[RFA] Patch to fix "reverse-step" command error

[RFA] Patch to fix "reverse-step" command error

[teawater]

[-- Attachment #1: Type: text/plain, Size: 581 bytes --]

Hi Michael,

This patch is for bug in
http://sourceware.org/ml/gdb-patches/2009-02/msg00246.html.

The reason of bug is when inferior reverse exection into dynsym
resolve code, gdb doesn't insert breakpoint in right address.
It just happen in some distributions for example SLED.

This patch make gdb insert breakpoint in right address or step in that case.

2008-03-02  Hui Zhu  <teawater@gmail.com>

       * infrun.c (handle_inferior_event): Make inferior insert
       breakpoint at right address or step in reverse and it stepped
       into dynsym resolve code.

Thanks,
Hui

[-- Attachment #2: fix-reverse-dynsym-error.txt --]
[-- Type: text/plain, Size: 1409 bytes --]

---
 infrun.c |   31 ++++++++++++++++++++++++++-----
 1 file changed, 26 insertions(+), 5 deletions(-)

--- a/infrun.c
+++ b/infrun.c
@@ -3412,13 +3412,34 @@ infrun: BPSTAT_WHAT_SET_LONGJMP_RESUME (
 
       if (pc_after_resolver)
 	{
-	  /* Set up a step-resume breakpoint at the address
-	     indicated by SKIP_SOLIB_RESOLVER.  */
 	  struct symtab_and_line sr_sal;
-	  init_sal (&sr_sal);
-	  sr_sal.pc = pc_after_resolver;
+	  if (execution_direction == EXEC_REVERSE)
+	    {
+	      if (ecs->stop_func_start != stop_pc)
+		{
+		  /* Normal (staticly linked) function call return.  */
+		  init_sal (&sr_sal);
+		  sr_sal.pc = ecs->stop_func_start;
+		  insert_step_resume_breakpoint_at_sal (sr_sal, null_frame_id);
+		}
+	      else
+		{
+		  /* We are stepping over a function call in reverse, and
+	             stop at the start address of the function.  Go back to
+	             single-stepping, which should take us back to the
+	             function call.  */
+		  ecs->event_thread->stepping_over_breakpoint = 1;
+		}
+	    }
+	  else
+	    {
+	      /* Set up a step-resume breakpoint at the address
+	         indicated by SKIP_SOLIB_RESOLVER.  */
+	      init_sal (&sr_sal);
+	      sr_sal.pc = pc_after_resolver;
 
-	  insert_step_resume_breakpoint_at_sal (sr_sal, null_frame_id);
+	      insert_step_resume_breakpoint_at_sal (sr_sal, null_frame_id);
+	    }
 	}
 
       keep_going (ecs);

Re: [RFA] Patch to fix "reverse-step" command error

[Hui Zhu]

Hi Marc,

I cannot reproduce this issue with cvs-head.
Could you please help me test with this issue?
If both of us cannot reproduce this issue,  I think I can give up this patch.

Thanks,
Hui


On Mon, Mar 2, 2009 at 14:10, teawater <teawater@gmail.com> wrote:
> Hi Michael,
>
> This patch is for bug in
> http://sourceware.org/ml/gdb-patches/2009-02/msg00246.html.
>
> The reason of bug is when inferior reverse exection into dynsym
> resolve code, gdb doesn't insert breakpoint in right address.
> It just happen in some distributions for example SLED.
>
> This patch make gdb insert breakpoint in right address or step in that case.
>
> 2008-03-02  Hui Zhu  <teawater@gmail.com>
>
>       * infrun.c (handle_inferior_event): Make inferior insert
>       breakpoint at right address or step in reverse and it stepped
>       into dynsym resolve code.
>
> Thanks,
> Hui
>

Re: [RFA] Patch to fix "reverse-step" command error

[Hui Zhu]

[-- Attachment #1: Type: text/plain, Size: 1469 bytes --]

Hi Marc and Michael,

Sorry I forget this issue just reproduce in SLED 10.

