linux-stable.git/tools/perf/util, branch v4.4.124

perf session: Don't rely on evlist in pipe mode

2018-03-22T08:23:26+00:00

[ Upstream commit 0973ad97c187e06aece61f685b9c3b2d93290a73 ]

Session sets a number parameters that rely on evlist. These parameters
are not used in pipe-mode and should not be set, since evlist is
unavailable. Fix that.

Signed-off-by: David Carrillo-Cisneros 
Acked-by: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Andi Kleen 
Cc: He Kuang 
Cc: Masami Hiramatsu 
Cc: Paul Turner 
Cc: Peter Zijlstra 
Cc: Simon Que 
Cc: Stephane Eranian 
Cc: Wang Nan 
Link: http://lkml.kernel.org/r/20170410201432.24807-6-davidcc@google.com
[ Check if file != NULL in perf_session__new(), like when used by builtin-top.c ]
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

perf inject: Copy events when reordering events in pipe mode

2018-03-22T08:23:26+00:00

[ Upstream commit 1e0d4f0200e4dbdfc38d818f329d8a0955f7c6f5 ]

__perf_session__process_pipe_events reuses the same memory buffer to
process all events in the pipe.

When reordering is needed (e.g. -b option), events are not immediately
flushed, but kept around until reordering is possible, causing
memory corruption.

The problem is usually observed by a "Unknown sample error" output. It
can easily be reproduced by:

  perf record -o - noploop | perf inject -b > output

Committer testing:

Before:

  $ perf record -o - stress -t 2 -c 2 | perf inject -b > /dev/null
  stress: info: [8297] dispatching hogs: 2 cpu, 0 io, 0 vm, 0 hdd
  stress: info: [8297] successful run completed in 2s
  [ perf record: Woken up 3 times to write data ]
  [ perf record: Captured and wrote 0.000 MB - ]
  Warning:
  Found 1 unknown events!

  Is this an older tool processing a perf.data file generated by a more recent tool?

  If that is not the case, consider reporting to linux-kernel@vger.kernel.org.

  $

After:

  $ perf record -o - stress -t 2 -c 2 | perf inject -b > /dev/null
  stress: info: [9027] dispatching hogs: 2 cpu, 0 io, 0 vm, 0 hdd
  stress: info: [9027] successful run completed in 2s
  [ perf record: Woken up 3 times to write data ]
  [ perf record: Captured and wrote 0.000 MB - ]
  no symbols found in /usr/bin/stress, maybe install a debug package?
  no symbols found in /usr/bin/stress, maybe install a debug package?
  $

Signed-off-by: David Carrillo-Cisneros 
Tested-by: Arnaldo Carvalho de Melo 
Acked-by: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Andi Kleen 
Cc: He Kuang 
Cc: Masami Hiramatsu 
Cc: Paul Turner 
Cc: Peter Zijlstra 
Cc: Simon Que 
Cc: Stephane Eranian 
Cc: Wang Nan 
Link: http://lkml.kernel.org/r/20170410201432.24807-3-davidcc@google.com
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

perf tools: Make perf_event__synthesize_mmap_events() scale

2018-03-22T08:23:20+00:00

[ Upstream commit 88b897a30c525c2eee6e7f16e1e8d0f18830845e ]

This patch significantly improves the execution time of
perf_event__synthesize_mmap_events() when running perf record on systems
where processes have lots of threads.

It just happens that cat /proc/pid/maps support uses a O(N^2) algorithm to
generate each map line in the maps file.  If you have 1000 threads, then you
have necessarily 1000 stacks.  For each vma, you need to check if it
corresponds to a thread's stack.  With a large number of threads, this can take
a very long time. I have seen latencies >> 10mn.

As of today, perf does not use the fact that a mapping is a stack, therefore we
can work around the issue by using /proc/pid/tasks/pid/maps.  This entry does
not try to map a vma to stack and is thus much faster with no loss of
functonality.

The proc-map-timeout logic is kept in case users still want some upper limit.

In V2, we fix the file path from /proc/pid/tasks/pid/maps to actual
/proc/pid/task/pid/maps, tasks -> task.  Thanks Arnaldo for catching this.

Committer note:

This problem seems to have been elliminated in the kernel since commit :
b18cb64ead40 ("fs/proc: Stop trying to report thread stacks").

