From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
To: David Vrabel <david.vrabel@citrix.com>
Cc: dcrisan@flexiant.com, xen-devel@lists.xensource.com,
Ian.Campbell@citrix.com, alex@alex.org.uk,
Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Subject: Re: [PATCH v2 1/2] xen/balloon: set a mapping for ballooned out pages
Date: Tue, 23 Jul 2013 09:58:26 -0400 [thread overview]
Message-ID: <20130723135826.GB3795@phenom.dumpdata.com> (raw)
In-Reply-To: <51ED6317.2060707@citrix.com>
On Mon, Jul 22, 2013 at 05:51:35PM +0100, David Vrabel wrote:
> On 22/07/13 17:28, Stefano Stabellini wrote:
> > Currently ballooned out pages are mapped to 0 and have INVALID_P2M_ENTRY
> > in the p2m. These ballooned out pages are used to map foreign grants
> > by gntdev and blkback (see alloc_xenballooned_pages).
> >
> > Allocate a page per cpu and map all the ballooned out pages to the
> > corresponding mfn. Set the p2m accordingly. This way reading from a
> > ballooned out page won't cause a kernel crash (see
> > http://lists.xen.org/archives/html/xen-devel/2012-12/msg01154.html).
>
> Generally in favour of this approach.
>
> I can't help thinking the "trade page" doesn't really mean anything but
> I can't think of a better, more descriptive name.
scratch page?
>
> > --- a/drivers/xen/balloon.c
> > +++ b/drivers/xen/balloon.c
> > @@ -36,6 +36,7 @@
> > * IN THE SOFTWARE.
> > */
> >
> > +#include <linux/cpu.h>
> > #include <linux/kernel.h>
> > #include <linux/sched.h>
> > #include <linux/errno.h>
> > @@ -88,6 +89,8 @@ EXPORT_SYMBOL_GPL(balloon_stats);
> >
> > /* We increase/decrease in batches which fit in a page */
> > static xen_pfn_t frame_list[PAGE_SIZE / sizeof(unsigned long)];
> > +static struct page** balloon_trade_pages;
>
> static DECLARE_PER_CPU(struct page *, balloon_trade_pages);
>
> > #ifdef CONFIG_HIGHMEM
> > #define inc_totalhigh_pages() (totalhigh_pages++)
> > @@ -423,7 +426,8 @@ static enum bp_state decrease_reservation(unsigned long nr_pages, gfp_t gfp)
> > if (xen_pv_domain() && !PageHighMem(page)) {
> > ret = HYPERVISOR_update_va_mapping(
> > (unsigned long)__va(pfn << PAGE_SHIFT),
> > - __pte_ma(0), 0);
> > + pfn_pte(page_to_pfn(get_balloon_trade_page()),
> > + PAGE_KERNEL_RO), 0);
>
> Preemption needs to be disabled while using the trade page, see
> suggestion below.
>
> > +struct page* get_balloon_trade_page(void)
> > +{
> > + BUG_ON(balloon_trade_pages[smp_processor_id()] == NULL);
> > + return balloon_trade_pages[smp_processor_id()];
> > +}
>
> I think you need a get_balloon_trade_page() and put_balloon_trade_page()
> pair that call get_cpu_var() and put_cpu_var() to ensure that preemption
> is disabled while using it.
>
> David
next prev parent reply other threads:[~2013-07-23 13:58 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-22 16:21 [PATCH v2 0/2] make ballooned out pages have a valid mapping at all times Stefano Stabellini
2013-07-22 16:28 ` [PATCH v2 1/2] xen/balloon: set a mapping for ballooned out pages Stefano Stabellini
2013-07-22 16:51 ` David Vrabel
2013-07-22 17:22 ` Ian Campbell
2013-07-22 17:26 ` David Vrabel
2013-07-22 17:32 ` Ian Campbell
2013-07-23 9:58 ` David Vrabel
2013-07-23 14:27 ` Ian Campbell
2013-07-23 15:03 ` David Vrabel
2013-07-23 17:04 ` Ian Campbell
2013-07-23 17:20 ` Stefano Stabellini
2013-07-23 17:33 ` David Vrabel
2013-07-23 12:49 ` Stefano Stabellini
2013-07-23 14:24 ` Ian Campbell
2013-07-23 13:58 ` Konrad Rzeszutek Wilk [this message]
2013-07-22 16:28 ` [PATCH v2 2/2] xen/m2p: use GNTTABOP_unmap_and_replace to reinstate the original mapping Stefano Stabellini
2013-07-22 17:02 ` David Vrabel
2013-07-23 10:35 ` Roger Pau Monné
2013-07-23 12:37 ` Stefano Stabellini
2013-07-23 12:47 ` Stefano Stabellini
2013-07-23 13:57 ` Konrad Rzeszutek Wilk
2013-07-23 14:23 ` Ian Campbell
2013-07-23 14:48 ` Alex Bligh
2013-07-23 14:51 ` Ian Campbell
2013-07-23 15:06 ` Alex Bligh
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=20130723135826.GB3795@phenom.dumpdata.com \
--to=konrad.wilk@oracle.com \
--cc=Ian.Campbell@citrix.com \
--cc=alex@alex.org.uk \
--cc=david.vrabel@citrix.com \
--cc=dcrisan@flexiant.com \
--cc=stefano.stabellini@eu.citrix.com \
--cc=xen-devel@lists.xensource.com \
/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).