I reproduced it with cvs-head and patch can fix it.  Michael, Please
help me review it.

The attachment is the new patch follow cvs-head.

2009-05-06  Hui Zhu  <teawater@gmail.com>

       * infrun.c (handle_inferior_event): Make inferior insert
       breakpoint at right address or step in reverse and it stepped
       into dynsym resolve code.

Thanks,
Hui

On Wed, May 6, 2009 at 13:31, Hui Zhu <teawater@gmail.com> wrote:
> Hi Marc,
>
> I cannot reproduce this issue with cvs-head.
> Could you please help me test with this issue?
> If both of us cannot reproduce this issue,  I think I can give up this patch.
>
> Thanks,
> Hui
>
>
> On Mon, Mar 2, 2009 at 14:10, teawater <teawater@gmail.com> wrote:
>> Hi Michael,
>>
>> This patch is for bug in
>> http://sourceware.org/ml/gdb-patches/2009-02/msg00246.html.
>>
>> The reason of bug is when inferior reverse exection into dynsym
>> resolve code, gdb doesn't insert breakpoint in right address.
>> It just happen in some distributions for example SLED.
>>
>> This patch make gdb insert breakpoint in right address or step in that case.
>>
>> 2008-03-02  Hui Zhu  <teawater@gmail.com>
>>
>>       * infrun.c (handle_inferior_event): Make inferior insert
>>       breakpoint at right address or step in reverse and it stepped
>>       into dynsym resolve code.
>>
>> Thanks,
>> Hui
>>
>

[-- Attachment #2: fix-reverse-dynsym-error.txt --]
[-- Type: text/plain, Size: 1409 bytes --]

---
 infrun.c |   31 ++++++++++++++++++++++++++-----
 1 file changed, 26 insertions(+), 5 deletions(-)

--- a/infrun.c
+++ b/infrun.c
@@ -3465,13 +3465,34 @@ infrun: BPSTAT_WHAT_SET_LONGJMP_RESUME (
 
       if (pc_after_resolver)
 	{
-	  /* Set up a step-resume breakpoint at the address
-	     indicated by SKIP_SOLIB_RESOLVER.  */
 	  struct symtab_and_line sr_sal;
-	  init_sal (&sr_sal);
-	  sr_sal.pc = pc_after_resolver;
+	  if (execution_direction == EXEC_REVERSE)
+	    {
+	      if (ecs->stop_func_start != stop_pc)
+		{
+		  /* Normal (staticly linked) function call return.  */
+		  init_sal (&sr_sal);
+		  sr_sal.pc = ecs->stop_func_start;
+		  insert_step_resume_breakpoint_at_sal (sr_sal, null_frame_id);
+		}
+	      else
+		{
+		  /* We are stepping over a function call in reverse, and
+	             stop at the start address of the function.  Go back to
+	             single-stepping, which should take us back to the
+	             function call.  */
+		  ecs->event_thread->stepping_over_breakpoint = 1;
+		}
+	    }
+	  else
+	    {
+	      /* Set up a step-resume breakpoint at the address
+	         indicated by SKIP_SOLIB_RESOLVER.  */
+	      init_sal (&sr_sal);
+	      sr_sal.pc = pc_after_resolver;
 
-	  insert_step_resume_breakpoint_at_sal (sr_sal, null_frame_id);
+	      insert_step_resume_breakpoint_at_sal (sr_sal, null_frame_id);
+	    }
 	}
 
       keep_going (ecs);

Re: [RFA] Patch to fix "reverse-step" command error

[Hui Zhu]

