- 30 Aug, 2018 16 commits
-
-
Arnaldo Carvalho de Melo authored
Since we copy all the payload for raw_syscalls:sys_enter plus add expanded pointers, we can use the syscall id to get its name, etc: # grep 'field:.* id' /sys/kernel/debug/tracing/events/raw_syscalls/sys_enter/format field:long id; offset:8; size:8; signed:1; # Before: # perf trace -e tools/perf/examples/bpf/augmented_syscalls.c,openat cat /etc/passwd > /dev/null 0.000 ( ): __augmented_syscalls__:dfd: CWD, filename: 0xec9f9da8, flags: CLOEXEC 0.006 ( 0.006 ms): cat/2395 openat(dfd: CWD, filename: 0xec9f9da8, flags: CLOEXEC) = 3 0.041 ( ): __augmented_syscalls__:dfd: CWD, filename: 0xecc01ce0, flags: CLOEXEC 0.042 ( 0.007 ms): cat/2395 openat(dfd: CWD, filename: 0xecc01ce0, flags: CLOEXEC) = 3 0.376 ( ): __augmented_syscalls__:dfd: CWD, filename: 0xac0a806b 0.379 ( 0.006 ms): cat/2395 openat(dfd: CWD, filename: 0xac0a806b) = 3 # After: # perf trace -e tools/perf/examples/bpf/augmented_syscalls.c,openat cat /etc/passwd > /dev/null 0.000 ( ): openat(dfd: CWD, filename: 0x31b6dda8, flags: CLOEXEC) 0.009 ( 0.009 ms): cat/3619 openat(dfd: CWD, filename: 0x31b6dda8, flags: CLOEXEC) = 3 0.051 ( ): openat(dfd: CWD, filename: 0x31d75ce0, flags: CLOEXEC) 0.054 ( 0.010 ms): cat/3619 openat(dfd: CWD, filename: 0x31d75ce0, flags: CLOEXEC) = 3 0.539 ( ): openat(dfd: CWD, filename: 0xca71506b) 0.543 ( 0.115 ms): cat/3619 openat(dfd: CWD, filename: 0xca71506b) = 3 # Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: https://lkml.kernel.org/n/tip-epz6y9i0eavmerc5ha98t7gn@git.kernel.orgSigned-off-by:Arnaldo Carvalho de Melo <acme@redhat.com>
-
Arnaldo Carvalho de Melo authored
When we attach a eBPF object to a tracepoint, if we return 1, then that tracepoint will be stored in the perf's ring buffer. In the augmented_syscalls.c case we want to just attach and _override_ the tracepoint payload with an augmented, extended one. In this example, tools/perf/examples/bpf/augmented_syscalls.c, we are attaching to the 'openat' syscall, and adding, after the syscalls:sys_enter_openat usual payload as defined by /sys/kernel/debug/tracing/events/syscalls/sys_enter_openat/format, a snapshot of its sole pointer arg: # grep 'field:.*\*' /sys/kernel/debug/tracing/events/syscalls/sys_enter_openat/format field:const char * filename; offset:24; size:8; signed:0; # For now this is not being considered, the next csets will make use of it, but as this is overriding the syscall tracepoint enter, we don't want that event appearing on the ring buffer, just our synthesized one. Before: # perf trace -e ~acme/git/perf/tools/perf/examples/bpf/augmented_syscalls.c,openat cat /etc/passwd > /dev/null 0.000 ( ): __augmented_syscalls__:dfd: CWD, filename: /etc/ld.so.cache, flags: CLOEXEC 0.006 ( ): syscalls:sys_enter_openat:dfd: CWD, filename: , flags: CLOEXEC 0.007 ( 0.004 ms): cat/24044 openat(dfd: CWD, filename: 0x216dda8, flags: CLOEXEC ) = 3 0.028 ( ): __augmented_syscalls__:dfd: CWD, filename: /lib64/libc.so.6, flags: CLOEXEC 0.030 ( ): syscalls:sys_enter_openat:dfd: CWD, filename: , flags: CLOEXEC 0.031 ( 0.006 ms): cat/24044 openat(dfd: CWD, filename: 0x2375ce0, flags: CLOEXEC ) = 3 0.291 ( ): __augmented_syscalls__:dfd: CWD, filename: /etc/passwd 0.293 ( ): syscalls:sys_enter_openat:dfd: CWD, filename: 0.294 ( 0.004 ms): cat/24044 openat(dfd: CWD, filename: 0x637db06b ) = 3 # After: # perf trace -e ~acme/git/perf/tools/perf/examples/bpf/augmented_syscalls.c,openat cat /etc/passwd > /dev/null 0.000 ( ): __augmented_syscalls__:dfd: CWD, filename: 0x9c6a1da8, flags: CLOEXEC 0.005 ( 0.015 ms): cat/27341 openat(dfd: CWD, filename: 0x9c6a1da8, flags: CLOEXEC ) = 3 0.040 ( ): __augmented_syscalls__:dfd: CWD, filename: 0x9c8a9ce0, flags: CLOEXEC 0.041 ( 0.006 ms): cat/27341 openat(dfd: CWD, filename: 0x9c8a9ce0, flags: CLOEXEC ) = 3 0.294 ( ): __augmented_syscalls__:dfd: CWD, filename: 0x482a706b 0.296 ( 0.067 ms): cat/27341 openat(dfd: CWD, filename: 0x482a706b ) = 3 # Now lets replace that __augmented_syscalls__ name with the syscall name, using: # grep 'field:.*syscall_nr' /sys/kernel/debug/tracing/events/syscalls/sys_enter_openat/format field:int __syscall_nr; offset:8; size:4; signed:1; # That the synthesized payload has exactly where the syscall enter tracepoint puts it. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: https://lkml.kernel.org/n/tip-og4r9k87mzp9hv7el046idmd@git.kernel.orgSigned-off-by: -
