This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 2/3] skip_prolgoue (amd64)
- From: Doug Evans <dje at google dot com>
- To: Tom Tromey <tromey at redhat dot com>
- Cc: Pedro Alves <palves at redhat dot com>, Yao Qi <yao at codesourcery dot com>, Mark Kettenis <mark dot kettenis at xs4all dot nl>, gdb-patches <gdb-patches at sourceware dot org>
- Date: Wed, 4 Dec 2013 10:30:30 -0800
- Subject: Re: [PATCH 2/3] skip_prolgoue (amd64)
- Authentication-results: sourceware.org; auth=none
- References: <1385735051-27558-1-git-send-email-yao at codesourcery dot com> <1385735051-27558-3-git-send-email-yao at codesourcery dot com> <201311291436 dot rATEaZ5Z030292 at glazunov dot sibelius dot xs4all dot nl> <201311291605 dot rATG5XVb030184 at glazunov dot sibelius dot xs4all dot nl> <52994E79 dot 4000004 at codesourcery dot com> <5299B9D0 dot 2020304 at redhat dot com> <529C37A2 dot 9000207 at codesourcery dot com> <529E9462 dot 9010001 at codesourcery dot com> <529F1B1F dot 2040606 at redhat dot com> <87ob4wr5hv dot fsf at fleche dot redhat dot com>
On Wed, Dec 4, 2013 at 7:38 AM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> "Pedro" == Pedro Alves <palves@redhat.com> writes:
>
> Pedro> Actually "non-stop", vs "all-stop" here isn't the ideal
> Pedro> predicate. The real predicate is "is any thread running".
> Pedro> "non-stop" is just being currently used in
> Pedro> prepare_execute_command as proxy for that, just because
> Pedro> that was the easiest.
>
> It seemed to me that the predicate must be "is any thread associated
> with this particular address space running?" -- but I wanted to ask if
> that makes sense, or if that was what you meant. This idea seems to
> open the door to finer-grained cache flushing.
It would be unfortunate to take an unnecessary perf hit when doing a
backtrace and then browsing the stack, or setting a breakpoint in
non-stop mode when, for example, I'm stopped in some thread for
whatever reason (and then other threads happen to have their own
events).
I think some kind of finer-grained flushing *could* be reasonable.
Implementing it may then make the cache "look like" it is a stack
cache and code cache. 1/2 :-)
[One could do something like record in each line an attribute
describing how/where it came from.]