Signed-off-by: Stephane Eranian 
Acked-by: Jiri Olsa 
Cc: Andy Lutomirski 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/20170315135059.GC2177@redhat.com
Link: http://lkml.kernel.org/r/1489598233-25586-1-git-send-email-eranian@google.com
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

perf sort: Fix segfault with basic block 'cycles' sort dimension

2018-03-22T08:23:20+00:00

[ Upstream commit 4b0b3aa6a2756e6115fdf275c521e4552a7082f3 ]

Skip the sample which doesn't have branch_info to avoid segmentation
fault:

The fault can be reproduced by:

  perf record -a
  perf report -F cycles

Signed-off-by: Changbin Du 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Andi Kleen 
Cc: Peter Zijlstra 
Fixes: 0e332f033a82 ("perf tools: Add support for cycles, weight branch_info field")
Link: http://lkml.kernel.org/r/20170313083148.23568-1-changbin.du@intel.com
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

perf symbols: Fix symbols__fixup_end heuristic for corner cases

2017-12-20T09:04:56+00:00

[ Upstream commit e7ede72a6d40cb3a30c087142d79381ca8a31dab ]

The current symbols__fixup_end() heuristic for the last entry in the rb
tree is suboptimal as it leads to not being able to recognize the symbol
in the call graph in a couple of corner cases, for example:

 i) If the symbol has a start address (f.e. exposed via kallsyms)
    that is at a page boundary, then the roundup(curr->start, 4096)
    for the last entry will result in curr->start == curr->end with
    a symbol length of zero.

ii) If the symbol has a start address that is shortly before a page
    boundary, then also here, curr->end - curr->start will just be
    very few bytes, where it's unrealistic that we could perform a
    match against.

Instead, change the heuristic to roundup(curr->start, 4096) + 4096, so
that we can catch such corner cases and have a better chance to find
that specific symbol. It's still just best effort as the real end of the
symbol is unknown to us (and could even be at a larger offset than the
current range), but better than the current situation.

Alexei reported that he recently run into case i) with a JITed eBPF
program (these are all page aligned) as the last symbol which wasn't
properly shown in the call graph (while other eBPF program symbols in
the rb tree were displayed correctly). Since this is a generic issue,
lets try to improve the heuristic a bit.

Reported-and-Tested-by: Alexei Starovoitov 
Signed-off-by: Daniel Borkmann 
Fixes: 2e538c4a1847 ("perf tools: Improve kernel/modules symbol lookup")
Link: http://lkml.kernel.org/r/bb5c80d27743be6f12afc68405f1956a330e1bc9.1489614365.git.daniel@iogearbox.net
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

perf tools: Only increase index if perf_evsel__new_idx() succeeds

2017-11-08T09:06:29+00:00

[ Upstream commit 75fc5ae5cc53fff71041ecadeb3354a2b4c9fe42 ]

Signed-off-by: Taeung Song 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Wang Nan 
Link: http://lkml.kernel.org/r/1485952447-7013-2-git-send-email-treeze.taeung@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

perf symbols: Robustify reading of build-id from sysfs

2017-08-07T02:19:46+00:00

[ Upstream commit 7934c98a6e04028eb34c1293bfb5a6b0ab630b66 ]

Markus reported that perf segfaults when reading /sys/kernel/notes from
a kernel linked with GNU gold, due to what looks like a gold bug, so do
some bounds checking to avoid crashing in that case.

Reported-by: Markus Trippelsdorf 
Report-Link: http://lkml.kernel.org/r/20161219161821.GA294@x4
Cc: Adrian Hunter 
Cc: David Ahern 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Wang Nan 
Link: http://lkml.kernel.org/n/tip-ryhgs6a6jxvz207j2636w31c@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

perf intel-pt: Ensure never to set 'last_ip' when packet 'count' is zero

2017-08-07T02:19:40+00:00

commit f952eaceb089b691eba7c4e13686e742a8f26bf5 upstream.

Intel PT uses IP compression based on the last IP. For decoding purposes,
'last IP' is not updated when a branch target has been suppressed, which is
indicated by IPBytes == 0. IPBytes is stored in the packet 'count', so
ensure never to set 'last_ip' when packet 'count' is zero.

Signed-off-by: Adrian Hunter 
Cc: Andi Kleen 
Link: http://lkml.kernel.org/r/1495786658-18063-7-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Greg Kroah-Hartman

perf intel-pt: Use FUP always when scanning for an IP

2017-08-07T02:19:40+00:00

commit 622b7a47b843c78626f40c1d1aeef8483383fba2 upstream.

The decoder will try to use branch packets to find an IP to start decoding
or to recover from errors. Currently the FUP packet is used only in the
case of an overflow, however there is no reason for that to be a special
case. So just use FUP always when scanning for an IP.

Signed-off-by: Adrian Hunter 
Cc: Andi Kleen 
Link: http://lkml.kernel.org/r/1495786658-18063-8-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Greg Kroah-Hartman

perf intel-pt: Fix last_ip usage

2017-08-07T02:19:40+00:00

commit ee14ac0ef6827cd6f9a572cc83dd0191ea17812c upstream.

Intel PT uses IP compression based on the last IP. For decoding
purposes, 'last IP' is considered to be reset to zero whenever there is
a synchronization packet (PSB). The decoder wasn't doing that, and was
treating the zero value to mean that there was no last IP, whereas
compression can be done against the zero value. Fix by setting last_ip
to zero when a PSB is received and keep track of have_last_ip.

Signed-off-by: Adrian Hunter 
Cc: Andi Kleen 
Link: http://lkml.kernel.org/r/1495786658-18063-6-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo 
Signed-off-by: Greg Kroah-Hartman