Re: ioreq server regression

[Boris Ostrovsky <boris.ostrovsky@oracle.com>]

On 12/12/2016 11:36 AM, Andrew Cooper wrote:
> On 12/12/16 16:01, Boris Ostrovsky wrote:
>> Looks like commit e7dabe59c ("x86/hvm: don't unconditionally create a
>> default ioreq server") breaks restore for HVM guests.
>>
>> I see "qemu: hardware error: Invalid ioreq type 0x53" in qemu log.
>>
>> -boris
>>
> Does this help?  The current creation_finished logic is wrong if a 2nd
> toolstack component performs a pause/unpause pair on the domain.

Yes, this seems to work.

-boris

>
> diff --git a/xen/common/domain.c b/xen/common/domain.c
> index 3abaca9..6043f89 100644
> --- a/xen/common/domain.c
> +++ b/xen/common/domain.c
> @@ -1004,13 +1004,6 @@ int domain_unpause_by_systemcontroller(struct
> domain *d)
>  {
>      int old, new, prev = d->controller_pause_count;
>  
> -    /*
> -     * We record this information here for populate_physmap to figure out
> -     * that the domain has finished being created. In fact, we're only
> -     * allowed to set the MEMF_no_tlbflush flag during VM creation.
> -     */
> -    d->creation_finished = true;
> -
>      do
>      {
>          old = prev;
> @@ -1022,6 +1015,14 @@ int domain_unpause_by_systemcontroller(struct
> domain *d)
>          prev = cmpxchg(&d->controller_pause_count, old, new);
>      } while ( prev != old );
>  
> +    /*
> +     * We record this information here for populate_physmap to figure out
> +     * that the domain has finished being created. In fact, we're only
> +     * allowed to set the MEMF_no_tlbflush flag during VM creation.
> +     */
> +    if ( prev == 0 )
> +        d->creation_finished = true;
> +
>      domain_unpause(d);
>  
>      return 0;
>



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel