All of lore.kernel.org
 help / color / mirror / Atom feed
From: mhiramat at kernel.org (Masami Hiramatsu)
Subject: [PATCH v6 20/21] [RESEND] perf-probe: Fix to convert array type collectly
Date: Sat, 17 Mar 2018 21:52:25 +0900	[thread overview]
Message-ID: <152129114502.31874.2474068470011496356.stgit@devbox> (raw)
In-Reply-To: <152129024033.31874.15800253385376959274.stgit@devbox>

Fix to convert array element type collectly.
Current perf-probe converts the type of array-elements
incorrectly. It always converts the types as a pointer
of array. This passes the "array" type DIE to the type
converter so that it can get correct "element of array"
type DIE from it.

E.g.
  ====
  $ cat hello.c
  #include <stdio.h>

  void foo(int a[])
  {
	  printf("%d\n", a[1]);
  }

  void main()
  {
	  int a[3] = {4, 5, 6};
	  printf("%d\n", a[0]);
	  foo(a);
  }

  $ gcc -g hello.c -o hello
  $ perf probe -x ./hello -D "foo a[1]"
  ====

Without this fix, above outputs
  ====
  p:probe_hello/foo /tmp/hello:0x4d3 a=+4(-8(%bp)):u64
  ====
The "u64" means "int *", but a[1] is "int".

With this,
  ====
  p:probe_hello/foo /tmp/hello:0x4d3 a=+4(-8(%bp)):s32
  ====
So, "int" correctly converted to "s32"

Fixes: b2a3c12b7442 ("perf probe: Support tracing an entry of array")
Signed-off-by: Masami Hiramatsu <mhiramat at kernel.org>
---
 tools/perf/util/probe-finder.c |   13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
index a5731de0e5eb..c37fbef1711d 100644
--- a/tools/perf/util/probe-finder.c
+++ b/tools/perf/util/probe-finder.c
@@ -423,20 +423,20 @@ static int convert_variable_fields(Dwarf_Die *vr_die, const char *varname,
 		pr_warning("Failed to get the type of %s.\n", varname);
 		return -ENOENT;
 	}
-	pr_debug2("Var real type: (%x)\n", (unsigned)dwarf_dieoffset(&type));
+	pr_debug2("Var real type: %s (%x)\n", dwarf_diename(&type),
+		  (unsigned)dwarf_dieoffset(&type));
 	tag = dwarf_tag(&type);
 
 	if (field->name[0] == '[' &&
 	    (tag == DW_TAG_array_type || tag == DW_TAG_pointer_type)) {
-		if (field->next)
-			/* Save original type for next field */
-			memcpy(die_mem, &type, sizeof(*die_mem));
+		/* Save original type for next field or type */
+		memcpy(die_mem, &type, sizeof(*die_mem));
 		/* Get the type of this array */
 		if (die_get_real_type(&type, &type) == NULL) {
 			pr_warning("Failed to get the type of %s.\n", varname);
 			return -ENOENT;
 		}
-		pr_debug2("Array real type: (%x)\n",
+		pr_debug2("Array real type: %s (%x)\n", dwarf_diename(&type),
 			 (unsigned)dwarf_dieoffset(&type));
 		if (tag == DW_TAG_pointer_type) {
 			ref = zalloc(sizeof(struct probe_trace_arg_ref));
@@ -448,9 +448,6 @@ static int convert_variable_fields(Dwarf_Die *vr_die, const char *varname,
 				*ref_ptr = ref;
 		}
 		ref->offset += dwarf_bytesize(&type) * field->index;
-		if (!field->next)
-			/* Save vr_die for converting types */
-			memcpy(die_mem, vr_die, sizeof(*die_mem));
 		goto next;
 	} else if (tag == DW_TAG_pointer_type) {
 		/* Check the pointer and dereference */

--
To unsubscribe from this list: send the line "unsubscribe linux-kselftest" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

