From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from eu-smtp-delivery-151.mimecast.com (eu-smtp-delivery-151.mimecast.com. [185.58.86.151]) by gmr-mx.google.com with ESMTPS id c24si83006ljk.7.2020.12.11.06.19.10 for (version=TLS1_2 cipher=ECDHE-ECDSA-AES128-GCM-SHA256 bits=128/128); Fri, 11 Dec 2020 06:19:10 -0800 (PST) From: David Laight Subject: RE: [patch 14/30] drm/i915/pmu: Replace open coded kstat_irqs() copy Date: Fri, 11 Dec 2020 14:19:05 +0000 Message-ID: References: <20201210192536.118432146@linutronix.de> <20201210194043.957046529@linutronix.de> <87y2i4h54i.fsf@nanos.tec.linutronix.de> In-Reply-To: <87y2i4h54i.fsf@nanos.tec.linutronix.de> MIME-Version: 1.0 Content-Language: en-US Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable To: 'Thomas Gleixner' , Tvrtko Ursulin , LKML Cc: Peter Zijlstra , Marc Zyngier , Jani Nikula , Joonas Lahtinen , Rodrigo Vivi , David Airlie , Daniel Vetter , "intel-gfx@lists.freedesktop.org" , "dri-devel@lists.freedesktop.org" , "James E.J. Bottomley" , Helge Deller , afzal mohammed , "linux-parisc@vger.kernel.org" , Russell King , "linux-arm-kernel@lists.infradead.org" , Mark Rutland , Catalin Marinas , Will Deacon , Christian Borntraeger , Heiko Carstens , "linux-s390@vger.kernel.org" , Pankaj Bharadiya , Chris Wilson , Wambui Karuga , Linus Walleij , "linux-gpio@vger.kernel.org" , Lee Jones , Jon Mason , Dave Jiang , Allen Hubbe , "linux-ntb@googlegroups.com" , Lorenzo Pieralisi , Rob Herring , Bjorn Helgaas , Michal Simek , "linux-pci@vger.kernel.org" , Karthikeyan Mitran , Hou Zhiqiang , Tariq Toukan , "David S. Miller" , Jakub Kicinski , "netdev@vger.kernel.org" , "linux-rdma@vger.kernel.org" , Saeed Mahameed , Leon Romanovsky , Boris Ostrovsky , Juergen Gross , Stefano Stabellini , "xen-devel@lists.xenproject.org" List-ID: From: Thomas Gleixner > Sent: 11 December 2020 12:58 .. > > After my failed hasty sketch from last night I had a different one whic= h > > was kind of heuristics based (re-reading the upper dword and retrying i= f > > it changed on 32-bit). >=20 > The problem is that there will be two seperate modifications for the low > and high word. Several ways how the compiler can translate this, but the > problem is the same for all of them: >=20 > CPU 0 CPU 1 > load low > load high > add low, 1 > addc high, 0 > store low load high > --> NMI load low > load high and compare > store high >=20 > You can't catch that. If this really becomes an issue you need a > sequence counter around it. Or just two copies of the high word. Provided the accesses are sequenced: writer: =09load high:low =09add small_value,high:low =09store high =09store low =09store high_copy reader: =09load high_copy =09load low =09load high =09if (high !=3D high_copy) =09=09low =3D 0; The read value is always stale, so it probably doesn't matter that the value you have is one that is between the value when you started and that when you finished. =09David - Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1= PT, UK Registration No: 1397386 (Wales)