Re: [patch] More suggestive displaced-stepping memory error message

[Pedro Alves <palves at redhat dot com>]

On 04/11/2012 12:31 AM, Yao Qi wrote:

> Probably, GDB can be smart enough to not
> use displaced stepping when get error on accessing scratch pad.  So far,
> displaced_step_prepare returns 1 for success and 0 for queued request.
> We can add one more return value -1 for `unable to do displaced
> stepping', so that caller can go non-displaced-stepping path.


We can't just go non-displaced-stepping path in non-stop with the
current code because nothing knows to pause all running threads
in order to be able to lift the breakpoint and step over it without other
threads missing it (like gdbserver knows, with linux-low.c:start_step_over
and all that).  My patch #1 "all-stop on top of non-stop" patch of
the itsets series goes in that direction.

-- 
Pedro Alves