On Thu, May 09, 2024 at 05:48:19PM +0800, Zhao Liu wrote: > Hi Daniel & Shaoqin, > > Since x86 also needs to implement PMU filter feature, though it uses > the different KVM ioctl, we can still make the QEMU API as general as > possible. > > To move forward with both ARM and x86, I'd like to discuss my API > thinking with you. ;-) > > On Mon, Apr 15, 2024 at 06:29:25PM +0100, Daniel P. Berrangé wrote: > > Date: Mon, 15 Apr 2024 18:29:25 +0100 > > From: "Daniel P. Berrangé" <berrange@xxxxxxxxxx> > > Subject: Re: [PATCH v9] arm/kvm: Enable support for > > KVM_ARM_VCPU_PMU_V3_FILTER > > > > On Mon, Apr 08, 2024 at 10:49:40PM -0400, Shaoqin Huang wrote: > > > The KVM_ARM_VCPU_PMU_V3_FILTER provides the ability to let the VMM decide > > > which PMU events are provided to the guest. Add a new option > > > `kvm-pmu-filter` as -cpu sub-option to set the PMU Event Filtering. > > > Without the filter, all PMU events are exposed from host to guest by > > > default. The usage of the new sub-option can be found from the updated > > > document (docs/system/arm/cpu-features.rst). > > > > > > Here is an example which shows how to use the PMU Event Filtering, when > > > we launch a guest by use kvm, add such command line: > > > > > > # qemu-system-aarch64 \ > > > -accel kvm \ > > > -cpu host,kvm-pmu-filter="D:0x11-0x11" > > > > I'm still against implementing this one-off custom parsed syntax > > for kvm-pmu-filter values. Once this syntax exists, we're locked > > into back-compatibility for multiple releases, and it will make > > a conversion to QAPI/JSON harder. > > Daniel, I understand you mean the new specific string format makes > external API support more complicated, right? > > What about the following options: > > 1. Firstly, add a feature flag option in "-cpu" to enable kvm_filter > feature for CPU: > > -cpu host,kvm-pmu-filter > > 2. Then use "-object kvm-pmu-event" to configure PMU event properties. > Since x86's PMU filter has very complex encoding rules, we need the > following three variants (one for general case, the other two are x86 > specific): > > - General format: > -object kvm-pmu-event,action=[allowed|denied],events=[event-list] > > e.g, as Shaoqin's example, > -object kvm-pmu-event,action=allowed,events=0x11-0x11,0x23-0x23 > -object kvm-pmu-event,action=denied,events=0x23-0x3a > > - x86 raw_event encoding format (for single raw format event encoding): > -object kvm-pmu-event,action=[allowed|denied],mode=0,select="0x01", > umask="0x3c",fixed-bitmap="0xffffffff" > > - x86 masked_event encoding format (for mutiple masked event encoding): > -object kvm-pmu-event,action=[allowed|denied],mode=masked,select="0x01", > mask="0x3c",match="0x11",exclude=true|false > > The whole API architecture looks more complex, but has the advantage of > being as general as possible and avoiding the introduction of new string > format parsing. > > What do you think? Because the most important thing about this feature > is the API design, welcome your comments! Please describe it in terms of a QAPI definition, as that's what we're striving for with all QEMU public interfaces. Once the QAPI design is agreed, then the -object mapping is trivial, as -object's JSON format supports arbitrary QAPI structures. With regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
