From: Avi Kivity <avi@redhat.com>
To: Christoffer Dall <c.dall@virtualopensystems.com>
Cc: android-virt@lists.cs.columbia.edu, kvm@vger.kernel.org
Subject: Re: [PATCH v8 11/15] ARM: KVM: World-switch implementation
Date: Thu, 21 Jun 2012 11:13:45 +0300 [thread overview]
Message-ID: <4FE2D7B9.9060401@redhat.com> (raw)
In-Reply-To: <CANM98qKf8z0J3hp0yUhuQ1cmAH8VvJCo_+M9_QAS-XiAdwP_XQ@mail.gmail.com>
On 06/20/2012 07:40 AM, Christoffer Dall wrote:
>
>>>
>>> cpu0 cpu1
>>> (vmid=0, gen=1) (gen=0)
>>> -------------------------- ----------------------
>>> gen == global_gen, return
>>>
>>> gen != global_gen
>>> increment global_gen
>>> smp_call_function
>>> reset_vm_context
>>> vmid=0
>>>
>>> enter with vmid=0 enter with vmid=0
>>
>> I can't see how the scenario above can happen. First, no-one can run
>> with vmid 0 - it is reserved for the host. If we bump global_gen on
>> cpuN, then since we do smp_call_function(x, x, wait=1) we are now sure
>> that after this call, all cpus(N-1) potentially being inside a VM will
>> have exited, called reset_vm_context, but before they can re-enter
>> into the guest, they will call update_vttbr, and if their generation
>> counter differs from global_gen, they will try to grab that spinlock
>> and everything should happen in order.
>>
>
> the whole vmid=0 confused me a bit. The point is since we moved the
> generation check outside the spin_lock we have to re-check, I see:
In fact I think the problem occured with the original code as well. The
problem is that the check is not atomic wrt guest entry, so
spin_lock
check/update
spin_unlock
enter guest
has a hole between spin_unlock and guest entry.
>
> /**
> + * check_new_vmid_gen - check that the VMID is still valid
> + * @kvm: The VM's VMID to checkt
> + *
> + * return true if there is a new generation of VMIDs being used
> + *
> + * The hardware supports only 256 values with the value zero reserved for the
> + * host, so we check if an assigned value belongs to a previous generation,
> + * which which requires us to assign a new value. If we're the first to use a
> + * VMID for the new generation, we must flush necessary caches and TLBs on all
> + * CPUs.
> + */
> +static bool check_new_vmid_gen(struct kvm *kvm)
> +{
> + if (likely(kvm->arch.vmid_gen == atomic64_read(&kvm_vmid_gen)))
> + return;
> +}
> +
> +/**
> * update_vttbr - Update the VTTBR with a valid VMID before the guest runs
> * @kvm The guest that we are about to run
> *
> @@ -324,15 +342,7 @@ static void update_vttbr(struct kvm *kvm)
> {
> phys_addr_t pgd_phys;
>
> - /*
> - * Check that the VMID is still valid.
> - * (The hardware supports only 256 values with the value zero
> - * reserved for the host, so we check if an assigned value belongs to
> - * a previous generation, which which requires us to assign a new
> - * value. If we're the first to use a VMID for the new generation,
> - * we must flush necessary caches and TLBs on all CPUs.)
> - */
> - if (likely(kvm->arch.vmid_gen == atomic64_read(&kvm_vmid_gen)))
> + if (!check_new_vmid_gen(kvm))
> return;
>
> spin_lock(&kvm_vmid_lock);
> @@ -504,6 +514,13 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu
> *vcpu, struct kvm_run *run)
> */
> preempt_disable();
> local_irq_disable();
> +
> + if (check_new_vmid_gen(kvm)) {
> + local_irq_enable();
> + preempt_enable();
> + continue;
> + }
> +
I see the same race with signals. Your signal_pending() check needs to
be after the local_irq_disable(), otherwise we can enter a guest with a
pending signal.
Better place the signal check before the vmid_gen check, to avoid the
possibility of a a signal being held up by other guests.
