linux.git/tools/perf/util/disasm.c, branch for-next

perf disasm: Allow configuring what disassemblers to use

2024-11-13T19:27:35+00:00

The perf tools annotation code used for a long time parsing the output
of binutils's objdump (or its reimplementations, like llvm's) to then
parse and augment it with samples, allow navigation, etc.

More recently disassemblers from the capstone and llvm (libraries, not
parsing the output of tools using those libraries to mimic binutils's
objdump output) were introduced.

So when all those methods are available, there is a static preference
for a series of attempts of disassembling a binary, with the 'llvm,
capstone, objdump' sequence being hard coded.

This patch allows users to change that sequence, specifying via a 'perf
config' 'annotate.disassemblers' entry which and in what order
disassemblers should be attempted.

As alluded to in the comments in the source code of this series, this
flexibility is useful for users and developers alike, elliminating the
requirement to rebuild the tool with some specific set of libraries to
see how the output of disassembling would be for one of these methods.

  root@x1:~# rm -f ~/.perfconfig
  root@x1:~# perf annotate -v --stdio2 update_load_avg
  
  symbol__disassemble:
    filename=/usr/lib/debug/lib/modules/6.11.4-201.fc40.x86_64/vmlinux,
    sym=update_load_avg, start=0xffffffffb6148fe0, en>
  annotating [0x6ff7170]
    /usr/lib/debug/lib/modules/6.11.4-201.fc40.x86_64/vmlinux :
    [0x7407ca0] update_load_avg
  Disassembled with llvm
  annotate.disassemblers=llvm,capstone,objdump
  Samples: 66  of event 'cpu_atom/cycles/P', 10000 Hz,
	Event count (approx.): 5185444, [percent: local period]
  update_load_avg()
    /usr/lib/debug/lib/modules/6.11.4-201.fc40.x86_64/vmlinux
  Percent       0xffffffff81148fe0 :
     1.61         pushq   %r15
                  pushq   %r14
     1.00         pushq   %r13
                  movl    %edx,%r13d
     1.90         pushq   %r12
                  pushq   %rbp
                  movq    %rsi,%rbp
                  pushq   %rbx
                  movq    %rdi,%rbx
                  subq    $0x18,%rsp
    15.14         movl    0x1a4(%rdi),%eax

  root@x1:~# perf config annotate.disassemblers=capstone
  root@x1:~# cat ~/.perfconfig
  # this file is auto-generated.
  [annotate]
	  disassemblers = capstone
  root@x1:~#
  root@x1:~# perf annotate -v --stdio2 update_load_avg
  
  Disassembled with capstone
  annotate.disassemblers=capstone
  Samples: 66  of event 'cpu_atom/cycles/P', 10000 Hz,
  Event count (approx.): 5185444, [percent: local period]
  update_load_avg()
  /usr/lib/debug/lib/modules/6.11.4-201.fc40.x86_64/vmlinux
  Percent       0xffffffff81148fe0 :
     1.61         pushq   %r15
                  pushq   %r14
     1.00         pushq   %r13
                  movl    %edx,%r13d
     1.90         pushq   %r12
                  pushq   %rbp
                  movq    %rsi,%rbp
                  pushq   %rbx
                  movq    %rdi,%rbx
                  subq    $0x18,%rsp
    15.14         movl    0x1a4(%rdi),%eax
  root@x1:~# perf config annotate.disassemblers=objdump,capstone
  root@x1:~# perf config annotate.disassemblers
  annotate.disassemblers=objdump,capstone
  root@x1:~# cat ~/.perfconfig
  # this file is auto-generated.
  [annotate]
	  disassemblers = objdump,capstone
  root@x1:~# perf annotate -v --stdio2 update_load_avg
  Executing: objdump  --start-address=0xffffffff81148fe0 \
		      --stop-address=0xffffffff811497aa  \
		      -d --no-show-raw-insn -S -C "$1"
  Disassembled with objdump
  annotate.disassemblers=objdump,capstone
  Samples: 66  of event 'cpu_atom/cycles/P', 10000 Hz,
  Event count (approx.): 5185444, [percent: local period]
  update_load_avg()
  /usr/lib/debug/lib/modules/6.11.4-201.fc40.x86_64/vmlinux
  Percent

