- 07 Sep, 2012 2 commits
-
-
Namhyung Kim authored
It prevents mindless git add from adding those files. Signed-off-by:
Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/1346982953-30824-1-git-send-email-namhyung@kernel.orgSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo: - Rename libtraceevent 'private' struct member to 'priv' so that it works in C++, from Steven Rostedt - Remove lots of exit()/die() calls from tools so that the main perf exit routine can take place, from David Ahern - Fix x86 build on x86-64, from David Ahern. - Remove some headers that prevented perf from building on Android, from David Ahern - {int,str,rb}list fixes from Suzuki K Poulose - perf.data header fixes from Namhyung Kim - Replace needless mempcpy with memcpy, to allow build on Android, from Irina Tirdea - Allow user to indicate objdump path, needed in cross environments, from Maciek Borzecki - Fix hardware cache event name generation, fix from Jiri Olsa - Add round trip test for sw, hw and cache event names, catching the problem Jiri fixed, after Jiri's patch, the test passes successfully. - Clean target should do clean for lib/traceevent too, fix from David Ahern - Check the right variable for allocation failure, fix from Namhyung Kim - Set up evsel->tp_format regardless of evsel->name being set already, fix from Namhyung Kim Signed-off-by:
Ingo Molnar <mingo@kernel.org>
-
- 06 Sep, 2012 8 commits
-
-
Jiri Olsa authored
If the event name is specified with all 3 components, the last one overwrites the previous one during the name composing within the parse_events_add_cache function. Fixing this by properly adjusting the string index. Reported-by:
Joel Uckelman <joel@lightboxtechnologies.com> Signed-off-by:
Jiri Olsa <jolsa@redhat.com> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Joel Uckelman <joel@lightboxtechnologies.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> LPU-Reference: 20120905175133.GA18352@krava.brq.redhat.com [ committer note: Remove the newline fix, done already in 42e1fb77 ] Signed-off-by:
-
Arnaldo Carvalho de Melo authored
That nicely catches the problem reported by Joel Uckelman in http://permalink.gmane.org/gmane.linux.kernel.perf.user/1016 : [root@sandy ~]# perf test 1: vmlinux symtab matches kallsyms: Ok 2: detect open syscall event: Ok 3: detect open syscall event on all cpus: Ok 4: read samples using the mmap interface: Ok 5: parse events tests: Ok 6: x86 rdpmc test: Ok 7: Validate PERF_RECORD_* events & perf_sample fields: Ok 8: Test perf pmu format parsing: Ok 9: Test dso data interface: Ok 10: roundtrip evsel->name check: FAILED! [root@sandy ~]# perf test -v 10 10: roundtrip evsel->name check: --- start --- L1-dcache-misses != L1-dcache-load-misses L1-dcache-misses != L1-dcache-store-misses L1-dcache-misses != L1-dcache-prefetch-misses L1-icache-misses != L1-icache-load-misses L1-icache-misses != L1-icache-prefetch-misses LLC-misses != LLC-load-misses LLC-misses != LLC-store-misses LLC-misses != LLC-prefetch-misses dTLB-misses != dTLB-load-misses dTLB-misses != dTLB-store-misses dTLB-misses != dTLB-prefetch-misses iTLB-misses != iTLB-load-misses branch-misses != branch-load-misses node-misses != node-load-misses node-misses != node-store-misses node-misses != node-prefetch-misses ---- end ---- roundtrip evsel->name check: FAILED! [root@sandy ~]# Now lemme apply Jiri's fix and try it again... Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Joel Uckelman <joel@lightboxtechnologies.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-bbewtxw0rfipp5qy1j3jtg5d@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
For debugging, etc. Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-fjimge1ovgh976qlt8dtmlp0@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
Noticed while developing a 'perf test' entry to verify that perf_evsel__name works. Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-xz6zgh38mp3cjnd2udh38z8f@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
