* [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table.
@ 2023-08-25 15:50 Carl Love
2023-08-25 15:53 ` [PATCH 1/2] Add gdb_compile options column-info and no-column-info Carl Love
` (2 more replies)
0 siblings, 3 replies; 8+ messages in thread
From: Carl Love @ 2023-08-25 15:50 UTC (permalink / raw)
To: Simon Marchi, blarsen, gdb-patches, UlrichWeigand, pedro
Cc: luis.machado, cel
GDB maintainers:
I received a few messages from Guinevere about these two patches. In
the end, we agreed the first patch was OK. There was a minor code
change suggested for the second patch that I have implemented. Since I
have not received any additional suggestions on the second patch, it is
time to ping the first patch and post the updated second patch for
review.
The first patch in the series has not been changed. I am reposting it
along with the second patch so it doesn't get lost.
The patch series has been tested on Power10 LE and on X86-64.
Please let me know if these two patches are acceptable for mainline.
Thanks.
Carl
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH 1/2] Add gdb_compile options column-info and no-column-info
2023-08-25 15:50 [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table Carl Love
@ 2023-08-25 15:53 ` Carl Love
2023-08-25 15:57 ` [PATCH 2/2 ver 8] Fix reverse stepping multiple contiguous PC ranges over the line table Carl Love
2023-08-29 12:28 ` [PATCH 0/2] " Guinevere Larsen
2 siblings, 0 replies; 8+ messages in thread
From: Carl Love @ 2023-08-25 15:53 UTC (permalink / raw)
To: Simon Marchi, blarsen, gdb-patches, UlrichWeigand, pedro; +Cc: luis.machado
Simon, GDB maintainers:
Version 2, updated the compiler check and handling for gcc version 6
and earlier. Retested on Power 10.
Per the comments on version 4 for the gdb.reverse/func-map-to-same-
line.exp, I have added support to proc gdb_compile to enable or disable
generating line information as part of the debug information. The two
new options are column-info and no-column-info.
This patch implements the new options for gdb_compile.
These options have been tested with patch 2 of 2 on PowerPC with the
GCC and clang compilers.
Please let me know if the patch is acceptable for mainline. Thanks.
Carl
------------------------------------------------------
Add gdb_compile options column-info and no-column-info
This patch adds two new options to gdb_compile to specify if the compile
should or should not generate the line table information. The
options are supported on clang and gcc version 7 and newer.
Patch has been tested on PowerPC with both gcc and clang.
---
gdb/testsuite/lib/gdb.exp | 34 ++++++++++++++++++++++++++++++++++
1 file changed, 34 insertions(+)
diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
index 8432e636ecc..e321e4e5ecc 100644
--- a/gdb/testsuite/lib/gdb.exp
+++ b/gdb/testsuite/lib/gdb.exp
@@ -4896,6 +4896,8 @@ proc quote_for_host { args } {
# debug information
# - text_segment=addr: Tell the linker to place the text segment at ADDR.
# - build-id: Ensure the final binary includes a build-id.
+# - no-column-info: Disable generation of column table information.
+# - column-info: Enable generation of column table information.
#
# And here are some of the not too obscure options understood by DejaGnu that
# influence the compilation:
@@ -5105,6 +5107,38 @@ proc gdb_compile {source dest type options} {
} else {
error "Don't know how to handle text_segment option."
}
+ } elseif { $opt == "column-info" } {
+ # If GCC or clang does not support column-info, compilation
+ # will fail and the usupported column-info option will be
+ # reported as such.
+ if {[test_compiler_info {gcc-*}]} {
+ lappend new_options "additional_flags=-gcolumn-info"
+
+ } elseif {[test_compiler_info {clang-*}]} {
+ lappend new_options "additional_flags=-gcolumn-info"
+
+ } else {
+ error "Don't know how to handle gcolumn-info option."
+ }
+
+ } elseif { $opt == "no-column-info" } {
+ if {[test_compiler_info {gcc-*}]} {
+ if {[test_compiler_info {gcc-[1-6]-*}]} {
+ # In this case, don't add the compile line option and
+ # the result will be the same as using no-column-info
+ # on a version that supports the option.
+ warning "gdb_compile option no-column-info not supported, ignoring."
+ } else {
+ lappend new_options "additional_flags=-gno-column-info"
+ }
+
+ } elseif {[test_compiler_info {clang-*}]} {
+ lappend new_options "additional_flags=-gno-column-info"
+
+ } else {
+ error "Don't know how to handle gno-column-info option."
+ }
+
} else {
lappend new_options $opt
}
--
2.37.2
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH 2/2 ver 8] Fix reverse stepping multiple contiguous PC ranges over the line table.
2023-08-25 15:50 [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table Carl Love
2023-08-25 15:53 ` [PATCH 1/2] Add gdb_compile options column-info and no-column-info Carl Love
@ 2023-08-25 15:57 ` Carl Love
2023-08-29 12:28 ` [PATCH 0/2] " Guinevere Larsen
2 siblings, 0 replies; 8+ messages in thread
From: Carl Love @ 2023-08-25 15:57 UTC (permalink / raw)
To: Simon Marchi, blarsen, gdb-patches, UlrichWeigand, pedro
Cc: luis.machado, cel
Guinevere, Simon, GDB maintainers:
I received a minor comment from Guinevere on the patch. I made the
suggested code change and was waiting to see if there were any other
comments. Haven't received any. So am posting the updated patch with
the change.
Version 8, made to code change in process_event_stop_test to set
ecs->event_thread->control.step_range_start from the call to
update_line_range_start (stop_pc, ecs) suggested by
Guinevere. Retested patch on PowerPC and Intel X86.
Version 7, addressed behavior of GDB when stepping backward thru a
function call as mentioned by Bruno. GDB would stop at the function
call and then a reverse step/next would cause GDB to stop at the first
instruction of the same line as the function call instead of stopping
in the previous line. The behavior was fixed and the various test
programs were updated to remove one of the two reverse step or next
instructions required to reach the previous line.
Version 6, fixed various code style issues in the GDB source. The
testcases were updated to use with_test_prefix for each gdb test in the
step and next test cases, switch using the standard_testfile, use
foreach_with_prefix to combine otherwise identical tests. Retested on
Power 10.
Version 5, changed comments in test case func-map-to-same-line.c.
Patch 1/2 implemented the new options for gdb_compile. Updated the
call to proc run_tests to use the new gdb_compile options in a
foreach_with_prefix loop.
Version 4, additional fixes for gcc version check, wrap function calls
using "with_test_prefix", move load_lib dwarf.exe. Fixed typo noted by
Luis.
Version 3, added the gcc version check as discussed further from
version 2 of the patch. Also updated the tests to check for supporting
reverse execution rather than requiring recording. I also noticed
there were a couple more instances of a requirement check, i.e. if []
which I changed to "require" per the current style for checking on the
test requirements.
The following patch fixes issues on PowerPC with the reverse-step and
reverse-next instructions when there are multiple assignment statements
on the same line and when there are multiple function calls on the same
line. The commit log below discusses these issues in further depth.
The discussion included what the correct operation should be for these
commands based on the GDB documentation. The proposed patch at that
time changed how the commands worked on other platforms such as X86 in
a way they no longer matched the documentation.
The issue is the line table contains multiple entries for the same
source line. The patch adds a function to search the line table to
find the address of the first instruction of a line. When setup up the
reverse stepping range, the function is called to make sure the start
of the range corresponds to the address of the first instruction for
the line. This approach was used. When Luis initially developed the
patch, he considered merging the contiguous ranges in the line table
when reading the line tables. He decided it was better to work with the
data directly in the line table rather than creating and using a
modified version of the line table.
The following patch fixes the execution of the reveres-step and
reverse-next commands for both senarios of multiple statements on the
same line for PowerPC and aarch64-linux. Unlike the previous patch, it
does not change the operation of the commands on other platforms, i.e.
X86. The patch adds new test cases for both scenarios to verify they
work correctly.
The patch has been tested on PowerPC, Intel X86 and aarch64-linux with
no new regression failures.
Please let me know if the patch is acceptable for mainline. Thanks.
Carl
--------------------------------
Fix reverse stepping multiple contiguous PC ranges over the line table.
There are a couple of scenarios where the GDB reverse-step and reverse-
next
commands do not work correctly.
Scenario 1 issue description by Luis Machado:
When running GDB's testsuite on aarch64-linux/Ubuntu 20.04 (also
spotted on
the ppc backend), I noticed some failures in gdb.reverse/solib-
precsave.exp
and gdb.reverse/solib-reverse.exp.
The failure happens around the following code:
38 b[1] = shr2(17); /* middle part two */
40 b[0] = 6; b[1] = 9; /* generic statement, end part two */
42 shr1 ("message 1\n"); /* shr1 one */
Normal execution:
- step from line 38 will land on line 40.
- step from line 40 will land on line 42.
Reverse execution:
- step from line 42 will land on line 40.
- step from line 40 will land on line 40.
- step from line 40 will land on line 38.
V
The problem here is that line 40 contains two contiguous but distinct
PC ranges in the line table, like so:
Line 40 - [0x7ec ~ 0x7f4]
Line 40 - [0x7f4 ~ 0x7fc]
The two distinct ranges are generated because GCC started outputting
source
column information, which GDB doesn't take into account at the moment.
When stepping forward from line 40, we skip both of these ranges and
land on
line 42. When stepping backward from line 42, we stop at the start PC
of the
second (or first, going backwards) range of line 40.
Since we've reached ecs->event_thread->control.step_range_start, we
stop
stepping backwards.