--
error compiling committee.c: too many arguments to function
next prev parent reply other threads:[~2012-06-21 8:13 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-06-15 19:06 [PATCH v8 00/15] KVM/ARM Implementation Christoffer Dall
2012-06-15 19:06 ` [PATCH v8 01/15] ARM: add mem_type prot_pte accessor Christoffer Dall
2012-06-15 19:07 ` [PATCH v8 02/15] KVM: use KVM_CAP_IRQ_ROUTING to protect the routing related code Christoffer Dall
2012-06-18 13:06 ` Avi Kivity
2012-06-15 19:07 ` [PATCH v8 03/15] KVM: Introduce __KVM_HAVE_IRQ_LINE Christoffer Dall
2012-06-18 13:07 ` Avi Kivity
2012-06-15 19:07 ` [PATCH v8 04/15] KVM: Guard mmu_notifier specific code with CONFIG_MMU_NOTIFIER Christoffer Dall
2012-06-18 13:08 ` Avi Kivity
2012-06-18 17:47 ` Christoffer Dall
2012-06-19 8:37 ` Avi Kivity
2012-06-28 21:28 ` Marcelo Tosatti
2012-06-15 19:07 ` [PATCH v8 05/15] ARM: KVM: Initial skeleton to compile KVM support Christoffer Dall
2012-06-15 19:07 ` [PATCH v8 06/15] ARM: KVM: Hypervisor identity mapping Christoffer Dall
2012-06-18 13:12 ` Avi Kivity
2012-06-18 17:55 ` Christoffer Dall
2012-06-19 8:38 ` Avi Kivity
2012-06-15 19:07 ` [PATCH v8 07/15] ARM: KVM: Hypervisor inititalization Christoffer Dall
2012-06-28 22:35 ` Marcelo Tosatti
2012-06-28 22:53 ` Christoffer Dall
2012-06-29 1:07 ` Marcelo Tosatti
2012-06-15 19:08 ` [PATCH v8 08/15] ARM: KVM: Module unloading support Christoffer Dall
2012-06-15 19:08 ` [PATCH v8 09/15] ARM: KVM: Memory virtualization setup Christoffer Dall
2012-06-21 12:29 ` Gleb Natapov
2012-06-21 19:48 ` Christoffer Dall
2012-06-28 22:34 ` Marcelo Tosatti
2012-06-28 22:51 ` Christoffer Dall
2012-06-15 19:08 ` [PATCH v8 10/15] ARM: KVM: Inject IRQs and FIQs from userspace Christoffer Dall
2012-06-18 13:32 ` Avi Kivity
2012-06-18 20:56 ` Christoffer Dall
2012-06-19 8:49 ` Avi Kivity
2012-06-20 3:17 ` Christoffer Dall
2012-06-15 19:08 ` [PATCH v8 11/15] ARM: KVM: World-switch implementation Christoffer Dall
2012-06-18 13:41 ` Avi Kivity
2012-06-18 22:05 ` Christoffer Dall
2012-06-19 9:16 ` Avi Kivity
2012-06-20 3:27 ` Christoffer Dall
2012-06-20 4:40 ` Christoffer Dall
2012-06-21 8:13 ` Avi Kivity [this message]
2012-06-21 17:54 ` Christoffer Dall
2012-07-02 13:07 ` Avi Kivity
2012-06-15 19:08 ` [PATCH v8 12/15] ARM: KVM: Emulation framework and CP15 emulation Christoffer Dall
2012-06-15 19:09 ` [PATCH v8 13/15] ARM: KVM: Handle guest faults in KVM Christoffer Dall
2012-06-18 13:45 ` Avi Kivity
2012-06-18 22:20 ` Christoffer Dall
2012-06-19 9:32 ` Avi Kivity
2012-06-19 10:41 ` Andrea Arcangeli
2012-06-20 15:13 ` Christoffer Dall
2012-06-20 17:49 ` Andrea Arcangeli
2012-06-15 19:09 ` [PATCH v8 14/15] ARM: KVM: Handle I/O aborts Christoffer Dall
2012-06-18 13:48 ` Avi Kivity
2012-06-18 22:28 ` Christoffer Dall
2012-06-15 19:09 ` [PATCH v8 15/15] ARM: KVM: Guest wait-for-interrupts (WFI) support Christoffer Dall
2012-06-28 21:49 ` [PATCH v8 00/15] KVM/ARM Implementation Marcelo Tosatti
2012-06-28 22:44 ` Christoffer Dall
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4FE2D7B9.9060401@redhat.com \
--to=avi@redhat.com \
--cc=android-virt@lists.cs.columbia.edu \
--cc=c.dall@virtualopensystems.com \
--cc=kvm@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).