It basically traverses the hardware and software event name arrays creating an evlist with all events, then it uses perf_evsel__name to check that the name is the expected one. With it I noticed this problem: [root@sandy ~]# perf test 10 10: roundtrip evsel->name check:invalid or unsupported event: 'CPU-migrations' Run 'perf list' for a list of valid events FAILED! Changed it to "cpu-migrations" in the software event arrays and it worked. This is to catch problems like the one reported by Joel Uckelman in http://permalink.gmane.org/gmane.linux.kernel.perf.user/1016 Hardware cache events will be checked in the following patch. Cc: David Ahern <dsahern@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lkml.kernel.org/n/tip-5jskfkuqvf2fi257zmni0ftz@git.kernel.orgSigned-off-by:
-
Namhyung Kim authored
Current perf_evlist__set_tracepoint_names is a misnomer because it finds and sets correspoding event_format in addition to the name. So skipping it when a event has set name already caused a trouble. Rename it and set name only a event doesn't have one. Reported-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
-
Namhyung Kim authored
For checking return value of the strdup, 'event' should be 'evsel'. Signed-off-by:
-
David Ahern authored
It's built as part of perf, so it should be cleaned too. Tested-by:
-
- 05 Sep, 2012 18 commits
-
-
Maciek Borzecki authored
When analyzing perf data from hosts of other architecture than one of the local host it's useful to call objdump that is part of a toolchain for that architecture. Instead of calling regular objdump, call one that user specified in command line. Signed-off-by:
Maciek Borzecki <maciek.borzecki@gmail.com> Acked-by:
-
Irina Tirdea authored
mempcpy is not supported by bionic in Android and will lead to compilation errors. Replacing mempcpy with memcpy so it will work in Android. Signed-off-by:
Irina Tirdea <irina.tirdea@intel.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Steven Rostedt <rostedt@goodmis.org> Link: http://lkml.kernel.org/r/CANg8OW+Y3ZMG-GdhYu2_yKOYH_XEMgw73PdCX_23UTnfYhmttA@mail.gmail.comSigned-off-by:
-
Namhyung Kim authored
Like others, the numbers can be saved in a different endian format than a host machine. Swap them if needed. Signed-off-by:
-
Namhyung Kim authored
The event name can be set already by processing a event_desc data. So check it before setting to prevent possible leak. Signed-off-by:
-
Namhyung Kim authored
Number of events (evsels) in a evlist is kept on nr_entries field so that we don't need to recalculate it. Signed-off-by:
-
David Ahern authored
perf does not have networking related functionality, and the inclusion of these headers is one of the causes of compile failures for Android: https://lkml.org/lkml/2012/8/23/316 https://lkml.org/lkml/2012/8/28/293 So, remove them. Signed-off-by:
-
Suzuki K. Poulose authored
The following commit: author David Ahern <dsahern@gmail.com> Tue, 31 Jul 2012 04:31:33 +0000 (22:31 -0600) committer Arnaldo Carvalho de Melo <acme@redhat.com> Fri, 3 Aug 2012 13:39:51 +0000 (10:39 -0300) commit ee8dd3ca causes a double free during a probe deletion as the node is never removed from the list via strlist__remove(), even though it gets 'deleted' (read free()'d). This causes a double free when we do strlist__delete() as the node is already deleted but present in the rblist. [suzukikp@suzukikp perf]$ sudo ./perf probe -a do_fork Added new event: probe:do_fork (on do_fork) You can now use it in all perf tools, such as: perf record -e probe:do_fork -aR sleep 1 [suzukikp@suzukikp perf]$ sudo ./perf probe -d do_fork Removed event: probe:do_fork *** glibc detected *** ./perf: double free or corruption (fasttop): 0x000000000133d600 *** ======= Backtrace: ========= /lib64/libc.so.6[0x38eec7dda6] ./perf(rblist__delete+0x5c)[0x47d3dc] ./perf(del_perf_probe_events+0xb6)[0x47b826] ./perf(cmd_probe+0x471)[0x42c8d1] ./perf[0x4150b3] ./perf(main+0x501)[0x4148e1] /lib64/libc.so.6(__libc_start_main+0xed)[0x38eec2169d] ./perf[0x414a61] Make sure we remove the node from the rblist before we delete the node. The rblist__remove_node() will invoke rblist->node_delete, which will take care of deleting the node with the suitable function provided by the user. Reported-by:
Ananth N. Mavinakayanahalli <ananth@in.ibm.com> Signed-off-by:
Suzuki K. Poulose <suzuki@in.ibm.com> Acked-by:
-
Suzuki K. Poulose authored
Similar to the one in : https://lkml.org/lkml/2012/8/29/27 Make sure we remove the node from the rblist before we delete the node. The rblist__remove_node() will invoke rblist->node_delete, which will take care of deleting the node with the suitable function provided by the user. Signed-off-by:
-
David Ahern authored
e.g., compiling i386 on x86_64 using: $ make -C tools/perf ARCH=i386 fails with: CC /tmp/pbuild/util/evsel.o In file included from util/evsel.c:21:0: util/perf_regs.h:5:23: fatal error: perf_regs.h: No such file or directory compilation terminated. Adding V=1 you see that the include argument for the arch is '-Iarch/i386/include' is wrong. It is supposed to be -Iarch/x86/include per the redefinition of ARCH in the Makefile. According to the make manual, http://www.gnu.org/software/make/manual/make.html#Override-Directive: "If a variable has been set with a command argument (see Overriding Variables), then ordinary assignments in the makefile are ignored. If you want to set the variable in the makefile even though it was set with a command argument, you can use an override directive ..." Make it so. Signed-off-by: -
David Ahern authored
Allows perf to clean up properly on exit. If perf-record is exiting due to failure, the on_exit should not run as the session has been deleted. Signed-off-by:
-
David Ahern authored
Allows perf to clean up properly on exit. Only exits left are exec failures which are appropriate and usage callbacks that list available options. Signed-off-by:
-
David Ahern authored
Allows perf to clean up properly on exit. Signed-off-by:
-
David Ahern authored
Allows perf to clean up properly on program termination. Signed-off-by:
-
David Ahern authored
Allows perf to clean up properly on exit. Signed-off-by:
-
David Ahern authored
Handle error from process callback and propagate back to caller. Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1346005487-62961-3-git-send-email-dsahern@gmail.comSigned-off-by:
-
David Ahern authored
Allows errors to propogate through event processing code and back to commands. Signed-off-by:
-
Steven Rostedt authored
Replace keyword "private" to "priv" in event-parse.h to allow it to be used in C++ programs. Signed-off-by:
Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Namhyung Kim <namhyung@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/1345735321.5069.62.camel@gandalf.local.homeSigned-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/rric/oprofileIngo Molnar authored
Pull oprofile fixes from Robert Richter. Signed-off-by:
-
- 28 Aug, 2012 10 commits
-
-
Oleg Nesterov authored
Nobody does set_orig_insn(verify => false), and I think nobody will. Remove this argument. IIUC set_orig_insn(verify => false) was needed to single-step without xol area. Signed-off-by:
Oleg Nesterov <oleg@redhat.com> Acked-by:
Srikar Dronamraju <srikar@linux.vnet.ibm.com>
-
Oleg Nesterov authored
Now that we have uprobe_dup_mmap() we can fold uprobe_reset_state() into the new hook and remove it. mmput()->uprobe_clear_state() can't be called before dup_mmap(). Signed-off-by:
-
Oleg Nesterov authored
Add the new MMF_HAS_UPROBES flag. It is set by install_breakpoint() and it is copied by dup_mmap(), uprobe_pre_sstep_notifier() checks it to avoid the slow path if the task was never probed. Perhaps it makes sense to check it in valid_vma(is_register => false) as well. This needs the new dup_mmap()->uprobe_dup_mmap() hook. We can't use uprobe_reset_state() or put MMF_HAS_UPROBES into MMF_INIT_MASK, we need oldmm->mmap_sem to avoid the race with uprobe_register() or mmap() from another thread. Currently we never clear this bit, it can be false-positive after uprobe_unregister() or uprobe_munmap() or if dup_mmap() hits the probed VM_DONTCOPY vma. But this is fine correctness-wise and has no effect unless the task hits the non-uprobe breakpoint. Signed-off-by:
-
Oleg Nesterov authored
-EEXIST from install_breakpoint() no longer makes sense, all callers should simply treat it as "success". Change the code to return zero and simplify register_for_each_vma(). Signed-off-by:
-
Oleg Nesterov authored
