Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1213038 > unrolled thread
| Started by | Wang Nan <wangnan0@huawei.com> |
|---|---|
| First post | 2015-08-25 15:30 +0200 |
| Last post | 2015-08-28 08:50 +0200 |
| Articles | 12 — 6 participants |
Back to article view | Back to linux.kernel
[PATCH 1/2] perf probe: Prevent segfault when reading probe point with absolute address Wang Nan <wangnan0@huawei.com> - 2015-08-25 15:30 +0200
[PATCH 2/2] perf probe: Support probing at absolute address Wang Nan <wangnan0@huawei.com> - 2015-08-25 15:30 +0200
RE: [PATCH 2/2] perf probe: Support probing at absolute address 平松雅巳 / HIRAMATU，MASAMI <masami.hiramatsu.pt@hitachi.com> - 2015-08-26 02:10 +0200
Re: [PATCH 2/2] perf probe: Support probing at absolute address "Wangnan (F)" <wangnan0@huawei.com> - 2015-08-26 04:40 +0200
RE: [PATCH 2/2] perf probe: Support probing at absolute address 平松雅巳 / HIRAMATU，MASAMI <masami.hiramatsu.pt@hitachi.com> - 2015-08-26 11:10 +0200
Re: [PATCH 2/2] perf probe: Support probing at absolute address "acme@kernel.org" <acme@kernel.org> - 2015-08-26 15:10 +0200
Re: [PATCH 2/2] perf probe: Support probing at absolute address "Wangnan (F)" <wangnan0@huawei.com> - 2015-08-26 15:20 +0200
Re: [PATCH 2/2] perf probe: Support probing at absolute address Arnaldo Carvalho de Melo <acme@kernel.org> - 2015-08-26 15:40 +0200
Re: [PATCH 2/2] perf probe: Support probing at absolute address "acme@kernel.org" <acme@kernel.org> - 2015-08-26 15:10 +0200
Re: [PATCH 1/2] perf probe: Prevent segfault when reading probe point with absolute address Arnaldo Carvalho de Melo <acme@kernel.org> - 2015-08-25 17:40 +0200
RE: [PATCH 1/2] perf probe: Prevent segfault when reading probe point with absolute address 平松雅巳 / HIRAMATU，MASAMI <masami.hiramatsu.pt@hitachi.com> - 2015-08-26 02:10 +0200
[tip:perf/core] perf probe: Prevent segfault when reading probe point with absolute address tip-bot for Wang Nan <tipbot@zytor.com> - 2015-08-28 08:50 +0200
| From | Wang Nan <wangnan0@huawei.com> |
|---|---|
| Date | 2015-08-25 15:30 +0200 |
| Subject | [PATCH 1/2] perf probe: Prevent segfault when reading probe point with absolute address |
| Message-ID | <q1msz-7Ih-17@gated-at.bofh.it> |
'perf probe -l' panic if there is a manually inserted probing point
with absolute address. For example:
# echo 'p:probe/abs_ffffffff811e6615 0xffffffff811e6615' > /sys/kernel/debug/tracing/kprobe_events
# perf probe -l
Segmentation fault (core dumped)
This patch fix this problem by considering the situation that
"tp->symbol == NULL" in find_perf_probe_point_from_dwarf() and
find_perf_probe_point_from_map().
After this patch:
# perf probe -l
probe:abs_ffffffff811e6615 (on SyS_write+5@fs/read_write.c)
And when debug info is missing:
# rm -rf ~/.debug
# mv /lib/modules/4.2.0-rc1+/build/vmlinux /lib/modules/4.2.0-rc1+/build/vmlinux.bak
# perf probe -l
probe:abs_ffffffff811e6615 (on sys_write+5)
Signed-off-by: Wang Nan <wangnan0@huawei.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/probe-event.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index f07374b..6c7e538 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -515,7 +515,7 @@ static int find_perf_probe_point_from_dwarf(struct probe_trace_point *tp,
if (ret < 0)
goto error;
addr += stext;
- } else {
+ } else if (tp->symbol) {
addr = kernel_get_symbol_address_by_name(tp->symbol, false);
if (addr == 0)
goto error;
@@ -1815,17 +1815,17 @@ static int find_perf_probe_point_from_map(struct probe_trace_point *tp,
{
struct symbol *sym = NULL;
struct map *map;
- u64 addr;
+ u64 addr = tp->address;
int ret = -ENOENT;
if (!is_kprobe) {
map = dso__new_map(tp->module);
if (!map)
goto out;
- addr = tp->address;
sym = map__find_symbol(map, addr, NULL);
} else {
- addr = kernel_get_symbol_address_by_name(tp->symbol, true);
+ if (tp->symbol)
+ addr = kernel_get_symbol_address_by_name(tp->symbol, true);
if (addr) {
addr += tp->offset;
sym = __find_kernel_function(addr, &map);
--
1.8.3.4
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [next] | [standalone]
| From | Wang Nan <wangnan0@huawei.com> |
|---|---|
| Date | 2015-08-25 15:30 +0200 |
| Subject | [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1msz-7Ih-25@gated-at.bofh.it> |
| In reply to | #1213038 |
It should be useful to allow 'perf probe' probe at absolute offset of
a target. For example, when (u)probing at a instruction of a shared
object in a embedded system where debuginfo is not avaliable but we
know the offset of that instruction by manually digging.
This patch enables following perf probe command syntax:
# perf probe +0xffffffff811e6615
And
# perf probe /lib/x86_64-linux-gnu/libc-2.19.so +0xeb860
In the above example, we don't need a anchor symbol, so it is possible
to compute absolute addresses using other methods and then use
'perf probe' to create the probing points.
Signed-off-by: Wang Nan <wangnan0@huawei.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/probe-event.c | 144 +++++++++++++++++++++++++++++++++++++----
tools/perf/util/probe-event.h | 3 +
tools/perf/util/probe-finder.c | 21 +-----
3 files changed, 138 insertions(+), 30 deletions(-)
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 6c7e538..59de69a4 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -1194,9 +1194,13 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
*ptr++ = '\0';
}
- tmp = strdup(arg);
- if (tmp == NULL)
- return -ENOMEM;
+ if (arg[0] == '\0')
+ tmp = NULL;
+ else {
+ tmp = strdup(arg);
+ if (tmp == NULL)
+ return -ENOMEM;
+ }
if (file_spec)
pp->file = tmp;
@@ -1283,11 +1287,6 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
return -EINVAL;
}
- if (pp->offset && !pp->function) {
- semantic_error("Offset requires an entry function.\n");
- return -EINVAL;
- }
-
if (pp->retprobe && !pp->function) {
semantic_error("Return probe requires an entry function.\n");
return -EINVAL;
@@ -1299,6 +1298,11 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
return -EINVAL;
}
+ if (!pp->function && !pp->offset && !pp->file) {
+ semantic_error("Absolute address should not be zero.\n");
+ return -EINVAL;
+ }
+
pr_debug("symbol:%s file:%s line:%d offset:%lu return:%d lazy:%s\n",
pp->function, pp->file, pp->line, pp->offset, pp->retprobe,
pp->lazy_line);
@@ -1609,7 +1613,7 @@ error:
static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
{
char *buf, *tmp;
- char offs[32] = "", line[32] = "", file[32] = "";
+ char offs[32] = "", line[32] = "", file[32] = "", addr[32] = "";
int ret, len;
buf = zalloc(MAX_CMDLEN);
@@ -1622,6 +1626,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
if (ret <= 0)
goto error;
}
+ if (!pp->function) {
+ ret = e_snprintf(addr, 32, "0x%lx", pp->offset);
+ if (ret <= 0)
+ goto error;
+ }
if (pp->line) {
ret = e_snprintf(line, 32, ":%d", pp->line);
if (ret <= 0)
@@ -1639,9 +1648,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
goto error;
}
- if (pp->function)
- ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s", pp->function,
- offs, pp->retprobe ? "%return" : "", line,
+ if (pp->function || pp->offset)
+ ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s",
+ pp->function ? : addr,
+ pp->function ? offs : "",
+ pp->retprobe ? "%return" : "", line,
file);
else
ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", file, line);
@@ -1786,6 +1797,11 @@ char *synthesize_probe_trace_command(struct probe_trace_event *tev)
if (tev->uprobes)
ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s:0x%lx",
tp->module, tp->address);
+ else if (tp->symbol[0] == '0' && tp->symbol[1] == 'x')
+ /* Absolute address. See try_to_find_absolute_address() */
+ ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s0x%lx",
+ tp->module ?: "", tp->module ? ":" : "",
+ tp->address);
else
ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s%s+%lu",
tp->module ?: "", tp->module ? ":" : "",
@@ -2572,6 +2588,87 @@ err_out:
goto out;
}
+static int try_to_find_absolute_address(struct perf_probe_event *pev,
+ struct probe_trace_event **tevs)
+{
+ struct perf_probe_point *pp = &pev->point;
+ struct probe_trace_event *tev;
+ struct probe_trace_point *tp;
+ int i, err;
+
+ if (perf_probe_event_need_dwarf(pev) || pev->point.function)
+ return -EINVAL;
+
+ /*
+ * This is 'perf probe /lib/libc.so +0xabcd'. Try to probe at
+ * absolute address.
+ *
+ * Only one tev can be generated by this.
+ */
+ *tevs = zalloc(sizeof(*tev));
+ if (!*tevs)
+ return -ENOMEM;
+
+ tev = *tevs;
+ tp = &tev->point;
+
+ /*
+ * Don't use tp->offset, use address directly, because
+ * in synthesize_probe_trace_command() address cannot be
+ * zero.
+ */
+ tp->address = pev->point.offset;
+ tp->retprobe = pp->retprobe;
+ tev->uprobes = pev->uprobes;
+
+ err = -ENOMEM;
+ /* Give it a '0x' leading symbol name */
+ if (asprintf(&tp->symbol, "0x%lx", tp->address) < 0)
+ goto errout;
+
+ /* For kprobe, check range */
+ if ((!tev->uprobes) &&
+ (kprobe_warn_out_range(tev->point.symbol,
+ tev->point.address))) {
+ err = -EACCES;
+ goto errout;
+ }
+
+ if (asprintf(&tp->realname, "abs_%lx", tp->address) < 0)
+ goto errout;
+
+ if (pev->target) {
+ tp->module = strdup(pev->target);
+ if (!tp->module)
+ goto errout;
+ }
+
+ if (tev->group) {
+ tev->group = strdup(pev->group);
+ if (!tev->group)
+ goto errout;
+ }
+
+ if (pev->event) {
+ tev->event = strdup(pev->event);
+ if (!tev->event)
+ goto errout;
+ }
+
+ tev->nargs = pev->nargs;
+ for (i = 0; i < tev->nargs; i++)
+ copy_to_probe_trace_arg(&tev->args[i], &pev->args[i]);
+
+ return 1;
+
+errout:
+ if (*tevs) {
+ clear_probe_trace_events(*tevs, 1);
+ *tevs = NULL;
+ }
+ return err;
+}
+
bool __weak arch__prefers_symtab(void) { return false; }
static int convert_to_probe_trace_events(struct perf_probe_event *pev,
@@ -2588,6 +2685,10 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
}
}
+ ret = try_to_find_absolute_address(pev, tevs);
+ if (ret > 0)
+ return ret;
+
if (arch__prefers_symtab() && !perf_probe_event_need_dwarf(pev)) {
ret = find_probe_trace_events_from_map(pev, tevs);
if (ret > 0)
@@ -2758,3 +2859,22 @@ end:
return ret;
}
+int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
+ struct perf_probe_arg *pvar)
+{
+ tvar->value = strdup(pvar->var);
+ if (tvar->value == NULL)
+ return -ENOMEM;
+ if (pvar->type) {
+ tvar->type = strdup(pvar->type);
+ if (tvar->type == NULL)
+ return -ENOMEM;
+ }
+ if (pvar->name) {
+ tvar->name = strdup(pvar->name);
+ if (tvar->name == NULL)
+ return -ENOMEM;
+ } else
+ tvar->name = NULL;
+ return 0;
+}
diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
index 83ee95e..174a3cf 100644
--- a/tools/perf/util/probe-event.h
+++ b/tools/perf/util/probe-event.h
@@ -156,4 +156,7 @@ int e_snprintf(char *str, size_t size, const char *format, ...)
/* Maximum index number of event-name postfix */
#define MAX_EVENT_INDEX 1024
+int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
+ struct perf_probe_arg *pvar);
+
#endif /*_PROBE_EVENT_H */
diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
index 7b80f8c..29c43c068 100644
--- a/tools/perf/util/probe-finder.c
