This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH v2 3/3] remove all_lwps from gdbserver: remove all_lwps
- From: Pedro Alves <palves at redhat dot com>
- To: Doug Evans <dje at google dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Thu, 06 Feb 2014 12:27:59 +0000
- Subject: Re: [PATCH v2 3/3] remove all_lwps from gdbserver: remove all_lwps
- Authentication-results: sourceware.org; auth=none
- References: <yjt2lhxyd4ii dot fsf at ruffy dot mtv dot corp dot google dot com>
On 01/29/2014 06:31 PM, Doug Evans wrote:
>
> Regression tested on amd64-linux.
> I think I caught all the places that need updating in the other ports,
> but I can't test them. Can anyone else?
Have you pushed this to a branch somewhere? I find that makes
it easier for people to try things.
> Another issue is:
> Several functions have a struct thread_info pointer and a struct
> lwp_info pointer. That's because lwp_info is recorded in the
> target_data field of thread_info. thread_info "has a" lwp_info,
> as opposed to lwp_info "is a" thread_info.
> This problem already existed. This patch makes it a bit worse
> because previously the ptid was recorded in both thread_info and
> lwp_info and the code didn't need a thread_info pointer to get the ptid.
> I could have left ptid in lwp_info but that would have meant
> recording the same data in two different places, also not good.
> The benefit of the current approach is clarity: there is only one place
> to get the ptid of a thread.
Fine with me.
The patch all looks good to me.
--
Pedro Alves