litmus-rt.git - The LITMUS^RT kernel.

Branch	Commit message	Author	Age
archive/unc-master-3.0	P-FP: fix BUG_ON releated to priority inheritance	Bjoern Brandenburg	12 years
archived-2013.1	uncachedev: mmap memory that is not cached by CPUs	Glenn Elliott	12 years
archived-private-master	Merge branch 'wip-2.6.34' into old-private-master	Andrea Bastoni	15 years
archived-semi-part	Merge branch 'wip-semi-part' of ssh://cvs/cvs/proj/litmus/repo/litmus2010 int...	Andrea Bastoni	14 years
demo	Further refinements	Jonathan Herman	13 years
ecrts-pgm-final	Merge branch 'wip-ecrts14-pgm' of ssh://rtsrv.cs.unc.edu/home/litmus/litmus-r...	Glenn Elliott	11 years
ecrts14-pgm-final	Merge branch 'wip-ecrts14-pgm' of ssh://rtsrv.cs.unc.edu/home/litmus/litmus-r...	Glenn Elliott	11 years
gpusync-rtss12	Final GPUSync implementation.	Glenn Elliott	12 years
gpusync/staging	Rename IKGLP R2DGLP.	Glenn Elliott	11 years
linux-tip	Merge branch 'slab/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/p...	Linus Torvalds	14 years
litmus2008-patch-series	add i386 feather-trace implementation	Bjoern B. Brandenburg	15 years
master	PSN-EDF: use inferred_sporadic_job_release_at	Bjoern Brandenburg	9 years
pgm	make it compile	Glenn Elliott	12 years
prop/litmus-signals	Infrastructure for Litmus signals.	Glenn Elliott	12 years
prop/robust-tie-break	Fixed bug in edf_higher_prio().	Glenn Elliott	12 years
staging	Fix tracepoint compilation error	Felipe Cerqueira	12 years
test	9/23/2016	Namhoon Kim	8 years
tracing-devel	Test kernel tracing events capabilities	Andrea Bastoni	15 years
v2.6.34-with-arm-patches	smsc911x: Add spinlocks around registers access	Catalin Marinas	15 years
v2015.1	Add ARM syscall def for get_current_budget	Bjoern Brandenburg	9 years
wip-2011.2-bbb	Litmus core: simplify np-section protocol	Bjoern B. Brandenburg	13 years
wip-2011.2-bbb-trace	Refactor sched_trace_log_message() -> debug_trace_log_message()	Andrea Bastoni	13 years
wip-2012.3-gpu	SOBLIV draining support for C-EDF.	Glenn Elliott	12 years
wip-2012.3-gpu-preport	pick up last C-RM file	Glenn Elliott	11 years
wip-2012.3-gpu-rtss13	Fix critical bug in GPU tracker.	Glenn Elliott	11 years
wip-2012.3-gpu-sobliv-budget-w-kshark	Proper sobliv draining and many bug fixes.	Glenn Elliott	12 years
wip-aedzl-final	Make it easier to compile AEDZL interfaces in liblitmus.	Glenn Elliott	14 years
wip-aedzl-revised	Add sched_trace data for Apative EDZL	Glenn Elliott	14 years
wip-arbit-deadline	Fix compilation bug.	Glenn Elliott	13 years
wip-aux-tasks	Description of refined aux task inheritance.	Glenn Elliott	12 years
wip-bbb	GSN-EDF & Core: improve debug TRACE'ing for NP sections	Bjoern B. Brandenburg	14 years
wip-bbb-prio-don	use correct timestamp	Bjoern B. Brandenburg	14 years
wip-better-break	Implement hash-based EDF tie-breaking.	Glenn Elliott	13 years
wip-binary-heap	Make C-EDF work with simplified binheap_delete	Glenn Elliott	13 years
wip-budget	Added support for choices in budget policy enforcement.	Glenn Elliott	15 years
wip-color	Summarize schedulability with final record	Jonathan Herman	13 years
wip-color-jlh	sched_color: Fixed two bugs causing crashing on experiment restart and a rare...	Jonathan Herman	13 years
wip-d10-hz1000	Enable HZ=1000 on District 10	Bjoern B. Brandenburg	14 years
wip-default-clustering	Feature: Make default C-EDF clustering compile-time configurable.	Glenn Elliott	14 years
wip-dissipation-jerickso	Update from 2.6.36 to 2.6.36.4	Jeremy Erickson	11 years
wip-dissipation2-jerickso	Update 2.6.36 to 2.6.36.4	Jeremy Erickson	11 years
wip-ecrts14-pgm	Merge branch 'wip-ecrts14-pgm' of ssh://rtsrv.cs.unc.edu/home/litmus/litmus-r...	Glenn Elliott	11 years
wip-edf-hsb	last tested version	Jonathan Herman	13 years
wip-edf-os	Lookup table EDF-os	Jeremy Erickson	12 years
wip-edf-tie-break	Merge branch 'wip-edf-tie-break' of ssh://rtsrv.cs.unc.edu/home/litmus/litmus...	Glenn Elliott	12 years
wip-edzl-critique	Use hr_timer's active checks instead of having own flag.	Glenn Elliott	14 years
wip-edzl-final	Implementation of the EDZL scheduler.	Glenn Elliott	14 years
wip-edzl-revised	Clean up comments.	Glenn Elliott	14 years
wip-events	Added support for tracing arbitrary actions.	Jonathan Herman	14 years
wip-extra-debug	DBG: add additional tracing	Bjoern B. Brandenburg	14 years
wip-fix-switch-jerickso	Attempt to fix race condition with plugin switching	Jeremy Erickson	14 years
wip-fix3	sched: show length of runqueue clock deactivation in /proc/sched_debug	Bjoern B. Brandenburg	14 years
wip-fmlp-dequeue	Improve FMLP queue management.	Glenn Elliott	14 years
wip-ft-irq-flag	Feather-Trace: keep track of interrupt-related interference.	Bjoern B. Brandenburg	13 years
wip-gpu-cleanup	Enable sched_trace log injection from userspace	Glenn Elliott	12 years
wip-gpu-interrupts	Remove option for threading of all softirqs.	Glenn Elliott	13 years
wip-gpu-rtas12	Generalized GPU cost predictors + EWMA. (untested)	Glenn Elliott	12 years
wip-gpu-rtss12	Final GPUSync implementation.	Glenn Elliott	13 years
wip-gpu-rtss12-srp	experimental changes to support GPUs under SRP	Glenn Elliott	13 years
wip-gpusync-merge	Cleanup priority tracking for budget enforcement.	Glenn Elliott	11 years
wip-ikglp	Move RSM and IKGLP imp. to own .c files	Glenn Elliott	13 years
wip-k-fmlp	Merge branch 'mpi-master' into wip-k-fmlp	Glenn Elliott	13 years
wip-kernel-coloring	Added recolor syscall	Namhoon Kim	7 years
wip-kernthreads	Kludge work-queue processing into klitirqd.	Glenn Elliott	14 years
wip-klmirqd-to-aux	Allow klmirqd threads to be given names.	Glenn Elliott	12 years
wip-kshark	Merge branch 'mpi-staging' into wip-kshark	Jonathan Herman	12 years
wip-litmus-3.2	Merge commit 'v3.2' into litmus-staging	Andrea Bastoni	12 years
wip-litmus2011.2	Cleanup: Coding conformance for affinity stuff.	Glenn Elliott	14 years
wip-litmus3.0-2011.2	Feather-Trace: keep track of interrupt-related interference.	Bjoern B. Brandenburg	13 years
wip-master-2.6.33-rt	Avoid deadlock when switching task policy to BACKGROUND (ugly)	Andrea Bastoni	15 years
wip-mc	Removed ARM-specific hacks which disabled less common mixed-criticality featu...	Jonathan Herman	12 years
wip-mc-bipasa	MC-EDF added	bipasa chattopadhyay	13 years
wip-mc-jerickso	Split C/D queues	Jeremy Erickson	14 years
wip-mc2-cache-slack	Manually patched mc^2 related code	Ming Yang	9 years
wip-mcrit-mac	cosmetic	Mac Mollison	14 years
wip-merge-3.0	Prevent Linux to send IPI and queue tasks on remote CPUs.	Andrea Bastoni	13 years
wip-merge-v3.0	Prevent Linux to send IPI and queue tasks on remote CPUs.	Andrea Bastoni	13 years
wip-migration-affinity	NULL affinity dereference in C-EDF.	Glenn Elliott	14 years
wip-mmap-uncache	share branch with others	Glenn Elliott	12 years
wip-modechange	RTSS 2017 submission	Namhoon Kim	8 years
wip-nested-locking	Appears to be working.	Bryan Ward	12 years
wip-omlp-gedf	First implementation of G-OMLP.	Glenn Elliott	15 years
wip-pai	Some cleanup of PAI	Glenn Elliott	13 years
wip-percore-lib	9/21/2016	Namhoon Kim	8 years
wip-performance	CONFIG_DONT_PREEMPT_ON_TIE: Don't preeempt a scheduled task on priority tie.	Glenn Elliott	14 years
wip-pgm	Add PGM support to C-FL	Glenn Elliott	11 years
wip-pgm-split	First draft of C-FL-split	Namhoon Kim	11 years
wip-pm-ovd	Add preemption-and-migration overhead tracing support	Andrea Bastoni	15 years
wip-prio-inh	P-EDF updated to use the generic pi framework.	Glenn Elliott	15 years
wip-prioq-dgl	BUG FIX: Support DGLs with PRIOQ_MUTEX	Glenn Elliott	12 years
wip-refactored-gedf	Generalizd architecture for GEDF-style scheduelrs to reduce code redundancy.	Glenn Elliott	14 years
wip-release-master-fix	bugfix: release master CPU must signal task was picked	Bjoern B. Brandenburg	14 years
wip-robust-tie-break	EDF priority tie-breaks.	Glenn Elliott	12 years
wip-rt-kshark	Move task time accounting into the complete_job method.	Jonathan Herman	12 years
wip-rtas12-pgm	Scheduling of PGM jobs.	Glenn Elliott	13 years
wip-semi-part	Fix compile error with newer GCC	Jeremy Erickson	12 years
wip-semi-part-edfos-jerickso	Use initial CPU set by client	Jeremy Erickson	12 years
wip-shared-lib	TODO: Fix condition checks in replicate_page_move_mapping()	Namhoon Kim	8 years
wip-shared-lib2	RTAS 2017 Submission ver.	Namhoon Kim	8 years
wip-shared-mem	Initial commit for shared library	Namhoon Kim	8 years
wip-splitting-jerickso	Fix release behavior	Jeremy Erickson	12 years
wip-splitting-omlp-jerickso	Bjoern's Dissertation Code with Priority Donation	Jeremy Erickson	12 years
wip-stage-binheap	An efficient binary heap implementation.	Glenn Elliott	13 years
wip-sun-port	Dynamic memory allocation and clean exit for FeatherTrace	Christopher Kenna	14 years
wip-timer-trace	bugfix: C-EDF, clear scheduled field of the correct CPU upon task_exit	Andrea Bastoni	14 years
wip-tracepoints	Add kernel-style events for sched_trace_XXX() functions	Andrea Bastoni	13 years

Tag	Download	Author	Age
2015.1	commit 8e51b37822...	Bjoern Brandenburg	10 years
2013.1	commit bcaacec1ca...	Glenn Elliott	12 years
2012.3	commit c158b5fbe4...	Jonathan Herman	12 years
2012.2	commit b53c479a0f...	Glenn Elliott	13 years
2012.1	commit 83b11ea1c6...	Bjoern B. Brandenburg	13 years
rtas12-mc-beta-exp	commit 8e236ee20f...	Christopher Kenna	13 years
2011.1	commit d11808b5c6...	Christopher Kenna	14 years
v2.6.37-rc4	commit e8a7e48bb2...	Linus Torvalds	14 years
v2.6.37-rc3	commit 3561d43fd2...	Linus Torvalds	14 years
v2.6.37-rc2	commit e53beacd23...	Linus Torvalds	14 years
v2.6.37-rc1	commit c8ddb2713c...	Linus Torvalds	14 years
v2.6.36	commit f6f94e2ab1...	Linus Torvalds	14 years
2010.2	commit 5c5456402d...	Bjoern B. Brandenburg	14 years
v2.6.36-rc8	commit cd07202cc8...	Linus Torvalds	14 years
v2.6.36-rc7	commit cb655d0f3d...	Linus Torvalds	14 years
v2.6.36-rc6	commit 899611ee7d...	Linus Torvalds	14 years
v2.6.36-rc5	commit b30a3f6257...	Linus Torvalds	14 years
v2.6.36-rc4	commit 49553c2ef8...	Linus Torvalds	14 years
v2.6.36-rc3	commit 2bfc96a127...	Linus Torvalds	14 years
v2.6.36-rc2	commit 76be97c1fc...	Linus Torvalds	14 years
v2.6.36-rc1	commit da5cabf80e...	Linus Torvalds	15 years
v2.6.35	commit 9fe6206f40...	Linus Torvalds	15 years
v2.6.35-rc6	commit b37fa16e78...	Linus Torvalds	15 years
v2.6.35-rc5	commit 1c5474a65b...	Linus Torvalds	15 years
v2.6.35-rc4	commit 815c4163b6...	Linus Torvalds	15 years
v2.6.35-rc3	commit 7e27d6e778...	Linus Torvalds	15 years
v2.6.35-rc2	commit e44a21b726...	Linus Torvalds	15 years
v2.6.35-rc1	commit 67a3e12b05...	Linus Torvalds	15 years
2010.1	commit 7c1ff4c544...	Andrea Bastoni	15 years
v2.6.34	commit e40152ee1e...	Linus Torvalds	15 years
v2.6.33.4	commit 4640b4e7d9...	Greg Kroah-Hartman	15 years
v2.6.34-rc7	commit b57f95a382...	Linus Torvalds	15 years
v2.6.34-rc6	commit 66f41d4c5c...	Linus Torvalds	15 years
v2.6.33.3	commit 3e7ad8ed97...	Greg Kroah-Hartman	15 years
v2.6.34-rc5	commit 01bf0b6457...	Linus Torvalds	15 years
v2.6.34-rc4	commit 0d0fb0f9c5...	Linus Torvalds	15 years
v2.6.33.2	commit 19f00f070c...	Greg Kroah-Hartman	15 years
v2.6.34-rc3	commit 2eaa9cfdf3...	Linus Torvalds	15 years
v2.6.34-rc2	commit 220bf991b0...	Linus Torvalds	15 years
v2.6.33.1	commit dbdafe5ccf...	Greg Kroah-Hartman	15 years
v2.6.34-rc1	commit 57d54889cd...	Linus Torvalds	15 years
v2.6.33	commit 60b341b778...	Linus Torvalds	15 years
v2.6.33-rc8	commit 724e6d3fe8...	Linus Torvalds	15 years
v2.6.33-rc7	commit 29275254ca...	Linus Torvalds	15 years
v2.6.33-rc6	commit abe94c756c...	Linus Torvalds	15 years
v2.6.33-rc5	commit 92dcffb916...	Linus Torvalds	15 years
v2.6.33-rc4	commit 7284ce6c9f...	Linus Torvalds	15 years
v2.6.33-rc3	commit 74d2e4f8d7...	Linus Torvalds	15 years
v2.6.33-rc2	commit 6b7b284958...	Linus Torvalds	15 years
v2.6.33-rc1	commit 55639353a0...	Linus Torvalds	15 years
v2.6.32	commit 22763c5cf3...	Linus Torvalds	15 years
v2.6.32-rc8	commit 648f4e3e50...	Linus Torvalds	15 years
v2.6.32-rc7	commit 156171c71a...	Linus Torvalds	15 years
v2.6.32-rc6	commit b419148e56...	Linus Torvalds	15 years
v2.6.32-rc5	commit 012abeea66...	Linus Torvalds	15 years
v2.6.32-rc4	commit 161291396e...	Linus Torvalds	15 years
v2.6.32-rc3	commit 374576a8b6...	Linus Torvalds	15 years
v2.6.32-rc1	commit 17d857be64...	Linus Torvalds	15 years
v2.6.32-rc2	commit 17d857be64...	Linus Torvalds	15 years
v2.6.31	commit 74fca6a428...	Linus Torvalds	15 years
v2.6.31-rc9	commit e07cccf404...	Linus Torvalds	15 years
v2.6.31-rc8	commit 326ba5010a...	Linus Torvalds	15 years
v2.6.31-rc7	commit 422bef879e...	Linus Torvalds	15 years
v2.6.31-rc6	commit 64f1607ffb...	Linus Torvalds	16 years
v2.6.31-rc5	commit ed680c4ad4...	Linus Torvalds	16 years
v2.6.31-rc4	commit 4be3bd7849...	Linus Torvalds	16 years
v2.6.31-rc3	commit 6847e154e3...	Linus Torvalds	16 years
v2.6.31-rc2	commit 8e4a718ff3...	Linus Torvalds	16 years
v2.6.31-rc1	commit 28d0325ce6...	Linus Torvalds	16 years
v2.6.30	commit 07a2039b8e...	Linus Torvalds	16 years
v2.6.30-rc8	commit 9fa7eb283c...	Linus Torvalds	16 years
v2.6.30-rc7	commit 59a3759d0f...	Linus Torvalds	16 years
v2.6.30-rc6	commit 1406de8e11...	Linus Torvalds	16 years
v2.6.30-rc5	commit 091bf7624d...	Linus Torvalds	16 years
v2.6.30-rc4	commit 091438dd56...	Linus Torvalds	16 years
v2.6.30-rc3	commit 0910697403...	Linus Torvalds	16 years
v2.6.30-rc2	commit 0882e8dd3a...	Linus Torvalds	16 years
v2.6.30-rc1	commit 577c9c456f...	Linus Torvalds	16 years
v2.6.29	commit 8e0ee43bc2...	Linus Torvalds	16 years
v2.6.29-rc8	commit 041b62374c...	Linus Torvalds	16 years
v2.6.29-rc7	commit fec6c6fec3...	Linus Torvalds	16 years
v2.6.29-rc6	commit 20f4d6c3a2...	Linus Torvalds	16 years
v2.6.29-rc5	commit d2f8d7ee1a...	Linus Torvalds	16 years
v2.6.29-rc4	commit 8e4921515c...	Linus Torvalds	16 years
v2.6.29-rc3	commit 18e352e4a7...	Linus Torvalds	16 years
v2.6.29-rc2	commit 1de9e8e70f...	Linus Torvalds	16 years
v2.6.29-rc1	commit c59765042f...	Linus Torvalds	16 years
v2.6.28	commit 4a6908a3a0...	Linus Torvalds	16 years
v2.6.28-rc9	commit 929096fe9f...	Linus Torvalds	16 years
v2.6.28-rc8	commit 8b1fae4e42...	Linus Torvalds	16 years
v2.6.28-rc7	commit 061e41fdb5...	Linus Torvalds	16 years
v2.6.28-rc6	commit 13d428afc0...	Linus Torvalds	16 years
v2.6.28-rc5	commit 9bf1a2445f...	Linus Torvalds	16 years
v2.6.28-rc4	commit f7160c7573...	Linus Torvalds	16 years
v2.6.28-rc3	commit 45beca08dd...	Linus Torvalds	16 years
v2.6.28-rc2	commit 0173a3265b...	Linus Torvalds	16 years
v2.6.28-rc1	commit 57f8f7b60d...	Linus Torvalds	16 years
v2.6.27	commit 3fa8749e58...	Linus Torvalds	16 years
v2.6.27-rc9	commit 4330ed8ed4...	Linus Torvalds	16 years
v2.6.27-rc8	commit 94aca1dac6...	Linus Torvalds	16 years
v2.6.27-rc7	commit 72d31053f6...	Linus Torvalds	16 years
v2.6.27-rc6	commit adee14b2e1...	Linus Torvalds	16 years
v2.6.27-rc5	commit 24342c34a0...	Linus Torvalds	16 years
v2.6.27-rc4	commit 6a55617ed5...	Linus Torvalds	17 years
v2.6.27-rc3	commit 30a2f3c60a...	Linus Torvalds	17 years
v2.6.27-rc2	commit 0967d61ea0...	Linus Torvalds	17 years
v2.6.27-rc1	commit 6e86841d05...	Linus Torvalds	17 years
v2.6.26	commit bce7f793da...	Linus Torvalds	17 years
v2.6.26-rc9	commit b7279469d6...	Linus Torvalds	17 years
v2.6.26-rc8	commit 543cf4cb3f...	Linus Torvalds	17 years
v2.6.26-rc7	commit d70ac829b7...	Linus Torvalds	17 years
v2.6.26-rc6	commit 5dd34572ad...	Linus Torvalds	17 years
v2.6.26-rc5	commit 53c8ba9540...	Linus Torvalds	17 years
v2.6.26-rc4	commit e490517a03...	Linus Torvalds	17 years
v2.6.26-rc3	commit b8291ad07a...	Linus Torvalds	17 years
v2.6.26-rc2	commit 492c2e476e...	Linus Torvalds	17 years
v2.6.26-rc1	commit 2ddcca36c8...	Linus Torvalds	17 years
v2.6.25	commit 4b119e21d0...	Linus Torvalds	17 years
v2.6.25-rc9	commit 120dd64cac...	Linus Torvalds	17 years
v2.6.25-rc8	commit 0e81a8ae37...	Linus Torvalds	17 years
v2.6.25-rc7	commit 05dda977f2...	Linus Torvalds	17 years
v2.6.25-rc6	commit a978b30af3...	Linus Torvalds	17 years
v2.6.25-rc5	commit cdeeeae056...	Linus Torvalds	17 years
v2.6.25-rc4	commit 29e8c3c304...	Linus Torvalds	17 years