                Disassembly of section .text:

                ffffffff81148fe0 :
                #define DO_ATTACH       0x4

                ffffffff81148fe0 :
                #define DO_ATTACH       0x4
                #define DO_DETACH       0x8

                /* Update task and its cfs_rq load average */
                static inline void update_load_avg(struct cfs_rq *cfs_rq,
						   struct sched_entity *se,
						   int flags)
                {
     1.61         push   %r15
                  push   %r14
     1.00         push   %r13
                  mov    %edx,%r13d
     1.90         push   %r12
                  push   %rbp
                  mov    %rsi,%rbp
                  push   %rbx
                  mov    %rdi,%rbx
                  sub    $0x18,%rsp
                }

                /* rq->task_clock normalized against any time
		   this cfs_rq has spent throttled */
                static inline u64 cfs_rq_clock_pelt(struct cfs_rq *cfs_rq)
                {
                if (unlikely(cfs_rq->throttle_count))
    15.14         mov    0x1a4(%rdi),%eax
  root@x1:~#

After adding a way to select the disassembler from the command line a
'perf test' comparing the output of the various diassemblers should be
introduced, to test these codebases.

Acked-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Athira Rajeev 
Cc: Jiri Olsa 
Cc: Kan Liang 
Cc: Namhyung Kim 
Cc: Steinar H. Gunderson 
Link: https://lore.kernel.org/r/20241111151734.1018476-4-acme@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf disasm: Define stubs for the LLVM and capstone disassemblers

2024-11-13T19:20:32+00:00

This reduces the number of ifdefs in the main symbol__disassemble()
method and paves the way for allowing the user to configure the
disassemblers of preference.

Acked-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Aditya Bodkhe 
Cc: Athira Rajeev 
Cc: Jiri Olsa 
Cc: Kan Liang 
Cc: Masami Hiramatsu (Google) 
Cc: Namhyung Kim 
Cc: Steinar H. Gunderson 
Link: https://lore.kernel.org/r/20241111151734.1018476-3-acme@kernel.org
[ Applied fixes from Masami Hiramatsu and Aditya Bodkhe for when capstone devel files are not available ]
Link: https://lore.kernel.org/r/B78FB6DF-24E9-4A3C-91C9-535765EC0E2A@ibm.com
Link: https://lore.kernel.org/r/173145729034.2747044.453926054000880254.stgit@mhiramat.roam.corp.google.com
Signed-off-by: Arnaldo Carvalho de Melo

perf disasm: Introduce symbol__disassemble_objdump()

2024-11-11T17:26:37+00:00

With the first disassemble method in perf, the parsing of objdump
output, just like we have for llvm and capstone.

This paves the way to allow the user to specify what disassemblers are
preferred and to also to at some point allow building without the
objdump method.

Acked-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Athira Rajeev 
Cc: Jiri Olsa 
Cc: Kan Liang 
Cc: Namhyung Kim 
Cc: Steinar H. Gunderson 
Link: https://lore.kernel.org/r/20241111151734.1018476-2-acme@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf dwarf-regs: Add EM_HOST and EF_HOST defines

2024-11-09T16:39:13+00:00

Computed from the build architecture defines, EM_HOST and EF_HOST give
values that can be used in dwarf register lookup. Place in
dwarf-regs.h so the value can be shared. Move some dwarf-regs.c
constants used for EM_HOST to dwarf-regs.h. Add CSky constants that
may be missing.

In disasm.c add an include of dwarf-regs.h as the included
arch/*/annotate/instructions.c files make use of the constants and we
want the elf.h/dwarf-regs.h dependency to be explicit.

