[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAKiXHKeZNdXXMakcYcMxQqKUrQX4Dj+M1XPt=sc-Wjeba5W8-Q@mail.gmail.com>
Date: Wed, 7 Jan 2026 16:12:08 +0100
From: Alessandro Ratti <alessandro@...5c.net>
To: Sean Christopherson <seanjc@...gle.com>
Cc: Alessandro Ratti <alessandro@...5c.net>, pbonzini@...hat.com, kvm@...r.kernel.org,
linux-kernel@...r.kernel.org,
syzbot+1522459a74d26b0ac33a@...kaller.appspotmail.com
Subject: Re: [PATCH] KVM: x86: Retry guest entry on -EBUSY from kvm_check_nested_events()
On Tue, 6 Jan 2026 at 18:34, Sean Christopherson <seanjc@...gle.com> wrote:
Hi Sean,
>
> On Tue, Jan 06, 2026, Sean Christopherson wrote:
> > On Sun, Jan 04, 2026, Alessandro Ratti wrote:
> > I'll post the below as part of a series, as there is at least one cleanup that
> > can be done on top to consolidate handling of EBUSY,
>
> Gah, I was wrong. I was thinking that morphing EBUSY to '0' could be moved into
> a common helper, but kvm_apic_accept_events() needs to bail immediately on EBUSY,
> i.e. needs to see EBUSY, not '0'.
>
> I'll post this as a standalone patch and then try to add WARNs in a separate
> series.
>
> > and I'm hopeful that the spirit of the WARN can be preserved, e.g. by
> > adding/extending WARNs in paths where KVM (re)injects events.
>
Thanks for the detailed review and for working out the correct fix.
I completely missed the mp_state interaction and learned a lot from
your explanation.
I'm interested in contributing more to KVM. If you have any pointers
on good places to start learning the codebase (e.g. particular areas,
bug reports, or documentation), I'd really appreciate it.
Thanks again for your time.
Alessandro
Powered by blists - more mailing lists