v2.6.25-rc3	commit bfa274e243...	Linus Torvalds	17 years
v2.6.25-rc2	commit 101142c37b...	Linus Torvalds	17 years
v2.6.25-rc1	commit 19af35546d...	Linus Torvalds	17 years
v2.6.24	commit 49914084e7...	Linus Torvalds	17 years
v2.6.24-rc8	commit cbd9c88369...	Linus Torvalds	17 years
v2.6.24-rc7	commit 3ce5445046...	Linus Torvalds	17 years
v2.6.24-rc6	commit ea67db4cdb...	Linus Torvalds	17 years
v2.6.24-rc5	commit 82d29bf6dc...	Linus Torvalds	17 years
v2.6.24-rc4	commit 09b56adc98...	Linus Torvalds	17 years
v2.6.24-rc3	commit d9f8bcbf67...	Linus Torvalds	17 years
v2.6.24-rc2	commit dbeeb816e8...	Linus Torvalds	17 years
v2.6.24-rc1	commit c9927c2bf4...	Linus Torvalds	17 years
v2.6.23	commit bbf25010f1...	Linus Torvalds	17 years
v2.6.23-rc9	commit 3146b39c18...	Linus Torvalds	17 years
v2.6.23-rc8	commit 4942de4a0e...	Linus Torvalds	17 years
v2.6.23-rc7	commit 81cfe79b9c...	Linus Torvalds	17 years
v2.6.23-rc6	commit 0d4cbb5e7f...	Linus Torvalds	17 years
v2.6.23-rc5	commit 40ffbfad6b...	Linus Torvalds	17 years
v2.6.23-rc4	commit b07d68b5ca...	Linus Torvalds	17 years
v2.6.23-rc3	commit 39d3520c92...	Linus Torvalds	18 years
v2.6.23-rc2	commit d4ac2477fa...	Linus Torvalds	18 years
v2.6.23-rc1	commit f695baf2df...	Linus Torvalds	18 years
v2.6.22	commit 7dcca30a32...	Linus Torvalds	18 years
v2.6.22-rc7	commit a38d6181ff...	Linus Torvalds	18 years
v2.6.22-rc6	commit 189548642c...	Linus Torvalds	18 years
v2.6.22-rc5	commit 188e1f81ba...	Linus Torvalds	18 years
v2.6.22-rc4	commit 5ecd3100e6...	Linus Torvalds	18 years
v2.6.22-rc3	commit c420bc9f09...	Linus Torvalds	18 years
v2.6.22-rc2	commit 55b637c6a0...	Linus Torvalds	18 years
v2.6.22-rc1	commit 39403865d2...	Linus Torvalds	18 years
v2.6.21	commit de46c33745...	Linus Torvalds	18 years
v2.6.21-rc7	commit 94a05509a9...	Linus Torvalds	18 years
v2.6.21-rc6	commit a21bd69e15...	Linus Torvalds	18 years
v2.6.21-rc5	commit e0f2e3a06b...	Linus Torvalds	18 years
v2.6.21-rc4	commit db98e0b434...	Linus Torvalds	18 years
v2.6.21-rc3	commit 08e15e81a4...	Linus Torvalds	18 years
v2.6.21-rc2	commit 606135a308...	Linus Torvalds	18 years
v2.6.21-rc1	commit c8f71b01a5...	Linus Torvalds	18 years
v2.6.20	commit 62d0cfcb27...	Linus Torvalds	18 years
v2.6.20-rc7	commit f56df2f4db...	Linus Torvalds	18 years
v2.6.20-rc6	commit 99abfeafb5...	Linus Torvalds	18 years
v2.6.20-rc5	commit a8b3485287...	Linus Torvalds	18 years
v2.6.20-rc4	commit bf81b46482...	Linus Torvalds	18 years
v2.6.20-rc3	commit 669df1b478...	Linus Torvalds	18 years
v2.6.20-rc2	commit 3bf8ba38f3...	Linus Torvalds	18 years
v2.6.20-rc1	commit cc016448b0...	Linus Torvalds	18 years
v2.6.19	commit 0215ffb08c...	Linus Torvalds	18 years
v2.6.19-rc6	commit 44597f65f6...	Linus Torvalds	18 years
v2.6.19-rc5	commit 80c2188127...	Linus Torvalds	18 years
v2.6.19-rc4	commit ae99a78af3...	Linus Torvalds	18 years
v2.6.19-rc3	commit 7059abedd2...	Linus Torvalds	18 years
v2.6.19-rc2	commit b4bd8c6643...	Linus Torvalds	18 years
v2.6.19-rc1	commit d223a60106...	Linus Torvalds	18 years
v2.6.18	commit e478bec0ba...	Linus Torvalds	18 years
v2.6.18-rc7	commit 95064a75eb...	Linus Torvalds	18 years
v2.6.18-rc6	commit c336923b66...	Linus Torvalds	18 years
v2.6.18-rc5	commit 60d4684068...	Linus Torvalds	18 years
v2.6.18-rc4	commit 9f737633e6...	Linus Torvalds	19 years
v2.6.18-rc3	commit b6ff50833a...	Linus Torvalds	19 years
v2.6.18-rc2	commit 82d6897fef...	Linus Torvalds	19 years
v2.6.18-rc1	commit 120bda20c6...	Linus Torvalds	19 years
v2.6.17	commit 427abfa28a...	Linus Torvalds	19 years
v2.6.17-rc6	commit 1def630a6a...	Linus Torvalds	19 years
v2.6.17-rc5	commit a8bd60705a...	Linus Torvalds	19 years
v2.6.17-rc4	commit d8c3291c73...	Linus Torvalds	19 years
v2.6.17-rc3	commit 2be4d50295...	Linus Torvalds	19 years
v2.6.17-rc2	commit 8bbde0e6d5...	Linus Torvalds	19 years
v2.6.17-rc1	commit 6246b6128b...	Linus Torvalds	19 years
v2.6.16	commit 7705a8792b...	Linus Torvalds	19 years
v2.6.16-rc6	commit 535744878e...	Linus Torvalds	19 years
v2.6.16-rc5	commit b9a33cebac...	Linus Torvalds	19 years
v2.6.16-rc4	commit bd71c2b174...	Linus Torvalds	19 years
v2.6.16-rc3	commit e9bb4c9929...	Linus Torvalds	19 years
v2.6.16-rc2	commit 826eeb53a6...	Linus Torvalds	19 years
v2.6.16-rc1	commit 2664b25051...	Linus Torvalds	19 years
v2.6.15	commit 88026842b0...	Linus Torvalds	19 years
v2.6.15-rc7	commit f89f5948fc...	Linus Torvalds	19 years
v2.6.15-rc6	commit df7addbb45...	Linus Torvalds	19 years
v2.6.15-rc5	commit 436b0f76f2...	Linus Torvalds	19 years
v2.6.15-rc4	commit 5666c0947e...	Linus Torvalds	19 years
v2.6.15-rc3	commit 624f54be20...	Linus Torvalds	19 years
v2.6.15-rc2	commit 3bedff1d73...	Linus Torvalds	19 years
v2.6.15-rc1	commit cd52d1ee9a...	Linus Torvalds	19 years
v2.6.14	commit 741b2252a5...	Linus Torvalds	19 years
v2.6.14-rc5	commit 93918e9afc...	Linus Torvalds	19 years
v2.6.14-rc4	commit 907a426179...	Linus Torvalds	19 years
v2.6.14-rc3	commit 1c9426e8a5...	Linus Torvalds	19 years
v2.6.14-rc2	commit 676d55ae30...	Linus Torvalds	19 years
v2.6.14-rc1	commit 2f4ba45a75...	Linus Torvalds	19 years
v2.6.13	commit 02b3e4e2d7...	Linus Torvalds	19 years
v2.6.13-rc7	commit 0572e3da3f...	Linus Torvalds	19 years
v2.6.13-rc6	commit 6fc32179de...	Linus Torvalds	20 years
v2.6.13-rc5	commit 9a351e30d7...	Linus Torvalds	20 years
v2.6.13-rc4	commit 6395352334...	Linus Torvalds	20 years
v2.6.11	tree c39ae07f39...
