[RFA take 2, 1 of 2] Improve many-shared-libraries performance

[Gary Benson <gbenson at redhat dot com>]

Hi all,

A week or so ago I proposed a patch to improve GDB's performance with
inferiors that load a lot of shared libraries.  My patch hinged around
avoiding calls to find_pc_section for solib event breakpoints.  There
were two separate places in handle_inferior_event that (indirectly)
call find_pc_section, and Pedro suggested I make them both lazy rather
than looking specifically for solib event breakpoints and avoiding the
calls only there.

This patch handles part of this, by calling find_pc_partial_function
lazily.  Since the two cases are orthogonal I wanted to get this out
there for review/approval while I work on the second case, which is
proving to be more difficult.  If nothing else this keeps the patches
simpler, but this patch has the potential to improve performance in
other areas in any event.

Is it ok to commit?

Cheers,
Gary

-- 
http://gbenson.net/

Attachment: patch
Description: Text document