[GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Arnaldo Carvalho de Melo]

Hi Ingo,

        Please pull from:

git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux-2.6 perf

Stephane and Tom,

	Now the build id code should be back working:

[root@doppio linux-2.6-tip]# rm -rf /root/.debug/
[root@doppio linux-2.6-tip]# perf record find / > /dev/null
^C[ perf record: Woken up 3 times to write data ]
[ perf record: Captured and wrote 0.667 MB perf.data (~29163 samples) ]
[root@doppio linux-2.6-tip]# perf buildid-list
349199a1e1b7cc25ac4f7004deed260f0146665e [kernel.kallsyms]
1f095dcd586045d7aa158b4774b8b053e7f580a4 /lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko
78913cef49182f29a0237d9867ff12811df00663 /lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko
5b2b076ef0b0be593d76c96fbd4393455380e589 /lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko
b6c457f3d69870e6f68f2c61978f8e3beb25221c /lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko
2249bf363819205e320e242af2e757906be5010e /lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko
a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f /bin/find
eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 /lib64/libc-2.10.2.so
5c68f7afeb33309c78037e374b0deee84dd441f6 /lib64/libpthread-2.10.2.so
[root@doppio linux-2.6-tip]# ls -la /root/.debug/
total 28
drwxr-xr-x   7 root root 4096 Mai 19 13:56 .
drwxr-x---. 31 root root 4096 Mai 19 13:56 ..
drwxr-xr-x   3 root root 4096 Mai 19 13:56 bin
drwxr-xr-x  76 root root 4096 Mai 19 13:56 .build-id
drwxr-xr-x   2 root root 4096 Mai 19 13:56 [kernel.kallsyms]
drwxr-xr-x   3 root root 4096 Mai 19 13:56 lib
drwxr-xr-x   4 root root 4096 Mai 19 13:56 lib64
[root@doppio linux-2.6-tip]# ls -la /root/.debug/.build-id/
03/ 0c/ 13/ 1e/ 20/ 28/ 2b/ 2d/ 34/ 3b/ 3f/ 47/ 56/ 5b/ 5e/ 6e/ 72/ 7e/
85/ 87/ 94/ 98/ a0/ a5/ a8/ ac/ bc/ c8/ cc/ d9/ dd/ e2/ e8/ eb/ ef/ f2/
fd/ 0b/ 0e/ 19/ 1f/ 22/ 29/ 2c/ 2e/ 39/ 3e/ 42/ 55/ 57/ 5c/ 62/ 6f/ 78/
80/ 86/ 91/ 97/ 99/ a1/ a7/ a9/ b6/ c6/ ca/ d8/ dc/ df/ e3/ e9/ ed/ f1/
fa/ fe/ 
[root@doppio linux-2.6-tip]# ls -la /root/.debug/lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 
-rwxr-xr-x 2 root root 1825560 Nov 20 09:29 /root/.debug/lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
[root@doppio linux-2.6-tip]# ls -la /root/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 
lrwxrwxrwx 1 root root 67 Mai 19 13:56 /root/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 -> ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
[root@doppio linux-2.6-tip]#
[root@doppio linux-2.6-tip]# perf archive
Now please run:

$ tar xvf perf.data.tar.bz2 -C ~/.debug

wherever you need to run 'perf report' on.
[root@doppio linux-2.6-tip]# 

ot@doppio linux-2.6-tip]# tar tvf perf.data.tar.bz2
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/34/9199a1e1b7cc25ac4f7004deed260f0146665e -> ../../[kernel.kallsyms]/349199a1e1b7cc25ac4f7004deed260f0146665e
-rw-r--r-- root/root   2842888 2010-05-19 13:56 [kernel.kallsyms]/349199a1e1b7cc25ac4f7004deed260f0146665e
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/1f/095dcd586045d7aa158b4774b8b053e7f580a4 -> ../../lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko/1f095dcd586045d7aa158b4774b8b053e7f580a4
-rw-r--r-- root/root   1670552 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko/1f095dcd586045d7aa158b4774b8b053e7f580a4
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/78/913cef49182f29a0237d9867ff12811df00663 -> ../../lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko/78913cef49182f29a0237d9867ff12811df00663
-rw-r--r-- root/root    326827 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko/78913cef49182f29a0237d9867ff12811df00663
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/5b/2b076ef0b0be593d76c96fbd4393455380e589 -> ../../lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko/5b2b076ef0b0be593d76c96fbd4393455380e589
-rw-r--r-- root/root    325915 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko/5b2b076ef0b0be593d76c96fbd4393455380e589
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/b6/c457f3d69870e6f68f2c61978f8e3beb25221c -> ../../lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko/b6c457f3d69870e6f68f2c61978f8e3beb25221c
-rw-r--r-- root/root   2587163 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko/b6c457f3d69870e6f68f2c61978f8e3beb25221c
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/22/49bf363819205e320e242af2e757906be5010e -> ../../lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko/2249bf363819205e320e242af2e757906be5010e
-rw-r--r-- root/root   8296246 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko/2249bf363819205e320e242af2e757906be5010e
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/a8/e4f743b40fb1fd8b85e2f9b88d93b661472b8f -> ../../bin/find/a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f
-rwxr-xr-x root/root    233032 2009-02-24 14:13 bin/find/a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 -> ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
-rwxr-xr-x root/root   1825560 2009-11-20 09:29 lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/5c/68f7afeb33309c78037e374b0deee84dd441f6 -> ../../lib64/libpthread-2.10.2.so/5c68f7afeb33309c78037e374b0deee84dd441f6
-rwxr-xr-x root/root    148528 2009-11-20 09:29 lib64/libpthread-2.10.2.so/5c68f7afeb33309c78037e374b0deee84dd441f6
[root@doppio linux-2.6-tip]# 

Regards,

- Arnaldo

Arnaldo Carvalho de Melo (2):
  perf symbols: Don't try to read the build-id twice
  perf session: Make read_build_id routines look at the host_machine too

 tools/perf/util/header.c  |   84 ++++++++++++++++++++++++++++-----------------
 tools/perf/util/session.c |    7 ++++
 tools/perf/util/session.h |    8 +---
 tools/perf/util/symbol.c  |   13 ++++++-
 tools/perf/util/symbol.h  |    1 +
 5 files changed, 73 insertions(+), 40 deletions(-)

[PATCH 1/2] perf symbols: Don't try to read the build-id twice

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

In __dsos__read_build_ids if the dso already had its build-id read,
don't try again.

Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Tom Zanussi <tzanussi@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/symbol.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index a06131f..87d9b1b 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -1131,6 +1131,10 @@ bool __dsos__read_build_ids(struct list_head *head, bool with_hits)
 	list_for_each_entry(pos, head, node) {
 		if (with_hits && !pos->hit)
 			continue;
+		if (pos->has_build_id) {
+			have_build_id = true;
+			continue;
+		}
 		if (filename__read_build_id(pos->long_name, pos->build_id,
 					    sizeof(pos->build_id)) > 0) {
 			have_build_id	  = true;
-- 
1.6.2.5

[PATCH 2/2] perf session: Make read_build_id routines look at the host_machine too

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

The changes made to support host and guest machines in a session, that
started when the 'perf kvm' tool was introduced ended up introducing a
bug where the host_machine was not having its DSOs traversed for
build-id processing.

Fix it by moving some methods to the right classes and considering the
host_machine when processing build-ids.

Reported-by: Tom Zanussi <tzanussi@gmail.com>
Reported-by: Stephane Eranian <eranian@google.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Tom Zanussi <tzanussi@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/header.c  |   84 ++++++++++++++++++++++++++++-----------------
 tools/perf/util/session.c |    7 ++++
 tools/perf/util/session.h |    8 +---
 tools/perf/util/symbol.c  |    9 ++++-
 tools/perf/util/symbol.h  |    1 +
 5 files changed, 69 insertions(+), 40 deletions(-)

diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index 8847bec..1f62435 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -221,29 +221,38 @@ static int __dsos__write_buildid_table(struct list_head *head, pid_t pid,
 	return 0;
 }
 
+static int machine__write_buildid_table(struct machine *self, int fd)
+{
+	int err;
+	u16 kmisc = PERF_RECORD_MISC_KERNEL,
+	    umisc = PERF_RECORD_MISC_USER;
+
+	if (!machine__is_host(self)) {
+		kmisc = PERF_RECORD_MISC_GUEST_KERNEL;
+		umisc = PERF_RECORD_MISC_GUEST_USER;
+	}
+
+	err = __dsos__write_buildid_table(&self->kernel_dsos, self->pid,
+					  kmisc, fd);
+	if (err == 0)
+		err = __dsos__write_buildid_table(&self->user_dsos,
+						  self->pid, umisc, fd);
+	return err;
+}
+
 static int dsos__write_buildid_table(struct perf_header *header, int fd)
 {
 	struct perf_session *session = container_of(header,
 			struct perf_session, header);
 	struct rb_node *nd;
-	int err = 0;
-	u16 kmisc, umisc;
+	int err = machine__write_buildid_table(&session->host_machine, fd);
+
+	if (err)
+		return err;
 
 	for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) {
 		struct machine *pos = rb_entry(nd, struct machine, rb_node);
-		if (machine__is_host(pos)) {
-			kmisc = PERF_RECORD_MISC_KERNEL;
-			umisc = PERF_RECORD_MISC_USER;
-		} else {
-			kmisc = PERF_RECORD_MISC_GUEST_KERNEL;
-			umisc = PERF_RECORD_MISC_GUEST_USER;
-		}
-
-		err = __dsos__write_buildid_table(&pos->kernel_dsos, pos->pid,
-						  kmisc, fd);
-		if (err == 0)
-			err = __dsos__write_buildid_table(&pos->user_dsos,
-							  pos->pid, umisc, fd);
+		err = machine__write_buildid_table(pos, fd);
 		if (err)
 			break;
 	}
@@ -363,12 +372,17 @@ static int __dsos__cache_build_ids(struct list_head *head, const char *debugdir)
 	return err;
 }
 