Reviewed-by: Masami Hiramatsu (Google) 
Signed-off-by: Ian Rogers 
Cc: Anup Patel 
Cc: Yang Jihong 
Cc: Palmer Dabbelt 
Cc: David S. Miller 
Cc: Albert Ou 
Cc: Shenlin Liang 
Cc: Nick Terrell 
Cc: Guilherme Amadio 
Cc: Steinar H. Gunderson 
Cc: Changbin Du 
Cc: Alexander Lobakin 
Cc: Przemek Kitszel 
Cc: Huacai Chen 
Cc: Guo Ren 
Cc: Masahiro Yamada 
Cc: Will Deacon 
Cc: James Clark 
Cc: Mike Leach 
Cc: Chen Pei 
Cc: Leo Yan 
Cc: Oliver Upton 
Cc: Aditya Gupta 
Cc: Kajol Jain 
Cc: Athira Rajeev 
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-riscv@lists.infradead.org
Cc: Bibo Mao 
Cc: John Garry 
Cc: Atish Patra 
Cc: Dima Kogan 
Cc: Paul Walmsley 
Cc: Dr. David Alan Gilbert 
Cc: linux-csky@vger.kernel.org
Link: https://lore.kernel.org/r/20241108234606.429459-4-irogers@google.com
Signed-off-by: Namhyung Kim

perf disasm: Fix not cleaning up disasm_line in symbol__disassemble_raw()

2024-10-23T22:36:14+00:00

In symbol__disassemble_raw(), the created disasm_line should be
discarded before returning an error. When creating disasm_line fails,
break the loop and then release the created lines.

Fixes: 0b971e6bf1c3 ("perf annotate: Add support to capture and parse raw instruction in powerpc using dso__data_read_offset utility")
Signed-off-by: Li Huafei 
Tested-by: Athira Rajeev 
Cc: sesse@google.com
Cc: kjain@linux.ibm.com
Link: https://lore.kernel.org/r/20241019154157.282038-3-lihuafei1@huawei.com
Signed-off-by: Namhyung Kim

perf disasm: Use disasm_line__free() to properly free disasm_line

2024-10-23T22:36:06+00:00

symbol__disassemble_capstone_powerpc() goto the 'err' label when it
failed in the loop that created disasm_line, and then used free()
directly to free disasm_line. Since the structure disasm_line contains
members that allocate memory dynamically, this can result in a memory
leak. In fact, we can simply break the loop when it fails in the middle
of the loop, and disasm_line__free() will then be called to properly
free the created line. Other error paths do not need to consider freeing
disasm_line.

Fixes: c5d60de1813a ("perf annotate: Add support to use libcapstone in powerpc")
Signed-off-by: Li Huafei 
Tested-by: Athira Rajeev 
Cc: sesse@google.com
Cc: kjain@linux.ibm.com
Link: https://lore.kernel.org/r/20241019154157.282038-2-lihuafei1@huawei.com
Signed-off-by: Namhyung Kim

perf disasm: Use disasm_line__free() to properly free disasm_line

2024-10-23T22:35:38+00:00

The structure disasm_line contains members that require dynamically
allocated memory and need to be freed correctly using
disasm_line__free().

This patch fixes the incorrect release in
symbol__disassemble_capstone().

Fixes: 6d17edc113de ("perf annotate: Use libcapstone to disassemble")
Signed-off-by: Li Huafei 
Tested-by: Athira Rajeev 
Cc: sesse@google.com
Cc: kjain@linux.ibm.com
Link: https://lore.kernel.org/r/20241019154157.282038-1-lihuafei1@huawei.com
Signed-off-by: Namhyung Kim

perf build: Rename HAVE_DWARF_SUPPORT to HAVE_LIBDW_SUPPORT

2024-10-18T17:17:40+00:00

In Makefile.config for unwinding the name dwarf implies either
libunwind or libdw. Make it clearer that HAVE_DWARF_SUPPORT is really
just defined when libdw is present by renaming to HAVE_LIBDW_SUPPORT.