+++ b/tools/perf/util/probe-finder.c
@@ -553,24 +553,9 @@ static int find_variable(Dwarf_Die *sc_die, struct probe_finder *pf)
char buf[32], *ptr;
int ret = 0;
- if (!is_c_varname(pf->pvar->var)) {
- /* Copy raw parameters */
- pf->tvar->value = strdup(pf->pvar->var);
- if (pf->tvar->value == NULL)
- return -ENOMEM;
- if (pf->pvar->type) {
- pf->tvar->type = strdup(pf->pvar->type);
- if (pf->tvar->type == NULL)
- return -ENOMEM;
- }
- if (pf->pvar->name) {
- pf->tvar->name = strdup(pf->pvar->name);
- if (pf->tvar->name == NULL)
- return -ENOMEM;
- } else
- pf->tvar->name = NULL;
- return 0;
- }
+ /* Copy raw parameters */
+ if (!is_c_varname(pf->pvar->var))
+ return copy_to_probe_trace_arg(pf->tvar, pf->pvar);
if (pf->pvar->name)
pf->tvar->name = strdup(pf->pvar->name);
--
1.8.3.4
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | 平松雅巳 / HIRAMATU，MASAMI <masami.hiramatsu.pt@hitachi.com> |
|---|---|
| Date | 2015-08-26 02:10 +0200 |
| Subject | RE: [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1wrU-5N8-11@gated-at.bofh.it> |
| In reply to | #1213040 |
PiBGcm9tOiBXYW5nIE5hbiBbbWFpbHRvOndhbmduYW4wQGh1YXdlaS5jb21dDQo+IA0KPiBJdCBz aG91bGQgYmUgdXNlZnVsIHRvIGFsbG93ICdwZXJmIHByb2JlJyBwcm9iZSBhdCBhYnNvbHV0ZSBv ZmZzZXQgb2YNCj4gYSB0YXJnZXQuIEZvciBleGFtcGxlLCB3aGVuICh1KXByb2JpbmcgYXQgYSBp bnN0cnVjdGlvbiBvZiBhIHNoYXJlZA0KPiBvYmplY3QgaW4gYSBlbWJlZGRlZCBzeXN0ZW0gd2hl cmUgZGVidWdpbmZvIGlzIG5vdCBhdmFsaWFibGUgYnV0IHdlDQo+IGtub3cgdGhlIG9mZnNldCBv ZiB0aGF0IGluc3RydWN0aW9uIGJ5IG1hbnVhbGx5IGRpZ2dpbmcuDQo+IA0KPiBUaGlzIHBhdGNo IGVuYWJsZXMgZm9sbG93aW5nIHBlcmYgcHJvYmUgY29tbWFuZCBzeW50YXg6DQo+IA0KPiAgIyBw ZXJmIHByb2JlICsweGZmZmZmZmZmODExZTY2MTUNCj4gDQo+IEFuZA0KPiANCj4gICMgcGVyZiBw cm9iZSAvbGliL3g4Nl82NC1saW51eC1nbnUvbGliYy0yLjE5LnNvICsweGViODYwDQoNCldoeSBk byB3ZSBuZWVkICIrIiBmb3IgdGhlIGFic29sdXRlIGFkZHJlc3M/DQpJdCBzZWVtcyB0aGF0IHdl IGNhbiBkbyBpdCBpZiB3ZSBmaW5kIHRoYXQgdGhlIGdpdmVuIHByb2JlIHBvaW50DQpzdGFydHMg d2l0aCAiMHgiLg0KDQpUaGFua3MsDQoNCj4gDQo+IEluIHRoZSBhYm92ZSBleGFtcGxlLCB3ZSBk b24ndCBuZWVkIGEgYW5jaG9yIHN5bWJvbCwgc28gaXQgaXMgcG9zc2libGUNCj4gdG8gY29tcHV0 ZSBhYnNvbHV0ZSBhZGRyZXNzZXMgdXNpbmcgb3RoZXIgbWV0aG9kcyBhbmQgdGhlbiB1c2UNCj4g J3BlcmYgcHJvYmUnIHRvIGNyZWF0ZSB0aGUgcHJvYmluZyBwb2ludHMuDQo+IA0KPiBTaWduZWQt b2ZmLWJ5OiBXYW5nIE5hbiA8d2FuZ25hbjBAaHVhd2VpLmNvbT4NCj4gQ2M6IEFybmFsZG8gQ2Fy dmFsaG8gZGUgTWVsbyA8YWNtZUBrZXJuZWwub3JnPg0KPiBDYzogSW5nbyBNb2xuYXIgPG1pbmdv QHJlZGhhdC5jb20+DQo+IENjOiBNYXNhbWkgSGlyYW1hdHN1IDxtYXNhbWkuaGlyYW1hdHN1LnB0 QGhpdGFjaGkuY29tPg0KPiBDYzogTmFtaHl1bmcgS2ltIDxuYW1oeXVuZ0BrZXJuZWwub3JnPg0K PiAtLS0NCj4gIHRvb2xzL3BlcmYvdXRpbC9wcm9iZS1ldmVudC5jICB8IDE0NCArKysrKysrKysr KysrKysrKysrKysrKysrKysrKysrKysrKysrLS0tLQ0KPiAgdG9vbHMvcGVyZi91dGlsL3Byb2Jl LWV2ZW50LmggIHwgICAzICsNCj4gIHRvb2xzL3BlcmYvdXRpbC9wcm9iZS1maW5kZXIuYyB8ICAy MSArLS0tLS0NCj4gIDMgZmlsZXMgY2hhbmdlZCwgMTM4IGluc2VydGlvbnMoKyksIDMwIGRlbGV0 aW9ucygtKQ0KPiANCj4gZGlmZiAtLWdpdCBhL3Rvb2xzL3BlcmYvdXRpbC9wcm9iZS1ldmVudC5j IGIvdG9vbHMvcGVyZi91dGlsL3Byb2JlLWV2ZW50LmMNCj4gaW5kZXggNmM3ZTUzOC4uNTlkZTY5 YTQgMTAwNjQ0DQo+IC0tLSBhL3Rvb2xzL3BlcmYvdXRpbC9wcm9iZS1ldmVudC5jDQo+ICsrKyBi L3Rvb2xzL3BlcmYvdXRpbC9wcm9iZS1ldmVudC5jDQo+IEBAIC0xMTk0LDkgKzExOTQsMTMgQEAg c3RhdGljIGludCBwYXJzZV9wZXJmX3Byb2JlX3BvaW50KGNoYXIgKmFyZywgc3RydWN0IHBlcmZf cHJvYmVfZXZlbnQgKnBldikNCj4gIAkJKnB0cisrID0gJ1wwJzsNCj4gIAl9DQo+IA0KPiAtCXRt cCA9IHN0cmR1cChhcmcpOw0KPiAtCWlmICh0bXAgPT0gTlVMTCkNCj4gLQkJcmV0dXJuIC1FTk9N RU07DQo+ICsJaWYgKGFyZ1swXSA9PSAnXDAnKQ0KPiArCQl0bXAgPSBOVUxMOw0KPiArCWVsc2Ug ew0KPiArCQl0bXAgPSBzdHJkdXAoYXJnKTsNCj4gKwkJaWYgKHRtcCA9PSBOVUxMKQ0KPiArCQkJ cmV0dXJuIC1FTk9NRU07DQo+ICsJfQ0KPiANCj4gIAlpZiAoZmlsZV9zcGVjKQ0KPiAgCQlwcC0+ ZmlsZSA9IHRtcDsNCj4gQEAgLTEyODMsMTEgKzEyODcsNiBAQCBzdGF0aWMgaW50IHBhcnNlX3Bl cmZfcHJvYmVfcG9pbnQoY2hhciAqYXJnLCBzdHJ1Y3QgcGVyZl9wcm9iZV9ldmVudCAqcGV2KQ0K PiAgCQlyZXR1cm4gLUVJTlZBTDsNCj4gIAl9DQo+IA0KPiAtCWlmIChwcC0+b2Zmc2V0ICYmICFw cC0+ZnVuY3Rpb24pIHsNCj4gLQkJc2VtYW50aWNfZXJyb3IoIk9mZnNldCByZXF1aXJlcyBhbiBl bnRyeSBmdW5jdGlvbi5cbiIpOw0KPiAtCQlyZXR1cm4gLUVJTlZBTDsNCj4gLQl9DQo+IC0NCj4g IAlpZiAocHAtPnJldHByb2JlICYmICFwcC0+ZnVuY3Rpb24pIHsNCj4gIAkJc2VtYW50aWNfZXJy b3IoIlJldHVybiBwcm9iZSByZXF1aXJlcyBhbiBlbnRyeSBmdW5jdGlvbi5cbiIpOw0KPiAgCQly ZXR1cm4gLUVJTlZBTDsNCj4gQEAgLTEyOTksNiArMTI5OCwxMSBAQCBzdGF0aWMgaW50IHBhcnNl X3BlcmZfcHJvYmVfcG9pbnQoY2hhciAqYXJnLCBzdHJ1Y3QgcGVyZl9wcm9iZV9ldmVudCAqcGV2 KQ0KPiAgCQlyZXR1cm4gLUVJTlZBTDsNCj4gIAl9DQo+IA0KPiArCWlmICghcHAtPmZ1bmN0aW9u ICYmICFwcC0+b2Zmc2V0ICYmICFwcC0+ZmlsZSkgew0KPiArCQlzZW1hbnRpY19lcnJvcigiQWJz b2x1dGUgYWRkcmVzcyBzaG91bGQgbm90IGJlIHplcm8uXG4iKTsNCj4gKwkJcmV0dXJuIC1FSU5W QUw7DQo+ICsJfQ0KPiArDQo+ICAJcHJfZGVidWcoInN5bWJvbDolcyBmaWxlOiVzIGxpbmU6JWQg b2Zmc2V0OiVsdSByZXR1cm46JWQgbGF6eTolc1xuIiwNCj4gIAkJIHBwLT5mdW5jdGlvbiwgcHAt PmZpbGUsIHBwLT5saW5lLCBwcC0+b2Zmc2V0LCBwcC0+cmV0cHJvYmUsDQo+ICAJCSBwcC0+bGF6 eV9saW5lKTsNCj4gQEAgLTE2MDksNyArMTYxMyw3IEBAIGVycm9yOg0KPiAgc3RhdGljIGNoYXIg KnN5bnRoZXNpemVfcGVyZl9wcm9iZV9wb2ludChzdHJ1Y3QgcGVyZl9wcm9iZV9wb2ludCAqcHAp DQo+ICB7DQo+ICAJY2hhciAqYnVmLCAqdG1wOw0KPiAtCWNoYXIgb2Zmc1szMl0gPSAiIiwgbGlu ZVszMl0gPSAiIiwgZmlsZVszMl0gPSAiIjsNCj4gKwljaGFyIG9mZnNbMzJdID0gIiIsIGxpbmVb MzJdID0gIiIsIGZpbGVbMzJdID0gIiIsIGFkZHJbMzJdID0gIiI7DQo+ICAJaW50IHJldCwgbGVu Ow0KPiANCj4gIAlidWYgPSB6YWxsb2MoTUFYX0NNRExFTik7DQo+IEBAIC0xNjIyLDYgKzE2MjYs MTEgQEAgc3RhdGljIGNoYXIgKnN5bnRoZXNpemVfcGVyZl9wcm9iZV9wb2ludChzdHJ1Y3QgcGVy Zl9wcm9iZV9wb2ludCAqcHApDQo+ICAJCWlmIChyZXQgPD0gMCkNCj4gIAkJCWdvdG8gZXJyb3I7 DQo+ICAJfQ0KPiArCWlmICghcHAtPmZ1bmN0aW9uKSB7DQo+ICsJCXJldCA9IGVfc25wcmludGYo YWRkciwgMzIsICIweCVseCIsIHBwLT5vZmZzZXQpOw0KPiArCQlpZiAocmV0IDw9IDApDQo+ICsJ CQlnb3RvIGVycm9yOw0KPiArCX0NCj4gIAlpZiAocHAtPmxpbmUpIHsNCj4gIAkJcmV0ID0gZV9z bnByaW50ZihsaW5lLCAzMiwgIjolZCIsIHBwLT5saW5lKTsNCj4gIAkJaWYgKHJldCA8PSAwKQ0K PiBAQCAtMTYzOSw5ICsxNjQ4LDExIEBAIHN0YXRpYyBjaGFyICpzeW50aGVzaXplX3BlcmZfcHJv YmVfcG9pbnQoc3RydWN0IHBlcmZfcHJvYmVfcG9pbnQgKnBwKQ0KPiAgCQkJZ290byBlcnJvcjsN Cj4gIAl9DQo+IA0KPiAtCWlmIChwcC0+ZnVuY3Rpb24pDQo+IC0JCXJldCA9IGVfc25wcmludGYo YnVmLCBNQVhfQ01ETEVOLCAiJXMlcyVzJXMlcyIsIHBwLT5mdW5jdGlvbiwNCj4gLQkJCQkgb2Zm cywgcHAtPnJldHByb2JlID8gIiVyZXR1cm4iIDogIiIsIGxpbmUsDQo+ICsJaWYgKHBwLT5mdW5j dGlvbiB8fCBwcC0+b2Zmc2V0KQ0KPiArCQlyZXQgPSBlX3NucHJpbnRmKGJ1ZiwgTUFYX0NNRExF TiwgIiVzJXMlcyVzJXMiLA0KPiArCQkJCSBwcC0+ZnVuY3Rpb24gPyA6IGFkZHIsDQo+ICsJCQkJ IHBwLT5mdW5jdGlvbiA/IG9mZnMgOiAiIiwNCj4gKwkJCQkgcHAtPnJldHByb2JlID8gIiVyZXR1 cm4iIDogIiIsIGxpbmUsDQo+ICAJCQkJIGZpbGUpOw0KPiAgCWVsc2UNCj4gIAkJcmV0ID0gZV9z bnByaW50ZihidWYsIE1BWF9DTURMRU4sICIlcyVzIiwgZmlsZSwgbGluZSk7DQo+IEBAIC0xNzg2 LDYgKzE3OTcsMTEgQEAgY2hhciAqc3ludGhlc2l6ZV9wcm9iZV90cmFjZV9jb21tYW5kKHN0cnVj dCBwcm9iZV90cmFjZV9ldmVudCAqdGV2KQ0KPiAgCWlmICh0ZXYtPnVwcm9iZXMpDQo+ICAJCXJl dCA9IGVfc25wcmludGYoYnVmICsgbGVuLCBNQVhfQ01ETEVOIC0gbGVuLCAiJXM6MHglbHgiLA0K PiAgCQkJCSB0cC0+bW9kdWxlLCB0cC0+YWRkcmVzcyk7DQo+ICsJZWxzZSBpZiAodHAtPnN5bWJv bFswXSA9PSAnMCcgJiYgdHAtPnN5bWJvbFsxXSA9PSAneCcpDQo+ICsJCS8qIEFic29sdXRlIGFk ZHJlc3MuIFNlZSB0cnlfdG9fZmluZF9hYnNvbHV0ZV9hZGRyZXNzKCkgKi8NCj4gKwkJcmV0ID0g ZV9zbnByaW50ZihidWYgKyBsZW4sIE1BWF9DTURMRU4gLSBsZW4sICIlcyVzMHglbHgiLA0KPiAr CQkJCSB0cC0+bW9kdWxlID86ICIiLCB0cC0+bW9kdWxlID8gIjoiIDogIiIsDQo+ICsJCQkJIHRw LT5hZGRyZXNzKTsNCj4gIAllbHNlDQo+ICAJCXJldCA9IGVfc25wcmludGYoYnVmICsgbGVuLCBN QVhfQ01ETEVOIC0gbGVuLCAiJXMlcyVzKyVsdSIsDQo+ICAJCQkJIHRwLT5tb2R1bGUgPzogIiIs IHRwLT5tb2R1bGUgPyAiOiIgOiAiIiwNCj4gQEAgLTI1NzIsNiArMjU4OCw4NyBAQCBlcnJfb3V0 Og0KPiAgCWdvdG8gb3V0Ow0KPiAgfQ0KPiANCj4gK3N0YXRpYyBpbnQgdHJ5X3RvX2ZpbmRfYWJz b2x1dGVfYWRkcmVzcyhzdHJ1Y3QgcGVyZl9wcm9iZV9ldmVudCAqcGV2LA0KPiArCQkJCQlzdHJ1 Y3QgcHJvYmVfdHJhY2VfZXZlbnQgKip0ZXZzKQ0KPiArew0KPiArCXN0cnVjdCBwZXJmX3Byb2Jl X3BvaW50ICpwcCA9ICZwZXYtPnBvaW50Ow0KPiArCXN0cnVjdCBwcm9iZV90cmFjZV9ldmVudCAq dGV2Ow0KPiArCXN0cnVjdCBwcm9iZV90cmFjZV9wb2ludCAqdHA7DQo+ICsJaW50IGksIGVycjsN Cj4gKw0KPiArCWlmIChwZXJmX3Byb2JlX2V2ZW50X25lZWRfZHdhcmYocGV2KSB8fCBwZXYtPnBv aW50LmZ1bmN0aW9uKQ0KPiArCQlyZXR1cm4gLUVJTlZBTDsNCj4gKw0KPiArCS8qDQo+ICsJICog VGhpcyBpcyAncGVyZiBwcm9iZSAvbGliL2xpYmMuc28gKzB4YWJjZCcuIFRyeSB0byBwcm9iZSBh dA0KPiArCSAqIGFic29sdXRlIGFkZHJlc3MuDQo+ICsJICoNCj4gKwkgKiBPbmx5IG9uZSB0ZXYg Y2FuIGJlIGdlbmVyYXRlZCBieSB0aGlzLg0KPiArCSAqLw0KPiArCSp0ZXZzID0gemFsbG9jKHNp emVvZigqdGV2KSk7DQo+ICsJaWYgKCEqdGV2cykNCj4gKwkJcmV0dXJuIC1FTk9NRU07DQo+ICsN Cj4gKwl0ZXYgPSAqdGV2czsNCj4gKwl0cCA9ICZ0ZXYtPnBvaW50Ow0KPiArDQo+ICsJLyoNCj4g KwkgKiBEb24ndCB1c2UgdHAtPm9mZnNldCwgdXNlIGFkZHJlc3MgZGlyZWN0bHksIGJlY2F1c2UN Cj4gKwkgKiBpbiBzeW50aGVzaXplX3Byb2JlX3RyYWNlX2NvbW1hbmQoKSBhZGRyZXNzIGNhbm5v dCBiZQ0KPiArCSAqIHplcm8uDQo+ICsJICovDQo+ICsJdHAtPmFkZHJlc3MgPSBwZXYtPnBvaW50 Lm9mZnNldDsNCj4gKwl0cC0+cmV0cHJvYmUgPSBwcC0+cmV0cHJvYmU7DQo+ICsJdGV2LT51cHJv YmVzID0gcGV2LT51cHJvYmVzOw0KPiArDQo+ICsJZXJyID0gLUVOT01FTTsNCj4gKwkvKiBHaXZl IGl0IGEgJzB4JyBsZWFkaW5nIHN5bWJvbCBuYW1lICovDQo+ICsJaWYgKGFzcHJpbnRmKCZ0cC0+ c3ltYm9sLCAiMHglbHgiLCB0cC0+YWRkcmVzcykgPCAwKQ0KPiArCQlnb3RvIGVycm91dDsNCj4g Kw0KPiArCS8qIEZvciBrcHJvYmUsIGNoZWNrIHJhbmdlICovDQo+ICsJaWYgKCghdGV2LT51cHJv YmVzKSAmJg0KPiArCSAgICAoa3Byb2JlX3dhcm5fb3V0X3JhbmdlKHRldi0+cG9pbnQuc3ltYm9s LA0KPiArCQkJCSAgIHRldi0+cG9pbnQuYWRkcmVzcykpKSB7DQo+ICsJCWVyciA9IC1FQUNDRVM7 DQo+ICsJCWdvdG8gZXJyb3V0Ow0KPiArCX0NCj4gKw0KPiArCWlmIChhc3ByaW50ZigmdHAtPnJl YWxuYW1lLCAiYWJzXyVseCIsIHRwLT5hZGRyZXNzKSA8IDApDQo+ICsJCWdvdG8gZXJyb3V0Ow0K PiArDQo+ICsJaWYgKHBldi0+dGFyZ2V0KSB7DQo+ICsJCXRwLT5tb2R1bGUgPSBzdHJkdXAocGV2 LT50YXJnZXQpOw0KPiArCQlpZiAoIXRwLT5tb2R1bGUpDQo+ICsJCQlnb3RvIGVycm91dDsNCj4g Kwl9DQo+ICsNCj4gKwlpZiAodGV2LT5ncm91cCkgew0KPiArCQl0ZXYtPmdyb3VwID0gc3RyZHVw KHBldi0+Z3JvdXApOw0KPiArCQlpZiAoIXRldi0+Z3JvdXApDQo+ICsJCQlnb3RvIGVycm91dDsN Cj4gKwl9DQo+ICsNCj4gKwlpZiAocGV2LT5ldmVudCkgew0KPiArCQl0ZXYtPmV2ZW50ID0gc3Ry ZHVwKHBldi0+ZXZlbnQpOw0KPiArCQlpZiAoIXRldi0+ZXZlbnQpDQo+ICsJCQlnb3RvIGVycm91 dDsNCj4gKwl9DQo+ICsNCj4gKwl0ZXYtPm5hcmdzID0gcGV2LT5uYXJnczsNCj4gKwlmb3IgKGkg PSAwOyBpIDwgdGV2LT5uYXJnczsgaSsrKQ0KPiArCQljb3B5X3RvX3Byb2JlX3RyYWNlX2FyZygm dGV2LT5hcmdzW2ldLCAmcGV2LT5hcmdzW2ldKTsNCj4gKw0KPiArCXJldHVybiAxOw0KPiArDQo+ ICtlcnJvdXQ6DQo+ICsJaWYgKCp0ZXZzKSB7DQo+ICsJCWNsZWFyX3Byb2JlX3RyYWNlX2V2ZW50 cygqdGV2cywgMSk7DQo+ICsJCSp0ZXZzID0gTlVMTDsNCj4gKwl9DQo+ICsJcmV0dXJuIGVycjsN Cj4gK30NCj4gKw0KPiAgYm9vbCBfX3dlYWsgYXJjaF9fcHJlZmVyc19zeW10YWIodm9pZCkgeyBy ZXR1cm4gZmFsc2U7IH0NCj4gDQo+ICBzdGF0aWMgaW50IGNvbnZlcnRfdG9fcHJvYmVfdHJhY2Vf ZXZlbnRzKHN0cnVjdCBwZXJmX3Byb2JlX2V2ZW50ICpwZXYsDQo+IEBAIC0yNTg4LDYgKzI2ODUs MTAgQEAgc3RhdGljIGludCBjb252ZXJ0X3RvX3Byb2JlX3RyYWNlX2V2ZW50cyhzdHJ1Y3QgcGVy Zl9wcm9iZV9ldmVudCAqcGV2LA0KPiAgCQl9DQo+ICAJfQ0KPiANCj4gKwlyZXQgPSB0cnlfdG9f ZmluZF9hYnNvbHV0ZV9hZGRyZXNzKHBldiwgdGV2cyk7DQo+ICsJaWYgKHJldCA+IDApDQo+ICsJ CXJldHVybiByZXQ7DQo+ICsNCj4gIAlpZiAoYXJjaF9fcHJlZmVyc19zeW10YWIoKSAmJiAhcGVy Zl9wcm9iZV9ldmVudF9uZWVkX2R3YXJmKHBldikpIHsNCj4gIAkJcmV0ID0gZmluZF9wcm9iZV90 cmFjZV9ldmVudHNfZnJvbV9tYXAocGV2LCB0ZXZzKTsNCj4gIAkJaWYgKHJldCA+IDApDQo+IEBA IC0yNzU4LDMgKzI4NTksMjIgQEAgZW5kOg0KPiAgCXJldHVybiByZXQ7DQo+ICB9DQo+IA0KPiAr aW50IGNvcHlfdG9fcHJvYmVfdHJhY2VfYXJnKHN0cnVjdCBwcm9iZV90cmFjZV9hcmcgKnR2YXIs DQo+ICsJCQkgICAgc3RydWN0IHBlcmZfcHJvYmVfYXJnICpwdmFyKQ0KPiArew0KPiArCXR2YXIt PnZhbHVlID0gc3RyZHVwKHB2YXItPnZhcik7DQo+ICsJaWYgKHR2YXItPnZhbHVlID09IE5VTEwp