From mboxrd@z Thu Jan 1 00:00:00 1970 From: will.deacon@arm.com (Will Deacon) Date: Wed, 28 Oct 2015 15:21:31 +0000 Subject: [PATCH] arm64: ftrace: function_graph: dump real return addr in call trace In-Reply-To: <5628340F.5080902@huawei.com> References: <1444911155-17480-1-git-send-email-huawei.libin@huawei.com> <6277407.jveniKQDxt@wuerfel> <20151015125133.GA29301@arm.com> <20151015101812.7cb6e917@gandalf.local.home> <20151020153227.GL11226@e104818-lin.cambridge.arm.com> <5628340F.5080902@huawei.com> Message-ID: <20151028152130.GF18966@arm.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thu, Oct 22, 2015 at 08:55:43AM +0800, libin wrote: > ? 2015/10/20 23:32, Catalin Marinas ??: > >On Thu, Oct 15, 2015 at 10:18:12AM -0400, Steven Rostedt wrote: > >>On Thu, 15 Oct 2015 13:51:33 +0100 > >>Will Deacon wrote: > >> > >>>Is this the same old problem caused by e306dfd06fcb ("ARM64: unwind: Fix > >>>PC calculation")? I've said previously that I'm happy to revert that if > >>>we're the only architecture with this behaviour, but Akashi resisted > >>>because there are other issues with ftrace that he was hoping to address > >>>and they would resolve this too. > >> > >>Just a reference, but this patch is pretty much exactly what x86 > >>currently has. I wonder if I should make that function generic for all > >>archs to use. > >> > >>If you accept this patch, I can look at what archs do and pull out the > >>common code and place it into the core code and have the archs call > >>that instead. > > > >The difference I see from the sh and x86 version is that we have this -4 > >on arm64, introduced by e306dfd06fcb as Will mentioned above (it seemed > >to have caused more problems that it solved). I think we should revert > >that commit first just to be in line with other architectures and then > >apply additional fixes as needed. > > > >Question for Li Bin: is your patch still needed if we revert commit > >e306dfd06fcb? > > > > It still be needed, but it can be implemented in generic for all archs > as Steve suggested. Well, there's still an argument for reverting e306dfd06fcb because it makes us behave differently to other architectures (in particular, arch/arm). Will