v2.6.11-tree	tree c39ae07f39...
v2.6.12	commit 9ee1c939d1...
v2.6.12-rc2	commit 1da177e4c3...
v2.6.12-rc3	commit a2755a80f4...
v2.6.12-rc4	commit 88d7bd8cb9...
v2.6.12-rc5	commit 2a24ab628a...
v2.6.12-rc6	commit 7cef5677ef...
v2.6.13-rc1	commit 4c91aedb75...
v2.6.13-rc2	commit a18bcb7450...
v2.6.13-rc3	commit c32511e271...

/* * Performance counter core code * * Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de> * Copyright (C) 2008-2009 Red Hat, Inc., Ingo Molnar * Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> * Copyright � 2009 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com> * * For licensing details see kernel-base/COPYING */ #include <linux/fs.h> #include <linux/mm.h> #include <linux/cpu.h> #include <linux/smp.h> #include <linux/file.h> #include <linux/poll.h> #include <linux/sysfs.h> #include <linux/dcache.h> #include <linux/percpu.h> #include <linux/ptrace.h> #include <linux/vmstat.h> #include <linux/hardirq.h> #include <linux/rculist.h> #include <linux/uaccess.h> #include <linux/syscalls.h> #include <linux/anon_inodes.h> #include <linux/kernel_stat.h> #include <linux/perf_counter.h> #include <asm/irq_regs.h> /* * Each CPU has a list of per CPU counters: */ DEFINE_PER_CPU(struct perf_cpu_context, perf_cpu_context); int perf_max_counters __read_mostly = 1; static int perf_reserved_percpu __read_mostly; static int perf_overcommit __read_mostly = 1; static atomic_t nr_counters __read_mostly; static atomic_t nr_mmap_counters __read_mostly; static atomic_t nr_comm_counters __read_mostly; /* * perf counter paranoia level: * 0 - not paranoid * 1 - disallow cpu counters to unpriv * 2 - disallow kernel profiling to unpriv */ int sysctl_perf_counter_paranoid __read_mostly; static inline bool perf_paranoid_cpu(void) { return sysctl_perf_counter_paranoid > 0; } static inline bool perf_paranoid_kernel(void) { return sysctl_perf_counter_paranoid > 1; } int sysctl_perf_counter_mlock __read_mostly = 512; /* 'free' kb per user */ /* * max perf counter sample rate */ int sysctl_perf_counter_sample_rate __read_mostly = 100000; static atomic64_t perf_counter_id; /* * Lock for (sysadmin-configurable) counter reservations: */ static DEFINE_SPINLOCK(perf_resource_lock); /* * Architecture provided APIs - weak aliases: */ extern __weak const struct pmu *hw_perf_counter_init(struct perf_counter *counter) { return NULL; } void __weak hw_perf_disable(void) { barrier(); } void __weak hw_perf_enable(void) { barrier(); } void __weak hw_perf_counter_setup(int cpu) { barrier(); } int __weak hw_perf_group_sched_in(struct perf_counter *group_leader, struct perf_cpu_context *cpuctx, struct perf_counter_context *ctx, int cpu) { return 0; } void __weak perf_counter_print_debug(void) { } static DEFINE_PER_CPU(int, disable_count); void __perf_disable(void) { __get_cpu_var(disable_count)++; } bool __perf_enable(void) { return !--__get_cpu_var(disable_count); } void perf_disable(void) { __perf_disable(); hw_perf_disable(); } void perf_enable(void) { if (__perf_enable()) hw_perf_enable(); } static void get_ctx(struct perf_counter_context *ctx) { WARN_ON(!atomic_inc_not_zero(&ctx->refcount)); } static void free_ctx(struct rcu_head *head) { struct perf_counter_context *ctx; ctx = container_of(head, struct perf_counter_context, rcu_head); kfree(ctx); } static void put_ctx(struct perf_counter_context *ctx) { if (atomic_dec_and_test(&ctx->refcount)) { if (ctx->parent_ctx) put_ctx(ctx->parent_ctx); if (ctx->task) put_task_struct(ctx->task); call_rcu(&ctx->rcu_head, free_ctx); } } static void unclone_ctx(struct perf_counter_context *ctx) { if (ctx->parent_ctx) { put_ctx(ctx->parent_ctx); ctx->parent_ctx = NULL; } } /* * If we inherit counters we want to return the parent counter id * to userspace. */ static u64 primary_counter_id(struct perf_counter *counter) { u64 id = counter->id; if (counter->parent) id = counter->parent->id; return id; } /* * Get the perf_counter_context for a task and lock it. * This has to cope with with the fact that until it is locked, * the context could get moved to another task. */ static struct perf_counter_context * perf_lock_task_context(struct task_struct *task, unsigned long *flags) { struct perf_counter_context *ctx; rcu_read_lock(); retry: ctx = rcu_dereference(task->perf_counter_ctxp); if (ctx) { /* * If this context is a clone of another, it might * get swapped for another underneath us by * perf_counter_task_sched_out, though the * rcu_read_lock() protects us from any context * getting freed. Lock the context and check if it * got swapped before we could get the lock, and retry * if so. If we locked the right context, then it * can't get swapped on us any more. */ spin_lock_irqsave(&ctx->lock, *flags); if (ctx != rcu_dereference(task->perf_counter_ctxp)) { spin_unlock_irqrestore(&ctx->lock, *flags); goto retry; } if (!atomic_inc_not_zero(&ctx->refcount)) { spin_unlock_irqrestore(&ctx->lock, *flags); ctx = NULL; } } rcu_read_unlock(); return ctx; } /* * Get the context for a task and increment its pin_count so it * can't get swapped to another task. This also increments its * reference count so that the context can't get freed. */ static struct perf_counter_context *perf_pin_task_context(struct task_struct *task) { struct perf_counter_context *ctx; unsigned long flags; ctx = perf_lock_task_context(task, &flags); if (ctx) { ++ctx->pin_count; spin_unlock_irqrestore(&ctx->lock, flags); } return ctx; } static void perf_unpin_context(struct perf_counter_context *ctx) { unsigned long flags; spin_lock_irqsave(&ctx->lock, flags); --ctx->pin_count; spin_unlock_irqrestore(&ctx->lock, flags); put_ctx(ctx); } /* * Add a counter from the lists for its context. * Must be called with ctx->mutex and ctx->lock held. */ static void list_add_counter(struct perf_counter *counter, struct perf_counter_context *ctx) { struct perf_counter *group_leader = counter->group_leader; /* * Depending on whether it is a standalone or sibling counter, * add it straight to the context's counter list, or to the group * leader's sibling list: */ if (group_leader == counter) list_add_tail(&counter->list_entry, &ctx->counter_list); else { list_add_tail(&counter->list_entry, &group_leader->sibling_list); group_leader->nr_siblings++; } list_add_rcu(&counter->event_entry, &ctx->event_list); ctx->nr_counters++; if (counter->attr.inherit_stat) ctx->nr_stat++; } /* * Remove a counter from the lists for its context. * Must be called with ctx->mutex and ctx->lock held. */ static void list_del_counter(struct perf_counter *counter, struct perf_counter_context *ctx) { struct perf_counter *sibling, *tmp; if (list_empty(&counter->list_entry)) return; ctx->nr_counters--; if (counter->attr.inherit_stat) ctx->nr_stat--; list_del_init(&counter->list_entry); list_del_rcu(&counter->event_entry); if (counter->group_leader != counter) counter->group_leader->nr_siblings--; /* * If this was a group counter with sibling counters then * upgrade the siblings to singleton counters by adding them * to the context list directly: */ list_for_each_entry_safe(sibling, tmp, &counter->sibling_list, list_entry) { list_move_tail(&sibling->list_entry, &ctx->counter_list); sibling->group_leader = sibling; } } static void counter_sched_out(struct perf_counter *counter, struct perf_cpu_context *cpuctx, struct perf_counter_context *ctx) { if (counter->state != PERF_COUNTER_STATE_ACTIVE) return; counter->state = PERF_COUNTER_STATE_INACTIVE; counter->tstamp_stopped = ctx->time; counter->pmu->disable(counter); counter->oncpu = -1; if (!is_software_counter(counter)) cpuctx->active_oncpu--; ctx->nr_active--; if (counter->attr.exclusive || !cpuctx->active_oncpu) cpuctx->exclusive = 0; } static void group_sched_out(struct perf_counter *group_counter, struct perf_cpu_context *cpuctx, struct perf_counter_context *ctx) { struct perf_counter *counter; if (group_counter->state != PERF_COUNTER_STATE_ACTIVE) return; counter_sched_out(group_counter, cpuctx, ctx); /* * Schedule out siblings (if any): */ list_for_each_entry(counter, &group_counter->sibling_list, list_entry) counter_sched_out(counter, cpuctx, ctx); if (group_counter->attr.exclusive) cpuctx->exclusive = 0; } /* * Cross CPU call to remove a performance counter * * We disable the counter on the hardware level first. After that we * remove it from the context list. */ static void __perf_counter_remove_from_context(void *info) { struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_counter *counter = info; struct perf_counter_context *ctx = counter->ctx; /* * If this is a task context, we need to check whether it is * the current task context of this cpu. If not it has been * scheduled out before the smp call arrived. */ if (ctx->task && cpuctx->task_ctx != ctx) return; spin_lock(&ctx->lock); /* * Protect the list operation against NMI by disabling the * counters on a global level. */ perf_disable(); counter_sched_out(counter, cpuctx, ctx); list_del_counter(counter, ctx); if (!ctx->task) { /* * Allow more per task counters with respect to the * reservation: */ cpuctx->max_pertask = min(perf_max_counters - ctx->nr_counters, perf_max_counters - perf_reserved_percpu); } perf_enable(); spin_unlock(&ctx->lock); } /* * Remove the counter from a task's (or a CPU's) list of counters. * * Must be called with ctx->mutex held. * * CPU counters are removed with a smp call. For task counters we only * call when the task is on a CPU. * * If counter->ctx is a cloned context, callers must make sure that * every task struct that counter->ctx->task could possibly point to * remains valid. This is OK when called from perf_release since * that only calls us on the top-level context, which can't be a clone. * When called from perf_counter_exit_task, it's OK because the * context has been detached from its task. */ static void perf_counter_remove_from_context(struct perf_counter *counter) { struct perf_counter_context *ctx = counter->ctx; struct task_struct *task = ctx->task; if (!task) { /* * Per cpu counters are removed via an smp call and * the removal is always sucessful. */ smp_call_function_single(counter->cpu, __perf_counter_remove_from_context, counter, 1); return; } retry: task_oncpu_function_call(task, __perf_counter_remove_from_context, counter); spin_lock_irq(&ctx->lock); /* * If the context is active we need to retry the smp call. */ if (ctx->nr_active && !list_empty(&counter->list_entry)) { spin_unlock_irq(&ctx->lock); goto retry; } /* * The lock prevents that this context is scheduled in so we * can remove the counter safely, if the call above did not * succeed. */ if (!list_empty(&counter->list_entry)) { list_del_counter(counter, ctx); } spin_unlock_irq(&ctx->lock); } static inline u64 perf_clock(void) { return cpu_clock(smp_processor_id()); } /* * Update the record of the current time in a context. */ static void update_context_time(struct perf_counter_context *ctx) { u64 now = perf_clock(); ctx->time += now - ctx->timestamp; ctx->timestamp = now; } /* * Update the total_time_enabled and total_time_running fields for a counter. */ static void update_counter_times(struct perf_counter *counter) { struct perf_counter_context *ctx = counter->ctx; u64 run_end; if (counter->state < PERF_COUNTER_STATE_INACTIVE) return; counter->total_time_enabled = ctx->time - counter->tstamp_enabled; if (counter->state == PERF_COUNTER_STATE_INACTIVE) run_end = counter->tstamp_stopped; else run_end = ctx->time; counter->total_time_running = run_end - counter->tstamp_running; } /* * Update total_time_enabled and total_time_running for all counters in a group. */ static void update_group_times(struct perf_counter *leader) { struct perf_counter *counter; update_counter_times(leader); list_for_each_entry(counter, &leader->sibling_list, list_entry) update_counter_times(counter); } /* * Cross CPU call to disable a performance counter */ static void __perf_counter_disable(void *info) { struct perf_counter *counter = info; struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_counter_context *ctx = counter->ctx; /* * If this is a per-task counter, need to check whether this * counter's task is the current task on this cpu. */ if (ctx->task && cpuctx->task_ctx != ctx) return; spin_lock(&ctx->lock); /* * If the counter is on, turn it off. * If it is in error state, leave it in error state. */ if (counter->state >= PERF_COUNTER_STATE_INACTIVE) { update_context_time(ctx); update_counter_times(counter); if (counter == counter->group_leader) group_sched_out(counter, cpuctx, ctx); else counter_sched_out(counter, cpuctx, ctx); counter->state = PERF_COUNTER_STATE_OFF; } spin_unlock(&ctx->lock); } /* * Disable a counter. * * If counter->ctx is a cloned context, callers must make sure that * every task struct that counter->ctx->task could possibly point to * remains valid. This condition is satisifed when called through * perf_counter_for_each_child or perf_counter_for_each because they * hold the top-level counter's child_mutex, so any descendant that * goes to exit will block in sync_child_counter. * When called from perf_pending_counter it's OK because counter->ctx * is the current context on this CPU and preemption is disabled, * hence we can't get into perf_counter_task_sched_out for this context. */ static void perf_counter_disable(struct