[Bug testsuite/26664] New: FAIL: gdb.threads/killed-outside.exp: prompt after first continue (timeout)

[Bug testsuite/26664] New: FAIL: gdb.threads/killed-outside.exp: prompt after first continue (timeout)

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=26664

            Bug ID: 26664
           Summary: FAIL: gdb.threads/killed-outside.exp: prompt after
                    first continue (timeout)
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: testsuite
          Assignee: unassigned at sourceware dot org
          Reporter: vries at gcc dot gnu.org
  Target Milestone: ---

On a SLE-15 SP1 machine (160 core, 3TB Mem, build area setup in /dev/shm), I
reliably run into:
...
(gdb) PASS: gdb.threads/killed-outside.exp: get pid of inferior
Executing on target: kill -9 145926    (timeout = 300)
spawn -ignore SIGHUP kill -9 145926^M
continue^M
Continuing.^M
Couldn't get registers: No such process.^M
(gdb) [Thread 0x7ffff74c4700 (LWP 145930) exited]^M
^M
Program terminated with signal SIGKILL, Killed.^M
The program no longer exists.^M
FAIL: gdb.threads/killed-outside.exp: prompt after first continue (timeout)
...

Using this patch:
...
diff --git a/gdb/testsuite/gdb.threads/killed-outside.exp
b/gdb/testsuite/gdb.threads/kille
d-outside.exp
index ff5a115728..49a909ebcb 100644
--- a/gdb/testsuite/gdb.threads/killed-outside.exp
+++ b/gdb/testsuite/gdb.threads/killed-outside.exp
@@ -43,6 +43,9 @@ set no_longer_exists_msg "The program no longer exists\."
 set not_being_run_msg "The program is not being run\."

 gdb_test_multiple "continue" "prompt after first continue" {
+    -re "Continuing\.\r\n$no_such_process_msg\r\n$gdb_prompt " {
+       fail $gdb_test_name
+    }
     -re
"Continuing\.\r\n$no_such_process_msg\r\n$no_such_process_msg\r\n$gdb_prompt "
{
        pass $gdb_test_name
        # Two times $no_such_process_msg.  The bug condition was triggered, go
...
we turn the timeout into a regular fail:
...
FAIL: gdb.threads/killed-outside.exp: prompt after first continue
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug testsuite/26664] FAIL: gdb.threads/killed-outside.exp: prompt after first continue (timeout)

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=26664

--- Comment #1 from Tom de Vries <vries at gcc dot gnu.org> ---
Here ( https://sourceware.org/pipermail/gdb-patches/2020-April/167781.html ) I
mentioned:
...
>> +gdb_test_multiple "continue" "prompt after first continue" {
>> +    -re "Continuing\.\r\n\r\n$killed_msg\r\n$no_longer_exists_msg\r\n$gdb_prompt $" {
>> +	pass $gdb_test_name
>> +	# Regular output, bug condition was not triggered, we're done.
>> +	return -1
>> +    }
>> +    -re "Continuing\.\r\n$no_such_process_msg\r\n$no_such_process_msg\r\n$gdb_prompt " {
>> +	pass $gdb_test_name
>> +	# Two times $no_such_process_msg.  The bug condition was triggered, go
>> +	# check for it.
>> +    }
>> +    -re "Continuing\.\r\n$no_such_process_msg\r\n$gdb_prompt $" {
>> +	pass $gdb_test_name
>> +	# One time $no_such_process_msg.  We're stuck here.  The bug condition
>> +	# was not triggered, but we're not getting correct gdb behaviour either:
>> +	# every subsequent continue produces one no_such_process_msg.  Give up.
>> +	return -1
> 
> I'm confused here -- the comment says we're not getting correct behavior,
> but this won't result in any FAIL?
> 

Um, yes. I made that decision because I was trying to trigger another
scenario. But agreed, it's debatable. Anyway, it also doesn't matter
anymore since rewriting the test to stop at a breakpoint eliminated the
need for this. The scenario I was trying to trigger now always reproduces.
...

So, that was incorrect, the scenario with one $no_such_process_msg still
reproduces, as this PR shows.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug testsuite/26664] FAIL: gdb.threads/killed-outside.exp: prompt after first continue (timeout)

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=26664

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |palves at redhat dot com

--- Comment #2 from Tom de Vries <vries at gcc dot gnu.org> ---
Hmm, this seems to get rid of the "Couldn't get registers: No such process":
...
diff --git a/gdb/infrun.c b/gdb/infrun.c
index 6532b06ae5..e909aab5e2 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -2330,7 +2330,7 @@ resume_1 (enum gdb_signal sig)
       step = 0;
     }

-  CORE_ADDR pc = regcache_read_pc (regcache);
+  CORE_ADDR pc = regcache_read_pc_protected (regcache);

   infrun_debug_printf ("step=%d, signal=%s, trap_expected=%d, "
                       "current thread [%s] at %s",
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug testsuite/26664] FAIL: gdb.threads/killed-outside.exp: prompt after first continue (timeout)

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=26664

--- Comment #3 from Tom de Vries <vries at gcc dot gnu.org> ---
https://sourceware.org/pipermail/gdb-patches/2021-October/182906.html

-- 
You are receiving this mail because:
You are on the CC list for the bug.