From mboxrd@z Thu Jan 1 00:00:00 1970 From: marc.zyngier@arm.com (Marc Zyngier) Date: Fri, 03 Mar 2017 10:17:28 +0000 Subject: [PATCH/RFC] arm64: pmu: add Qualcomm Technologies extensions In-Reply-To: <2c6fcb54-482a-14dc-0fbb-58df0f9a68b5@codeaurora.org> (Neil Leeder's message of "Thu, 2 Mar 2017 14:30:53 -0500") References: <1488385085-19238-1-git-send-email-nleeder@codeaurora.org> <20170301181032.GL28874@leverpostej> <4b90b76f-f08d-f248-8153-287b2552dceb@arm.com> <2c6fcb54-482a-14dc-0fbb-58df0f9a68b5@codeaurora.org> Message-ID: <8737eu4qbr.fsf@on-the-bus.cambridge.arm.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thu, Mar 02 2017 at 7:30:53 pm GMT, "Leeder, Neil" wrote: > Hi Mark Z., > > On 3/2/2017 4:05 AM, Marc Zyngier wrote: >> On 01/03/17 21:36, Leeder, Neil wrote: >>> On 3/1/2017 1:10 PM, Mark Rutland wrote: >>>> KVM already has (architected) PMU support, and without a corresponding >>>> KVM patch this is at best insufficient. I don't imagine the KVM folk >>>> will be too thrilled about the prospect of emulating an IMPLEMENTATION >>>> DEFINED CPU feature like this. >>> >>> Does KVM handle ARMv7 PMU implementations? If so, do you know what it >>> does for the scorpion_* and krait_* implementations in >>> arch/arm/kernel/perf_events_v7.c? These extensions in ARMv8 are very >>> similar to the krait extensions, with some 64-bit tweaks, so could be >>> handled by KVM the same way it handles the ARMv7 cases. >> >> No, KVM doesn't handle the ARMv7 PMU at all. I'm not aware of the >> virtualization extensions being available on Scorpion or Krait, which >> makes it a moot point. What it handles is the PMUv3 architecture. > > Thank you for the explanation. > > This driver is specifically for Qualcomm Technologies server > chips. They will not be in a heterogenous environment with > non-Qualcomm processors, so there should be no migration issues. How do you know that? I'm afraid this is not something you or I can guarantee (and even less enforce). > If we were to provide a patch which added KVM support for the 4 > additional registers here, would you consider reviewing it, or is > adding implementation-defined registers a show-stopper? At the moment, it seems that there is a consensus against adding support for an IMPDEF PMU. Thanks, M. -- Jazz is not dead, it just smell funny.