From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:35487) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1dQT3N-0002Aa-1r for qemu-devel@nongnu.org; Thu, 29 Jun 2017 02:35:50 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1dQT3J-00023N-U6 for qemu-devel@nongnu.org; Thu, 29 Jun 2017 02:35:49 -0400 Received: from roura.ac.upc.es ([147.83.33.10]:44881) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1dQT3J-000239-Hc for qemu-devel@nongnu.org; Thu, 29 Jun 2017 02:35:45 -0400 From: =?utf-8?Q?Llu=C3=ADs_Vilanova?= References: <149855711979.4578.11884555663631849051.stgit@frigg.lan> <20170627175815.GA17639@flamenco> <878tkce4sm.fsf@frigg.lan> <20170628230725.GA9893@flamenco> Date: Thu, 29 Jun 2017 09:35:32 +0300 In-Reply-To: <20170628230725.GA9893@flamenco> (Emilio G. Cota's message of "Wed, 28 Jun 2017 19:07:25 -0400") Message-ID: <87tw2z713f.fsf@frigg.lan> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v9 0/7] trace: [tcg] Optimize per-vCPU tracing states with separate TB caches List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Emilio G. Cota" Cc: qemu-devel@nongnu.org, Stefan Hajnoczi , Eduardo Habkost Emilio G Cota writes: > On Wed, Jun 28, 2017 at 14:21:29 +0300, Llu=C3=ADs Vilanova wrote: >> Emilio G Cota writes: >> > I wanted to save you some time and sent a v9 yesterday with these >> > same changes -- although I see some changes in my v8 didn't make it >> > to your v9. For this iteration I only added more perf numbers to the >> > last patch, see here: >> > https://lists.gnu.org/archive/html/qemu-devel/2017-06/msg05764.html >>=20 >> Woops, sorry about the duplicate effort. >>=20 >> I just dropped the rename of variable trace_ds, since I prefer names who= se >> purpose can be more easily understood. > - I think trace_vcpu_dstate is way too long (IMO a short name + a comment > in the struct is better). > - You missed a change I made to patch 3: >> [cota: use CPU_TRACE_DSTATE_MAX_EVENTS instead of trace_get_vcpu_event_c= ount()] > Was that intentional? No it wasn't. I'll look into it for the next series. > - Would be nice to include the perf numbers I got, along with yours. I'll add them to the cover and commit message. Thanks, Lluis