DQo+ICsJCXJldHVybiAtRU5PTUVNOw0KPiArCWlmIChwdmFyLT50eXBlKSB7DQo+ICsJCXR2YXIt PnR5cGUgPSBzdHJkdXAocHZhci0+dHlwZSk7DQo+ICsJCWlmICh0dmFyLT50eXBlID09IE5VTEwp DQo+ICsJCQlyZXR1cm4gLUVOT01FTTsNCj4gKwl9DQo+ICsJaWYgKHB2YXItPm5hbWUpIHsNCj4g KwkJdHZhci0+bmFtZSA9IHN0cmR1cChwdmFyLT5uYW1lKTsNCj4gKwkJaWYgKHR2YXItPm5hbWUg PT0gTlVMTCkNCj4gKwkJCXJldHVybiAtRU5PTUVNOw0KPiArCX0gZWxzZQ0KPiArCQl0dmFyLT5u YW1lID0gTlVMTDsNCj4gKwlyZXR1cm4gMDsNCj4gK30NCj4gZGlmZiAtLWdpdCBhL3Rvb2xzL3Bl cmYvdXRpbC9wcm9iZS1ldmVudC5oIGIvdG9vbHMvcGVyZi91dGlsL3Byb2JlLWV2ZW50LmgNCj4g aW5kZXggODNlZTk1ZS4uMTc0YTNjZiAxMDA2NDQNCj4gLS0tIGEvdG9vbHMvcGVyZi91dGlsL3By b2JlLWV2ZW50LmgNCj4gKysrIGIvdG9vbHMvcGVyZi91dGlsL3Byb2JlLWV2ZW50LmgNCj4gQEAg LTE1Niw0ICsxNTYsNyBAQCBpbnQgZV9zbnByaW50ZihjaGFyICpzdHIsIHNpemVfdCBzaXplLCBj b25zdCBjaGFyICpmb3JtYXQsIC4uLikNCj4gIC8qIE1heGltdW0gaW5kZXggbnVtYmVyIG9mIGV2 ZW50LW5hbWUgcG9zdGZpeCAqLw0KPiAgI2RlZmluZSBNQVhfRVZFTlRfSU5ERVgJMTAyNA0KPiAN Cj4gK2ludCBjb3B5X3RvX3Byb2JlX3RyYWNlX2FyZyhzdHJ1Y3QgcHJvYmVfdHJhY2VfYXJnICp0 dmFyLA0KPiArCQkJICAgIHN0cnVjdCBwZXJmX3Byb2JlX2FyZyAqcHZhcik7DQo+ICsNCj4gICNl bmRpZiAvKl9QUk9CRV9FVkVOVF9IICovDQo+IGRpZmYgLS1naXQgYS90b29scy9wZXJmL3V0aWwv cHJvYmUtZmluZGVyLmMgYi90b29scy9wZXJmL3V0aWwvcHJvYmUtZmluZGVyLmMNCj4gaW5kZXgg N2I4MGY4Yy4uMjljNDNjMDY4IDEwMDY0NA0KPiAtLS0gYS90b29scy9wZXJmL3V0aWwvcHJvYmUt ZmluZGVyLmMNCj4gKysrIGIvdG9vbHMvcGVyZi91dGlsL3Byb2JlLWZpbmRlci5jDQo+IEBAIC01 NTMsMjQgKzU1Myw5IEBAIHN0YXRpYyBpbnQgZmluZF92YXJpYWJsZShEd2FyZl9EaWUgKnNjX2Rp ZSwgc3RydWN0IHByb2JlX2ZpbmRlciAqcGYpDQo+ICAJY2hhciBidWZbMzJdLCAqcHRyOw0KPiAg CWludCByZXQgPSAwOw0KPiANCj4gLQlpZiAoIWlzX2NfdmFybmFtZShwZi0+cHZhci0+dmFyKSkg ew0KPiAtCQkvKiBDb3B5IHJhdyBwYXJhbWV0ZXJzICovDQo+IC0JCXBmLT50dmFyLT52YWx1ZSA9 IHN0cmR1cChwZi0+cHZhci0+dmFyKTsNCj4gLQkJaWYgKHBmLT50dmFyLT52YWx1ZSA9PSBOVUxM KQ0KPiAtCQkJcmV0dXJuIC1FTk9NRU07DQo+IC0JCWlmIChwZi0+cHZhci0+dHlwZSkgew0KPiAt CQkJcGYtPnR2YXItPnR5cGUgPSBzdHJkdXAocGYtPnB2YXItPnR5cGUpOw0KPiAtCQkJaWYgKHBm LT50dmFyLT50eXBlID09IE5VTEwpDQo+IC0JCQkJcmV0dXJuIC1FTk9NRU07DQo+IC0JCX0NCj4g LQkJaWYgKHBmLT5wdmFyLT5uYW1lKSB7DQo+IC0JCQlwZi0+dHZhci0+bmFtZSA9IHN0cmR1cChw Zi0+cHZhci0+bmFtZSk7DQo+IC0JCQlpZiAocGYtPnR2YXItPm5hbWUgPT0gTlVMTCkNCj4gLQkJ CQlyZXR1cm4gLUVOT01FTTsNCj4gLQkJfSBlbHNlDQo+IC0JCQlwZi0+dHZhci0+bmFtZSA9IE5V TEw7DQo+IC0JCXJldHVybiAwOw0KPiAtCX0NCj4gKwkvKiBDb3B5IHJhdyBwYXJhbWV0ZXJzICov DQo+ICsJaWYgKCFpc19jX3Zhcm5hbWUocGYtPnB2YXItPnZhcikpDQo+ICsJCXJldHVybiBjb3B5 X3RvX3Byb2JlX3RyYWNlX2FyZyhwZi0+dHZhciwgcGYtPnB2YXIpOw0KPiANCj4gIAlpZiAocGYt PnB2YXItPm5hbWUpDQo+ICAJCXBmLT50dmFyLT5uYW1lID0gc3RyZHVwKHBmLT5wdmFyLT5uYW1l KTsNCj4gLS0NCj4gMS44LjMuNA0KDQo= -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | "Wangnan (F)" <wangnan0@huawei.com> |
|---|---|
| Date | 2015-08-26 04:40 +0200 |
| Subject | Re: [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1yN3-LO-7@gated-at.bofh.it> |
| In reply to | #1213451 |
On 2015/8/26 8:02, 平松雅巳 / HIRAMATU，MASAMI wrote:
>> From: Wang Nan [mailto:wangnan0@huawei.com]
>>
>> It should be useful to allow 'perf probe' probe at absolute offset of
>> a target. For example, when (u)probing at a instruction of a shared
>> object in a embedded system where debuginfo is not avaliable but we
>> know the offset of that instruction by manually digging.
>>
>> This patch enables following perf probe command syntax:
>>
>> # perf probe +0xffffffff811e6615
>>
>> And
>>
>> # perf probe /lib/x86_64-linux-gnu/libc-2.19.so +0xeb860
> Why do we need "+" for the absolute address?
> It seems that we can do it if we find that the given probe point
> starts with "0x".
>
> Thanks,
I will change 2/2 as you suggection.
However, we can only ensure that in kernel side symbol never leading
with '0x'. Although I don't think symbol leading with 0x is useful,
it is still possible for a userspace program compiled and linked by
a language other than C produces such symbol. '+' helps us separate
address and function name semantically, make us don't rely on assumption
on function names. If in future we do meet '0x' symbols, I think we still
need the '+' syntax back. But we can do it at that time.
Thank you.
>> In the above example, we don't need a anchor symbol, so it is possible
>> to compute absolute addresses using other methods and then use
>> 'perf probe' to create the probing points.
>>
>> Signed-off-by: Wang Nan <wangnan0@huawei.com>
>> Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
>> Cc: Ingo Molnar <mingo@redhat.com>
>> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
>> Cc: Namhyung Kim <namhyung@kernel.org>
>> ---
>> tools/perf/util/probe-event.c | 144 +++++++++++++++++++++++++++++++++++++----
>> tools/perf/util/probe-event.h | 3 +
>> tools/perf/util/probe-finder.c | 21 +-----
>> 3 files changed, 138 insertions(+), 30 deletions(-)
>>
>> diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
>> index 6c7e538..59de69a4 100644
>> --- a/tools/perf/util/probe-event.c
>> +++ b/tools/perf/util/probe-event.c
>> @@ -1194,9 +1194,13 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
>> *ptr++ = '\0';
>> }
>>
>> - tmp = strdup(arg);
>> - if (tmp == NULL)
>> - return -ENOMEM;
>> + if (arg[0] == '\0')
>> + tmp = NULL;
>> + else {
>> + tmp = strdup(arg);
>> + if (tmp == NULL)
>> + return -ENOMEM;
>> + }
>>
>> if (file_spec)
>> pp->file = tmp;
>> @@ -1283,11 +1287,6 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
>> return -EINVAL;
>> }
>>
>> - if (pp->offset && !pp->function) {
>> - semantic_error("Offset requires an entry function.\n");
>> - return -EINVAL;
>> - }
>> -
>> if (pp->retprobe && !pp->function) {
>> semantic_error("Return probe requires an entry function.\n");
>> return -EINVAL;
>> @@ -1299,6 +1298,11 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
>> return -EINVAL;
>> }
>>
>> + if (!pp->function && !pp->offset && !pp->file) {
>> + semantic_error("Absolute address should not be zero.\n");
>> + return -EINVAL;
>> + }
>> +
>> pr_debug("symbol:%s file:%s line:%d offset:%lu return:%d lazy:%s\n",
>> pp->function, pp->file, pp->line, pp->offset, pp->retprobe,
>> pp->lazy_line);
>> @@ -1609,7 +1613,7 @@ error:
>> static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
>> {
>> char *buf, *tmp;
>> - char offs[32] = "", line[32] = "", file[32] = "";
>> + char offs[32] = "", line[32] = "", file[32] = "", addr[32] = "";
>> int ret, len;
>>
>> buf = zalloc(MAX_CMDLEN);
>> @@ -1622,6 +1626,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
>> if (ret <= 0)
>> goto error;
>> }
>> + if (!pp->function) {
>> + ret = e_snprintf(addr, 32, "0x%lx", pp->offset);
>> + if (ret <= 0)
>> + goto error;
>> + }
>> if (pp->line) {
>> ret = e_snprintf(line, 32, ":%d", pp->line);
>> if (ret <= 0)
>> @@ -1639,9 +1648,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
>> goto error;
>> }
>>
>> - if (pp->function)
>> - ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s", pp->function,
>> - offs, pp->retprobe ? "%return" : "", line,
>> + if (pp->function || pp->offset)
>> + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s",
>> + pp->function ? : addr,
>> + pp->function ? offs : "",
>> + pp->retprobe ? "%return" : "", line,
>> file);
>> else
>> ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", file, line);
>> @@ -1786,6 +1797,11 @@ char *synthesize_probe_trace_command(struct probe_trace_event *tev)
>> if (tev->uprobes)
>> ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s:0x%lx",
>> tp->module, tp->address);
>> + else if (tp->symbol[0] == '0' && tp->symbol[1] == 'x')
>> + /* Absolute address. See try_to_find_absolute_address() */
>> + ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s0x%lx",
>> + tp->module ?: "", tp->module ? ":" : "",
>> + tp->address);
>> else
>> ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s%s+%lu",
>> tp->module ?: "", tp->module ? ":" : "",
>> @@ -2572,6 +2588,87 @@ err_out:
>> goto out;
>> }
>>
>> +static int try_to_find_absolute_address(struct perf_probe_event *pev,
>> + struct probe_trace_event **tevs)
>> +{
>> + struct perf_probe_point *pp = &pev->point;
>> + struct probe_trace_event *tev;
>> + struct probe_trace_point *tp;
>> + int i, err;
>> +
>> + if (perf_probe_event_need_dwarf(pev) || pev->point.function)
>> + return -EINVAL;
>> +
>> + /*
>> + * This is 'perf probe /lib/libc.so +0xabcd'. Try to probe at
>> + * absolute address.
>> + *
>> + * Only one tev can be generated by this.
>> + */
>> + *tevs = zalloc(sizeof(*tev));
>> + if (!*tevs)
>> + return -ENOMEM;
>> +
>> + tev = *tevs;
>> + tp = &tev->point;
>> +
>> + /*
>> + * Don't use tp->offset, use address directly, because
>> + * in synthesize_probe_trace_command() address cannot be
>> + * zero.
>> + */
>> + tp->address = pev->point.offset;
>> + tp->retprobe = pp->retprobe;
>> + tev->uprobes = pev->uprobes;
>> +
>> + err = -ENOMEM;
>> + /* Give it a '0x' leading symbol name */
>> + if (asprintf(&tp->symbol, "0x%lx", tp->address) < 0)
>> + goto errout;
>> +
>> + /* For kprobe, check range */
>> + if ((!tev->uprobes) &&
>> + (kprobe_warn_out_range(tev->point.symbol,
>> + tev->point.address))) {
>> + err = -EACCES;
>> + goto errout;
>> + }
>> +
>> + if (asprintf(&tp->realname, "abs_%lx", tp->address) < 0)
>> + goto errout;
>> +
>> + if (pev->target) {
>> + tp->module = strdup(pev->target);
>> + if (!tp->module)
>> + goto errout;
>> + }
>> +
>> + if (tev->group) {
>> + tev->group = strdup(pev->group);
>> + if (!tev->group)
>> + goto errout;
>> + }
>> +
>> + if (pev->event) {
>> + tev->event = strdup(pev->event);
>> + if (!tev->event)
>> + goto errout;
>> + }
>> +
>> + tev->nargs = pev->nargs;
>> + for (i = 0; i < tev->nargs; i++)
>> + copy_to_probe_trace_arg(&tev->args[i], &pev->args[i]);
>> +
>> + return 1;
>> +
>> +errout:
>> + if (*tevs) {
>> + clear_probe_trace_events(*tevs, 1);
>> + *tevs = NULL;
>> + }
>> + return err;
>> +}
>> +
>> bool __weak arch__prefers_symtab(void) { return false; }
>>
>> static int convert_to_probe_trace_events(struct perf_probe_event *pev,
>> @@ -2588,6 +2685,10 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
>> }
>> }
>>
>> + ret = try_to_find_absolute_address(pev, tevs);
>> + if (ret > 0)
>> + return ret;
>> +
>> if (arch__prefers_symtab() && !perf_probe_event_need_dwarf(pev)) {
>> ret = find_probe_trace_events_from_map(pev, tevs);
>> if (ret > 0)
>> @@ -2758,3 +2859,22 @@ end:
>> return ret;
>> }
>>
>> +int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
>> + struct perf_probe_arg *pvar)
>> +{
>> + tvar->value = strdup(pvar->var);
>> + if (tvar->value == NULL)
>> + return -ENOMEM;
>> + if (pvar->type) {
>> + tvar->type = strdup(pvar->type);
>> + if (tvar->type == NULL)
>> + return -ENOMEM;
>> + }
>> + if (pvar->name) {
>> + tvar->name = strdup(pvar->name);
>> + if (tvar->name == NULL)
>> + return -ENOMEM;
>> + } else
>> + tvar->name = NULL;
>> + return 0;
>> +}
>> diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
>> index 83ee95e..174a3cf 100644
>> --- a/tools/perf/util/probe-event.h
>> +++ b/tools/perf/util/probe-event.h
>> @@ -156,4 +156,7 @@ int e_snprintf(char *str, size_t size, const char *format, ...)
>> /* Maximum index number of event-name postfix */
>> #define MAX_EVENT_INDEX 1024
>>
>> +int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
>> + struct perf_probe_arg *pvar);
>> +
>> #endif /*_PROBE_EVENT_H */
>> diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
>> index 7b80f8c..29c43c068 100644
>> --- a/tools/perf/util/probe-finder.c
>> +++ b/tools/perf/util/probe-finder.c
>> @@ -553,24 +553,9 @@ static int find_variable(Dwarf_Die *sc_die, struct probe_finder *pf)
>> char buf[32], *ptr;
>> int ret = 0;
>>
>> - if (!is_c_varname(pf->pvar->var)) {
>> - /* Copy raw parameters */
>> - pf->tvar->value = strdup(pf->pvar->var);
>> - if (pf->tvar->value == NULL)
>> - return -ENOMEM;
>> - if (pf->pvar->type) {
>> - pf->tvar->type = strdup(pf->pvar->type);
>> - if (pf->tvar->type == NULL)
>> - return -ENOMEM;
>> - }
>> - if (pf->pvar->name) {
>> - pf->tvar->name = strdup(pf->pvar->name);
>> - if (pf->tvar->name == NULL)
>> - return -ENOMEM;
>> - } else
>> - pf->tvar->name = NULL;
>> - return 0;
>> - }
>> + /* Copy raw parameters */
>> + if (!is_c_varname(pf->pvar->var))
>> + return copy_to_probe_trace_arg(pf->tvar, pf->pvar);
>>
>> if (pf->pvar->name)
>> pf->tvar->name = strdup(pf->pvar->name);
>> --
>> 1.8.3.4
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | 平松雅巳 / HIRAMATU，MASAMI <masami.hiramatsu.pt@hitachi.com> |
|---|---|
| Date | 2015-08-26 11:10 +0200 |
| Subject | RE: [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1ESu-1fe-13@gated-at.bofh.it> |
| In reply to | #1213503 |
PiBGcm9tOiBXYW5nbmFuIChGKSBbbWFpbHRvOndhbmduYW4wQGh1YXdlaS5jb21dIA0KPiANCj4g T24gMjAxNS84LzI2IDg6MDIsIOW5s+advumbheW3syAvIEhJUkFNQVRV77yMTUFTQU1JIHdyb3Rl Og0KPiA+PiBGcm9tOiBXYW5nIE5hbiBbbWFpbHRvOndhbmduYW4wQGh1YXdlaS5jb21dDQo+ID4+ DQo+ID4+IEl0IHNob3VsZCBiZSB1c2VmdWwgdG8gYWxsb3cgJ3BlcmYgcHJvYmUnIHByb2JlIGF0 IGFic29sdXRlIG9mZnNldCBvZg0KPiA+PiBhIHRhcmdldC4gRm9yIGV4YW1wbGUsIHdoZW4gKHUp cHJvYmluZyBhdCBhIGluc3RydWN0aW9uIG9mIGEgc2hhcmVkDQo+ID4+IG9iamVjdCBpbiBhIGVt YmVkZGVkIHN5c3RlbSB3aGVyZSBkZWJ1Z2luZm8gaXMgbm90IGF2YWxpYWJsZSBidXQgd2UNCj4g Pj4ga25vdyB0aGUgb2Zmc2V0IG9mIHRoYXQgaW5zdHJ1Y3Rpb24gYnkgbWFudWFsbHkgZGlnZ2lu Zy4NCj4gPj4NCj4gPj4gVGhpcyBwYXRjaCBlbmFibGVzIGZvbGxvd2luZyBwZXJmIHByb2JlIGNv bW1hbmQgc3ludGF4Og0KPiA+Pg0KPiA+PiAgICMgcGVyZiBwcm9iZSArMHhmZmZmZmZmZjgxMWU2 NjE1DQo+ID4+DQo+ID4+IEFuZA0KPiA+Pg0KPiA+PiAgICMgcGVyZiBwcm9iZSAvbGliL3g4Nl82 NC1saW51eC1nbnUvbGliYy0yLjE5LnNvICsweGViODYwDQo+ID4gV2h5IGRvIHdlIG5lZWQgIisi IGZvciB0aGUgYWJzb2x1dGUgYWRkcmVzcz8NCj4gPiBJdCBzZWVtcyB0aGF0IHdlIGNhbiBkbyBp dCBpZiB3ZSBmaW5kIHRoYXQgdGhlIGdpdmVuIHByb2JlIHBvaW50DQo+ID4gc3RhcnRzIHdpdGgg IjB4Ii4NCj4gPg0KPiA+IFRoYW5rcywNCj4gDQo+IEkgd2lsbCBjaGFuZ2UgMi8yIGFzIHlvdSBz dWdnZWN0aW9uLg0KPiANCj4gSG93ZXZlciwgd2UgY2FuIG9ubHkgZW5zdXJlIHRoYXQgaW4ga2Vy bmVsIHNpZGUgc3ltYm9sIG5ldmVyIGxlYWRpbmcNCj4gd2l0aCAnMHgnLiBBbHRob3VnaCBJIGRv bid0IHRoaW5rIHN5bWJvbCBsZWFkaW5nIHdpdGggMHggaXMgdXNlZnVsLA0KPiBpdCBpcyBzdGls bCBwb3NzaWJsZSBmb3IgYSB1c2Vyc3BhY2UgcHJvZ3JhbSBjb21waWxlZCBhbmQgbGlua2VkIGJ5 DQo+IGEgbGFuZ3VhZ2Ugb3RoZXIgdGhhbiBDIHByb2R1Y2VzIHN1Y2ggc3ltYm9sLiAnKycgaGVs cHMgdXMgc2VwYXJhdGUNCj4gYWRkcmVzcyBhbmQgZnVuY3Rpb24gbmFtZSBzZW1hbnRpY2FsbHks IG1ha2UgdXMgZG9uJ3QgcmVseSBvbiBhc3N1bXB0aW9uDQo+IG9uIGZ1bmN0aW9uIG5hbWVzLiBJ ZiBpbiBmdXR1cmUgd2UgZG8gbWVldCAnMHgnIHN5bWJvbHMsIEkgdGhpbmsgd2Ugc3RpbGwNCj4g bmVlZCB0aGUgJysnIHN5bnRheCBiYWNrLiBCdXQgd2UgY2FuIGRvIGl0IGF0IHRoYXQgdGltZS4N Cg0KQWdyZWVkLCBpZiB3ZSByZWFsbHkgZmluZCB0aGUgY2FzZSwgd2UgY2FuIGFkZCAnKycgcHJl Zml4IHN1cHBvcnQNCmp1c3QgZm9yIGFuIG9wdGlvbmFsIGV4cHJlc3Npb24gOikNCg0KVGhhbmsg eW91IQ0KDQo+IA0KPiBUaGFuayB5b3UuDQo+IA0KPiANCj4gPj4gSW4gdGhlIGFib3ZlIGV4YW1w bGUsIHdlIGRvbid0IG5lZWQgYSBhbmNob3Igc3ltYm9sLCBzbyBpdCBpcyBwb3NzaWJsZQ0KPiA+ PiB0byBjb21wdXRlIGFic29sdXRlIGFkZHJlc3NlcyB1c2luZyBvdGhlciBtZXRob2RzIGFuZCB0 aGVuIHVzZQ0KPiA+PiAncGVyZiBwcm9iZScgdG8gY3JlYXRlIHRoZSBwcm9iaW5nIHBvaW50cy4N Cj4gPj4NCj4gPj4gU2lnbmVkLW9mZi1ieTogV2FuZyBOYW4gPHdhbmduYW4wQGh1YXdlaS5jb20+ DQo+ID4+IENjOiBBcm5hbGRvIENhcnZhbGhvIGRlIE1lbG8gPGFjbWVAa2VybmVsLm9yZz4NCj4g Pj4gQ2M6IEluZ28gTW9sbmFyIDxtaW5nb0ByZWRoYXQuY29tPg0KPiA+PiBDYzogTWFzYW1pIEhp cmFtYXRzdSA8bWFzYW1pLmhpcmFtYXRzdS5wdEBoaXRhY2hpLmNvbT4NCj4gPj4gQ2M6IE5hbWh5 dW5nIEtpbSA8bmFtaHl1bmdAa2VybmVsLm9yZz4NCj4gPj4gLS0tDQo+ID4+ICAgdG9vbHMvcGVy Zi91dGlsL3Byb2JlLWV2ZW50LmMgIHwgMTQ0ICsrKysrKysrKysrKysrKysrKysrKysrKysrKysr KysrKysrKystLS0tDQo+ID4+ICAgdG9vbHMvcGVyZi91dGlsL3Byb2JlLWV2ZW50LmggIHwgICAz ICsNCj4gPj4gICB0b29scy9wZXJmL3V0aWwvcHJvYmUtZmluZGVyLmMgfCAgMjEgKy0tLS0tDQo+ ID4+ICAgMyBmaWxlcyBjaGFuZ2VkLCAxMzggaW5zZXJ0aW9ucygrKSwgMzAgZGVsZXRpb25zKC0p DQo+ID4+DQo+ID4+IGRpZmYgLS1naXQgYS90b29scy9wZXJmL3V0aWwvcHJvYmUtZXZlbnQuYyBi L3Rvb2xzL3BlcmYvdXRpbC9wcm9iZS1ldmVudC5jDQo+ID4+IGluZGV4IDZjN2U1MzguLjU5ZGU2 OWE0IDEwMDY0NA0KPiA+PiAtLS0gYS90b29scy9wZXJmL3V0aWwvcHJvYmUtZXZlbnQuYw0KPiA+ PiArKysgYi90b29scy9wZXJmL3V0aWwvcHJvYmUtZXZlbnQuYw0KPiA+PiBAQCAtMTE5NCw5ICsx MTk0LDEzIEBAIHN0YXRpYyBpbnQgcGFyc2VfcGVyZl9wcm9iZV9wb2ludChjaGFyICphcmcsIHN0 cnVjdCBwZXJmX3Byb2JlX2V2ZW50ICpwZXYpDQo+ID4+ICAgCQkqcHRyKysgPSAnXDAnOw0KPiA+ PiAgIAl9DQo+ID4+DQo+ID4+IC0JdG1wID0gc3RyZHVwKGFyZyk7DQo+ID4+IC0JaWYgKHRtcCA9 PSBOVUxMKQ0KPiA+PiAtCQlyZXR1cm4gLUVOT01FTTsNCj4gPj4gKwlpZiAoYXJnWzBdID09ICdc MCcpDQo+ID4+ICsJCXRtcCA9IE5VTEw7DQo+ID4+ICsJZWxzZSB7DQo+ID4+ICsJCXRtcCA9IHN0 cmR1cChhcmcpOw0KPiA+PiArCQlpZiAodG1wID09IE5VTEwpDQo+ID4+ICsJCQlyZXR1cm4gLUVO T01FTTsNCj4gPj4gKwl9DQo+ID4+DQo+ID4+ICAgCWlmIChmaWxlX3NwZWMpDQo+ID4+ICAgCQlw cC0+ZmlsZSA9IHRtcDsNCj4gPj4gQEAgLTEyODMsMTEgKzEyODcsNiBAQCBzdGF0aWMgaW50IHBh cnNlX3BlcmZfcHJvYmVfcG9pbnQoY2hhciAqYXJnLCBzdHJ1Y3QgcGVyZl9wcm9iZV9ldmVudCAq cGV2KQ0KPiA+PiAgIAkJcmV0dXJuIC1FSU5WQUw7DQo+ID4+ICAgCX0NCj4gPj4NCj4gPj4gLQlp ZiAocHAtPm9mZnNldCAmJiAhcHAtPmZ1bmN0aW9uKSB7DQo+ID4+IC0JCXNlbWFudGljX2Vycm9y KCJPZmZzZXQgcmVxdWlyZXMgYW4gZW50cnkgZnVuY3Rpb24uXG4iKTsNCj4gPj4gLQkJcmV0dXJu IC1FSU5WQUw7DQo+ID4+IC0JfQ0KPiA+PiAtDQo+ID4+ICAgCWlmIChwcC0+cmV0cHJvYmUgJiYg IXBwLT5mdW5jdGlvbikgew0KPiA+PiAgIAkJc2VtYW50aWNfZXJyb3IoIlJldHVybiBwcm9iZSBy ZXF1aXJlcyBhbiBlbnRyeSBmdW5jdGlvbi5cbiIpOw0KPiA+PiAgIAkJcmV0dXJuIC1FSU5WQUw7 DQo+ID4+IEBAIC0xMjk5LDYgKzEyOTgsMTEgQEAgc3RhdGljIGludCBwYXJzZV9wZXJmX3Byb2Jl X3BvaW50KGNoYXIgKmFyZywgc3RydWN0IHBlcmZfcHJvYmVfZXZlbnQgKnBldikNCj4gPj4gICAJ CXJldHVybiAtRUlOVkFMOw0KPiA+PiAgIAl9DQo+ID4+DQo+ID4+ICsJaWYgKCFwcC0+ZnVuY3Rp b24gJiYgIXBwLT5vZmZzZXQgJiYgIXBwLT5maWxlKSB7DQo+ID4+ICsJCXNlbWFudGljX2Vycm9y KCJBYnNvbHV0ZSBhZGRyZXNzIHNob3VsZCBub3QgYmUgemVyby5cbiIpOw0KPiA+PiArCQlyZXR1 cm4gLUVJTlZBTDsNCj4gPj4gKwl9DQo+ID4+ICsNCj4gPj4gICAJcHJfZGVidWcoInN5bWJvbDol cyBmaWxlOiVzIGxpbmU6JWQgb2Zmc2V0OiVsdSByZXR1cm46JWQgbGF6eTolc1xuIiwNCj4gPj4g ICAJCSBwcC0+ZnVuY3Rpb24sIHBwLT5maWxlLCBwcC0+bGluZSwgcHAtPm9mZnNldCwgcHAtPnJl dHByb2JlLA0KPiA+PiAgIAkJIHBwLT5sYXp5X2xpbmUpOw0KPiA+PiBAQCAtMTYwOSw3ICsxNjEz LDcgQEAgZXJyb3I6DQo+ID4+ICAgc3RhdGljIGNoYXIgKnN5bnRoZXNpemVfcGVyZl9wcm9iZV9w b2ludChzdHJ1Y3QgcGVyZl9wcm9iZV9wb2ludCAqcHApDQo+ID4+ICAgew0KPiA+PiAgIAljaGFy ICpidWYsICp0bXA7DQo+ID4+IC0JY2hhciBvZmZzWzMyXSA9ICIiLCBsaW5lWzMyXSA9ICIiLCBm aWxlWzMyXSA9ICIiOw0KPiA+PiArCWNoYXIgb2Zmc1szMl0gPSAiIiwgbGluZVszMl0gPSAiIiwg ZmlsZVszMl0gPSAiIiwgYWRkclszMl0gPSAiIjsNCj4gPj4gICAJaW50IHJldCwgbGVuOw0KPiA+ Pg0KPiA+PiAgIAlidWYgPSB6YWxsb2MoTUFYX0NNRExFTik7DQo+ID4+IEBAIC0xNjIyLDYgKzE2 MjYsMTEgQEAgc3RhdGljIGNoYXIgKnN5bnRoZXNpemVfcGVyZl9wcm9iZV9wb2ludChzdHJ1Y3Qg cGVyZl9wcm9iZV9wb2ludCAqcHApDQo+ID4+ICAgCQlpZiAocmV0IDw9IDApDQo+ID4+ICAgCQkJ Z290byBlcnJvcjsNCj4gPj4gICAJfQ0KPiA+PiArCWlmICghcHAtPmZ1bmN0aW9uKSB7DQo+ID4+ ICsJCXJldCA9IGVfc25wcmludGYoYWRkciwgMzIsICIweCVseCIsIHBwLT5vZmZzZXQpOw0KPiA+ PiArCQlpZiAocmV0IDw9IDApDQo+ID4+ICsJCQlnb3RvIGVycm9yOw0KPiA+PiArCX0NCj4gPj4g ICAJaWYgKHBwLT5saW5lKSB7DQo+ID4+ICAgCQlyZXQgPSBlX3NucHJpbnRmKGxpbmUsIDMyLCAi OiVkIiwgcHAtPmxpbmUpOw0KPiA+PiAgIAkJaWYgKHJldCA8PSAwKQ0KPiA+PiBAQCAtMTYzOSw5 ICsxNjQ4LDExIEBAIHN0YXRpYyBjaGFyICpzeW50aGVzaXplX3BlcmZfcHJvYmVfcG9pbnQoc3Ry dWN0IHBlcmZfcHJvYmVfcG9pbnQgKnBwKQ0KPiA+PiAgIAkJCWdvdG8gZXJyb3I7DQo+ID4+ICAg CX0NCj4gPj4NCj4gPj4gLQlpZiAocHAtPmZ1bmN0aW9uKQ0KPiA+PiAtCQlyZXQgPSBlX3NucHJp bnRmKGJ1ZiwgTUFYX0NNRExFTiwgIiVzJXMlcyVzJXMiLCBwcC0+ZnVuY3Rpb24sDQo+ID4+IC0J CQkJIG9mZnMsIHBwLT5yZXRwcm9iZSA/ICIlcmV0dXJuIiA6ICIiLCBsaW5lLA0KPiA+PiArCWlm IChwcC0+ZnVuY3Rpb24gfHwgcHAtPm9mZnNldCkNCj4gPj4gKwkJcmV0ID0gZV9zbnByaW50Zihi dWYsIE1BWF9DTURMRU4sICIlcyVzJXMlcyVzIiwNCj4gPj4gKwkJCQkgcHAtPmZ1bmN0aW9uID8g OiBhZGRyLA0KPiA+PiArCQkJCSBwcC0+ZnVuY3Rpb24gPyBvZmZzIDogIiIsDQo+ID4+ICsJCQkJ IHBwLT5yZXRwcm9iZSA/ICIlcmV0dXJuIiA6ICIiLCBsaW5lLA0KPiA+PiAgIAkJCQkgZmlsZSk7 DQo+ID4+ICAgCWVsc2UNCj4gPj4gICAJCXJldCA9IGVfc25wcmludGYoYnVmLCBNQVhfQ01ETEVO LCAiJXMlcyIsIGZpbGUsIGxpbmUpOw0KPiA+PiBAQCAtMTc4Niw2ICsxNzk3LDExIEBAIGNoYXIg