From: Tamas K Lengyel <tklengyel@sec.in.tum.de>
To: xen-devel@lists.xen.org
Cc: ian.campbell@citrix.com, tim@xen.org, julien.grall@linaro.org,
ian.jackson@eu.citrix.com, stefano.stabellini@citrix.com,
andres@lagarcavilla.org, jbeulich@suse.com,
dgdegra@tycho.nsa.gov, Tamas K Lengyel <tklengyel@sec.in.tum.de>
Subject: [PATCH v3 12/15] xen/arm: Shatter large pages when using mem_acces
Date: Mon, 1 Sep 2014 16:22:06 +0200 [thread overview]
Message-ID: <1409581329-2607-13-git-send-email-tklengyel@sec.in.tum.de> (raw)
In-Reply-To: <1409581329-2607-1-git-send-email-tklengyel@sec.in.tum.de>
When using mem_acccess+mem_event, the coarsest granularity is required
in the pagetables, therefore we shatter each large page (1GB/2MB) as we
apply the permission changes.
Signed-off-by: Tamas K Lengyel <tklengyel@sec.in.tum.de>
---
xen/arch/arm/p2m.c | 25 +++++++++++++++++++++----
1 file changed, 21 insertions(+), 4 deletions(-)
diff --git a/xen/arch/arm/p2m.c b/xen/arch/arm/p2m.c
index e8f5671..0d447e2 100644
--- a/xen/arch/arm/p2m.c
+++ b/xen/arch/arm/p2m.c
@@ -1132,12 +1132,30 @@ static bool_t p2m_set_entry(struct domain *d, paddr_t paddr, p2m_access_t p2ma)
pte = &first[first_table_offset(paddr)];
if ( !p2m_table(*pte) )
- goto done;
+ {
+ /* This is a mapping of a 1GB page, shatter it */
+ rc = p2m_create_table(d, pte, FIRST_SHIFT - PAGE_SHIFT, 1);
+ if ( rc < 0 )
+ goto err;
+
+ p2m->stats.shattered[1]++;
+ p2m->stats.mappings[1]--;
+ p2m->stats.mappings[2] += LPAE_ENTRIES;
+ }
second = map_domain_page(pte->p2m.base);
pte = &second[second_table_offset(paddr)];
if ( !p2m_table(*pte) )
- goto done;
+ {
+ /* This is a mapping of 2MB page, shatter it */
+ rc = p2m_create_table(d, pte, SECOND_SHIFT - PAGE_SHIFT, 1);
+ if ( rc < 0 )
+ goto err;
+
+ p2m->stats.shattered[2]++;
+ p2m->stats.mappings[2]--;
+ p2m->stats.mappings[3] += LPAE_ENTRIES;
+ }
third = map_domain_page(pte->p2m.base);
pte = &third[third_table_offset(paddr)];
@@ -1148,7 +1166,6 @@ static bool_t p2m_set_entry(struct domain *d, paddr_t paddr, p2m_access_t p2ma)
if ( !p2m_table(*pte) )
pte->bits = 0;
-done:
if ( p2m_valid(*pte) )
{
ASSERT(pte->p2m.type != p2m_invalid);
@@ -1159,11 +1176,11 @@ done:
rc = 1;
}
+err:
if (third) unmap_domain_page(third);
if (second) unmap_domain_page(second);
if (first) unmap_domain_page(first);
-err:
return rc;
}
--
2.1.0.rc1
next prev parent reply other threads:[~2014-09-01 14:22 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-09-01 14:21 [PATCH v3 00/15] Mem_event and mem_access for ARM Tamas K Lengyel
2014-09-01 14:21 ` [PATCH v3 01/15] xen: Relocate mem_access and mem_event into common Tamas K Lengyel
2014-09-01 15:06 ` Jan Beulich
2014-09-01 15:15 ` Tamas K Lengyel
2014-09-01 14:21 ` [PATCH v3 02/15] xen: Relocate struct npfec definition " Tamas K Lengyel
