From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:41114) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ayKoL-0006C8-K4 for qemu-devel@nongnu.org; Thu, 05 May 2016 11:03:45 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ayKo9-000390-JB for qemu-devel@nongnu.org; Thu, 05 May 2016 11:03:24 -0400 Received: from mail-wm0-x230.google.com ([2a00:1450:400c:c09::230]:36003) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ayKo8-0002zy-Al for qemu-devel@nongnu.org; Thu, 05 May 2016 11:03:17 -0400 Received: by mail-wm0-x230.google.com with SMTP id n129so24181145wmn.1 for ; Thu, 05 May 2016 08:02:57 -0700 (PDT) References: <1459870344-16773-1-git-send-email-alex.bennee@linaro.org> <1459870344-16773-5-git-send-email-alex.bennee@linaro.org> <572B5684.1030001@gmail.com> From: Alex =?utf-8?Q?Benn=C3=A9e?= In-reply-to: <572B5684.1030001@gmail.com> Date: Thu, 05 May 2016 16:03:04 +0100 Message-ID: <87k2j8r1hz.fsf@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Subject: Re: [Qemu-devel] [RFC v2 04/11] tcg: comment on which functions have to be called with tb_lock held List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Sergey Fedorov Cc: mttcg@listserver.greensocs.com, fred.konrad@greensocs.com, a.rigo@virtualopensystems.com, cota@braap.org, qemu-devel@nongnu.org, mark.burton@greensocs.com, pbonzini@redhat.com, jan.kiszka@siemens.com, rth@twiddle.net, peter.maydell@linaro.org, claudio.fontana@huawei.com, Peter Crosthwaite , Andreas =?utf-8?Q?F?= =?utf-8?Q?=C3=A4rber?= Sergey Fedorov writes: > On 05/04/16 18:32, Alex Bennée wrote: >> diff --git a/exec.c b/exec.c >> index f46e596..17f390e 100644 >> --- a/exec.c >> +++ b/exec.c >> @@ -826,6 +826,7 @@ int cpu_breakpoint_insert(CPUState *cpu, vaddr pc, int flags, >> { >> CPUBreakpoint *bp; >> >> + /* TODO: locking (RCU?) */ >> bp = g_malloc(sizeof(*bp)); >> >> bp->pc = pc; > > This comment is a little inconsistent. We should make access to > breakpoint and watchpoint lists to be thread-safe in all the functions > using them. So if we note this, it should be noted in all such places. > Also, it's probably not a good idea to put such comment just above > g_malloc() invocation, it could be a bit confusing. A bit more details > would also be nice. Good point. I could really do with some tests to exercise the debugging interface. I did some when I wrote the arm kvm GDB stuff (see 261f4d6d3e5445f887e070f047968e756c30cf06) but it is a) not plumbed in and b) not really a stress test which is what you want to be sure your handling is thread safe. > > Kind regards, > Sergey -- Alex Bennée