KnN5bnRoZXNpemVfcHJvYmVfdHJhY2VfY29tbWFuZChzdHJ1Y3QgcHJvYmVfdHJhY2VfZXZlbnQg KnRldikNCj4gPj4gICAJaWYgKHRldi0+dXByb2JlcykNCj4gPj4gICAJCXJldCA9IGVfc25wcmlu dGYoYnVmICsgbGVuLCBNQVhfQ01ETEVOIC0gbGVuLCAiJXM6MHglbHgiLA0KPiA+PiAgIAkJCQkg dHAtPm1vZHVsZSwgdHAtPmFkZHJlc3MpOw0KPiA+PiArCWVsc2UgaWYgKHRwLT5zeW1ib2xbMF0g PT0gJzAnICYmIHRwLT5zeW1ib2xbMV0gPT0gJ3gnKQ0KPiA+PiArCQkvKiBBYnNvbHV0ZSBhZGRy ZXNzLiBTZWUgdHJ5X3RvX2ZpbmRfYWJzb2x1dGVfYWRkcmVzcygpICovDQo+ID4+ICsJCXJldCA9 IGVfc25wcmludGYoYnVmICsgbGVuLCBNQVhfQ01ETEVOIC0gbGVuLCAiJXMlczB4JWx4IiwNCj4g Pj4gKwkJCQkgdHAtPm1vZHVsZSA/OiAiIiwgdHAtPm1vZHVsZSA/ICI6IiA6ICIiLA0KPiA+PiAr CQkJCSB0cC0+YWRkcmVzcyk7DQo+ID4+ICAgCWVsc2UNCj4gPj4gICAJCXJldCA9IGVfc25wcmlu dGYoYnVmICsgbGVuLCBNQVhfQ01ETEVOIC0gbGVuLCAiJXMlcyVzKyVsdSIsDQo+ID4+ICAgCQkJ CSB0cC0+bW9kdWxlID86ICIiLCB0cC0+bW9kdWxlID8gIjoiIDogIiIsDQo+ID4+IEBAIC0yNTcy LDYgKzI1ODgsODcgQEAgZXJyX291dDoNCj4gPj4gICAJZ290byBvdXQ7DQo+ID4+ICAgfQ0KPiA+ Pg0KPiA+PiArc3RhdGljIGludCB0cnlfdG9fZmluZF9hYnNvbHV0ZV9hZGRyZXNzKHN0cnVjdCBw ZXJmX3Byb2JlX2V2ZW50ICpwZXYsDQo+ID4+ICsJCQkJCXN0cnVjdCBwcm9iZV90cmFjZV9ldmVu dCAqKnRldnMpDQo+ID4+ICt7DQo+ID4+ICsJc3RydWN0IHBlcmZfcHJvYmVfcG9pbnQgKnBwID0g JnBldi0+cG9pbnQ7DQo+ID4+ICsJc3RydWN0IHByb2JlX3RyYWNlX2V2ZW50ICp0ZXY7DQo+ID4+ ICsJc3RydWN0IHByb2JlX3RyYWNlX3BvaW50ICp0cDsNCj4gPj4gKwlpbnQgaSwgZXJyOw0KPiA+ PiArDQo+ID4+ICsJaWYgKHBlcmZfcHJvYmVfZXZlbnRfbmVlZF9kd2FyZihwZXYpIHx8IHBldi0+ cG9pbnQuZnVuY3Rpb24pDQo+ID4+ICsJCXJldHVybiAtRUlOVkFMOw0KPiA+PiArDQo+ID4+ICsJ LyoNCj4gPj4gKwkgKiBUaGlzIGlzICdwZXJmIHByb2JlIC9saWIvbGliYy5zbyArMHhhYmNkJy4g VHJ5IHRvIHByb2JlIGF0DQo+ID4+ICsJICogYWJzb2x1dGUgYWRkcmVzcy4NCj4gPj4gKwkgKg0K PiA+PiArCSAqIE9ubHkgb25lIHRldiBjYW4gYmUgZ2VuZXJhdGVkIGJ5IHRoaXMuDQo+ID4+ICsJ ICovDQo+ID4+ICsJKnRldnMgPSB6YWxsb2Moc2l6ZW9mKCp0ZXYpKTsNCj4gPj4gKwlpZiAoISp0 ZXZzKQ0KPiA+PiArCQlyZXR1cm4gLUVOT01FTTsNCj4gPj4gKw0KPiA+PiArCXRldiA9ICp0ZXZz Ow0KPiA+PiArCXRwID0gJnRldi0+cG9pbnQ7DQo+ID4+ICsNCj4gPj4gKwkvKg0KPiA+PiArCSAq IERvbid0IHVzZSB0cC0+b2Zmc2V0LCB1c2UgYWRkcmVzcyBkaXJlY3RseSwgYmVjYXVzZQ0KPiA+ PiArCSAqIGluIHN5bnRoZXNpemVfcHJvYmVfdHJhY2VfY29tbWFuZCgpIGFkZHJlc3MgY2Fubm90 IGJlDQo+ID4+ICsJICogemVyby4NCj4gPj4gKwkgKi8NCj4gPj4gKwl0cC0+YWRkcmVzcyA9IHBl di0+cG9pbnQub2Zmc2V0Ow0KPiA+PiArCXRwLT5yZXRwcm9iZSA9IHBwLT5yZXRwcm9iZTsNCj4g Pj4gKwl0ZXYtPnVwcm9iZXMgPSBwZXYtPnVwcm9iZXM7DQo+ID4+ICsNCj4gPj4gKwllcnIgPSAt RU5PTUVNOw0KPiA+PiArCS8qIEdpdmUgaXQgYSAnMHgnIGxlYWRpbmcgc3ltYm9sIG5hbWUgKi8N Cj4gPj4gKwlpZiAoYXNwcmludGYoJnRwLT5zeW1ib2wsICIweCVseCIsIHRwLT5hZGRyZXNzKSA8 IDApDQo+ID4+ICsJCWdvdG8gZXJyb3V0Ow0KPiA+PiArDQo+ID4+ICsJLyogRm9yIGtwcm9iZSwg Y2hlY2sgcmFuZ2UgKi8NCj4gPj4gKwlpZiAoKCF0ZXYtPnVwcm9iZXMpICYmDQo+ID4+ICsJICAg IChrcHJvYmVfd2Fybl9vdXRfcmFuZ2UodGV2LT5wb2ludC5zeW1ib2wsDQo+ID4+ICsJCQkJICAg dGV2LT5wb2ludC5hZGRyZXNzKSkpIHsNCj4gPj4gKwkJZXJyID0gLUVBQ0NFUzsNCj4gPj4gKwkJ Z290byBlcnJvdXQ7DQo+ID4+ICsJfQ0KPiA+PiArDQo+ID4+ICsJaWYgKGFzcHJpbnRmKCZ0cC0+ cmVhbG5hbWUsICJhYnNfJWx4IiwgdHAtPmFkZHJlc3MpIDwgMCkNCj4gPj4gKwkJZ290byBlcnJv dXQ7DQo+ID4+ICsNCj4gPj4gKwlpZiAocGV2LT50YXJnZXQpIHsNCj4gPj4gKwkJdHAtPm1vZHVs ZSA9IHN0cmR1cChwZXYtPnRhcmdldCk7DQo+ID4+ICsJCWlmICghdHAtPm1vZHVsZSkNCj4gPj4g KwkJCWdvdG8gZXJyb3V0Ow0KPiA+PiArCX0NCj4gPj4gKw0KPiA+PiArCWlmICh0ZXYtPmdyb3Vw KSB7DQo+ID4+ICsJCXRldi0+Z3JvdXAgPSBzdHJkdXAocGV2LT5ncm91cCk7DQo+ID4+ICsJCWlm ICghdGV2LT5ncm91cCkNCj4gPj4gKwkJCWdvdG8gZXJyb3V0Ow0KPiA+PiArCX0NCj4gPj4gKw0K PiA+PiArCWlmIChwZXYtPmV2ZW50KSB7DQo+ID4+ICsJCXRldi0+ZXZlbnQgPSBzdHJkdXAocGV2 LT5ldmVudCk7DQo+ID4+ICsJCWlmICghdGV2LT5ldmVudCkNCj4gPj4gKwkJCWdvdG8gZXJyb3V0 Ow0KPiA+PiArCX0NCj4gPj4gKw0KPiA+PiArCXRldi0+bmFyZ3MgPSBwZXYtPm5hcmdzOw0KPiA+ PiArCWZvciAoaSA9IDA7IGkgPCB0ZXYtPm5hcmdzOyBpKyspDQo+ID4+ICsJCWNvcHlfdG9fcHJv YmVfdHJhY2VfYXJnKCZ0ZXYtPmFyZ3NbaV0sICZwZXYtPmFyZ3NbaV0pOw0KPiA+PiArDQo+ID4+ ICsJcmV0dXJuIDE7DQo+ID4+ICsNCj4gPj4gK2Vycm91dDoNCj4gPj4gKwlpZiAoKnRldnMpIHsN Cj4gPj4gKwkJY2xlYXJfcHJvYmVfdHJhY2VfZXZlbnRzKCp0ZXZzLCAxKTsNCj4gPj4gKwkJKnRl dnMgPSBOVUxMOw0KPiA+PiArCX0NCj4gPj4gKwlyZXR1cm4gZXJyOw0KPiA+PiArfQ0KPiA+PiAr DQo+ID4+ICAgYm9vbCBfX3dlYWsgYXJjaF9fcHJlZmVyc19zeW10YWIodm9pZCkgeyByZXR1cm4g ZmFsc2U7IH0NCj4gPj4NCj4gPj4gICBzdGF0aWMgaW50IGNvbnZlcnRfdG9fcHJvYmVfdHJhY2Vf ZXZlbnRzKHN0cnVjdCBwZXJmX3Byb2JlX2V2ZW50ICpwZXYsDQo+ID4+IEBAIC0yNTg4LDYgKzI2 ODUsMTAgQEAgc3RhdGljIGludCBjb252ZXJ0X3RvX3Byb2JlX3RyYWNlX2V2ZW50cyhzdHJ1Y3Qg cGVyZl9wcm9iZV9ldmVudCAqcGV2LA0KPiA+PiAgIAkJfQ0KPiA+PiAgIAl9DQo+ID4+DQo+ID4+ ICsJcmV0ID0gdHJ5X3RvX2ZpbmRfYWJzb2x1dGVfYWRkcmVzcyhwZXYsIHRldnMpOw0KPiA+PiAr CWlmIChyZXQgPiAwKQ0KPiA+PiArCQlyZXR1cm4gcmV0Ow0KPiA+PiArDQo+ID4+ICAgCWlmIChh cmNoX19wcmVmZXJzX3N5bXRhYigpICYmICFwZXJmX3Byb2JlX2V2ZW50X25lZWRfZHdhcmYocGV2 KSkgew0KPiA+PiAgIAkJcmV0ID0gZmluZF9wcm9iZV90cmFjZV9ldmVudHNfZnJvbV9tYXAocGV2 LCB0ZXZzKTsNCj4gPj4gICAJCWlmIChyZXQgPiAwKQ0KPiA+PiBAQCAtMjc1OCwzICsyODU5LDIy IEBAIGVuZDoNCj4gPj4gICAJcmV0dXJuIHJldDsNCj4gPj4gICB9DQo+ID4+DQo+ID4+ICtpbnQg Y29weV90b19wcm9iZV90cmFjZV9hcmcoc3RydWN0IHByb2JlX3RyYWNlX2FyZyAqdHZhciwNCj4g Pj4gKwkJCSAgICBzdHJ1Y3QgcGVyZl9wcm9iZV9hcmcgKnB2YXIpDQo+ID4+ICt7DQo+ID4+ICsJ dHZhci0+dmFsdWUgPSBzdHJkdXAocHZhci0+dmFyKTsNCj4gPj4gKwlpZiAodHZhci0+dmFsdWUg PT0gTlVMTCkNCj4gPj4gKwkJcmV0dXJuIC1FTk9NRU07DQo+ID4+ICsJaWYgKHB2YXItPnR5cGUp IHsNCj4gPj4gKwkJdHZhci0+dHlwZSA9IHN0cmR1cChwdmFyLT50eXBlKTsNCj4gPj4gKwkJaWYg KHR2YXItPnR5cGUgPT0gTlVMTCkNCj4gPj4gKwkJCXJldHVybiAtRU5PTUVNOw0KPiA+PiArCX0N Cj4gPj4gKwlpZiAocHZhci0+bmFtZSkgew0KPiA+PiArCQl0dmFyLT5uYW1lID0gc3RyZHVwKHB2 YXItPm5hbWUpOw0KPiA+PiArCQlpZiAodHZhci0+bmFtZSA9PSBOVUxMKQ0KPiA+PiArCQkJcmV0 dXJuIC1FTk9NRU07DQo+ID4+ICsJfSBlbHNlDQo+ID4+ICsJCXR2YXItPm5hbWUgPSBOVUxMOw0K PiA+PiArCXJldHVybiAwOw0KPiA+PiArfQ0KPiA+PiBkaWZmIC0tZ2l0IGEvdG9vbHMvcGVyZi91 dGlsL3Byb2JlLWV2ZW50LmggYi90b29scy9wZXJmL3V0aWwvcHJvYmUtZXZlbnQuaA0KPiA+PiBp bmRleCA4M2VlOTVlLi4xNzRhM2NmIDEwMDY0NA0KPiA+PiAtLS0gYS90b29scy9wZXJmL3V0aWwv cHJvYmUtZXZlbnQuaA0KPiA+PiArKysgYi90b29scy9wZXJmL3V0aWwvcHJvYmUtZXZlbnQuaA0K PiA+PiBAQCAtMTU2LDQgKzE1Niw3IEBAIGludCBlX3NucHJpbnRmKGNoYXIgKnN0ciwgc2l6ZV90 IHNpemUsIGNvbnN0IGNoYXIgKmZvcm1hdCwgLi4uKQ0KPiA+PiAgIC8qIE1heGltdW0gaW5kZXgg bnVtYmVyIG9mIGV2ZW50LW5hbWUgcG9zdGZpeCAqLw0KPiA+PiAgICNkZWZpbmUgTUFYX0VWRU5U X0lOREVYCTEwMjQNCj4gPj4NCj4gPj4gK2ludCBjb3B5X3RvX3Byb2JlX3RyYWNlX2FyZyhzdHJ1 Y3QgcHJvYmVfdHJhY2VfYXJnICp0dmFyLA0KPiA+PiArCQkJICAgIHN0cnVjdCBwZXJmX3Byb2Jl X2FyZyAqcHZhcik7DQo+ID4+ICsNCj4gPj4gICAjZW5kaWYgLypfUFJPQkVfRVZFTlRfSCAqLw0K PiA+PiBkaWZmIC0tZ2l0IGEvdG9vbHMvcGVyZi91dGlsL3Byb2JlLWZpbmRlci5jIGIvdG9vbHMv cGVyZi91dGlsL3Byb2JlLWZpbmRlci5jDQo+ID4+IGluZGV4IDdiODBmOGMuLjI5YzQzYzA2OCAx MDA2NDQNCj4gPj4gLS0tIGEvdG9vbHMvcGVyZi91dGlsL3Byb2JlLWZpbmRlci5jDQo+ID4+ICsr KyBiL3Rvb2xzL3BlcmYvdXRpbC9wcm9iZS1maW5kZXIuYw0KPiA+PiBAQCAtNTUzLDI0ICs1NTMs OSBAQCBzdGF0aWMgaW50IGZpbmRfdmFyaWFibGUoRHdhcmZfRGllICpzY19kaWUsIHN0cnVjdCBw cm9iZV9maW5kZXIgKnBmKQ0KPiA+PiAgIAljaGFyIGJ1ZlszMl0sICpwdHI7DQo+ID4+ICAgCWlu dCByZXQgPSAwOw0KPiA+Pg0KPiA+PiAtCWlmICghaXNfY192YXJuYW1lKHBmLT5wdmFyLT52YXIp KSB7DQo+ID4+IC0JCS8qIENvcHkgcmF3IHBhcmFtZXRlcnMgKi8NCj4gPj4gLQkJcGYtPnR2YXIt PnZhbHVlID0gc3RyZHVwKHBmLT5wdmFyLT52YXIpOw0KPiA+PiAtCQlpZiAocGYtPnR2YXItPnZh bHVlID09IE5VTEwpDQo+ID4+IC0JCQlyZXR1cm4gLUVOT01FTTsNCj4gPj4gLQkJaWYgKHBmLT5w dmFyLT50eXBlKSB7DQo+ID4+IC0JCQlwZi0+dHZhci0+dHlwZSA9IHN0cmR1cChwZi0+cHZhci0+ dHlwZSk7DQo+ID4+IC0JCQlpZiAocGYtPnR2YXItPnR5cGUgPT0gTlVMTCkNCj4gPj4gLQkJCQly ZXR1cm4gLUVOT01FTTsNCj4gPj4gLQkJfQ0KPiA+PiAtCQlpZiAocGYtPnB2YXItPm5hbWUpIHsN Cj4gPj4gLQkJCXBmLT50dmFyLT5uYW1lID0gc3RyZHVwKHBmLT5wdmFyLT5uYW1lKTsNCj4gPj4g LQkJCWlmIChwZi0+dHZhci0+bmFtZSA9PSBOVUxMKQ0KPiA+PiAtCQkJCXJldHVybiAtRU5PTUVN Ow0KPiA+PiAtCQl9IGVsc2UNCj4gPj4gLQkJCXBmLT50dmFyLT5uYW1lID0gTlVMTDsNCj4gPj4g LQkJcmV0dXJuIDA7DQo+ID4+IC0JfQ0KPiA+PiArCS8qIENvcHkgcmF3IHBhcmFtZXRlcnMgKi8N Cj4gPj4gKwlpZiAoIWlzX2NfdmFybmFtZShwZi0+cHZhci0+dmFyKSkNCj4gPj4gKwkJcmV0dXJu IGNvcHlfdG9fcHJvYmVfdHJhY2VfYXJnKHBmLT50dmFyLCBwZi0+cHZhcik7DQo+ID4+DQo+ID4+ ICAgCWlmIChwZi0+cHZhci0+bmFtZSkNCj4gPj4gICAJCXBmLT50dmFyLT5uYW1lID0gc3RyZHVw KHBmLT5wdmFyLT5uYW1lKTsNCj4gPj4gLS0NCj4gPj4gMS44LjMuNA0KPiANCg0K -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | "acme@kernel.org" <acme@kernel.org> |
|---|---|
| Date | 2015-08-26 15:10 +0200 |
| Subject | Re: [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1ICK-6BL-7@gated-at.bofh.it> |
| In reply to | #1213503 |
Em Wed, Aug 26, 2015 at 10:38:18AM +0800, Wangnan (F) escreveu:
>
> On 2015/8/26 8:02, 平松雅巳 / HIRAMATU，MASAMI wrote:
> >>From: Wang Nan [mailto:wangnan0@huawei.com]
> >>
> >>It should be useful to allow 'perf probe' probe at absolute offset of
> >>a target. For example, when (u)probing at a instruction of a shared
> >>object in a embedded system where debuginfo is not avaliable but we
> >>know the offset of that instruction by manually digging.
