public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
* [PATCH] [gdb.base] Convert multi-line function call into in foll-exec.c
@ 2021-06-03 14:57 Kumar N, Bhuvanendra
  2021-06-03 15:38 ` Simon Marchi
  0 siblings, 1 reply; 5+ messages in thread
From: Kumar N, Bhuvanendra @ 2021-06-03 14:57 UTC (permalink / raw)
  To: Simon Marchi, gdb-patches
  Cc: George, Jini Susan, Achra, Nitika, Sharma, Alok Kumar, E,
	Nagajyothi, Tomar, Sourabh Singh

[-- Attachment #1: Type: text/plain, Size: 6668 bytes --]

[AMD Official Use Only - Internal Distribution Only]

Hi Simon and all,

Could you please review this GDB patch. Patch is attached with this email and also inlined here. Thanks in advance

Problem Description:
Following 8 test points started to fail after the clang backend patch(https://reviews.llvm.org/D91734).

FAIL: gdb.base/foll-exec.exp: step through execlp call
FAIL: gdb.base/foll-exec.exp: step after execlp call
FAIL: gdb.base/foll-exec.exp: print execd-program/global_i (after execlp)
FAIL: gdb.base/foll-exec.exp: print execd-program/local_j (after execlp)
FAIL: gdb.base/foll-exec.exp: print follow-exec/local_k (after execlp)
FAIL: gdb.base/foll-exec.exp: step through execl call
FAIL: gdb.base/foll-exec.exp: step after execl call
FAIL: gdb.base/foll-exec.exp: print execd-program/local_j (after execl)

Sample gdb.base/foll-exec.c test file snippet :
. . .
42   execlp (prog, /* tbreak-execlp */
43           prog,
44           "execlp arg1 from foll-exec",
45           (char *) 0);
46
47   printf ("foll-exec is about to execl(execd-prog)...\n");
. . .

Resolution:

We had discussed 2 ways of handling this issue earlier, i.e.


  1.  Adding few extra "next" command during these multi-line function call
  2.  combine these multi-line function call into single line in foll-exec.c


This PR has taken the latter approach and converted the multi-line function call into single line in foll-exec.c, thereby there is no change in .debug_line entries now and test case works as expected.

NOTE: earlier I had sent a test patch(sent on 2021-04-15) with first approach and it was suggested to follow the second approach, hence now I am sending this PR with the second approach

regards,
bhuvan

Patch content inlined:

From 77e09b9b9a19dfb35a748fd57d7a16eb52f1b6ee Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E2=80=9Cbhkumarn=E2=80=9D?= Bhuvanendra.KumarN@amd.com<mailto:Bhuvanendra.KumarN@amd.com>
Date: Thu, 3 Jun 2021 17:50:28 +0530
Subject: [PATCH] [gdb.base] Convert multi-line function call into
single line.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

After this clang backend patch(https://reviews.llvm.org/D91734), 8 test points
started to FAIL in this test case. As mentioned in this PR, "...this test is
trying to "next" over a function call; gcc attributes all parameter evaluation
to the function name, while clang will attribute each parameter to its own
location. And when the parameters span across multiple source lines, the
is_stmt heuristic kicks in, so we stop on each line with actual parameters...".

gdb.base/foll-exec.c test file snippet :
. . .
42   execlp (prog, /* tbreak-execlp */
43           prog,
44           "execlp arg1 from foll-exec",
45           (char *) 0);
46
47   printf ("foll-exec is about to execl(execd-prog)...\n");
. . .

Line table: (before clang backend patch for the above code snippet) :
0x000000b0: 84 address += 8,  line += 2
            0x000000000020196a     42      3      1   0             0
0x000000b1: 08 DW_LNS_const_add_pc (0x0000000000000011)
0x000000b2: 41 address += 3,  line += 5
            0x000000000020197e     47      3      1   0             0

Line table: (after clang backend patch for the above code snippet) :
0x000000b5: 84 address += 8,  line += 2
            0x0000000000201958     42     11      1   0             0
0x000000b6: 05 DW_LNS_set_column (4)
0x000000b8: 75 address += 7,  line += 1
            0x000000000020195f     43      4      1   0             0
0x000000b9: 05 DW_LNS_set_column (3)
0x000000bb: 73 address += 7,  line += -1
            0x0000000000201966     42      3      1   0             0
0x000000bc: 08 DW_LNS_const_add_pc (0x0000000000000011)
0x000000bd: 4f address += 4,  line += 5
            0x000000000020197b     47      3      1   0             0

Following 8 test points started to fail after the above clang backend patch.