-static int dsos__cache_build_ids(struct perf_header *self)
+static int machine__cache_build_ids(struct machine *self, const char *debugdir)
+{
+	int ret = __dsos__cache_build_ids(&self->kernel_dsos, debugdir);
+	ret |= __dsos__cache_build_ids(&self->user_dsos, debugdir);
+	return ret;
+}
+
+static int perf_session__cache_build_ids(struct perf_session *self)
 {
-	struct perf_session *session = container_of(self,
-			struct perf_session, header);
 	struct rb_node *nd;
-	int ret = 0;
+	int ret;
 	char debugdir[PATH_MAX];
 
 	snprintf(debugdir, sizeof(debugdir), "%s/%s", getenv("HOME"),
@@ -377,25 +391,30 @@ static int dsos__cache_build_ids(struct perf_header *self)
 	if (mkdir(debugdir, 0755) != 0 && errno != EEXIST)
 		return -1;
 
-	for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) {
+	ret = machine__cache_build_ids(&self->host_machine, debugdir);
+
+	for (nd = rb_first(&self->machines); nd; nd = rb_next(nd)) {
 		struct machine *pos = rb_entry(nd, struct machine, rb_node);
-		ret |= __dsos__cache_build_ids(&pos->kernel_dsos, debugdir);
-		ret |= __dsos__cache_build_ids(&pos->user_dsos, debugdir);
+		ret |= machine__cache_build_ids(pos, debugdir);
 	}
 	return ret ? -1 : 0;
 }
 
-static bool dsos__read_build_ids(struct perf_header *self, bool with_hits)
+static bool machine__read_build_ids(struct machine *self, bool with_hits)
+{
+	bool ret = __dsos__read_build_ids(&self->kernel_dsos, with_hits);
+	ret |= __dsos__read_build_ids(&self->user_dsos, with_hits);
+	return ret;
+}
+
+static bool perf_session__read_build_ids(struct perf_session *self, bool with_hits)
 {
-	bool ret = false;
-	struct perf_session *session = container_of(self,
-			struct perf_session, header);
 	struct rb_node *nd;
+	bool ret = machine__read_build_ids(&self->host_machine, with_hits);
 
-	for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) {
+	for (nd = rb_first(&self->machines); nd; nd = rb_next(nd)) {
 		struct machine *pos = rb_entry(nd, struct machine, rb_node);
-		ret |= __dsos__read_build_ids(&pos->kernel_dsos, with_hits);
-		ret |= __dsos__read_build_ids(&pos->user_dsos, with_hits);
+		ret |= machine__read_build_ids(pos, with_hits);
 	}
 
 	return ret;
@@ -404,12 +423,14 @@ static bool dsos__read_build_ids(struct perf_header *self, bool with_hits)
 static int perf_header__adds_write(struct perf_header *self, int fd)
 {
 	int nr_sections;
+	struct perf_session *session;
 	struct perf_file_section *feat_sec;
 	int sec_size;
 	u64 sec_start;
 	int idx = 0, err;
 
-	if (dsos__read_build_ids(self, true))
+	session = container_of(self, struct perf_session, header);
+	if (perf_session__read_build_ids(session, true))
 		perf_header__set_feat(self, HEADER_BUILD_ID);
 
 	nr_sections = bitmap_weight(self->adds_features, HEADER_FEAT_BITS);
@@ -450,7 +471,7 @@ static int perf_header__adds_write(struct perf_header *self, int fd)
 		}
 		buildid_sec->size = lseek(fd, 0, SEEK_CUR) -
 					  buildid_sec->offset;
-		dsos__cache_build_ids(self);
+		perf_session__cache_build_ids(session);
 	}
 
 	lseek(fd, sec_start, SEEK_SET);
@@ -490,7 +511,6 @@ int perf_header__write(struct perf_header *self, int fd, bool at_exit)
 
 	lseek(fd, sizeof(f_header), SEEK_SET);
 
-
 	for (i = 0; i < self->attrs; i++) {
 		attr = self->attr[i];
 
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index e4eaa6d..8f83a18 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -895,3 +895,10 @@ size_t perf_session__fprintf_dsos(struct perf_session *self, FILE *fp)
 	       __dsos__fprintf(&self->host_machine.user_dsos, fp) +
 	       machines__fprintf_dsos(&self->machines, fp);
 }
+
+size_t perf_session__fprintf_dsos_buildid(struct perf_session *self, FILE *fp,
+					  bool with_hits)
+{
+	size_t ret = machine__fprintf_dsos_buildid(&self->host_machine, fp, with_hits);
+	return ret + machines__fprintf_dsos_buildid(&self->machines, fp, with_hits);
+}
diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h
index e7fce48..55c6881 100644
--- a/tools/perf/util/session.h
+++ b/tools/perf/util/session.h
@@ -132,12 +132,8 @@ void perf_session__process_machines(struct perf_session *self,
 
 size_t perf_session__fprintf_dsos(struct perf_session *self, FILE *fp);
 
-static inline
-size_t perf_session__fprintf_dsos_buildid(struct perf_session *self, FILE *fp,
-					  bool with_hits)
-{
-	return machines__fprintf_dsos_buildid(&self->machines, fp, with_hits);
-}
+size_t perf_session__fprintf_dsos_buildid(struct perf_session *self,
+					  FILE *fp, bool with_hits);
 
 static inline
 size_t perf_session__fprintf_nr_events(struct perf_session *self, FILE *fp)
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 87d9b1b..96bff0e 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -1937,6 +1937,12 @@ static size_t __dsos__fprintf_buildid(struct list_head *head, FILE *fp,
 	return ret;
 }
 
