Re: [PATCH v7 0/5] perf: script: Intro capstone disasm engine to show instruction trace

From: Adrian Hunter
Date: Fri Feb 16 2024 - 06:20:55 EST

On 9/02/24 20:49, Namhyung Kim wrote:
> Hello,
>
> For some reason, I cannot see Changbin's email in my mailbox. :(
>
> On Tue, Feb 6, 2024 at 3:50 AM Adrian Hunter <adrian.hunter@xxxxxxxxx> wrote:
>>
>> On 6/02/24 13:29, Changbin Du wrote:
>>> This series introduces capstone disassembler engine to print instructions of
>>> Intel PT trace, which was printed via the XED tool.
>>>
>>> The advantages compared to XED tool:
>>> * Support arm, arm64, x86-32, x86_64, s390 (more could be supported),
>>> xed only for x86_64.
>>> * More friendly to read. Immediate address operands are shown as symbol+offs.
>>>
>>> Display raw instructions:
>>> $ sudo perf record --event intel_pt//u -- ls
>>> $ sudo perf script --insn-trace
>>> perf 17423 [000] 423271.557970005: 7f2d95f16217 __GI___ioctl+0x7 (/lib/x86_64-linux-gnu/libc-2.27.so) insn: 48 3d 01 f0 ff ff
>>> perf 17423 [000] 423271.557970005: 7f2d95f1621d __GI___ioctl+0xd (/lib/x86_64-linux-gnu/libc-2.27.so) insn: 73 01
>>> perf 17423 [000] 423271.557970338: 7f2d95f1621f __GI___ioctl+0xf (/lib/x86_64-linux-gnu/libc-2.27.so) insn: c3
>>> perf 17423 [000] 423271.557970338: 5593ad3346d7 perf_evsel__enable_cpu+0x97 (/work/linux/tools/perf/perf) insn: 85 c0
>>> perf 17423 [000] 423271.557970338: 5593ad3346d9 perf_evsel__enable_cpu+0x99 (/work/linux/tools/perf/perf) insn: 75 12
>>> perf 17423 [000] 423271.557970338: 5593ad3346db perf_evsel__enable_cpu+0x9b (/work/linux/tools/perf/perf) insn: 49 8b 84 24 a8 00 00 00
>>> perf 17423 [000] 423271.557970338: 5593ad3346e3 perf_evsel__enable_cpu+0xa3 (/work/linux/tools/perf/perf) insn: 48 8b 50 20
>>>
>>> Display mnemonic instructions:
>>> $ sudo perf script --insn-trace=disasm
>>> perf 17423 [000] 423271.557970005: 7f2d95f16217 __GI___ioctl+0x7 (/lib/x86_64-linux-gnu/libc-2.27.so) cmpq $-0xfff, %rax
>>> perf 17423 [000] 423271.557970005: 7f2d95f1621d __GI___ioctl+0xd (/lib/x86_64-linux-gnu/libc-2.27.so) jae __GI___ioctl+0x10
>>> perf 17423 [000] 423271.557970338: 7f2d95f1621f __GI___ioctl+0xf (/lib/x86_64-linux-gnu/libc-2.27.so) retq
>>> perf 17423 [000] 423271.557970338: 5593ad3346d7 perf_evsel__enable_cpu+0x97 (/work/linux/tools/perf/perf) testl %eax, %eax
>>> perf 17423 [000] 423271.557970338: 5593ad3346d9 perf_evsel__enable_cpu+0x99 (/work/linux/tools/perf/perf) jne perf_evsel__enable_cpu+0xad
>>> perf 17423 [000] 423271.557970338: 5593ad3346db perf_evsel__enable_cpu+0x9b (/work/linux/tools/perf/perf) movq 0xa8(%r12), %rax
>>> perf 17423 [000] 423271.557970338: 5593ad3346e3 perf_evsel__enable_cpu+0xa3 (/work/linux/tools/perf/perf) movq 0x20(%rax), %rdx
>>> perf 17423 [000] 423271.557970338: 5593ad3346e7 perf_evsel__enable_cpu+0xa7 (/work/linux/tools/perf/perf) cmpl %edx, %ebx
>>> perf 17423 [000] 423271.557970338: 5593ad3346e9 perf_evsel__enable_cpu+0xa9 (/work/linux/tools/perf/perf) jl perf_evsel__enable_cpu+0x60
>>> perf 17423 [000] 423271.557970338: 5593ad3346eb perf_evsel__enable_cpu+0xab (/work/linux/tools/perf/perf) xorl %eax, %eax
>>>
>>> v7:
>>> - remove redundant space for raw instruction.
