From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Shishkin Subject: Re: [PATCH v13 08/12] KVM: x86: Add Intel PT context switch for each vcpu Date: Wed, 31 Oct 2018 13:46:24 +0200 Message-ID: <87o9ba5ofj.fsf@ashishki-desk.ger.corp.intel.com> References: <1540368316-12998-1-git-send-email-luwei.kang@intel.com> <1540368316-12998-9-git-send-email-luwei.kang@intel.com> <87a7n37iuf.fsf@ashishki-desk.ger.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain Cc: Luwei Kang , kvm@vger.kernel.org, x86@kernel.org, mingo@redhat.com, bp@alien8.de, hpa@zytor.com, rkrcmar@redhat.com, joro@8bytes.org, songliubraving@fb.com, peterz@infradead.org, kstewart@linuxfoundation.org, gregkh@linuxfoundation.org, thomas.lendacky@amd.com, konrad.wilk@oracle.com, mattst88@gmail.com, Janakarajan.Natarajan@amd.com, dwmw@amazon.co.uk, jpoimboe@redhat.com, marcorr@google.com, ubizjak@gmail.com, sean.j.christopherson@intel.com, jmattson@google.com, linux-kernel@vger.kernel.org, Chao Peng To: Paolo Bonzini , Thomas Gleixner Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org Paolo Bonzini writes: > On 30/10/2018 11:00, Thomas Gleixner wrote: >> So at least we need a way for perf on the host to programmatically detect, >> that 'guest traces itself' is enabled, so it can inject that information >> into the host data and post processing can tell that. W/o something like >> that it's going to be a FAQ. > > In guest-tracing mode there will be already a TIP.PGD and TIP.PGE packet > respectively before vmentry and after vmexit, caused by the RTIT_CTL > WRMSRs in pt_guest_enter and pt_guest_exit. The target IP of the > packets will come from kvm-intel.ko. Most people aren't tracing the kernel, so they'd just get a PGD with no address and a PGE after the kvm is done without any indication of what happened in between. > In system mode instead you get a Paging Information Packet on > vmentry/vmexit, with bit 0 set in the third byte. You won't get it if > guest-side tracing is on (because tracing has been disabled by > pt_guest_enter and won't be re-enabled until pt_guest_exit). I don't > think it's correct to "fake" the PIP in guest-tracing mode, because > TIP.PGD should be followed immediately by TIP.PGE. Indeed, we should most definitely not fake PIP. Perf has RECORD_AUX, which already has PARTIAL flag that was introduced specifically because of kvm. Regards, -- Alex