---------------------------------------------------------
Scenario 2 issue described by Pedro Alves:
The following explanation of the issue was taken from the gdb mailing
list
discussion of the withdrawn patch to change the behavior of the
reverse-step
and reverse-next commands. Specifically, message from Pedro Alves
<pedro@palves.net> where he demonstrates the issue where you have
multiple
function calls on the same source code line:
https://sourceware.org/pipermail/gdb-patches/2023-January/196110.html
The source line looks like:
func1 (); func2 ();
so stepping backwards over that line should always stop at the first
instruction of the line, not in the middle. Let's simplify this.
Here's the full source code of my example:
(gdb) list 1
1 void func1 ()
2 {
3 }
4
5 void func2 ()
6 {
7 }
8
9 int main ()
10 {
11 func1 (); func2 ();
12 }
Compiled with:
$ gcc reverse.c -o reverse -g3 -O0
$ gcc -v
...
gcc version 11.3.0 (Ubuntu 11.3.0-1ubuntu1~22.04)
Now let's debug it with target record, using current gdb git master
(f3d8ae90b236),
without your patch:
$ gdb ~/reverse
GNU gdb (GDB) 14.0.50.20230124-git
...
Reading symbols from /home/pedro/reverse...
(gdb) start
Temporary breakpoint 1 at 0x1147: file reverse.c, line 11.
Starting program: /home/pedro/reverse
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-
gnu/libthread_db.so.1".
Temporary breakpoint 1, main () at reverse.c:11
11 func1 (); func2 ();
(gdb) record
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
=> 0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
(gdb) n
12 }
So far so good, a "next" stepped over the whole of line 11 and stopped
at line 12.
Let's confirm where we are now:
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
=> 0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
Good, we're at the first instruction of line 12.
Now let's undo the "next", with "reverse-next":
(gdb) reverse-next
11 func1 (); func2 ();
Seemingly stopped at line 11. Let's see exactly where:
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
=> 0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
(gdb)
And lo, we stopped in the middle of line 11! That is a bug, we should
have
stepped back all the way to the beginning of the line. The "reverse-
next"
should have fully undone the prior "next" command.
The above issues were fixed by introducing a new function that looks
for
adjacent PC ranges for the same line, until we notice a line change.
Then
we take that as the start PC of the range. The new start PC for the
range
is used for the control.step_range_start when setting up a step range.
The test case gdb.reverse/map-to-same-line.exp is added to test the fix
for the issues in scenario 1.
The test case gdb.reverse/func-map-to-same-line.exp was added to test
the
fix for scenario 2 when the binary was compiled with and without line
table information.
bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28426
Co-authored-by: Luis Machado <luis.machado@arm.com>
Co-authored-by: Carl Love <cel@us.ibm.com>
------------------------------
Fix reverse stepping multiple contiguous PC ranges over the line table.
There are a couple of scenarios where the GDB reverse-step and reverse-next
commands do not work correctly.
Scenario 1 issue description by Luis Machado:
When running GDB's testsuite on aarch64-linux/Ubuntu 20.04 (also spotted on
the ppc backend), I noticed some failures in gdb.reverse/solib-precsave.exp
and gdb.reverse/solib-reverse.exp.
The failure happens around the following code:
38 b[1] = shr2(17); /* middle part two */
40 b[0] = 6; b[1] = 9; /* generic statement, end part two */
42 shr1 ("message 1\n"); /* shr1 one */
Normal execution:
- step from line 38 will land on line 40.
- step from line 40 will land on line 42.
Reverse execution:
- step from line 42 will land on line 40.
- step from line 40 will land on line 40.
- step from line 40 will land on line 38.
V
The problem here is that line 40 contains two contiguous but distinct
PC ranges in the line table, like so:
Line 40 - [0x7ec ~ 0x7f4]
Line 40 - [0x7f4 ~ 0x7fc]
The two distinct ranges are generated because GCC started outputting source
column information, which GDB doesn't take into account at the moment.
When stepping forward from line 40, we skip both of these ranges and land on
line 42. When stepping backward from line 42, we stop at the start PC of the
second (or first, going backwards) range of line 40.
Since we've reached ecs->event_thread->control.step_range_start, we stop
stepping backwards.
---------------------------------------------------------
Scenario 2 issue described by Pedro Alves:
The following explanation of the issue was taken from the gdb mailing list
discussion of the withdrawn patch to change the behavior of the reverse-step
and reverse-next commands. Specifically, message from Pedro Alves
<pedro@palves.net> where he demonstrates the issue where you have multiple
function calls on the same source code line:
https://sourceware.org/pipermail/gdb-patches/2023-January/196110.html
The source line looks like:
func1 (); func2 ();
so stepping backwards over that line should always stop at the first
instruction of the line, not in the middle. Let's simplify this.
Here's the full source code of my example:
(gdb) list 1
1 void func1 ()
2 {
3 }
4
5 void func2 ()
6 {
7 }
8
9 int main ()
10 {
11 func1 (); func2 ();
12 }
Compiled with:
$ gcc reverse.c -o reverse -g3 -O0
$ gcc -v
...
gcc version 11.3.0 (Ubuntu 11.3.0-1ubuntu1~22.04)
Now let's debug it with target record, using current gdb git master (f3d8ae90b236),
without your patch:
$ gdb ~/reverse
GNU gdb (GDB) 14.0.50.20230124-git
...
Reading symbols from /home/pedro/reverse...
(gdb) start
Temporary breakpoint 1 at 0x1147: file reverse.c, line 11.
Starting program: /home/pedro/reverse
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Temporary breakpoint 1, main () at reverse.c:11
11 func1 (); func2 ();
(gdb) record
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
=> 0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
(gdb) n
12 }
So far so good, a "next" stepped over the whole of line 11 and stopped at line 12.
Let's confirm where we are now:
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
=> 0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
Good, we're at the first instruction of line 12.
Now let's undo the "next", with "reverse-next":
(gdb) reverse-next
11 func1 (); func2 ();
Seemingly stopped at line 11. Let's see exactly where:
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
=> 0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
(gdb)
And lo, we stopped in the middle of line 11! That is a bug, we should have
stepped back all the way to the beginning of the line. The "reverse-next"
should have fully undone the prior "next" command.
The above issues were fixed by introducing a new function that looks for
adjacent PC ranges for the same line, until we notice a line change. Then
we take that as the start PC of the range. The new start PC for the range
is used for the control.step_range_start when setting up a step range.
The test case gdb.reverse/map-to-same-line.exp is added to test the fix
for the issues in scenario 1.
The test case gdb.reverse/func-map-to-same-line.exp was added to test the
fix for scenario 2 when the binary was compiled with and without line
table information.
bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28426
Co-authored-by: Luis Machado <luis.machado@arm.com>
Co-authored-by: Carl Love <cel@us.ibm.com>
---
gdb/infcmd.c | 13 ++
gdb/infrun.c | 57 +++++++
gdb/symtab.c | 49 ++++++
gdb/symtab.h | 16 ++
gdb/testsuite/gdb.mi/mi-reverse.exp | 5 +-
.../gdb.reverse/finish-reverse-next.exp | 42 ++---
.../gdb.reverse/func-map-to-same-line.c | 37 +++++
.../gdb.reverse/func-map-to-same-line.exp | 139 ++++++++++++++++
gdb/testsuite/gdb.reverse/map-to-same-line.c | 58 +++++++
.../gdb.reverse/map-to-same-line.exp | 153 ++++++++++++++++++
10 files changed, 535 insertions(+), 34 deletions(-)
create mode 100644 gdb/testsuite/gdb.reverse/func-map-to-same-line.c
create mode 100644 gdb/testsuite/gdb.reverse/func-map-to-same-line.exp
create mode 100644 gdb/testsuite/gdb.reverse/map-to-same-line.c
create mode 100644 gdb/testsuite/gdb.reverse/map-to-same-line.exp
diff --git a/gdb/infcmd.c b/gdb/infcmd.c
index fd85d27466a..42dff4e43f9 100644
--- a/gdb/infcmd.c
+++ b/gdb/infcmd.c
@@ -975,6 +975,19 @@ prepare_one_step (thread_info *tp, struct step_command_fsm *sm)
&tp->control.step_range_start,
&tp->control.step_range_end);
+ if (execution_direction == EXEC_REVERSE)
+ {
+ symtab_and_line sal = find_pc_line (pc, 0);
+ symtab_and_line sal_start
+ = find_pc_line (tp->control.step_range_start, 0);
+
+ if (sal.line == sal_start.line)
+ /* Executing in reverse, the step_range_start address is in
+ the same line. We want to stop in the previous line so
+ move step_range_start before the current line. */
+ tp->control.step_range_start--;
+ }
+
/* There's a problem in gcc (PR gcc/98780) that causes missing line
table entries, which results in a too large stepping range.
Use inlined_subroutine info to make the range more narrow. */
diff --git a/gdb/infrun.c b/gdb/infrun.c
index 4730d290442..5132ebb6647 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -115,6 +115,9 @@ static struct async_event_handler *infrun_async_inferior_event_token;
Starts off as -1, indicating "never enabled/disabled". */
static int infrun_is_async = -1;
+static CORE_ADDR update_line_range_start (CORE_ADDR pc,
+ struct execution_control_state *ecs);
+
/* See infrun.h. */
void
@@ -6886,6 +6889,27 @@ handle_signal_stop (struct execution_control_state *ecs)
process_event_stop_test (ecs);
}
+/* Return the address for the beginning of the line. */
+
+CORE_ADDR
+update_line_range_start (CORE_ADDR pc, struct execution_control_state *ecs)
+{
+ /* The line table may have multiple entries for the same source code line.
+ Given the PC, check the line table and return the PC that corresponds
+ to the line table entry for the source line that PC is in. */
+ CORE_ADDR start_line_pc = ecs->event_thread->control.step_range_start;
+ gdb::optional<CORE_ADDR> real_range_start;
+
+ /* Call find_line_range_start to get the smallest address in the
+ linetable for multiple Line X entries in the line table. */
+ real_range_start = find_line_range_start (pc);
+
+ if (real_range_start.has_value ())
+ start_line_pc = *real_range_start;
+
+ return start_line_pc;
+}
+
/* Come here when we've got some debug event / signal we can explain
(IOW, not a random signal), and test whether it should cause a
stop, or whether we should resume the inferior (transparently).
@@ -7687,6 +7711,28 @@ process_event_stop_test (struct execution_control_state *ecs)
if (stop_pc_sal.is_stmt)
{
+ if (execution_direction == EXEC_REVERSE)
+ {
+ /* We are stepping backwards make sure we have reached the
+ beginning of the line. */
+ CORE_ADDR stop_pc = ecs->event_thread->stop_pc ();
+ CORE_ADDR start_line_pc
+ = update_line_range_start (stop_pc, ecs);
+
+ if (stop_pc != start_line_pc)
+ {
+ /* Have not reached the beginning of the source code line.
+ Set a step range. Execution should stop in any function
+ calls we execute back into before reaching the beginning
+ of the line. */
+ ecs->event_thread->control.step_range_start = start_line_pc;
+ ecs->event_thread->control.step_range_end = stop_pc;
+ set_step_info (ecs->event_thread, frame, stop_pc_sal);
+ keep_going (ecs);
+ return;
+ }
+ }
+
/* We are at the start of a statement.
So stop. Note that we don't stop if we step into the middle of a
@@ -7749,6 +7795,17 @@ process_event_stop_test (struct execution_control_state *ecs)
set_step_info (ecs->event_thread, frame, stop_pc_sal);
infrun_debug_printf ("keep going");
+
+ if (execution_direction == EXEC_REVERSE)
+ {
+ CORE_ADDR stop_pc = ecs->event_thread->stop_pc ();
+
+ /* Make sure the stop_pc is set to the beginning of the line. */
+ if (stop_pc != ecs->event_thread->control.step_range_start)
+ ecs->event_thread->control.step_range_start
+ = update_line_range_start (stop_pc, ecs);
+ }
+
keep_going (ecs);
}
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 0117a2a59d7..c38bb01b747 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -3284,6 +3284,55 @@ find_pc_line (CORE_ADDR pc, int notcurrent)
return sal;
}
+/* Compare two symtab_and_line entries. Return true if both have
+ the same line number and the same symtab pointer. That means we
+ are dealing with two entries from the same line and from the same
+ source file.
+
+ Return false otherwise. */
+
+static bool
+sal_line_symtab_matches_p (const symtab_and_line &sal1,
+ const symtab_and_line &sal2)
+{
+ return sal1.line == sal2.line && sal1.symtab == sal2.symtab;
+}
+
+/* See symtah.h. */
+
+gdb::optional<CORE_ADDR>
+find_line_range_start (CORE_ADDR pc)
+{
+ struct symtab_and_line current_sal = find_pc_line (pc, 0);
+
+ if (current_sal.line == 0)
+ return {};
+
+ struct symtab_and_line prev_sal = find_pc_line (current_sal.pc - 1, 0);
+
+ /* If the previous entry is for a different line, that means we are already
+ at the entry with the start PC for this line. */
+ if (!sal_line_symtab_matches_p (prev_sal, current_sal))
+ return current_sal.pc;
+
+ /* Otherwise, keep looking for entries for the same line but with
+ smaller PC's. */
+ bool done = false;
+ CORE_ADDR prev_pc;
+ while (!done)
+ {
+ prev_pc = prev_sal.pc;
+
+ prev_sal = find_pc_line (prev_pc - 1, 0);
+
+ /* Did we notice a line change? If so, we are done with the search. */
+ if (!sal_line_symtab_matches_p (prev_sal, current_sal))
+ done = true;
+ }
+
+ return prev_pc;
+}
+
/* See symtab.h. */
struct symtab *
diff --git a/gdb/symtab.h b/gdb/symtab.h
index ee4729b14cd..82bc013622f 100644
--- a/gdb/symtab.h
+++ b/gdb/symtab.h
@@ -2354,6 +2354,22 @@ extern struct symtab_and_line find_pc_line (CORE_ADDR, int);
extern struct symtab_and_line find_pc_sect_line (CORE_ADDR,
struct obj_section *, int);
+/* Given PC, and assuming it is part of a range of addresses that is part of a
+ line, go back through the linetable and find the starting PC of that
+ line.
+
+ For example, suppose we have 3 PC ranges for line X:
+
+ Line X - [0x0 - 0x8]
+ Line X - [0x8 - 0x10]
+ Line X - [0x10 - 0x18]
+
+ If we call the function with PC == 0x14, we want to return 0x0, as that is
+ the starting PC of line X, and the ranges are contiguous.
+*/
+
+extern gdb::optional<CORE_ADDR> find_line_range_start (CORE_ADDR pc);
+
/* Wrapper around find_pc_line to just return the symtab. */
extern struct symtab *find_pc_line_symtab (CORE_ADDR);
diff --git a/gdb/testsuite/gdb.mi/mi-reverse.exp b/gdb/testsuite/gdb.mi/mi-reverse.exp
index baa53a495d7..0630b8b6c7f 100644
--- a/gdb/testsuite/gdb.mi/mi-reverse.exp
+++ b/gdb/testsuite/gdb.mi/mi-reverse.exp
@@ -96,11 +96,8 @@ proc test_controlled_execution_reverse {} {
"reverse finish from callme"
# Test exec-reverse-next
- # It takes two steps to get back to the previous line,
- # as the first step moves us to the start of the current line,
- # and the one after that moves back to the previous line.
- mi_execute_to "exec-next --reverse 2" \
+ mi_execute_to "exec-next --reverse" \
"end-stepping-range" "main" "" \
"basics.c" $line_main_hello "" \
"reverse next to get over the call to do_nothing"
diff --git a/gdb/testsuite/gdb.reverse/finish-reverse-next.exp b/gdb/testsuite/gdb.reverse/finish-reverse-next.exp
index 1f53b649a7d..303f325eb18 100644
--- a/gdb/testsuite/gdb.reverse/finish-reverse-next.exp
+++ b/gdb/testsuite/gdb.reverse/finish-reverse-next.exp
@@ -76,14 +76,10 @@ gdb_continue_to_breakpoint \
repeat_cmd_until "stepi" "CALL VIA LEP" "{" "stepi into function1 call" "100"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*function1 \\(a, b\\); // CALL VIA LEP.*" \
"reverse-finish function1 LEP call from LEP "
-gdb_test "reverse-next" ".*function1 \\(a, b\\); // CALL VIA LEP" \
- "reverse next 1 LEP entry point function call from LEP"
gdb_test "reverse-next" ".*b = 5;.*" "reverse next 2, at b = 5, call from LEP"
@@ -109,14 +105,10 @@ gdb_continue_to_breakpoint \
gdb_test "step" ".*int ret = 0;.*" "step test 1"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*function1 \\(a, b\\); // CALL VIA LEP.*" \
"reverse-finish function1 LEP call from function body"
-gdb_test "reverse-next" ".*function1 \\(a, b\\); // CALL VIA LEP.*" \
- "reverse next 1 LEP from function body"
gdb_test "reverse-next" ".*b = 5;.*" \
"reverse next 2 at b = 5, from function body"
@@ -144,14 +136,11 @@ gdb_continue_to_breakpoint \
repeat_cmd_until "stepi" "CALL VIA GEP" "{" "stepi into funp call"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
+
gdb_test "reverse-finish" ".*funp \\(a, b\\);.*" \
"function1 GEP call call from GEP"
-gdb_test "reverse-next" ".*funp \\(a, b\\);.*" \
- "reverse next 1 GEP entry point function call from GEP"
gdb_test "reverse-next" ".*b = 50;.*" "reverse next 2 at b = 50, call from GEP"
gdb_test "reverse-continue" ".*" "setup for test 4"
@@ -180,14 +169,10 @@ repeat_cmd_until "stepi" "CALL VIA GEP" "{" "stepi into funp call again"
gdb_test "stepi" "{" "stepi to between GEP and LEP"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*funp \\(a, b\\);.*" \
"function1 GEP call call from GEP again"
-gdb_test "reverse-next" ".*funp \\(a, b\\);.*" \
- "reverse next 1 GEP entry point function call from GEP again"
gdb_test "reverse-next" ".*b = 50;.*" \
"reverse next 2 at b = 50, call from GEP again"
@@ -212,13 +197,10 @@ gdb_continue_to_breakpoint \
gdb_test "step" ".*int ret = 0;.*" "step test 2"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*funp \\(a, b\\);.*" \
"reverse-finish function1 GEP call, from function body "
-gdb_test "reverse-next" ".*funp \\(a, b\\);.*" \
- "reverse next 1 GEP call from function body"
+
gdb_test "reverse-next" ".*b = 50;.*" \
"reverse next 2 at b = 50 from function body"
diff --git a/gdb/testsuite/gdb.reverse/func-map-to-same-line.c b/gdb/testsuite/gdb.reverse/func-map-to-same-line.c
new file mode 100644
index 00000000000..17fe17af267
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/func-map-to-same-line.c
@@ -0,0 +1,37 @@
+/* Copyright 2023 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+ This test is used to test the reverse-step and reverse-next instruction
+ execution for a source line that contains multiple function calls. */
+
+void
+func1 (void)
+{
+} /* END FUNC1 */
+
+void
+func2 (void)
+{
+} /* END FUNC2 */
+
+int
+main (void)
+{
+ int a, b;
+ a = 1;
+ b = 2;
+ func1 (); func2 ();
+ a = a + b; /* START REVERSE TEST */
+}
diff --git a/gdb/testsuite/gdb.reverse/func-map-to-same-line.exp b/gdb/testsuite/gdb.reverse/func-map-to-same-line.exp
new file mode 100644
index 00000000000..2890b5b1a70
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/func-map-to-same-line.exp
@@ -0,0 +1,139 @@
+# Copyright 2023 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+# This file is part of the GDB testsuite. It tests reverse stepping.
+# Lots of code borrowed from "step-test.exp".
+
+# This test checks to make sure there is no regression failures for
+# the reverse-next command when stepping back over two functions in
+# the same line.
+
+require supports_reverse
+
+# This test uses the gcc no-column-info command which was added in gcc 7.1.
+
+proc run_tests {} {
+ global testfile
+
+ clean_restart ${testfile}
+
+ if { ![runto_main] } {
+ return
+ }
+
+ with_test_prefix "next-test" {
+ gdb_test_no_output "record" "turn on process record"
+
+ # This regression test verifies the reverse-step and reverse-next
+ # commands work properly when executing backwards thru a source line
+ # containing two function calls on the same source line, i.e. func1 ();
+ # func2 ();. This test is compiled so the dwarf info not contain the
+ # line table information.
+
+ # Test 1, reverse-next command
+ # Set breakpoint at the line after the function calls.
+ set bp_start_reverse_test [gdb_get_line_number "START REVERSE TEST"]
+
+ gdb_breakpoint $bp_start_reverse_test temporary
+
+ # Continue to break point for reverse-next test.
+ # Command definition: reverse-next [count]
+ # Run backward to the beginning of the previous line executed in the
+ # current (innermost) stack frame. If the line contains function
+ # calls,they will be “un-executed” without stopping. Starting from
+ # the first line of a function, reverse-next will take you back to
+ # the caller of that function, before the function was called, just
+ # as the normal next command would take you from the last line of a
+ # function back to its return to its caller 2 .
+
+ gdb_continue_to_breakpoint \
+ "stopped at command reverse-next test start location" \
+ ".*$bp_start_reverse_test\r\n.*"
+
+ # The reverse-next should step all the way back to the beginning of the
+ # line, i.e. at the beginning of the func1 call.
+ gdb_test "reverse-next" ".*func1 \\(\\); func2 \\(\\);.*" \
+ " reverse-next to line with two functions"
+
+ # We should be stopped at the first instruction of the line. A
+ # reverse-step should step back and stop at the beginning of the
+ # previous line b = 2, i.e. not in func1 ().
+ gdb_test "reverse-stepi" ".*b = 2;.*" \
+ "reverse-stepi to previous line b = 2"
+ }
+
+ # Setup for test 2
+ clean_restart ${testfile}
+
+ if { ![runto_main] } {
+ return
+ }
+
+ with_test_prefix "step-test" {
+ gdb_test_no_output "record" "turn on process record"
+
+ # Test 2, reverse-step command
+ # Set breakpoint at the line after the function calls.
+ gdb_breakpoint $bp_start_reverse_test temporary
+
+ # Continue to the start of the reverse-step test.
+ # Command definition: reverse-step [count]
+ # Run the program backward until control reaches the start of a
+ # different source line; then stop it, and return control to gdb.
+ # Like the step command, reverse-step will only stop at the beginning
+ # of a source line. It “un-executes” the previously executed source
+ # line. If the previous source line included calls to debuggable
+ # functions, reverse-step will step (backward) into the called
+ # function, stopping at the beginning of the last statement in the
+ # called function (typically a return statement). Also, as with the
+ # step command, if non-debuggable functions are called, reverse-step
+ # will run thru them backward without stopping.
+
+ gdb_continue_to_breakpoint \
+ "stopped at command reverse-step test start location" \
+ ".*$bp_start_reverse_test\r\n.*"
+
+ # The first reverse step should take us call of func2 ().
+ gdb_test "reverse-step" ".*END FUNC2.*" \
+ "reverse-step into func2 "
+
+ # The second reverse step should take us into func1 ().
+ gdb_test "reverse-step" ".*END FUNC1.*" \
+ "reverse-step into func1 "
+
+ # The third reverse step should take us call of func1 ().
+ gdb_test "reverse-step" ".*func1 \\(\\); func2 \\(\\);.*" \
+ "reverse-step to line func1(); func2(), at call for func1 "
+
+ # We should be stopped at the first instruction of the line. A reverse
+ # stepi should take us to b = 2 ().
+ gdb_test "reverse-stepi" ".*b = 2;.*" \
+ "reverse-stepi to line b = 2 "
+ }
+}
+
+standard_testfile .c
+
+# test with and without gcc column info enabled
+foreach_with_prefix column_info_flag {column-info no-column-info} {
+ set options [list debug $column_info_flag]
+
+ if {[prepare_for_testing "failed to prepare" ${testfile} ${srcfile} \
+ $options]} {
+ return -1
+ }
+
+ run_tests
+}
diff --git a/gdb/testsuite/gdb.reverse/map-to-same-line.c b/gdb/testsuite/gdb.reverse/map-to-same-line.c
new file mode 100644
index 00000000000..5ae0a89b329
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/map-to-same-line.c
@@ -0,0 +1,58 @@
+/* Copyright 2023 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/ >. */
+
+/* The purpose of this test is to create a DWARF line table that contains two
+ or more entries for the same line. When stepping (forwards or backwards),
+ GDB should step over the entire line and not just a particular entry in the
+ line table. */
+
+int
+main (void)
+{ /* TAG: main prologue */
+ asm ("main_label: .globl main_label");
+ int i = 1, j = 2, k;
+ float f1 = 2.0, f2 = 4.1, f3;
+ const char *str_1 = "foo", *str_2 = "bar", *str_3;
+
+ asm ("line1: .globl line1");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 1 */
+
+ asm ("line2: .globl line2");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 2 */
+
+ asm ("line3: .globl line3");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 3 */
+
+ asm ("line4: .globl line4");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 4 */
+
+ asm ("line5: .globl line5");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 5 */
+
+ asm ("line6: .globl line6");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 6 */
+
+ asm ("line7: .globl line7");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 7 */
+
+ asm ("line8: .globl line8");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 8 */
+
+ asm ("main_return: .globl main_return");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: main return */
+
+ asm ("end_of_sequence: .globl end_of_sequence");
+ return 0; /* TAG: main return */
+}
diff --git a/gdb/testsuite/gdb.reverse/map-to-same-line.exp b/gdb/testsuite/gdb.reverse/map-to-same-line.exp
new file mode 100644
index 00000000000..63f8c9c76b3
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/map-to-same-line.exp
@@ -0,0 +1,153 @@
+# Copyright 2023 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/ >.
+
+# When stepping (forwards or backwards), GDB should step over the entire line
+# and not just a particular entry in the line table. This test was added to
+# verify the find_line_range_start function properly sets the step range for a
+# line that consists of multiple statements, i.e. multiple entries in the line
+# table. This test creates a DWARF line table that contains two entries for
+# the same line to do the needed testing.
+
+# This test can only be run on targets which support DWARF-2 and use gas.
+load_lib dwarf.exp
+require dwarf2_support
+
+# The DWARF assembler requires the gcc compiler.
+require is_c_compiler_gcc
+
+# This test suitable only for process that can do reverse execution
+require supports_reverse
+
+standard_testfile .c .S
+
+if { [prepare_for_testing "failed to prepare" ${testfile} ${srcfile}] } {
+ return -1
+}
+
+set asm_file [standard_output_file $srcfile2]
+Dwarf::assemble $asm_file {
+ global srcdir subdir srcfile
+ declare_labels integer_label L
+
+ # Find start address and length of program
+ lassign [function_range main [list ${srcdir}/${subdir}/$srcfile]] \
+ main_start main_len
+ set main_end "$main_start + $main_len"
+
+ cu {} {
+ compile_unit {
+ {language @DW_LANG_C}
+ {name map-to-same-line.c}
+ {stmt_list $L DW_FORM_sec_offset}
+ {low_pc 0 addr}
+ } {
+ subprogram {
+ {external 1 flag}
+ {name main}
+ {low_pc $main_start addr}
+ {high_pc $main_len DW_FORM_data4}
+ }
+ }
+ }
+
+ lines {version 2 default_is_stmt 1} L {
+ include_dir "${srcdir}/${subdir}"
+ file_name "$srcfile" 1
+
+ # Generate the line table program with distinct source lines being
+ # mapped to the same line entry. Line 1, 5 and 8 contain 1 statement
+ # each. Line 2 contains 2 statements. Line 3 contains 3 statements.
+ program {
+ DW_LNE_set_address $main_start
+ line [gdb_get_line_number "TAG: main prologue"]
+ DW_LNS_copy
+ DW_LNE_set_address line1
+ line [gdb_get_line_number "TAG: line 1" ]
+ DW_LNS_copy
+ DW_LNE_set_address line2
+ line [gdb_get_line_number "TAG: line 2" ]
+ DW_LNS_copy
+ DW_LNE_set_address line3
+ line [gdb_get_line_number "TAG: line 2" ]
+ DW_LNS_copy
+ DW_LNE_set_address line4
+ line [gdb_get_line_number "TAG: line 3" ]
+ DW_LNS_copy
+ DW_LNE_set_address line5
+ line [gdb_get_line_number "TAG: line 3" ]
+ DW_LNS_copy
+ DW_LNE_set_address line6
+ line [gdb_get_line_number "TAG: line 3" ]
+ DW_LNS_copy
+ DW_LNE_set_address line7
+ line [gdb_get_line_number "TAG: line 5" ]
+ DW_LNS_copy
+ DW_LNE_set_address line8
+ line [gdb_get_line_number "TAG: line 8" ]
+ DW_LNS_copy
+ DW_LNE_set_address main_return
+ line [gdb_get_line_number "TAG: main return"]
+ DW_LNS_copy
+ DW_LNE_set_address end_of_sequence
+ DW_LNE_end_sequence
+ }
+ }
+}
+
+if { [prepare_for_testing "failed to prepare" ${testfile} \
+ [list $srcfile $asm_file] {nodebug} ] } {
+ return -1
+}
+
+if { ![runto_main] } {
+ return
+}
+
+# Print the line table
+gdb_test_multiple "maint info line-table ${testfile}" "" {
+ -re "\r\n$decimal\[ \t\]+$decimal\[ \t\]+($hex)\[ \t\]+Y\[^\r\n\]*" {
+ lappend is_stmt $expect_out(1,string)
+ exp_continue
+ }
+ -re -wrap "" {
+ }
+}
+
+# Do the reverse-step and reverse-next tests
+foreach_with_prefix cmd {step next} {
+ gdb_test_no_output "record" "turn on process record, test $cmd"
+
+ set bp_main_return [gdb_get_line_number "TAG: main return" $srcfile]
+ gdb_breakpoint $srcfile:$bp_main_return
+ gdb_continue_to_breakpoint "run to end of main, reverse-$cmd test" ".*$srcfile:$bp_main_return.*"
+ gdb_test "display \$pc" ".*pc =.*" "display pc, reverse-$cmd test"
+
+ # At this point, GDB has already recorded the execution up until the return
+ # statement. Reverse and test if GDB transitions between lines in the
+ # expected order. It should reverse-step or reverse-next across lines 8,
+ # 5, 3, 2 and 1.