FAIL: gdb.base/foll-exec.exp: step through execlp call
FAIL: gdb.base/foll-exec.exp: step after execlp call
FAIL: gdb.base/foll-exec.exp: print execd-program/global_i (after execlp)
FAIL: gdb.base/foll-exec.exp: print execd-program/local_j (after execlp)
FAIL: gdb.base/foll-exec.exp: print follow-exec/local_k (after execlp)
FAIL: gdb.base/foll-exec.exp: step through execl call
FAIL: gdb.base/foll-exec.exp: step after execl call
FAIL: gdb.base/foll-exec.exp: print execd-program/local_j (after execl)

As we can note, reason for these new test failures is due to additional
.debug_line entries getting created in case of clang compiler, hence to fix
this issue, test case required either additional "next" command during
these multi-line function call or combine these multi-line function call into
single line. This PR has taken the latter approach and converted the multi-line
function call into single line in foll-exec.c, thereby there is no change in
.debug_line entries now and test case works as expected.
---
gdb/testsuite/ChangeLog            |  5 +++++
gdb/testsuite/gdb.base/foll-exec.c | 11 ++---------
2 files changed, 7 insertions(+), 9 deletions(-)

diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 7959f58c3c4..1873f9ddf05 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2021-06-03  Bhuvanendra Kumar N  Bhuvanendra.KumarN@amd.com<mailto:Bhuvanendra.KumarN@amd.com>
+
+          * gdb.base/foll-exec.c: convert the multi-line function call into
+          single line.
+
2021-06-02  Bernd Edlinger  bernd.edlinger@hotmail.de<mailto:bernd.edlinger@hotmail.de>
            * gdb.dwarf2/per-bfd-sharing.exp: Fix temp-dir leakage.
diff --git a/gdb/testsuite/gdb.base/foll-exec.c b/gdb/testsuite/gdb.base/foll-exec.c
index fea62b06db4..f1b97aca4a4 100644
--- a/gdb/testsuite/gdb.base/foll-exec.c
+++ b/gdb/testsuite/gdb.base/foll-exec.c
@@ -39,18 +39,11 @@ int main (int argc, char ** argv)
   memcpy (prog + len - 9, "execd-prog", 10);
   prog[len + 1] = 0;
-  execlp (prog, /* tbreak-execlp */
-             prog,
-             "execlp arg1 from foll-exec",
-             (char *) 0);
+  execlp (prog, /* tbreak-execlp */ prog, "execlp arg1 from foll-exec", (char *) 0);
   printf ("foll-exec is about to execl(execd-prog)...\n");
-  execl (prog,  /* tbreak-execl */
-           prog,
-           "execl arg1 from foll-exec",
-           "execl arg2 from foll-exec",
-           (char *) 0);
+  execl (prog, /* tbreak-execl */ prog, "execl arg1 from foll-exec", "execl arg2 from foll-exec", (char *) 0);
   {
     static char * argv[] = {
--
2.17.1


[-- Attachment #2: gdb.base-Convert-multi-line-function-call-into.patch --]
[-- Type: application/octet-stream, Size: 4726 bytes --]

From 77e09b9b9a19dfb35a748fd57d7a16eb52f1b6ee Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E2=80=9Cbhkumarn=E2=80=9D?= <Bhuvanendra.KumarN@amd.com>
Date: Thu, 3 Jun 2021 17:50:28 +0530
Subject: [PATCH] [gdb.base] Convert multi-line function call into
 single line.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

After this clang backend patch(https://reviews.llvm.org/D91734), 8 test points
started to FAIL in this test case. As mentioned in this PR, "...this test is
trying to "next" over a function call; gcc attributes all parameter evaluation
to the function name, while clang will attribute each parameter to its own
location. And when the parameters span across multiple source lines, the
is_stmt heuristic kicks in, so we stop on each line with actual parameters...".

gdb.base/foll-exec.c test file snippet :
. . .
 42   execlp (prog, /* tbreak-execlp */
 43           prog,
 44           "execlp arg1 from foll-exec",
 45           (char *) 0);
 46
 47   printf ("foll-exec is about to execl(execd-prog)...\n");
. . .

Line table: (before clang backend patch for the above code snippet) :
0x000000b0: 84 address += 8,  line += 2
            0x000000000020196a     42      3      1   0             0
0x000000b1: 08 DW_LNS_const_add_pc (0x0000000000000011)
0x000000b2: 41 address += 3,  line += 5
            0x000000000020197e     47      3      1   0             0

Line table: (after clang backend patch for the above code snippet) :
0x000000b5: 84 address += 8,  line += 2
            0x0000000000201958     42     11      1   0             0
0x000000b6: 05 DW_LNS_set_column (4)
0x000000b8: 75 address += 7,  line += 1
            0x000000000020195f     43      4      1   0             0
0x000000b9: 05 DW_LNS_set_column (3)
0x000000bb: 73 address += 7,  line += -1
            0x0000000000201966     42      3      1   0             0
0x000000bc: 08 DW_LNS_const_add_pc (0x0000000000000011)
0x000000bd: 4f address += 4,  line += 5
            0x000000000020197b     47      3      1   0             0

Following 8 test points started to fail after the above clang backend patch.

FAIL: gdb.base/foll-exec.exp: step through execlp call
FAIL: gdb.base/foll-exec.exp: step after execlp call
FAIL: gdb.base/foll-exec.exp: print execd-program/global_i (after execlp)
FAIL: gdb.base/foll-exec.exp: print execd-program/local_j (after execlp)
FAIL: gdb.base/foll-exec.exp: print follow-exec/local_k (after execlp)
FAIL: gdb.base/foll-exec.exp: step through execl call
FAIL: gdb.base/foll-exec.exp: step after execl call
FAIL: gdb.base/foll-exec.exp: print execd-program/local_j (after execl)

As we can note, reason for these new test failures is due to additional
.debug_line entries getting created in case of clang compiler, hence to fix
this issue, test case required either additional “next” command during
these multi-line function call or combine these multi-line function call into
single line. This PR has taken the latter approach and converted the multi-line
function call into single line in foll-exec.c, thereby there is no change in
.debug_line entries now and test case works as expected.
---
 gdb/testsuite/ChangeLog            |  5 +++++
 gdb/testsuite/gdb.base/foll-exec.c | 11 ++---------
 2 files changed, 7 insertions(+), 9 deletions(-)

diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 7959f58c3c4..1873f9ddf05 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2021-06-03  Bhuvanendra Kumar N  <Bhuvanendra.KumarN@amd.com>
+
+	* gdb.base/foll-exec.c: convert the multi-line function call into
+	single line.
+
 2021-06-02  Bernd Edlinger  <bernd.edlinger@hotmail.de>
 
 	* gdb.dwarf2/per-bfd-sharing.exp: Fix temp-dir leakage.
diff --git a/gdb/testsuite/gdb.base/foll-exec.c b/gdb/testsuite/gdb.base/foll-exec.c
index fea62b06db4..f1b97aca4a4 100644
--- a/gdb/testsuite/gdb.base/foll-exec.c
+++ b/gdb/testsuite/gdb.base/foll-exec.c
@@ -39,18 +39,11 @@ int main (int argc, char ** argv)
   memcpy (prog + len - 9, "execd-prog", 10);
   prog[len + 1] = 0;
 
-  execlp (prog, /* tbreak-execlp */
-	  prog,
-	  "execlp arg1 from foll-exec",
-	  (char *) 0);
+  execlp (prog, /* tbreak-execlp */ prog, "execlp arg1 from foll-exec", (char *) 0);
 
   printf ("foll-exec is about to execl(execd-prog)...\n");
 
-  execl (prog,	/* tbreak-execl */
-	 prog,
-	 "execl arg1 from foll-exec",
-	 "execl arg2 from foll-exec",
-	 (char *) 0);
+  execl (prog,	/* tbreak-execl */ prog, "execl arg1 from foll-exec", "execl arg2 from foll-exec", (char *) 0);
 
   {
     static char * argv[] = {
-- 
2.17.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2021-06-07 15:19 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-06-03 14:57 [PATCH] [gdb.base] Convert multi-line function call into in foll-exec.c Kumar N, Bhuvanendra
2021-06-03 15:38 ` Simon Marchi
2021-06-07  6:31   ` Kumar N, Bhuvanendra
2021-06-07 15:07     ` Simon Marchi
2021-06-07 15:19       ` Kumar N, Bhuvanendra

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).