* KVM: x86: Cancel pvclock_gtod_work on module removal
@ 2021-05-05 21:48 Thomas Gleixner
2021-05-06 11:35 ` Paolo Bonzini
0 siblings, 1 reply; 2+ messages in thread
From: Thomas Gleixner @ 2021-05-05 21:48 UTC (permalink / raw)
To: kvm; +Cc: Paolo Bonzini, Sean Christopherson, x86, LKML
Nothing prevents the following:
pvclock_gtod_notify()
queue_work(system_long_wq, &pvclock_gtod_work);
...
remove_module(kvm);
...
work_queue_run()
pvclock_gtod_work() <- UAF
Ditto for any other operation on that workqueue list head which touches
pvclock_gtod_work after module removal.
Cancel the work in kvm_arch_exit() to prevent that.
Fixes: 16e8d74d2da9 ("KVM: x86: notifier for clocksource changes")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
Found by inspection because of:
https://lkml.kernel.org/r/0000000000001d43ac05c0f5c6a0@google.com
See also:
https://lkml.kernel.org/r/20210505105940.190490250@infradead.org
TL;DR: Scheduling work with tk_core.seq write held is a bad idea.
---
arch/x86/kvm/x86.c | 1 +
1 file changed, 1 insertion(+)
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -8168,6 +8168,7 @@ void kvm_arch_exit(void)
cpuhp_remove_state_nocalls(CPUHP_AP_X86_KVM_CLK_ONLINE);
#ifdef CONFIG_X86_64
pvclock_gtod_unregister_notifier(&pvclock_gtod_notifier);
+ cancel_work_sync(&pvclock_gtod_work);
#endif
kvm_x86_ops.hardware_enable = NULL;
kvm_mmu_module_exit();
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: KVM: x86: Cancel pvclock_gtod_work on module removal
2021-05-05 21:48 KVM: x86: Cancel pvclock_gtod_work on module removal Thomas Gleixner
@ 2021-05-06 11:35 ` Paolo Bonzini
0 siblings, 0 replies; 2+ messages in thread
From: Paolo Bonzini @ 2021-05-06 11:35 UTC (permalink / raw)
To: Thomas Gleixner, kvm; +Cc: Sean Christopherson, x86, LKML
On 05/05/21 23:48, Thomas Gleixner wrote:
> Nothing prevents the following:
>
> pvclock_gtod_notify()
> queue_work(system_long_wq, &pvclock_gtod_work);
> ...
> remove_module(kvm);
> ...
> work_queue_run()
> pvclock_gtod_work() <- UAF
>
> Ditto for any other operation on that workqueue list head which touches
> pvclock_gtod_work after module removal.
>
> Cancel the work in kvm_arch_exit() to prevent that.
>
> Fixes: 16e8d74d2da9 ("KVM: x86: notifier for clocksource changes")
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> ---
> Found by inspection because of:
> https://lkml.kernel.org/r/0000000000001d43ac05c0f5c6a0@google.com
> See also:
> https://lkml.kernel.org/r/20210505105940.190490250@infradead.org
>
> TL;DR: Scheduling work with tk_core.seq write held is a bad idea.
> ---
> arch/x86/kvm/x86.c | 1 +
> 1 file changed, 1 insertion(+)
>
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -8168,6 +8168,7 @@ void kvm_arch_exit(void)
> cpuhp_remove_state_nocalls(CPUHP_AP_X86_KVM_CLK_ONLINE);
> #ifdef CONFIG_X86_64
> pvclock_gtod_unregister_notifier(&pvclock_gtod_notifier);
> + cancel_work_sync(&pvclock_gtod_work);
> #endif
> kvm_x86_ops.hardware_enable = NULL;
> kvm_mmu_module_exit();
>
Queued, thanks (with added Cc to stable).
Paolo
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2021-05-06 11:35 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-05-05 21:48 KVM: x86: Cancel pvclock_gtod_work on module removal Thomas Gleixner
2021-05-06 11:35 ` Paolo Bonzini
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).