+ foreach line {8 5 3 2 1} {
+ gdb_test "reverse-$cmd" ".*TAG: line $line.*" "reverse $cmd to line $line"
+ }
+
+ if {$cmd =="step"} {
+ ## Clean restart, test reverse-next command
+ clean_restart ${testfile}
+
+ if { ![runto_main] } {
+ return
+ }
+ }
+}
--
2.37.2
----------------------------------------------------
From 298805d68852a5a9e2a80aa983d665a64afda08a Mon Sep 17 00:00:00 2001
From: Carl Love <cel@us.ibm.com>
Date: Thu, 27 Apr 2023 11:34:01 -0400
Subject: [PATCH 2/2] Fix reverse stepping multiple contiguous PC ranges over
the line table.
There are a couple of scenarios where the GDB reverse-step and reverse-next
commands do not work correctly.
Scenario 1 issue description by Luis Machado:
When running GDB's testsuite on aarch64-linux/Ubuntu 20.04 (also spotted on
the ppc backend), I noticed some failures in gdb.reverse/solib-precsave.exp
and gdb.reverse/solib-reverse.exp.
The failure happens around the following code:
38 b[1] = shr2(17); /* middle part two */
40 b[0] = 6; b[1] = 9; /* generic statement, end part two */
42 shr1 ("message 1\n"); /* shr1 one */
Normal execution:
- step from line 38 will land on line 40.
- step from line 40 will land on line 42.
Reverse execution:
- step from line 42 will land on line 40.
- step from line 40 will land on line 40.
- step from line 40 will land on line 38.
V
The problem here is that line 40 contains two contiguous but distinct
PC ranges in the line table, like so:
Line 40 - [0x7ec ~ 0x7f4]
Line 40 - [0x7f4 ~ 0x7fc]
The two distinct ranges are generated because GCC started outputting source
column information, which GDB doesn't take into account at the moment.
When stepping forward from line 40, we skip both of these ranges and land on
line 42. When stepping backward from line 42, we stop at the start PC of the
second (or first, going backwards) range of line 40.
Since we've reached ecs->event_thread->control.step_range_start, we stop
stepping backwards.
---------------------------------------------------------
Scenario 2 issue described by Pedro Alves:
The following explanation of the issue was taken from the gdb mailing list
discussion of the withdrawn patch to change the behavior of the reverse-step
and reverse-next commands. Specifically, message from Pedro Alves
<pedro@palves.net> where he demonstrates the issue where you have multiple
function calls on the same source code line:
https://sourceware.org/pipermail/gdb-patches/2023-January/196110.html
The source line looks like:
func1 (); func2 ();
so stepping backwards over that line should always stop at the first
instruction of the line, not in the middle. Let's simplify this.
Here's the full source code of my example:
(gdb) list 1
1 void func1 ()
2 {
3 }
4
5 void func2 ()
6 {
7 }
8
9 int main ()
10 {
11 func1 (); func2 ();
12 }
Compiled with:
$ gcc reverse.c -o reverse -g3 -O0
$ gcc -v
...
gcc version 11.3.0 (Ubuntu 11.3.0-1ubuntu1~22.04)
Now let's debug it with target record, using current gdb git master (f3d8ae90b236),
without your patch:
$ gdb ~/reverse
GNU gdb (GDB) 14.0.50.20230124-git
...
Reading symbols from /home/pedro/reverse...
(gdb) start
Temporary breakpoint 1 at 0x1147: file reverse.c, line 11.
Starting program: /home/pedro/reverse
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Temporary breakpoint 1, main () at reverse.c:11
11 func1 (); func2 ();
(gdb) record
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
=> 0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
(gdb) n
12 }
So far so good, a "next" stepped over the whole of line 11 and stopped at line 12.
Let's confirm where we are now:
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
=> 0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
Good, we're at the first instruction of line 12.
Now let's undo the "next", with "reverse-next":
(gdb) reverse-next
11 func1 (); func2 ();
Seemingly stopped at line 11. Let's see exactly where:
(gdb) disassemble /s
Dump of assembler code for function main:
reverse.c:
10 {
0x000055555555513f <+0>: endbr64
0x0000555555555143 <+4>: push %rbp
0x0000555555555144 <+5>: mov %rsp,%rbp
11 func1 (); func2 ();
0x0000555555555147 <+8>: mov $0x0,%eax
0x000055555555514c <+13>: call 0x555555555129 <func1>
=> 0x0000555555555151 <+18>: mov $0x0,%eax
0x0000555555555156 <+23>: call 0x555555555134 <func2>
0x000055555555515b <+28>: mov $0x0,%eax
12 }
0x0000555555555160 <+33>: pop %rbp
0x0000555555555161 <+34>: ret
End of assembler dump.
(gdb)
And lo, we stopped in the middle of line 11! That is a bug, we should have
stepped back all the way to the beginning of the line. The "reverse-next"
should have fully undone the prior "next" command.
The above issues were fixed by introducing a new function that looks for
adjacent PC ranges for the same line, until we notice a line change. Then
we take that as the start PC of the range. The new start PC for the range
is used for the control.step_range_start when setting up a step range.
The test case gdb.reverse/map-to-same-line.exp is added to test the fix
for the issues in scenario 1.
The test case gdb.reverse/func-map-to-same-line.exp was added to test the
fix for scenario 2 when the binary was compiled with and without line
table information.
bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28426
Co-authored-by: Luis Machado <luis.machado@arm.com>
Co-authored-by: Carl Love <cel@us.ibm.com>
---
gdb/infcmd.c | 13 ++
gdb/infrun.c | 57 +++++++
gdb/symtab.c | 49 ++++++
gdb/symtab.h | 16 ++
gdb/testsuite/gdb.mi/mi-reverse.exp | 5 +-
.../gdb.reverse/finish-reverse-next.exp | 42 ++---
.../gdb.reverse/func-map-to-same-line.c | 37 +++++
.../gdb.reverse/func-map-to-same-line.exp | 139 ++++++++++++++++
gdb/testsuite/gdb.reverse/map-to-same-line.c | 58 +++++++
.../gdb.reverse/map-to-same-line.exp | 153 ++++++++++++++++++
10 files changed, 535 insertions(+), 34 deletions(-)
create mode 100644 gdb/testsuite/gdb.reverse/func-map-to-same-line.c
create mode 100644 gdb/testsuite/gdb.reverse/func-map-to-same-line.exp
create mode 100644 gdb/testsuite/gdb.reverse/map-to-same-line.c
create mode 100644 gdb/testsuite/gdb.reverse/map-to-same-line.exp
diff --git a/gdb/infcmd.c b/gdb/infcmd.c
index fd85d27466a..42dff4e43f9 100644
--- a/gdb/infcmd.c
+++ b/gdb/infcmd.c
@@ -975,6 +975,19 @@ prepare_one_step (thread_info *tp, struct step_command_fsm *sm)
&tp->control.step_range_start,
&tp->control.step_range_end);
+ if (execution_direction == EXEC_REVERSE)
+ {
+ symtab_and_line sal = find_pc_line (pc, 0);
+ symtab_and_line sal_start
+ = find_pc_line (tp->control.step_range_start, 0);
+
+ if (sal.line == sal_start.line)
+ /* Executing in reverse, the step_range_start address is in
+ the same line. We want to stop in the previous line so
+ move step_range_start before the current line. */
+ tp->control.step_range_start--;
+ }
+
/* There's a problem in gcc (PR gcc/98780) that causes missing line
table entries, which results in a too large stepping range.
Use inlined_subroutine info to make the range more narrow. */
diff --git a/gdb/infrun.c b/gdb/infrun.c
index 4730d290442..5132ebb6647 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -115,6 +115,9 @@ static struct async_event_handler *infrun_async_inferior_event_token;
Starts off as -1, indicating "never enabled/disabled". */
static int infrun_is_async = -1;
+static CORE_ADDR update_line_range_start (CORE_ADDR pc,
+ struct execution_control_state *ecs);
+
/* See infrun.h. */
void
@@ -6886,6 +6889,27 @@ handle_signal_stop (struct execution_control_state *ecs)
process_event_stop_test (ecs);
}
+/* Return the address for the beginning of the line. */
+
+CORE_ADDR
+update_line_range_start (CORE_ADDR pc, struct execution_control_state *ecs)
+{
+ /* The line table may have multiple entries for the same source code line.