+size_t machine__fprintf_dsos_buildid(struct machine *self, FILE *fp, bool with_hits)
+{
+	return __dsos__fprintf_buildid(&self->kernel_dsos, fp, with_hits) +
+	       __dsos__fprintf_buildid(&self->user_dsos, fp, with_hits);
+}
+
 size_t machines__fprintf_dsos_buildid(struct rb_root *self, FILE *fp, bool with_hits)
 {
 	struct rb_node *nd;
@@ -1944,8 +1950,7 @@ size_t machines__fprintf_dsos_buildid(struct rb_root *self, FILE *fp, bool with_
 
 	for (nd = rb_first(self); nd; nd = rb_next(nd)) {
 		struct machine *pos = rb_entry(nd, struct machine, rb_node);
-		ret += __dsos__fprintf_buildid(&pos->kernel_dsos, fp, with_hits);
-		ret += __dsos__fprintf_buildid(&pos->user_dsos, fp, with_hits);
+		ret += machine__fprintf_dsos_buildid(pos, fp, with_hits);
 	}
 	return ret;
 }
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index 032469e..5d25b5e 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -170,6 +170,7 @@ int machine__load_vmlinux_path(struct machine *self, enum map_type type,
 
 size_t __dsos__fprintf(struct list_head *head, FILE *fp);
 
+size_t machine__fprintf_dsos_buildid(struct machine *self, FILE *fp, bool with_hits);
 size_t machines__fprintf_dsos(struct rb_root *self, FILE *fp);
 size_t machines__fprintf_dsos_buildid(struct rb_root *self, FILE *fp, bool with_hits);
 
-- 
1.6.2.5

Re: [GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Stephane Eranian]

Hi,

On Wed, May 19, 2010 at 7:01 PM, Arnaldo Carvalho de Melo
<acme@infradead.org> wrote:
> Hi Ingo,
>
>        Please pull from:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux-2.6 perf
>
> Stephane and Tom,
>
>        Now the build id code should be back working:
>
> [root@doppio linux-2.6-tip]# rm -rf /root/.debug/
> [root@doppio linux-2.6-tip]# perf record find / > /dev/null
> ^C[ perf record: Woken up 3 times to write data ]
> [ perf record: Captured and wrote 0.667 MB perf.data (~29163 samples) ]
> [root@doppio linux-2.6-tip]# perf buildid-list
> 349199a1e1b7cc25ac4f7004deed260f0146665e [kernel.kallsyms]


It is working again for me too. Thanks for the quick fix.

Related to .debug, I think it would be useful to have a choice as to
where all that stuff gets saved. It is hardcoded to $HOME today.

We could provide:
- an environment variable, e.g., PERFHOME or PERFDEBUGDIR
- an option to perf record, perf report, perf annotate, i.e., all commands
  that use .debug

I looked at that today and I have a draft patch to add both the variable
+ the option.

The problem with the environment variable is that you tend to forget to
set it. As you suggest below, you may run perf report on another machine
too and you don't necessarily want to overwrite the .debug you may already
have in your $HOME dir. I think the option has the advantage of making this
explicit. It would still default to $HOME if not set.


> 1f095dcd586045d7aa158b4774b8b053e7f580a4 /lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko
> 78913cef49182f29a0237d9867ff12811df00663 /lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko
> 5b2b076ef0b0be593d76c96fbd4393455380e589 /lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko
> b6c457f3d69870e6f68f2c61978f8e3beb25221c /lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko
> 2249bf363819205e320e242af2e757906be5010e /lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko
> a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f /bin/find
> eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 /lib64/libc-2.10.2.so
> 5c68f7afeb33309c78037e374b0deee84dd441f6 /lib64/libpthread-2.10.2.so
> [root@doppio linux-2.6-tip]# ls -la /root/.debug/
> total 28
> drwxr-xr-x   7 root root 4096 Mai 19 13:56 .
> drwxr-x---. 31 root root 4096 Mai 19 13:56 ..
> drwxr-xr-x   3 root root 4096 Mai 19 13:56 bin
> drwxr-xr-x  76 root root 4096 Mai 19 13:56 .build-id
> drwxr-xr-x   2 root root 4096 Mai 19 13:56 [kernel.kallsyms]
> drwxr-xr-x   3 root root 4096 Mai 19 13:56 lib
> drwxr-xr-x   4 root root 4096 Mai 19 13:56 lib64
> [root@doppio linux-2.6-tip]# ls -la /root/.debug/.build-id/
> 03/ 0c/ 13/ 1e/ 20/ 28/ 2b/ 2d/ 34/ 3b/ 3f/ 47/ 56/ 5b/ 5e/ 6e/ 72/ 7e/
> 85/ 87/ 94/ 98/ a0/ a5/ a8/ ac/ bc/ c8/ cc/ d9/ dd/ e2/ e8/ eb/ ef/ f2/
> fd/ 0b/ 0e/ 19/ 1f/ 22/ 29/ 2c/ 2e/ 39/ 3e/ 42/ 55/ 57/ 5c/ 62/ 6f/ 78/
> 80/ 86/ 91/ 97/ 99/ a1/ a7/ a9/ b6/ c6/ ca/ d8/ dc/ df/ e3/ e9/ ed/ f1/
> fa/ fe/
> [root@doppio linux-2.6-tip]# ls -la /root/.debug/lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> -rwxr-xr-x 2 root root 1825560 Nov 20 09:29 /root/.debug/lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> [root@doppio linux-2.6-tip]# ls -la /root/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> lrwxrwxrwx 1 root root 67 Mai 19 13:56 /root/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 -> ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> [root@doppio linux-2.6-tip]#
> [root@doppio linux-2.6-tip]# perf archive
> Now please run:
>
> $ tar xvf perf.data.tar.bz2 -C ~/.debug
>
> wherever you need to run 'perf report' on.
> [root@doppio linux-2.6-tip]#
>
> ot@doppio linux-2.6-tip]# tar tvf perf.data.tar.bz2
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/34/9199a1e1b7cc25ac4f7004deed260f0146665e -> ../../[kernel.kallsyms]/349199a1e1b7cc25ac4f7004deed260f0146665e
> -rw-r--r-- root/root   2842888 2010-05-19 13:56 [kernel.kallsyms]/349199a1e1b7cc25ac4f7004deed260f0146665e
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/1f/095dcd586045d7aa158b4774b8b053e7f580a4 -> ../../lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko/1f095dcd586045d7aa158b4774b8b053e7f580a4
> -rw-r--r-- root/root   1670552 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko/1f095dcd586045d7aa158b4774b8b053e7f580a4
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/78/913cef49182f29a0237d9867ff12811df00663 -> ../../lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko/78913cef49182f29a0237d9867ff12811df00663
> -rw-r--r-- root/root    326827 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko/78913cef49182f29a0237d9867ff12811df00663
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/5b/2b076ef0b0be593d76c96fbd4393455380e589 -> ../../lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko/5b2b076ef0b0be593d76c96fbd4393455380e589
> -rw-r--r-- root/root    325915 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko/5b2b076ef0b0be593d76c96fbd4393455380e589
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/b6/c457f3d69870e6f68f2c61978f8e3beb25221c -> ../../lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko/b6c457f3d69870e6f68f2c61978f8e3beb25221c
> -rw-r--r-- root/root   2587163 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko/b6c457f3d69870e6f68f2c61978f8e3beb25221c
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/22/49bf363819205e320e242af2e757906be5010e -> ../../lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko/2249bf363819205e320e242af2e757906be5010e
> -rw-r--r-- root/root   8296246 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko/2249bf363819205e320e242af2e757906be5010e
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/a8/e4f743b40fb1fd8b85e2f9b88d93b661472b8f -> ../../bin/find/a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f
> -rwxr-xr-x root/root    233032 2009-02-24 14:13 bin/find/a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 -> ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> -rwxr-xr-x root/root   1825560 2009-11-20 09:29 lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/5c/68f7afeb33309c78037e374b0deee84dd441f6 -> ../../lib64/libpthread-2.10.2.so/5c68f7afeb33309c78037e374b0deee84dd441f6
> -rwxr-xr-x root/root    148528 2009-11-20 09:29 lib64/libpthread-2.10.2.so/5c68f7afeb33309c78037e374b0deee84dd441f6
> [root@doppio linux-2.6-tip]#
>
> Regards,
>
> - Arnaldo
>
> Arnaldo Carvalho de Melo (2):
>  perf symbols: Don't try to read the build-id twice
>  perf session: Make read_build_id routines look at the host_machine too
>
>  tools/perf/util/header.c  |   84 ++++++++++++++++++++++++++++-----------------
>  tools/perf/util/session.c |    7 ++++
>  tools/perf/util/session.h |    8 +---
>  tools/perf/util/symbol.c  |   13 ++++++-
>  tools/perf/util/symbol.h  |    1 +
>  5 files changed, 73 insertions(+), 40 deletions(-)
>
>

Re: [GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Arnaldo Carvalho de Melo]

Em Wed, May 19, 2010 at 09:36:45PM +0200, Stephane Eranian escreveu:
> Hi,
> 
> On Wed, May 19, 2010 at 7:01 PM, Arnaldo Carvalho de Melo
> <acme@infradead.org> wrote:
> > Hi Ingo,
> >
> >        Please pull from:
> >
> > git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux-2.6 perf
> >
> > Stephane and Tom,
> >
> >        Now the build id code should be back working:
> >
> > [root@doppio linux-2.6-tip]# rm -rf /root/.debug/
> > [root@doppio linux-2.6-tip]# perf record find / > /dev/null
> > ^C[ perf record: Woken up 3 times to write data ]
> > [ perf record: Captured and wrote 0.667 MB perf.data (~29163 samples) ]
> > [root@doppio linux-2.6-tip]# perf buildid-list
> > 349199a1e1b7cc25ac4f7004deed260f0146665e [kernel.kallsyms]
> 
> 
> It is working again for me too. Thanks for the quick fix.

Great, and thanks for the "quick" part, myself I thought it took more
time than needed :-)
 
> Related to .debug, I think it would be useful to have a choice as to
> where all that stuff gets saved. It is hardcoded to $HOME today.
> 
> We could provide:
> - an environment variable, e.g., PERFHOME or PERFDEBUGDIR
> - an option to perf record, perf report, perf annotate, i.e., all commands
>   that use .debug
> 
> I looked at that today and I have a draft patch to add both the variable
> + the option.
>

Yeah, the NO_NEWT make target was out of our uneasyness about making
newt the sole report output, and my patch removing tons of code stolen
but not used (greedy us!) was also in that direction, trying to get the
git ~/.CONFIG way into perf.

We have to have that and consistent accross the tools command line
options to override where this is stored/used.

> The problem with the environment variable is that you tend to forget to
> set it. As you suggest below, you may run perf report on another machine
> too and you don't necessarily want to overwrite the .debug you may already

Well, we wouldn't be overwriting anything since we hope that what is
used to create the buildid is reasonably safe against collision, so even
if the DSOs are for different arches, it should be rather difficult to
get 'perf unarchive' causing problems.

> have in your $HOME dir. I think the option has the advantage of making this
> explicit. It would still default to $HOME if not set.

> 
> > 1f095dcd586045d7aa158b4774b8b053e7f580a4 /lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko
> > 78913cef49182f29a0237d9867ff12811df00663 /lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko
> > 5b2b076ef0b0be593d76c96fbd4393455380e589 /lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko
> > b6c457f3d69870e6f68f2c61978f8e3beb25221c /lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko
> > 2249bf363819205e320e242af2e757906be5010e /lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko
> > a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f /bin/find
> > eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 /lib64/libc-2.10.2.so
> > 5c68f7afeb33309c78037e374b0deee84dd441f6 /lib64/libpthread-2.10.2.so
> > [root@doppio linux-2.6-tip]# ls -la /root/.debug/
> > total 28
> > drwxr-xr-x   7 root root 4096 Mai 19 13:56 .
> > drwxr-x---. 31 root root 4096 Mai 19 13:56 ..
> > drwxr-xr-x   3 root root 4096 Mai 19 13:56 bin
> > drwxr-xr-x  76 root root 4096 Mai 19 13:56 .build-id
> > drwxr-xr-x   2 root root 4096 Mai 19 13:56 [kernel.kallsyms]
> > drwxr-xr-x   3 root root 4096 Mai 19 13:56 lib
> > drwxr-xr-x   4 root root 4096 Mai 19 13:56 lib64
> > [root@doppio linux-2.6-tip]# ls -la /root/.debug/.build-id/
> > 03/ 0c/ 13/ 1e/ 20/ 28/ 2b/ 2d/ 34/ 3b/ 3f/ 47/ 56/ 5b/ 5e/ 6e/ 72/ 7e/
> > 85/ 87/ 94/ 98/ a0/ a5/ a8/ ac/ bc/ c8/ cc/ d9/ dd/ e2/ e8/ eb/ ef/ f2/
> > fd/ 0b/ 0e/ 19/ 1f/ 22/ 29/ 2c/ 2e/ 39/ 3e/ 42/ 55/ 57/ 5c/ 62/ 6f/ 78/
> > 80/ 86/ 91/ 97/ 99/ a1/ a7/ a9/ b6/ c6/ ca/ d8/ dc/ df/ e3/ e9/ ed/ f1/
> > fa/ fe/
> > [root@doppio linux-2.6-tip]# ls -la /root/.debug/lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> > -rwxr-xr-x 2 root root 1825560 Nov 20 09:29 /root/.debug/lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> > [root@doppio linux-2.6-tip]# ls -la /root/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> > lrwxrwxrwx 1 root root 67 Mai 19 13:56 /root/.debug/.build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 -> ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> > [root@doppio linux-2.6-tip]#
> > [root@doppio linux-2.6-tip]# perf archive
> > Now please run:
> >
> > $ tar xvf perf.data.tar.bz2 -C ~/.debug
> >
> > wherever you need to run 'perf report' on.
> > [root@doppio linux-2.6-tip]#
> >
> > ot@doppio linux-2.6-tip]# tar tvf perf.data.tar.bz2
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/34/9199a1e1b7cc25ac4f7004deed260f0146665e -> ../../[kernel.kallsyms]/349199a1e1b7cc25ac4f7004deed260f0146665e
> > -rw-r--r-- root/root   2842888 2010-05-19 13:56 [kernel.kallsyms]/349199a1e1b7cc25ac4f7004deed260f0146665e
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/1f/095dcd586045d7aa158b4774b8b053e7f580a4 -> ../../lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko/1f095dcd586045d7aa158b4774b8b053e7f580a4
> > -rw-r--r-- root/root   1670552 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/net/ipv4/netfilter/nf_nat.ko/1f095dcd586045d7aa158b4774b8b053e7f580a4
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/78/913cef49182f29a0237d9867ff12811df00663 -> ../../lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko/78913cef49182f29a0237d9867ff12811df00663
> > -rw-r--r-- root/root    326827 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/md/dm-crypt.ko/78913cef49182f29a0237d9867ff12811df00663
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/5b/2b076ef0b0be593d76c96fbd4393455380e589 -> ../../lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko/5b2b076ef0b0be593d76c96fbd4393455380e589
> > -rw-r--r-- root/root    325915 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/ata/sata_sil.ko/5b2b076ef0b0be593d76c96fbd4393455380e589
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/b6/c457f3d69870e6f68f2c61978f8e3beb25221c -> ../../lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko/b6c457f3d69870e6f68f2c61978f8e3beb25221c
> > -rw-r--r-- root/root   2587163 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/net/e1000e/e1000e.ko/b6c457f3d69870e6f68f2c61978f8e3beb25221c
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/22/49bf363819205e320e242af2e757906be5010e -> ../../lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko/2249bf363819205e320e242af2e757906be5010e
> > -rw-r--r-- root/root   8296246 2010-04-30 11:43 lib/modules/2.6.34-rc6/kernel/drivers/gpu/drm/i915/i915.ko/2249bf363819205e320e242af2e757906be5010e
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/a8/e4f743b40fb1fd8b85e2f9b88d93b661472b8f -> ../../bin/find/a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f
> > -rwxr-xr-x root/root    233032 2009-02-24 14:13 bin/find/a8e4f743b40fb1fd8b85e2f9b88d93b661472b8f
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/eb/4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1 -> ../../lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> > -rwxr-xr-x root/root   1825560 2009-11-20 09:29 lib64/libc-2.10.2.so/eb4ec8fa8b2a5eb18cad173c92f27ed8887ed1c1
> > lrwxrwxrwx root/root         0 2010-05-19 13:56 .build-id/5c/68f7afeb33309c78037e374b0deee84dd441f6 -> ../../lib64/libpthread-2.10.2.so/5c68f7afeb33309c78037e374b0deee84dd441f6
> > -rwxr-xr-x root/root    148528 2009-11-20 09:29 lib64/libpthread-2.10.2.so/5c68f7afeb33309c78037e374b0deee84dd441f6
> > [root@doppio linux-2.6-tip]#
> >
> > Regards,
> >
> > - Arnaldo
> >
> > Arnaldo Carvalho de Melo (2):
> >  perf symbols: Don't try to read the build-id twice
> >  perf session: Make read_build_id routines look at the host_machine too
> >
> >  tools/perf/util/header.c  |   84 ++++++++++++++++++++++++++++-----------------
> >  tools/perf/util/session.c |    7 ++++
> >  tools/perf/util/session.h |    8 +---
> >  tools/perf/util/symbol.c  |   13 ++++++-
> >  tools/perf/util/symbol.h  |    1 +
> >  5 files changed, 73 insertions(+), 40 deletions(-)
> >
> >

