Re: More on 2.6.28 and balloon driver

[Jeremy Fitzhardinge <jeremy@goop.org>]

Dan Magenheimer wrote:
>> It appears that in the upstream balloon driver,
>> the call to HYPERVISOR_update_va_mapping is missing
>> from decrease_reservation.  I think as a result,
>> the balloon driver is eating memory but not
>> releasing it to Xen, thus rendering the balloon
>> driver essentially useless.  (Can be observed via xentop.)
>>     
>
> Limited testing, but it appears that adding it
> back in with this simple patch to linux-2.6.28
> makes ballooning work properly.
>   

Thanks.  That's a pretty embarrassing oversight :/

> As the code is lifted from the original Xen balloon driver,
> I'm not sure I need this, but:
>
> Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com>
>   

Yeah, the S-o-b line is a way to keep track of who a patch went via.  
It's "I assert that I'm allowed to post this patch".  In theory you'd 
include the previous S-o-b lines (if any), but a reference to the source 
repo/version would be enough in this case.

    J

> Dan
>
> P.S. Attachement is same, provided in case my mailer
> messes up this >80-char-width patch.
>
>
> --- linux-2.6.28/drivers/xen/balloon.c.orig	2009-01-07 12:35:48.000000000 -0700
> +++ linux-2.6.28/drivers/xen/balloon.c	2009-01-07 12:36:06.000000000 -0700
> @@ -296,6 +296,11 @@
>  		frame_list[i] = pfn_to_mfn(pfn);
>  
>  		scrub_page(page);
> +
> +		ret = HYPERVISOR_update_va_mapping(
> +			(unsigned long)__va(pfn << PAGE_SHIFT),
> +			__pte_ma(0), 0);
> +		BUG_ON(ret);
>  	}
>  
>  	/* Ensure that ballooned highmem pages don't have kmaps. */
> ------------------------------------------------------------------------
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
>