From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753033AbaKLOhj (ORCPT ); Wed, 12 Nov 2014 09:37:39 -0500 Received: from mga11.intel.com ([192.55.52.93]:16442 "EHLO mga11.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752946AbaKLOhh convert rfc822-to-8bit (ORCPT ); Wed, 12 Nov 2014 09:37:37 -0500 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.97,862,1389772800"; d="scan'208";a="415420760" From: "Liang, Kan" To: Jiri Olsa CC: "acme@kernel.org" , "a.p.zijlstra@chello.nl" , "eranian@google.com" , "linux-kernel@vger.kernel.org" , "mingo@redhat.com" , "paulus@samba.org" , "ak@linux.intel.com" Subject: RE: [PATCH 2/2] perf tools: Construct LBR call chain Thread-Topic: [PATCH 2/2] perf tools: Construct LBR call chain Thread-Index: AQHP+dQE2S0iXFhhE0OnjR0eOmOh6JxcNBgAgADU+tA= Date: Wed, 12 Nov 2014 14:37:13 +0000 Message-ID: <37D7C6CF3E00A74B8858931C1DB2F07701661906@SHSMSX103.ccr.corp.intel.com> References: <1415285886-16949-1-git-send-email-kan.liang@intel.com> <1415285886-16949-3-git-send-email-kan.liang@intel.com> <20141112085914.GC25338@krava> In-Reply-To: <20141112085914.GC25338@krava> Accept-Language: zh-CN, en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.239.127.40] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 8BIT MIME-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > > + > > + printf("... chain: nr:%" PRIu64 "\n", total_nr); > > + > > + for (i = 0; i < callchain_nr + 1; i++) > > printf("..... %2d: %016" PRIx64 "\n", > > i, sample->callchain->ips[i]); > > so if there's lbr callstack info we dont display user stack part from standard > callchain? I think the dump code should dump out all the info.. > Right, we don't display user stack part from fp if there is lbr callstack info. The lbr callstack info can only be captured when the user set --call-graph lbr. If --call-graph is set to fp and dwarf, there will be no lbr callstack info. If the user set lbr, I think he really want the lbr info. So I think if we display both lbr and fp, the fp chain might be meaningless and it will confuse them. If the user want to do compare, they can do perf record twice with different --call-graph. Thanks, Kan > jirka