xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
To: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Cc: ian.campbell@citrix.com, xen-devel@lists.xen.org
Subject: Re: [PATCH] xen/balloon: do not modify the p2m for auto_translate guests
Date: Fri, 13 Dec 2013 09:57:04 -0500	[thread overview]
Message-ID: <20131213145704.GA4770@phenom.dumpdata.com> (raw)
In-Reply-To: <alpine.DEB.2.02.1312131445220.8667@kaball.uk.xensource.com>

On Fri, Dec 13, 2013 at 02:47:45PM +0000, Stefano Stabellini wrote:
> On Fri, 13 Dec 2013, Konrad Rzeszutek Wilk wrote:
> > On Wed, Dec 11, 2013 at 04:58:42PM +0000, Stefano Stabellini wrote:
> > > decrease_reservation doesn't modify the p2m for auto_translate guests,
> > > but increase_reservation does.
> > > Fix that by avoiding any p2m modifications in both increase_reservation
> > > and decrease_reservation for auto_translated guests.
> > > 
> > > Avoid allocating or using scratch pages for auto_translated guests.
> > > 
> > > Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> > > ---
> > >  drivers/xen/balloon.c |   63 ++++++++++++++++++++++++++-----------------------
> > >  1 file changed, 34 insertions(+), 29 deletions(-)
> > > 
> > > diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c
> > > index 55ea73f..4c02e2b 100644
> > > --- a/drivers/xen/balloon.c
> > > +++ b/drivers/xen/balloon.c
> > > @@ -350,17 +350,19 @@ static enum bp_state increase_reservation(unsigned long nr_pages)
> > >  
> > >  		pfn = page_to_pfn(page);
> > >  
> > > -		set_phys_to_machine(pfn, frame_list[i]);
> > > -
> > >  #ifdef CONFIG_XEN_HAVE_PVMMU
> > > -		/* Link back into the page tables if not highmem. */
> > > -		if (xen_pv_domain() && !PageHighMem(page)) {
> > > -			int ret;
> > > -			ret = HYPERVISOR_update_va_mapping(
> > > -				(unsigned long)__va(pfn << PAGE_SHIFT),
> > > -				mfn_pte(frame_list[i], PAGE_KERNEL),
> > > -				0);
> > > -			BUG_ON(ret);
> > > +		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
> > > +			set_phys_to_machine(pfn, frame_list[i]);
> > > +
> > > +			/* Link back into the page tables if not highmem. */
> > > +			if (!PageHighMem(page)) {
> > > +				int ret;
> > > +				ret = HYPERVISOR_update_va_mapping(
> > > +						(unsigned long)__va(pfn << PAGE_SHIFT),
> > > +						mfn_pte(frame_list[i], PAGE_KERNEL),
> > > +						0);
> > > +				BUG_ON(ret);
> > > +			}
> > >  		}
> > >  #endif
> > >  
> > > @@ -378,7 +380,6 @@ static enum bp_state decrease_reservation(unsigned long nr_pages, gfp_t gfp)
> > >  	enum bp_state state = BP_DONE;
> > >  	unsigned long  pfn, i;
> > >  	struct page   *page;
> > > -	struct page   *scratch_page;
> > >  	int ret;
> > >  	struct xen_memory_reservation reservation = {
> > >  		.address_bits = 0,
> > > @@ -411,27 +412,29 @@ static enum bp_state decrease_reservation(unsigned long nr_pages, gfp_t gfp)
> > >  
> > >  		scrub_page(page);
> > >  
> > > +#ifdef CONFIG_XEN_HAVE_PVMMU
> > >  		/*
> > >  		 * Ballooned out frames are effectively replaced with
> > >  		 * a scratch frame.  Ensure direct mappings and the
> > >  		 * p2m are consistent.
> > >  		 */
> > > -		scratch_page = get_balloon_scratch_page();
> > > -#ifdef CONFIG_XEN_HAVE_PVMMU
> > > -		if (xen_pv_domain() && !PageHighMem(page)) {
> > > -			ret = HYPERVISOR_update_va_mapping(
> > > -				(unsigned long)__va(pfn << PAGE_SHIFT),
> > > -				pfn_pte(page_to_pfn(scratch_page),
> > > -					PAGE_KERNEL_RO), 0);
> > > -			BUG_ON(ret);
> > > -		}
> > > -#endif
> > >  		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
> > >  			unsigned long p;
> > > +			struct page   *scratch_page = get_balloon_scratch_page();
> > > +
> > > +			if (!PageHighMem(page)) {
> > 
> > How come you removed the 'xen_pv_domain' part?
> 
> I thought that it's meaningless here. What we care about is whether it
> is an auto_translated guests or not, everything else is irrelevant.
> It also goes in the direction of using specific tests for things instead
> of catch all tests like xen_pv_domain that don't mean much anymore.

Please mention that in the commit description. Actually looking at
the description you could say:

"xen/balloon: Seperate the auto-translate logic properly.

The auto-xlat logic vs the non-xlat means that we don't need to for
auto-xlat guests (like PVH, HVM or ARM):
 - use P2M
 - use scratch page.

However the code in increase_reservation does modify the p2m for
auto_translate guests, but not in decrease_reservation.

Fix that by avoiding any p2m modifications in both increase_reservation
and decrease_reservation for auto_translated guests.

And also avoid allocating or using scratch pages for auto_translated guests.

Lastly, since !auto-xlat is really another way of saying 'xen_pv'
remove the redundant 'xen_pv_domain' check.
"

or so.

      reply	other threads:[~2013-12-13 14:57 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-12-11 16:58 [PATCH] xen/balloon: do not modify the p2m for auto_translate guests Stefano Stabellini
2013-12-11 17:12 ` Konrad Rzeszutek Wilk
2013-12-13 14:35 ` Konrad Rzeszutek Wilk
2013-12-13 14:47   ` Stefano Stabellini
2013-12-13 14:57     ` Konrad Rzeszutek Wilk [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20131213145704.GA4770@phenom.dumpdata.com \
    --to=konrad.wilk@oracle.com \
    --cc=ian.campbell@citrix.com \
    --cc=stefano.stabellini@eu.citrix.com \
    --cc=xen-devel@lists.xen.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).