat system call functions where all parameter types\nhave a size of four or less than four bytes, or are pointer types\nto COMPAT_SYSCALL_DEFINE.\nThe implicit casts within COMPAT_SYSCALL_DEFINE will perform proper\nzero and sign extension to 64 bit of all parameters if needed.\n\nSigned-off-by: Heiko Carstens \u003cheiko.carstens@de.ibm.com\u003e\n" }, { "commit": "291fdb0bcebd5e8db6af767c1fdc522167dad73d", "tree": "019dbcd82cdc4a2e9d6c2426f0c52b8c2eb945e0", "parents": [ "c2e7c3d0ef0d1a3765792a35ae9e6f91a3025214" ], "author": { "name": "Heiko Carstens", "email": "heiko.carstens@de.ibm.com", "time": "Tue Mar 04 12:39:03 2014 +0100" }, "committer": { "name": "Heiko Carstens", "email": "heiko.carstens@de.ibm.com", "time": "Thu Mar 06 15:35:09 2014 +0100" }, "message": "ipc/compat_sys_msgrcv: change msgtyp type from long to compat_long_t\n\nChange the type of compat_sys_msgrcv\u0027s msgtyp parameter from long\nto compat_long_t, since compat user space passes only a 32 bit signed\nvalue.\nLet the compat wrapper do proper sign extension to 64 bit of this\nparameter.\n\nSigned-off-by: Heiko Carstens \u003cheiko.carstens@de.ibm.com\u003e\n" }, { "commit": "f3713fd9cff733d9df83116422d8e4af6e86b2bb", "tree": "58a1a64720424b7a4b32bc6d26ca23000c2ce414", "parents": [ "08088cb9ac0a9c28d4cf3efa4f6848a9b053ccfd" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Tue Feb 25 15:01:45 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Feb 25 15:25:45 2014 -0800" }, "message": "ipc,mqueue: remove limits for the amount of system-wide queues\n\nCommit 93e6f119c0ce (\"ipc/mqueue: cleanup definition names and\nlocations\") added global hardcoded limits to the amount of message\nqueues that can be created. While these limits are per-namespace,\nreality is that it ends up breaking userspace applications.\nHistorically users have, at least in theory, been able to create up to\nINT_MAX queues, and limiting it to just 1024 is way too low and dramatic\nfor some workloads and use cases. For instance, Madars reports:\n\n \"This update imposes bad limits on our multi-process application. As\n our app uses approaches that each process opens its own set of queues\n (usually something about 3-5 queues per process). In some scenarios\n we might run up to 3000 processes or more (which of-course for linux\n is not a problem). Thus we might need up to 9000 queues or more. All\n processes run under one user.\"\n\nOther affected users can be found in launchpad bug #1155695:\n https://bugs.launchpad.net/ubuntu/+source/manpages/+bug/1155695\n\nInstead of increasing this limit, revert it entirely and fallback to the\noriginal way of dealing queue limits -- where once a user\u0027s resource\nlimit is reached, and all memory is used, new queues cannot be created.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nReported-by: Madars Vitolins \u003cm@silodev.com\u003e\nAcked-by: Doug Ledford \u003cdledford@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e\t[3.5+]\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "81993e81a994504f4c8b97d3410c9a052cdbcc9d", "tree": "7471cde617f11dad2f29187daa764e1d119a9731", "parents": [ "5cb480f6b488128140c940abff3c36f524a334a8" ], "author": { "name": "H. Peter Anvin", "email": "hpa@linux.intel.com", "time": "Sat Feb 01 18:54:11 2014 -0800" }, "committer": { "name": "H. Peter Anvin", "email": "hpa@linux.intel.com", "time": "Sun Feb 02 14:09:12 2014 -0800" }, "message": "compat: Get rid of (get|put)_compat_time(val|spec)\n\nWe have two APIs for compatiblity timespec/val, with confusingly\nsimilar names. compat_(get|put)_time(val|spec) *do* handle the case\nwhere COMPAT_USE_64BIT_TIME is set, whereas\n(get|put)_compat_time(val|spec) do not. This is an accident waiting\nto happen.\n\nClean it up by favoring the full-service version; the limited version\nis replaced with double-underscore versions static to kernel/compat.c.\n\nA common pattern is to convert a struct timespec to kernel format in\nan allocation on the user stack. Unfortunately it is open-coded in\nseveral places. Since this allocation isn\u0027t actually needed if\nCOMPAT_USE_64BIT_TIME is true (since user format \u003d\u003d kernel format)\nencapsulate that whole pattern into the function\ncompat_convert_timespec(). An equivalent function should be written\nfor struct timeval if it is needed in the future.\n\nFinally, get rid of compat_(get|put)_timeval_convert(): each was only\nused once, and the latter was not even doing what the function said\n(no conversion actually was being done.) Moving the conversion into\ncompat_sys_settimeofday() itself makes the code much more similar to\nsys_settimeofday() itself.\n\nv3: Remove unused compat_convert_timeval().\n\nv2: Drop bogus \"const\" in the destination argument for\n compat_convert_time*().\n\nCc: Mauro Carvalho Chehab \u003cm.chehab@samsung.com\u003e\nCc: Alexander Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: Hans Verkuil \u003chans.verkuil@cisco.com\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nCc: Heiko Carstens \u003cheiko.carstens@de.ibm.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Mateusz Guzik \u003cmguzik@redhat.com\u003e\nCc: Rafael Aquini \u003caquini@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Stephen Rothwell \u003csfr@canb.auug.org.au\u003e\nCc: Dan Carpenter \u003cdan.carpenter@oracle.com\u003e\nCc: Arnd Bergmann \u003carnd@arndb.de\u003e\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Ingo Molnar \u003cmingo@kernel.org\u003e\nCc: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nCc: Catalin Marinas \u003ccatalin.marinas@arm.com\u003e\nCc: Will Deacon \u003cwill.deacon@arm.com\u003e\nTested-by: H.J. Lu \u003chjl.tools@gmail.com\u003e\nSigned-off-by: H. Peter Anvin \u003chpa@linux.intel.com\u003e\n" }, { "commit": "e7ca2552369c1dfe0216c626baf82c3d83ec36bb", "tree": "dfaa96f46b9c9cf907c41bc6c825d250a8595010", "parents": [ "ffa571dafbaec0c74e374ce0ea7b4212b6cbc94c" ], "author": { "name": "Mateusz Guzik", "email": "mguzik@redhat.com", "time": "Mon Jan 27 17:07:11 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:40 2014 -0800" }, "message": "ipc: fix compat msgrcv with negative msgtyp\n\nCompat function takes msgtyp argument as u32 and passes it down to\ndo_msgrcv which results in casting to long, thus the sign is lost and we\nget a big positive number instead.\n\nCast the argument to signed type before passing it down.\n\nSigned-off-by: Mateusz Guzik \u003cmguzik@redhat.com\u003e\nReported-by: Gabriellla Schmidt \u003cgsc@bruker.de\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ffa571dafbaec0c74e374ce0ea7b4212b6cbc94c", "tree": "21f19b4f4ddcf219c7b024c2cf7ffbac9d0a80b8", "parents": [ "daf948c7d1a080041ae19aca07625efec670695a" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Jan 27 17:07:10 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:40 2014 -0800" }, "message": "ipc,msg: document barriers\n\nBoth expunge_all() and pipeline_send() rely on both a nil msg value and\na full barrier to guarantee the correct ordering when waking up a task.\n\nWhile its counterpart at the receiving end is well documented for the\nlockless recv algorithm, we still need to document these specific\nsmp_mb() calls.\n\n[akpm@linux-foundation.org: fix typo, per Mike]\n[akpm@linux-foundation.org: mroe tpyos]\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Aswin Chandramouleeswaran \u003caswin@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "daf948c7d1a080041ae19aca07625efec670695a", "tree": "af17f8452bcd1e6c709f4875c4a04d95d4b6b62f", "parents": [ "8dc5cd04f97b5d6cad64df1e7dc5c49110b4d5e3" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Jan 27 17:07:09 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:40 2014 -0800" }, "message": "ipc: delete seq_max field in struct ipc_ids\n\nThis field is only used to reset the ids seq number if it exceeds the\nsmaller of INT_MAX/SEQ_MULTIPLIER and USHRT_MAX, and can therefore be\nmoved out of the structure and into its own macro. Since each\nipc_namespace contains a table of 3 pointers to struct ipc_ids we can\nsave space in instruction text:\n\n text data bss dec hex filename\n 56232 2348 24 58604 e4ec ipc/built-in.o\n 56216 2348 24 58588 e4dc ipc/built-in.o-after\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nReviewed-by: Jonathan Gonzalez \u003cjgonzalez@linets.cl\u003e\nCc: Aswin Chandramouleeswaran \u003caswin@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8dc5cd04f97b5d6cad64df1e7dc5c49110b4d5e3", "tree": "d2576a46ef38f8f3dbb8fdef98e922812f3d2837", "parents": [ "95d4eb2822d5d67fb14197d90d3e4d4146671cfb" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Jan 27 17:07:08 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:40 2014 -0800" }, "message": "ipc: simplify sysvipc_proc_open() return\n\nGet rid of silly/useless label jumping.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Aswin Chandramouleeswaran \u003caswin@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "95d4eb2822d5d67fb14197d90d3e4d4146671cfb", "tree": "5fd76fa75d3af39e90bfa3bdda5a91731888e0b9", "parents": [ "3ab08fe20475658bab65118d599d03cd8ca44dd1" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Jan 27 17:07:07 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:40 2014 -0800" }, "message": "ipc: remove useless return statement\n\nOnly found in ipc_rmid().\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Aswin Chandramouleeswaran \u003caswin@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3ab08fe20475658bab65118d599d03cd8ca44dd1", "tree": "25f4f43e58aad6a4c286a098f1f1f9d69452c30d", "parents": [ "8001c85810dd2277d75ae60376e840456afa9b7e" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Jan 27 17:07:06 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:39 2014 -0800" }, "message": "ipc: remove braces for single statements\n\nDeal with checkpatch messages:\n WARNING: braces {} are not necessary for single statement blocks\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Aswin Chandramouleeswaran \u003caswin@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8001c85810dd2277d75ae60376e840456afa9b7e", "tree": "60a03cdecbbbb6e2bdca6b132b4b332f90a728ad", "parents": [ "239521f31d7496a5322ee664ed8bbd1027b98c4b" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Jan 27 17:07:05 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:39 2014 -0800" }, "message": "ipc: standardize code comments\n\nIPC commenting style is all over the place, *specially* in util.c. This\npatch orders things a bit.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Aswin Chandramouleeswaran \u003caswin@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "239521f31d7496a5322ee664ed8bbd1027b98c4b", "tree": "b340031d42d28ebe22edf55a6cf3a6ea0e2c0d99", "parents": [ "72a8ff2f9245128c254387c58f948f1f0152ea46" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Jan 27 17:07:04 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:39 2014 -0800" }, "message": "ipc: whitespace cleanup\n\nThe ipc code does not adhere the typical linux coding style.\nThis patch fixes lots of simple whitespace errors.\n\n- mostly autogenerated by\n scripts/checkpatch.pl -f --fix \\\n\t--types\u003dpointer_location,spacing,space_before_tab\n- one manual fixup (keep structure members tab-aligned)\n- removal of additional space_before_tab that were not found by --fix\n\nTested with some of my msg and sem test apps.\n\nAndrew: Could you include it in -mm and move it towards Linus\u0027 tree?\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSuggested-by: Li Bin \u003chuawei.libin@huawei.com\u003e\nCc: Joe Perches \u003cjoe@perches.com\u003e\nAcked-by: Rafael Aquini \u003caquini@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "72a8ff2f9245128c254387c58f948f1f0152ea46", "tree": "176eec6a1d264e9e679d7bd73eb3f5b77975e4a3", "parents": [ "0f3d2b0135f4bdbfe47a99753923a64efd373d11" ], "author": { "name": "Rafael Aquini", "email": "aquini@redhat.com", "time": "Mon Jan 27 17:07:02 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:39 2014 -0800" }, "message": "ipc: change kern_ipc_perm.deleted type to bool\n\nstruct kern_ipc_perm.deleted is meant to be used as a boolean toggle, and\nthe changes introduced by this patch are just to make the case explicit.\n\nSigned-off-by: Rafael Aquini \u003caquini@redhat.com\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Greg Thelen \u003cgthelen@google.com\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "0f3d2b0135f4bdbfe47a99753923a64efd373d11", "tree": "4b211b7502e2412aedd94400bdcb012a2b5814b3", "parents": [ "78f5009cc35eb5e52d276a046d90ee2f41b60f8c" ], "author": { "name": "Rafael Aquini", "email": "aquini@redhat.com", "time": "Mon Jan 27 17:07:01 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:39 2014 -0800" }, "message": "ipc: introduce ipc_valid_object() helper to sort out IPC_RMID races\n\nAfter the locking semantics for the SysV IPC API got improved, a couple\nof IPC_RMID race windows were opened because we ended up dropping the\n\u0027kern_ipc_perm.deleted\u0027 check performed way down in ipc_lock(). The\nspotted races got sorted out by re-introducing the old test within the\nracy critical sections.\n\nThis patch introduces ipc_valid_object() to consolidate the way we cope\nwith IPC_RMID races by using the same abstraction across the API\nimplementation.\n\nSigned-off-by: Rafael Aquini \u003caquini@redhat.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Greg Thelen \u003cgthelen@google.com\u003e\nReviewed-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "78f5009cc35eb5e52d276a046d90ee2f41b60f8c", "tree": "ea7a64dd1bf5b00e5c6c1bb4c709ac51aa3f7675", "parents": [ "729abd2ba7007cc2be9e77718ba52d0866d3f60f" ], "author": { "name": "Petr Mladek", "email": "pmladek@suse.cz", "time": "Mon Jan 27 17:07:00 2014 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 27 21:02:39 2014 -0800" }, "message": "ipc/sem.c: avoid overflow of semop undo (semadj) value\n\nWhen trying to understand semop code, I found a small mistake in the check\nfor semadj (undo) value overflow. The new undo value is not stored\nimmediately and next potential checks are done against the old value.\n\nThe failing scenario is not much practical. One semop call has to do more\noperations on the same semaphore. Also semval and semadj must have\ndifferent values, so there has to be some operations without SEM_UNDO\nflag. For example:\n\n\tstruct sembuf depositor_op[1];\n\tstruct sembuf collector_op[2];\n\n\tdepositor_op[0].sem_num \u003d 0;\n\tdepositor_op[0].sem_op \u003d 20000;\n\tdepositor_op[0].sem_flg \u003d 0;\n\n\tcollector_op[0].sem_num \u003d 0;\n\tcollector_op[0].sem_op \u003d -10000;\n\tcollector_op[0].sem_flg \u003d SEM_UNDO;\n\tcollector_op[1].sem_num \u003d 0;\n\tcollector_op[1].sem_op \u003d -10000;\n\tcollector_op[1].sem_flg \u003d SEM_UNDO;\n\n\tif (semop(semid, depositor_op, 1) \u003d\u003d -1)\n\t\t{ perror(\"Failed to do 1st deposit\"); return 1; }\n\n\tif (semop(semid, collector_op, 2) \u003d\u003d -1)\n\t\t{ perror(\"Failed to do 1st collect\"); return 1; }\n\n\tif (semop(semid, depositor_op, 1) \u003d\u003d -1)\n\t\t{ perror(\"Failed to do 2nd deposit\"); return 1; }\n\n\tif (semop(semid, collector_op, 2) \u003d\u003d -1)\n\t\t{ perror(\"Failed to do 2nd collect\"); return 1; }\n\n\treturn 0;\n\nIt passes without error now but the semadj value has overflown in the 2nd\ncollector operation.\n\n[akpm@linux-foundation.org: restore lessened scope of local `undo\u0027]\n[davidlohr@hp.com: correct header comment for perform_atomic_semop]\nSigned-off-by: Petr Mladek \u003cpmladek@suse.cz\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Jiri Kosina \u003cjkosina@suse.cz\u003e\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3a72660b07d86d60457ca32080b1ce8c2b628ee2", "tree": "edc6f42a551b165943c4590389482c70d6f6639f", "parents": [ "b7a9f420ed737cb7cd4075ba06ac1a6f0da9f878" ], "author": { "name": "Jesper Nilsson", "email": "jesper.nilsson@axis.com", "time": "Thu Nov 21 14:32:08 2013 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Nov 21 16:42:28 2013 -0800" }, "message": "ipc,shm: correct error return value in shmctl (SHM_UNLOCK)\n\nCommit 2caacaa82a51 (\"ipc,shm: shorten critical region for shmctl\")\nrestructured the ipc shm to shorten critical region, but introduced a\npath where the return value could be -EPERM, even if the operation\nactually was performed.\n\nBefore the commit, the err return value was reset by the return value\nfrom security_shm_shmctl() after the if (!ns_capable(...)) statement.\n\nNow, we still exit the if statement with err set to -EPERM, and in the\ncase of SHM_UNLOCK, it is not reset at all, and used as the return value\nfrom shmctl.\n\nTo fix this, we only set err when errors occur, leaving the fallthrough\ncase alone.\n\nSigned-off-by: Jesper Nilsson \u003cjesper.nilsson@axis.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: \u003cstable@vger.kernel.org\u003e\t[3.12.x]\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a399b29dfbaaaf91162b2dc5a5875dd51bbfa2a1", "tree": "dfff401b36b593b284cd21e4051ad7a23a944845", "parents": [ "30b0a105d9f7141e4cbf72ae5511832457d89788" ], "author": { "name": "Greg Thelen", "email": "gthelen@google.com", "time": "Thu Nov 21 14:32:00 2013 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Nov 21 16:42:27 2013 -0800" }, "message": "ipc,shm: fix shm_file deletion races\n\nWhen IPC_RMID races with other shm operations there\u0027s potential for\nuse-after-free of the shm object\u0027s associated file (shm_file).\n\nHere\u0027s the race before this patch:\n\n TASK 1 TASK 2\n ------ ------\n shm_rmid()\n ipc_lock_object()\n shmctl()\n shp \u003d shm_obtain_object_check()\n\n shm_destroy()\n shum_unlock()\n fput(shp-\u003eshm_file)\n ipc_lock_object()\n shmem_lock(shp-\u003eshm_file)\n \u003cOOPS\u003e\n\nThe oops is caused because shm_destroy() calls fput() after dropping the\nipc_lock. fput() clears the file\u0027s f_inode, f_path.dentry, and\nf_path.mnt, which causes various NULL pointer references in task 2. I\nreliably see the oops in task 2 if with shmlock, shmu\n\nThis patch fixes the races by:\n1) set shm_file\u003dNULL in shm_destroy() while holding ipc_object_lock().\n2) modify at risk operations to check shm_file while holding\n ipc_object_lock().\n\nExample workloads, which each trigger oops...\n\nWorkload 1:\n while true; do\n id\u003d$(shmget 1 4096)\n shm_rmid $id \u0026\n shmlock $id \u0026\n wait\n done\n\n The oops stack shows accessing NULL f_inode due to racing fput:\n _raw_spin_lock\n shmem_lock\n SyS_shmctl\n\nWorkload 2:\n while true; do\n id\u003d$(shmget 1 4096)\n shmat $id 4096 \u0026\n shm_rmid $id \u0026\n wait\n done\n\n The oops stack is similar to workload 1 due to NULL f_inode:\n touch_atime\n shmem_mmap\n shm_mmap\n mmap_region\n do_mmap_pgoff\n do_shmat\n SyS_shmat\n\nWorkload 3:\n while true; do\n id\u003d$(shmget 1 4096)\n shmlock $id\n shm_rmid $id \u0026\n shmunlock $id \u0026\n wait\n done\n\n The oops stack shows second fput tripping on an NULL f_inode. The\n first fput() completed via from shm_destroy(), but a racing thread did\n a get_file() and queued this fput():\n locks_remove_flock\n __fput\n ____fput\n task_work_run\n do_notify_resume\n int_signal\n\nFixes: c2c737a0461e (\"ipc,shm: shorten critical region for shmat\")\nFixes: 2caacaa82a51 (\"ipc,shm: shorten critical region for shmctl\")\nSigned-off-by: Greg Thelen \u003cgthelen@google.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e # 3.10.17+ 3.11.6+\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "5cbb3d216e2041700231bcfc383ee5f8b7fc8b74", "tree": "a738fa82dbcefa9bd283c08bc67f38827be63937", "parents": [ "9bc9ccd7db1c9f043f75380b5a5b94912046a60e", "4e9b45a19241354daec281d7a785739829b52359" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Nov 13 15:45:43 2013 +0900" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Nov 13 15:45:43 2013 +0900" }, "message": "Merge branch \u0027akpm\u0027 (patches from Andrew Morton)\n\nMerge first patch-bomb from Andrew Morton:\n \"Quite a lot of other stuff is banked up awaiting further\n next-\u003emainline merging, but this batch contains:\n\n - Lots of random misc patches\n - OCFS2\n - Most of MM\n - backlight updates\n - lib/ updates\n - printk updates\n - checkpatch updates\n - epoll tweaking\n - rtc updates\n - hfs\n - hfsplus\n - documentation\n - procfs\n - update gcov to gcc-4.7 format\n - IPC\"\n\n* emailed patches from Andrew Morton \u003cakpm@linux-foundation.org\u003e: (269 commits)\n ipc, msg: fix message length check for negative values\n ipc/util.c: remove unnecessary work pending test\n devpts: plug the memory leak in kill_sb\n ./Makefile: export initial ramdisk compression config option\n init/Kconfig: add option to disable kernel compression\n drivers: w1: make w1_slave::flags long to avoid memory corruption\n drivers/w1/masters/ds1wm.cuse dev_get_platdata()\n drivers/memstick/core/ms_block.c: fix unreachable state in h_msb_read_page()\n drivers/memstick/core/mspro_block.c: fix attributes array allocation\n drivers/pps/clients/pps-gpio.c: remove redundant of_match_ptr\n kernel/panic.c: reduce 1 byte usage for print tainted buffer\n gcov: reuse kbasename helper\n kernel/gcov/fs.c: use pr_warn()\n kernel/module.c: use pr_foo()\n gcov: compile specific gcov implementation based on gcc version\n gcov: add support for gcc 4.7 gcov format\n gcov: move gcov structs definitions to a gcc version specific file\n kernel/taskstats.c: return -ENOMEM when alloc memory fails in add_del_listener()\n kernel/taskstats.c: add nla_nest_cancel() for failure processing between nla_nest_start() and nla_nest_end()\n kernel/sysctl_binary.c: use scnprintf() instead of snprintf()\n ...\n" }, { "commit": "9bc9ccd7db1c9f043f75380b5a5b94912046a60e", "tree": "dd0a1b3396ae9414f668b0110cc39d11268ad3ed", "parents": [ "f0230294271f511b41797305b685365a9e569a09", "bdd3536618443809d18868563eeafa63b9d29603" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Nov 13 15:34:18 2013 +0900" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Nov 13 15:34:18 2013 +0900" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs\n\nPull vfs updates from Al Viro:\n \"All kinds of stuff this time around; some more notable parts:\n\n - RCU\u0027d vfsmounts handling\n - new primitives for coredump handling\n - files_lock is gone\n - Bruce\u0027s delegations handling series\n - exportfs fixes\n\n plus misc stuff all over the place\"\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (101 commits)\n ecryptfs: -\u003ef_op is never NULL\n locks: break delegations on any attribute modification\n locks: break delegations on link\n locks: break delegations on rename\n locks: helper functions for delegation breaking\n locks: break delegations on unlink\n namei: minor vfs_unlink cleanup\n locks: implement delegations\n locks: introduce new FL_DELEG lock flag\n vfs: take i_mutex on renamed file\n vfs: rename I_MUTEX_QUOTA now that it\u0027s not used for quotas\n vfs: don\u0027t use PARENT/CHILD lock classes for non-directories\n vfs: pull ext4\u0027s double-i_mutex-locking into common code\n exportfs: fix quadratic behavior in filehandle lookup\n exportfs: better variable name\n exportfs: move most of reconnect_path to helper function\n exportfs: eliminate unused \"noprogress\" counter\n exportfs: stop retrying once we race with rename/remove\n exportfs: clear DISCONNECTED on all parents sooner\n exportfs: more detailed comment for path_reconnect\n ...\n" }, { "commit": "4e9b45a19241354daec281d7a785739829b52359", "tree": "aa06af392271e191e6d9c5e7941cd1edbd8332e9", "parents": [ "206fa940977260ede421151aae067e2509356116" ], "author": { "name": "Mathias Krause", "email": "minipli@googlemail.com", "time": "Tue Nov 12 15:11:47 2013 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Nov 13 12:09:36 2013 +0900" }, "message": "ipc, msg: fix message length check for negative values\n\nOn 64 bit systems the test for negative message sizes is bogus as the\nsize, which may be positive when evaluated as a long, will get truncated\nto an int when passed to load_msg(). So a long might very well contain a\npositive value but when truncated to an int it would become negative.\n\nThat in combination with a small negative value of msg_ctlmax (which will\nbe promoted to an unsigned type for the comparison against msgsz, making\nit a big positive value and therefore make it pass the check) will lead to\ntwo problems: 1/ The kmalloc() call in alloc_msg() will allocate a too\nsmall buffer as the addition of alen is effectively a subtraction. 2/ The\ncopy_from_user() call in load_msg() will first overflow the buffer with\nuserland data and then, when the userland access generates an access\nviolation, the fixup handler copy_user_handle_tail() will try to fill the\nremainder with zeros -- roughly 4GB. That almost instantly results in a\nsystem crash or reset.\n\n ,-[ Reproducer (needs to be run as root) ]--\n | #include \u003csys/stat.h\u003e\n | #include \u003csys/msg.h\u003e\n | #include \u003cunistd.h\u003e\n | #include \u003cfcntl.h\u003e\n |\n | int main(void) {\n | long msg \u003d 1;\n | int fd;\n |\n | fd \u003d open(\"/proc/sys/kernel/msgmax\", O_WRONLY);\n | write(fd, \"-1\", 2);\n | close(fd);\n |\n | msgsnd(0, \u0026msg, 0xfffffff0, IPC_NOWAIT);\n |\n | return 0;\n | }\n \u0027---\n\nFix the issue by preventing msgsz from getting truncated by consistently\nusing size_t for the message length. This way the size checks in\ndo_msgsnd() could still be passed with a negative value for msg_ctlmax but\nwe would fail on the buffer allocation in that case and error out.\n\nAlso change the type of m_ts from int to size_t to avoid similar nastiness\nin other code paths -- it is used in similar constructs, i.e. signed vs.\nunsigned checks. It should never become negative under normal\ncircumstances, though.\n\nSetting msg_ctlmax to a negative value is an odd configuration and should\nbe prevented. As that might break existing userland, it will be handled\nin a separate commit so it could easily be reverted and reworked without\nreintroducing the above described bug.\n\nHardening mechanisms for user copy operations would have catched that bug\nearly -- e.g. checking slab object sizes on user copy operations as the\nusercopy feature of the PaX patch does. Or, for that matter, detect the\nlong vs. int sign change due to truncation, as the size overflow plugin\nof the very same patch does.\n\n[akpm@linux-foundation.org: fix i386 min() warnings]\nSigned-off-by: Mathias Krause \u003cminipli@googlemail.com\u003e\nCc: Pax Team \u003cpageexec@freemail.hu\u003e\nCc: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nCc: Brad Spengler \u003cspender@grsecurity.net\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e\t[ v2.3.27+ -- yes, that old ;) ]\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "206fa940977260ede421151aae067e2509356116", "tree": "07322ef889d492cf1cfea0ca0a36c86d6e30b026", "parents": [ "66da0e1f9034140ae2f571ef96e254a25083906c" ], "author": { "name": "Xie XiuQi", "email": "xiexiuqi@huawei.com", "time": "Tue Nov 12 15:11:46 2013 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Nov 13 12:09:36 2013 +0900" }, "message": "ipc/util.c: remove unnecessary work pending test\n\nRemove unnecessary work pending test before calling schedule_work(). It\nhas been tested in queue_work_on() already. No functional changed.\n\nSigned-off-by: Xie XiuQi \u003cxiexiuqi@huawei.com\u003e\nCc: Tejun Heo \u003ctj@kernel.org\u003e\nReviewed-by: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "b21996e36c8e3b92a84e972378bde80b43acd890", "tree": "0ed5eb8a3d11434ad6aa9c7efe24823bdb52b1e7", "parents": [ "9accbb977ab78234b8f298df5f306ed08d06bedb" ], "author": { "name": "J. Bruce Fields", "email": "bfields@redhat.com", "time": "Tue Sep 20 09:14:34 2011 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sat Nov 09 00:16:42 2013 -0500" }, "message": "locks: break delegations on unlink\n\nWe need to break delegations on any operation that changes the set of\nlinks pointing to an inode. Start with unlink.\n\nSuch operations also hold the i_mutex on a parent directory. Breaking a\ndelegation may require waiting for a timeout (by default 90 seconds) in\nthe case of a unresponsive NFS client. To avoid blocking all directory\noperations, we therefore drop locks before waiting for the delegation.\nThe logic then looks like:\n\n\tacquire locks\n\t...\n\ttest for delegation; if found:\n\t\ttake reference on inode\n\t\trelease locks\n\t\twait for delegation break\n\t\tdrop reference on inode\n\t\tretry\n\nIt is possible this could never terminate. (Even if we take precautions\nto prevent another delegation being acquired on the same inode, we could\nget a different inode on each retry.) But this seems very unlikely.\n\nThe initial test for a delegation happens after the lock on the target\ninode is acquired, but the directory inode may have been acquired\nfurther up the call stack. We therefore add a \"struct inode **\"\nargument to any intervening functions, which we use to pass the inode\nback up to the caller in the case it needs a delegation synchronously\nbroken.\n\nCc: David Howells \u003cdhowells@redhat.com\u003e\nCc: Tyler Hicks \u003ctyhicks@canonical.com\u003e\nCc: Dustin Kirkland \u003cdustin.kirkland@gazzang.com\u003e\nAcked-by: Jeff Layton \u003cjlayton@redhat.com\u003e\nSigned-off-by: J. Bruce Fields \u003cbfields@redhat.com\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "9bf76ca325d5e9208eb343f7bd4cc666f703ed30", "tree": "0c6e992844c24162838013caa4f4f3356d6594dd", "parents": [ "9dc8c89dfbbac5546101379d8d2aa0fa30d39888" ], "author": { "name": "Mathias Krause", "email": "minipli@googlemail.com", "time": "Sun Nov 03 12:36:28 2013 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Nov 03 10:53:11 2013 -0800" }, "message": "ipc, msg: forbid negative values for \"msg{max,mnb,mni}\"\n\nNegative message lengths make no sense -- so don\u0027t do negative queue\nlenghts or identifier counts. Prevent them from getting negative.\n\nAlso change the underlying data types to be unsigned to avoid hairy\nsurprises with sign extensions in cases where those variables get\nevaluated in unsigned expressions with bigger data types, e.g size_t.\n\nIn case a user still wants to have \"unlimited\" sizes she could just use\nINT_MAX instead.\n\nSigned-off-by: Mathias Krause \u003cminipli@googlemail.com\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6e224f94597842c5eb17f1fc2208d20b6f7f7d49", "tree": "f2a2ed86aa847b16b0d952411aed975f3a978908", "parents": [ "18ccee263c7e250a57f01c9434658f11f4118a64" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Wed Oct 16 13:46:45 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 16 21:35:52 2013 -0700" }, "message": "ipc/sem.c: synchronize semop and semctl with IPC_RMID\n\nAfter acquiring the semlock spinlock, operations must test that the\narray is still valid.\n\n - semctl() and exit_sem() would walk stale linked lists (ugly, but\n should be ok: all lists are empty)\n\n - semtimedop() would sleep forever - and if woken up due to a signal -\n access memory after free.\n\nThe patch also:\n - standardizes the tests for .deleted, so that all tests in one\n function leave the function with the same approach.\n - unconditionally tests for .deleted immediately after every call to\n sem_lock - even it it means that for semctl(GETALL), .deleted will be\n tested twice.\n\nBoth changes make the review simpler: After every sem_lock, there must\nbe a test of .deleted, followed by a goto to the cleanup code (if the\nfunction uses \"goto cleanup\").\n\nThe only exception is semctl_down(): If sem_ids().rwsem is locked, then\nthe presence in ids-\u003eipcs_idr is equivalent to !.deleted, thus no\nadditional test is required.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Mike Galbraith \u003cefault@gmx.de\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "18ccee263c7e250a57f01c9434658f11f4118a64", "tree": "1655243afdaed94d9e2bdb50af14e45e1cc92834", "parents": [ "9c56751271e7a917783fb57ec49fe8382e0dc867" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Wed Oct 16 13:46:45 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 16 21:35:52 2013 -0700" }, "message": "ipc: update locking scheme comments\n\nThe initial documentation was a bit incomplete, update accordingly.\n\n[akpm@linux-foundation.org: make it more readable in 80 columns]\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4271b05a227dc6175b66c3d9941aeab09048aeb2", "tree": "0c535ee0400940f5ad905128c9ecb405c9d65980", "parents": [ "0e8c665699e953fa58dc1b0b0d09e5dce7343cc7" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Sep 30 13:45:26 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Sep 30 14:31:03 2013 -0700" }, "message": "ipc,msg: prevent race with rmid in msgsnd,msgrcv\n\nThis fixes a race in both msgrcv() and msgsnd() between finding the msg\nand actually dealing with the queue, as another thread can delete shmid\nunderneath us if we are preempted before acquiring the\nkern_ipc_perm.lock.\n\nManfred illustrates this nicely:\n\nAssume a preemptible kernel that is preempted just after\n\n msq \u003d msq_obtain_object_check(ns, msqid)\n\nin do_msgrcv(). The only lock that is held is rcu_read_lock().\n\nNow the other thread processes IPC_RMID. When the first task is\nresumed, then it will happily wait for messages on a deleted queue.\n\nFix this by checking for if the queue has been deleted after taking the\nlock.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nReported-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Mike Galbraith \u003cefault@gmx.de\u003e\nCc: \u003cstable@vger.kernel.org\u003e \t[3.11]\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "0e8c665699e953fa58dc1b0b0d09e5dce7343cc7", "tree": "43f6f73159d7be2fbd7930315a053e27c4c8d189", "parents": [ "fb31ba30fb10fe0ee11739f51669d581b4a1412c" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Sep 30 13:45:25 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Sep 30 14:31:03 2013 -0700" }, "message": "ipc/sem.c: update sem_otime for all operations\n\nIn commit 0a2b9d4c7967 (\"ipc/sem.c: move wake_up_process out of the\nspinlock section\"), the update of semaphore\u0027s sem_otime(last semop time)\nwas moved to one central position (do_smart_update).\n\nBut since do_smart_update() is only called for operations that modify\nthe array, this means that wait-for-zero semops do not update sem_otime\nanymore.\n\nThe fix is simple:\nNon-alter operations must update sem_otime.\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nReported-by: Jia He \u003cjiakernel@gmail.com\u003e\nTested-by: Jia He \u003cjiakernel@gmail.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Mike Galbraith \u003cefault@gmx.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d8c633766ad88527f25d9f81a5c2f083d78a2b39", "tree": "68d0ce69d04d245b8c4236b84892ea740232d6f2", "parents": [ "6d07b68ce16ae9535955ba2059dedba5309c3ca1" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Sep 30 13:45:07 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Sep 30 14:31:01 2013 -0700" }, "message": "ipc/sem.c: synchronize the proc interface\n\nThe proc interface is not aware of sem_lock(), it instead calls\nipc_lock_object() directly. This means that simple semop() operations\ncan run in parallel with the proc interface. Right now, this is\nuncritical, because the implementation doesn\u0027t do anything that requires\na proper synchronization.\n\nBut it is dangerous and therefore should be fixed.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Mike Galbraith \u003cefault@gmx.de\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6d07b68ce16ae9535955ba2059dedba5309c3ca1", "tree": "f8eccb1e9ff34ffb41225d247329d7e73148b6fe", "parents": [ "5e9d527591421ccdb16acb8c23662231135d8686" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Sep 30 13:45:06 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Sep 30 14:31:01 2013 -0700" }, "message": "ipc/sem.c: optimize sem_lock()\n\nOperations that need access to the whole array must guarantee that there\nare no simple operations ongoing. Right now this is achieved by\nspin_unlock_wait(sem-\u003elock) on all semaphores.\n\nIf complex_count is nonzero, then this spin_unlock_wait() is not\nnecessary, because it was already performed in the past by the thread\nthat increased complex_count and even though sem_perm.lock was dropped\ninbetween, no simple operation could have started, because simple\noperations cannot start when complex_count is non-zero.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Mike Galbraith \u003cbitbucket@online.de\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "5e9d527591421ccdb16acb8c23662231135d8686", "tree": "dbe055ecd0bb6e497dceab2c9c0136151d04063a", "parents": [ "f6ea3adb70b20ae36277a1b0eaaf4da9f6479a28" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Sep 30 13:45:04 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Sep 30 14:31:01 2013 -0700" }, "message": "ipc/sem.c: fix race in sem_lock()\n\nThe exclusion of complex operations in sem_lock() is insufficient: after\nacquiring the per-semaphore lock, a simple op must first check that\nsem_perm.lock is not locked and only after that test check\ncomplex_count. The current code does it the other way around - and that\ncreates a race. Details are below.\n\nThe patch is a complete rewrite of sem_lock(), based in part on the code\nfrom Mike Galbraith. It removes all gotos and all loops and thus the\nrisk of livelocks.\n\nI have tested the patch (together with the next one) on my i3 laptop and\nit didn\u0027t cause any problems.\n\nThe bug is probably also present in 3.10 and 3.11, but for these kernels\nit might be simpler just to move the test of sma-\u003ecomplex_count after\nthe spin_is_locked() test.\n\nDetails of the bug:\n\nAssume:\n - sma-\u003ecomplex_count \u003d 0.\n - Thread 1: semtimedop(complex op that must sleep)\n - Thread 2: semtimedop(simple op).\n\nPseudo-Trace:\n\nThread 1: sem_lock(): acquire sem_perm.lock\nThread 1: sem_lock(): check for ongoing simple ops\n\t\t\tNothing ongoing, thread 2 is still before sem_lock().\nThread 1: try_atomic_semop()\n\t\u003c\u003c\u003c preempted.\n\nThread 2: sem_lock():\n static inline int sem_lock(struct sem_array *sma, struct sembuf *sops,\n int nsops)\n {\n int locknum;\n again:\n if (nsops \u003d\u003d 1 \u0026\u0026 !sma-\u003ecomplex_count) {\n struct sem *sem \u003d sma-\u003esem_base + sops-\u003esem_num;\n\n /* Lock just the semaphore we are interested in. */\n spin_lock(\u0026sem-\u003elock);\n\n /*\n * If sma-\u003ecomplex_count was set while we were spinning,\n * we may need to look at things we did not lock here.\n */\n if (unlikely(sma-\u003ecomplex_count)) {\n spin_unlock(\u0026sem-\u003elock);\n goto lock_array;\n }\n \u003c\u003c\u003c\u003c\u003c\u003c\u003c\u003c\u003c\n\t\u003c\u003c\u003c complex_count is still 0.\n\t\u003c\u003c\u003c\n \u003c\u003c\u003c Here it is preempted\n \u003c\u003c\u003c\u003c\u003c\u003c\u003c\u003c\u003c\n\nThread 1: try_atomic_semop() returns, notices that it must sleep.\nThread 1: increases sma-\u003ecomplex_count.\nThread 1: drops sem_perm.lock\nThread 2:\n /*\n * Another process is holding the global lock on the\n * sem_array; we cannot enter our critical section,\n * but have to wait for the global lock to be released.\n */\n if (unlikely(spin_is_locked(\u0026sma-\u003esem_perm.lock))) {\n spin_unlock(\u0026sem-\u003elock);\n spin_unlock_wait(\u0026sma-\u003esem_perm.lock);\n goto again;\n }\n\t\u003c\u003c\u003c sem_perm.lock already dropped, thus no \"goto again;\"\n\n locknum \u003d sops-\u003esem_num;\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Mike Galbraith \u003cbitbucket@online.de\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e\t[3.10+]\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "53dad6d3a8e5ac1af8bacc6ac2134ae1a8b085f1", "tree": "fc9349452c9bae7e86dbbbeed99c07bde4bad0c4", "parents": [ "4a10c2ac2f368583138b774ca41fac4207911983" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr@hp.com", "time": "Mon Sep 23 17:04:45 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 24 09:36:53 2013 -0700" }, "message": "ipc: fix race with LSMs\n\nCurrently, IPC mechanisms do security and auditing related checks under\nRCU. However, since security modules can free the security structure,\nfor example, through selinux_[sem,msg_queue,shm]_free_security(), we can\nrace if the structure is freed before other tasks are done with it,\ncreating a use-after-free condition. Manfred illustrates this nicely,\nfor instance with shared mem and selinux:\n\n -\u003e do_shmat calls rcu_read_lock()\n -\u003e do_shmat calls shm_object_check().\n Checks that the object is still valid - but doesn\u0027t acquire any locks.\n Then it returns.\n -\u003e do_shmat calls security_shm_shmat (e.g. selinux_shm_shmat)\n -\u003e selinux_shm_shmat calls ipc_has_perm()\n -\u003e ipc_has_perm accesses ipc_perms-\u003esecurity\n\nshm_close()\n -\u003e shm_close acquires rw_mutex \u0026 shm_lock\n -\u003e shm_close calls shm_destroy\n -\u003e shm_destroy calls security_shm_free (e.g. selinux_shm_free_security)\n -\u003e selinux_shm_free_security calls ipc_free_security(\u0026shp-\u003eshm_perm)\n -\u003e ipc_free_security calls kfree(ipc_perms-\u003esecurity)\n\nThis patch delays the freeing of the security structures after all RCU\nreaders are done. Furthermore it aligns the security life cycle with\nthat of the rest of IPC - freeing them based on the reference counter.\nFor situations where we need not free security, the current behavior is\nkept. Linus states:\n\n \"... the old behavior was suspect for another reason too: having the\n security blob go away from under a user sounds like it could cause\n various other problems anyway, so I think the old code was at least\n _prone_ to bugs even if it didn\u0027t have catastrophic behavior.\"\n\nI have tested this patch with IPC testcases from LTP on both my\nquad-core laptop and on a 64 core NUMA server. In both cases selinux is\nenabled, and tests pass for both voluntary and forced preemption models.\nWhile the mentioned races are theoretical (at least no one as reported\nthem), I wanted to make sure that this new logic doesn\u0027t break anything\nwe weren\u0027t aware of.\n\nSuggested-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr@hp.com\u003e\nAcked-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "20b8875abcf2daa1dda5cf70bd6369df5e85d4c1", "tree": "c2f22cc30656da1d32ffec2871db887e989cae10", "parents": [ "7a25dd9e042b2b94202a67e5551112f4ac87285a" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:31 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:45 2013 -0700" }, "message": "ipc: drop ipc_lock_check\n\nNo remaining users, we now use ipc_obtain_object_check().\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7a25dd9e042b2b94202a67e5551112f4ac87285a", "tree": "17a51b7eaf606565c6018d46192d1700f8b6a4b8", "parents": [ "32a2750010981216fb788c5190fb0e646abfab30" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:30 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:44 2013 -0700" }, "message": "ipc, shm: drop shm_lock_check\n\nThis function was replaced by a the lockless shm_obtain_object_check(),\nand no longer has any users.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "32a2750010981216fb788c5190fb0e646abfab30", "tree": "8a3d02037101eabc254f2e63836323e4373e18a2", "parents": [ "530fcd16d87cd2417c472a581ba5a1e501556c86" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:29 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:44 2013 -0700" }, "message": "ipc: drop ipc_lock_by_ptr\n\nAfter previous cleanups and optimizations, this function is no longer\nheavily used and we don\u0027t have a good reason to keep it. Update the few\nremaining callers and get rid of it.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "530fcd16d87cd2417c472a581ba5a1e501556c86", "tree": "c741f07d57e5b6fdf5920c6ba87e59b779073aaf", "parents": [ "05603c44a7627793219b0bd9a7b236099dc9cd9d" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:28 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:44 2013 -0700" }, "message": "ipc, shm: guard against non-existant vma in shmdt(2)\n\nWhen !CONFIG_MMU there\u0027s a chance we can derefence a NULL pointer when the\nVM area isn\u0027t found - check the return value of find_vma().\n\nAlso, remove the redundant -EINVAL return: retval is set to the proper\nreturn code and *only* changed to 0, when we actually unmap the segments.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "05603c44a7627793219b0bd9a7b236099dc9cd9d", "tree": "7efa266f81f6b68861bbea6cbda6baec48267408", "parents": [ "4718787d1f626f45ddb239912bc07266b9880044" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:26 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:43 2013 -0700" }, "message": "ipc: document general ipc locking scheme\n\nAs suggested by Andrew, add a generic initial locking scheme used\nthroughout all sysv ipc mechanisms. Documenting the ids rwsem, how rcu\ncan be enough to do the initial checks and when to actually acquire the\nkern_ipc_perm.lock spinlock.\n\nI found that adding it to util.c was generic enough.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4718787d1f626f45ddb239912bc07266b9880044", "tree": "160a8a8b414dac9426518987f3ff15bc679cbdb3", "parents": [ "d9a605e40b1376eb02b067d7690580255a0df68f" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:25 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:42 2013 -0700" }, "message": "ipc,msg: drop msg_unlock\n\nThere is only one user left, drop this function and just call\nipc_unlock_object() and rcu_read_unlock().\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d9a605e40b1376eb02b067d7690580255a0df68f", "tree": "b21254f7172ae8db6faffd9b7941d579fa421478", "parents": [ "c2c737a0461e61a34676bd0bd1bc1a70a1b4e396" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:24 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:42 2013 -0700" }, "message": "ipc: rename ids-\u003erw_mutex\n\nSince in some situations the lock can be shared for readers, we shouldn\u0027t\nbe calling it a mutex, rename it to rwsem.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c2c737a0461e61a34676bd0bd1bc1a70a1b4e396", "tree": "f8d9d21cbc43ac12a67bbc428cab083adb31cd30", "parents": [ "f42569b1388b1408b574a5e93a23a663647d4181" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:23 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:42 2013 -0700" }, "message": "ipc,shm: shorten critical region for shmat\n\nSimilar to other system calls, acquire the kern_ipc_perm lock after doing\nthe initial permission and security checks.\n\n[sasha.levin@oracle.com: dont leave do_shmat with rcu lock held]\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Sasha Levin \u003csasha.levin@oracle.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f42569b1388b1408b574a5e93a23a663647d4181", "tree": "cc3a37852d3ad59bfc76988d7c4b31973a468187", "parents": [ "2caacaa82a51b78fc0c800e206473874094287ed" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:22 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:42 2013 -0700" }, "message": "ipc,shm: cleanup do_shmat pasta\n\nClean up some of the messy do_shmat() spaghetti code, getting rid of\nout_free and out_put_dentry labels. This makes shortening the critical\nregion of this function in the next patch a little easier to do and read.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2caacaa82a51b78fc0c800e206473874094287ed", "tree": "07dfe1f340d35d907de80f4a80296c72896cef2c", "parents": [ "c97cb9ccab8c85428ec21eff690642ad2ce1fa8a" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:21 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:41 2013 -0700" }, "message": "ipc,shm: shorten critical region for shmctl\n\nWith the *_INFO, *_STAT, IPC_RMID and IPC_SET commands already optimized,\ndeal with the remaining SHM_LOCK and SHM_UNLOCK commands. Take the\nshm_perm lock after doing the initial auditing and security checks. The\nrest of the logic remains unchanged.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c97cb9ccab8c85428ec21eff690642ad2ce1fa8a", "tree": "c435476a079d3bff28ec8c67377227effba57ca9", "parents": [ "68eccc1dc345539d589ae78ee43b835c1a06a134" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:20 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:39 2013 -0700" }, "message": "ipc,shm: make shmctl_nolock lockless\n\nWhile the INFO cmd doesn\u0027t take the ipc lock, the STAT commands do acquire\nit unnecessarily. We can do the permissions and security checks only\nholding the rcu lock.\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "68eccc1dc345539d589ae78ee43b835c1a06a134", "tree": "16943d2674537a5d13a14794763a3df7e51411ab", "parents": [ "3b1c4ad37741e53804ffe0a30dd01e08b2ab6241" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:18 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:39 2013 -0700" }, "message": "ipc,shm: introduce shmctl_nolock\n\nSimilar to semctl and msgctl, when calling msgctl, the *_INFO and *_STAT\ncommands can be performed without acquiring the ipc object.\n\nAdd a shmctl_nolock() function and move the logic of *_INFO and *_STAT out\nof msgctl(). Since we are just moving functionality, this change still\ntakes the lock and it will be properly lockless in the next patch.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3b1c4ad37741e53804ffe0a30dd01e08b2ab6241", "tree": "72bdaef5ae179d87eed60879c75128e2384f6aa4", "parents": [ "79ccf0f8c8e04e8b9eda6645ba0f63b0915a3075" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:17 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:39 2013 -0700" }, "message": "ipc: drop ipcctl_pre_down\n\nNow that sem, msgque and shm, through *_down(), all use the lockless\nvariant of ipcctl_pre_down(), go ahead and delete it.\n\n[akpm@linux-foundation.org: fix function name in kerneldoc, cleanups]\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "79ccf0f8c8e04e8b9eda6645ba0f63b0915a3075", "tree": "7178c7dbb1edd12689359f49da36d4e04de93cf8", "parents": [ "8b8d52ac382b17a19906b930cd69e2edb0aca8ba" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:16 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:39 2013 -0700" }, "message": "ipc,shm: shorten critical region in shmctl_down\n\nInstead of holding the ipc lock for the entire function, use the\nipcctl_pre_down_nolock and only acquire the lock for specific commands:\nRMID and SET.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8b8d52ac382b17a19906b930cd69e2edb0aca8ba", "tree": "19c4c3afbb5c349e9a6ee570c7dd10c73d2f4a90", "parents": [ "6e19eded3684dc184181093af3bff2ff440f5b53" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Wed Sep 11 14:26:15 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Sep 11 15:59:38 2013 -0700" }, "message": "ipc,shm: introduce lockless functions to obtain the ipc object\n\nThis is the third and final patchset that deals with reducing the amount\nof contention we impose on the ipc lock (kern_ipc_perm.lock). These\nchanges mostly deal with shared memory, previous work has already been\ndone for semaphores and message queues:\n\n http://lkml.org/lkml/2013/3/20/546 (sems)\n http://lkml.org/lkml/2013/5/15/584 (mqueues)\n\nWith these patches applied, a custom shm microbenchmark stressing shmctl\ndoing IPC_STAT with 4 threads a million times, reduces the execution\ntime by 50%. A similar run, this time with IPC_SET, reduces the\nexecution time from 3 mins and 35 secs to 27 seconds.\n\nPatches 1-8: replaces blindly taking the ipc lock for a smarter\ncombination of rcu and ipc_obtain_object, only acquiring the spinlock\nwhen updating.\n\nPatch 9: renames the ids rw_mutex to rwsem, which is what it already was.\n\nPatch 10: is a trivial mqueue leftover cleanup\n\nPatch 11: adds a brief lock scheme description, requested by Andrew.\n\nThis patch:\n\nAdd shm_obtain_object() and shm_obtain_object_check(), which will allow us\nto get the ipc object without acquiring the lock. Just as with other\nforms of ipc, these functions are basically wrappers around\nipc_obtain_object*().\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c7c4591db64dbd1e504bc4e2806d7ef290a3c81b", "tree": "a2fb124f9760eec668d20541383e762822d7cc7b", "parents": [ "11c7b03d42a847db90862d0f9d8be6ce9b2f0553", "c7b96acf1456ef127fef461fcfedb54b81fecfbb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Sep 07 14:35:32 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Sep 07 14:35:32 2013 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace\n\nPull namespace changes from Eric Biederman:\n \"This is an assorted mishmash of small cleanups, enhancements and bug\n fixes.\n\n The major theme is user namespace mount restrictions. nsown_capable\n is killed as it encourages not thinking about details that need to be\n considered. A very hard to hit pid namespace exiting bug was finally\n tracked and fixed. A couple of cleanups to the basic namespace\n infrastructure.\n\n Finally there is an enhancement that makes per user namespace\n capabilities usable as capabilities, and an enhancement that allows\n the per userns root to nice other processes in the user namespace\"\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace:\n userns: Kill nsown_capable it makes the wrong thing easy\n capabilities: allow nice if we are privileged\n pidns: Don\u0027t have unshare(CLONE_NEWPID) imply CLONE_THREAD\n userns: Allow PR_CAPBSET_DROP in a user namespace.\n namespaces: Simplify copy_namespaces so it is clear what is going on.\n pidns: Fix hang in zap_pid_ns_processes by sending a potentially extra wakeup\n sysfs: Restrict mounting sysfs\n userns: Better restrictions on when proc and sysfs can be mounted\n vfs: Don\u0027t copy mount bind mounts of /proc/\u003cpid\u003e/ns/mnt between namespaces\n kernel/nsproxy.c: Improving a snippet of code.\n proc: Restrict mounting the proc filesystem\n vfs: Lock in place mounts from more privileged users\n" }, { "commit": "bebcb928c820d0ee83aca4b192adc195e43e66a2", "tree": "28ad45577c7e65e28938eed04b0459674a11f0e3", "parents": [ "ec1882a9391c55332ebf3d1654f40b76e4a6c010" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Tue Sep 03 16:00:08 2013 +0200" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 03 10:42:56 2013 -0700" }, "message": "ipc/msg.c: Fix lost wakeup in msgsnd().\n\nThe check if the queue is full and adding current to the wait queue of\npending msgsnd() operations (ss_add()) must be atomic.\n\nOtherwise:\n - the thread that performs msgsnd() finds a full queue and decides to\n sleep.\n - the thread that performs msgrcv() first reads all messages from the\n queue and then sleeps, because the queue is empty.\n - the msgrcv() calls do not perform any wakeups, because the msgsnd()\n task has not yet called ss_add().\n - then the msgsnd()-thread first calls ss_add() and then sleeps.\n\nNet result: msgsnd() and msgrcv() both sleep forever.\n\nObserved with msgctl08 from ltp with a preemptible kernel.\n\nFix: Call ipc_lock_object() before performing the check.\n\nThe patch also moves security_msg_queue_msgsnd() under ipc_lock_object:\n - msgctl(IPC_SET) explicitely mentions that it tries to expunge any\n pending operations that are not allowed anymore with the new\n permissions. If security_msg_queue_msgsnd() is called without locks,\n then there might be races.\n - it makes the patch much simpler.\n\nReported-and-tested-by: Vineet Gupta \u003cVineet.Gupta1@synopsys.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: stable@vger.kernel.org # for 3.11\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c7b96acf1456ef127fef461fcfedb54b81fecfbb", "tree": "1cc9387d23e96685453e545bda6d5a5efea8fa63", "parents": [ "f54fb863c6bbcbafdfc332b4a4260abb5a002137" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Wed Mar 20 12:49:49 2013 -0700" }, "committer": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Fri Aug 30 23:44:11 2013 -0700" }, "message": "userns: Kill nsown_capable it makes the wrong thing easy\n\nnsown_capable is a special case of ns_capable essentially for just CAP_SETUID and\nCAP_SETGID. For the existing users it doesn\u0027t noticably simplify things and\nfrom the suggested patches I have seen it encourages people to do the wrong\nthing. So remove nsown_capable.\n\nAcked-by: Serge Hallyn \u003cserge.hallyn@canonical.com\u003e\nSigned-off-by: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\n" }, { "commit": "368ae537e056acd3f751fa276f48423f06803922", "tree": "c0e8dec82905b231fc74b732fd0966d62d530b22", "parents": [ "aaaafb7f953c60d9c9eeb1b10ecdbe6970421b24" ], "author": { "name": "Svenning Sørensen", "email": "sss@secomea.dk", "time": "Wed Aug 28 16:35:17 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 28 19:26:38 2013 -0700" }, "message": "IPC: bugfix for msgrcv with msgtyp \u003c 0\n\nAccording to \u0027man msgrcv\u0027: \"If msgtyp is less than 0, the first message of\nthe lowest type that is less than or equal to the absolute value of msgtyp\nshall be received.\"\n\nBug: The kernel only returns a message if its type is 1; other messages\nwith type \u003c abs(msgtype) will never get returned.\n\nFix: After having traversed the list to find the first message with the\nlowest type, we need to actually return that message.\n\nThis regression was introduced by commit daaf74cf0867 (\"ipc: refactor\nmsg list search into separate function\")\n\nSigned-off-by: Svenning Soerensen \u003csss@secomea.dk\u003e\nReviewed-by: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "758a6ba39ef6df4cdc615e5edd7bd86eab81a5f7", "tree": "519c1f2e672e888c04b5447a85ad78aa08c1c693", "parents": [ "d12e1e50e47e0900dbbf52237b7e171f4f15ea1e" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Jul 08 16:01:26 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:28 2013 -0700" }, "message": "ipc/sem.c: rename try_atomic_semop() to perform_atomic_semop(), docu update\n\nCleanup: Some minor points that I noticed while writing the previous\npatches\n\n1) The name try_atomic_semop() is misleading: The function performs the\n operation (if it is possible).\n\n2) Some documentation updates.\n\nNo real code change, a rename and documentation changes.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d12e1e50e47e0900dbbf52237b7e171f4f15ea1e", "tree": "83b54585a883ee062353eb464d93944d08ab4a1c", "parents": [ "f269f40ad5aeee229ed70044926f44318abe41ef" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Jul 08 16:01:25 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:28 2013 -0700" }, "message": "ipc/sem.c: replace shared sem_otime with per-semaphore value\n\nsem_otime contains the time of the last semaphore operation that\ncompleted successfully. Every operation updates this value, thus access\nfrom multiple cpus can cause thrashing.\n\nTherefore the patch replaces the variable with a per-semaphore variable.\nThe per-array sem_otime is only calculated when required.\n\nNo performance improvement on a single-socket i3 - only important for\nlarger systems.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f269f40ad5aeee229ed70044926f44318abe41ef", "tree": "878fd6dabbf34c154809ed21f6d64cb90c7e394c", "parents": [ "1a82e9e1d0f1b45f47a97c9e2349020536ff8987" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Jul 08 16:01:24 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:28 2013 -0700" }, "message": "ipc/sem.c: always use only one queue for alter operations\n\nThere are two places that can contain alter operations:\n - the global queue: sma-\u003epending_alter\n - the per-semaphore queues: sma-\u003esem_base[].pending_alter.\n\nSince one of the queues must be processed first, this causes an odd\npriorization of the wakeups: complex operations have priority over\nsimple ops.\n\nThe patch restores the behavior of linux \u003c\u003d3.0.9: The longest waiting\noperation has the highest priority.\n\nThis is done by using only one queue:\n - if there are complex ops, then sma-\u003epending_alter is used.\n - otherwise, the per-semaphore queues are used.\n\nAs a side effect, do_smart_update_queue() becomes much simpler: no more\ngoto logic.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "1a82e9e1d0f1b45f47a97c9e2349020536ff8987", "tree": "956810c32e02f6ae1527db015c6ae622800bd720", "parents": [ "f5c936c0f267ec58641451cf8b8d39b4c207ee4d" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Jul 08 16:01:23 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:28 2013 -0700" }, "message": "ipc/sem: separate wait-for-zero and alter tasks into seperate queues\n\nIntroduce separate queues for operations that do not modify the\nsemaphore values. Advantages:\n\n - Simpler logic in check_restart().\n - Faster update_queue(): Right now, all wait-for-zero operations are\n always tested, even if the semaphore value is not 0.\n - wait-for-zero gets again priority, as in linux \u003c\u003d3.0.9\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f5c936c0f267ec58641451cf8b8d39b4c207ee4d", "tree": "714671efc570653a64e602b70524741b0a488907", "parents": [ "196aa0132fc7261f34b10ae1bfb44abc1bc69b3c" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Jul 08 16:01:22 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:28 2013 -0700" }, "message": "ipc/sem.c: cacheline align the semaphore structures\n\nAs now each semaphore has its own spinlock and parallel operations are\npossible, give each semaphore its own cacheline.\n\nOn a i3 laptop, this gives up to 28% better performance:\n\n #semscale 10 | grep \"interleave 2\"\n - before:\n Cpus 1, interleave 2 delay 0: 36109234 in 10 secs\n Cpus 2, interleave 2 delay 0: 55276317 in 10 secs\n Cpus 3, interleave 2 delay 0: 62411025 in 10 secs\n Cpus 4, interleave 2 delay 0: 81963928 in 10 secs\n\n -after:\n Cpus 1, interleave 2 delay 0: 35527306 in 10 secs\n Cpus 2, interleave 2 delay 0: 70922909 in 10 secs \u003c\u003c\u003c + 28%\n Cpus 3, interleave 2 delay 0: 80518538 in 10 secs\n Cpus 4, interleave 2 delay 0: 89115148 in 10 secs \u003c\u003c\u003c + 8.7%\n\ni3, with 2 cores and with hyperthreading enabled. Interleave 2 in order\nuse first the full cores. HT partially hides the delay from cacheline\ntrashing, thus the improvement is \"only\" 8.7% if 4 threads are running.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "196aa0132fc7261f34b10ae1bfb44abc1bc69b3c", "tree": "022d2de428ef8d8c642017adf448790ab803bd96", "parents": [ "9ad66ae65fc8d3e7e3344310fb0aa835910264fe" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Mon Jul 08 16:01:20 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:28 2013 -0700" }, "message": "ipc/util.c, ipc_rcu_alloc: cacheline align allocation\n\nEnforce that ipc_rcu_alloc returns a cacheline aligned pointer on SMP.\n\nRationale:\n\nThe SysV sem code tries to move the main spinlock into a seperate\ncacheline (____cacheline_aligned_in_smp). This works only if\nipc_rcu_alloc returns cacheline aligned pointers. vmalloc and kmalloc\nreturn cacheline algined pointers, the implementation of ipc_rcu_alloc\nbreaks that.\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9ad66ae65fc8d3e7e3344310fb0aa835910264fe", "tree": "25dbc032d43dbc9e5cb85c93408fa06e27e5e29d", "parents": [ "41a0d523d0f626e9da0dc01de47f1b89058033cf" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:19 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc: remove unused functions\n\nWe can now drop the msg_lock and msg_lock_check functions along with a\nbogus comment introduced previously in semctl_down.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "41a0d523d0f626e9da0dc01de47f1b89058033cf", "tree": "af4021f97c49c59784dbab62fb1dad36fe6d6f3c", "parents": [ "3dd1f784ed6603d7ab1043e51e6371235edf2313" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:18 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc,msg: shorten critical region in msgrcv\n\ndo_msgrcv() is the last msg queue function that abuses the ipc lock Take\nit only when needed when actually updating msq.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3dd1f784ed6603d7ab1043e51e6371235edf2313", "tree": "21bd41e3f445ca3ba3b96b24f53e9e4adc2ea1fb", "parents": [ "ac0ba20ea6f2201a1589d6dc26ad1a4f0f967bb8" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:17 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc,msg: shorten critical region in msgsnd\n\ndo_msgsnd() is another function that does too many things with the ipc\nobject lock acquired. Take it only when needed when actually updating\nmsq.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ac0ba20ea6f2201a1589d6dc26ad1a4f0f967bb8", "tree": "bc3e2e4c4140f07ee1861b724ba7e1b295d7bbc6", "parents": [ "a5001a0d9768568de5d613c3b3a5b9c7721299da" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:16 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc,msg: make msgctl_nolock lockless\n\nWhile the INFO cmd doesn\u0027t take the ipc lock, the STAT commands do\nacquire it unnecessarily. We can do the permissions and security checks\nonly holding the rcu lock.\n\nThis function now mimics semctl_nolock().\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a5001a0d9768568de5d613c3b3a5b9c7721299da", "tree": "8bdb9f698da1f9e1469322892ba6f6e7c796533b", "parents": [ "2cafed30f150f7314f98717b372df8173516cae0" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:15 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc,msg: introduce lockless functions to obtain the ipc object\n\nAdd msq_obtain_object() and msq_obtain_object_check(), which will allow\nus to get the ipc object without acquiring the lock. Just as with\nsemaphores, these functions are basically wrappers around\nipc_obtain_object*().\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2cafed30f150f7314f98717b372df8173516cae0", "tree": "21f2440948b12d2905d1c2ded7fcc7f88e980360", "parents": [ "15724ecb7e9bab35fc694c666ad563adba820cc3" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:14 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc,msg: introduce msgctl_nolock\n\nSimilar to semctl, when calling msgctl, the *_INFO and *_STAT commands\ncan be performed without acquiring the ipc object.\n\nAdd a msgctl_nolock() function and move the logic of *_INFO and *_STAT\nout of msgctl(). This change still takes the lock and it will be\nproperly lockless in the next patch\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "15724ecb7e9bab35fc694c666ad563adba820cc3", "tree": "d9ae089252e84327367056a29c0cf6b16f844c2e", "parents": [ "7b4cc5d8411bd4e9d61d8714f53859740cf830c2" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:13 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc,msg: shorten critical region in msgctl_down\n\nInstead of holding the ipc lock for the entire function, use the\nipcctl_pre_down_nolock and only acquire the lock for specific commands:\nRMID and SET.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7b4cc5d8411bd4e9d61d8714f53859740cf830c2", "tree": "b95f3b875a5f4c927b0f27cc3c7ddcba5cc8e1e8", "parents": [ "cf9d5d78d05bca96df7618dfc3a5ee4414dcae58" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:12 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc: move locking out of ipcctl_pre_down_nolock\n\nThis function currently acquires both the rw_mutex and the rcu lock on\nsuccessful lookups, leaving the callers to explicitly unlock them,\ncreating another two level locking situation.\n\nMake the callers (including those that still use ipcctl_pre_down())\nexplicitly lock and unlock the rwsem and rcu lock.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "cf9d5d78d05bca96df7618dfc3a5ee4414dcae58", "tree": "b03375cb8d7a2482ad659bd9b675589f427c2336", "parents": [ "1ca7003ab41152d673d9e359632283d05294f3d6" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:11 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc: close open coded spin lock calls\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "1ca7003ab41152d673d9e359632283d05294f3d6", "tree": "b587df3d3a2aecc47acd74712348c2895c292bb6", "parents": [ "dbfcd91f06f0e2d5564b2fd184e9c2a43675f9ab" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:10 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:27 2013 -0700" }, "message": "ipc: introduce ipc object locking helpers\n\nSimple helpers around the (kern_ipc_perm *)-\u003elock spinlock.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "dbfcd91f06f0e2d5564b2fd184e9c2a43675f9ab", "tree": "dff823ad3c61e2b199fb42c71abe81264d60e5dd", "parents": [ "c103a4dc4a32f53f095b66cd798d648c652f05b4" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Mon Jul 08 16:01:09 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:26 2013 -0700" }, "message": "ipc: move rcu lock out of ipc_addid\n\nThis patchset continues the work that began in the sysv ipc semaphore\nscaling series, see\n\n https://lkml.org/lkml/2013/3/20/546\n\nJust like semaphores used to be, sysv shared memory and msg queues also\nabuse the ipc lock, unnecessarily holding it for operations such as\npermission and security checks.\n\nThis patchset mostly deals with mqueues, and while shared mem can be\ndone in a very similar way, I want to get these patches out in the open\nfirst. It also does some pending cleanups, mostly focused on the two\nlevel locking we have in ipc code, taking care of ipc_addid() and\nipcctl_pre_down_nolock() - yes there are still functions that need to be\nupdated as well.\n\nThis patch:\n\nMake all callers explicitly take and release the RCU read lock.\n\nThis addresses the two level locking seen in newary(), newseg() and\nnewqueue(). For the last two, explicitly unlock the ipc object and the\nrcu lock, instead of calling the custom shm_unlock and msg_unlock\nfunctions. The next patch will deal with the open coded locking for\n-\u003eperm.lock\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c103a4dc4a32f53f095b66cd798d648c652f05b4", "tree": "4cc3cbd0007255ab708747b6a5b9c3669e1388a8", "parents": [ "f7da04c9e363e479258135ac825734d78aecd2b0" ], "author": { "name": "Andrew Morton", "email": "akpm@linux-foundation.org", "time": "Mon Jul 08 16:01:08 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:26 2013 -0700" }, "message": "ipc/shmc.c: eliminate ugly 80-col tricks\n\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "79f6530cb59e2a0af6953742a33cc29e98ca631c", "tree": "3778b26699b0f217a3c888853faaf0e15c760fc2", "parents": [ "f9f0a7d0dcbd19e9705e8b96a4b408f035e25c93" ], "author": { "name": "Jeff Layton", "email": "jlayton@redhat.com", "time": "Mon Jul 08 15:59:36 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 09 10:33:19 2013 -0700" }, "message": "audit: fix mq_open and mq_unlink to add the MQ root as a hidden parent audit_names record\n\nThe old audit PATH records for mq_open looked like this:\n\n type\u003dPATH msg\u003daudit(1366282323.982:869): item\u003d1 name\u003d(null) inode\u003d6777\n dev\u003d00:0c mode\u003d041777 ouid\u003d0 ogid\u003d0 rdev\u003d00:00\n obj\u003dsystem_u:object_r:tmpfs_t:s15:c0.c1023\n type\u003dPATH msg\u003daudit(1366282323.982:869): item\u003d0 name\u003d\"test_mq\" inode\u003d26732\n dev\u003d00:0c mode\u003d0100700 ouid\u003d0 ogid\u003d0 rdev\u003d00:00\n obj\u003dstaff_u:object_r:user_tmpfs_t:s15:c0.c1023\n\n...with the audit related changes that went into 3.7, they now look like this:\n\n type\u003dPATH msg\u003daudit(1366282236.776:3606): item\u003d2 name\u003d(null) inode\u003d66655\n dev\u003d00:0c mode\u003d0100700 ouid\u003d0 ogid\u003d0 rdev\u003d00:00\n obj\u003dstaff_u:object_r:user_tmpfs_t:s15:c0.c1023\n type\u003dPATH msg\u003daudit(1366282236.776:3606): item\u003d1 name\u003d(null) inode\u003d6926\n dev\u003d00:0c mode\u003d041777 ouid\u003d0 ogid\u003d0 rdev\u003d00:00\n obj\u003dsystem_u:object_r:tmpfs_t:s15:c0.c1023\n type\u003dPATH msg\u003daudit(1366282236.776:3606): item\u003d0 name\u003d\"test_mq\"\n\nBoth of these look wrong to me. As Steve Grubb pointed out:\n\n \"What we need is 1 PATH record that identifies the MQ. The other PATH\n records probably should not be there.\"\n\nFix it to record the mq root as a parent, and flag it such that it\nshould be hidden from view when the names are logged, since the root of\nthe mq filesystem isn\u0027t terribly interesting. With this change, we get\na single PATH record that looks more like this:\n\n type\u003dPATH msg\u003daudit(1368021604.836:484): item\u003d0 name\u003d\"test_mq\" inode\u003d16914\n dev\u003d00:0c mode\u003d0100644 ouid\u003d0 ogid\u003d0 rdev\u003d00:00\n obj\u003dunconfined_u:object_r:user_tmpfs_t:s0\n\nIn order to do this, a new audit_inode_parent_hidden() function is\nadded. If we do it this way, then we avoid having the existing callers\nof audit_inode needing to do any sort of flag conversion if auditing is\ninactive.\n\nSigned-off-by: Jeff Layton \u003cjlayton@redhat.com\u003e\nReported-by: Jiri Jaburek \u003cjjaburek@redhat.com\u003e\nCc: Steve Grubb \u003csgrubb@redhat.com\u003e\nCc: Eric Paris \u003ceparis@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ab465df9dda42a997f7537c875127eeb6a88158c", "tree": "28a1ec28c15f1b59c938b7757c8f56eed0de0ecd", "parents": [ "89ff77837a67994e4a4a20bb648687fbcc3083f2" ], "author": { "name": "Manfred Spraul", "email": "manfred@colorfullife.com", "time": "Sun May 26 11:08:52 2013 +0200" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun May 26 15:14:51 2013 -0700" }, "message": "ipc/sem.c: Fix missing wakeups in do_smart_update_queue()\n\ndo_smart_update_queue() is called when an operation (semop,\nsemctl(SETVAL), semctl(SETALL), ...) modified the array. It must check\nwhich of the sleeping tasks can proceed.\n\ndo_smart_update_queue() missed a few wakeups:\n - if a sleeping complex op was completed, then all per-semaphore queues\n must be scanned - not only those that were modified by *sops\n - if a sleeping simple op proceeded, then the global queue must be\n scanned again\n\nAnd:\n - the test for \"|sops \u003d\u003d NULL) before scanning the global queue is not\n required: If the global queue is empty, then it doesn\u0027t need to be\n scanned - regardless of the reason for calling do_smart_update_queue()\n\nThe patch is not optimized, i.e. even completing a wait-for-zero\noperation causes a rescan. This is done to keep the patch as simple as\npossible.\n\nSigned-off-by: Manfred Spraul \u003cmanfred@colorfullife.com\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "091d0d55b286c9340201b4ed4470be87fc568228", "tree": "5417ab8864fbabe1a9931f3a9a81355cd3d3ebaa", "parents": [ "de2657f94acd4f0df44626db7c4d2b71babc8cd3" ], "author": { "name": "Li Zefan", "email": "lizefan@huawei.com", "time": "Thu May 09 15:08:15 2013 +0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 09 14:22:47 2013 -0700" }, "message": "shm: fix null pointer deref when userspace specifies invalid hugepage size\n\nDave reported an oops triggered by trinity:\n\n BUG: unable to handle kernel NULL pointer dereference at 0000000000000008\n IP: newseg+0x10d/0x390\n PGD cf8c1067 PUD cf8c2067 PMD 0\n Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC\n CPU: 2 PID: 7636 Comm: trinity-child2 Not tainted 3.9.0+#67\n ...\n Call Trace:\n ipcget+0x182/0x380\n SyS_shmget+0x5a/0x60\n tracesys+0xdd/0xe2\n\nThis bug was introduced by commit af73e4d9506d (\"hugetlbfs: fix mmap\nfailure in unaligned size request\").\n\nReported-by: Dave Jones \u003cdavej@redhat.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e\nSigned-off-by: Li Zefan \u003clizfan@huawei.com\u003e\nReviewed-by: Naoya Horiguchi \u003cn-horiguchi@ah.jp.nec.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "de2657f94acd4f0df44626db7c4d2b71babc8cd3", "tree": "4499951b79ff2d199bca0d87ef379d829cc1e400", "parents": [ "ebc2e5e6a408a0e6ed63c0ba98c2c8a232c6b4f4" ], "author": { "name": "Rik van Riel", "email": "riel@redhat.com", "time": "Thu May 09 16:59:59 2013 -0400" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 09 14:17:47 2013 -0700" }, "message": "ipc,sem: fix semctl(..., GETNCNT)\n\nThe semctl GETNCNT returns the number of semops waiting for the\nspecified semaphore to become nonzero. After commit 9f1bc2c9022c\n(\"ipc,sem: have only one list in struct sem_queue\"), the semops waiting\non just one semaphore are waiting on that semaphore\u0027s list.\n\nIn order to return the correct count, we have to walk that list too, in\naddition to the sem_array\u0027s list for complex operations.\n\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ebc2e5e6a408a0e6ed63c0ba98c2c8a232c6b4f4", "tree": "eb376fff2b32ae63b6d5ba8faac2f26a899d51df", "parents": [ "07e074503eba3ee657ab50a8c9497ddf90039e7e" ], "author": { "name": "Rik van Riel", "email": "riel@redhat.com", "time": "Thu May 09 16:53:28 2013 -0400" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 09 14:17:47 2013 -0700" }, "message": "ipc,sem: fix semctl(..., GETZCNT)\n\nThe semctl GETZCNT returns the number of semops waiting for the\nspecified semaphore to become zero. After commit 9f1bc2c9022c\n(\"ipc,sem: have only one list in struct sem_queue\"), the semops waiting\non just one semaphore are waiting on that semaphore\u0027s list.\n\nIn order to return the correct count, we have to walk that list too, in\naddition to the sem_array\u0027s list for complex operations.\n\nThis bug broke dbench; it works again with this patch applied.\n\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nReported-by: Kent Overstreet \u003ckoverstreet@google.com\u003e\nTested-by: Kent Overstreet \u003ckoverstreet@google.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "af73e4d9506d3b797509f3c030e7dcd554f7d9c4", "tree": "f9f1bf7483495b66b6cf2cfb3c676791133733b1", "parents": [ "1ab4ce762370b82870834899e49c08129d7ae271" ], "author": { "name": "Naoya Horiguchi", "email": "n-horiguchi@ah.jp.nec.com", "time": "Tue May 07 16:18:13 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue May 07 18:38:27 2013 -0700" }, "message": "hugetlbfs: fix mmap failure in unaligned size request\n\nThe current kernel returns -EINVAL unless a given mmap length is\n\"almost\" hugepage aligned. This is because in sys_mmap_pgoff() the\ngiven length is passed to vm_mmap_pgoff() as it is without being aligned\nwith hugepage boundary.\n\nThis is a regression introduced in commit 40716e29243d (\"hugetlbfs: fix\nalignment of huge page requests\"), where alignment code is pushed into\nhugetlb_file_setup() and the variable len in caller side is not changed.\n\nTo fix this, this patch partially reverts that commit, and adds\nalignment code in caller side. And it also introduces hstate_sizelog()\nin order to get proper hstate to specified hugepage size.\n\nAddresses https://bugzilla.kernel.org/show_bug.cgi?id\u003d56881\n\n[akpm@linux-foundation.org: fix warning when CONFIG_HUGETLB_PAGE\u003dn]\nSigned-off-by: Naoya Horiguchi \u003cn-horiguchi@ah.jp.nec.com\u003e\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReported-by: \u003ciceman_dvd@yahoo.com\u003e\nCc: Steven Truelove \u003csteven.truelove@utoronto.ca\u003e\nCc: Jianguo Wu \u003cwujianguo@huawei.com\u003e\nCc: Hugh Dickins \u003chughd@google.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "941b0304a74b240c607ff098401fd4ef70c9d1cc", "tree": "92cb88d1851360c5e890b7daf737b264cf3181de", "parents": [ "c728b9c87b59fb943c4cba0552d38152787a4ab6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 11:04:29 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 17:24:59 2013 -0700" }, "message": "ipc: simplify rcu_read_lock() in semctl_nolock()\n\nThis trivially combines two rcu_read_lock() calls in both sides of a\nif-statement into one single one in front of the if-statement.\n\nSplit out as an independent cleanup from the previous commit.\n\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c728b9c87b59fb943c4cba0552d38152787a4ab6", "tree": "072a4bac6686701ef46eca2fabb06306c400391b", "parents": [ "321310ced2d6cc0175c76fa512fa8a829ee35223" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 11:04:29 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 17:21:58 2013 -0700" }, "message": "ipc: simplify semtimedop/semctl_main() common error path handling\n\nWith various straight RCU lock/unlock movements, one common exit path\npattern had become\n\n\trcu_read_unlock();\n\tgoto out_wakeup;\n\nand in fact there were no cases where we wanted to exit to out_wakeup\n_without_ releasing the RCU read lock.\n\nSo replace that pattern with \"goto out_rcu_wakeup\", and remove the old\nout_wakeup.\n\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "321310ced2d6cc0175c76fa512fa8a829ee35223", "tree": "0722d86daa8778a69207965cc6bc005d0def9f08", "parents": [ "fbfd1d2862a8316c7191bc551c6a842e6918abb0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 10:47:57 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 17:20:14 2013 -0700" }, "message": "ipc: move sem_obtain_lock() rcu locking into the only caller\n\nsem_obtain_lock() was another of those functions that returned with the\nRCU lock held for reading in the success case. Move the RCU locking to\nthe caller (semtimedop()), making it more obvious. We already did RCU\nlocking elsewhere in that function.\n\nSide note: why does semtimedop() re-do the semphore lookup after the\nsleep, rather than just getting a reference to the semaphore it already\nlooked up originally?\n\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "fbfd1d2862a8316c7191bc551c6a842e6918abb0", "tree": "06c5a391238bcf93bb3d7f2c587d069eeb6809e5", "parents": [ "4091fd942e96af5a0b1dfa6aac5f44153ebf7cdb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 10:25:11 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 17:19:59 2013 -0700" }, "message": "ipc: fix double sem unlock in semctl error path\n\nFix another ipc locking buglet introduced by the scalability patches:\nwhen semctl_down() was changed to delay the semaphore locking, one error\npath for security_sem_semctl() went through the semaphore unlock logic\neven though the semaphore had never been locked.\n\nIntroduced by commit 16df3674efe3 (\"ipc,sem: do not hold ipc lock more\nthan necessary\")\n\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4091fd942e96af5a0b1dfa6aac5f44153ebf7cdb", "tree": "0455a439ad1a9884018286670e49c4e18cfaa359", "parents": [ "73b29505c36eeb4751eccad41f6aad78562521f8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 10:13:40 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 17:19:39 2013 -0700" }, "message": "ipc: move the rcu_read_lock() from sem_lock_and_putref() into callers\n\nThis is another ipc semaphore locking cleanup, trying to make the\nlocking more straightforward. We move the rcu read locking into the\ncallers of sem_lock_and_putref(), which in general means that we now\nmostly do the rcu_read_lock() and rcu_read_unlock() in the same\nfunction.\n\nMostly. We still have the ipc_addid/newary/freeary mess, and things\nlike ipcctl_pre_down_nolock().\n\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "73b29505c36eeb4751eccad41f6aad78562521f8", "tree": "228603617659029e082a0968b2e0a6490c121b0d", "parents": [ "6d49dab8ae06c6d35a4d0967364a9ecbe8fdea2c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 03 15:22:00 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 11:24:21 2013 -0700" }, "message": "ipc: sem_putref() does not need the semaphore lock any more\n\nipc_rcu_putref() uses atomics for the refcount, and the games to lock\nand unlock the semaphore just to try to keep the reference counting\nworking are no longer useful.\n\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6d49dab8ae06c6d35a4d0967364a9ecbe8fdea2c", "tree": "7f28ea59e2c52d912faf696d38e9577ffb1c29c3", "parents": [ "ce857229e0c3adc211944a13a5579ef84fd7b4af" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 03 15:04:40 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 04 11:24:00 2013 -0700" }, "message": "ipc: move rcu_read_unlock() out of sem_unlock() and into callers\n\nThe IPC locking is a mess, and sem_unlock() unlocks not only the\nsemaphore spinlock, it also drops the rcu read lock. Unlike sem_lock(),\nwhich just gets the spin-lock, and expects the caller to get the rcu\nread lock.\n\nThis all makes things very hard to follow, and it\u0027s very confusing when\nyou take the rcu read lock in one function, and then release it in\nanother. And it has caused actual bugs: the sem_obtain_lock() function\nended up dropping the RCU read lock twice in one error path, because it\nfirst did the sem_unlock(), and then did a rcu_read_unlock() to match\nthe rcu_read_lock() it had done.\n\nThis is just a totally mindless \"remove rcu_read_unlock() from\nsem_unlock() and add it immediately after each caller\" (except for the\naforementioned bug where we did too many rcu_read_unlock(), and in\nfind_alloc_undo() where we just got the rcu_read_lock() to correct for\nthe fact that sem_unlock would immediately drop it again).\n\nWe can (and should) clean things up further, but this fixes the bug with\nthe minimal amount of subtlety.\n\nReviewed-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ce857229e0c3adc211944a13a5579ef84fd7b4af", "tree": "f12310c2c6492dd7ef4b25dd0c6052f7a114b9cf", "parents": [ "20a2078ce7705a6e0722ef5184336eb8657a58d8" ], "author": { "name": "Al Viro", "email": "viro@ZenIV.linux.org.uk", "time": "Fri May 03 00:30:49 2013 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 02 19:51:31 2013 -0700" }, "message": "ipc: fix GETALL/IPC_RM race for sysv semaphores\n\nWe can step on WARN_ON_ONCE() in sem_getref() if a semaphore is removed\njust as we are about to call sem_getref() from semctl_main(); results\nare not pretty.\n\nWe should fail with -EIDRM, same as if IPC_RM happened while we\u0027d been\ndoing allocation there. This also expands sem_getref() at its only\ncallsite (and fixed there), while sem_getref_and_unlock() is simply\nkilled off - it has no callers at all.\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "600fe9751aeb6f6b72de84076a05c5b8c04152c0", "tree": "3c02f2627ab4dad3cee03d3c03955376830b5065", "parents": [ "4ada8db38a44654446fe35ceb20a1972220e0f69" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Apr 29 12:42:09 2013 -0400" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 02 08:03:33 2013 -0700" }, "message": "ipc_schedule_free() can do vfree() directly now\n\nCommit 32fcfd40715e (\"make vfree() safe to call from interrupt\ncontexts\") made it safe to do vfree directly from the RCU callback,\nwhich allows us to simplify ipc/util.c a lot by getting rid of the\ndifferences between vmalloc/kmalloc memory.\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "20b4fb485227404329e41ad15588afad3df23050", "tree": "f3e099f0ab3da8a93b447203e294d2bb22f6dc05", "parents": [ "b9394d8a657cd3c064fa432aa0905c1b58b38fe9", "ac3e3c5b1164397656df81b9e9ab4991184d3236" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 17:51:54 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 17:51:54 2013 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs\n\nPull VFS updates from Al Viro,\n\nMisc cleanups all over the place, mainly wrt /proc interfaces (switch\ncreate_proc_entry to proc_create(), get rid of the deprecated\ncreate_proc_read_entry() in favor of using proc_create_data() and\nseq_file etc).\n\n7kloc removed.\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (204 commits)\n don\u0027t bother with deferred freeing of fdtables\n proc: Move non-public stuff from linux/proc_fs.h to fs/proc/internal.h\n proc: Make the PROC_I() and PDE() macros internal to procfs\n proc: Supply a function to remove a proc entry by PDE\n take cgroup_open() and cpuset_open() to fs/proc/base.c\n ppc: Clean up scanlog\n ppc: Clean up rtas_flash driver somewhat\n hostap: proc: Use remove_proc_subtree()\n drm: proc: Use remove_proc_subtree()\n drm: proc: Use minor-\u003eindex to label things, not PDE-\u003ename\n drm: Constify drm_proc_list[]\n zoran: Don\u0027t print proc_dir_entry data in debug\n reiserfs: Don\u0027t access the proc_dir_entry in r_open(), r_start() r_show()\n proc: Supply an accessor for getting the data from a PDE\u0027s parent\n airo: Use remove_proc_subtree()\n rtl8192u: Don\u0027t need to save device proc dir PDE\n rtl8187se: Use a dir under /proc/net/r8180/\n proc: Add proc_mkdir_data()\n proc: Move some bits from linux/proc_fs.h to linux/{of.h,signal.h,tty.h}\n proc: Move PDE_NET() to fs/proc/proc_net.c\n ...\n" }, { "commit": "0bb80f240520c4148b623161e7856858c021696d", "tree": "eeb27af589bc12d5864bd2594e5eea59aabe37db", "parents": [ "c3bef7bcaaa7d9f6704fcd81a171c9f0c91a2259" ], "author": { "name": "David Howells", "email": "dhowells@redhat.com", "time": "Fri Apr 12 01:50:06 2013 +0100" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Wed May 01 17:29:39 2013 -0400" }, "message": "proc: Split the namespace stuff out into linux/proc_ns.h\n\nSplit the proc namespace stuff out into linux/proc_ns.h.\n\nSigned-off-by: David Howells \u003cdhowells@redhat.com\u003e\ncc: netdev@vger.kernel.org\ncc: Serge E. Hallyn \u003cserge.hallyn@ubuntu.com\u003e\ncc: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "d69f3bad4675ac519d41ca2b11e1c00ca115cecd", "tree": "343d1fab484751290b2044b347372f8a8468e762", "parents": [ "41239fe82d85c135684b09f1e65622d6c1dbe8dc" ], "author": { "name": "Robin Holt", "email": "holt@sgi.com", "time": "Tue Apr 30 19:15:54 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:58 2013 -0700" }, "message": "ipc: sysv shared memory limited to 8TiB\n\nTrying to run an application which was trying to put data into half of\nmemory using shmget(), we found that having a shmall value below 8EiB-8TiB\nwould prevent us from using anything more than 8TiB. By setting\nkernel.shmall greater than 8EiB-8TiB would make the job work.\n\nIn the newseg() function, ns-\u003eshm_tot which, at 8TiB is INT_MAX.\n\nipc/shm.c:\n 458 static int newseg(struct ipc_namespace *ns, struct ipc_params *params)\n 459 {\n...\n 465 int numpages \u003d (size + PAGE_SIZE -1) \u003e\u003e PAGE_SHIFT;\n...\n 474 if (ns-\u003eshm_tot + numpages \u003e ns-\u003eshm_ctlall)\n 475 return -ENOSPC;\n\n[akpm@linux-foundation.org: make ipc/shm.c:newseg()\u0027s numpages size_t, not int]\nSigned-off-by: Robin Holt \u003cholt@sgi.com\u003e\nReported-by: Alex Thorlton \u003cathorlton@sgi.com\u003e\nCc: \u003cstable@vger.kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "41239fe82d85c135684b09f1e65622d6c1dbe8dc", "tree": "c7c44c3f87444d9636bc911ce98401ab974c21dd", "parents": [ "6062a8dc0517bce23e3c2f7d2fea5e22411269a3" ], "author": { "name": "Nikola Pajkovsky", "email": "npajkovs@redhat.com", "time": "Tue Apr 30 19:15:49 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:58 2013 -0700" }, "message": "ipc/msg.c: use list_for_each_entry_[safe] for list traversing\n\nThe ipc/msg.c code does its list operations by hand and it open-codes the\naccesses, instead of using for_each_entry_[safe].\n\nSigned-off-by: Nikola Pajkovsky \u003cnpajkovs@redhat.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6062a8dc0517bce23e3c2f7d2fea5e22411269a3", "tree": "e1dd1553167fccb726a8aa9352b27ba14f188374", "parents": [ "9f1bc2c9022c1d4944c4a1a44c2f365487420aca" ], "author": { "name": "Rik van Riel", "email": "riel@surriel.com", "time": "Tue Apr 30 19:15:44 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:58 2013 -0700" }, "message": "ipc,sem: fine grained locking for semtimedop\n\nIntroduce finer grained locking for semtimedop, to handle the common case\nof a program wanting to manipulate one semaphore from an array with\nmultiple semaphores.\n\nIf the call is a semop manipulating just one semaphore in an array with\nmultiple semaphores, only take the lock for that semaphore itself.\n\nIf the call needs to manipulate multiple semaphores, or another caller is\nin a transaction that manipulates multiple semaphores, the sem_array lock\nis taken, as well as all the locks for the individual semaphores.\n\nOn a 24 CPU system, performance numbers with the semop-multi\ntest with N threads and N semaphores, look like this:\n\n\tvanilla\t\tDavidlohr\u0027s\tDavidlohr\u0027s +\tDavidlohr\u0027s +\nthreads\t\t\tpatches\t\trwlock patches\tv3 patches\n10\t610652\t\t726325\t\t1783589\t\t2142206\n20\t341570\t\t365699\t\t1520453\t\t1977878\n30\t288102\t\t307037\t\t1498167\t\t2037995\n40\t290714\t\t305955\t\t1612665\t\t2256484\n50\t288620\t\t312890\t\t1733453\t\t2650292\n60\t289987\t\t306043\t\t1649360\t\t2388008\n70\t291298\t\t306347\t\t1723167\t\t2717486\n80\t290948\t\t305662\t\t1729545\t\t2763582\n90\t290996\t\t306680\t\t1736021\t\t2757524\n100\t292243\t\t306700\t\t1773700\t\t3059159\n\n[davidlohr.bueso@hp.com: do not call sem_lock when bogus sma]\n[davidlohr.bueso@hp.com: make refcounter atomic]\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nSuggested-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Chegu Vinod \u003cchegu_vinod@hp.com\u003e\nCc: Jason Low \u003cjason.low2@hp.com\u003e\nReviewed-by: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nTested-by: Emmanuel Benisty \u003cbenisty.e@gmail.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9f1bc2c9022c1d4944c4a1a44c2f365487420aca", "tree": "a1764801b992d87e1c488035135f10ed80413944", "parents": [ "c460b662d5cae467f1c341c59b02a5c5e68fed0b" ], "author": { "name": "Rik van Riel", "email": "riel@surriel.com", "time": "Tue Apr 30 19:15:39 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:58 2013 -0700" }, "message": "ipc,sem: have only one list in struct sem_queue\n\nHaving only one list in struct sem_queue, and only queueing simple\nsemaphore operations on the list for the semaphore involved, allows us to\nintroduce finer grained locking for semtimedop.\n\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Chegu Vinod \u003cchegu_vinod@hp.com\u003e\nCc: Emmanuel Benisty \u003cbenisty.e@gmail.com\u003e\nCc: Jason Low \u003cjason.low2@hp.com\u003e\nCc: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c460b662d5cae467f1c341c59b02a5c5e68fed0b", "tree": "0cc2f44f5a0abd0a7a143d8bd57630e27bc52311", "parents": [ "16df3674efe39f3ab63e7052f1244dd3d50e7f84" ], "author": { "name": "Rik van Riel", "email": "riel@surriel.com", "time": "Tue Apr 30 19:15:35 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:58 2013 -0700" }, "message": "ipc,sem: open code and rename sem_lock\n\nRename sem_lock() to sem_obtain_lock(), so we can introduce a sem_lock()\nlater that only locks the sem_array and does nothing else.\n\nOpen code the locking from ipc_lock() in sem_obtain_lock() so we can\nintroduce finer grained locking for the sem_array in the next patch.\n\n[akpm@linux-foundation.org: propagate the ipc_obtain_object() errno out of sem_obtain_lock()]\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nCc: Chegu Vinod \u003cchegu_vinod@hp.com\u003e\nCc: Emmanuel Benisty \u003cbenisty.e@gmail.com\u003e\nCc: Jason Low \u003cjason.low2@hp.com\u003e\nCc: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "16df3674efe39f3ab63e7052f1244dd3d50e7f84", "tree": "45ecd49ba72e0b84dbaac338d79173ba1ea4c972", "parents": [ "444d0f621b64716f7868dcbde448e0c66ece4e61" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Tue Apr 30 19:15:29 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:58 2013 -0700" }, "message": "ipc,sem: do not hold ipc lock more than necessary\n\nInstead of holding the ipc lock for permissions and security checks, among\nothers, only acquire it when necessary.\n\nSome numbers....\n\n1) With Rik\u0027s semop-multi.c microbenchmark we can see the following\n results:\n\nBaseline (3.9-rc1):\ncpus 4, threads: 256, semaphores: 128, test duration: 30 secs\ntotal operations: 151452270, ops/sec 5048409\n\n+ 59.40% a.out [kernel.kallsyms] [k] _raw_spin_lock\n+ 6.14% a.out [kernel.kallsyms] [k] sys_semtimedop\n+ 3.84% a.out [kernel.kallsyms] [k] avc_has_perm_flags\n+ 3.64% a.out [kernel.kallsyms] [k] __audit_syscall_exit\n+ 2.06% a.out [kernel.kallsyms] [k] copy_user_enhanced_fast_string\n+ 1.86% a.out [kernel.kallsyms] [k] ipc_lock\n\nWith this patchset:\ncpus 4, threads: 256, semaphores: 128, test duration: 30 secs\ntotal operations: 273156400, ops/sec 9105213\n\n+ 18.54% a.out [kernel.kallsyms] [k] _raw_spin_lock\n+ 11.72% a.out [kernel.kallsyms] [k] sys_semtimedop\n+ 7.70% a.out [kernel.kallsyms] [k] ipc_has_perm.isra.21\n+ 6.58% a.out [kernel.kallsyms] [k] avc_has_perm_flags\n+ 6.54% a.out [kernel.kallsyms] [k] __audit_syscall_exit\n+ 4.71% a.out [kernel.kallsyms] [k] ipc_obtain_object_check\n\n2) While on an Oracle swingbench DSS (data mining) workload the\n improvements are not as exciting as with Rik\u0027s benchmark, we can see\n some positive numbers. For an 8 socket machine the following are the\n percentages of %sys time incurred in the ipc lock:\n\nBaseline (3.9-rc1):\n100 swingbench users: 8,74%\n400 swingbench users: 21,86%\n800 swingbench users: 84,35%\n\nWith this patchset:\n100 swingbench users: 8,11%\n400 swingbench users: 19,93%\n800 swingbench users: 77,69%\n\n[riel@redhat.com: fix two locking bugs]\n[sasha.levin@oracle.com: prevent releasing RCU read lock twice in semctl_main]\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: Chegu Vinod \u003cchegu_vinod@hp.com\u003e\nAcked-by: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Jason Low \u003cjason.low2@hp.com\u003e\nCc: Emmanuel Benisty \u003cbenisty.e@gmail.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "444d0f621b64716f7868dcbde448e0c66ece4e61", "tree": "3947e0fe791c53c81d87a929af6d1b8e18a5c6aa", "parents": [ "4d2bff5eb86e8d7b4a20934cccb93bdeebed3558" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Tue Apr 30 19:15:24 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:58 2013 -0700" }, "message": "ipc: introduce lockless pre_down ipcctl\n\nVarious forms of ipc use ipcctl_pre_down() to retrieve an ipc object and\ncheck permissions, mostly for IPC_RMID and IPC_SET commands.\n\nIntroduce ipcctl_pre_down_nolock(), a lockless version of this function.\nThe locking version is retained, yet modified to call the nolock version\nwithout affecting its semantics, thus transparent to all ipc callers.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nSuggested-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nCc: Chegu Vinod \u003cchegu_vinod@hp.com\u003e\nCc: Emmanuel Benisty \u003cbenisty.e@gmail.com\u003e\nCc: Jason Low \u003cjason.low2@hp.com\u003e\nCc: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4d2bff5eb86e8d7b4a20934cccb93bdeebed3558", "tree": "3667eec32ddcf8686335f64db553613c319f0625", "parents": [ "7bb4deff61bdab3338534841cb6d0508314a41d6" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Tue Apr 30 19:15:19 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:57 2013 -0700" }, "message": "ipc: introduce obtaining a lockless ipc object\n\nThrough ipc_lock() and therefore ipc_lock_check() we currently return the\nlocked ipc object. This is not necessary for all situations and can,\ntherefore, cause unnecessary ipc lock contention.\n\nIntroduce analogous ipc_obtain_object() and ipc_obtain_object_check()\nfunctions that only lookup and return the ipc object.\n\nBoth these functions must be called within the RCU read critical section.\n\n[akpm@linux-foundation.org: propagate the ipc_obtain_object() errno from ipc_lock()]\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: Chegu Vinod \u003cchegu_vinod@hp.com\u003e\nAcked-by: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Emmanuel Benisty \u003cbenisty.e@gmail.com\u003e\nCc: Jason Low \u003cjason.low2@hp.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7bb4deff61bdab3338534841cb6d0508314a41d6", "tree": "2a3be8bf66313a085b9e7c084bbcee5cad8a9b2a", "parents": [ "1e3c941c52eab70c8acb2f77829c24673445c858" ], "author": { "name": "Davidlohr Bueso", "email": "davidlohr.bueso@hp.com", "time": "Tue Apr 30 19:15:14 2013 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed May 01 08:12:57 2013 -0700" }, "message": "ipc: remove bogus lock comment for ipc_checkid\n\nThis series makes the sysv semaphore code more scalable, by reducing the\ntime the semaphore lock is held, and making the locking more scalable for\nsemaphore arrays with multiple semaphores.\n\nThe first four patches were written by Davidlohr Buesso, and reduce the\nhold time of the semaphore lock.\n\nThe last three patches change the sysv semaphore code locking to be more\nfine grained, providing a performance boost when multiple semaphores in a\nsemaphore array are being manipulated simultaneously.\n\nOn a 24 CPU system, performance numbers with the semop-multi\ntest with N threads and N semaphores, look like this:\n\n\tvanilla\t\tDavidlohr\u0027s\tDavidlohr\u0027s +\tDavidlohr\u0027s +\n\tthreads\t\t\tpatches\t\trwlock patches\tv3 patches\n\t10\t610652\t\t726325\t\t1783589\t\t2142206\n\t20\t341570\t\t365699\t\t1520453\t\t1977878\n\t30\t288102\t\t307037\t\t1498167\t\t2037995\n\t40\t290714\t\t305955\t\t1612665\t\t2256484\n\t50\t288620\t\t312890\t\t1733453\t\t2650292\n\t60\t289987\t\t306043\t\t1649360\t\t2388008\n\t70\t291298\t\t306347\t\t1723167\t\t2717486\n\t80\t290948\t\t305662\t\t1729545\t\t2763582\n\t90\t290996\t\t306680\t\t1736021\t\t2757524\n\t100\t292243\t\t306700\t\t1773700\t\t3059159\n\nThis patch:\n\nThere is no reason to be holding the ipc lock while reading ipcp-\u003eseq,\nhence remove misleading comment.\n\nAlso simplify the return value for the function.\n\nSigned-off-by: Davidlohr Bueso \u003cdavidlohr.bueso@hp.com\u003e\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Chegu Vinod \u003cchegu_vinod@hp.com\u003e\nCc: Emmanuel Benisty \u003cbenisty.e@gmail.com\u003e\nCc: Jason Low \u003cjason.low2@hp.com\u003e\nCc: Michel Lespinasse \u003cwalken@google.com\u003e\nCc: Peter Hurley \u003cpeter@hurleysoftware.com\u003e\nCc: Stanislav Kinsbursky \u003cskinsbursky@parallels.com\u003e\nTested-by: Sedat Dilek \u003csedat.dilek@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" } ], "next": "1e3c941c52eab70c8acb2f77829c24673445c858" }