This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 2/2] Enable range stepping for ARM on GDBServer
- From: Yao Qi <qiyaoltc at gmail dot com>
- To: Pedro Alves <palves at redhat dot com>
- Cc: Antoine Tremblay <antoine dot tremblay at ericsson dot com>, "gdb-patches at sourceware dot org" <gdb-patches at sourceware dot org>
- Date: Sun, 18 Sep 2016 20:58:45 +0100
- Subject: Re: [PATCH 2/2] Enable range stepping for ARM on GDBServer
- Authentication-results: sourceware.org; auth=none
- References: <20160831171406.24057-1-antoine.tremblay@ericsson.com> <20160831171406.24057-2-antoine.tremblay@ericsson.com> <b5bc1e35-1840-a7ab-6392-cb29c4ee4b8f@redhat.com> <wwokwpiwyfct.fsf@ericsson.com> <3fdb7193-60c7-49c9-ccf5-bc040aa157ea@redhat.com>
On Wed, Aug 31, 2016 at 7:39 PM, Pedro Alves <palves@redhat.com> wrote:
> On 08/31/2016 07:15 PM, Antoine Tremblay wrote:
>
> It sounds like gdbserver's event starvation avoidance isn't really
> working on sss targets with range stepping enabled. E.g., are we
> doing the randomization too late?
>
We only randomize the events to be reported to GDB, however, we
don't randomize the pending events. Every time, GDBserver select
pending events from the first one in thread list, threads in the end of
the list are likely starved. This issue can be fixed by introducing pending
events randomization.
My patches are ready, and being regression tested. During the
regression tests, some other issues are found (without my patches), so
I need to take a look at them first.
--
Yao (齐尧)