WARNING: multiple messages have this Message-ID (diff)
From: mhiramat@kernel.org (Masami Hiramatsu)
Subject: [PATCH v6 20/21] [RESEND] perf-probe: Fix to convert array type collectly
Date: Sat, 17 Mar 2018 21:52:25 +0900	[thread overview]
Message-ID: <152129114502.31874.2474068470011496356.stgit@devbox> (raw)
Message-ID: <20180317125225.3iJqx5DH_Jv4qkrB_CFhDrUDT5e_IlM3TdhfMANFIOY@z> (raw)
In-Reply-To: <152129024033.31874.15800253385376959274.stgit@devbox>

Fix to convert array element type collectly.
Current perf-probe converts the type of array-elements
incorrectly. It always converts the types as a pointer
of array. This passes the "array" type DIE to the type
converter so that it can get correct "element of array"
type DIE from it.

E.g.
  ====
  $ cat hello.c
  #include <stdio.h>

  void foo(int a[])
  {
	  printf("%d\n", a[1]);
  }

  void main()
  {
	  int a[3] = {4, 5, 6};
	  printf("%d\n", a[0]);
	  foo(a);
  }

  $ gcc -g hello.c -o hello
  $ perf probe -x ./hello -D "foo a[1]"
  ====

Without this fix, above outputs
  ====
  p:probe_hello/foo /tmp/hello:0x4d3 a=+4(-8(%bp)):u64
  ====
The "u64" means "int *", but a[1] is "int".

With this,
  ====
  p:probe_hello/foo /tmp/hello:0x4d3 a=+4(-8(%bp)):s32
  ====
So, "int" correctly converted to "s32"

Fixes: b2a3c12b7442 ("perf probe: Support tracing an entry of array")
Signed-off-by: Masami Hiramatsu <mhiramat at kernel.org>
---
 tools/perf/util/probe-finder.c |   13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
index a5731de0e5eb..c37fbef1711d 100644
--- a/tools/perf/util/probe-finder.c
+++ b/tools/perf/util/probe-finder.c
@@ -423,20 +423,20 @@ static int convert_variable_fields(Dwarf_Die *vr_die, const char *varname,
 		pr_warning("Failed to get the type of %s.\n", varname);
 		return -ENOENT;
 	}
-	pr_debug2("Var real type: (%x)\n", (unsigned)dwarf_dieoffset(&type));
+	pr_debug2("Var real type: %s (%x)\n", dwarf_diename(&type),
+		  (unsigned)dwarf_dieoffset(&type));
 	tag = dwarf_tag(&type);
 
 	if (field->name[0] == '[' &&
 	    (tag == DW_TAG_array_type || tag == DW_TAG_pointer_type)) {
-		if (field->next)
-			/* Save original type for next field */
-			memcpy(die_mem, &type, sizeof(*die_mem));
+		/* Save original type for next field or type */
+		memcpy(die_mem, &type, sizeof(*die_mem));
 		/* Get the type of this array */
 		if (die_get_real_type(&type, &type) == NULL) {
 			pr_warning("Failed to get the type of %s.\n", varname);
 			return -ENOENT;
 		}
-		pr_debug2("Array real type: (%x)\n",
+		pr_debug2("Array real type: %s (%x)\n", dwarf_diename(&type),
 			 (unsigned)dwarf_dieoffset(&type));
 		if (tag == DW_TAG_pointer_type) {
 			ref = zalloc(sizeof(struct probe_trace_arg_ref));
@@ -448,9 +448,6 @@ static int convert_variable_fields(Dwarf_Die *vr_die, const char *varname,
 				*ref_ptr = ref;
 		}
 		ref->offset += dwarf_bytesize(&type) * field->index;
-		if (!field->next)
-			/* Save vr_die for converting types */
-			memcpy(die_mem, vr_die, sizeof(*die_mem));
 		goto next;
 	} else if (tag == DW_TAG_pointer_type) {
 		/* Check the pointer and dereference */

--
To unsubscribe from this list: send the line "unsubscribe linux-kselftest" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

