kernel/git/stable/linux.git/tools/perf, branch linux-5.12.y

perf script python: Fix buffer size to report iregs in perf script

2021-07-20T14:02:28+00:00

[ Upstream commit dea8cfcc33695f70f56023b416cf88ae44c8a45a ] Commit 48a1f565261d2ab1 ("perf script python: Add more PMU fields to event handler dict") added functionality to report fields like weight, iregs, uregs etc via perf report. That commit predefined buffer size to 512 bytes to print those fields. But in PowerPC, since we added extended regs support in: 068aeea3773a6f4c ("perf powerpc: Support exposing Performance Monitor Counter SPRs as part of extended regs") d735599a069f6936 ("powerpc/perf: Add extended regs support for power10 platform") Now iregs can carry more bytes of data and this predefined buffer size can result to data loss in perf script output. This patch resolves this issue by making the buffer size dynamic, based on the number of registers needed to print. It also changes the regs_map() return type from int to void, as it is not being used by the set_regs_in_dict(), its only caller. Fixes: 068aeea3773a6f4c ("perf powerpc: Support exposing Performance Monitor Counter SPRs as part of extended regs") Signed-off-by: Kajol Jain Tested-by: Nageswara R Sastry Cc: Athira Jajeev Cc: Jiri Olsa Cc: Madhavan Srinivasan Cc: Paul Clarke Cc: Ravi Bangoria Cc: linuxppc-dev@lists.ozlabs.org Link: http://lore.kernel.org/lkml/20210628062341.155839-1-kjain@linux.ibm.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf llvm: Return -ENOMEM when asprintf() fails

2021-07-14T15:00:29+00:00

[ Upstream commit c435c166dcf526ac827bc964d82cc0d5e7a1fd0b ] Zhihao sent a patch but it made llvm__compile_bpf() return what asprintf() returns on error, which is just -1, but since this function returns -errno, fix it by returning -ENOMEM for this case instead. Fixes: cb76371441d098 ("perf llvm: Allow passing options to llc ...") Fixes: 5eab5a7ee032ac ("perf llvm: Display eBPF compiling command ...") Reported-by: Hulk Robot Reported-by: Zhihao Cheng Cc: Alexei Starovoitov Cc: Andrii Nakryiko Cc: Daniel Borkmann Cc: Ingo Molnar Cc: Jiri Olsa Cc: Nathan Chancellor Cc: Nick Desaulniers Cc: Peter Zijlstra Cc: Yu Kuai Cc: clang-built-linux@googlegroups.com Link: http://lore.kernel.org/lkml/20210609115945.2193194-1-chengzhihao1@huawei.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf scripting python: Fix tuple_set_u64()

2021-07-14T15:00:14+00:00

[ Upstream commit d04c1ff0b3ddd5c0fbbe640996c8eaad279ed1c5 ] tuple_set_u64() produces a signed value instead of an unsigned value. That works for database export but not other cases. Rename to tuple_set_d64() for database export and fix tuple_set_u64(). Fixes: df919b400ad3f ("perf scripting python: Extend interface to export data in a database-friendly way") Signed-off-by: Adrian Hunter Cc: Andi Kleen Cc: Jiri Olsa Signed-off-by: Arnaldo Carvalho de Melo Link: https://lore.kernel.org/r/20210525095112.1399-2-adrian.hunter@intel.com Signed-off-by: Sasha Levin

perf beauty: Update copy of linux/socket.h with the kernel sources

2021-06-23T12:44:11+00:00

commit ef83f9efe8461b8fd71eb60b53dbb6a5dd7b39e9 upstream. To pick the changes in: ea6932d70e223e02 ("net: make get_net_ns return error if NET_NS is disabled") That don't result in any changes in the tables generated from that header. This silences this perf build warning: Warning: Kernel ABI header at 'tools/perf/trace/beauty/include/linux/socket.h' differs from latest version at 'include/linux/socket.h' diff -u tools/perf/trace/beauty/include/linux/socket.h include/linux/socket.h Cc: Changbin Du Cc: David S. Miller Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Greg Kroah-Hartman

perf metricgroup: Return error code from metricgroup__add_metric_sys_event_iter()