perf_counter *counter) { struct perf_counter_context *ctx = counter->ctx; struct task_struct *task = ctx->task; if (!task) { /* * Disable the counter on the cpu that it's on */ smp_call_function_single(counter->cpu, __perf_counter_disable, counter, 1); return; } retry: task_oncpu_function_call(task, __perf_counter_disable, counter); spin_lock_irq(&ctx->lock); /* * If the counter is still active, we need to retry the cross-call. */ if (counter->state == PERF_COUNTER_STATE_ACTIVE) { spin_unlock_irq(&ctx->lock); goto retry; } /* * Since we have the lock this context can't be scheduled * in, so we can change the state safely. */ if (counter->state == PERF_COUNTER_STATE_INACTIVE) { update_counter_times(counter); counter->state = PERF_COUNTER_STATE_OFF; } spin_unlock_irq(&ctx->lock); } static int counter_sched_in(struct perf_counter *counter, struct perf_cpu_context *cpuctx, struct perf_counter_context *ctx, int cpu) { if (counter->state <= PERF_COUNTER_STATE_OFF) return 0; counter->state = PERF_COUNTER_STATE_ACTIVE; counter->oncpu = cpu; /* TODO: put 'cpu' into cpuctx->cpu */ /* * The new state must be visible before we turn it on in the hardware: */ smp_wmb(); if (counter->pmu->enable(counter)) { counter->state = PERF_COUNTER_STATE_INACTIVE; counter->oncpu = -1; return -EAGAIN; } counter->tstamp_running += ctx->time - counter->tstamp_stopped; if (!is_software_counter(counter)) cpuctx->active_oncpu++; ctx->nr_active++; if (counter->attr.exclusive) cpuctx->exclusive = 1; return 0; } static int group_sched_in(struct perf_counter *group_counter, struct perf_cpu_context *cpuctx, struct perf_counter_context *ctx, int cpu) { struct perf_counter *counter, *partial_group; int ret; if (group_counter->state == PERF_COUNTER_STATE_OFF) return 0; ret = hw_perf_group_sched_in(group_counter, cpuctx, ctx, cpu); if (ret) return ret < 0 ? ret : 0; if (counter_sched_in(group_counter, cpuctx, ctx, cpu)) return -EAGAIN; /* * Schedule in siblings as one group (if any): */ list_for_each_entry(counter, &group_counter->sibling_list, list_entry) { if (counter_sched_in(counter, cpuctx, ctx, cpu)) { partial_group = counter; goto group_error; } } return 0; group_error: /* * Groups can be scheduled in as one unit only, so undo any * partial group before returning: */ list_for_each_entry(counter, &group_counter->sibling_list, list_entry) { if (counter == partial_group) break; counter_sched_out(counter, cpuctx, ctx); } counter_sched_out(group_counter, cpuctx, ctx); return -EAGAIN; } /* * Return 1 for a group consisting entirely of software counters, * 0 if the group contains any hardware counters. */ static int is_software_only_group(struct perf_counter *leader) { struct perf_counter *counter; if (!is_software_counter(leader)) return 0; list_for_each_entry(counter, &leader->sibling_list, list_entry) if (!is_software_counter(counter)) return 0; return 1; } /* * Work out whether we can put this counter group on the CPU now. */ static int group_can_go_on(struct perf_counter *counter, struct perf_cpu_context *cpuctx, int can_add_hw) { /* * Groups consisting entirely of software counters can always go on. */ if (is_software_only_group(counter)) return 1; /* * If an exclusive group is already on, no other hardware * counters can go on. */ if (cpuctx->exclusive) return 0; /* * If this group is exclusive and there are already * counters on the CPU, it can't go on. */ if (counter->attr.exclusive && cpuctx->active_oncpu) return 0; /* * Otherwise, try to add it if all previous groups were able * to go on. */ return can_add_hw; } static void add_counter_to_ctx(struct perf_counter *counter, struct perf_counter_context *ctx) { list_add_counter(counter, ctx); counter->tstamp_enabled = ctx->time; counter->tstamp_running = ctx->time; counter->tstamp_stopped = ctx->time; } /* * Cross CPU call to install and enable a performance counter * * Must be called with ctx->mutex held */ static void __perf_install_in_context(void *info) { struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_counter *counter = info; struct perf_counter_context *ctx = counter->ctx; struct perf_counter *leader = counter->group_leader; int cpu = smp_processor_id(); int err; /* * If this is a task context, we need to check whether it is * the current task context of this cpu. If not it has been * scheduled out before the smp call arrived. * Or possibly this is the right context but it isn't * on this cpu because it had no counters. */ if (ctx->task && cpuctx->task_ctx != ctx) { if (cpuctx->task_ctx || ctx->task != current) return; cpuctx->task_ctx = ctx; } spin_lock(&ctx->lock); ctx->is_active = 1; update_context_time(ctx); /* * Protect the list operation against NMI by disabling the * counters on a global level. NOP for non NMI based counters. */ perf_disable(); add_counter_to_ctx(counter, ctx); /* * Don't put the counter on if it is disabled or if * it is in a group and the group isn't on. */ if (counter->state != PERF_COUNTER_STATE_INACTIVE || (leader != counter && leader->state != PERF_COUNTER_STATE_ACTIVE)) goto unlock; /* * An exclusive counter can't go on if there are already active * hardware counters, and no hardware counter can go on if there * is already an exclusive counter on. */ if (!group_can_go_on(counter, cpuctx, 1)) err = -EEXIST; else err = counter_sched_in(counter, cpuctx, ctx, cpu); if (err) { /* * This counter couldn't go on. If it is in a group * then we have to pull the whole group off. * If the counter group is pinned then put it in error state. */ if (leader != counter) group_sched_out(leader, cpuctx, ctx); if (leader->attr.pinned) { update_group_times(leader); leader->state = PERF_COUNTER_STATE_ERROR; } } if (!err && !ctx->task && cpuctx->max_pertask) cpuctx->max_pertask--; unlock: perf_enable(); spin_unlock(&ctx->lock); } /* * Attach a performance counter to a context * * First we add the counter to the list with the hardware enable bit * in counter->hw_config cleared. * * If the counter is attached to a task which is on a CPU we use a smp * call to enable it in the task context. The task might have been * scheduled away, but we check this in the smp call again. * * Must be called with ctx->mutex held. */ static void perf_install_in_context(struct perf_counter_context *ctx, struct perf_counter *counter, int cpu) { struct task_struct *task = ctx->task; if (!task) { /* * Per cpu counters are installed via an smp call and * the install is always sucessful. */ smp_call_function_single(cpu, __perf_install_in_context, counter, 1); return; } retry: task_oncpu_function_call(task, __perf_install_in_context, counter); spin_lock_irq(&ctx->lock); /* * we need to retry