Re: [GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Stephane Eranian]

On Wed, May 19, 2010 at 10:11 PM, Arnaldo Carvalho de Melo
<acme@infradead.org> wrote:
>
> Em Wed, May 19, 2010 at 09:36:45PM +0200, Stephane Eranian escreveu:
> > Hi,
> >
> > On Wed, May 19, 2010 at 7:01 PM, Arnaldo Carvalho de Melo
> > <acme@infradead.org> wrote:
> > > Hi Ingo,
> > >
> > >        Please pull from:
> > >
> > > git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux-2.6 perf
> > >
> > > Stephane and Tom,
> > >
> > >        Now the build id code should be back working:
> > >
> > > [root@doppio linux-2.6-tip]# rm -rf /root/.debug/
> > > [root@doppio linux-2.6-tip]# perf record find / > /dev/null
> > > ^C[ perf record: Woken up 3 times to write data ]
> > > [ perf record: Captured and wrote 0.667 MB perf.data (~29163 samples) ]
> > > [root@doppio linux-2.6-tip]# perf buildid-list
> > > 349199a1e1b7cc25ac4f7004deed260f0146665e [kernel.kallsyms]
> >
> >
> > It is working again for me too. Thanks for the quick fix.
>
> Great, and thanks for the "quick" part, myself I thought it took more
> time than needed :-)

Well I could not understand the logical code flow with buildids, that's why
I posted the bug report. I am glad you fixed this.

>
> > Related to .debug, I think it would be useful to have a choice as to
> > where all that stuff gets saved. It is hardcoded to $HOME today.
> >
> > We could provide:
> > - an environment variable, e.g., PERFHOME or PERFDEBUGDIR
> > - an option to perf record, perf report, perf annotate, i.e., all commands
> >   that use .debug
> >
> > I looked at that today and I have a draft patch to add both the variable
> > + the option.
> >
>
> Yeah, the NO_NEWT make target was out of our uneasyness about making
> newt the sole report output, and my patch removing tons of code stolen
> but not used (greedy us!) was also in that direction, trying to get the
> git ~/.CONFIG way into perf.

That's not because I compile with newt support that I want curses output each
time I use perf report. I could not find an option to turn it off.
Unless you tell
me that if it detects stdout is not a tty it does not use newt. But
that would not
be good enough I think.
Ok, so you are advocating that perf configuration options be saved in
~/.perfconfig
and that would include the destination dir for all the "symbolization"
files. BTW,
I think the name .debug is confusing here. It contains symbolization info not
debug info. For all I know the binaries may have no debug info at all.

>
> We have to have that and consistent accross the tools command line
> options to override where this is stored/used.
>
> > The problem with the environment variable is that you tend to forget to
> > set it. As you suggest below, you may run perf report on another machine
> > too and you don't necessarily want to overwrite the .debug you may already
>
> Well, we wouldn't be overwriting anything since we hope that what is
> used to create the buildid is reasonably safe against collision, so even
> if the DSOs are for different arches, it should be rather difficult to
> get 'perf unarchive' causing problems.


I think you're right. If you have the same file name, e.g.,
/usr/bin/date but compiled
differently, then you should have different buildids, thus you will
not clash on either
leaf files in your tree. Directories may already exist but that's
okay. Am I right?