Once install_breakpoint() fails uprobe_mmap() "ignores" all other uprobes and returns the error. It was never really needed to to stop after the first error, and in fact it was always wrong at least in -ENOTSUPP case. Change uprobe_mmap() to ignore the errors and always return 0. This is not what we want in the long term, but until we teach the callers to handle the failure it would be better to remove the pointless complications. And this doesn't look too bad, the only "reasonable" error is ENOMEM but in this case the caller should be oom-killed in the likely case or the system has more serious problems. However it makes sense to stop if fatal_signal_pending() == T. In particular this helps if the task was oom-killed. Signed-off-by:
-
Oleg Nesterov authored
1. Kill dup_mmap()->uprobe_mmap(), it was only needed to calculate new_mm->uprobes_state.count removed by the previous patch. If the forking process has a pending uprobe (int3) in vma, it will be copied by copy_page_range(), note that it checks vma->anon_vma so "Don't copy ptes" is not possible after install_breakpoint() which does anon_vma_prepare(). 2. Remove is_swbp_at_addr() and "int count" in uprobe_mmap(). Again, this was needed for uprobes_state.count. As a side effect this fixes the bug pointed out by Srikar, this code lacked the necessary put_uprobe(). 3. uprobe_munmap() becomes a nop after the previous patch. Remove the meaningless code but do not remove the helper, we will need it. Signed-off-by:
-
Oleg Nesterov authored
uprobes_state->count is only needed to avoid the slow path in uprobe_pre_sstep_notifier(). It is also checked in uprobe_munmap() but ironically its only goal to decrement this counter. However, it is very broken. Just some examples: - uprobe_mmap() can race with uprobe_unregister() and wrongly increment the counter if it hits the non-uprobe "int3". Note that install_breakpoint() checks ->consumers first and returns -EEXIST if it is NULL. "atomic_sub() if error" in uprobe_mmap() looks obviously wrong too. - uprobe_munmap() can race with uprobe_register() and wrongly decrement the counter by the same reason. - Suppose an appication tries to increase the mmapped area via sys_mremap(). vma_adjust() does uprobe_munmap(whole_vma) first, this can nullify the counter temporarily and race with another thread which can hit the bp, the application will be killed by SIGTRAP. - Suppose an application mmaps 2 consecutive areas in the same file and one (or both) of these areas has uprobes. In the likely case mmap_region()->vma_merge() suceeds. Like above, this leads to uprobe_munmap/uprobe_mmap from vma_merge()->vma_adjust() but then mmap_region() does another uprobe_mmap(resulting_vma) and doubles the counter. This patch only removes this counter and fixes the compile errors, then we will try to cleanup the changed code and add something else instead. Signed-off-by:
-
Sebastian Andrzej Siewior authored
by the time we get here (after we pass cleanup_ret) uprobe is always is set. If it is NULL we leave very early in the code. Signed-off-by:
Sebastian Andrzej Siewior <bigeasy@linutronix.de> Signed-off-by:
-
Srikar Dronamraju authored
Since read_opcode() reads from the referenced page and doesnt modify the page contents nor the page attributes, there is no need to lock the page. Signed-off-by:
-
Ingo Molnar authored
Pick up the latest fixes because upcoming uprobes changes will rely on it. Signed-off-by:
-
- 27 Aug, 2012 2 commits
-
-
Robert Richter authored
Under certain workloads we see the following warnings: WQ on CPU0, prefer CPU1 WQ on CPU0, prefer CPU2 WQ on CPU0, prefer CPU3 It warns the user that the wq to access a per-cpu buffers runs not on the same cpu. This happens if the wq is rescheduled on a different cpu than where the buffer is located. This was probably implemented to detect performance issues. Not sure if there actually is one as the buffers are copied to a single buffer anyway which should be the actual bottleneck. We wont change WQ implementation. Since a user can do nothing the warning is pointless. Removing it. Cc: Andi Kleen <andi@firstfloor.org> Signed-off-by:Robert Richter <robert.richter@amd.com>
-
Robert Richter authored
If oprofilefs_ulong_from_user() is called with count equals zero, *val remains unchanged. Depending on the implementation it might be uninitialized. Fixing users of oprofilefs_ulong_ from_user(). We missed these s390 changes with: 913050b9 oprofile: Fix uninitialized memory access when writing to writing to oprofilefs Cc: stable@vger.kernel.org # 3.3+ Signed-off-by:
-
