* [PATCH 1/2] perf, tools: Add support for skipping itrace instructions
@ 2016-02-27 0:32 Andi Kleen
2016-02-27 0:32 ` [PATCH 2/2] perf, tools, script: Add support for printing assembler Andi Kleen
0 siblings, 1 reply; 9+ messages in thread
From: Andi Kleen @ 2016-02-27 0:32 UTC (permalink / raw
To: acme
Cc: jolsa, namhyung, adrian.hunter, alexander.shishkin, linux-kernel,
Andi Kleen
From: Andi Kleen <ak@linux.intel.com>
When using perf script to look at PT traces it is often
useful to ignore the initialization code at the beginning.
On larger traces which may have many millions of instructions
in initialization code doing that in a pipeline can be very
slow, with perf script spending a lot of CPU time calling
printf and writing data.
This patch adds an extension to the --itrace argument
that skips 'n' events (instructions, branches or transactions)
at the beginning. This is much more efficient.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
tools/perf/Documentation/intel-pt.txt | 7 +++++++
tools/perf/util/auxtrace.c | 7 +++++++
tools/perf/util/auxtrace.h | 2 ++
tools/perf/util/intel-pt.c | 14 ++++++++++++--
4 files changed, 28 insertions(+), 2 deletions(-)
diff --git a/tools/perf/Documentation/intel-pt.txt b/tools/perf/Documentation/intel-pt.txt
index be764f9..c6c8318 100644
--- a/tools/perf/Documentation/intel-pt.txt
+++ b/tools/perf/Documentation/intel-pt.txt
@@ -672,6 +672,7 @@ The letters are:
d create a debug log
g synthesize a call chain (use with i or x)
l synthesize last branch entries (use with i or x)
+ s skip initial number of events
"Instructions" events look like they were recorded by "perf record -e
instructions".
@@ -730,6 +731,12 @@ from one sample to the next.
To disable trace decoding entirely, use the option --no-itrace.
+It is also possible to skip events generated (instructions, branches, transactions)
+at the beginning. This is useful to ignore initialization code.
+
+ --itrace=i0nss1000000
+
+skips the first million instructions.
dump option
-----------
diff --git a/tools/perf/util/auxtrace.c b/tools/perf/util/auxtrace.c
index 360fda0..a013c1f 100644
--- a/tools/perf/util/auxtrace.c
+++ b/tools/perf/util/auxtrace.c
@@ -939,6 +939,7 @@ void itrace_synth_opts__set_default(struct itrace_synth_opts *synth_opts)
synth_opts->period = PERF_ITRACE_DEFAULT_PERIOD;
synth_opts->callchain_sz = PERF_ITRACE_DEFAULT_CALLCHAIN_SZ;
synth_opts->last_branch_sz = PERF_ITRACE_DEFAULT_LAST_BRANCH_SZ;
+ synth_opts->initial_skip = 0;
}
/*
@@ -1063,6 +1064,12 @@ int itrace_parse_synth_opts(const struct option *opt, const char *str,
synth_opts->last_branch_sz = val;
}
break;
+ case 's':
+ synth_opts->initial_skip = strtoul(p, &endptr, 10);
+ if (p == endptr)
+ goto out_err;
+ p = endptr;
+ break;
case ' ':
case ',':
break;
diff --git a/tools/perf/util/auxtrace.h b/tools/perf/util/auxtrace.h
index b86f90db..235d3c3 100644
--- a/tools/perf/util/auxtrace.h
+++ b/tools/perf/util/auxtrace.h
@@ -68,6 +68,7 @@ enum itrace_period_type {
* @last_branch_sz: branch context size
* @period: 'instructions' events period
* @period_type: 'instructions' events period type
+ * @initial_skip: skip N events at the beginning.
*/
struct itrace_synth_opts {
bool set;
@@ -86,6 +87,7 @@ struct itrace_synth_opts {
unsigned int last_branch_sz;
unsigned long long period;
enum itrace_period_type period_type;
+ unsigned long initial_skip;
};
/**
diff --git a/tools/perf/util/intel-pt.c b/tools/perf/util/intel-pt.c
index 05d8158..a0c5e7d 100644
--- a/tools/perf/util/intel-pt.c
+++ b/tools/perf/util/intel-pt.c
@@ -100,6 +100,8 @@ struct intel_pt {
u64 cyc_bit;
u64 noretcomp_bit;
unsigned max_non_turbo_ratio;
+
+ unsigned long num_events;
};
enum switch_state {
@@ -972,6 +974,10 @@ static int intel_pt_synth_branch_sample(struct intel_pt_queue *ptq)
if (pt->branches_filter && !(pt->branches_filter & ptq->flags))
return 0;
+ if (pt->synth_opts.initial_skip &&
+ pt->num_events++ >= pt->synth_opts.initial_skip)
+ return 0;
+
event->sample.header.type = PERF_RECORD_SAMPLE;
event->sample.header.misc = PERF_RECORD_MISC_USER;
event->sample.header.size = sizeof(struct perf_event_header);
@@ -1196,14 +1202,18 @@ static int intel_pt_sample(struct intel_pt_queue *ptq)
ptq->have_sample = false;
if (pt->sample_instructions &&
- (state->type & INTEL_PT_INSTRUCTION)) {
+ (state->type & INTEL_PT_INSTRUCTION) &&
+ (!pt->synth_opts.initial_skip ||
+ pt->num_events++ >= pt->synth_opts.initial_skip)) {
err = intel_pt_synth_instruction_sample(ptq);
if (err)
return err;
}
if (pt->sample_transactions &&
- (state->type & INTEL_PT_TRANSACTION)) {
+ (state->type & INTEL_PT_TRANSACTION) &&
+ (!pt->synth_opts.initial_skip ||
+ pt->num_events++ >= pt->synth_opts.initial_skip)) {
err = intel_pt_synth_transaction_sample(ptq);
if (err)
return err;
--
2.5.0
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH 2/2] perf, tools, script: Add support for printing assembler
2016-02-27 0:32 [PATCH 1/2] perf, tools: Add support for skipping itrace instructions Andi Kleen
@ 2016-02-27 0:32 ` Andi Kleen
2016-02-29 10:27 ` Jiri Olsa
2016-02-29 14:56 ` Arnaldo Carvalho de Melo
0 siblings, 2 replies; 9+ messages in thread
From: Andi Kleen @ 2016-02-27 0:32 UTC (permalink / raw
To: acme
Cc: jolsa, namhyung, adrian.hunter, alexander.shishkin, linux-kernel,
Andi Kleen
From: Andi Kleen <ak@linux.intel.com>
When dumping PT traces with perf script it is very useful to see the
assembler for each sample, so that it is easily possible to follow
the control flow.
As using objdump is difficult and inefficient from perf script this
patch uses the udis86 library to implement assembler output.
The library can be downloaded from http://udis86.sourceforge.net/
The library is probed as an external dependency in the usual way. Then perf
script calls into it when needed, and handles callbacks to resolve
symbols.
% perf record -e intel_pt//u true
% perf script -F sym,symoff,ip,asm --itrace=i0ns | head
7fc7188b4190 _start+0x0 mov %rsp, %rdi
7fc7188b4193 _start+0x3 call _dl_start
7fc7188b7710 _dl_start+0x0 push %rbp
7fc7188b7711 _dl_start+0x1 mov %rsp, %rbp
7fc7188b7714 _dl_start+0x4 push %r15
7fc7188b7716 _dl_start+0x6 push %r14
7fc7188b7718 _dl_start+0x8 push %r13
7fc7188b771a _dl_start+0xa push %r12
7fc7188b771c _dl_start+0xc mov %rdi, %r12
7fc7188b771f _dl_start+0xf push %rbx
Current issues:
- Some jump references do not get resolved to symbols.
- udis86 release does not support STAC/CLAC, which are used in the kernel,
but there is a pending patch for it.
Cc: adrian.hunter@intel.com
Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
tools/build/Makefile.feature | 6 +-
tools/build/feature/Makefile | 8 ++-
tools/build/feature/test-all.c | 5 ++
tools/build/feature/test-udis86.c | 8 +++
tools/perf/Documentation/perf-script.txt | 4 +-
tools/perf/builtin-script.c | 108 +++++++++++++++++++++++++++++--
tools/perf/config/Makefile | 5 ++
7 files changed, 134 insertions(+), 10 deletions(-)
create mode 100644 tools/build/feature/test-udis86.c
diff --git a/tools/build/Makefile.feature b/tools/build/Makefile.feature
index 02db3cd..7de4fcb 100644
--- a/tools/build/Makefile.feature
+++ b/tools/build/Makefile.feature
@@ -54,7 +54,8 @@ FEATURE_TESTS ?= \
zlib \
lzma \
get_cpuid \
- bpf
+ bpf \
+ udis86
FEATURE_DISPLAY ?= \
dwarf \
@@ -73,7 +74,8 @@ FEATURE_DISPLAY ?= \
zlib \
lzma \
get_cpuid \
- bpf
+ bpf \
+ udis86
# Set FEATURE_CHECK_(C|LD)FLAGS-all for all FEATURE_TESTS features.
# If in the future we need per-feature checks/flags for features not
diff --git a/tools/build/feature/Makefile b/tools/build/feature/Makefile
index bf8f035..c89941e 100644
--- a/tools/build/feature/Makefile
+++ b/tools/build/feature/Makefile
@@ -35,7 +35,8 @@ FILES= \
test-zlib.bin \
test-lzma.bin \
test-bpf.bin \
- test-get_cpuid.bin
+ test-get_cpuid.bin \
+ test-udis86.bin
FILES := $(addprefix $(OUTPUT),$(FILES))
@@ -50,7 +51,7 @@ __BUILD = $(CC) $(CFLAGS) -Wall -Werror -o $@ $(patsubst %.bin,%.c,$(@F)) $(LDFL
###############################
$(OUTPUT)test-all.bin:
- $(BUILD) -fstack-protector-all -O2 -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell $(PKG_CONFIG) --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl -lz -llzma
+ $(BUILD) -fstack-protector-all -O2 -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell $(PKG_CONFIG) --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl -lz -llzma -ludis86
$(OUTPUT)test-hello.bin:
$(BUILD)
@@ -96,6 +97,9 @@ $(OUTPUT)test-numa_num_possible_cpus.bin:
$(OUTPUT)test-libunwind.bin:
$(BUILD) -lelf
+$(OUTPUT)test-udis86.bin:
+ $(BUILD) -ludis86
+
$(OUTPUT)test-libunwind-debug-frame.bin:
$(BUILD) -lelf
diff --git a/tools/build/feature/test-all.c b/tools/build/feature/test-all.c
index 81025ca..d67b1d5 100644
--- a/tools/build/feature/test-all.c
+++ b/tools/build/feature/test-all.c
@@ -129,6 +129,10 @@
# include "test-bpf.c"
#undef main
+#define main main_test_udis86
+# include "test-udis86.c"
+#endif
+
int main(int argc, char *argv[])
{
main_test_libpython();
@@ -158,6 +162,7 @@ int main(int argc, char *argv[])
main_test_lzma();
main_test_get_cpuid();
main_test_bpf();
+ main_test_udis86();
return 0;
}
diff --git a/tools/build/feature/test-udis86.c b/tools/build/feature/test-udis86.c
new file mode 100644
index 0000000..623c545
--- /dev/null
+++ b/tools/build/feature/test-udis86.c
@@ -0,0 +1,8 @@
+#include <udis86.h>
+
+int main(void)
+{
+ ud_t ud;
+ ud_init(&ud);
+ return 0;
+}
diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt
index 382ddfb..c834f4d 100644
--- a/tools/perf/Documentation/perf-script.txt
+++ b/tools/perf/Documentation/perf-script.txt
@@ -116,7 +116,7 @@ OPTIONS
--fields::
Comma separated list of fields to print. Options are:
comm, tid, pid, time, cpu, event, trace, ip, sym, dso, addr, symoff,
- srcline, period, iregs, brstack, brstacksym, flags.
+ srcline, period, iregs, brstack, brstacksym, flags, asm.
Field list can be prepended with the type, trace, sw or hw,
to indicate to which event type the field list applies.
e.g., -f sw:comm,tid,time,ip,sym and -f trace:time,cpu,trace
@@ -185,6 +185,8 @@ OPTIONS
The brstacksym is identical to brstack, except that the FROM and TO addresses are printed in a symbolic form if possible.
+ When asm is specified the assembler instruction of each sample is printed in disassembled form.