+ Given the PC, check the line table and return the PC that corresponds
+ to the line table entry for the source line that PC is in. */
+ CORE_ADDR start_line_pc = ecs->event_thread->control.step_range_start;
+ gdb::optional<CORE_ADDR> real_range_start;
+
+ /* Call find_line_range_start to get the smallest address in the
+ linetable for multiple Line X entries in the line table. */
+ real_range_start = find_line_range_start (pc);
+
+ if (real_range_start.has_value ())
+ start_line_pc = *real_range_start;
+
+ return start_line_pc;
+}
+
/* Come here when we've got some debug event / signal we can explain
(IOW, not a random signal), and test whether it should cause a
stop, or whether we should resume the inferior (transparently).
@@ -7687,6 +7711,28 @@ process_event_stop_test (struct execution_control_state *ecs)
if (stop_pc_sal.is_stmt)
{
+ if (execution_direction == EXEC_REVERSE)
+ {
+ /* We are stepping backwards make sure we have reached the
+ beginning of the line. */
+ CORE_ADDR stop_pc = ecs->event_thread->stop_pc ();
+ CORE_ADDR start_line_pc
+ = update_line_range_start (stop_pc, ecs);
+
+ if (stop_pc != start_line_pc)
+ {
+ /* Have not reached the beginning of the source code line.
+ Set a step range. Execution should stop in any function
+ calls we execute back into before reaching the beginning
+ of the line. */
+ ecs->event_thread->control.step_range_start = start_line_pc;
+ ecs->event_thread->control.step_range_end = stop_pc;
+ set_step_info (ecs->event_thread, frame, stop_pc_sal);
+ keep_going (ecs);
+ return;
+ }
+ }
+
/* We are at the start of a statement.
So stop. Note that we don't stop if we step into the middle of a
@@ -7749,6 +7795,17 @@ process_event_stop_test (struct execution_control_state *ecs)
set_step_info (ecs->event_thread, frame, stop_pc_sal);
infrun_debug_printf ("keep going");
+
+ if (execution_direction == EXEC_REVERSE)
+ {
+ CORE_ADDR stop_pc = ecs->event_thread->stop_pc ();
+
+ /* Make sure the stop_pc is set to the beginning of the line. */
+ if (stop_pc != ecs->event_thread->control.step_range_start)
+ ecs->event_thread->control.step_range_start
+ = update_line_range_start (stop_pc, ecs);
+ }
+
keep_going (ecs);
}
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 0117a2a59d7..c38bb01b747 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -3284,6 +3284,55 @@ find_pc_line (CORE_ADDR pc, int notcurrent)
return sal;
}
+/* Compare two symtab_and_line entries. Return true if both have
+ the same line number and the same symtab pointer. That means we
+ are dealing with two entries from the same line and from the same
+ source file.
+
+ Return false otherwise. */
+
+static bool
+sal_line_symtab_matches_p (const symtab_and_line &sal1,
+ const symtab_and_line &sal2)
+{
+ return sal1.line == sal2.line && sal1.symtab == sal2.symtab;
+}
+
+/* See symtah.h. */
+
+gdb::optional<CORE_ADDR>
+find_line_range_start (CORE_ADDR pc)
+{
+ struct symtab_and_line current_sal = find_pc_line (pc, 0);
+
+ if (current_sal.line == 0)
+ return {};
+
+ struct symtab_and_line prev_sal = find_pc_line (current_sal.pc - 1, 0);
+
+ /* If the previous entry is for a different line, that means we are already
+ at the entry with the start PC for this line. */
+ if (!sal_line_symtab_matches_p (prev_sal, current_sal))
+ return current_sal.pc;
+
+ /* Otherwise, keep looking for entries for the same line but with
+ smaller PC's. */
+ bool done = false;
+ CORE_ADDR prev_pc;
+ while (!done)
+ {
+ prev_pc = prev_sal.pc;
+
+ prev_sal = find_pc_line (prev_pc - 1, 0);
+
+ /* Did we notice a line change? If so, we are done with the search. */
+ if (!sal_line_symtab_matches_p (prev_sal, current_sal))
+ done = true;
+ }
+
+ return prev_pc;
+}
+
/* See symtab.h. */
struct symtab *
diff --git a/gdb/symtab.h b/gdb/symtab.h
index ee4729b14cd..82bc013622f 100644
--- a/gdb/symtab.h
+++ b/gdb/symtab.h
@@ -2354,6 +2354,22 @@ extern struct symtab_and_line find_pc_line (CORE_ADDR, int);
extern struct symtab_and_line find_pc_sect_line (CORE_ADDR,
struct obj_section *, int);
+/* Given PC, and assuming it is part of a range of addresses that is part of a
+ line, go back through the linetable and find the starting PC of that
+ line.
+
+ For example, suppose we have 3 PC ranges for line X:
+
+ Line X - [0x0 - 0x8]
+ Line X - [0x8 - 0x10]
+ Line X - [0x10 - 0x18]
+
+ If we call the function with PC == 0x14, we want to return 0x0, as that is
+ the starting PC of line X, and the ranges are contiguous.
+*/
+
+extern gdb::optional<CORE_ADDR> find_line_range_start (CORE_ADDR pc);
+
/* Wrapper around find_pc_line to just return the symtab. */
extern struct symtab *find_pc_line_symtab (CORE_ADDR);
diff --git a/gdb/testsuite/gdb.mi/mi-reverse.exp b/gdb/testsuite/gdb.mi/mi-reverse.exp
index baa53a495d7..0630b8b6c7f 100644
--- a/gdb/testsuite/gdb.mi/mi-reverse.exp
+++ b/gdb/testsuite/gdb.mi/mi-reverse.exp
@@ -96,11 +96,8 @@ proc test_controlled_execution_reverse {} {
"reverse finish from callme"
# Test exec-reverse-next
- # It takes two steps to get back to the previous line,
- # as the first step moves us to the start of the current line,
- # and the one after that moves back to the previous line.
- mi_execute_to "exec-next --reverse 2" \
+ mi_execute_to "exec-next --reverse" \
"end-stepping-range" "main" "" \
"basics.c" $line_main_hello "" \
"reverse next to get over the call to do_nothing"
diff --git a/gdb/testsuite/gdb.reverse/finish-reverse-next.exp b/gdb/testsuite/gdb.reverse/finish-reverse-next.exp
index 1f53b649a7d..303f325eb18 100644
--- a/gdb/testsuite/gdb.reverse/finish-reverse-next.exp
+++ b/gdb/testsuite/gdb.reverse/finish-reverse-next.exp
@@ -76,14 +76,10 @@ gdb_continue_to_breakpoint \
repeat_cmd_until "stepi" "CALL VIA LEP" "{" "stepi into function1 call" "100"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*function1 \\(a, b\\); // CALL VIA LEP.*" \
"reverse-finish function1 LEP call from LEP "
-gdb_test "reverse-next" ".*function1 \\(a, b\\); // CALL VIA LEP" \
- "reverse next 1 LEP entry point function call from LEP"
gdb_test "reverse-next" ".*b = 5;.*" "reverse next 2, at b = 5, call from LEP"
@@ -109,14 +105,10 @@ gdb_continue_to_breakpoint \
gdb_test "step" ".*int ret = 0;.*" "step test 1"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*function1 \\(a, b\\); // CALL VIA LEP.*" \
"reverse-finish function1 LEP call from function body"
-gdb_test "reverse-next" ".*function1 \\(a, b\\); // CALL VIA LEP.*" \
- "reverse next 1 LEP from function body"
gdb_test "reverse-next" ".*b = 5;.*" \
"reverse next 2 at b = 5, from function body"
@@ -144,14 +136,11 @@ gdb_continue_to_breakpoint \
repeat_cmd_until "stepi" "CALL VIA GEP" "{" "stepi into funp call"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
+
gdb_test "reverse-finish" ".*funp \\(a, b\\);.*" \
"function1 GEP call call from GEP"
-gdb_test "reverse-next" ".*funp \\(a, b\\);.*" \
- "reverse next 1 GEP entry point function call from GEP"
gdb_test "reverse-next" ".*b = 50;.*" "reverse next 2 at b = 50, call from GEP"
gdb_test "reverse-continue" ".*" "setup for test 4"
@@ -180,14 +169,10 @@ repeat_cmd_until "stepi" "CALL VIA GEP" "{" "stepi into funp call again"
gdb_test "stepi" "{" "stepi to between GEP and LEP"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*funp \\(a, b\\);.*" \
"function1 GEP call call from GEP again"
-gdb_test "reverse-next" ".*funp \\(a, b\\);.*" \
- "reverse next 1 GEP entry point function call from GEP again"
gdb_test "reverse-next" ".*b = 50;.*" \
"reverse next 2 at b = 50, call from GEP again"
@@ -212,13 +197,10 @@ gdb_continue_to_breakpoint \
gdb_test "step" ".*int ret = 0;.*" "step test 2"
# The reverse-finish command should stop on the function call instruction
-# which is the last instruction in the source code line. A reverse-next
-# instruction should then stop at the first instruction in the same source
-# code line. Another revers-next instruction stops at the previous source
-# code line.
+# which is the last instruction in the source code line. Another revers-next
+# instruction stops at the previous source code line.
gdb_test "reverse-finish" ".*funp \\(a, b\\);.*" \
"reverse-finish function1 GEP call, from function body "
-gdb_test "reverse-next" ".*funp \\(a, b\\);.*" \
- "reverse next 1 GEP call from function body"
+
gdb_test "reverse-next" ".*b = 50;.*" \
"reverse next 2 at b = 50 from function body"
diff --git a/gdb/testsuite/gdb.reverse/func-map-to-same-line.c b/gdb/testsuite/gdb.reverse/func-map-to-same-line.c
new file mode 100644
index 00000000000..17fe17af267
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/func-map-to-same-line.c
@@ -0,0 +1,37 @@
+/* Copyright 2023 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+ This test is used to test the reverse-step and reverse-next instruction
+ execution for a source line that contains multiple function calls. */
+
+void
+func1 (void)
+{
+} /* END FUNC1 */
+
+void
+func2 (void)
+{
+} /* END FUNC2 */
+
+int
+main (void)
+{
+ int a, b;
+ a = 1;
+ b = 2;
+ func1 (); func2 ();
+ a = a + b; /* START REVERSE TEST */
+}
diff --git a/gdb/testsuite/gdb.reverse/func-map-to-same-line.exp b/gdb/testsuite/gdb.reverse/func-map-to-same-line.exp
new file mode 100644
index 00000000000..2890b5b1a70
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/func-map-to-same-line.exp
@@ -0,0 +1,139 @@
+# Copyright 2023 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+# This file is part of the GDB testsuite. It tests reverse stepping.
