Third version patches for upstreaming grant table version 2

Third version patches for upstreaming grant table version 2

[annie li]

Hi

The following patches introduce and implement grant table version 2. 
This is the third version patches, and based on v3.2.0-rc1+. Changes 
from the previous patches are following:

Patch 1:
* removing gnttab_v1_ops declaration
* renaming update_grant_entry_v1 to gnttab_update_entry_v1

Patch3:
* changing following code

xen_raw_printk();
panic();

back to

panic();
* re-arrange "else if" format in gnttab_request_version function.


Descriptions for those patches:

1. In those patches, the grant table code supports both grant table v1 and v2 version, v2 is an extension from v1. Grant table of guest domain can be either v1 or v2 version, and every grant table entry on one guest should be the same version. 

2. Full page structure of grant table v2 play the same role as grant table v1. Although full page structure is different from v1, grant table 2 is totally backwards compatible with v1. Grant table is shared between guest and Xen, domu and dom0 all have their own grant table shared with Xen, and their grant table version should be set before any grants are activated. When domu grants an entry to dom0 to map a frame,following are steps: 
* domu introduces a grant entry by reference 
* domu informs dom0 the gref 
* dom0 sends hypercall to map frame through this reference, Xen copy shared entry to active entry and update frame 
* dom0 does its work and release the frame, Xen releases the entry. 
* domu redo those steps for a new cycle. 
Xen mapping process can be found in function __gnttab_map_grant_ref in Xen code: xen/common/grant_table.c

3. If dom0 supports grant table v2, guests run on it can either supports v1 or v2. Xen is responsible to judge what  version the guests are using. This is implemented in Xen code: xen/common/grant_table.c. Key word is:  rd->grant_table->gt_version. 

4. Grant table v2 has been supported by Xen for a long time, and receiver-side copying mechanism bases on this implementation. Netback and netfront driver can pick up this new feature to get better network performance and better CPU accounting.


Diff:

 arch/x86/xen/grant-table.c          |   42 ++++-
 drivers/xen/grant-table.c           |  354 
++++++++++++++++++++++++++++++-----
 include/xen/grant_table.h           |   10 +-
 include/xen/interface/grant_table.h |  167 ++++++++++++++++-
 include/xen/interface/xen.h         |    2 +
 5 files changed, 519 insertions(+), 56 deletions(-)

Shortlog:
Annie Li (4):
      xen/granttable: Introducing grant table V2 stucture
      xen/granttable: Refactor some code
      xen/granttable: Grant tables V2 implementation
      xen/granttable: Keep code format clean

Thanks
Annie

[PATCH 1/4] xen/granttable: Introducing grant table V2 stucture

[annie.li]

From: Annie Li <annie.li@oracle.com>

This patch introduces new structures of grant table V2, grant table V2 is an
extension from V1. Grant table is shared between guest and Xen, and Xen is
responsible to do corresponding work for grant operations, such as: figure
out guest's grant table version, perform different actions based on
different grant table version, etc. Although full-page structure of V2
is different from V1, it play the same role as V1.

Signed-off-by: Annie Li <annie.li@oracle.com>
---
 arch/x86/xen/grant-table.c          |    7 +-
 drivers/xen/grant-table.c           |  181 +++++++++++++++++++++++++++--------
 include/xen/grant_table.h           |    4 +-
 include/xen/interface/grant_table.h |  167 +++++++++++++++++++++++++++++++-
 include/xen/interface/xen.h         |    2 +
 5 files changed, 310 insertions(+), 51 deletions(-)

diff --git a/arch/x86/xen/grant-table.c b/arch/x86/xen/grant-table.c
index 6bbfd7a..c6ab2e7 100644
--- a/arch/x86/xen/grant-table.c
+++ b/arch/x86/xen/grant-table.c
@@ -64,10 +64,10 @@ static int unmap_pte_fn(pte_t *pte, struct page *pmd_page,
 
 int arch_gnttab_map_shared(unsigned long *frames, unsigned long nr_gframes,
 			   unsigned long max_nr_gframes,
-			   struct grant_entry **__shared)
+			   void **__shared)
 {
 	int rc;
-	struct grant_entry *shared = *__shared;
+	void *shared = *__shared;
 
 	if (shared == NULL) {
 		struct vm_struct *area =
@@ -83,8 +83,7 @@ int arch_gnttab_map_shared(unsigned long *frames, unsigned long nr_gframes,
 	return rc;
 }
 