WARNING: multiple messages have this Message-ID (diff)
From: Masami Hiramatsu <mhiramat@kernel.org>
To: Steven Rostedt <rostedt@goodmis.org>, linux-kernel@vger.kernel.org
Cc: mhiramat@kernel.org, Ingo Molnar <mingo@redhat.com>,
	Namhyung Kim <namhyung@kernel.org>,
	Tom Zanussi <tom.zanussi@linux.intel.com>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	linux-trace-users@vger.kernel.org,
	linux-kselftest@vger.kernel.org, shuah@kernel.org,
	Ravi Bangoria <ravi.bangoria@linux.vnet.ibm.com>
Subject: [PATCH v6 20/21] [RESEND] perf-probe: Fix to convert array type collectly
Date: Sat, 17 Mar 2018 21:52:25 +0900	[thread overview]
Message-ID: <152129114502.31874.2474068470011496356.stgit@devbox> (raw)
In-Reply-To: <152129024033.31874.15800253385376959274.stgit@devbox>

Fix to convert array element type collectly.
Current perf-probe converts the type of array-elements
incorrectly. It always converts the types as a pointer
of array. This passes the "array" type DIE to the type
converter so that it can get correct "element of array"
type DIE from it.

E.g.
  ====
  $ cat hello.c
  #include <stdio.h>

  void foo(int a[])
  {
	  printf("%d\n", a[1]);
  }

  void main()
  {
	  int a[3] = {4, 5, 6};
	  printf("%d\n", a[0]);
	  foo(a);
  }

  $ gcc -g hello.c -o hello
  $ perf probe -x ./hello -D "foo a[1]"
  ====

Without this fix, above outputs
  ====
  p:probe_hello/foo /tmp/hello:0x4d3 a=+4(-8(%bp)):u64
  ====
The "u64" means "int *", but a[1] is "int".

With this,
  ====
  p:probe_hello/foo /tmp/hello:0x4d3 a=+4(-8(%bp)):s32
  ====
So, "int" correctly converted to "s32"

Fixes: b2a3c12b7442 ("perf probe: Support tracing an entry of array")
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
---
 tools/perf/util/probe-finder.c |   13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c
index a5731de0e5eb..c37fbef1711d 100644
--- a/tools/perf/util/probe-finder.c
+++ b/tools/perf/util/probe-finder.c
@@ -423,20 +423,20 @@ static int convert_variable_fields(Dwarf_Die *vr_die, const char *varname,
 		pr_warning("Failed to get the type of %s.\n", varname);
 		return -ENOENT;
 	}
-	pr_debug2("Var real type: (%x)\n", (unsigned)dwarf_dieoffset(&type));
+	pr_debug2("Var real type: %s (%x)\n", dwarf_diename(&type),
+		  (unsigned)dwarf_dieoffset(&type));
 	tag = dwarf_tag(&type);
 
 	if (field->name[0] == '[' &&
 	    (tag == DW_TAG_array_type || tag == DW_TAG_pointer_type)) {
-		if (field->next)
-			/* Save original type for next field */
-			memcpy(die_mem, &type, sizeof(*die_mem));
+		/* Save original type for next field or type */
+		memcpy(die_mem, &type, sizeof(*die_mem));
 		/* Get the type of this array */
 		if (die_get_real_type(&type, &type) == NULL) {
 			pr_warning("Failed to get the type of %s.\n", varname);
 			return -ENOENT;
 		}
-		pr_debug2("Array real type: (%x)\n",
+		pr_debug2("Array real type: %s (%x)\n", dwarf_diename(&type),
 			 (unsigned)dwarf_dieoffset(&type));
 		if (tag == DW_TAG_pointer_type) {
 			ref = zalloc(sizeof(struct probe_trace_arg_ref));
@@ -448,9 +448,6 @@ static int convert_variable_fields(Dwarf_Die *vr_die, const char *varname,
 				*ref_ptr = ref;
 		}
 		ref->offset += dwarf_bytesize(&type) * field->index;
-		if (!field->next)
-			/* Save vr_die for converting types */
-			memcpy(die_mem, vr_die, sizeof(*die_mem));
 		goto next;
 	} else if (tag == DW_TAG_pointer_type) {
 		/* Check the pointer and dereference */


  parent reply	other threads:[~2018-03-17 12:52 UTC|newest]