+# Lots of code borrowed from "step-test.exp".
+
+# This test checks to make sure there is no regression failures for
+# the reverse-next command when stepping back over two functions in
+# the same line.
+
+require supports_reverse
+
+# This test uses the gcc no-column-info command which was added in gcc 7.1.
+
+proc run_tests {} {
+ global testfile
+
+ clean_restart ${testfile}
+
+ if { ![runto_main] } {
+ return
+ }
+
+ with_test_prefix "next-test" {
+ gdb_test_no_output "record" "turn on process record"
+
+ # This regression test verifies the reverse-step and reverse-next
+ # commands work properly when executing backwards thru a source line
+ # containing two function calls on the same source line, i.e. func1 ();
+ # func2 ();. This test is compiled so the dwarf info not contain the
+ # line table information.
+
+ # Test 1, reverse-next command
+ # Set breakpoint at the line after the function calls.
+ set bp_start_reverse_test [gdb_get_line_number "START REVERSE TEST"]
+
+ gdb_breakpoint $bp_start_reverse_test temporary
+
+ # Continue to break point for reverse-next test.
+ # Command definition: reverse-next [count]
+ # Run backward to the beginning of the previous line executed in the
+ # current (innermost) stack frame. If the line contains function
+ # calls,they will be “un-executed” without stopping. Starting from
+ # the first line of a function, reverse-next will take you back to
+ # the caller of that function, before the function was called, just
+ # as the normal next command would take you from the last line of a
+ # function back to its return to its caller 2 .
+
+ gdb_continue_to_breakpoint \
+ "stopped at command reverse-next test start location" \
+ ".*$bp_start_reverse_test\r\n.*"
+
+ # The reverse-next should step all the way back to the beginning of the
+ # line, i.e. at the beginning of the func1 call.
+ gdb_test "reverse-next" ".*func1 \\(\\); func2 \\(\\);.*" \
+ " reverse-next to line with two functions"
+
+ # We should be stopped at the first instruction of the line. A
+ # reverse-step should step back and stop at the beginning of the
+ # previous line b = 2, i.e. not in func1 ().
+ gdb_test "reverse-stepi" ".*b = 2;.*" \
+ "reverse-stepi to previous line b = 2"
+ }
+
+ # Setup for test 2
+ clean_restart ${testfile}
+
+ if { ![runto_main] } {
+ return
+ }
+
+ with_test_prefix "step-test" {
+ gdb_test_no_output "record" "turn on process record"
+
+ # Test 2, reverse-step command
+ # Set breakpoint at the line after the function calls.
+ gdb_breakpoint $bp_start_reverse_test temporary
+
+ # Continue to the start of the reverse-step test.
+ # Command definition: reverse-step [count]
+ # Run the program backward until control reaches the start of a
+ # different source line; then stop it, and return control to gdb.
+ # Like the step command, reverse-step will only stop at the beginning
+ # of a source line. It “un-executes” the previously executed source
+ # line. If the previous source line included calls to debuggable
+ # functions, reverse-step will step (backward) into the called
+ # function, stopping at the beginning of the last statement in the
+ # called function (typically a return statement). Also, as with the
+ # step command, if non-debuggable functions are called, reverse-step
+ # will run thru them backward without stopping.
+
+ gdb_continue_to_breakpoint \
+ "stopped at command reverse-step test start location" \
+ ".*$bp_start_reverse_test\r\n.*"
+
+ # The first reverse step should take us call of func2 ().
+ gdb_test "reverse-step" ".*END FUNC2.*" \
+ "reverse-step into func2 "
+
+ # The second reverse step should take us into func1 ().
+ gdb_test "reverse-step" ".*END FUNC1.*" \
+ "reverse-step into func1 "
+
+ # The third reverse step should take us call of func1 ().
+ gdb_test "reverse-step" ".*func1 \\(\\); func2 \\(\\);.*" \
+ "reverse-step to line func1(); func2(), at call for func1 "
+
+ # We should be stopped at the first instruction of the line. A reverse
+ # stepi should take us to b = 2 ().
+ gdb_test "reverse-stepi" ".*b = 2;.*" \
+ "reverse-stepi to line b = 2 "
+ }
+}
+
+standard_testfile .c
+
+# test with and without gcc column info enabled
+foreach_with_prefix column_info_flag {column-info no-column-info} {
+ set options [list debug $column_info_flag]
+
+ if {[prepare_for_testing "failed to prepare" ${testfile} ${srcfile} \
+ $options]} {
+ return -1
+ }
+
+ run_tests
+}
diff --git a/gdb/testsuite/gdb.reverse/map-to-same-line.c b/gdb/testsuite/gdb.reverse/map-to-same-line.c
new file mode 100644
index 00000000000..5ae0a89b329
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/map-to-same-line.c
@@ -0,0 +1,58 @@
+/* Copyright 2023 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/ >. */
+
+/* The purpose of this test is to create a DWARF line table that contains two
+ or more entries for the same line. When stepping (forwards or backwards),
+ GDB should step over the entire line and not just a particular entry in the
+ line table. */
+
+int
+main (void)
+{ /* TAG: main prologue */
+ asm ("main_label: .globl main_label");
+ int i = 1, j = 2, k;
+ float f1 = 2.0, f2 = 4.1, f3;
+ const char *str_1 = "foo", *str_2 = "bar", *str_3;
+
+ asm ("line1: .globl line1");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 1 */
+
+ asm ("line2: .globl line2");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 2 */
+
+ asm ("line3: .globl line3");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 3 */
+
+ asm ("line4: .globl line4");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 4 */
+
+ asm ("line5: .globl line5");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 5 */
+
+ asm ("line6: .globl line6");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 6 */
+
+ asm ("line7: .globl line7");
+ k = i; f3 = f1; str_3 = str_1; /* TAG: line 7 */
+
+ asm ("line8: .globl line8");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: line 8 */
+
+ asm ("main_return: .globl main_return");
+ k = j; f3 = f2; str_3 = str_2; /* TAG: main return */
+
+ asm ("end_of_sequence: .globl end_of_sequence");
+ return 0; /* TAG: main return */
+}
diff --git a/gdb/testsuite/gdb.reverse/map-to-same-line.exp b/gdb/testsuite/gdb.reverse/map-to-same-line.exp
new file mode 100644
index 00000000000..63f8c9c76b3
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/map-to-same-line.exp
@@ -0,0 +1,153 @@
+# Copyright 2023 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/ >.
+
+# When stepping (forwards or backwards), GDB should step over the entire line
+# and not just a particular entry in the line table. This test was added to
+# verify the find_line_range_start function properly sets the step range for a
+# line that consists of multiple statements, i.e. multiple entries in the line
+# table. This test creates a DWARF line table that contains two entries for
+# the same line to do the needed testing.
+
+# This test can only be run on targets which support DWARF-2 and use gas.
+load_lib dwarf.exp
+require dwarf2_support
+
+# The DWARF assembler requires the gcc compiler.
+require is_c_compiler_gcc
+
+# This test suitable only for process that can do reverse execution
+require supports_reverse
+
+standard_testfile .c .S
+
+if { [prepare_for_testing "failed to prepare" ${testfile} ${srcfile}] } {
+ return -1
+}
+
+set asm_file [standard_output_file $srcfile2]
+Dwarf::assemble $asm_file {
+ global srcdir subdir srcfile
+ declare_labels integer_label L
+
+ # Find start address and length of program
+ lassign [function_range main [list ${srcdir}/${subdir}/$srcfile]] \
+ main_start main_len
+ set main_end "$main_start + $main_len"
+
+ cu {} {
+ compile_unit {
+ {language @DW_LANG_C}
+ {name map-to-same-line.c}
+ {stmt_list $L DW_FORM_sec_offset}
+ {low_pc 0 addr}
+ } {
+ subprogram {
+ {external 1 flag}
+ {name main}
+ {low_pc $main_start addr}
+ {high_pc $main_len DW_FORM_data4}
+ }
+ }
+ }
+
+ lines {version 2 default_is_stmt 1} L {
+ include_dir "${srcdir}/${subdir}"
+ file_name "$srcfile" 1
+
+ # Generate the line table program with distinct source lines being
+ # mapped to the same line entry. Line 1, 5 and 8 contain 1 statement
+ # each. Line 2 contains 2 statements. Line 3 contains 3 statements.