the smp call. */ if (ctx->is_active && list_empty(&counter->list_entry)) { spin_unlock_irq(&ctx->lock); goto retry; } /* * The lock prevents that this context is scheduled in so we * can add the counter safely, if it the call above did not * succeed. */ if (list_empty(&counter->list_entry)) add_counter_to_ctx(counter, ctx); spin_unlock_irq(&ctx->lock); } /* * Cross CPU call to enable a performance counter */ static void __perf_counter_enable(void *info) { struct perf_counter *counter = info; struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_counter_context *ctx = counter->ctx; struct perf_counter *leader = counter->group_leader; int err; /* * If this is a per-task counter, need to check whether this * counter's task is the current task on this cpu. */ if (ctx->task && cpuctx->task_ctx != ctx) { if (cpuctx->task_ctx || ctx->task != current) return; cpuctx->task_ctx = ctx; } spin_lock(&ctx->lock); ctx->is_active = 1; update_context_time(ctx); if (counter->state >= PERF_COUNTER_STATE_INACTIVE) goto unlock; counter->state = PERF_COUNTER_STATE_INACTIVE; counter->tstamp_enabled = ctx->time - counter->total_time_enabled; /* * If the counter is in a group and isn't the group leader, * then don't put it on unless the group is on. */ if (leader != counter && leader->state != PERF_COUNTER_STATE_ACTIVE) goto unlock; if (!group_can_go_on(counter, cpuctx, 1)) { err = -EEXIST; } else { perf_disable(); if (counter == leader) err = group_sched_in(counter, cpuctx, ctx, smp_processor_id()); else err = counter_sched_in(counter, cpuctx, ctx, smp_processor_id()); perf_enable(); } if (err) { /* * If this counter can't go on and it's part of a * group, then the whole group has to come off. */ if (leader != counter) group_sched_out(leader, cpuctx, ctx); if (leader->attr.pinned) { update_group_times(leader); leader->state = PERF_COUNTER_STATE_ERROR; } } unlock: spin_unlock(&ctx->lock); } /* * Enable a counter. * * If counter->ctx is a cloned context, callers must make sure that * every task struct that counter->ctx->task could possibly point to * remains valid. This condition is satisfied when called through * perf_counter_for_each_child or perf_counter_for_each as described * for perf_counter_disable. */ static void perf_counter_enable(struct perf_counter *counter) { struct perf_counter_context *ctx = counter->ctx; struct task_struct *task = ctx->task; if (!task) { /* * Enable the counter on the cpu that it's on */ smp_call_function_single(counter->cpu, __perf_counter_enable, counter, 1); return; } spin_lock_irq(&ctx->lock); if (counter->state >= PERF_COUNTER_STATE_INACTIVE) goto out; /* * If the counter is in error state, clear that first. * That way, if we see the counter in error state below, we * know that it has gone back into error state, as distinct * from the task having been scheduled away before the * cross-call arrived. */ if (counter->state == PERF_COUNTER_STATE_ERROR) counter->state = PERF_COUNTER_STATE_OFF; retry: spin_unlock_irq(&ctx->lock); task_oncpu_function_call(task, __perf_counter_enable, counter); spin_lock_irq(&ctx->lock); /* * If the context is active and the counter is still off, * we need to retry the cross-call. */ if (ctx->is_active && counter->state == PERF_COUNTER_STATE_OFF) goto retry; /* * Since we have the lock this context can't be scheduled * in, so we can change the state safely. */ if (counter->state == PERF_COUNTER_STATE_OFF) { counter->state = PERF_COUNTER_STATE_INACTIVE; counter->tstamp_enabled = ctx->time - counter->total_time_enabled; } out: spin_unlock_irq(&ctx->lock); } static int perf_counter_refresh(struct perf_counter *counter, int refresh) { /* * not supported on inherited counters */ if (counter->attr.inherit) return -EINVAL; atomic_add(refresh, &counter->event_limit); perf_counter_enable(counter); return 0; } void __perf_counter_sched_out(struct perf_counter_context *ctx, struct perf_cpu_context *cpuctx) { struct perf_counter *counter; spin_lock(&ctx->lock); ctx->is_active = 0; if (likely(!ctx->nr_counters)) goto out; update_context_time(ctx); perf_disable(); if (ctx->nr_active) { list_for_each_entry(counter, &ctx->counter_list, list_entry) { if (counter != counter->group_leader) counter_sched_out(counter, cpuctx, ctx); else group_sched_out(counter, cpuctx, ctx); } } perf_enable(); out: spin_unlock(&ctx->lock); } /* * Test whether two contexts are equivalent, i.e. whether they * have both been cloned from the same version of the same context * and they both have the same number of enabled counters. * If the number of enabled counters is the same, then the set * of enabled counters should be the same, because these are both * inherited contexts, therefore we can't access individual counters * in them directly with an fd; we can only enable/disable all * counters via prctl, or enable/disable all counters in a family * via ioctl, which will have the same effect on both contexts. */ static int context_equiv(struct perf_counter_context *ctx1, struct perf_counter_context *ctx2) { return ctx1->parent_ctx && ctx1->parent_ctx == ctx2->parent_ctx && ctx1->parent_gen == ctx2->parent_gen && !ctx1->pin_count && !ctx2->pin_count; } static void __perf_counter_read(void *counter); static void __perf_counter_sync_stat(struct perf_counter *counter, struct perf_counter *next_counter) { u64 value; if (!counter->attr.inherit_stat) return; /* * Update the counter value, we cannot use perf_counter_read() * because we're in the middle of a context switch and have IRQs * disabled, which upsets smp_call_function_single(), however * we know the counter must be on the current CPU, therefore we * don't need to use it. */ switch (counter->state) { case PERF_COUNTER_STATE_ACTIVE: __perf_counter_read(counter); break; case PERF_COUNTER_STATE_INACTIVE: update_counter_times(counter); break; default: break; } /* * In order to keep per-task stats reliable we need to flip the counter * values when we flip the contexts. */ value = atomic64_read(&next_counter->count); value = atomic64_xchg(&counter->count, value); atomic64_set(&next_counter->count, value); swap(counter->total_time_enabled, next_counter->total_time_enabled); swap(counter->total_time_running, next_counter->total_time_running); /* * Since we swizzled the values, update the user visible data too. */ perf_counter_update_userpage(counter); perf_counter_update_userpage(next_counter); } #define list_next_entry(pos, member) \ list_entry(pos->member.next, typeof(*pos), member) static void perf_counter_sync_stat(struct perf_counter_context *ctx, struct perf_counter_context *next_ctx) { struct perf_counter *counter, *next_counter; if (!ctx->nr_stat)