-void arch_gnttab_unmap_shared(struct grant_entry *shared,
-			      unsigned long nr_gframes)
+void arch_gnttab_unmap_shared(void *shared, unsigned long nr_gframes)
 {
 	apply_to_page_range(&init_mm, (unsigned long)shared,
 			    PAGE_SIZE * nr_gframes, unmap_pte_fn, NULL);
diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c
index bf1c094..18355a5 100644
--- a/drivers/xen/grant-table.c
+++ b/drivers/xen/grant-table.c
@@ -53,7 +53,7 @@
 /* External tools reserve first few grant table entries. */
 #define NR_RESERVED_ENTRIES 8
 #define GNTTAB_LIST_END 0xffffffff
-#define GREFS_PER_GRANT_FRAME (PAGE_SIZE / sizeof(struct grant_entry))
+#define GREFS_PER_GRANT_FRAME (PAGE_SIZE / sizeof(struct grant_entry_v1))
 
 static grant_ref_t **gnttab_list;
 static unsigned int nr_grant_frames;
@@ -64,7 +64,63 @@ static DEFINE_SPINLOCK(gnttab_list_lock);
 unsigned long xen_hvm_resume_frames;
 EXPORT_SYMBOL_GPL(xen_hvm_resume_frames);
 
-static struct grant_entry *shared;
+static union {
+	struct grant_entry_v1 *v1;
+	void *addr;
+} gnttab_shared;
+
+/*This is a structure of function pointers for grant table*/
+struct gnttab_ops {
+	/*
+	 * Mapping a list of frames for storing grant entries. First input
+	 * parameter is used to storing grant table address when grant table
+	 * being setup, second parameter is the number of frames to map grant
+	 * table. Returning GNTST_okay means success and negative value means
+	 * failure.
+	 */
+	int (*map_frames)(unsigned long *, unsigned int);
+	/*
+	 * Release a list of frames which are mapped in map_frames for grant
+	 * entry status.
+	 */
+	void (*unmap_frames)(void);
+	/*
+	 * Introducing a valid entry into the grant table, granting the frame
+	 * of this grant entry to domain for accessing, or transfering, or
+	 * transitively accessing. First input parameter is reference of this
+	 * introduced grant entry, second one is domid of granted domain, third
+	 * one is the frame to be granted, and the last one is status of the
+	 * grant entry to be updated.
+	 */
+	void (*update_entry)(grant_ref_t, domid_t, unsigned long, unsigned);
+	/*
+	 * Stop granting a grant entry to domain for accessing. First input
+	 * parameter is reference of a grant entry whose grant access will be
+	 * stopped, second one is not in use now. If the grant entry is
+	 * currently mapped for reading or writing, just return failure(==0)
+	 * directly and don't tear down the grant access. Otherwise, stop grant
+	 * access for this entry and return success(==1).
+	 */
+	int (*end_foreign_access_ref)(grant_ref_t, int);
+	/*
+	 * Stop granting a grant entry to domain for transfer. If tranfer has
+	 * not started, just reclaim the grant entry and return failure(==0).
+	 * Otherwise, wait for the transfer to complete and then return the
+	 * frame.
+	 */
+	unsigned long (*end_foreign_transfer_ref)(grant_ref_t);
+	/*
+	 * Query the status of a grant entry. Input parameter is reference of
+	 * queried grant entry, return value is the status of queried entry.
+	 * Detailed status(writing/reading) can be gotten from the return value
+	 * by bit operations.
+	 */
+	int (*query_foreign_access)(grant_ref_t);
+};
+
+static struct gnttab_ops *gnttab_interface;
+
+static int grant_table_version;
 
 static struct gnttab_free_callback *gnttab_free_callback_list;
 