+
-k::
--vmlinux=<file>::
vmlinux pathname
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index c691214..9fd8cb3 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -24,6 +24,10 @@
#include <linux/bitmap.h>
#include "asm/bug.h"
+#ifdef HAVE_UDIS86
+#include <udis86.h>
+#endif
+
static char const *script_name;
static char const *generate_script_lang;
static bool debug_mode;
@@ -58,6 +62,7 @@ enum perf_output_field {
PERF_OUTPUT_IREGS = 1U << 14,
PERF_OUTPUT_BRSTACK = 1U << 15,
PERF_OUTPUT_BRSTACKSYM = 1U << 16,
+ PERF_OUTPUT_ASM = 1U << 17,
};
struct output_option {
@@ -81,6 +86,7 @@ struct output_option {
{.str = "iregs", .field = PERF_OUTPUT_IREGS},
{.str = "brstack", .field = PERF_OUTPUT_BRSTACK},
{.str = "brstacksym", .field = PERF_OUTPUT_BRSTACKSYM},
+ {.str = "asm", .field = PERF_OUTPUT_ASM},
};
/* default set to maintain compatibility with current format */
@@ -264,7 +270,11 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel,
"selected. Hence, no address to lookup the source line number.\n");
return -EINVAL;
}
-
+ if (PRINT_FIELD(ASM) && !PRINT_FIELD(IP)) {
+ pr_err("Display of assembler requested but sample IP is not\n"
+ "selected.\n");
+ return -EINVAL;
+ }
if ((PRINT_FIELD(PID) || PRINT_FIELD(TID)) &&
perf_evsel__check_stype(evsel, PERF_SAMPLE_TID, "TID",
PERF_OUTPUT_TID|PERF_OUTPUT_PID))
@@ -405,6 +415,89 @@ static void print_sample_iregs(union perf_event *event __maybe_unused,
}
}
+#ifdef HAVE_UDIS86
+
+struct perf_ud {
+ ud_t ud_obj;
+ struct thread *thread;
+ u8 cpumode;
+ int cpu;
+};
+
+static const char *dis_resolve(struct ud *u, uint64_t addr, int64_t *off)
+{
+ struct perf_ud *ud = container_of(u, struct perf_ud, ud_obj);
+ struct addr_location al;
+
+ memset(&al, 0, sizeof(struct addr_location));
+
+ thread__find_addr_map(ud->thread, ud->cpumode, MAP__FUNCTION, addr, &al);
+ if (!al.map)
+ thread__find_addr_map(ud->thread, ud->cpumode, MAP__VARIABLE,
+ addr, &al);
+ al.cpu = ud->cpu;
+ al.sym = NULL;
+
+ if (al.map)
+ al.sym = map__find_symbol(al.map, al.addr, NULL);
+
+ if (!al.sym)
+ return NULL;
+
+ if (addr < al.sym->end)
+ *off = addr - al.sym->start;
+ else
+ *off = addr - al.map->start - al.sym->start;
+ return al.sym->name;
+}
+#endif
+
+static void print_sample_asm(union perf_event *event __maybe_unused,
+ struct perf_sample *sample __maybe_unused,
+ struct thread *thread __maybe_unused,
+ struct perf_event_attr *attr __maybe_unused,
+ struct addr_location *al __maybe_unused,
+ struct machine *machine __maybe_unused)
+{
+#ifdef HAVE_UDIS86
+ static bool ud_initialized = false;
+ static struct perf_ud ud;
+ u8 buffer[32];
+ int len;
+ u64 offset;
+
+ if (!ud_initialized) {
+ ud_initialized = true;
+ ud_init(&ud.ud_obj);
+ ud_set_syntax(&ud.ud_obj, UD_SYN_ATT);
+ ud_set_sym_resolver(&ud.ud_obj, dis_resolve);
+ }
+ ud.thread = thread;
+ ud.cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
+ ud.cpu = sample->cpu;
+
+ if (!al->map || !al->map->dso)
+ return;
+ if (al->map->dso->data.status == DSO_DATA_STATUS_ERROR)
+ return;
+
+ /* Load maps to ensure dso->is_64_bit has been updated */
+ map__load(al->map, machine->symbol_filter);
+
+ offset = al->map->map_ip(al->map, sample->ip);
+ len = dso__data_read_offset(al->map->dso, machine,
+ offset, buffer, 32);
+ if (len <= 0)
+ return;
+
+ ud_set_mode(&ud.ud_obj, al->map->dso->is_64_bit ? 64 : 32);
+ ud_set_pc(&ud.ud_obj, sample->ip);
+ ud_set_input_buffer(&ud.ud_obj, buffer, len);
+ ud_disassemble(&ud.ud_obj);
+ printf("\t%s", ud_insn_asm(&ud.ud_obj));
+#endif
+}
+
static void print_sample_start(struct perf_sample *sample,
struct thread *thread,
struct perf_evsel *evsel)
@@ -636,7 +729,8 @@ static int perf_evlist__max_name_len(struct perf_evlist *evlist)
static void process_event(struct perf_script *script, union perf_event *event,
struct perf_sample *sample, struct perf_evsel *evsel,
- struct addr_location *al)
+ struct addr_location *al,
+ struct machine *machine)
{
struct thread *thread = al->thread;
struct perf_event_attr *attr = &evsel->attr;
@@ -664,7 +758,7 @@ static void process_event(struct perf_script *script, union perf_event *event,
if (is_bts_event(attr)) {
print_sample_bts(event, sample, evsel, thread, al);
- return;
+ goto print_rest;
}
if (PRINT_FIELD(TRACE))
@@ -687,11 +781,15 @@ static void process_event(struct perf_script *script, union perf_event *event,
if (PRINT_FIELD(IREGS))
print_sample_iregs(event, sample, thread, attr);
+print_rest:
if (PRINT_FIELD(BRSTACK))
print_sample_brstack(event, sample, thread, attr);
else if (PRINT_FIELD(BRSTACKSYM))
print_sample_brstacksym(event, sample, thread, attr);
+ if (PRINT_FIELD(ASM))
+ print_sample_asm(event, sample, thread, attr, al, machine);
+
printf("\n");
}
@@ -798,7 +896,7 @@ static int process_sample_event(struct perf_tool *tool,
if (scripting_ops)
scripting_ops->process_event(event, sample, evsel, &al);
else
- process_event(scr, event, sample, evsel, &al);
+ process_event(scr, event, sample, evsel, &al, machine);
out_put:
addr_location__put(&al);
@@ -1913,7 +2011,7 @@ int cmd_script(int argc, const char **argv, const char *prefix __maybe_unused)
"comma separated output fields prepend with 'type:'. "
"Valid types: hw,sw,trace,raw. "
"Fields: comm,tid,pid,time,cpu,event,trace,ip,sym,dso,"
- "addr,symoff,period,iregs,brstack,brstacksym,flags", parse_output_fields),
+ "addr,symoff,period,iregs,brstack,brstacksym,flags,asm", parse_output_fields),
OPT_BOOLEAN('a', "all-cpus", &system_wide,
"system-wide collection from all CPUs"),
OPT_STRING('S', "symbols", &symbol_conf.sym_list_str, "symbol[,symbol...]",
diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile
index 511141b..19bac7c 100644
--- a/tools/perf/config/Makefile
+++ b/tools/perf/config/Makefile
@@ -576,6 +576,11 @@ ifneq ($(filter -lbfd,$(EXTLIBS)),)
CFLAGS += -DHAVE_LIBBFD_SUPPORT
endif
+ifeq ($(feature-udis86), 1)
+ CFLAGS += -DHAVE_UDIS86
+ EXTLIBS += -ludis86
+endif
+
ifndef NO_ZLIB
ifeq ($(feature-zlib), 1)
CFLAGS += -DHAVE_ZLIB_SUPPORT
--
2.5.0
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] perf, tools, script: Add support for printing assembler
2016-02-27 0:32 ` [PATCH 2/2] perf, tools, script: Add support for printing assembler Andi Kleen
@ 2016-02-29 10:27 ` Jiri Olsa
2016-02-29 14:56 ` Arnaldo Carvalho de Melo
1 sibling, 0 replies; 9+ messages in thread
From: Jiri Olsa @ 2016-02-29 10:27 UTC (permalink / raw
To: Andi Kleen
Cc: acme, jolsa, namhyung, adrian.hunter, alexander.shishkin,
linux-kernel, Andi Kleen
On Fri, Feb 26, 2016 at 04:32:08PM -0800, Andi Kleen wrote:
> From: Andi Kleen <ak@linux.intel.com>
>
> When dumping PT traces with perf script it is very useful to see the
> assembler for each sample, so that it is easily possible to follow
> the control flow.
>
> As using objdump is difficult and inefficient from perf script this
> patch uses the udis86 library to implement assembler output.
> The library can be downloaded from http://udis86.sourceforge.net/
>
> The library is probed as an external dependency in the usual way. Then perf
> script calls into it when needed, and handles callbacks to resolve
> symbols.
>
> % perf record -e intel_pt//u true
> % perf script -F sym,symoff,ip,asm --itrace=i0ns | head
> 7fc7188b4190 _start+0x0 mov %rsp, %rdi
> 7fc7188b4193 _start+0x3 call _dl_start
> 7fc7188b7710 _dl_start+0x0 push %rbp
> 7fc7188b7711 _dl_start+0x1 mov %rsp, %rbp
> 7fc7188b7714 _dl_start+0x4 push %r15
> 7fc7188b7716 _dl_start+0x6 push %r14
> 7fc7188b7718 _dl_start+0x8 push %r13
> 7fc7188b771a _dl_start+0xa push %r12
> 7fc7188b771c _dl_start+0xc mov %rdi, %r12
> 7fc7188b771f _dl_start+0xf push %rbx
>
> Current issues:
> - Some jump references do not get resolved to symbols.
> - udis86 release does not support STAC/CLAC, which are used in the kernel,
> but there is a pending patch for it.
>
> Cc: adrian.hunter@intel.com
> Signed-off-by: Andi Kleen <ak@linux.intel.com>
> ---
> tools/build/Makefile.feature | 6 +-
> tools/build/feature/Makefile | 8 ++-
> tools/build/feature/test-all.c | 5 ++
> tools/build/feature/test-udis86.c | 8 +++
hi,
could you please split the udis86 detection from the rest?
thanks,
jirka
> tools/perf/Documentation/perf-script.txt | 4 +-
> tools/perf/builtin-script.c | 108 +++++++++++++++++++++++++++++--
> tools/perf/config/Makefile | 5 ++
> 7 files changed, 134 insertions(+), 10 deletions(-)
> create mode 100644 tools/build/feature/test-udis86.c
SNIP
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] perf, tools, script: Add support for printing assembler
2016-02-27 0:32 ` [PATCH 2/2] perf, tools, script: Add support for printing assembler Andi Kleen
2016-02-29 10:27 ` Jiri Olsa
@ 2016-02-29 14:56 ` Arnaldo Carvalho de Melo
2016-02-29 17:35 ` Andi Kleen
1 sibling, 1 reply; 9+ messages in thread
From: Arnaldo Carvalho de Melo @ 2016-02-29 14:56 UTC (permalink / raw
To: Andi Kleen
Cc: jolsa, namhyung, Adrian Hunter, alexander.shishkin, linux-kernel,
Andi Kleen
Em Fri, Feb 26, 2016 at 04:32:08PM -0800, Andi Kleen escreveu:
> From: Andi Kleen <ak@linux.intel.com>
>
> When dumping PT traces with perf script it is very useful to see the
> assembler for each sample, so that it is easily possible to follow
> the control flow.
>
> As using objdump is difficult and inefficient from perf script this
> patch uses the udis86 library to implement assembler output.
> The library can be downloaded from http://udis86.sourceforge.net/
>
> The library is probed as an external dependency in the usual way. Then perf
> script calls into it when needed, and handles callbacks to resolve
> symbols.
What kernel have you used for testing this? 4.5-rc? I'm having trouble
with intel_pt with 4.5.0-rc4 :-\
And while this looks like a great feature to have, have you considered
using what is in tools/perf/util/intel-pt-decoder/ somehow, so that we
don't end up adding one more dependency on another external library?
It would be great if we could, if done that way, perhaps at some point
we could stop using objdump somehow :-\
- Arnaldo
> % perf record -e intel_pt//u true
> % perf script -F sym,symoff,ip,asm --itrace=i0ns | head
> 7fc7188b4190 _start+0x0 mov %rsp, %rdi
> 7fc7188b4193 _start+0x3 call _dl_start
> 7fc7188b7710 _dl_start+0x0 push %rbp
> 7fc7188b7711 _dl_start+0x1 mov %rsp, %rbp
> 7fc7188b7714 _dl_start+0x4 push %r15
> 7fc7188b7716 _dl_start+0x6 push %r14
> 7fc7188b7718 _dl_start+0x8 push %r13
> 7fc7188b771a _dl_start+0xa push %r12
> 7fc7188b771c _dl_start+0xc mov %rdi, %r12
> 7fc7188b771f _dl_start+0xf push %rbx
>
> Current issues:
> - Some jump references do not get resolved to symbols.
> - udis86 release does not support STAC/CLAC, which are used in the kernel,
> but there is a pending patch for it.
>
> Cc: adrian.hunter@intel.com
> Signed-off-by: Andi Kleen <ak@linux.intel.com>
> ---
> tools/build/Makefile.feature | 6 +-
> tools/build/feature/Makefile | 8 ++-
> tools/build/feature/test-all.c | 5 ++
> tools/build/feature/test-udis86.c | 8 +++
> tools/perf/Documentation/perf-script.txt | 4 +-
> tools/perf/builtin-script.c | 108 +++++++++++++++++++++++++++++--
> tools/perf/config/Makefile | 5 ++
> 7 files changed, 134 insertions(+), 10 deletions(-)
> create mode 100644 tools/build/feature/test-udis86.c
>
> diff --git a/tools/build/Makefile.feature b/tools/build/Makefile.feature
> index 02db3cd..7de4fcb 100644
> --- a/tools/build/Makefile.feature
> +++ b/tools/build/Makefile.feature
> @@ -54,7 +54,8 @@ FEATURE_TESTS ?= \
> zlib \
> lzma \
> get_cpuid \
> - bpf
> + bpf \
> + udis86
>
> FEATURE_DISPLAY ?= \
> dwarf \
> @@ -73,7 +74,8 @@ FEATURE_DISPLAY ?= \
> zlib \
> lzma \
> get_cpuid \
> - bpf
> + bpf \
> + udis86
>
> # Set FEATURE_CHECK_(C|LD)FLAGS-all for all FEATURE_TESTS features.
> # If in the future we need per-feature checks/flags for features not
> diff --git a/tools/build/feature/Makefile b/tools/build/feature/Makefile
> index bf8f035..c89941e 100644
> --- a/tools/build/feature/Makefile
> +++ b/tools/build/feature/Makefile
> @@ -35,7 +35,8 @@ FILES= \
> test-zlib.bin \
> test-lzma.bin \
> test-bpf.bin \
> - test-get_cpuid.bin
> + test-get_cpuid.bin \
> + test-udis86.bin
>
> FILES := $(addprefix $(OUTPUT),$(FILES))
>
> @@ -50,7 +51,7 @@ __BUILD = $(CC) $(CFLAGS) -Wall -Werror -o $@ $(patsubst %.bin,%.c,$(@F)) $(LDFL
> ###############################
>
> $(OUTPUT)test-all.bin:
> - $(BUILD) -fstack-protector-all -O2 -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell $(PKG_CONFIG) --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl -lz -llzma
> + $(BUILD) -fstack-protector-all -O2 -D_FORTIFY_SOURCE=2 -ldw -lelf -lnuma -lelf -laudit -I/usr/include/slang -lslang $(shell $(PKG_CONFIG) --libs --cflags gtk+-2.0 2>/dev/null) $(FLAGS_PERL_EMBED) $(FLAGS_PYTHON_EMBED) -DPACKAGE='"perf"' -lbfd -ldl -lz -llzma -ludis86
>
> $(OUTPUT)test-hello.bin:
> $(BUILD)
> @@ -96,6 +97,9 @@ $(OUTPUT)test-numa_num_possible_cpus.bin:
> $(OUTPUT)test-libunwind.bin:
> $(BUILD) -lelf
>
> +$(OUTPUT)test-udis86.bin:
> + $(BUILD) -ludis86
> +
> $(OUTPUT)test-libunwind-debug-frame.bin:
> $(BUILD) -lelf
>
> diff --git a/tools/build/feature/test-all.c b/tools/build/feature/test-all.c
> index 81025ca..d67b1d5 100644
> --- a/tools/build/feature/test-all.c
> +++ b/tools/build/feature/test-all.c
> @@ -129,6 +129,10 @@
> # include "test-bpf.c"
> #undef main
>
> +#define main main_test_udis86
> +# include "test-udis86.c"
> +#endif
> +
> int main(int argc, char *argv[])
> {
> main_test_libpython();
> @@ -158,6 +162,7 @@ int main(int argc, char *argv[])
> main_test_lzma();
> main_test_get_cpuid();
> main_test_bpf();
> + main_test_udis86();
>
> return 0;
> }
> diff --git a/tools/build/feature/test-udis86.c b/tools/build/feature/test-udis86.c
> new file mode 100644
> index 0000000..623c545
> --- /dev/null
> +++ b/tools/build/feature/test-udis86.c
> @@ -0,0 +1,8 @@
> +#include <udis86.h>
> +
> +int main(void)
> +{
> + ud_t ud;
> + ud_init(&ud);
> + return 0;
> +}
> diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt
> index 382ddfb..c834f4d 100644
> --- a/tools/perf/Documentation/perf-script.txt
> +++ b/tools/perf/Documentation/perf-script.txt
> @@ -116,7 +116,7 @@ OPTIONS
> --fields::
> Comma separated list of fields to print. Options are:
> comm, tid, pid, time, cpu, event, trace, ip, sym, dso, addr, symoff,
> - srcline, period, iregs, brstack, brstacksym, flags.
> + srcline, period, iregs, brstack, brstacksym, flags, asm.
> Field list can be prepended with the type, trace, sw or hw,
> to indicate to which event type the field list applies.
> e.g., -f sw:comm,tid,time,ip,sym and -f trace:time,cpu,trace
> @@ -185,6 +185,8 @@ OPTIONS
>
> The brstacksym is identical to brstack, except that the FROM and TO addresses are printed in a symbolic form if possible.
>
> + When asm is specified the assembler instruction of each sample is printed in disassembled form.
> +
> -k::
> --vmlinux=<file>::
> vmlinux pathname
> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> index c691214..9fd8cb3 100644
> --- a/tools/perf/builtin-script.c
> +++ b/tools/perf/builtin-script.c
> @@ -24,6 +24,10 @@
> #include <linux/bitmap.h>
> #include "asm/bug.h"
>
> +#ifdef HAVE_UDIS86
> +#include <udis86.h>
> +#endif
> +
> static char const *script_name;
> static char const *generate_script_lang;
> static bool debug_mode;
> @@ -58,6 +62,7 @@ enum perf_output_field {
> PERF_OUTPUT_IREGS = 1U << 14,
> PERF_OUTPUT_BRSTACK = 1U << 15,
> PERF_OUTPUT_BRSTACKSYM = 1U << 16,
> + PERF_OUTPUT_ASM = 1U << 17,
> };
>
> struct output_option {
> @@ -81,6 +86,7 @@ struct output_option {
> {.str = "iregs", .field = PERF_OUTPUT_IREGS},
> {.str = "brstack", .field = PERF_OUTPUT_BRSTACK},
> {.str = "brstacksym", .field = PERF_OUTPUT_BRSTACKSYM},
> + {.str = "asm", .field = PERF_OUTPUT_ASM},
> };
>
> /* default set to maintain compatibility with current format */
> @@ -264,7 +270,11 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel,
> "selected. Hence, no address to lookup the source line number.\n");
> return -EINVAL;
> }
> -
> + if (PRINT_FIELD(ASM) && !PRINT_FIELD(IP)) {
> + pr_err("Display of assembler requested but sample IP is not\n"
> + "selected.\n");
> + return -EINVAL;
> + }
> if ((PRINT_FIELD(PID) || PRINT_FIELD(TID)) &&
> perf_evsel__check_stype(evsel, PERF_SAMPLE_TID, "TID",
> PERF_OUTPUT_TID|PERF_OUTPUT_PID))
> @@ -405,6 +415,89 @@ static void print_sample_iregs(union perf_event *event __maybe_unused,
> }
> }
>
> +#ifdef HAVE_UDIS86
> +
> +struct perf_ud {
> + ud_t ud_obj;
> + struct thread *thread;
> + u8 cpumode;
> + int cpu;
> +};
> +
> +static const char *dis_resolve(struct ud *u, uint64_t addr, int64_t *off)
> +{
> + struct perf_ud *ud = container_of(u, struct perf_ud, ud_obj);
> + struct addr_location al;
> +
> + memset(&al, 0, sizeof(struct addr_location));
> +
> + thread__find_addr_map(ud->thread, ud->cpumode, MAP__FUNCTION, addr, &al);
> + if (!al.map)
> + thread__find_addr_map(ud->thread, ud->cpumode, MAP__VARIABLE,
> + addr, &al);
> + al.cpu = ud->cpu;
> + al.sym = NULL;
> +
> + if (al.map)
> + al.sym = map__find_symbol(al.map, al.addr, NULL);
> +
> + if (!al.sym)
> + return NULL;
> +
> + if (addr < al.sym->end)
> + *off = addr - al.sym->start;
> + else
> + *off = addr - al.map->start - al.sym->start;
> + return al.sym->name;
> +}
> +#endif
> +
> +static void print_sample_asm(union perf_event *event __maybe_unused,
> + struct perf_sample *sample __maybe_unused,
> + struct thread *thread __maybe_unused,
> + struct perf_event_attr *attr __maybe_unused,
> + struct addr_location *al __maybe_unused,
> + struct machine *machine __maybe_unused)
> +{
> +#ifdef HAVE_UDIS86
> + static bool ud_initialized = false;
> + static struct perf_ud ud;
> + u8 buffer[32];
> + int len;
> + u64 offset;
> +
> + if (!ud_initialized) {
> + ud_initialized = true;
> + ud_init(&ud.ud_obj);
> + ud_set_syntax(&ud.ud_obj, UD_SYN_ATT);
> + ud_set_sym_resolver(&ud.ud_obj, dis_resolve);
> + }
> + ud.thread = thread;
> + ud.cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK;
> + ud.cpu = sample->cpu;
> +
> + if (!al->map || !al->map->dso)
> + return;
> + if (al->map->dso->data.status == DSO_DATA_STATUS_ERROR)
> + return;
> +
> + /* Load maps to ensure dso->is_64_bit has been updated */
> + map__load(al->map, machine->symbol_filter);
> +
> + offset = al->map->map_ip(al->map, sample->ip);
> + len = dso__data_read_offset(al->map->dso, machine,
> + offset, buffer, 32);
> + if (len <= 0)
> + return;
> +
> + ud_set_mode(&ud.ud_obj, al->map->dso->is_64_bit ? 64 : 32);
> + ud_set_pc(&ud.ud_obj, sample->ip);
> + ud_set_input_buffer(&ud.ud_obj, buffer, len);
> + ud_disassemble(&ud.ud_obj);
> + printf("\t%s", ud_insn_asm(&ud.ud_obj));
> +#endif
> +}
> +
> static void print_sample_start(struct perf_sample *sample,
> struct thread *thread,
> struct perf_evsel *evsel)
> @@ -636,7 +729,8 @@ static int perf_evlist__max_name_len(struct perf_evlist *evlist)
>
> static void process_event(struct perf_script *script, union perf_event *event,
> struct perf_sample *sample, struct perf_evsel *evsel,
> - struct addr_location *al)
> + struct addr_location *al,
> + struct machine *machine)
> {
> struct thread *thread = al->thread;
> struct perf_event_attr *attr = &evsel->attr;
> @@ -664,7 +758,7 @@ static void process_event(struct perf_script *script, union perf_event *event,
>
> if (is_bts_event(attr)) {
> print_sample_bts(event, sample, evsel, thread, al);
> - return;
> + goto print_rest;
> }
>
> if (PRINT_FIELD(TRACE))
> @@ -687,11 +781,15 @@ static void process_event(struct perf_script *script, union perf_event *event,
> if (PRINT_FIELD(IREGS))
> print_sample_iregs(event, sample, thread, attr);
>
> +print_rest:
> if (PRINT_FIELD(BRSTACK))
> print_sample_brstack(event, sample, thread, attr);
> else if (PRINT_FIELD(BRSTACKSYM))
> print_sample_brstacksym(event, sample, thread, attr);
>
> + if (PRINT_FIELD(ASM))
> + print_sample_asm(event, sample, thread, attr, al, machine);
> +
> printf("\n");
> }
>
> @@ -798,7 +896,7 @@ static int process_sample_event(struct perf_tool *tool,
> if (scripting_ops)
> scripting_ops->process_event(event, sample, evsel, &al);
> else
> - process_event(scr, event, sample, evsel, &al);