+ program {
+ DW_LNE_set_address $main_start
+ line [gdb_get_line_number "TAG: main prologue"]
+ DW_LNS_copy
+ DW_LNE_set_address line1
+ line [gdb_get_line_number "TAG: line 1" ]
+ DW_LNS_copy
+ DW_LNE_set_address line2
+ line [gdb_get_line_number "TAG: line 2" ]
+ DW_LNS_copy
+ DW_LNE_set_address line3
+ line [gdb_get_line_number "TAG: line 2" ]
+ DW_LNS_copy
+ DW_LNE_set_address line4
+ line [gdb_get_line_number "TAG: line 3" ]
+ DW_LNS_copy
+ DW_LNE_set_address line5
+ line [gdb_get_line_number "TAG: line 3" ]
+ DW_LNS_copy
+ DW_LNE_set_address line6
+ line [gdb_get_line_number "TAG: line 3" ]
+ DW_LNS_copy
+ DW_LNE_set_address line7
+ line [gdb_get_line_number "TAG: line 5" ]
+ DW_LNS_copy
+ DW_LNE_set_address line8
+ line [gdb_get_line_number "TAG: line 8" ]
+ DW_LNS_copy
+ DW_LNE_set_address main_return
+ line [gdb_get_line_number "TAG: main return"]
+ DW_LNS_copy
+ DW_LNE_set_address end_of_sequence
+ DW_LNE_end_sequence
+ }
+ }
+}
+
+if { [prepare_for_testing "failed to prepare" ${testfile} \
+ [list $srcfile $asm_file] {nodebug} ] } {
+ return -1
+}
+
+if { ![runto_main] } {
+ return
+}
+
+# Print the line table
+gdb_test_multiple "maint info line-table ${testfile}" "" {
+ -re "\r\n$decimal\[ \t\]+$decimal\[ \t\]+($hex)\[ \t\]+Y\[^\r\n\]*" {
+ lappend is_stmt $expect_out(1,string)
+ exp_continue
+ }
+ -re -wrap "" {
+ }
+}
+
+# Do the reverse-step and reverse-next tests
+foreach_with_prefix cmd {step next} {
+ gdb_test_no_output "record" "turn on process record, test $cmd"
+
+ set bp_main_return [gdb_get_line_number "TAG: main return" $srcfile]
+ gdb_breakpoint $srcfile:$bp_main_return
+ gdb_continue_to_breakpoint "run to end of main, reverse-$cmd test" ".*$srcfile:$bp_main_return.*"
+ gdb_test "display \$pc" ".*pc =.*" "display pc, reverse-$cmd test"
+
+ # At this point, GDB has already recorded the execution up until the return
+ # statement. Reverse and test if GDB transitions between lines in the
+ # expected order. It should reverse-step or reverse-next across lines 8,
+ # 5, 3, 2 and 1.
+ foreach line {8 5 3 2 1} {
+ gdb_test "reverse-$cmd" ".*TAG: line $line.*" "reverse $cmd to line $line"
+ }
+
+ if {$cmd =="step"} {
+ ## Clean restart, test reverse-next command
+ clean_restart ${testfile}
+
+ if { ![runto_main] } {
+ return
+ }
+ }
+}
--
2.37.2
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table.
2023-08-25 15:50 [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table Carl Love
2023-08-25 15:53 ` [PATCH 1/2] Add gdb_compile options column-info and no-column-info Carl Love
2023-08-25 15:57 ` [PATCH 2/2 ver 8] Fix reverse stepping multiple contiguous PC ranges over the line table Carl Love
@ 2023-08-29 12:28 ` Guinevere Larsen
2023-08-29 18:55 ` Carl Love
2 siblings, 1 reply; 8+ messages in thread
From: Guinevere Larsen @ 2023-08-29 12:28 UTC (permalink / raw)
To: Carl Love, Simon Marchi, gdb-patches, UlrichWeigand, pedro; +Cc: luis.machado
On 25/08/2023 17:50, Carl Love wrote:
> GDB maintainers:
>
> I received a few messages from Guinevere about these two patches. In
> the end, we agreed the first patch was OK. There was a minor code
> change suggested for the second patch that I have implemented. Since I
> have not received any additional suggestions on the second patch, it is
> time to ping the first patch and post the updated second patch for
> review.
>
> The first patch in the series has not been changed. I am reposting it
> along with the second patch so it doesn't get lost.
>
> The patch series has been tested on Power10 LE and on X86-64.
>
> Please let me know if these two patches are acceptable for mainline.
> Thanks.
>
> Carl
>
I have tested this patch with both GCC and Clang. I can verify it fixes
the issue, adds no regression, and even fixes the final issues when
testing gdb.reverse with clang!
Tested-By: Guinevere Larsen <blarsen@redhat.com>
--
Cheers,
Guinevere Larsen
She/Her/Hers
^ permalink raw reply [flat|nested] 8+ messages in thread
* RE: [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table.
2023-08-29 12:28 ` [PATCH 0/2] " Guinevere Larsen
@ 2023-08-29 18:55 ` Carl Love
2023-09-05 15:12 ` Carl Love
0 siblings, 1 reply; 8+ messages in thread
From: Carl Love @ 2023-08-29 18:55 UTC (permalink / raw)
To: Guinevere Larsen, Simon Marchi, gdb-patches, UlrichWeigand, pedro
Cc: luis.machado, cel
Guinevere:
On Tue, 2023-08-29 at 14:28 +0200, Guinevere Larsen wrote:
> I have tested this patch with both GCC and Clang. I can verify it
> fixes
> the issue, adds no regression, and even fixes the final issues when
> testing gdb.reverse with clang!
>
> Tested-By: Guinevere Larsen <blarsen@redhat.com>
Thanks for the help testing and that it fixes the clang failures for
test gdb.reverse.
Carl
^ permalink raw reply [flat|nested] 8+ messages in thread
* RE: [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table.
2023-08-29 18:55 ` Carl Love
@ 2023-09-05 15:12 ` Carl Love
2023-09-25 22:15 ` Carl Love
0 siblings, 1 reply; 8+ messages in thread
From: Carl Love @ 2023-09-05 15:12 UTC (permalink / raw)
To: Guinevere Larsen, Simon Marchi, gdb-patches, UlrichWeigand, pedro
Cc: luis.machado
GDB maintainers:
Ping. I was wondering if anyoneany of the maintainers has had time to
review these patches? Thanks for your time and effort on the patch.
On Tue, 2023-08-29 at 11:55 -0700, Carl Love wrote:
> Guinevere:
>
> On Tue, 2023-08-29 at 14:28 +0200, Guinevere Larsen wrote:
> > I have tested this patch with both GCC and Clang. I can verify it
> > fixes
> > the issue, adds no regression, and even fixes the final issues
> > when
> > testing gdb.reverse with clang!
> >
> > Tested-By: Guinevere Larsen <blarsen@redhat.com>
Carl
^ permalink raw reply [flat|nested] 8+ messages in thread
* RE: [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table.
2023-09-05 15:12 ` Carl Love
@ 2023-09-25 22:15 ` Carl Love
2023-11-02 17:30 ` PING " Carl Love
0 siblings, 1 reply; 8+ messages in thread
From: Carl Love @ 2023-09-25 22:15 UTC (permalink / raw)
To: Guinevere Larsen, Simon Marchi, gdb-patches, UlrichWeigand, pedro
Cc: luis.machado, cel
GDB maintainers:
Ping.
On Tue, 2023-09-05 at 08:12 -0700, Carl Love wrote:
> GDB maintainers:
>
> Ping. I was wondering if anyoneany of the maintainers has had time
> to
> review these patches? Thanks for your time and effort on the patch.
>
> On Tue, 2023-08-29 at 11:55 -0700, Carl Love wrote:
> > Guinevere:
> >
> > On Tue, 2023-08-29 at 14:28 +0200, Guinevere Larsen wrote:
> > > I have tested this patch with both GCC and Clang. I can verify it
> > > fixes
> > > the issue, adds no regression, and even fixes the final issues
> > > when
> > > testing gdb.reverse with clang!
> > >
> > > Tested-By: Guinevere Larsen <blarsen@redhat.com>
>
> Carl
^ permalink raw reply [flat|nested] 8+ messages in thread
* PING Re: [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table.
2023-09-25 22:15 ` Carl Love
@ 2023-11-02 17:30 ` Carl Love
0 siblings, 0 replies; 8+ messages in thread
From: Carl Love @ 2023-11-02 17:30 UTC (permalink / raw)
To: Carl Love, Guinevere Larsen, Simon Marchi, gdb-patches,
UlrichWeigand, pedro
Cc: luis.machado
GDB maintainers:
Ping I haven't seen any responses to the lastest patch set. There are
two patches in the set:
[PATCH 1/2] Add gdb_compile options column-info and no-column-info
[PATCH 2/2 ver 8] Fix reverse stepping multiple contiguous PC ranges
over the line table.
Please let me know if these patches are OK. Thanks.
Carl
----------------------------------------------------------
On Mon, 2023-09-25 at 15:15 -0700, Carl Love wrote:
> GDB maintainers:
>
> Ping.
>
>
> On Tue, 2023-09-05 at 08:12 -0700, Carl Love wrote:
> > GDB maintainers:
> >
> > Ping. I was wondering if anyoneany of the maintainers has had time
> > to
> > review these patches? Thanks for your time and effort on the
> > patch.
> >
> > On Tue, 2023-08-29 at 11:55 -0700, Carl Love wrote:
> > > Guinevere:
> > >
> > > On Tue, 2023-08-29 at 14:28 +0200, Guinevere Larsen wrote:
> > > > I have tested this patch with both GCC and Clang. I can verify
> > > > it
> > > > fixes
> > > > the issue, adds no regression, and even fixes the final issues
> > > > when
> > > > testing gdb.reverse with clang!
> > > >
> > > > Tested-By: Guinevere Larsen <blarsen@redhat.com>
> >
> > Carl
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2023-11-02 17:30 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-08-25 15:50 [PATCH 0/2] Fix reverse stepping multiple contiguous PC ranges over the line table Carl Love
2023-08-25 15:53 ` [PATCH 1/2] Add gdb_compile options column-info and no-column-info Carl Love
2023-08-25 15:57 ` [PATCH 2/2 ver 8] Fix reverse stepping multiple contiguous PC ranges over the line table Carl Love
2023-08-29 12:28 ` [PATCH 0/2] " Guinevere Larsen
2023-08-29 18:55 ` Carl Love
2023-09-05 15:12 ` Carl Love
2023-09-25 22:15 ` Carl Love
2023-11-02 17:30 ` PING " Carl Love
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).