@@ -142,23 +198,23 @@ static void put_free_entry(grant_ref_t ref)
 	spin_unlock_irqrestore(&gnttab_list_lock, flags);
 }
 
-static void update_grant_entry(grant_ref_t ref, domid_t domid,
-			       unsigned long frame, unsigned flags)
+/*
+ * Introducing a valid entry into the grant table:
+ *  1. Write ent->domid.
+ *  2. Write ent->frame:
+ *      GTF_permit_access:   Frame to which access is permitted.
+ *      GTF_accept_transfer: Pseudo-phys frame slot being filled by new
+ *                           frame, or zero if none.
+ *  3. Write memory barrier (WMB).
+ *  4. Write ent->flags, inc. valid type.
+ */
+static void gnttab_update_entry_v1(grant_ref_t ref, domid_t domid,
+				   unsigned long frame, unsigned flags)
 {
-	/*
-	 * Introducing a valid entry into the grant table:
-	 *  1. Write ent->domid.
-	 *  2. Write ent->frame:
-	 *      GTF_permit_access:   Frame to which access is permitted.
-	 *      GTF_accept_transfer: Pseudo-phys frame slot being filled by new
-	 *                           frame, or zero if none.
-	 *  3. Write memory barrier (WMB).
-	 *  4. Write ent->flags, inc. valid type.
-	 */
-	shared[ref].frame = frame;
-	shared[ref].domid = domid;
+	gnttab_shared.v1[ref].domid = domid;
+	gnttab_shared.v1[ref].frame = frame;
 	wmb();
-	shared[ref].flags = flags;
+	gnttab_shared.v1[ref].flags = flags;
 }
 
 /*
@@ -167,7 +223,7 @@ static void update_grant_entry(grant_ref_t ref, domid_t domid,
 void gnttab_grant_foreign_access_ref(grant_ref_t ref, domid_t domid,
 				     unsigned long frame, int readonly)
 {
-	update_grant_entry(ref, domid, frame,
+	gnttab_interface->update_entry(ref, domid, frame,
 			   GTF_permit_access | (readonly ? GTF_readonly : 0));
 }
 EXPORT_SYMBOL_GPL(gnttab_grant_foreign_access_ref);
@@ -187,31 +243,37 @@ int gnttab_grant_foreign_access(domid_t domid, unsigned long frame,
 }
 EXPORT_SYMBOL_GPL(gnttab_grant_foreign_access);
 
-int gnttab_query_foreign_access(grant_ref_t ref)
+static int gnttab_query_foreign_access_v1(grant_ref_t ref)
 {
-	u16 nflags;
-
-	nflags = shared[ref].flags;
+	return gnttab_shared.v1[ref].flags & (GTF_reading|GTF_writing);
+}
 
-	return nflags & (GTF_reading|GTF_writing);
+int gnttab_query_foreign_access(grant_ref_t ref)
+{
+	return gnttab_interface->query_foreign_access(ref);
 }
 EXPORT_SYMBOL_GPL(gnttab_query_foreign_access);
 
-int gnttab_end_foreign_access_ref(grant_ref_t ref, int readonly)
+static int gnttab_end_foreign_access_ref_v1(grant_ref_t ref, int readonly)
 {
 	u16 flags, nflags;
 