PING

On Wed, May 6, 2009 at 16:13, Hui Zhu <teawater@gmail.com> wrote:
> Hi Marc and Michael,
>
> Sorry I forget this issue just reproduce in SLED 10.
>
> I reproduced it with cvs-head and patch can fix it.  Michael, Please
> help me review it.
>
> The attachment is the new patch follow cvs-head.
>
> 2009-05-06  Hui Zhu  <teawater@gmail.com>
>
>       * infrun.c (handle_inferior_event): Make inferior insert
>       breakpoint at right address or step in reverse and it stepped
>       into dynsym resolve code.
>
> Thanks,
> Hui
>
> On Wed, May 6, 2009 at 13:31, Hui Zhu <teawater@gmail.com> wrote:
>> Hi Marc,
>>
>> I cannot reproduce this issue with cvs-head.
>> Could you please help me test with this issue?
>> If both of us cannot reproduce this issue,  I think I can give up this patch.
>>
>> Thanks,
>> Hui
>>
>>
>> On Mon, Mar 2, 2009 at 14:10, teawater <teawater@gmail.com> wrote:
>>> Hi Michael,
>>>
>>> This patch is for bug in
>>> http://sourceware.org/ml/gdb-patches/2009-02/msg00246.html.
>>>
>>> The reason of bug is when inferior reverse exection into dynsym
>>> resolve code, gdb doesn't insert breakpoint in right address.
>>> It just happen in some distributions for example SLED.
>>>
>>> This patch make gdb insert breakpoint in right address or step in that case.
>>>
>>> 2008-03-02  Hui Zhu  <teawater@gmail.com>
>>>
>>>       * infrun.c (handle_inferior_event): Make inferior insert
>>>       breakpoint at right address or step in reverse and it stepped
>>>       into dynsym resolve code.
>>>
>>> Thanks,
>>> Hui
>>>
>>
>

Re: [RFA] Patch to fix "reverse-step" command error

[Hui Zhu]

PING

On Mon, May 11, 2009 at 15:07, Hui Zhu<teawater@gmail.com> wrote:
> PING
>
> On Wed, May 6, 2009 at 16:13, Hui Zhu <teawater@gmail.com> wrote:
>> Hi Marc and Michael,
>>
>> Sorry I forget this issue just reproduce in SLED 10.
>>
>> I reproduced it with cvs-head and patch can fix it.  Michael, Please
>> help me review it.
>>
>> The attachment is the new patch follow cvs-head.
>>
>> 2009-05-06  Hui Zhu  <teawater@gmail.com>
>>
>>       * infrun.c (handle_inferior_event): Make inferior insert
>>       breakpoint at right address or step in reverse and it stepped
>>       into dynsym resolve code.
>>
>> Thanks,
>> Hui
>>
>> On Wed, May 6, 2009 at 13:31, Hui Zhu <teawater@gmail.com> wrote:
>>> Hi Marc,
>>>
>>> I cannot reproduce this issue with cvs-head.
>>> Could you please help me test with this issue?
>>> If both of us cannot reproduce this issue,  I think I can give up this patch.
>>>
>>> Thanks,
>>> Hui
>>>
>>>
>>> On Mon, Mar 2, 2009 at 14:10, teawater <teawater@gmail.com> wrote:
>>>> Hi Michael,
>>>>
>>>> This patch is for bug in
>>>> http://sourceware.org/ml/gdb-patches/2009-02/msg00246.html.
>>>>
>>>> The reason of bug is when inferior reverse exection into dynsym
>>>> resolve code, gdb doesn't insert breakpoint in right address.
>>>> It just happen in some distributions for example SLED.
>>>>
>>>> This patch make gdb insert breakpoint in right address or step in that case.
>>>>
>>>> 2008-03-02  Hui Zhu  <teawater@gmail.com>
>>>>
>>>>       * infrun.c (handle_inferior_event): Make inferior insert
>>>>       breakpoint at right address or step in reverse and it stepped
>>>>       into dynsym resolve code.
>>>>
>>>> Thanks,
>>>> Hui
>>>>
>>>
>>
>