2021-06-23T12:44:05+00:00

[ Upstream commit fe7a98b9d9b36e5c8a22d76b67d29721f153f66e ] The error code is not set at all in the sys event iter function. This may lead to an uninitialized value of "ret" in metricgroup__add_metric() when no CPU metric is added. Fix by properly setting the error code. It is not necessary to init "ret" to 0 in metricgroup__add_metric(), as if we have no CPU or sys event metric matching, then "has_match" should be 0 and "ret" is set to -EINVAL. However gcc cannot detect that it may not have been set after the map_for_each_metric() loop for CPU metrics, which is strange. Fixes: be335ec28efa8 ("perf metricgroup: Support adding metrics for system PMUs") Signed-off-by: John Garry Acked-by: Ian Rogers Cc: Alexander Shishkin Cc: Jiri Olsa Cc: Kajol Jain Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lore.kernel.org/lkml/1623335580-187317-3-git-send-email-john.garry@huawei.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf metricgroup: Fix find_evsel_group() event selector

2021-06-23T12:44:05+00:00

[ Upstream commit fc96ec4d5d4155c61cbafd49fb2dd403c899a9f4 ] The following command segfaults on my x86 broadwell: $ ./perf stat -M frontend_bound,retiring,backend_bound,bad_speculation sleep 1 WARNING: grouped events cpus do not match, disabling group: anon group { raw 0x10e } anon group { raw 0x10e } perf: util/evsel.c:1596: get_group_fd: Assertion `!(!leader->core.fd)' failed. Aborted (core dumped) The issue shows itself as a use-after-free in evlist__check_cpu_maps(), whereby the leader of an event selector (evsel) has been deleted (yet we still attempt to verify for an evsel). Fundamentally the problem comes from metricgroup__setup_events() -> find_evsel_group(), and has developed from the previous fix attempt in commit 9c880c24cb0d ("perf metricgroup: Fix for metrics containing duration_time"). The problem now is that the logic in checking if an evsel is in the same group is subtly broken for the "cycles" event. For the "cycles" event, the pmu_name is NULL; however the logic in find_evsel_group() may set an event matched against "cycles" as used, when it should not be. This leads to a condition where an evsel is set, yet its leader is not. Fix the check for evsel pmu_name by not matching evsels when either has a NULL pmu_name. There is still a pre-existing metric issue whereby the ordering of the metrics may break the 'stat' function, as discussed at: https://lore.kernel.org/lkml/49c6fccb-b716-1bf0-18a6-cace1cdb66b9@huawei.com/ Fixes: 9c880c24cb0d ("perf metricgroup: Fix for metrics containing duration_time") Signed-off-by: John Garry Tested-by: Arnaldo Carvalho de Melo # On a Thinkpad T450S Cc: Alexander Shishkin Cc: Ian Rogers Cc: Jiri Olsa Cc: Kajol Jain Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lore.kernel.org/lkml/1623335580-187317-2-git-send-email-john.garry@huawei.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf session: Correct buffer copying when peeking events

2021-06-16T10:05:16+00:00

[ Upstream commit 197eecb6ecae0b04bd694432f640ff75597fed9c ] When peeking an event, it has a short path and a long path. The short path uses the session pointer "one_mmap_addr" to directly fetch the event; and the long path needs to read out the event header and the following event data from file and fill into the buffer pointer passed through the argument "buf". The issue is in the long path that it copies the event header and event data into the same destination address which pointer "buf", this means the event header is overwritten. We are just lucky to run into the short path in most cases, so we don't hit the issue in the long path. This patch adds the offset "hdr_sz" to the pointer "buf" when copying the event data, so that it can reserve the event header which can be used properly by its caller. Fixes: 5a52f33adf02 ("perf session: Add perf_session__peek_event()") Signed-off-by: Leo Yan Acked-by: Adrian Hunter Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lore.kernel.org/lkml/20210605052957.1070720-1-leo.yan@linaro.org Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf probe: Fix NULL pointer dereference in convert_variable_location()

2021-06-10T11:41:35+00:00

[ Upstream commit 3cb17cce1e76ccc5499915a4d7e095a1ad6bf7ff ] If we just check whether the variable can be converted, 'tvar' should be a null pointer. However, the null pointer check is missing in the 'Constant value' execution path. The following cases can trigger this problem: $ cat test.c #include void main(void) { int a; const int b = 1; asm volatile("mov %1, %0" : "=r"(a): "i"(b)); printf("a: %d\n", a); } $ gcc test.c -o test -O -g $ sudo ./perf probe -x ./test -L "main" 0 void main(void) { 2 int a; const int b = 1; asm volatile("mov %1, %0" : "=r"(a): "i"(b)); 6 printf("a: %d\n", a); } $ sudo ./perf probe -x ./test -V "main:6" Segmentation fault The check on 'tvar' is added. If 'tavr' is a null pointer, we return 0 to indicate that the variable can be converted. Now, we can successfully show the variables that can be accessed. $ sudo ./perf probe -x ./test -V "main:6" Available variables at main:6 @ char* __fmt int a int b However, the variable 'b' cannot be tracked. $ sudo ./perf probe -x ./test -D "main:6 b" Failed to find the location of the 'b' variable at this address. Perhaps it has been optimized out. Use -V with the --range option to show 'b' location range. Error: Failed to add events. This is because __die_find_variable_cb() did not successfully match variable 'b', which has the DW_AT_const_value attribute instead of DW_AT_location. We added support for DW_AT_const_value in __die_find_variable_cb(). With this modification, we can successfully track the variable 'b'. $ sudo ./perf probe -x ./test -D "main:6 b" p:probe_test/main_L6 /home/lhf/test:0x1156 b=\1:s32 Fixes: 66f69b219716 ("perf probe: Support DW_AT_const_value constant value") Signed-off-by: Li Huafei Tested-by: Arnaldo Carvalho de Melo Cc: Alexander Shishkin Cc: Frank Ch. Eigler Cc: Jianlin Lv Cc: Jiri Olsa Cc: Mark Rutland Cc: Masami Hiramatsu Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Yang Jihong Cc: Zhang Jinhao http://lore.kernel.org/lkml/20210601092750.169601-1-lihuafei1@huawei.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Sasha Levin

perf jevents: Fix getting maximum number of fds

2021-06-03T07:09:38+00:00

commit 75ea44e356b5de8c817f821c9dd68ae329e82add upstream. On some hosts, rlim.rlim_max can be returned as RLIM_INFINITY. By casting it to int, it is interpreted as -1, which will cause get_maxfds to return 0, causing "Invalid argument" errors in nftw() calls. Fix this by casting the second argument of min() to rlim_t instead. Fixes: 80eeb67fe577 ("perf jevents: Program to convert JSON file") Signed-off-by: Felix Fietkau Cc: Alexander Shishkin Cc: Andi Kleen Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Sukadev Bhattiprolu Link: http://lore.kernel.org/lkml/20210525160758.97829-1-nbd@nbd.name Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Greg Kroah-Hartman

perf debug: Move debug initialization earlier

2021-06-03T07:09:37+00:00

commit c59870e2110e1229a6e4b2457aece6ffe8d68d99 upstream. This avoids segfaults during option handlers that use pr_err. For example, "perf --debug nopager list" segfaults before this change. Fixes: 8abceacff87d (perf debug: Add debug_set_file function) Signed-off-by: Ian Rogers Tested-by: Arnaldo Carvalho de Melo Cc: Alexander Shishkin Cc: Jiri Olsa Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20210519164447.2672030-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Greg Kroah-Hartman