explicit checkpointing

Chris Jones jones.chris.g at gmail.com
Mon May 5 22:14:19 PDT 2014


On Mon, May 5, 2014 at 9:19 PM, Robert O'Callahan <robert at ocallahan.org>wrote:

> Am I missing anything? It sounds very easy.
>
>
​That's as far as I'd planned things out, and matches my thinking :).​

The one potential subtlety is ensuring we don't confuse gdb too much.  We
need to ensure that all the debugger breakpoints are cleared before we make
a new checkpoint (and potentially, we'll need to save/restore them too),
and after restoring a checkpoint, we'll probably have to notify gdb of all
the new task statuses.  (Remember, to the gdb client it'll look like the
whole world changed behind its back.)

gdb is pretty good about managing breakpoints, so I don't expect that to be
a problem.

Cheers,
Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/rr-dev/attachments/20140505/f18a4740/attachment.html>


More information about the rr-dev mailing list