> + process_event(scr, event, sample, evsel, &al, machine);
>
> out_put:
> addr_location__put(&al);
> @@ -1913,7 +2011,7 @@ int cmd_script(int argc, const char **argv, const char *prefix __maybe_unused)
> "comma separated output fields prepend with 'type:'. "
> "Valid types: hw,sw,trace,raw. "
> "Fields: comm,tid,pid,time,cpu,event,trace,ip,sym,dso,"
> - "addr,symoff,period,iregs,brstack,brstacksym,flags", parse_output_fields),
> + "addr,symoff,period,iregs,brstack,brstacksym,flags,asm", parse_output_fields),
> OPT_BOOLEAN('a', "all-cpus", &system_wide,
> "system-wide collection from all CPUs"),
> OPT_STRING('S', "symbols", &symbol_conf.sym_list_str, "symbol[,symbol...]",
> diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile
> index 511141b..19bac7c 100644
> --- a/tools/perf/config/Makefile
> +++ b/tools/perf/config/Makefile
> @@ -576,6 +576,11 @@ ifneq ($(filter -lbfd,$(EXTLIBS)),)
> CFLAGS += -DHAVE_LIBBFD_SUPPORT
> endif
>
> +ifeq ($(feature-udis86), 1)
> + CFLAGS += -DHAVE_UDIS86
> + EXTLIBS += -ludis86
> +endif
> +
> ifndef NO_ZLIB
> ifeq ($(feature-zlib), 1)
> CFLAGS += -DHAVE_ZLIB_SUPPORT
> --
> 2.5.0
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] perf, tools, script: Add support for printing assembler
2016-02-29 14:56 ` Arnaldo Carvalho de Melo
@ 2016-02-29 17:35 ` Andi Kleen
2016-02-29 19:25 ` Arnaldo Carvalho de Melo
0 siblings, 1 reply; 9+ messages in thread
From: Andi Kleen @ 2016-02-29 17:35 UTC (permalink / raw
To: Arnaldo Carvalho de Melo
Cc: Andi Kleen, jolsa, namhyung, Adrian Hunter, alexander.shishkin,
linux-kernel, Andi Kleen
> What kernel have you used for testing this? 4.5-rc? I'm having trouble
> with intel_pt with 4.5.0-rc4 :-\
Works for me (on -rc5). What problem do you see?
>
> And while this looks like a great feature to have, have you considered
> using what is in tools/perf/util/intel-pt-decoder/ somehow, so that we
> don't end up adding one more dependency on another external library?
This needs a text disassembler.
The PT decoder doesn't have a disassembler unfortunately. The kernel
instruction decoder is also not a text disassembler. Changing it to do
that would be a quite large project.
> It would be great if we could, if done that way, perhaps at some point
> we could stop using objdump somehow :-\
objdump does a lot of work to line up the source code correctly.
Reproducing it would be a lot of work in perf too, and need
large scales changes. Perhaps at some time, but not any time
soon.
-Andi
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH 2/2] perf, tools, script: Add support for printing assembler
2016-02-29 17:35 ` Andi Kleen
@ 2016-02-29 19:25 ` Arnaldo Carvalho de Melo
2016-03-01 15:35 ` Intel PT not work on 4.5-rc6 as well after suspend/resume assembler Arnaldo Carvalho de Melo
0 siblings, 1 reply; 9+ messages in thread
From: Arnaldo Carvalho de Melo @ 2016-02-29 19:25 UTC (permalink / raw
To: Andi Kleen
Cc: jolsa, namhyung, Adrian Hunter, alexander.shishkin, linux-kernel,
Andi Kleen
Em Mon, Feb 29, 2016 at 06:35:28PM +0100, Andi Kleen escreveu:
> > What kernel have you used for testing this? 4.5-rc? I'm having trouble
> > with intel_pt with 4.5.0-rc4 :-\
>
> Works for me (on -rc5). What problem do you see?
Doesn't work for me on (4.5.0-rc4), but after your report of it working
my -rc4 .config and built -rc6+ with it, now it seems to work as
expected, will test your patch with it later.
> > And while this looks like a great feature to have, have you considered
> > using what is in tools/perf/util/intel-pt-decoder/ somehow, so that we
> > don't end up adding one more dependency on another external library?
>
> This needs a text disassembler.
>
> The PT decoder doesn't have a disassembler unfortunately. The kernel
> instruction decoder is also not a text disassembler. Changing it to do
> that would be a quite large project.
Pity, using this patch will not help with that and will add yet another
form of disassembly.
But then its a nice feature to have :-\ I'll test it.
> > It would be great if we could, if done that way, perhaps at some point
> > we could stop using objdump somehow :-\
>
> objdump does a lot of work to line up the source code correctly.
> Reproducing it would be a lot of work in perf too, and need
> large scales changes. Perhaps at some time, but not any time
> soon.
- Arnaldo
^ permalink raw reply [flat|nested] 9+ messages in thread
* Intel PT not work on 4.5-rc6 as well after suspend/resume assembler
2016-02-29 19:25 ` Arnaldo Carvalho de Melo
@ 2016-03-01 15:35 ` Arnaldo Carvalho de Melo
2016-03-01 23:49 ` Andi Kleen
0 siblings, 1 reply; 9+ messages in thread
From: Arnaldo Carvalho de Melo @ 2016-03-01 15:35 UTC (permalink / raw
To: Adrian Hunter
Cc: Andi Kleen, jolsa, namhyung, alexander.shishkin, linux-kernel,
Andi Kleen
Em Mon, Feb 29, 2016 at 04:25:49PM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Mon, Feb 29, 2016 at 06:35:28PM +0100, Andi Kleen escreveu:
> > > What kernel have you used for testing this? 4.5-rc? I'm having trouble
> > > with intel_pt with 4.5.0-rc4 :-\
> > Works for me (on -rc5). What problem do you see?
> Doesn't work for me on (4.5.0-rc4), but after your report of it working
> my -rc4 .config and built -rc6+ with it, now it seems to work as
> expected, will test your patch with it later.