Arnaldo Carvalho de Melo authored
If the tracepoint payload is bigger than what a syscall expected from what is in its format file in tracefs, then that will be used as augmented args, i.e. the expansion of syscall arg pointers, with things like a filename, structs, etc. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: https://lkml.kernel.org/n/tip-bsbqx7xi2ot4q9bf570f7tqs@git.kernel.orgSigned-off-by:
-
Kim Phillips authored
Starting with binutils 2.28, aarch64 objdump adds comments to the disassembly output to show the alternative names of a condition code [1]. It is assumed that commas in objdump comments could occur in other arches now or in the future, so this fix is arch-independent. The fix could have been done with arm64 specific jump__parse and jump__scnprintf functions, but the jump__scnprintf instruction would have to have its comment character be a literal, since the scnprintf functions cannot receive a struct arch easily. This inconvenience also applies to the generic jump__scnprintf, which is why we add a raw_comment pointer to struct ins_operands, so the __parse function assigns it to be re-used by its corresponding __scnprintf function. Example differences in 'perf annotate --stdio2' output on an aarch64 perf.data file: BEFORE: → b.cs ffff200008133d1c <unwind_frame+0x18c> // b.hs, dffff7ecc47b AFTER : ↓ b.cs 18c BEFORE: → b.cc ffff200008d8d9cc <get_alloc_profile+0x31c> // b.lo, b.ul, dffff727295b AFTER : ↓ b.cc 31c The branch target labels 18c and 31c also now appear in the output: BEFORE: add x26, x29, #0x80 AFTER : 18c: add x26, x29, #0x80 BEFORE: add x21, x21, #0x8 AFTER : 31c: add x21, x21, #0x8 The Fixes: tag below is added so stable branches will get the update; it doesn't necessarily mean that commit was broken at the time, rather it didn't withstand the aarch64 objdump update. Tested no difference in output for sample x86_64, power arch perf.data files. [1] https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=commit;h=bb7eff5206e4795ac79c177a80fe9f4630aaf730Signed-off-by:
Kim Phillips <kim.phillips@arm.com> Tested-by:
-
Sandipan Das authored
This makes sure that the SyS symbols are ignored for any powerpc system, not just the big endian ones. Reported-by:
Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> Signed-off-by:
Sandipan Das <sandipan@linux.ibm.com> Reviewed-by:
Kamalesh Babulal <kamalesh@linux.vnet.ibm.com> Acked-by:
-
Chris Phlipot authored
Some implementations of libc do not support the 'm' width modifier as part of the scanf string format specifier. This can cause the parsing to fail. Since the parser never checks if the scanf parsing was successesful, this can result in a crash. Change the comm string to be allocated as a fixed size instead of dynamically using 'm' scanf width modifier. This can be safely done since comm size is limited to 16 bytes by TASK_COMM_LEN within the kernel. This change prevents perf from crashing when linked against bionic as well as reduces the total number of heap allocations and frees invoked while accomplishing the same task. Signed-off-by:
Chris Phlipot <cphlipot0@gmail.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20180830021950.15563-1-cphlipot0@gmail.comSigned-off-by:
-
Chris Phlipot authored
In the write to the output_fd in the error condition of record_saved_cmdline(), we are writing 8 bytes from a memory location on the stack that contains a primitive that is only 4 bytes in size. Change the primitive to 8 bytes in size to match the size of the write in order to avoid reading unknown memory from the stack. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