2014-09-01 15:44 ` Jan Beulich
2014-09-01 14:21 ` [PATCH v3 03/15] xen: Relocate mem_event_op domctl and access_op memop " Tamas K Lengyel
2014-09-01 15:46 ` Jan Beulich
2014-09-01 16:25 ` Tamas K Lengyel
2014-09-02 6:30 ` Jan Beulich
2014-09-02 7:43 ` Tamas K Lengyel
2014-09-01 18:11 ` Julien Grall
2014-09-01 20:51 ` Tamas K Lengyel
2014-09-02 6:53 ` Jan Beulich
2014-09-02 7:41 ` Tamas K Lengyel
2014-09-01 14:21 ` [PATCH v3 04/15] xen/mem_event: Clean out superfluous white-spaces Tamas K Lengyel
2014-09-01 14:21 ` [PATCH v3 05/15] xen/mem_event: Relax error condition on debug builds Tamas K Lengyel
2014-09-01 15:47 ` Jan Beulich
2014-09-01 14:22 ` [PATCH v3 06/15] xen/mem_event: Abstract architecture specific sanity checks Tamas K Lengyel
2014-09-01 14:22 ` [PATCH v3 07/15] xen/mem_access: Abstract architecture specific sanity check Tamas K Lengyel
2014-09-01 15:50 ` Jan Beulich
2014-09-01 14:22 ` [PATCH v3 08/15] xen/arm: p2m type definitions and changes Tamas K Lengyel
2014-09-01 14:22 ` [PATCH v3 09/15] xen/arm: Add set access required domctl Tamas K Lengyel
2014-09-01 19:10 ` Julien Grall
2014-09-02 7:48 ` Tamas K Lengyel
2014-09-02 8:17 ` Jan Beulich
2014-09-02 9:23 ` Tamas K Lengyel
2014-09-01 14:22 ` [PATCH v3 10/15] xen/arm: Data abort exception (R/W) mem_events Tamas K Lengyel
2014-09-01 21:07 ` Julien Grall
2014-09-02 9:06 ` Tamas K Lengyel
2014-09-03 20:20 ` Julien Grall
2014-09-03 21:56 ` Tamas K Lengyel
2014-09-08 20:41 ` Julien Grall
2014-09-09 9:20 ` Ian Campbell
2014-09-09 13:08 ` Tamas K Lengyel
2014-09-01 14:22 ` [PATCH v3 11/15] xen/arm: Instruction prefetch abort (X) mem_event handling Tamas K Lengyel
2014-09-01 14:22 ` Tamas K Lengyel [this message]
2014-09-01 14:22 ` [PATCH v3 13/15] xen/arm: Enable the compilation of mem_access and mem_event on ARM Tamas K Lengyel
2014-09-03 14:38 ` Daniel De Graaf
2014-09-01 14:22 ` [PATCH v3 14/15] tools/libxc: Allocate magic page for mem access " Tamas K Lengyel
2014-09-01 14:22 ` [PATCH v3 15/15] tools/tests: Enable xen-access " Tamas K Lengyel
2014-09-01 21:26 ` Julien Grall
2014-09-02 8:49 ` Tamas K Lengyel
2014-09-02 12:15 ` Tamas K Lengyel
2014-09-03 20:27 ` Julien Grall
2014-09-03 22:06 ` Tamas K Lengyel
2014-09-01 19:56 ` [PATCH v3 00/15] Mem_event and mem_access for ARM Julien Grall
2014-09-02 9:47 ` Tamas K Lengyel
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=1409581329-2607-13-git-send-email-tklengyel@sec.in.tum.de \
--to=tklengyel@sec.in.tum.de \
--cc=andres@lagarcavilla.org \
--cc=dgdegra@tycho.nsa.gov \
--cc=ian.campbell@citrix.com \
--cc=ian.jackson@eu.citrix.com \
--cc=jbeulich@suse.com \
--cc=julien.grall@linaro.org \
--cc=stefano.stabellini@citrix.com \
--cc=tim@xen.org \
--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).