Re: [PATCH 1/1] gdb: Avoid warning for the jump command inside an inline function.

[Bruno Larsen <blarsen@redhat.com>]

On 21/03/2023 15:05, Willgerodt, Felix wrote:
>> -----Original Message-----
>> From: Bruno Larsen <blarsen@redhat.com>
>> Sent: Freitag, 17. März 2023 14:34
>> To: Willgerodt, Felix <felix.willgerodt@intel.com>; gdb-
>> patches@sourceware.org
>> Subject: Re: [PATCH 1/1] gdb: Avoid warning for the jump command inside
>> an inline function.
>>
>> On 17/03/2023 13:56, Willgerodt, Felix wrote:
>>>> -----Original Message-----
>>>> From: Bruno Larsen <blarsen@redhat.com>
>>>> Sent: Freitag, 17. März 2023 11:14
>>>> To: Willgerodt, Felix <felix.willgerodt@intel.com>; gdb-
>>>> patches@sourceware.org
>>>> Cc: Cristian Sandu <cristian.sandu@intel.com>
>>>> Subject: Re: [PATCH 1/1] gdb: Avoid warning for the jump command
>> inside
>>>> an inline function.
>>>>
>>>> On 24/01/2023 16:19, Felix Willgerodt via Gdb-patches wrote:
>>>>> When stopped inside an inline function, trying to jump to a different line
>>>>> of the same function currently results in a warning about jumping to
>>>> another
>>>>> function.  Fix this by taking inline functions into account.
>>>>>
>>>>> Before:
>>>>>      Breakpoint 1, function_inline (x=510) at jump-inline.cpp:22
>>>>>      22        a = a + x;             /* inline-funct */
>>>>>      (gdb) j 21
>>>>>      Line 21 is not in `function_inline(int)'.  Jump anyway? (y or n)
>>>>>
>>>>> After:
>>>>>      Breakpoint 2, function_inline (x=510) at jump-inline.cpp:22
>>>>>      22        a = a + x;            /* inline-funct */
>>>>>      (gdb) j 21
>>>>>      Continuing at 0x400679.
>>>>>
>>>>>      Breakpoint 1, function_inline (x=510) at jump-inline.cpp:21
>>>>>      21        a += 1020 + a;                /* increment-funct */
>>>>>
>>>>> This was regression-tested on X86-64 Linux.
>>>>>
>>>>> Co-Authored-by: Cristian Sandu <cristian.sandu@intel.com>
>>>>> ---
>>>>>     gdb/infcmd.c                           |  3 +-
>>>>>     gdb/testsuite/gdb.base/jump-inline.c   | 30 +++++++++++++++++
>>>>>     gdb/testsuite/gdb.base/jump-inline.exp | 45
>>>> ++++++++++++++++++++++++++
>>>>>     3 files changed, 77 insertions(+), 1 deletion(-)
>>>>>     create mode 100644 gdb/testsuite/gdb.base/jump-inline.c
>>>>>     create mode 100644 gdb/testsuite/gdb.base/jump-inline.exp
>>>>>
>>>>> diff --git a/gdb/infcmd.c b/gdb/infcmd.c
>>>>> index fd88b8ca328..40414bc9260 100644
>>>>> --- a/gdb/infcmd.c
>>>>> +++ b/gdb/infcmd.c
>>>>> @@ -1091,7 +1091,8 @@ jump_command (const char *arg, int from_tty)
>>>>>
>>>>>       /* See if we are trying to jump to another function.  */
>>>>>       fn = get_frame_function (get_current_frame ());
>>>>> -  sfn = find_pc_function (sal.pc);
>>>>> +  sfn = find_pc_sect_containing_function (sal.pc,
>>>>> +					  find_pc_mapped_section (sal.pc));
>>>> Hi Felix,
>>>>
>>>> Thanks for doing this, it is a good improvement, but I don't know if
>>>> this is the best way to go about it. Is there a reason why
>>>> find_pc_function should not return inlined functions?
>>>>
>>>> I feel like most of the time we want to know the function, knowing if
>>>> we're in an inlined one would be desirable, but I might be wrong. Does
>>>> anyone know?
>>>>
>>> Hi Bruno,
>>>
>>> I don't know the details, but the comments in symtab.h are rather explicit
>>> about it, so I assume there is a reason:
>>>
>>>
>>> /* lookup the function symbol corresponding to the address.  The
>>>      return value will not be an inlined function; the containing
>>>      function will be returned instead.  */
>>>
>>> extern struct symbol *find_pc_function (CORE_ADDR);
>>>
>>> /* lookup the function symbol corresponding to the address and
>>>      section.  The return value will be the closest enclosing function,
>>>      which might be an inline function.  */
>>>
>>> extern struct symbol *find_pc_sect_containing_function
>>>     (CORE_ADDR pc, struct obj_section *section);
>> Hi Felix,
>>
>> I thought it was mostly a descriptive comment, rather than prescriptive.
>> I tested changing find_pc_function locally and there were only 2
>> regressions, which might just be broken assumptions, but our testsuite
>> is probably not very comprehensive on inlined functions, so I don't know
>> how representative this test actually is.
>>
> Hi Bruno,
>
> Could you share what changes you have done locally?
> Did you basically just change find_pc_function to call
> find_pc_sec_containing_function?

Hi Felix

Yeah, this is exactly what I did.

>
> I saw that there is another comment in blockframe.c about
> "backwards compatibility", but that has been in the code for ages and
> I couldn't find anything interesting related to it.
Backwards compatibility is part of the reason I'm hesitant to trust my 
results. Fedora has very new everything, so I don't get to trigger bugs 
in old compilers or setups if we dont have a test manually recreating 
that. The other part is some other compiler/language/combination of 
things that could go wrong. I hope Pedro can remember the original 
reasoning!
>
> When git blaming the comments in symtab.h, I saw they were added
> by Pedro a couple years ago:
> https://sourceware.org/git/?p=binutils-gdb.git;a=commit;f=gdb/symtab.h;h=cd2bb709940d33668fe6dbe8d4ffee0ed44c25e6
> Maybe he can help shed some light on this?
>
> Thanks,
> Felix
> Intel Deutschland GmbH
> Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
> Tel: +49 89 99 8853-0, www.intel.de <http://www.intel.de>
> Managing Directors: Christin Eisenschmid, Sharon Heck, Tiffany Doon Silva
> Chairperson of the Supervisory Board: Nicole Lau
> Registered Office: Munich
> Commercial Register: Amtsgericht Muenchen HRB 186928


-- 
Cheers,
Bruno