> >>
> >>This patch enables following perf probe command syntax:
> >>
> >> # perf probe +0xffffffff811e6615
> >>
> >>And
> >>
> >> # perf probe /lib/x86_64-linux-gnu/libc-2.19.so +0xeb860
> >Why do we need "+" for the absolute address?
> >It seems that we can do it if we find that the given probe point
> >starts with "0x".
> >
> >Thanks,
>
> I will change 2/2 as you suggection.
>
> However, we can only ensure that in kernel side symbol never leading
> with '0x'. Although I don't think symbol leading with 0x is useful,
> it is still possible for a userspace program compiled and linked by
> a language other than C produces such symbol. '+' helps us separate
> address and function name semantically, make us don't rely on assumption
> on function names. If in future we do meet '0x' symbols, I think we still
> need the '+' syntax back. But we can do it at that time.
Agreed, I also think that using '+' is better, but will not dwell on
this so as to make progress :-)
- Arnaldo
> Thank you.
>
>
> >>In the above example, we don't need a anchor symbol, so it is possible
> >>to compute absolute addresses using other methods and then use
> >>'perf probe' to create the probing points.
> >>
> >>Signed-off-by: Wang Nan <wangnan0@huawei.com>
> >>Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
> >>Cc: Ingo Molnar <mingo@redhat.com>
> >>Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
> >>Cc: Namhyung Kim <namhyung@kernel.org>
> >>---
> >> tools/perf/util/probe-event.c | 144 +++++++++++++++++++++++++++++++++++++----
> >> tools/perf/util/probe-event.h | 3 +
> >> tools/perf/util/probe-finder.c | 21 +-----
> >> 3 files changed, 138 insertions(+), 30 deletions(-)
> >>
> >>diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
> >>index 6c7e538..59de69a4 100644
> >>--- a/tools/perf/util/probe-event.c
> >>+++ b/tools/perf/util/probe-event.c
> >>@@ -1194,9 +1194,13 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
> >> *ptr++ = '\0';
> >> }
> >>
> >>- tmp = strdup(arg);
> >>- if (tmp == NULL)
> >>- return -ENOMEM;
> >>+ if (arg[0] == '\0')
> >>+ tmp = NULL;
> >>+ else {
> >>+ tmp = strdup(arg);
> >>+ if (tmp == NULL)
> >>+ return -ENOMEM;
> >>+ }
> >>
> >> if (file_spec)
> >> pp->file = tmp;
> >>@@ -1283,11 +1287,6 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
> >> return -EINVAL;
> >> }
> >>
> >>- if (pp->offset && !pp->function) {
> >>- semantic_error("Offset requires an entry function.\n");
> >>- return -EINVAL;
> >>- }
> >>-
> >> if (pp->retprobe && !pp->function) {
> >> semantic_error("Return probe requires an entry function.\n");
> >> return -EINVAL;
> >>@@ -1299,6 +1298,11 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
> >> return -EINVAL;
> >> }
> >>
> >>+ if (!pp->function && !pp->offset && !pp->file) {
> >>+ semantic_error("Absolute address should not be zero.\n");
> >>+ return -EINVAL;
> >>+ }
> >>+
> >> pr_debug("symbol:%s file:%s line:%d offset:%lu return:%d lazy:%s\n",
> >> pp->function, pp->file, pp->line, pp->offset, pp->retprobe,
> >> pp->lazy_line);
> >>@@ -1609,7 +1613,7 @@ error:
> >> static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
> >> {
> >> char *buf, *tmp;
> >>- char offs[32] = "", line[32] = "", file[32] = "";
> >>+ char offs[32] = "", line[32] = "", file[32] = "", addr[32] = "";
> >> int ret, len;
> >>
> >> buf = zalloc(MAX_CMDLEN);
> >>@@ -1622,6 +1626,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
> >> if (ret <= 0)
> >> goto error;
> >> }
> >>+ if (!pp->function) {
> >>+ ret = e_snprintf(addr, 32, "0x%lx", pp->offset);
> >>+ if (ret <= 0)
> >>+ goto error;
> >>+ }
> >> if (pp->line) {
> >> ret = e_snprintf(line, 32, ":%d", pp->line);
> >> if (ret <= 0)
> >>@@ -1639,9 +1648,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
> >> goto error;
> >> }
> >>
> >>- if (pp->function)
> >>- ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s", pp->function,
> >>- offs, pp->retprobe ? "%return" : "", line,
> >>+ if (pp->function || pp->offset)
> >>+ ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s",
> >>+ pp->function ? : addr,
> >>+ pp->function ? offs : "",
> >>+ pp->retprobe ? "%return" : "", line,
> >> file);
> >> else
> >> ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", file, line);
> >>@@ -1786,6 +1797,11 @@ char *synthesize_probe_trace_command(struct probe_trace_event *tev)
> >> if (tev->uprobes)
> >> ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s:0x%lx",
> >> tp->module, tp->address);
> >>+ else if (tp->symbol[0] == '0' && tp->symbol[1] == 'x')
> >>+ /* Absolute address. See try_to_find_absolute_address() */
> >>+ ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s0x%lx",
> >>+ tp->module ?: "", tp->module ? ":" : "",
> >>+ tp->address);
> >> else
> >> ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s%s+%lu",
> >> tp->module ?: "", tp->module ? ":" : "",
> >>@@ -2572,6 +2588,87 @@ err_out:
> >> goto out;
> >> }
> >>
> >>+static int try_to_find_absolute_address(struct perf_probe_event *pev,
> >>+ struct probe_trace_event **tevs)
> >>+{
> >>+ struct perf_probe_point *pp = &pev->point;
> >>+ struct probe_trace_event *tev;
> >>+ struct probe_trace_point *tp;
> >>+ int i, err;
> >>+
> >>+ if (perf_probe_event_need_dwarf(pev) || pev->point.function)
> >>+ return -EINVAL;
> >>+
> >>+ /*
> >>+ * This is 'perf probe /lib/libc.so +0xabcd'. Try to probe at
> >>+ * absolute address.
> >>+ *
> >>+ * Only one tev can be generated by this.
> >>+ */
> >>+ *tevs = zalloc(sizeof(*tev));
> >>+ if (!*tevs)
> >>+ return -ENOMEM;
> >>+
> >>+ tev = *tevs;
> >>+ tp = &tev->point;
> >>+
> >>+ /*
> >>+ * Don't use tp->offset, use address directly, because
> >>+ * in synthesize_probe_trace_command() address cannot be
> >>+ * zero.
> >>+ */
> >>+ tp->address = pev->point.offset;
> >>+ tp->retprobe = pp->retprobe;
> >>+ tev->uprobes = pev->uprobes;
> >>+
> >>+ err = -ENOMEM;
> >>+ /* Give it a '0x' leading symbol name */
> >>+ if (asprintf(&tp->symbol, "0x%lx", tp->address) < 0)
> >>+ goto errout;
> >>+
> >>+ /* For kprobe, check range */
> >>+ if ((!tev->uprobes) &&
> >>+ (kprobe_warn_out_range(tev->point.symbol,
> >>+ tev->point.address))) {
> >>+ err = -EACCES;
> >>+ goto errout;
> >>+ }
> >>+
> >>+ if (asprintf(&tp->realname, "abs_%lx", tp->address) < 0)
> >>+ goto errout;
> >>+
> >>+ if (pev->target) {
> >>+ tp->module = strdup(pev->target);
> >>+ if (!tp->module)
> >>+ goto errout;
> >>+ }
> >>+
> >>+ if (tev->group) {
> >>+ tev->group = strdup(pev->group);
> >>+ if (!tev->group)
> >>+ goto errout;
> >>+ }
> >>+
> >>+ if (pev->event) {
> >>+ tev->event = strdup(pev->event);
> >>+ if (!tev->event)
> >>+ goto errout;
> >>+ }
> >>+
> >>+ tev->nargs = pev->nargs;
> >>+ for (i = 0; i < tev->nargs; i++)
> >>+ copy_to_probe_trace_arg(&tev->args[i], &pev->args[i]);
> >>+
> >>+ return 1;
> >>+
> >>+errout:
> >>+ if (*tevs) {
> >>+ clear_probe_trace_events(*tevs, 1);
> >>+ *tevs = NULL;
> >>+ }
> >>+ return err;
> >>+}
> >>+
> >> bool __weak arch__prefers_symtab(void) { return false; }
> >>
> >> static int convert_to_probe_trace_events(struct perf_probe_event *pev,
> >>@@ -2588,6 +2685,10 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
> >> }
> >> }
> >>
> >>+ ret = try_to_find_absolute_address(pev, tevs);
> >>+ if (ret > 0)
> >>+ return ret;
> >>+
> >> if (arch__prefers_symtab() && !perf_probe_event_need_dwarf(pev)) {
> >> ret = find_probe_trace_events_from_map(pev, tevs);
> >> if (ret > 0)
> >>@@ -2758,3 +2859,22 @@ end:
> >> return ret;
> >> }
> >>
> >>+int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
> >>+ struct perf_probe_arg *pvar)
> >>+{
> >>+ tvar->value = strdup(pvar->var);
> >>+ if (tvar->value == NULL)
> >>+ return -ENOMEM;
> >>+ if (pvar->type) {
> >>+ tvar->type = strdup(pvar->type);
> >>+ if (tvar->type == NULL)
> >>+ return -ENOMEM;
> >>+ }
> >>+ if (pvar->name) {
> >>+ tvar->name = strdup(pvar->name);
> >>+ if (tvar->name == NULL)
> >>+ return -ENOMEM;
> >>+ } else
> >>+ tvar->name = NULL;
> >>+ return 0;
> >>+}
> >>diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
> >>index 83ee95e..174a3cf 100644
> >>--- a/tools/perf/util/probe-event.h
> >>+++ b/tools/perf/util/probe-event.h
> >>@@ -156,4 +156,7 @@ int e_snprintf(char *str, size_t size, const char *format, ...)
> >> /* Maximum index number of event-name postfix */
> >> #define MAX_EVENT_INDEX 1024
> >>
> >>+int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
> >>+ struct perf_probe_arg *pvar);
> >>+
> >> #endif /*_PROBE_EVENT_H */
> >>diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
> >>index 7b80f8c..29c43c068 100644
> >>--- a/tools/perf/util/probe-finder.c
> >>+++ b/tools/perf/util/probe-finder.c
> >>@@ -553,24 +553,9 @@ static int find_variable(Dwarf_Die *sc_die, struct probe_finder *pf)
> >> char buf[32], *ptr;
> >> int ret = 0;
> >>
> >>- if (!is_c_varname(pf->pvar->var)) {
> >>- /* Copy raw parameters */
> >>- pf->tvar->value = strdup(pf->pvar->var);
> >>- if (pf->tvar->value == NULL)
> >>- return -ENOMEM;
> >>- if (pf->pvar->type) {
> >>- pf->tvar->type = strdup(pf->pvar->type);
> >>- if (pf->tvar->type == NULL)
> >>- return -ENOMEM;
> >>- }
> >>- if (pf->pvar->name) {
> >>- pf->tvar->name = strdup(pf->pvar->name);
> >>- if (pf->tvar->name == NULL)
> >>- return -ENOMEM;
> >>- } else
> >>- pf->tvar->name = NULL;
> >>- return 0;
> >>- }
> >>+ /* Copy raw parameters */
> >>+ if (!is_c_varname(pf->pvar->var))
> >>+ return copy_to_probe_trace_arg(pf->tvar, pf->pvar);
> >>
> >> if (pf->pvar->name)
> >> pf->tvar->name = strdup(pf->pvar->name);
> >>--
> >>1.8.3.4
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | "Wangnan (F)" <wangnan0@huawei.com> |
|---|---|
| Date | 2015-08-26 15:20 +0200 |
| Subject | Re: [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1IMp-6N8-3@gated-at.bofh.it> |
| In reply to | #1213838 |
On 2015/8/26 21:02, acme@kernel.org wrote:
> Em Wed, Aug 26, 2015 at 10:38:18AM +0800, Wangnan (F) escreveu:
>> On 2015/8/26 8:02, 平松雅巳 / HIRAMATU，MASAMI wrote:
>>>> From: Wang Nan [mailto:wangnan0@huawei.com]
>>>>
>>>> It should be useful to allow 'perf probe' probe at absolute offset of
>>>> a target. For example, when (u)probing at a instruction of a shared
>>>> object in a embedded system where debuginfo is not avaliable but we
>>>> know the offset of that instruction by manually digging.
>>>>
>>>> This patch enables following perf probe command syntax:
>>>>
>>>> # perf probe +0xffffffff811e6615
>>>>
>>>> And
>>>>
>>>> # perf probe /lib/x86_64-linux-gnu/libc-2.19.so +0xeb860
>>> Why do we need "+" for the absolute address?
>>> It seems that we can do it if we find that the given probe point
>>> starts with "0x".
>>>
>>> Thanks,
>> I will change 2/2 as you suggection.
>>
>> However, we can only ensure that in kernel side symbol never leading
>> with '0x'. Although I don't think symbol leading with 0x is useful,
>> it is still possible for a userspace program compiled and linked by
>> a language other than C produces such symbol. '+' helps us separate
>> address and function name semantically, make us don't rely on assumption
>> on function names. If in future we do meet '0x' symbols, I think we still
>> need the '+' syntax back. But we can do it at that time.
> Agreed, I also think that using '+' is better, but will not dwell on
> this so as to make progress :-)
>
> - Arnaldo
>
Maybe we should support both of them, making '+0x1234' the core
syntax, and '0x1234' style as a syntax sugar. However I have worked
on this problem for nearly a full day but my main work should be BPF
related things...
Since Masami has acked all of the 6 v3 patches, if we still need '+' I can
bring it back with a new patch when I have time. However, same to you,
I don't think this should be a blocking problem.
Thank you.
>> Thank you.
>>
>>
>>>> In the above example, we don't need a anchor symbol, so it is possible
>>>> to compute absolute addresses using other methods and then use
>>>> 'perf probe' to create the probing points.
