[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] [PATCH v3 extra 11/11] x86/mm: move description of x86 page table API to pv/mm.c



Signed-off-by: Wei Liu <wei.liu2@xxxxxxxxxx>
---
 xen/arch/x86/mm.c    | 65 ----------------------------------------------------
 xen/arch/x86/pv/mm.c | 65 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 65 insertions(+), 65 deletions(-)

diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index ade3ed2c48..75c84d2275 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -18,71 +18,6 @@
  * along with this program; If not, see <http://www.gnu.org/licenses/>.
  */
 
-/*
- * A description of the x86 page table API:
- *
- * Domains trap to do_mmu_update with a list of update requests.
- * This is a list of (ptr, val) pairs, where the requested operation
- * is *ptr = val.
- *
- * Reference counting of pages:
- * ----------------------------
- * Each page has two refcounts: tot_count and type_count.
- *
- * TOT_COUNT is the obvious reference count. It counts all uses of a
- * physical page frame by a domain, including uses as a page directory,
- * a page table, or simple mappings via a PTE. This count prevents a
- * domain from releasing a frame back to the free pool when it still holds
- * a reference to it.
- *
- * TYPE_COUNT is more subtle. A frame can be put to one of three
- * mutually-exclusive uses: it might be used as a page directory, or a
- * page table, or it may be mapped writable by the domain [of course, a
- * frame may not be used in any of these three ways!].
- * So, type_count is a count of the number of times a frame is being
- * referred to in its current incarnation. Therefore, a page can only
- * change its type when its type count is zero.
- *
- * Pinning the page type:
- * ----------------------
- * The type of a page can be pinned/unpinned with the commands
- * MMUEXT_[UN]PIN_L?_TABLE. Each page can be pinned exactly once (that is,
- * pinning is not reference counted, so it can't be nested).
- * This is useful to prevent a page's type count falling to zero, at which
- * point safety checks would need to be carried out next time the count
- * is increased again.
- *
- * A further note on writable page mappings:
- * -----------------------------------------
- * For simplicity, the count of writable mappings for a page may not
- * correspond to reality. The 'writable count' is incremented for every
- * PTE which maps the page with the _PAGE_RW flag set. However, for
- * write access to be possible the page directory entry must also have
- * its _PAGE_RW bit set. We do not check this as it complicates the
- * reference counting considerably [consider the case of multiple
- * directory entries referencing a single page table, some with the RW
- * bit set, others not -- it starts getting a bit messy].
- * In normal use, this simplification shouldn't be a problem.
- * However, the logic can be added if required.
- *
- * One more note on read-only page mappings:
- * -----------------------------------------
- * We want domains to be able to map pages for read-only access. The
- * main reason is that page tables and directories should be readable
- * by a domain, but it would not be safe for them to be writable.
- * However, domains have free access to rings 1 & 2 of the Intel
- * privilege model. In terms of page protection, these are considered
- * to be part of 'supervisor mode'. The WP bit in CR0 controls whether
- * read-only restrictions are respected in supervisor mode -- if the
- * bit is clear then any mapped page is writable.
- *
- * We get round this by always setting the WP bit and disallowing
- * updates to it. This is very unlikely to cause a problem for guest
- * OS's, which will generally use the WP bit to simplify copy-on-write
- * implementation (in that case, OS wants a fault when it writes to
- * an application-supplied buffer).
- */
-
 #include <xen/init.h>
 #include <xen/kernel.h>
 #include <xen/lib.h>
diff --git a/xen/arch/x86/pv/mm.c b/xen/arch/x86/pv/mm.c
index ad35808c51..39e6a3bc9a 100644
--- a/xen/arch/x86/pv/mm.c
+++ b/xen/arch/x86/pv/mm.c
@@ -20,6 +20,71 @@
  * along with this program; If not, see <http://www.gnu.org/licenses/>.
  */
 
+/*
+ * A description of the x86 page table API:
+ *
+ * Domains trap to do_mmu_update with a list of update requests.
+ * This is a list of (ptr, val) pairs, where the requested operation
+ * is *ptr = val.
+ *
+ * Reference counting of pages:
+ * ----------------------------
+ * Each page has two refcounts: tot_count and type_count.
+ *
+ * TOT_COUNT is the obvious reference count. It counts all uses of a
+ * physical page frame by a domain, including uses as a page directory,
+ * a page table, or simple mappings via a PTE. This count prevents a
+ * domain from releasing a frame back to the free pool when it still holds
+ * a reference to it.
+ *
+ * TYPE_COUNT is more subtle. A frame can be put to one of three
+ * mutually-exclusive uses: it might be used as a page directory, or a
+ * page table, or it may be mapped writable by the domain [of course, a
+ * frame may not be used in any of these three ways!].
+ * So, type_count is a count of the number of times a frame is being
+ * referred to in its current incarnation. Therefore, a page can only
+ * change its type when its type count is zero.
+ *
+ * Pinning the page type:
+ * ----------------------
+ * The type of a page can be pinned/unpinned with the commands
+ * MMUEXT_[UN]PIN_L?_TABLE. Each page can be pinned exactly once (that is,
+ * pinning is not reference counted, so it can't be nested).
+ * This is useful to prevent a page's type count falling to zero, at which
+ * point safety checks would need to be carried out next time the count
+ * is increased again.
+ *
+ * A further note on writable page mappings:
+ * -----------------------------------------
+ * For simplicity, the count of writable mappings for a page may not
+ * correspond to reality. The 'writable count' is incremented for every
+ * PTE which maps the page with the _PAGE_RW flag set. However, for
+ * write access to be possible the page directory entry must also have
+ * its _PAGE_RW bit set. We do not check this as it complicates the
+ * reference counting considerably [consider the case of multiple
+ * directory entries referencing a single page table, some with the RW
+ * bit set, others not -- it starts getting a bit messy].
+ * In normal use, this simplification shouldn't be a problem.
+ * However, the logic can be added if required.
+ *
+ * One more note on read-only page mappings:
+ * -----------------------------------------
+ * We want domains to be able to map pages for read-only access. The
+ * main reason is that page tables and directories should be readable
+ * by a domain, but it would not be safe for them to be writable.
+ * However, domains have free access to rings 1 & 2 of the Intel
+ * privilege model. In terms of page protection, these are considered
+ * to be part of 'supervisor mode'. The WP bit in CR0 controls whether
+ * read-only restrictions are respected in supervisor mode -- if the
+ * bit is clear then any mapped page is writable.
+ *
+ * We get round this by always setting the WP bit and disallowing
+ * updates to it. This is very unlikely to cause a problem for guest
+ * OS's, which will generally use the WP bit to simplify copy-on-write
+ * implementation (in that case, OS wants a fault when it writes to
+ * an application-supplied buffer).
+ */
+
 #include <xen/event.h>
 #include <xen/guest_access.h>
 
-- 
2.11.0


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

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.