[Xen-devel] Re: [PATCH][RESEND] vmx-io-events.patch

[Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>]

This patch doesn't apply cleanly.

 -- Keir

On 7 Jun 2005, at 01:13, Arun Sharma wrote:

> - Handle the case where the VMX domains get events from ports other 
> than
>   IOPACKET_PORT (because of paravirtualized drivers)
>
> - Use clear_bit() to operate on evtchn_upcall_pending
>
> Signed-off-by: Xiaofeng Ling <xiaofeng.ling@xxxxxxxxx>
> Signed-off-by: Kevin Tian <kevin.tian@xxxxxxxxx>
> Signed-off-by: Arun Sharma <arun.sharma@xxxxxxxxx>
>
> --- trunk/xen/arch/x86/vmx_io.c (revision 1179)
> +++ trunk/xen/arch/x86/vmx_io.c (working copy)
> @@ -282,7 +282,7 @@
>
>      /* Note: VMX domains may need upcalls as well */
>      if (!ed->vcpu_info->evtchn_pending_sel)
> -        ed->vcpu_info->evtchn_upcall_pending = 0;
> +        clear_bit(0, &ed->vcpu_info->evtchn_upcall_pending);
>
>      /* clear the pending bit for IOPACKET_PORT */
>      return test_and_clear_bit(IOPACKET_PORT,
> @@ -310,10 +310,16 @@
>      extern void do_block();
>
>      do {
> -        do_block();
> +        if(!test_bit(IOPACKET_PORT,
> +            &current->domain->shared_info->evtchn_pending[0]))
> +            do_block();
>          vmx_check_events(current);
>          if (!test_bit(ARCH_VMX_IO_WAIT, 
> &current->arch.arch_vmx.flags))
>              break;
> +        /* Events other than IOPACKET_PORT might have woken us up. In 
> that
> +           case, safely go back to sleep. */
> +        clear_bit(IOPACKET_PORT>>5, 
> &current->vcpu_info->evtchn_pending_sel);
> +        clear_bit(0, &current->vcpu_info->evtchn_upcall_pending);
>      } while(1);
>  }

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel