This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: per core memory for core files
- From: Tom Tromey <tromey at redhat dot com>
- To: Mike Stump <mikestump at comcast dot net>
- Cc: gdb-patches at sourceware dot org
- Date: Fri, 22 Jun 2012 12:05:12 -0600
- Subject: Re: per core memory for core files
- References: <FD83DF53-4260-460B-8745-37BA2C44D7D0@comcast.net>
>>>>> "Mike" == Mike Stump <mikestump@comcast.net> writes:
Mike> On machines with core unique memory, there is no way to represent that
Mike> memory presently. We can represent per core registers with .reg/<n>,
Mike> so, we logically, we want to use .data/<n> for memory relevant to the
Mike> core that .reg/<n> is relevant for. This patch does this.
I didn't see a reply to this.
The patch needs a ChangeLog entry.
Does it need a documentation change as well? I can't tell.
I actually don't understand the patch very much, I just thought I'd
reply since nobody else had.
Mike> + /* core specific sections never satisfy, unless named. */
GNU coding style.
Mike> {
Mike> + LONGEST r;
Mike> switch (object)
gdb puts a blank line between declarations and code.
Mike> + tid=strchr (p->the_bfd_section->name, '/');
GNU coding style.
Mike> + /* Figure out the thread for the section, and see if that
Mike> + thread is the current thread. */
Mike> + thread_id = atoi (tid+1);
GNU coding style.
Tom