Re: [PATCH] target/arm/kvm: Retry KVM_CREATE_VM call if it fails EINTR

[Peter Maydell]

On Mon, 26 Sept 2022 at 19:08, Vitaly Chikunov <vt@altlinux.org> wrote:
>
> Peter,
>
> On Mon, Sep 26, 2022 at 02:36:44PM +0100, Peter Maydell wrote:
> > Occasionally the KVM_CREATE_VM ioctl can return EINTR, even though
> > there is no pending signal to be taken. In commit 94ccff13382055
> > we added a retry-on-EINTR loop to the KVM_CREATE_VM call in the
> > generic KVM code. Adopt the same approach for the use of the
> > ioctl in the Arm-specific KVM code (where we use it to create a
> > scratch VM for probing for various things).
> >
> > For more information, see the mailing list thread:
> > https://lore.kernel.org/qemu-devel/8735e0s1zw.wl-maz@kernel.org/
> >
> > Reported-by: Vitaly Chikunov <vt@altlinux.org>
> > Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> > ---
> > The view in the thread seems to be that this is a kernel bug (because
> > in QEMU's case there shouldn't be a signal to be delivered at this
> > point because of our signal handling strategy); so I've adopted the
> > same "just retry-on-EINTR for this specific ioctl" approach that
> > commit 94ccff13 did, rather than, for instance, something wider like
> > "make kvm_ioctl() and friends always retry on EINTR".
> > ---
> >  target/arm/kvm.c | 4 +++-
> >  1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/target/arm/kvm.c b/target/arm/kvm.c
> > index e5c1bd50d29..2982d216176 100644
> > --- a/target/arm/kvm.c
> > +++ b/target/arm/kvm.c
> > @@ -79,7 +79,9 @@ bool kvm_arm_create_scratch_host_vcpu(const uint32_t 
> > *cpus_to_try,
> >      if (max_vm_pa_size < 0) {
> >          max_vm_pa_size = 0;
> >      }
> > -    vmfd = ioctl(kvmfd, KVM_CREATE_VM, max_vm_pa_size);
> > +    do {
> > +        vmfd = ioctl(kvmfd, KVM_CREATE_VM, max_vm_pa_size);
> > +    } while (vmfd == -EINTR);
>
> This does not seem correct. ioctl(2) returns -1 on error and will set errno
> to EINTR (in this case).

Oops, thanks for the catch. I was copying the pattern from kvm-all.c,
but there we are calling kvm_ioctl(), which does "if ioctl
returns -1, return -errno", and here we are calling ioctl() directly.
(We can't use kvm_ioctl() here because this stunt VM doesn't have
a KVMState.) I'll respin the patch with the obvious fix.

thanks
-- PMM