[PATCH v3 00/29] Windows code sharing + bug fix

[Hannes Domani]

 Am Mittwoch, 8. April 2020, 22:33:18 MESZ hat Tom Tromey <tromey@adacore.com> Folgendes geschrieben:

> >>>>> "Tom" == Tom Tromey <tromey@adacore.com> writes:
>
> Tom> This is v3 of the series to share a lot of the Windows code between
> Tom> gdb and gdbserver.  It also fixes a bug that a customer reported; in
> Tom> fact this fix was the origin of this series.  See patch #11 for
> Tom> details on the bug.
>
> Tom> I think this addresses all the review comments.  It's somewhat hard to
> Tom> be sure since they were done in gerrit, and extracting comments from
> Tom> that is a pain.  Also I think I had sent v2 before I updated my
> Tom> scripts to preserve the gerrit change ID, so it is in gerrit twice.
> Tom> Anyway, the reviews happened around Nov 2019, so you can find some in
> Tom> the mailing list archives.
>
> Tom> I tested this largely by hand.  I also ran it through the AdaCore test
> Tom> suite, where it did ok.  (There were some issues, but I think they are
> Tom> largely unrelated; some things like gdb printing paths that the test
> Tom> suite didn't really recognize.)
>
> Tom> I've also updated this to handle the WOW64 stuff, but only in a
> Tom> minimal way -- I didn't try to make this work in gdbserver.
>
> I'm checking this in now.
> Let me know if there are any issues.

I seems I kinda missed to actually use Wow64SuspendThread when I implemented
the WOW64 stuff.
I never noticed any problems, and SuspendThread actually seems to work fine
for WOW64 processes, but that might just be luck (or depend on the Windows
version).

You moved the SuspendThread stuff now into nat/windows-nat.c, so that
complicates the situation a bit (but it's also the reason I noticed).

Maybe I should just try to implement the WOW64 stuff for gdbserver as well, and
then try to move more shared stuff into nat/windows-nat.c?


Hannes