>>> - rename sample__fprintf_insn() as sample__fprintf_insn_asm() and make it empty if libcapstone is not available.
>>> v6:
>>> - make '-F +disasm' fatal if libcapstone is not supported. (Adrian Hunter)
>>> - display with perf version --build-options. (Adrian Hunter)
>>> - exclude libcapstone from make_minimal. (Adrian Hunter)
>>> v5:
>>> - fixes and improments suggested by Adrian Hunter
>>> v4:
>>> - rename 'insn_disam' to 'disasm' (Adrian Hunter)
>>> v3:
>>> - fix s390 detection. (Thomas Richter)
>>> v2:
>>> - add a new field 'insn_disam' instead of changing the default output.
>>> - preserve the old --xed option.
>>>
>>> Changbin Du (5):
>>> perf: build: introduce the libcapstone
>>> perf: util: use capstone disasm engine to show assembly instructions
>>> perf: script: add field 'disasm' to display mnemonic instructions
>>> perf: script: add raw|disasm arguments to --insn-trace option
>>> perf: script: prefer capstone to XED
>>>
>>> tools/build/Makefile.feature | 2 +
>>> tools/build/feature/Makefile | 4 +
>>> tools/build/feature/test-all.c | 4 +
>>> tools/build/feature/test-libcapstone.c | 11 ++
>>> tools/perf/Documentation/perf-intel-pt.txt | 14 ++-
>>> tools/perf/Documentation/perf-script.txt | 20 +--
>>> tools/perf/Makefile.config | 21 ++++
>>> tools/perf/Makefile.perf | 3 +
>>> tools/perf/builtin-script.c | 45 +++++--
>>> tools/perf/builtin-version.c | 1 +
>>> tools/perf/tests/make | 4 +-
>>> tools/perf/ui/browsers/res_sample.c | 2 +-
>>> tools/perf/ui/browsers/scripts.c | 2 +-
>>> tools/perf/util/Build | 1 +
>>> tools/perf/util/print_insn.c | 135 +++++++++++++++++++++
>>> tools/perf/util/print_insn.h | 16 +++
>>> 16 files changed, 258 insertions(+), 27 deletions(-)
>>> create mode 100644 tools/build/feature/test-libcapstone.c
>>> create mode 100644 tools/perf/util/print_insn.c
>>> create mode 100644 tools/perf/util/print_insn.h
>>>
>>
>> For whole patch set:
>>
>> Reviewed-by: Adrian Hunter <adrian.hunter@xxxxxxxxx>
>
> Anyway, I see a build failure on Fedora 40.
>
> In file included from /usr/include/capstone/capstone.h:325,
> from util/print_insn.c:29:
> /usr/include/capstone/bpf.h:94:14: error: 'bpf_insn' defined as wrong
> kind of tag
> 94 | typedef enum bpf_insn {
> | ^~~~~~~~

libcapstone header files seem to trample on other namespaces.
We might have to wrap the libcapstone API at some point, but
for now I suggest avoiding including the conflicting headers
in the same C file. Below works for me:

diff --git a/tools/perf/util/print_insn.c b/tools/perf/util/print_insn.c
index c10ccda529b13..459e0e93d7b1b 100644
--- a/tools/perf/util/print_insn.c
+++ b/tools/perf/util/print_insn.c
@@ -7,7 +7,7 @@
#include <string.h>
#include <stdbool.h>
#include "debug.h"
-#include "event.h"
+#include "sample.h"
#include "symbol.h"
#include "machine.h"
#include "thread.h"
diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h
index 0df775b5c1105..df344262eaee0 100644
--- a/tools/perf/util/thread.h
+++ b/tools/perf/util/thread.h
@@ -13,7 +13,6 @@
#include <strlist.h>
#include <intlist.h>
#include "rwsem.h"
-#include "event.h"
#include "callchain.h"
#include <internal/rc_check.h>