From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 26598 invoked by alias); 26 Sep 2008 23:37:30 -0000 Received: (qmail 26590 invoked by uid 22791); 26 Sep 2008 23:37:29 -0000 X-Spam-Check-By: sourceware.org Received: from mail.codesourcery.com (HELO mail.codesourcery.com) (65.74.133.4) by sourceware.org (qpsmtpd/0.31) with ESMTP; Fri, 26 Sep 2008 23:36:53 +0000 Received: (qmail 24514 invoked from network); 26 Sep 2008 23:36:51 -0000 Received: from unknown (HELO orlando.local) (pedro@127.0.0.2) by mail.codesourcery.com with ESMTPA; 26 Sep 2008 23:36:51 -0000 From: Pedro Alves To: gdb-patches@sourceware.org, tromey@redhat.com Subject: Re: [RFA/Ada] Implement Ada tasking support (take 2) Date: Fri, 26 Sep 2008 23:37:00 -0000 User-Agent: KMail/1.9.9 Cc: Joel Brobecker References: <20080924174850.GC3607@adacore.com> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200809270036.56455.pedro@codesourcery.com> X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2008-09/txt/msg00536.txt.bz2 On Saturday 27 September 2008 00:22:04, Tom Tromey wrote: > I thought I heard an argument to the effect that thread-specificity > could be more efficiently implemented at a low level, at least in > theory, on some platforms. =A0(I don't know. =A0And that is mighty vague.) That is usually brought about in the remote target context. Currently, a thread specific breakpoint is implemented by letting any thread hit it, and then having GDB's inferior control side check if it was the thread the user specified --- if so, stop; or, if some other thread hit it --- then resume immediatelly. If the thread associated with the breakpoint is communicated to the lower target layers, we can minimize this overhead by having the target itself filter those uninteresting breakpoint hits. This extra overhead is of course more visible when you have a serial communication protocol in the middle. Currently, the remote protocol has no support for thread-specific breakpoints, though, so usually talking about this ends up with a 'IWBN'. --=20 Pedro Alves