Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-03 Thread Paul E. McKenney
On Tue, Jun 02, 2009 at 09:53:29PM -0400, Gregory Haskins wrote: Paul E. McKenney wrote: On Tue, Jun 02, 2009 at 02:23:14PM -0400, Gregory Haskins wrote: Paul E. McKenney wrote: On Tue, Jun 02, 2009 at 11:15:38AM -0400, Gregory Haskins wrote: Assigning an irqfd

Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-03 Thread Gregory Haskins
Thanks again for the review, Paul. IIUC, you think the design is ok as it is. Davide, In light of this, would you like to submit patch 1/2 formally with your SOB at your earliest convenience? Or would you prefer that I submit it and you can simply ack it? Either is fine with me. -Greg

Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-03 Thread Davide Libenzi
On Wed, 3 Jun 2009, Gregory Haskins wrote: Thanks again for the review, Paul. IIUC, you think the design is ok as it is. Davide, In light of this, would you like to submit patch 1/2 formally with your SOB at your earliest convenience? Or would you prefer that I submit it and you

[KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-02 Thread Gregory Haskins
Assigning an irqfd object to a kvm object creates a relationship that we currently manage by having the kvm oject acquire/hold a file* reference to the underlying eventfd. The lifetime of these objects is properly maintained by decoupling the two objects whenever the irqfd is closed or kvm is

Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-02 Thread Davide Libenzi
On Tue, 2 Jun 2009, Gregory Haskins wrote: @@ -64,12 +101,28 @@ irqfd_wakeup(wait_queue_t *wait, unsigned mode, int sync, void *key) { struct _irqfd *irqfd = container_of(wait, struct _irqfd, wait); - /* - * The wake_up is called with interrupts disabled. Therefore we

Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-02 Thread Gregory Haskins
Davide Libenzi wrote: On Tue, 2 Jun 2009, Gregory Haskins wrote: @@ -64,12 +101,28 @@ irqfd_wakeup(wait_queue_t *wait, unsigned mode, int sync, void *key) { struct _irqfd *irqfd = container_of(wait, struct _irqfd, wait); -/* - * The wake_up is called with interrupts

Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-02 Thread Paul E. McKenney
On Tue, Jun 02, 2009 at 11:15:38AM -0400, Gregory Haskins wrote: Assigning an irqfd object to a kvm object creates a relationship that we currently manage by having the kvm oject acquire/hold a file* reference to the underlying eventfd. The lifetime of these objects is properly maintained by

Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-02 Thread Gregory Haskins
Paul E. McKenney wrote: On Tue, Jun 02, 2009 at 11:15:38AM -0400, Gregory Haskins wrote: Assigning an irqfd object to a kvm object creates a relationship that we currently manage by having the kvm oject acquire/hold a file* reference to the underlying eventfd. The lifetime of these

Re: [KVM-RFC PATCH 2/2] kvm: use POLLHUP to close an irqfd instead of an explicit ioctl

2009-06-02 Thread Gregory Haskins
Paul E. McKenney wrote: On Tue, Jun 02, 2009 at 02:23:14PM -0400, Gregory Haskins wrote: Paul E. McKenney wrote: On Tue, Jun 02, 2009 at 11:15:38AM -0400, Gregory Haskins wrote: Assigning an irqfd object to a kvm object creates a relationship that we currently manage by