-	nflags = shared[ref].flags;
+	nflags = gnttab_shared.v1[ref].flags;
 	do {
 		flags = nflags;
 		if (flags & (GTF_reading|GTF_writing)) {
 			printk(KERN_ALERT "WARNING: g.e. still in use!\n");
 			return 0;
 		}
-	} while ((nflags = sync_cmpxchg(&shared[ref].flags, flags, 0)) != flags);
+	} while ((nflags = sync_cmpxchg(&gnttab_shared.v1[ref].flags, flags, 0)) != flags);
 
 	return 1;
 }
+
+int gnttab_end_foreign_access_ref(grant_ref_t ref, int readonly)
+{
+	return gnttab_interface->end_foreign_access_ref(ref, readonly);
+}
 EXPORT_SYMBOL_GPL(gnttab_end_foreign_access_ref);
 
 void gnttab_end_foreign_access(grant_ref_t ref, int readonly,
@@ -246,11 +308,11 @@ EXPORT_SYMBOL_GPL(gnttab_grant_foreign_transfer);
 void gnttab_grant_foreign_transfer_ref(grant_ref_t ref, domid_t domid,
 				       unsigned long pfn)
 {
-	update_grant_entry(ref, domid, pfn, GTF_accept_transfer);
+	gnttab_interface->update_entry(ref, domid, pfn, GTF_accept_transfer);
 }
 EXPORT_SYMBOL_GPL(gnttab_grant_foreign_transfer_ref);
 
-unsigned long gnttab_end_foreign_transfer_ref(grant_ref_t ref)
+static unsigned long gnttab_end_foreign_transfer_ref_v1(grant_ref_t ref)
 {
 	unsigned long frame;
 	u16           flags;
@@ -259,24 +321,29 @@ unsigned long gnttab_end_foreign_transfer_ref(grant_ref_t ref)
 	 * If a transfer is not even yet started, try to reclaim the grant
 	 * reference and return failure (== 0).
 	 */
-	while (!((flags = shared[ref].flags) & GTF_transfer_committed)) {
-		if (sync_cmpxchg(&shared[ref].flags, flags, 0) == flags)
+	while (!((flags = gnttab_shared.v1[ref].flags) & GTF_transfer_committed)) {
+		if (sync_cmpxchg(&gnttab_shared.v1[ref].flags, flags, 0) == flags)
 			return 0;
 		cpu_relax();
 	}
 
 	/* If a transfer is in progress then wait until it is completed. */
 	while (!(flags & GTF_transfer_completed)) {
-		flags = shared[ref].flags;
+		flags = gnttab_shared.v1[ref].flags;
 		cpu_relax();
 	}
 
 	rmb();	/* Read the frame number /after/ reading completion status. */
-	frame = shared[ref].frame;
+	frame = gnttab_shared.v1[ref].frame;
 	BUG_ON(frame == 0);
 
 	return frame;
 }
+
+unsigned long gnttab_end_foreign_transfer_ref(grant_ref_t ref)
+{
+	return gnttab_interface->end_foreign_transfer_ref(ref);
+}
 EXPORT_SYMBOL_GPL(gnttab_end_foreign_transfer_ref);
 
 unsigned long gnttab_end_foreign_transfer(grant_ref_t ref)
@@ -520,6 +587,23 @@ int gnttab_unmap_refs(struct gnttab_unmap_grant_ref *unmap_ops,
 }
 EXPORT_SYMBOL_GPL(gnttab_unmap_refs);
 
+static int gnttab_map_frames_v1(unsigned long *frames, unsigned int nr_gframes)
+{
+	int rc;
+
+	rc = arch_gnttab_map_shared(frames, nr_gframes,
+				    gnttab_max_grant_frames(),
+				    &gnttab_shared.addr);
+	BUG_ON(rc);
+
+	return 0;
+}
+
+static void gnttab_unmap_frames_v1(void)
+{
+	arch_gnttab_unmap_shared(gnttab_shared.addr, nr_grant_frames);
+}
+
 static int gnttab_map(unsigned int start_idx, unsigned int end_idx)
 {
 	struct gnttab_setup_table setup;
@@ -567,19 +651,35 @@ static int gnttab_map(unsigned int start_idx, unsigned int end_idx)
 