We were emitting 4 lines, two of them misleading: make: Entering directory '/home/acme/git/perf/tools/perf' <SNIP> INSTALL lib INSTALL include/bpf INSTALL lib INSTALL examples/bpf <SNIP> make: Leaving directory '/home/acme/git/perf/tools/perf' Make it more compact by showing just two lines: make: Entering directory '/home/acme/git/perf/tools/perf' INSTALL bpf-headers INSTALL bpf-examples make: Leaving directory '/home/acme/git/perf/tools/perf' Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: https://lkml.kernel.org/n/tip-0nvkyciqdkrgy829lony5925@git.kernel.orgSigned-off-by: -
Hisao Tanabe authored
If evsel is NULL, we should return NULL to avoid a NULL pointer dereference a bit later in the code. Signed-off-by:
Hisao Tanabe <xtanabe@gmail.com> Acked-by:
Namhyung Kim <namhyung@kernel.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Wang Nan <wangnan0@huawei.com> Fixes: 03e0a7df ("perf tools: Introduce bpf-output event") LPU-Reference: 20180824154556.23428-1-xtanabe@gmail.com Link: https://lkml.kernel.org/n/tip-e5plzjhx6595a5yjaf22jss3@git.kernel.orgSigned-off-by:
-
Kim Phillips authored
The new syscall table support for arm64 mistakenly used the system's asm-generic/unistd.h file when processing the tools/arch/arm64/include/uapi/asm/unistd.h file's include directive: #include <asm-generic/unistd.h> See "Committer notes" section of commit 2b588243 "perf arm64: Generate system call table from asm/unistd.h" for more details. This patch removes the committer's temporary workaround, and instructs the host compiler to search the build tree's include path for the right copy of the unistd.h file, instead of the one on the system's /usr/include path. It thus fixes the committer's test that cross-builds an arm64 perf on an x86 platform running Ubuntu 14.04.5 LTS with an old toolchain: $ tools/perf/arch/arm64/entry/syscalls/mksyscalltbl /gcc-linaro-5.4.1-2017.05-x86_64_aarch64-linux-gnu/bin/aarch64-linux-gnu-gcc gcc `pwd`/tools tools/arch/arm64/include/uapi/asm/unistd.h | grep bpf [280] = "bpf", Signed-off-by:
-
Jiri Olsa authored
We can safely enable the breakpoint back for both the fail and success paths by checking only the bp->attr.disabled, which either holds the new 'requested' disabled state or the original breakpoint state. Committer testing: At the end of the series, the 'perf test' entry introduced as the first patch now runs to completion without finding the fixed issues: # perf test "bp modify" 62: x86 bp modify : Ok # In verbose mode: # perf test -v "bp modify" 62: x86 bp modify : --- start --- test child forked, pid 5161 rip 5950a0, bp_1 0x5950a0 in bp_1 rip 5950a0, bp_1 0x5950a0 in bp_1 test child finished with 0 ---- end ---- x86 bp modify: Ok Suggested-by:
Oleg Nesterov <oleg@redhat.com> Acked-by:
Jiri Olsa <jolsa@kernel.org> Tested-by:
-
Jiri Olsa authored
Currently we enable the breakpoint back only if the breakpoint modification was successful. If it fails we can leave the breakpoint in disabled state with attr->disabled == 0. We can safely enable the breakpoint back for both the fail and success paths by checking the bp->attr.disabled, which either holds the new 'requested' disabled state or the original breakpoint state. Suggested-by:
Frederic Weisbecker <frederic@kernel.org> Acked-by:
-
Jiri Olsa authored
Once the breakpoint was succesfully modified, the attr->disabled value is in bp->attr.disabled. So there's no reason to set it again, removing that. Signed-off-by:
-
Jiri Olsa authored
We need to change the breakpoint even if the attr with new fields has disabled set to true. Current code prevents following user code to change the breakpoint address: ptrace(PTRACE_POKEUSER, child, offsetof(struct user, u_debugreg[0]), addr_1) ptrace(PTRACE_POKEUSER, child, offsetof(struct user, u_debugreg[0]), addr_2) ptrace(PTRACE_POKEUSER, child, offsetof(struct user, u_debugreg[7]), dr7) The first PTRACE_POKEUSER creates the breakpoint with attr.disabled set to true: ptrace_set_breakpoint_addr(nr = 0) struct perf_event *bp = t->ptrace_bps[nr]; ptrace_register_breakpoint(..., disabled = true) ptrace_fill_bp_fields(..., disabled) register_user_hw_breakpoint So the second PTRACE_POKEUSER will be omitted: ptrace_set_breakpoint_addr(nr = 0) struct perf_event *bp = t->ptrace_bps[nr]; struct perf_event_attr attr = bp->attr; modify_user_hw_breakpoint(bp, &attr) if (!attr->disabled) modify_user_hw_breakpoint_check Reported-by:Milind Chabbi <chabbi.milind@gmail.com> Signed-off-by:
-
Jiri Olsa authored
Adding to tests that aims on kernel breakpoint modification bugs. First test creates HW breakpoint, tries to change it and checks it was properly changed. It aims on kernel issue that prevents HW breakpoint to be changed via ptrace interface. The first test forks, the child sets itself as ptrace tracee and waits in signal for parent to trace it, then it calls bp_1 and quits. The parent does following steps: - creates a new breakpoint (id 0) for bp_2 function - changes that breakpoint to bp_1 function - waits for the breakpoint to hit and checks it has proper rip of bp_1 function This test aims on an issue in kernel preventing to change disabled breakpoints Second test mimics the first one except for few steps in the parent: - creates a new breakpoint (id 0) for bp_1 function - changes that breakpoint to bogus (-1) address - waits for the breakpoint to hit and checks it has proper rip of bp_1 function This test aims on an issue in kernel disabling enabled breakpoint after unsuccesful change. Committer testing: # uname -a Linux jouet 4.18.0-rc8-00002-g1236568e #12 SMP Tue Aug 7 14:08:26 -03 2018 x86_64 x86_64 x86_64 GNU/Linux # perf test -v "bp modify" 62: x86 bp modify : --- start --- test child forked, pid 25671 in bp_1 tracee exited prematurely 2 FAILED arch/x86/tests/bp-modify.c:209 modify test 1 failed test child finished with -1 ---- end ---- x86 bp modify: FAILED! # Signed-off-by:
-
Martin Liška authored
The patch changes the parsing of: callq *0x8(%rbx) from: 0.26 │ → callq *8 to: 0.26 │ → callq *0x8(%rbx) in this case an address is followed by a register, thus one can't parse only the address. Committer testing: 1) run 'perf record sleep 10' 2) before applying the patch, run: perf annotate --stdio2 > /tmp/before 3) after applying the patch, run: perf annotate --stdio2 > /tmp/after 4) diff /tmp/before /tmp/after: --- /tmp/before 2018-08-28 11:16:03.238384143 -0300 +++ /tmp/after 2018-08-28 11:15:39.335341042 -0300 @@ -13274,7 +13274,7 @@ ↓ jle 128 hash_value = hash_table->hash_func (key); mov 0x8(%rsp),%rdi - 0.91 → callq *30 + 0.91 → callq *0x30(%r12) mov $0x2,%r8d cmp $0x2,%eax node_hash = hash_table->hashes[node_index]; @@ -13848,7 +13848,7 @@ mov %r14,%rdi sub %rbx,%r13 mov %r13,%rdx - → callq *38 + → callq *0x38(%r15) cmp %rax,%r13 1.91 ↓ je 240 1b4: mov $0xffffffff,%r13d @@ -14026,7 +14026,7 @@ mov %rcx,-0x500(%rbp) mov %r15,%rsi mov %r14,%rdi - → callq *38 + → callq *0x38(%rax) mov -0x500(%rbp),%rcx cmp %rax,%rcx ↓ jne 9b0 <SNIP tons of other such cases> Signed-off-by:Martin Liška <mliska@suse.cz> Tested-by:
-
- 23 Aug, 2018 1 commit
-
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo-4.19-20180820' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo: LLVM/clang/eBPF: (Arnaldo Carvalho de Melo) - Allow passing options to llc in addition to to clang. Hardware tracing: (Jack Henschel) - Improve error message for PMU address filters, clarifying availability of that feature in hardware having hardware tracing such as Intel PT. Python interface: (Jiri Olsa) - Fix read_on_cpu() interface. ELF/DWARF libraries: (Jiri Olsa) - Fix handling of the combo compressed module file + decompressed associated debuginfo file. Build (Rasmus Villemoes) - Disable parallelism for 'make clean', avoiding multiple submakes deleting the same files and causing the build to fail on systems such as Yocto. Kernel ABI copies: (Arnaldo Carvalho de Melo) - Update tools's copy of x86's cpufeatures.h. - Update arch/x86/lib/memcpy_64.S copy used in 'perf bench mem memcpy'. Miscellaneous: (Steven Rostedt) - Change libtraceevent to SPDX License format. Signed-off-by:
Ingo Molnar <mingo@kernel.org>
-
- 20 Aug, 2018 21 commits
-
-
Arnaldo Carvalho de Melo authored
To bring in the change made in this cset: Fixes: a7bea830 ("x86/asm/64: Use 32-bit XOR to zero registers") CC /tmp/build/perf/bench/mem-memcpy-x86-64-asm.o LD /tmp/build/perf/bench/perf-in.o LD /tmp/build/perf/perf-in.o LINK /tmp/build/perf/perf Silencing this perf build warning: Warning: Kernel ABI header at 'tools/arch/x86/lib/memcpy_64.S' differs from latest version at 'arch/x86/lib/memcpy_64.S' diff -u tools/arch/x86/lib/memcpy_64.S arch/x86/lib/memcpy_64.S Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jan Beulich <JBeulich@suse.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: https://lkml.kernel.org/n/tip-sad22dudoz71qr3tsnlqtkia@git.kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
To get the changes in the following csets: 301d328a ("x86/cpufeatures: Add EPT_AD feature bit") 706d5168 ("x86/speculation: Support Enhanced IBRS on future CPUs") No tools were affected, copy it to silence this perf tool build warning: Warning: Kernel ABI header at 'tools/arch/x86/include/asm/cpufeatures.h' differs from latest version at 'arch/x86/include/asm/cpufeatures.h' diff -u tools/arch/x86/include/asm/cpufeatures.h arch/x86/include/asm/cpufeatures.h Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Feiner <pfeiner@google.com> Cc: Sai Praneeth <sai.praneeth.prakhya@intel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Wang Nan <wangnan0@huawei.com> Link: https://lkml.kernel.org/n/tip-bvs8wgd5wp4lz9f0xf1iug5r@git.kernel.orgSigned-off-by:
-
Jiri Olsa authored
Jaroslav reported errors from valgrind over perf python script: # echo 0 > /sys/devices/system/cpu/cpu4/online # valgrind ./test.py ==7524== Memcheck, a memory error detector ... ==7524== Command: ./test.py ==7524== pid 7526 exited ==7524== Invalid read of size 8 ==7524== at 0xCC2C2B3: perf_mmap__read_forward (evlist.c:780) ==7524== by 0xCC2A681: pyrf_evlist__read_on_cpu (python.c:959) ... ==7524== Address 0x65c4868 is 16 bytes after a block of size 459,36.. ==7524== at 0x4C2B955: calloc (vg_replace_malloc.c:711) ==7524== by 0xCC2F484: zalloc (util.h:35) ==7524== by 0xCC2F484: perf_evlist__alloc_mmap (evlist.c:978) ... The reason for this is in the python interface, that allows a script to pass arbitrary cpu number, which is then used to access struct perf_evlist::mmap array. That's obviously wrong and works only when if all cpus are available and fails if some cpu is missing, like in the example above. This patch makes pyrf_evlist__read_on_cpu() search the evlist's maps array for the proper map to access. It's linear search at the moment. Based on the way how is the read_on_cpu used, I don't think we need to be fast in here. But we could add some hash in the middle to make it fast/er. We don't allow python interface to set write_backward event attribute, so it's safe to check only evlist's mmaps. Reported-by:
Jaroslav Škarvada <jskarvad@redhat.com> Signed-off-by:
-
Jiri Olsa authored
Store the real cpu number in 'struct perf_mmap', which will be used by python interface that allows user to read a particular memory map for given cpu. Signed-off-by:
-
Jiri Olsa authored
Having comp carrying the compression ID, we no longer need return the extension. Removing it and updating the automated test. Signed-off-by:
-
Jiri Olsa authored
Add implementation of the is_compressed callback for gzip. Signed-off-by:
-
Jiri Olsa authored
Add implementation of the is_compressed callback for lzma. Signed-off-by:
-
Jiri Olsa authored
Add is_compressed callback to the compressions array, that returns 0 if the file is compressed or != 0 if not. The new callback is used to recognize the situation when we have a 'compressed' object, like: /lib/modules/.../drivers/net/ethernet/intel/igb/igb.ko.xz but we need to read its debug data from debuginfo files, which might not be compressed, like: /root/.debug/.build-id/d6/...c4b301f/debug So even for a 'compressed' object we read debug data from a plain uncompressed object. To keep this transparent, we detect this in decompress_kmodule() and return the file descriptor to the uncompressed file. Signed-off-by:
-
Jiri Olsa authored
We will add a compression check in the following patch and it makes it easier if the file processing is done in a single place. It also makes the current code simpler. The decompress_kmodule function now returns the fd of the uncompressed file and the file name in the pathname arg, if it's provided. Signed-off-by:
-
Jiri Olsa authored
Once we parsed out the compression ID, we dont need to iterate all available compressions and we can call it directly. Signed-off-by:
-
Jiri Olsa authored
Add comp to 'struct dso' to hold the compression index. It will be used in the following patches. Signed-off-by:
-
Jiri Olsa authored
Store a decompression ID in 'struct kmod_path', so it can be later stored in 'struct dso'. Switch 'struct kmod_path's 'comp' from 'bool' to 'int' to return the compressions array index. Add 0 index item into compressions array, so that the comp usage stays as it was: 0 - no compression, != 0 compression index. Update the kmod_path tests. Committer notes: Use a designated initializer + terminating comma, e.g. { .fmt = NULL, }, to fix the build in several distros: centos:6: util/dso.c:201: error: missing initializer centos:6: util/dso.c:201: error: (near initialization for 'compressions[0].decompress') debian:9: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] fedora:25: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] fedora:26: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] fedora:27: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] oraclelinux:6: util/dso.c:201: error: missing initializer oraclelinux:6: util/dso.c:201: error: (near initialization for 'compressions[0].decompress') ubuntu:12.04.5: util/dso.c:201:2: error: missing initializer [-Werror=missing-field-initializers] ubuntu:12.04.5: util/dso.c:201:2: error: (near initialization for 'compressions[0].decompress') [-Werror=missing-field-initializers] ubuntu:16.04: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] ubuntu:16.10: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] ubuntu:16.10: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] ubuntu:17.10: util/dso.c:201:24: error: missing field 'decompress' initializer [-Werror,-Wmissing-field-initializers] Signed-off-by: -
Jiri Olsa authored
There's no outside user of it. Signed-off-by:
-
Jiri Olsa authored
There's no outside user of it. Signed-off-by:
-
Jiri Olsa authored
There's no need to call dso__needs_decompress() twice in the function. Signed-off-by:
-
Jiri Olsa authored
There's no need to call dso__needs_decompress() twice in the function. Signed-off-by:
-
Jiri Olsa authored
There's no need to call dso__needs_decompress() twice in the function. Signed-off-by:
-
Steven Rostedt (VMware) authored
Replace the GPL text with SPDX tags in the tools/lib/traceevent files. Signed-off-by:
Steven Rostedt <rostedt@goodmis.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Tzvetomir Stoyanov (VMware) <tz.stoyanov@gmail.com> Cc: Yordan Karadzhov (VMware) <y.karadz@gmail.com> Cc: linux-trace-devel@vger.kernel.org Link: http://lkml.kernel.org/r/20180816111015.125e0f25@gandalf.local.homeSigned-off-by:
-
Arnaldo Carvalho de Melo authored
The newly added 'llvm.opts' variable allows passing options directly to llc, like needed to get sane DWARF in BPF ELF debug sections: With: [root@seventh perf]# cat ~/.perfconfig [llvm] dump-obj = true clang-opt = -g [root@seventh perf]# We get: [root@seventh perf]# perf trace -e tools/perf/examples/bpf/hello.c cat /etc/passwd > /dev/null LLVM: dumping tools/perf/examples/bpf/hello.o 0.000 __bpf_stdout__:Hello, world 0.015 __bpf_stdout__:Hello, world 0.187 __bpf_stdout__:Hello, world [root@seventh perf]# pahole tools/perf/examples/bpf/hello.o struct clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c) { clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566e.org clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c); /* 0 4 */ clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566e.org clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c); /* 4 4 */ clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566e.org clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c); /* 8 4 */ clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566e.org clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c); /* 12 4 */ clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566e.org clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c); /* 16 4 */ clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566e.org clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c); /* 20 4 */ clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566e.org clang version 8.0.0 (http://llvm.org/git/clang.git 8587270a739ee30c926a76d5657e65e85b560f6e) (http://llvm.org/git/llvm.git 0566eefef9c3777bd780ec4cbb9efa764633b76c); /* 24 4 */ /* size: 28, cachelines: 1, members: 7 */ /* last cacheline: 28 bytes */ }; [root@seventh perf]# Adding these options to be passed to llvm's llc: [root@seventh perf]# cat ~/.perfconfig [llvm] dump-obj = true clang-opt = -g opts = -mattr=dwarfris [root@seventh perf]# We get sane output: [root@seventh perf]# perf trace -e tools/perf/examples/bpf/hello.c cat /etc/passwd > /dev/null LLVM: dumping tools/perf/examples/bpf/hello.o 0.000 __bpf_stdout__:Hello, world 0.015 __bpf_stdout__:Hello, world 0.185 __bpf_stdout__:Hello, world [root@seventh perf]# pahole tools/perf/examples/bpf/hello.o struct bpf_map { unsigned int type; /* 0 4 */ unsigned int key_size; /* 4 4 */ unsigned int value_size; /* 8 4 */ unsigned int max_entries; /* 12 4 */ unsigned int map_flags; /* 16 4 */ unsigned int inner_map_idx; /* 20 4 */ unsigned int numa_node; /* 24 4 */ /* size: 28, cachelines: 1, members: 7 */ /* last cacheline: 28 bytes */ }; [root@seventh perf]# Cc: Alexei Starovoitov <ast@fb.com> Cc: Daniel Borkmann <daniel@iogearbox.net> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Martin KaFai Lau <kafai@fb.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com>, Cc: Yonghong Song <yhs@fb.com> Link: https://lkml.kernel.org/n/tip-0lrwmrip4dru1651rm8xa7tq@git.kernel.orgSigned-off-by: -
Jack Henschel authored
This is the second version of a patch that improves the error message of the perf events parser when the PMU hardware does not support address filters. Previously, the perf returned the following error: $ perf record -e intel_pt// --filter 'filter sys_write' --filter option should follow a -e tracepoint or HW tracer option This implies there is some syntax error present in the command line, which is not true. Rather, notify the user that the CPU does not have support for this feature. For example, Intel chips based on the Broadwell micro-archticture have the Intel PT PMU, but do not support address filtering. Now, perf prints the following error message: $ perf record -e intel_pt// --filter 'filter sys_write' This CPU does not support address filtering Signed-off-by:
Jack Henschel <jackdev@mailbox.org> Tested-by:
-
Rasmus Villemoes authored
The Yocto build system does a 'make clean' when rebuilding due to changed dependencies, and that consistently fails for me (causing the whole BSP build to fail) with errors such as | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory | find: '[...]/perf/1.0-r9/perf-1.0/plugin_mac80211.so': No such file or directory | find: find: '[...]/perf/1.0-r9/perf-1.0/libtraceevent.a''[...]/perf/1.0-r9/perf-1.0/libtraceevent.a': No such file or directory: No such file or directory | [...] | find: cannot delete '/mnt/xfs/devel/pil/yocto/tmp-glibc/work/wandboard-oe-linux-gnueabi/perf/1.0-r9/perf-1.0/util/.pstack.o.cmd': No such file or directory Apparently (despite the comment), 'make clean' ends up launching multiple sub-makes that all want to remove the same things - perhaps this only happens in combination with a O=... parameter. In any case, we don't lose much by explicitly disabling the parallelism for the clean target, and it makes automated builds much more reliable. Signed-off-by:
Rasmus Villemoes <linux@rasmusvillemoes.dk> Acked-by:
-
- 18 Aug, 2018 1 commit
-
-
Ingo Molnar authored
Merge tag 'perf-core-for-mingo-4.19-20180815' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo: kernel: - kallsyms, x86: Export addresses of PTI entry trampolines (Alexander Shishkin) - kallsyms: Simplify update_iter_mod() (Adrian Hunter) - x86: Add entry trampolines to kcore (Adrian Hunter) Hardware tracing: - Fix auxtrace queue resize (Adrian Hunter) Arch specific: - Fix uninitialized ARM SPE record error variable (Kim Phillips) - Fix trace event post-processing in powerpc (Sandipan Das) Build: - Fix check-headers.sh AND list path of execution (Alexander Kapshuk) - Remove -mcet and -fcf-protection when building the python binding with older clang versions (Arnaldo Carvalho de Melo) - Make check-headers.sh check based on kernel dir (Jiri Olsa) - Move syscall_64.tbl check into check-headers.sh (Jiri Olsa) Infrastructure: - Check for null when copying nsinfo. (Benno Evers) Libraries: - Rename libtraceevent prefixes, prep work for making it a shared library generaly available (Tzvetomir Stoyanov (VMware)) Signed-off-by:
-
- 14 Aug, 2018 1 commit
-
-
Adrian Hunter authored
Without program headers for PTI entry trampoline pages, the trampoline virtual addresses do not map to anything. Example before: sudo gdb --quiet vmlinux /proc/kcore Reading symbols from vmlinux...done. [New process 1] Core was generated by `BOOT_IMAGE=/boot/vmlinuz-4.16.0 root=UUID=a6096b83-b763-4101-807e-f33daff63233'. #0 0x0000000000000000 in irq_stack_union () (gdb) x /21ib 0xfffffe0000006000 0xfffffe0000006000: Cannot access memory at address 0xfffffe0000006000 (gdb) quit After: sudo gdb --quiet vmlinux /proc/kcore [sudo] password for ahunter: Reading symbols from vmlinux...done. [New process 1] Core was generated by `BOOT_IMAGE=/boot/vmlinuz-4.16.0-fix-4-00005-gd6e65a8b4072 root=UUID=a6096b83-b7'. #0 0x0000000000000000 in irq_stack_union () (gdb) x /21ib 0xfffffe0000006000 0xfffffe0000006000: swapgs 0xfffffe0000006003: mov %rsp,-0x3e12(%rip) # 0xfffffe00000021f8 0xfffffe000000600a: xchg %ax,%ax 0xfffffe000000600c: mov %cr3,%rsp 0xfffffe000000600f: bts $0x3f,%rsp 0xfffffe0000006014: and $0xffffffffffffe7ff,%rsp 0xfffffe000000601b: mov %rsp,%cr3 0xfffffe000000601e: mov -0x3019(%rip),%rsp # 0xfffffe000000300c 0xfffffe0000006025: pushq $0x2b 0xfffffe0000006027: pushq -0x3e35(%rip) # 0xfffffe00000021f8 0xfffffe000000602d: push %r11 0xfffffe000000602f: pushq $0x33 0xfffffe0000006031: push %rcx 0xfffffe0000006032: push %rdi 0xfffffe0000006033: mov $0xffffffff91a00010,%rdi 0xfffffe000000603a: callq 0xfffffe0000006046 0xfffffe000000603f: pause 0xfffffe0000006041: lfence 0xfffffe0000006044: jmp 0xfffffe000000603f 0xfffffe0000006046: mov %rdi,(%rsp) 0xfffffe000000604a: retq (gdb) quit In addition, entry trampolines all map to the same page. Represent that by giving the corresponding program headers in kcore the same offset. This has the benefit that, when perf tools uses /proc/kcore as a source for kernel object code, samples from different CPU trampolines are aggregated together. Note, such aggregation is normal for profiling i.e. people want to profile the object code, not every different virtual address the object code might be mapped to (across different processes for example). Notes by PeterZ: This also adds the KCORE_REMAP functionality. Signed-off-by:Adrian Hunter <adrian.hunter@intel.com> Acked-by:
Andi Kleen <ak@linux.intel.com> Acked-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: H. Peter Anvin <hpa@zytor.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Joerg Roedel <joro@8bytes.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: x86@kernel.org Link: http://lkml.kernel.org/r/1528289651-4113-4-git-send-email-adrian.hunter@intel.comSigned-off-by:
-