Thread overview: 103+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-17 12:37 [PATCH v6 00/21] tracing: probeevent: Improve fetcharg features mhiramat
2018-03-17 12:37 ` Masami Hiramatsu
2018-03-17 12:37 ` Masami Hiramatsu
2018-03-17 12:38 ` [PATCH v6 01/21] [BUGFIX] tracing: probeevent: Fix to support minus offset from symbol mhiramat
2018-03-17 12:38   ` Masami Hiramatsu
2018-03-17 12:38   ` Masami Hiramatsu
2018-03-17 12:38 ` [PATCH v6 02/21] selftests: ftrace: Add probe event argument syntax testcase mhiramat
2018-03-17 12:38   ` Masami Hiramatsu
2018-03-17 12:38   ` Masami Hiramatsu
2018-03-17 12:39 ` [PATCH v6 03/21] selftests: ftrace: Add a testcase for string type with kprobe_event mhiramat
2018-03-17 12:39   ` Masami Hiramatsu
2018-03-17 12:39   ` Masami Hiramatsu
2018-03-17 12:40 ` [PATCH v6 04/21] selftests: ftrace: Add a testcase for probepoint mhiramat
2018-03-17 12:40   ` Masami Hiramatsu
2018-03-17 12:40   ` Masami Hiramatsu
2018-03-23 16:19   ` rostedt
2018-03-23 16:19     ` Steven Rostedt
2018-03-23 16:19     ` Steven Rostedt
2018-03-17 12:41 ` [PATCH v6 05/21] tracing: probeevent: Cleanup print argument functions mhiramat
2018-03-17 12:41   ` Masami Hiramatsu
2018-03-17 12:41   ` Masami Hiramatsu
2018-03-23 16:36   ` rostedt
2018-03-23 16:36     ` Steven Rostedt
2018-03-23 16:36     ` Steven Rostedt
2018-03-26  4:17     ` mhiramat
2018-03-26  4:17       ` Masami Hiramatsu
2018-03-26  4:17       ` Masami Hiramatsu
2018-03-26 17:28       ` rostedt
2018-03-26 17:28         ` Steven Rostedt
2018-03-26 17:28         ` Steven Rostedt
2018-03-28  4:21         ` mhiramat
2018-03-28  4:21           ` Masami Hiramatsu
2018-03-28  4:21           ` Masami Hiramatsu
2018-03-17 12:41 ` [PATCH v6 06/21] tracing: probeevent: Cleanup argument field definition mhiramat
2018-03-17 12:41   ` Masami Hiramatsu
2018-03-17 12:41   ` Masami Hiramatsu
2018-03-17 12:42 ` [PATCH v6 07/21] tracing: probeevent: Remove NOKPROBE_SYMBOL from print functions mhiramat
2018-03-17 12:42   ` Masami Hiramatsu
2018-03-17 12:42   ` Masami Hiramatsu
2018-03-17 12:43 ` [PATCH v6 08/21] tracing: probeevent: Introduce new argument fetching code mhiramat
2018-03-17 12:43   ` Masami Hiramatsu
2018-03-17 12:43   ` Masami Hiramatsu
2018-03-17 12:44 ` [PATCH v6 09/21] tracing: probeevent: Unify fetch type tables mhiramat
2018-03-17 12:44   ` Masami Hiramatsu
2018-03-17 12:44   ` Masami Hiramatsu
2018-03-17 12:44 ` [PATCH v6 10/21] tracing: probeevent: Return consumed bytes of dynamic area mhiramat
2018-03-17 12:44   ` Masami Hiramatsu
2018-03-17 12:44   ` Masami Hiramatsu
2018-04-02 20:02   ` rostedt
2018-04-02 20:02     ` Steven Rostedt
2018-04-02 20:02     ` Steven Rostedt
2018-04-03 14:41     ` mhiramat
2018-04-03 14:41       ` Masami Hiramatsu
2018-04-03 14:41       ` Masami Hiramatsu