 	BUG_ON(rc || setup.status);
 
-	rc = arch_gnttab_map_shared(frames, nr_gframes, gnttab_max_grant_frames(),
-				    &shared);
-	BUG_ON(rc);
+	rc = gnttab_interface->map_frames(frames, nr_gframes);
 
 	kfree(frames);
 
-	return 0;
+	return rc;
+}
+
+static struct gnttab_ops gnttab_v1_ops = {
+	.map_frames			= gnttab_map_frames_v1,
+	.unmap_frames			= gnttab_unmap_frames_v1,
+	.update_entry			= gnttab_update_entry_v1,
+	.end_foreign_access_ref		= gnttab_end_foreign_access_ref_v1,
+	.end_foreign_transfer_ref	= gnttab_end_foreign_transfer_ref_v1,
+	.query_foreign_access		= gnttab_query_foreign_access_v1,
+};
+
+static void gnttab_request_version(void)
+{
+	grant_table_version = 1;
+	gnttab_interface = &gnttab_v1_ops;
+	printk(KERN_INFO "Grant tables using version %d layout.\n",
+		grant_table_version);
 }
 
 int gnttab_resume(void)
 {
 	unsigned int max_nr_gframes;
 
+	gnttab_request_version();
 	max_nr_gframes = gnttab_max_grant_frames();
 	if (max_nr_gframes < nr_grant_frames)
 		return -ENOSYS;
@@ -587,9 +687,10 @@ int gnttab_resume(void)
 	if (xen_pv_domain())
 		return gnttab_map(0, nr_grant_frames - 1);
 
-	if (!shared) {
-		shared = ioremap(xen_hvm_resume_frames, PAGE_SIZE * max_nr_gframes);
-		if (shared == NULL) {
+	if (gnttab_shared.addr == NULL) {
+		gnttab_shared.addr = ioremap(xen_hvm_resume_frames,
+						PAGE_SIZE * max_nr_gframes);
+		if (gnttab_shared.addr == NULL) {
 			printk(KERN_WARNING
 					"Failed to ioremap gnttab share frames!");
 			return -ENOMEM;
@@ -603,7 +704,7 @@ int gnttab_resume(void)
 
 int gnttab_suspend(void)
 {
-	arch_gnttab_unmap_shared(shared, nr_grant_frames);
+	gnttab_interface->unmap_frames();
 	return 0;
 }
 
diff --git a/include/xen/grant_table.h b/include/xen/grant_table.h
index 11e2dfc..c7a40f8 100644
--- a/include/xen/grant_table.h
+++ b/include/xen/grant_table.h
@@ -145,8 +145,8 @@ gnttab_set_unmap_op(struct gnttab_unmap_grant_ref *unmap, phys_addr_t addr,
 
 int arch_gnttab_map_shared(unsigned long *frames, unsigned long nr_gframes,
 			   unsigned long max_nr_gframes,
-			   struct grant_entry **__shared);
-void arch_gnttab_unmap_shared(struct grant_entry *shared,
+			   void **__shared);
+void arch_gnttab_unmap_shared(void *shared,
 			      unsigned long nr_gframes);
 
 extern unsigned long xen_hvm_resume_frames;
diff --git a/include/xen/interface/grant_table.h b/include/xen/interface/grant_table.h
index 39e5717..a17d844 100644
--- a/include/xen/interface/grant_table.h
+++ b/include/xen/interface/grant_table.h
@@ -85,12 +85,22 @@
  */
 