Re: [GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Arnaldo Carvalho de Melo]

Em Wed, May 19, 2010 at 10:55:55PM +0200, Stephane Eranian escreveu:
> On Wed, May 19, 2010 at 10:11 PM, Arnaldo Carvalho de Melo <acme@infradead.org> wrote:
> > Yeah, the NO_NEWT make target was out of our uneasyness about making
> > newt the sole report output, and my patch removing tons of code stolen
> > but not used (greedy us!) was also in that direction, trying to get the
> > git ~/.CONFIG way into perf.

> That's not because I compile with newt support that I want curses output each
> time I use perf report. I could not find an option to turn it off.

> Unless you tell me that if it detects stdout is not a tty it does not
> use newt. But that would not be good enough I think.

Agreed, I'll add something like what git has for formatting, i.e. some
perf report --tty and --tui

With --tui being the default. If someone don't want it they can either
use 'make NO_NEWT=1', use | less, redirect to a file, remove the
newt-devel/libnewt-dev or set it in ~/.perfconfig.

> Ok, so you are advocating that perf configuration options be saved in
> ~/.perfconfig and that would include the destination dir for all the
> "symbolization" files. BTW, I think the name .debug is confusing here.
> It contains symbolization info not debug info. For all I know the
> binaries may have no debug info at all.

Yeah, I even called a symtab cache when I introduced it, but the idea
was that perhaps it could be used by systemtap and other tools that use
build id indexed debugging info, even if it is just a symtab.

But because it doesn't always have just a symtab, I decided to use
something more generic.

perf annotate, for instance, can use the DWARF info there if present.

> > We have to have that and consistent accross the tools command line
> > options to override where this is stored/used.
> >
> > > The problem with the environment variable is that you tend to forget to
> > > set it. As you suggest below, you may run perf report on another machine
> > > too and you don't necessarily want to overwrite the .debug you may already
> >
> > Well, we wouldn't be overwriting anything since we hope that what is
> > used to create the buildid is reasonably safe against collision, so even
> > if the DSOs are for different arches, it should be rather difficult to
> > get 'perf unarchive' causing problems.

> I think you're right. If you have the same file name, e.g.,
> /usr/bin/date but compiled differently, then you should have different
> buildids, thus you will not clash on either leaf files in your tree.
> Directories may already exist but that's okay. Am I right?

Yes, you are. I decided to have DSO names used as a directory where we
store files named by its build ids so that, among others, the following
use case becomes straight forward:

vi app.c
make

# caches app _at this point in time_ by build-id creates perf.data
perf record ./app

# do some changes, optimizations, whatever
vi app.c
make

# caches app _with another build-id_, renames existing perf.data file to
# perf.data.old
perf record ./app

# compares perf.data and perf.data.old, finding the two binary versions
perf diff

	Right now perf.data.old is deleted if present and perf.data is
renamed to it, so we have just two files for perf diff, but we can as
well implement code in 'perf diff' to consider N files, generating
plots, etc, and the build-id cache is what allows this to be done really
seamlessly.

- Arnaldo

Re: [GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Tom Zanussi]

On Wed, 2010-05-19 at 14:01 -0300, Arnaldo Carvalho de Melo wrote:
> Hi Ingo,
> 
>         Please pull from:
> 
> git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux-2.6 perf
> 
> Stephane and Tom,
> 
> 	Now the build id code should be back working:

Thanks for providing the fix!

Tom

Re: [GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Stephane Eranian]

Arnaldo,

On Wed, May 19, 2010 at 7:01 PM, Arnaldo Carvalho de Melo
<acme@infradead.org> wrote:
> [root@doppio linux-2.6-tip]#
> [root@doppio linux-2.6-tip]# perf archive
> Now please run:
>
> $ tar xvf perf.data.tar.bz2 -C ~/.debug
>
> wherever you need to run 'perf report' on.

Supposedly, on the remote machine with the perf.data +tarball files,
I should be able to run perf report, perf annotate.

I tried that. The perf report works but not perf annotate. It can't
find the binary image of my program. Yet there is a copy in the
.debug subdir. It seems, it is looking for the image using the
absolute file path in perf.data, instead.

Am I missing something here?

Re: [GIT PULL 0/2] perf session fix host_machine handling wrt build ids

[Arnaldo Carvalho de Melo]

Em Thu, May 20, 2010 at 11:54:17AM +0200, Stephane Eranian escreveu:
> Arnaldo,
> 
> On Wed, May 19, 2010 at 7:01 PM, Arnaldo Carvalho de Melo
> <acme@infradead.org> wrote:
> > [root@doppio linux-2.6-tip]#
> > [root@doppio linux-2.6-tip]# perf archive
> > Now please run:
> >
> > $ tar xvf perf.data.tar.bz2 -C ~/.debug
> >
> > wherever you need to run 'perf report' on.
> 
> Supposedly, on the remote machine with the perf.data +tarball files,
> I should be able to run perf report, perf annotate.
> 
> I tried that. The perf report works but not perf annotate. It can't
> find the binary image of my program. Yet there is a copy in the
> .debug subdir. It seems, it is looking for the image using the
> absolute file path in perf.data, instead.
> 
> Am I missing something here?

No, you are not, annotate should look by build-id, then pass this to
objdump, lemme cook a patch.

- Arnaldo