[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241003113050.22875-1-mankku@gmail.com>
Date: Thu, 3 Oct 2024 14:29:25 +0300
From: Markku Ahvenjärvi <mankku@...il.com>
To: seanjc@...gle.com
Cc: bp@...en8.de,
dave.hansen@...ux.intel.com,
hpa@...or.com,
janne.karhunen@...il.com,
kvm@...r.kernel.org,
linux-kernel@...r.kernel.org,
mankku@...il.com,
mingo@...hat.com,
pbonzini@...hat.com,
tglx@...utronix.de,
x86@...nel.org
Subject: Re: [PATCH 1/1] KVM: nVMX: update VPPR on vmlaunch/vmresume
Hi Sean,
> Talking to myself :-)
A good monologue. :)
As you said, PPR is updated constantly with kvm_vcpu_has_events() anyways.
Actually implicit updates, especially in kvm_apic_has_interrupt(), makes it a
bit difficult to follow when these updates take place. It is quite easy to
accidentally make an update in architecturally incorrect place. Could be
useful to make it explicit.
> Assuming it actually fixes your issue, this is what I'm planning on posting. I
> suspect KVM botches something when the deprivileged host is active, but given
> that the below will allow for additional cleanups, and practically speaking doesn't
> have any downsides, I don't see any reason to withhold the hack-a-fix. Though
> hopefully we'll someday figure out exactly what's broken.
It fixes the issue. Thanks a lot Sean!
Kind regards,
Markku
Powered by blists - more mailing lists