Humm, after suspend/resume it is back not working, using your example
for this case:
[root@jouet ~]# uname -a
Linux jouet 4.5.0-rc6+ #1 SMP Mon Feb 29 16:09:29 BRT 2016 x86_64 x86_64 x86_64 GNU/Linux
[root@jouet ~]# perf record -e intel_pt//u true
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.024 MB perf.data ]
[root@jouet ~]# perf script
[root@jouet ~]# perf report -D | tail -20
. ... raw event: size 48 bytes
. 0000: 0f 00 00 00 00 00 30 00 98 03 00 00 98 03 00 00 ......0.........
. 0010: b6 49 00 00 b6 49 00 00 2c 78 8d 7d 9d 04 00 00 .I...I..,x.}....
. 0020: 03 00 00 00 00 00 00 00 1d 77 1b 00 00 00 00 00 .........w......
3 5074462799916 0x62f0 [0x30]: PERF_RECORD_SWITCH_CPU_WIDE IN prev pid/tid: 920/920
Aggregated stats: (excludes AUX area (e.g. instruction trace) decoded / synthesized events)
TOTAL events: 281
MMAP events: 124
COMM events: 2
EXIT events: 1
MMAP2 events: 4
ITRACE_START events: 1
SWITCH_CPU_WIDE events: 146
FINISHED_ROUND events: 2
AUXTRACE_INFO events: 1
intel_pt//u stats:
dummy:u stats:
dummy:u stats:
[root@jouet ~]#
CPU info for this machine:
processor : 3
vendor_id : GenuineIntel
cpu family : 6
model : 61
model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz
stepping : 4
microcode : 0x22
cpu MHz : 2759.046
cache size : 4096 KB
physical id : 0
siblings : 4
core id : 1
cpu cores : 2
apicid : 3
initial apicid : 3
fpu : yes
fpu_exception : yes
cpuid level : 20
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch epb intel_pt tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm rdseed adx smap xsaveopt dtherm ida arat pln pts
bugs :
bogomips : 5187.95
clflush size : 64
cache_alignment : 64
address sizes : 39 bits physical, 48 bits virtual
power management:
[root@jouet ~]#
Trying to reboot the machine to see if it gets back working...
- Arnaldo
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: Intel PT not work on 4.5-rc6 as well after suspend/resume assembler
2016-03-01 15:35 ` Intel PT not work on 4.5-rc6 as well after suspend/resume assembler Arnaldo Carvalho de Melo
@ 2016-03-01 23:49 ` Andi Kleen
2016-03-02 12:02 ` Arnaldo Carvalho de Melo
0 siblings, 1 reply; 9+ messages in thread
From: Andi Kleen @ 2016-03-01 23:49 UTC (permalink / raw
To: Arnaldo Carvalho de Melo
Cc: Adrian Hunter, Andi Kleen, jolsa, namhyung, alexander.shishkin,
linux-kernel
On Tue, Mar 01, 2016 at 12:35:24PM -0300, Arnaldo Carvalho de Melo wrote:
> Em Mon, Feb 29, 2016 at 04:25:49PM -0300, Arnaldo Carvalho de Melo escreveu:
> > Em Mon, Feb 29, 2016 at 06:35:28PM +0100, Andi Kleen escreveu:
> > > > What kernel have you used for testing this? 4.5-rc? I'm having trouble
> > > > with intel_pt with 4.5.0-rc4 :-\
>
> > > Works for me (on -rc5). What problem do you see?
>
> > Doesn't work for me on (4.5.0-rc4), but after your report of it working
> > my -rc4 .config and built -rc6+ with it, now it seems to work as
> > expected, will test your patch with it later.
>
> Humm, after suspend/resume it is back not working, using your example
> for this case:
Hi Arnaldo,
PT is always disabled when VT-x is active.
I guess the problem is here that it doesn't get reenabled when context switching
out of the guest. Need to add some code to KVM to do this (similar to how PEBS is handled)
However it should work again when you stop the KVM guest and restart the trace.
-Andi
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: Intel PT not work on 4.5-rc6 as well after suspend/resume assembler
2016-03-01 23:49 ` Andi Kleen
@ 2016-03-02 12:02 ` Arnaldo Carvalho de Melo
0 siblings, 0 replies; 9+ messages in thread
From: Arnaldo Carvalho de Melo @ 2016-03-02 12:02 UTC (permalink / raw
To: Andi Kleen
Cc: Adrian Hunter, Andi Kleen, jolsa, namhyung, alexander.shishkin,
linux-kernel
Em Tue, Mar 01, 2016 at 03:49:53PM -0800, Andi Kleen escreveu:
> On Tue, Mar 01, 2016 at 12:35:24PM -0300, Arnaldo Carvalho de Melo wrote:
> > Em Mon, Feb 29, 2016 at 04:25:49PM -0300, Arnaldo Carvalho de Melo escreveu:
> > > Em Mon, Feb 29, 2016 at 06:35:28PM +0100, Andi Kleen escreveu:
> > > > > What kernel have you used for testing this? 4.5-rc? I'm having trouble
> > > > > with intel_pt with 4.5.0-rc4 :-\
> > > > Works for me (on -rc5). What problem do you see?
> > > Doesn't work for me on (4.5.0-rc4), but after your report of it working
> > > my -rc4 .config and built -rc6+ with it, now it seems to work as
> > > expected, will test your patch with it later.
> > Humm, after suspend/resume it is back not working, using your example
> > for this case:
> PT is always disabled when VT-x is active.
> I guess the problem is here that it doesn't get reenabled when context switching
> out of the guest. Need to add some code to KVM to do this (similar to how PEBS is handled)
> However it should work again when you stop the KVM guest and restart the trace.
Right, if there are no KVM guests running, Intel PT is back working,
which is rather incovenient :-\
- Arnaldo
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2016-03-02 12:02 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-02-27 0:32 [PATCH 1/2] perf, tools: Add support for skipping itrace instructions Andi Kleen
2016-02-27 0:32 ` [PATCH 2/2] perf, tools, script: Add support for printing assembler Andi Kleen
2016-02-29 10:27 ` Jiri Olsa
2016-02-29 14:56 ` Arnaldo Carvalho de Melo
2016-02-29 17:35 ` Andi Kleen
2016-02-29 19:25 ` Arnaldo Carvalho de Melo
2016-03-01 15:35 ` Intel PT not work on 4.5-rc6 as well after suspend/resume assembler Arnaldo Carvalho de Melo
2016-03-01 23:49 ` Andi Kleen
2016-03-02 12:02 ` Arnaldo Carvalho de Melo
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.