Signed-off-by: Ian Rogers 
Acked-by: Namhyung Kim 
Tested-by: Leo Yan 
Cc: Anup Patel 
Cc: Yang Jihong 
Cc: Palmer Dabbelt 
Cc: David S. Miller 
Cc: Albert Ou 
Cc: Shenlin Liang 
Cc: Nick Terrell 
Cc: Guilherme Amadio 
Cc: Steinar H. Gunderson 
Cc: Changbin Du 
Cc: Alexander Lobakin 
Cc: Przemek Kitszel 
Cc: Huacai Chen 
Cc: Guo Ren 
Cc: Masahiro Yamada 
Cc: Masami Hiramatsu (Google) 
Cc: Will Deacon 
Cc: James Clark 
Cc: Mike Leach 
Cc: Chen Pei 
Cc: Leo Yan 
Cc: Oliver Upton 
Cc: Aditya Gupta 
Cc: Kajol Jain 
Cc: Athira Rajeev 
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-riscv@lists.infradead.org
Cc: Bibo Mao 
Cc: John Garry 
Cc: Atish Patra 
Cc: Dima Kogan 
Cc: Paul Walmsley 
Cc: Dr. David Alan Gilbert 
Cc: linux-csky@vger.kernel.org
Link: https://lore.kernel.org/r/20241017001354.56973-11-irogers@google.com
Signed-off-by: Namhyung Kim

perf disasm: Fix capstone memory leak

2024-10-17T19:43:14+00:00

The insn argument passed to cs_disasm needs freeing. To support
accurately having count, add an additional free_count variable.

Fixes: c5d60de1813a ("perf annotate: Add support to use libcapstone in powerpc")
Signed-off-by: Ian Rogers 
Reviewed-by: James Clark 
Cc: David S. Miller 
Cc: Steinar H. Gunderson 
Cc: Alexander Lobakin 
Cc: Masami Hiramatsu (Google) 
Cc: Kajol Jain 
Cc: Athira Rajeev 
Cc: Hemant Kumar 
Link: https://lore.kernel.org/r/20241016235622.52166-2-irogers@google.com
Signed-off-by: Namhyung Kim

perf annotate: LLVM-based disassembler

2024-09-03T13:39:20+00:00

Support using LLVM as a disassembler method, allowing helperless
annotation in non-distro builds. (It is also much faster than
using libbfd or bfd objdump on binaries with a lot of debug
information.)

This is nearly identical to the output of llvm-objdump; there are
some very rare whitespace differences, some minor changes to demangling
(since we use perf's regular demangling and not LLVM's own) and
the occasional case where llvm-objdump makes a different choice
when multiple symbols share the same address.

It should work across all of LLVM's supported architectures, although
I've only tested 64-bit x86, and finding the right triple from perf's
idea of machine architecture can sometimes be a bit tricky. Ideally, we
should have some way of finding the triplet just from the file itself.

Committer notes:

Address this on 32-bit systems by using PRIu64 from inttypes.h

     3    17.58 almalinux:9-i386              : FAIL gcc version 11.4.1 20231218 (Red Hat 11.4.1-3) (GCC)
      util/llvm-c-helpers.cpp: In function ‘char* make_symbol_relative_string(dso*, const char*, u64, u64)’:
      util/llvm-c-helpers.cpp:150:52: error: format ‘%lx’ expects argument of type ‘long unsigned int’, but argument 5 has type ‘u64’ {aka
  +‘long long unsigned int’} [-Werror=format=]
        150 |                 snprintf(buf, sizeof(buf), "%s+0x%lx",
            |                                                  ~~^
            |                                                    |
            |                                                    long unsigned int
            |                                                  %llx
        151 |                          demangled ? demangled : sym_name, addr - base_addr);
            |                                                            ~~~~~~~~~~~~~~~~
            |                                                                 |
            |                                                                 u64 {aka long long unsigned int}
      cc1plus: all warnings being treated as errors

Signed-off-by: Steinar H. Gunderson 
Cc: Ian Rogers 
Link: https://lore.kernel.org/r/20240803152008.2818485-3-sesse@google.com
Signed-off-by: Arnaldo Carvalho de Melo