2018-03-17 12:45 ` [PATCH v6 11/21] tracing: probeevent: Append traceprobe_ for exported function mhiramat
2018-03-17 12:45   ` Masami Hiramatsu
2018-03-17 12:45   ` Masami Hiramatsu
2018-03-17 12:46 ` [PATCH v6 12/21] tracing: probeevent: Unify fetch_insn processing common part mhiramat
2018-03-17 12:46   ` Masami Hiramatsu
2018-03-17 12:46   ` Masami Hiramatsu
2018-03-17 12:47 ` [PATCH v6 13/21] tracing: probeevent: Add symbol type mhiramat
2018-03-17 12:47   ` Masami Hiramatsu
2018-03-17 12:47   ` Masami Hiramatsu
2018-03-17 12:47 ` [PATCH v6 14/21] x86: ptrace: Add function argument access API mhiramat
2018-03-17 12:47   ` Masami Hiramatsu
2018-03-17 12:47   ` Masami Hiramatsu
2018-03-17 12:48 ` [PATCH v6 15/21] tracing: probeevent: Add $argN for accessing function args mhiramat
2018-03-17 12:48   ` Masami Hiramatsu
2018-03-17 12:48   ` Masami Hiramatsu
2018-03-17 12:49 ` [PATCH v6 16/21] tracing: probeevent: Add array type support mhiramat
2018-03-17 12:49   ` Masami Hiramatsu
2018-03-17 12:49   ` Masami Hiramatsu
2018-03-17 12:50 ` [PATCH v6 17/21] selftests: ftrace: Add a testcase for symbol type mhiramat
2018-03-17 12:50   ` Masami Hiramatsu
2018-03-17 12:50   ` Masami Hiramatsu
2018-03-17 12:50 ` [PATCH v6 18/21] selftests: ftrace: Add a testcase for $argN with kprobe_event mhiramat
2018-03-17 12:50   ` Masami Hiramatsu
2018-03-17 12:50   ` Masami Hiramatsu
2018-03-17 12:51 ` [PATCH v6 19/21] selftests: ftrace: Add a testcase for array type " mhiramat
2018-03-17 12:51   ` Masami Hiramatsu
2018-03-17 12:51   ` Masami Hiramatsu
2018-03-17 12:52 ` mhiramat [this message]
2018-03-17 12:52   ` [PATCH v6 20/21] [RESEND] perf-probe: Fix to convert array type collectly Masami Hiramatsu
2018-03-17 12:52   ` Masami Hiramatsu
2018-03-20  6:36   ` [tip:perf/core] perf probe: Use right type to access array elements tip-bot for Masami Hiramatsu
2018-03-17 12:53 ` [PATCH v6 21/21] perf-probe: Add array argument support mhiramat
2018-03-17 12:53   ` Masami Hiramatsu
2018-03-17 12:53   ` Masami Hiramatsu
2018-03-19  7:59   ` ravi.bangoria
2018-03-19  7:59     ` Ravi Bangoria
2018-03-19  7:59     ` Ravi Bangoria
2018-03-22 10:23     ` mhiramat
2018-03-22 10:23       ` Masami Hiramatsu
2018-03-22 10:23       ` Masami Hiramatsu
2018-03-22 10:49       ` ravi.bangoria
2018-03-22 10:49         ` Ravi Bangoria
2018-03-22 10:49         ` Ravi Bangoria
2018-03-26  3:53         ` mhiramat
2018-03-26  3:53           ` Masami Hiramatsu
2018-03-26  3:53           ` Masami Hiramatsu
2018-03-17 14:06 ` [PATCH v6 00/21] tracing: probeevent: Improve fetcharg features mhiramat
2018-03-17 14:06   ` Masami Hiramatsu
2018-03-17 14:06   ` Masami Hiramatsu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=152129114502.31874.2474068470011496356.stgit@devbox \
    --to=unknown@example.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.