>>>>
>>>> Signed-off-by: Wang Nan <wangnan0@huawei.com>
>>>> Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
>>>> Cc: Ingo Molnar <mingo@redhat.com>
>>>> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
>>>> Cc: Namhyung Kim <namhyung@kernel.org>
>>>> ---
>>>> tools/perf/util/probe-event.c | 144 +++++++++++++++++++++++++++++++++++++----
>>>> tools/perf/util/probe-event.h | 3 +
>>>> tools/perf/util/probe-finder.c | 21 +-----
>>>> 3 files changed, 138 insertions(+), 30 deletions(-)
>>>>
>>>> diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
>>>> index 6c7e538..59de69a4 100644
>>>> --- a/tools/perf/util/probe-event.c
>>>> +++ b/tools/perf/util/probe-event.c
>>>> @@ -1194,9 +1194,13 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
>>>> *ptr++ = '\0';
>>>> }
>>>>
>>>> - tmp = strdup(arg);
>>>> - if (tmp == NULL)
>>>> - return -ENOMEM;
>>>> + if (arg[0] == '\0')
>>>> + tmp = NULL;
>>>> + else {
>>>> + tmp = strdup(arg);
>>>> + if (tmp == NULL)
>>>> + return -ENOMEM;
>>>> + }
>>>>
>>>> if (file_spec)
>>>> pp->file = tmp;
>>>> @@ -1283,11 +1287,6 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
>>>> return -EINVAL;
>>>> }
>>>>
>>>> - if (pp->offset && !pp->function) {
>>>> - semantic_error("Offset requires an entry function.\n");
>>>> - return -EINVAL;
>>>> - }
>>>> -
>>>> if (pp->retprobe && !pp->function) {
>>>> semantic_error("Return probe requires an entry function.\n");
>>>> return -EINVAL;
>>>> @@ -1299,6 +1298,11 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
>>>> return -EINVAL;
>>>> }
>>>>
>>>> + if (!pp->function && !pp->offset && !pp->file) {
>>>> + semantic_error("Absolute address should not be zero.\n");
>>>> + return -EINVAL;
>>>> + }
>>>> +
>>>> pr_debug("symbol:%s file:%s line:%d offset:%lu return:%d lazy:%s\n",
>>>> pp->function, pp->file, pp->line, pp->offset, pp->retprobe,
>>>> pp->lazy_line);
>>>> @@ -1609,7 +1613,7 @@ error:
>>>> static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
>>>> {
>>>> char *buf, *tmp;
>>>> - char offs[32] = "", line[32] = "", file[32] = "";
>>>> + char offs[32] = "", line[32] = "", file[32] = "", addr[32] = "";
>>>> int ret, len;
>>>>
>>>> buf = zalloc(MAX_CMDLEN);
>>>> @@ -1622,6 +1626,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
>>>> if (ret <= 0)
>>>> goto error;
>>>> }
>>>> + if (!pp->function) {
>>>> + ret = e_snprintf(addr, 32, "0x%lx", pp->offset);
>>>> + if (ret <= 0)
>>>> + goto error;
>>>> + }
>>>> if (pp->line) {
>>>> ret = e_snprintf(line, 32, ":%d", pp->line);
>>>> if (ret <= 0)
>>>> @@ -1639,9 +1648,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
>>>> goto error;
>>>> }
>>>>
>>>> - if (pp->function)
>>>> - ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s", pp->function,
>>>> - offs, pp->retprobe ? "%return" : "", line,
>>>> + if (pp->function || pp->offset)
>>>> + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s",
>>>> + pp->function ? : addr,
>>>> + pp->function ? offs : "",
>>>> + pp->retprobe ? "%return" : "", line,
>>>> file);
>>>> else
>>>> ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", file, line);
>>>> @@ -1786,6 +1797,11 @@ char *synthesize_probe_trace_command(struct probe_trace_event *tev)
>>>> if (tev->uprobes)
>>>> ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s:0x%lx",
>>>> tp->module, tp->address);
>>>> + else if (tp->symbol[0] == '0' && tp->symbol[1] == 'x')
>>>> + /* Absolute address. See try_to_find_absolute_address() */
>>>> + ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s0x%lx",
>>>> + tp->module ?: "", tp->module ? ":" : "",
>>>> + tp->address);
>>>> else
>>>> ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s%s+%lu",
>>>> tp->module ?: "", tp->module ? ":" : "",
>>>> @@ -2572,6 +2588,87 @@ err_out:
>>>> goto out;
>>>> }
>>>>
>>>> +static int try_to_find_absolute_address(struct perf_probe_event *pev,
>>>> + struct probe_trace_event **tevs)
>>>> +{
>>>> + struct perf_probe_point *pp = &pev->point;
>>>> + struct probe_trace_event *tev;
>>>> + struct probe_trace_point *tp;
>>>> + int i, err;
>>>> +
>>>> + if (perf_probe_event_need_dwarf(pev) || pev->point.function)
>>>> + return -EINVAL;
>>>> +
>>>> + /*
>>>> + * This is 'perf probe /lib/libc.so +0xabcd'. Try to probe at
>>>> + * absolute address.
>>>> + *
>>>> + * Only one tev can be generated by this.
>>>> + */
>>>> + *tevs = zalloc(sizeof(*tev));
>>>> + if (!*tevs)
>>>> + return -ENOMEM;
>>>> +
>>>> + tev = *tevs;
>>>> + tp = &tev->point;
>>>> +
>>>> + /*
>>>> + * Don't use tp->offset, use address directly, because
>>>> + * in synthesize_probe_trace_command() address cannot be
>>>> + * zero.
>>>> + */
>>>> + tp->address = pev->point.offset;
>>>> + tp->retprobe = pp->retprobe;
>>>> + tev->uprobes = pev->uprobes;
>>>> +
>>>> + err = -ENOMEM;
>>>> + /* Give it a '0x' leading symbol name */
>>>> + if (asprintf(&tp->symbol, "0x%lx", tp->address) < 0)
>>>> + goto errout;
>>>> +
>>>> + /* For kprobe, check range */
>>>> + if ((!tev->uprobes) &&
>>>> + (kprobe_warn_out_range(tev->point.symbol,
>>>> + tev->point.address))) {
>>>> + err = -EACCES;
>>>> + goto errout;
>>>> + }
>>>> +
>>>> + if (asprintf(&tp->realname, "abs_%lx", tp->address) < 0)
>>>> + goto errout;
>>>> +
>>>> + if (pev->target) {
>>>> + tp->module = strdup(pev->target);
>>>> + if (!tp->module)
>>>> + goto errout;
>>>> + }
>>>> +
>>>> + if (tev->group) {
>>>> + tev->group = strdup(pev->group);
>>>> + if (!tev->group)
>>>> + goto errout;
>>>> + }
>>>> +
>>>> + if (pev->event) {
>>>> + tev->event = strdup(pev->event);
>>>> + if (!tev->event)
>>>> + goto errout;
>>>> + }
>>>> +
>>>> + tev->nargs = pev->nargs;
>>>> + for (i = 0; i < tev->nargs; i++)
>>>> + copy_to_probe_trace_arg(&tev->args[i], &pev->args[i]);
>>>> +
>>>> + return 1;
>>>> +
>>>> +errout:
>>>> + if (*tevs) {
>>>> + clear_probe_trace_events(*tevs, 1);
>>>> + *tevs = NULL;
>>>> + }
>>>> + return err;
>>>> +}
>>>> +
>>>> bool __weak arch__prefers_symtab(void) { return false; }
>>>>
>>>> static int convert_to_probe_trace_events(struct perf_probe_event *pev,
>>>> @@ -2588,6 +2685,10 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
>>>> }
>>>> }
>>>>
>>>> + ret = try_to_find_absolute_address(pev, tevs);
>>>> + if (ret > 0)
>>>> + return ret;
>>>> +
>>>> if (arch__prefers_symtab() && !perf_probe_event_need_dwarf(pev)) {
>>>> ret = find_probe_trace_events_from_map(pev, tevs);
>>>> if (ret > 0)
>>>> @@ -2758,3 +2859,22 @@ end:
>>>> return ret;
>>>> }
>>>>
>>>> +int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
>>>> + struct perf_probe_arg *pvar)
>>>> +{
>>>> + tvar->value = strdup(pvar->var);
>>>> + if (tvar->value == NULL)
>>>> + return -ENOMEM;
>>>> + if (pvar->type) {
>>>> + tvar->type = strdup(pvar->type);
>>>> + if (tvar->type == NULL)
>>>> + return -ENOMEM;
>>>> + }
>>>> + if (pvar->name) {
>>>> + tvar->name = strdup(pvar->name);
>>>> + if (tvar->name == NULL)
>>>> + return -ENOMEM;
>>>> + } else
>>>> + tvar->name = NULL;
>>>> + return 0;
>>>> +}
>>>> diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
>>>> index 83ee95e..174a3cf 100644
>>>> --- a/tools/perf/util/probe-event.h
>>>> +++ b/tools/perf/util/probe-event.h
>>>> @@ -156,4 +156,7 @@ int e_snprintf(char *str, size_t size, const char *format, ...)
>>>> /* Maximum index number of event-name postfix */
>>>> #define MAX_EVENT_INDEX 1024
>>>>
>>>> +int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
>>>> + struct perf_probe_arg *pvar);
>>>> +
>>>> #endif /*_PROBE_EVENT_H */
>>>> diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
>>>> index 7b80f8c..29c43c068 100644
>>>> --- a/tools/perf/util/probe-finder.c
>>>> +++ b/tools/perf/util/probe-finder.c
>>>> @@ -553,24 +553,9 @@ static int find_variable(Dwarf_Die *sc_die, struct probe_finder *pf)
>>>> char buf[32], *ptr;
>>>> int ret = 0;
>>>>
>>>> - if (!is_c_varname(pf->pvar->var)) {
>>>> - /* Copy raw parameters */
>>>> - pf->tvar->value = strdup(pf->pvar->var);
>>>> - if (pf->tvar->value == NULL)
>>>> - return -ENOMEM;
>>>> - if (pf->pvar->type) {
>>>> - pf->tvar->type = strdup(pf->pvar->type);
>>>> - if (pf->tvar->type == NULL)
>>>> - return -ENOMEM;
>>>> - }
>>>> - if (pf->pvar->name) {
>>>> - pf->tvar->name = strdup(pf->pvar->name);
>>>> - if (pf->tvar->name == NULL)
>>>> - return -ENOMEM;
>>>> - } else
>>>> - pf->tvar->name = NULL;
>>>> - return 0;
>>>> - }
>>>> + /* Copy raw parameters */
>>>> + if (!is_c_varname(pf->pvar->var))
>>>> + return copy_to_probe_trace_arg(pf->tvar, pf->pvar);
>>>>
>>>> if (pf->pvar->name)
>>>> pf->tvar->name = strdup(pf->pvar->name);
>>>> --
>>>> 1.8.3.4
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Arnaldo Carvalho de Melo <acme@kernel.org> |
|---|---|
| Date | 2015-08-26 15:40 +0200 |
| Subject | Re: [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1J5M-79A-13@gated-at.bofh.it> |
| In reply to | #1213846 |
Em Wed, Aug 26, 2015 at 09:19:27PM +0800, Wangnan (F) escreveu: > On 2015/8/26 21:02, acme@kernel.org wrote: > >Em Wed, Aug 26, 2015 at 10:38:18AM +0800, Wangnan (F) escreveu: > >>On 2015/8/26 8:02, 平松雅巳 / HIRAMATU，MASAMI wrote: > >>>>From: Wang Nan [mailto:wangnan0@huawei.com] > >>>> # perf probe /lib/x86_64-linux-gnu/libc-2.19.so +0xeb860 > >>>Why do we need "+" for the absolute address? > >>>It seems that we can do it if we find that the given probe point > >>>starts with "0x". > >>I will change 2/2 as you suggestion. > >>However, we can only ensure that in kernel side symbol never leading > >>with '0x'. Although I don't think symbol leading with 0x is useful, > >>it is still possible for a userspace program compiled and linked by > >>a language other than C produces such symbol. '+' helps us separate > >>address and function name semantically, make us don't rely on assumption > >>on function names. If in future we do meet '0x' symbols, I think we still > >>need the '+' syntax back. But we can do it at that time. > >Agreed, I also think that using '+' is better, but will not dwell on > >this so as to make progress :-) > Maybe we should support both of them, making '+0x1234' the core > syntax, and '0x1234' style as a syntax sugar. However I have worked > on this problem for nearly a full day but my main work should be BPF > related things... > Since Masami has acked all of the 6 v3 patches, if we still need '+' I can > bring it back with a new patch when I have time. However, same to you, > I don't think this should be a blocking problem. Agreed, we don't have to agree on everything all the time, lets keep Masami happy this time :-) - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | "acme@kernel.org" <acme@kernel.org> |
|---|---|
| Date | 2015-08-26 15:10 +0200 |
| Subject | Re: [PATCH 2/2] perf probe: Support probing at absolute address |
| Message-ID | <q1ICK-6BL-29@gated-at.bofh.it> |
| In reply to | #1213451 |
Em Wed, Aug 26, 2015 at 12:02:40AM +0000, 平松雅巳 / HIRAMATU，MASAMI escreveu:
> > From: Wang Nan [mailto:wangnan0@huawei.com]
> >
> > It should be useful to allow 'perf probe' probe at absolute offset of
> > a target. For example, when (u)probing at a instruction of a shared
> > object in a embedded system where debuginfo is not avaliable but we
> > know the offset of that instruction by manually digging.
> >
> > This patch enables following perf probe command syntax:
> >
> > # perf probe +0xffffffff811e6615
> >
> > And
> >
> > # perf probe /lib/x86_64-linux-gnu/libc-2.19.so +0xeb860
>
> Why do we need "+" for the absolute address?
I guess 'absolute' was the misleading term here, 'offset' would be
better, perhaps, or even 'raw' address.
> It seems that we can do it if we find that the given probe point
> starts with "0x".
Which is kinda ok, since that is the most common way of expressing
offsets, but i don't know why we would preclude using, say, decimal,
octal, etc.
- Arnaldo
> Thanks,
>
> >
> > In the above example, we don't need a anchor symbol, so it is possible
> > to compute absolute addresses using other methods and then use
> > 'perf probe' to create the probing points.
> >
> > Signed-off-by: Wang Nan <wangnan0@huawei.com>
> > Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
> > Cc: Ingo Molnar <mingo@redhat.com>
> > Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
> > Cc: Namhyung Kim <namhyung@kernel.org>
> > ---
> > tools/perf/util/probe-event.c | 144 +++++++++++++++++++++++++++++++++++++----
> > tools/perf/util/probe-event.h | 3 +
> > tools/perf/util/probe-finder.c | 21 +-----
> > 3 files changed, 138 insertions(+), 30 deletions(-)
> >
> > diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
> > index 6c7e538..59de69a4 100644
> > --- a/tools/perf/util/probe-event.c
> > +++ b/tools/perf/util/probe-event.c
> > @@ -1194,9 +1194,13 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
> > *ptr++ = '\0';
> > }
> >
> > - tmp = strdup(arg);
> > - if (tmp == NULL)
> > - return -ENOMEM;
> > + if (arg[0] == '\0')
> > + tmp = NULL;
> > + else {
> > + tmp = strdup(arg);
> > + if (tmp == NULL)
> > + return -ENOMEM;
> > + }
> >
> > if (file_spec)
> > pp->file = tmp;
> > @@ -1283,11 +1287,6 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
> > return -EINVAL;
> > }
> >
> > - if (pp->offset && !pp->function) {
> > - semantic_error("Offset requires an entry function.\n");
> > - return -EINVAL;
> > - }
> > -
> > if (pp->retprobe && !pp->function) {
> > semantic_error("Return probe requires an entry function.\n");
> > return -EINVAL;
> > @@ -1299,6 +1298,11 @@ static int parse_perf_probe_point(char *arg, struct perf_probe_event *pev)
> > return -EINVAL;
> > }
> >
> > + if (!pp->function && !pp->offset && !pp->file) {
> > + semantic_error("Absolute address should not be zero.\n");
> > + return -EINVAL;
> > + }
> > +
> > pr_debug("symbol:%s file:%s line:%d offset:%lu return:%d lazy:%s\n",
> > pp->function, pp->file, pp->line, pp->offset, pp->retprobe,
> > pp->lazy_line);
> > @@ -1609,7 +1613,7 @@ error:
> > static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
> > {
> > char *buf, *tmp;
> > - char offs[32] = "", line[32] = "", file[32] = "";
> > + char offs[32] = "", line[32] = "", file[32] = "", addr[32] = "";
> > int ret, len;
> >
> > buf = zalloc(MAX_CMDLEN);
> > @@ -1622,6 +1626,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
> > if (ret <= 0)
> > goto error;
> > }
> > + if (!pp->function) {
> > + ret = e_snprintf(addr, 32, "0x%lx", pp->offset);
> > + if (ret <= 0)
> > + goto error;
> > + }
> > if (pp->line) {
> > ret = e_snprintf(line, 32, ":%d", pp->line);
> > if (ret <= 0)
> > @@ -1639,9 +1648,11 @@ static char *synthesize_perf_probe_point(struct perf_probe_point *pp)
> > goto error;
> > }
> >
> > - if (pp->function)
> > - ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s", pp->function,
> > - offs, pp->retprobe ? "%return" : "", line,
> > + if (pp->function || pp->offset)
> > + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s",
> > + pp->function ? : addr,
> > + pp->function ? offs : "",
> > + pp->retprobe ? "%return" : "", line,
> > file);
> > else
> > ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", file, line);
> > @@ -1786,6 +1797,11 @@ char *synthesize_probe_trace_command(struct probe_trace_event *tev)
> > if (tev->uprobes)
> > ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s:0x%lx",
> > tp->module, tp->address);
> > + else if (tp->symbol[0] == '0' && tp->symbol[1] == 'x')
> > + /* Absolute address. See try_to_find_absolute_address() */
> > + ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s0x%lx",
> > + tp->module ?: "", tp->module ? ":" : "",
> > + tp->address);
> > else
> > ret = e_snprintf(buf + len, MAX_CMDLEN - len, "%s%s%s+%lu",
> > tp->module ?: "", tp->module ? ":" : "",
> > @@ -2572,6 +2588,87 @@ err_out:
> > goto out;
> > }
> >
> > +static int try_to_find_absolute_address(struct perf_probe_event *pev,
> > + struct probe_trace_event **tevs)
> > +{
> > + struct perf_probe_point *pp = &pev->point;
> > + struct probe_trace_event *tev;
> > + struct probe_trace_point *tp;
> > + int i, err;
> > +
> > + if (perf_probe_event_need_dwarf(pev) || pev->point.function)
> > + return -EINVAL;
> > +
> > + /*
> > + * This is 'perf probe /lib/libc.so +0xabcd'. Try to probe at
> > + * absolute address.