Re: [RFA] Patch to fix "reverse-step" command error

[Michael Snyder]

Hui Zhu wrote:
> PING

Hui, Marc, I just posted a new patch that should replace this one.
Please review http://sourceware.org/ml/gdb-patches/2009-06/msg00551.html

Thanks,
Michael

RE: [RFA] Patch to fix "reverse-step" command error

[Marc Khouzam]

> -----Original Message-----
> From: Michael Snyder [mailto:msnyder@vmware.com] 
> Sent: June-21-09 7:12 PM
> To: Hui Zhu
> Cc: Marc Khouzam; gdb-patches@sourceware.org
> Subject: Re: [RFA] Patch to fix "reverse-step" command error
> 
> Hui Zhu wrote:
> > PING
> 
> Hui, Marc, I just posted a new patch that should replace this one.
> Please review 
> http://sourceware.org/ml/gdb-patches/2009-06/msg00551.html

Thanks for that.

I don't have my SLED right now, which is where I could see the
problem.  I may be able to get my hands on one the week after next.
If Hui hasn't been able to test this until that time, then 
I will give it a try myself.

Thanks again.


> 
> Thanks,
> Michael
> 

Re: [RFA] Patch to fix "reverse-step" command error

[Hui Zhu]

On Mon, Jun 22, 2009 at 09:51, Marc Khouzam<marc.khouzam@ericsson.com> wrote:
>> -----Original Message-----
>> From: Michael Snyder [mailto:msnyder@vmware.com]
>> Sent: June-21-09 7:12 PM
>> To: Hui Zhu
>> Cc: Marc Khouzam; gdb-patches@sourceware.org
>> Subject: Re: [RFA] Patch to fix "reverse-step" command error
>>
>> Hui Zhu wrote:
>> > PING
>>
>> Hui, Marc, I just posted a new patch that should replace this one.
>> Please review
>> http://sourceware.org/ml/gdb-patches/2009-06/msg00551.html
>
> Thanks for that.
>
> I don't have my SLED right now, which is where I could see the
> problem.  I may be able to get my hands on one the week after next.
> If Hui hasn't been able to test this until that time, then
> I will give it a try myself.
>
> Thanks again.
>
Don't warry about it, Marc.  I keep SLED 10 in virtualbox.
I will try this patch in SLED 10.

Thanks,
Hui

Re: [RFA] Patch to fix "reverse-step" command error

[Michael Snyder]

Marc Khouzam wrote:
>> -----Original Message-----
>> From: Michael Snyder [mailto:msnyder@vmware.com] 
>> Sent: June-21-09 7:12 PM
>> To: Hui Zhu
>> Cc: Marc Khouzam; gdb-patches@sourceware.org
>> Subject: Re: [RFA] Patch to fix "reverse-step" command error
>>
>> Hui Zhu wrote:
>>> PING
>> Hui, Marc, I just posted a new patch that should replace this one.
>> Please review 
>> http://sourceware.org/ml/gdb-patches/2009-06/msg00551.html
> 
> Thanks for that.
> 
> I don't have my SLED right now, which is where I could see the
> problem.  I may be able to get my hands on one the week after next.
> If Hui hasn't been able to test this until that time, then 
> I will give it a try myself.

By the way, I was able to see the mis-behavior on RHEL4.

[RFA] Patch to fix "reverse-step" command error

[teawater]

Hi Michael,

This patch is for bug in
http://sourceware.org/ml/gdb-patches/2009-02/msg00246.html.

The reason of bug is when inferior reverse exection into dynsym
resolve code, gdb doesn't insert breakpoint in right address.
It just happen in some distributions for example SLED.

This patch make gdb insert breakpoint in right address or step in that case.

2008-03-02  Hui Zhu  <teawater@gmail.com>

	* infrun.c (handle_inferior_event): Make inferior insert
	breakpoint at right address or step in reverse and it stepped
	into dynsym resolve code.

Thanks,
Hui