Re: [PATCHv2 5/6] gdb: add inferior-specific breakpoints and watchpoints

[Pedro Alves <pedro@palves.net>]

On 2023-02-03 4:55 p.m., Andrew Burgess wrote:
> Pedro Alves <pedro@palves.net> writes:
> 
>> On 2023-01-20 9:46 a.m., Andrew Burgess via Gdb-patches wrote:
>>> This commit extends the breakpoint mechanism to allow for inferior
>>> specific breakpoints (and watchpoints).
>>>
>>> As GDB gains better support for multiple connections, and so for
>>> running multiple (possibly unrelated) inferiors, then it is not hard
>>> to imagine that a user might wish to create breakpoints that apply to
>>> any thread in a single inferior.  To achieve this currently, the user
>>> would need to create a condition possibly making use of the $_inferior
>>> convenience variable, which, though functional, isn't the most user
>>> friendly.
>>
>> So IIUC, if you do:
>>
>> (gdb) break foo inferior 1
>>
>> and foo exists in other inferiors, the resulting breakpoint will still
>> have locations for the other inferiors.  Is that so?
>>
>> I think that instead, we should not get any location for other inferiors.
>> I.e., we should do location pre-filtering, simply by not looking for
>> locations in pspaces other than the inferior's.  I think that can be
>> done easily by simply passing the inferior's pspace as search_pspace argument
>> to decode_line_full.
> 
> Thanks for all your great feedback.  I'll work through the rest of your
> comments, but I want to follow up on this one.
> 
> Baris already raised the point of filtering per-inferior breakpoints so
> a to avoid adding extra locations.

Ah, missed that.  Apologies.

> 
> I 100% agree that this is something that we should do, and this is on my
> plan if/when I land this series.
> 
> But I don't see any particular difference between per-inferior and
> per-thread breakpoints in this regard.  Please correct me if I'm wrong,
> but a per-thread breakpoint will insert into every inferior despite only
> ever triggering for a single thread in a single inferior (per thread
> breakpoints are on the global thread-id).

That is a very good point.  Given that, I agree you don't need to change it
in this patch.

I filed it as a bug, here:
 https://sourceware.org/bugzilla/show_bug.cgi?id=30086

> 
> So, because this oddity already exists in GDB, my plan was to land this
> feature, then follow up with a patch to limit both per-thread and
> per-inferior breakpoints to only insert into the correct pspace.

Agreed.