> > + *
> > + * Only one tev can be generated by this.
> > + */
> > + *tevs = zalloc(sizeof(*tev));
> > + if (!*tevs)
> > + return -ENOMEM;
> > +
> > + tev = *tevs;
> > + tp = &tev->point;
> > +
> > + /*
> > + * Don't use tp->offset, use address directly, because
> > + * in synthesize_probe_trace_command() address cannot be
> > + * zero.
> > + */
> > + tp->address = pev->point.offset;
> > + tp->retprobe = pp->retprobe;
> > + tev->uprobes = pev->uprobes;
> > +
> > + err = -ENOMEM;
> > + /* Give it a '0x' leading symbol name */
> > + if (asprintf(&tp->symbol, "0x%lx", tp->address) < 0)
> > + goto errout;
> > +
> > + /* For kprobe, check range */
> > + if ((!tev->uprobes) &&
> > + (kprobe_warn_out_range(tev->point.symbol,
> > + tev->point.address))) {
> > + err = -EACCES;
> > + goto errout;
> > + }
> > +
> > + if (asprintf(&tp->realname, "abs_%lx", tp->address) < 0)
> > + goto errout;
> > +
> > + if (pev->target) {
> > + tp->module = strdup(pev->target);
> > + if (!tp->module)
> > + goto errout;
> > + }
> > +
> > + if (tev->group) {
> > + tev->group = strdup(pev->group);
> > + if (!tev->group)
> > + goto errout;
> > + }
> > +
> > + if (pev->event) {
> > + tev->event = strdup(pev->event);
> > + if (!tev->event)
> > + goto errout;
> > + }
> > +
> > + tev->nargs = pev->nargs;
> > + for (i = 0; i < tev->nargs; i++)
> > + copy_to_probe_trace_arg(&tev->args[i], &pev->args[i]);
> > +
> > + return 1;
> > +
> > +errout:
> > + if (*tevs) {
> > + clear_probe_trace_events(*tevs, 1);
> > + *tevs = NULL;
> > + }
> > + return err;
> > +}
> > +
> > bool __weak arch__prefers_symtab(void) { return false; }
> >
> > static int convert_to_probe_trace_events(struct perf_probe_event *pev,
> > @@ -2588,6 +2685,10 @@ static int convert_to_probe_trace_events(struct perf_probe_event *pev,
> > }
> > }
> >
> > + ret = try_to_find_absolute_address(pev, tevs);
> > + if (ret > 0)
> > + return ret;
> > +
> > if (arch__prefers_symtab() && !perf_probe_event_need_dwarf(pev)) {
> > ret = find_probe_trace_events_from_map(pev, tevs);
> > if (ret > 0)
> > @@ -2758,3 +2859,22 @@ end:
> > return ret;
> > }
> >
> > +int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
> > + struct perf_probe_arg *pvar)
> > +{
> > + tvar->value = strdup(pvar->var);
> > + if (tvar->value == NULL)
> > + return -ENOMEM;
> > + if (pvar->type) {
> > + tvar->type = strdup(pvar->type);
> > + if (tvar->type == NULL)
> > + return -ENOMEM;
> > + }
> > + if (pvar->name) {
> > + tvar->name = strdup(pvar->name);
> > + if (tvar->name == NULL)
> > + return -ENOMEM;
> > + } else
> > + tvar->name = NULL;
> > + return 0;
> > +}
> > diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h
> > index 83ee95e..174a3cf 100644
> > --- a/tools/perf/util/probe-event.h
> > +++ b/tools/perf/util/probe-event.h
> > @@ -156,4 +156,7 @@ int e_snprintf(char *str, size_t size, const char *format, ...)
> > /* Maximum index number of event-name postfix */
> > #define MAX_EVENT_INDEX 1024
> >
> > +int copy_to_probe_trace_arg(struct probe_trace_arg *tvar,
> > + struct perf_probe_arg *pvar);
> > +
> > #endif /*_PROBE_EVENT_H */
> > diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
> > index 7b80f8c..29c43c068 100644
> > --- a/tools/perf/util/probe-finder.c
> > +++ b/tools/perf/util/probe-finder.c
> > @@ -553,24 +553,9 @@ static int find_variable(Dwarf_Die *sc_die, struct probe_finder *pf)
> > char buf[32], *ptr;
> > int ret = 0;
> >
> > - if (!is_c_varname(pf->pvar->var)) {
> > - /* Copy raw parameters */
> > - pf->tvar->value = strdup(pf->pvar->var);
> > - if (pf->tvar->value == NULL)
> > - return -ENOMEM;
> > - if (pf->pvar->type) {
> > - pf->tvar->type = strdup(pf->pvar->type);
> > - if (pf->tvar->type == NULL)
> > - return -ENOMEM;
> > - }
> > - if (pf->pvar->name) {
> > - pf->tvar->name = strdup(pf->pvar->name);
> > - if (pf->tvar->name == NULL)
> > - return -ENOMEM;
> > - } else
> > - pf->tvar->name = NULL;
> > - return 0;
> > - }
> > + /* Copy raw parameters */
> > + if (!is_c_varname(pf->pvar->var))
> > + return copy_to_probe_trace_arg(pf->tvar, pf->pvar);
> >
> > if (pf->pvar->name)
> > pf->tvar->name = strdup(pf->pvar->name);
> > --
> > 1.8.3.4
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | Arnaldo Carvalho de Melo <acme@kernel.org> |
|---|---|
| Date | 2015-08-25 17:40 +0200 |
| Subject | Re: [PATCH 1/2] perf probe: Prevent segfault when reading probe point with absolute address |
| Message-ID | <q1oum-2cl-29@gated-at.bofh.it> |
| In reply to | #1213038 |
Em Tue, Aug 25, 2015 at 01:27:35PM +0000, Wang Nan escreveu: > 'perf probe -l' panic if there is a manually inserted probing point > with absolute address. For example: Thanks, tested and applied. - Arnaldo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | 平松雅巳 / HIRAMATU，MASAMI <masami.hiramatsu.pt@hitachi.com> |
|---|---|
| Date | 2015-08-26 02:10 +0200 |
| Subject | RE: [PATCH 1/2] perf probe: Prevent segfault when reading probe point with absolute address |
| Message-ID | <q1wrU-5N8-7@gated-at.bofh.it> |
| In reply to | #1213038 |
PiBGcm9tOiBXYW5nIE5hbiBbbWFpbHRvOndhbmduYW4wQGh1YXdlaS5jb21dDQo+IA0KPiAncGVy ZiBwcm9iZSAtbCcgcGFuaWMgaWYgdGhlcmUgaXMgYSBtYW51YWxseSBpbnNlcnRlZCBwcm9iaW5n IHBvaW50DQo+IHdpdGggYWJzb2x1dGUgYWRkcmVzcy4gRm9yIGV4YW1wbGU6DQo+IA0KPiAgIyBl Y2hvICdwOnByb2JlL2Fic19mZmZmZmZmZjgxMWU2NjE1IDB4ZmZmZmZmZmY4MTFlNjYxNScgPiAv c3lzL2tlcm5lbC9kZWJ1Zy90cmFjaW5nL2twcm9iZV9ldmVudHMNCj4gICMgcGVyZiBwcm9iZSAt bA0KPiAgU2VnbWVudGF0aW9uIGZhdWx0IChjb3JlIGR1bXBlZCkNCj4gDQo+IFRoaXMgcGF0Y2gg Zml4IHRoaXMgcHJvYmxlbSBieSBjb25zaWRlcmluZyB0aGUgc2l0dWF0aW9uIHRoYXQNCj4gInRw LT5zeW1ib2wgPT0gTlVMTCIgaW4gZmluZF9wZXJmX3Byb2JlX3BvaW50X2Zyb21fZHdhcmYoKSBh bmQNCj4gZmluZF9wZXJmX3Byb2JlX3BvaW50X2Zyb21fbWFwKCkuDQoNCk9vcHMsIE9LIHRoaXMg aXMgYSBidWcuDQoNClRoYW5rcyENCg0KDQpBY2tlZC1ieTogTWFzYW1pIEhpcmFtYXRzdSA8bWFz YW1pLmhpcmFtYXRzdS5wdEBoaXRhY2hpLmNvbT4NCg0KPiANCj4gQWZ0ZXIgdGhpcyBwYXRjaDoN Cj4gDQo+ICAjIHBlcmYgcHJvYmUgLWwNCj4gIHByb2JlOmFic19mZmZmZmZmZjgxMWU2NjE1IChv biBTeVNfd3JpdGUrNUBmcy9yZWFkX3dyaXRlLmMpDQo+IA0KPiBBbmQgd2hlbiBkZWJ1ZyBpbmZv IGlzIG1pc3Npbmc6DQo+IA0KPiAgIyBybSAtcmYgfi8uZGVidWcNCj4gICMgbXYgL2xpYi9tb2R1 bGVzLzQuMi4wLXJjMSsvYnVpbGQvdm1saW51eCAvbGliL21vZHVsZXMvNC4yLjAtcmMxKy9idWls ZC92bWxpbnV4LmJhaw0KPiAgIyBwZXJmIHByb2JlIC1sDQo+ICBwcm9iZTphYnNfZmZmZmZmZmY4 MTFlNjYxNSAob24gc3lzX3dyaXRlKzUpDQo+IA0KPiBTaWduZWQtb2ZmLWJ5OiBXYW5nIE5hbiA8 d2FuZ25hbjBAaHVhd2VpLmNvbT4NCj4gQ2M6IEFybmFsZG8gQ2FydmFsaG8gZGUgTWVsbyA8YWNt ZUBrZXJuZWwub3JnPg0KPiBDYzogSW5nbyBNb2xuYXIgPG1pbmdvQHJlZGhhdC5jb20+DQo+IENj OiBNYXNhbWkgSGlyYW1hdHN1IDxtYXNhbWkuaGlyYW1hdHN1LnB0QGhpdGFjaGkuY29tPg0KPiBD YzogTmFtaHl1bmcgS2ltIDxuYW1oeXVuZ0BrZXJuZWwub3JnPg0KPiAtLS0NCj4gIHRvb2xzL3Bl cmYvdXRpbC9wcm9iZS1ldmVudC5jIHwgOCArKysrLS0tLQ0KPiAgMSBmaWxlIGNoYW5nZWQsIDQg aW5zZXJ0aW9ucygrKSwgNCBkZWxldGlvbnMoLSkNCj4gDQo+IGRpZmYgLS1naXQgYS90b29scy9w ZXJmL3V0aWwvcHJvYmUtZXZlbnQuYyBiL3Rvb2xzL3BlcmYvdXRpbC9wcm9iZS1ldmVudC5jDQo+ IGluZGV4IGYwNzM3NGIuLjZjN2U1MzggMTAwNjQ0DQo+IC0tLSBhL3Rvb2xzL3BlcmYvdXRpbC9w cm9iZS1ldmVudC5jDQo+ICsrKyBiL3Rvb2xzL3BlcmYvdXRpbC9wcm9iZS1ldmVudC5jDQo+IEBA IC01MTUsNyArNTE1LDcgQEAgc3RhdGljIGludCBmaW5kX3BlcmZfcHJvYmVfcG9pbnRfZnJvbV9k d2FyZihzdHJ1Y3QgcHJvYmVfdHJhY2VfcG9pbnQgKnRwLA0KPiAgCQlpZiAocmV0IDwgMCkNCj4g IAkJCWdvdG8gZXJyb3I7DQo+ICAJCWFkZHIgKz0gc3RleHQ7DQo+IC0JfSBlbHNlIHsNCj4gKwl9 IGVsc2UgaWYgKHRwLT5zeW1ib2wpIHsNCj4gIAkJYWRkciA9IGtlcm5lbF9nZXRfc3ltYm9sX2Fk ZHJlc3NfYnlfbmFtZSh0cC0+c3ltYm9sLCBmYWxzZSk7DQo+ICAJCWlmIChhZGRyID09IDApDQo+ ICAJCQlnb3RvIGVycm9yOw0KPiBAQCAtMTgxNSwxNyArMTgxNSwxNyBAQCBzdGF0aWMgaW50IGZp bmRfcGVyZl9wcm9iZV9wb2ludF9mcm9tX21hcChzdHJ1Y3QgcHJvYmVfdHJhY2VfcG9pbnQgKnRw LA0KPiAgew0KPiAgCXN0cnVjdCBzeW1ib2wgKnN5bSA9IE5VTEw7DQo+ICAJc3RydWN0IG1hcCAq bWFwOw0KPiAtCXU2NCBhZGRyOw0KPiArCXU2NCBhZGRyID0gdHAtPmFkZHJlc3M7DQo+ICAJaW50 IHJldCA9IC1FTk9FTlQ7DQo+IA0KPiAgCWlmICghaXNfa3Byb2JlKSB7DQo+ICAJCW1hcCA9IGRz b19fbmV3X21hcCh0cC0+bW9kdWxlKTsNCj4gIAkJaWYgKCFtYXApDQo+ICAJCQlnb3RvIG91dDsN Cj4gLQkJYWRkciA9IHRwLT5hZGRyZXNzOw0KPiAgCQlzeW0gPSBtYXBfX2ZpbmRfc3ltYm9sKG1h cCwgYWRkciwgTlVMTCk7DQo+ICAJfSBlbHNlIHsNCj4gLQkJYWRkciA9IGtlcm5lbF9nZXRfc3lt Ym9sX2FkZHJlc3NfYnlfbmFtZSh0cC0+c3ltYm9sLCB0cnVlKTsNCj4gKwkJaWYgKHRwLT5zeW1i b2wpDQo+ICsJCQlhZGRyID0ga2VybmVsX2dldF9zeW1ib2xfYWRkcmVzc19ieV9uYW1lKHRwLT5z eW1ib2wsIHRydWUpOw0KPiAgCQlpZiAoYWRkcikgew0KPiAgCQkJYWRkciArPSB0cC0+b2Zmc2V0 Ow0KPiAgCQkJc3ltID0gX19maW5kX2tlcm5lbF9mdW5jdGlvbihhZGRyLCAmbWFwKTsNCj4gLS0N Cj4gMS44LjMuNA0KDQo= -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [next] | [standalone]
| From | tip-bot for Wang Nan <tipbot@zytor.com> |
|---|---|
| Date | 2015-08-28 08:50 +0200 |
| Subject | [tip:perf/core] perf probe: Prevent segfault when reading probe point with absolute address |
| Message-ID | <q2lE6-3R6-7@gated-at.bofh.it> |
| In reply to | #1213038 |
Commit-ID: e486367f0101d43a044404ed2fa35aa043b19667
Gitweb: http://git.kernel.org/tip/e486367f0101d43a044404ed2fa35aa043b19667
Author: Wang Nan <wangnan0@huawei.com>
AuthorDate: Tue, 25 Aug 2015 13:27:35 +0000
Committer: Arnaldo Carvalho de Melo <acme@redhat.com>
CommitDate: Wed, 26 Aug 2015 10:34:11 -0300
perf probe: Prevent segfault when reading probe point with absolute address
'perf probe -l' panic if there is a manually inserted probing point with
absolute address. For example:
# echo 'p:probe/abs_ffffffff811e6615 0xffffffff811e6615' > /sys/kernel/debug/tracing/kprobe_events
# perf probe -l
Segmentation fault (core dumped)
This patch fix this problem by considering the situation that
"tp->symbol == NULL" in find_perf_probe_point_from_dwarf() and
find_perf_probe_point_from_map().
After this patch:
# perf probe -l
probe:abs_ffffffff811e6615 (on SyS_write+5@fs/read_write.c)
And when debug info is missing:
# rm -rf ~/.debug
# mv /lib/modules/4.2.0-rc1+/build/vmlinux /lib/modules/4.2.0-rc1+/build/vmlinux.bak
# perf probe -l
probe:abs_ffffffff811e6615 (on sys_write+5)
Signed-off-by: Wang Nan <wangnan0@huawei.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: pi3orama@163.com
Link: http://lkml.kernel.org/r/1440509256-193590-1-git-send-email-wangnan0@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
tools/perf/util/probe-event.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index f07374b..6c7e538 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -515,7 +515,7 @@ static int find_perf_probe_point_from_dwarf(struct probe_trace_point *tp,
if (ret < 0)
goto error;
addr += stext;
- } else {
+ } else if (tp->symbol) {
addr = kernel_get_symbol_address_by_name(tp->symbol, false);
if (addr == 0)
goto error;
@@ -1815,17 +1815,17 @@ static int find_perf_probe_point_from_map(struct probe_trace_point *tp,
{
struct symbol *sym = NULL;
struct map *map;
- u64 addr;
+ u64 addr = tp->address;
int ret = -ENOENT;
if (!is_kprobe) {
map = dso__new_map(tp->module);
if (!map)
goto out;
- addr = tp->address;
sym = map__find_symbol(map, addr, NULL);
} else {
- addr = kernel_get_symbol_address_by_name(tp->symbol, true);
+ if (tp->symbol)
+ addr = kernel_get_symbol_address_by_name(tp->symbol, true);
if (addr) {
addr += tp->offset;
sym = __find_kernel_function(addr, &map);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web