 /*
+ * Reference to a grant entry in a specified domain's grant table.
+ */
+typedef uint32_t grant_ref_t;
+
+/*
  * A grant table comprises a packed array of grant entries in one or more
  * page frames shared between Xen and a guest.
  * [XEN]: This field is written by Xen and read by the sharing guest.
  * [GST]: This field is written by the guest and read by Xen.
  */
-struct grant_entry {
+
+/*
+ * Version 1 of the grant table entry structure is maintained purely
+ * for backwards compatibility.  New guests should use version 2.
+ */
+struct grant_entry_v1 {
     /* GTF_xxx: various type and flag information.  [XEN,GST] */
     uint16_t flags;
     /* The domain being granted foreign privileges. [GST] */
@@ -108,10 +118,13 @@ struct grant_entry {
  *  GTF_permit_access: Allow @domid to map/access @frame.
  *  GTF_accept_transfer: Allow @domid to transfer ownership of one page frame
  *                       to this guest. Xen writes the page number to @frame.
+ *  GTF_transitive: Allow @domid to transitively access a subrange of
+ *                  @trans_grant in @trans_domid.  No mappings are allowed.
  */
 #define GTF_invalid         (0U<<0)
 #define GTF_permit_access   (1U<<0)
 #define GTF_accept_transfer (2U<<0)
+#define GTF_transitive      (3U<<0)
 #define GTF_type_mask       (3U<<0)
 
 /*
@@ -119,6 +132,9 @@ struct grant_entry {
  *  GTF_readonly: Restrict @domid to read-only mappings and accesses. [GST]
  *  GTF_reading: Grant entry is currently mapped for reading by @domid. [XEN]
  *  GTF_writing: Grant entry is currently mapped for writing by @domid. [XEN]
+ *  GTF_sub_page: Grant access to only a subrange of the page.  @domid
+ *                will only be allowed to copy from the grant, and not
+ *                map it. [GST]
  */
 #define _GTF_readonly       (2)
 #define GTF_readonly        (1U<<_GTF_readonly)
@@ -126,6 +142,8 @@ struct grant_entry {
 #define GTF_reading         (1U<<_GTF_reading)
 #define _GTF_writing        (4)
 #define GTF_writing         (1U<<_GTF_writing)
+#define _GTF_sub_page       (8)
+#define GTF_sub_page        (1U<<_GTF_sub_page)
 
 /*
  * Subflags for GTF_accept_transfer:
@@ -142,15 +160,81 @@ struct grant_entry {
 #define _GTF_transfer_completed (3)
 #define GTF_transfer_completed  (1U<<_GTF_transfer_completed)
 
+/*
+ * Version 2 grant table entries.  These fulfil the same role as
+ * version 1 entries, but can represent more complicated operations.
+ * Any given domain will have either a version 1 or a version 2 table,
+ * and every entry in the table will be the same version.
+ *
+ * The interface by which domains use grant references does not depend
+ * on the grant table version in use by the other domain.
+ */
 
-/***********************************
- * GRANT TABLE QUERIES AND USES
+/*
+ * Version 1 and version 2 grant entries share a common prefix.  The
+ * fields of the prefix are documented as part of struct
+ * grant_entry_v1.
  */
+struct grant_entry_header {
+    uint16_t flags;
+    domid_t  domid;
+};
 
 /*
- * Reference to a grant entry in a specified domain's grant table.
+ * Version 2 of the grant entry structure, here is an union because three
+ * different types are suppotted: full_page, sub_page and transitive.
+ */
+union grant_entry_v2 {
+    struct grant_entry_header hdr;
+
+    /*
+     * This member is used for V1-style full page grants, where either:
+     *
+     * -- hdr.type is GTF_accept_transfer, or
+     * -- hdr.type is GTF_permit_access and GTF_sub_page is not set.
+     *
+     * In that case, the frame field has the same semantics as the
+     * field of the same name in the V1 entry structure.
+     */
+    struct {
+	struct grant_entry_header hdr;
+	uint32_t pad0;
+	uint64_t frame;
+    } full_page;
+
+    /*
+     * If the grant type is GTF_grant_access and GTF_sub_page is set,
+     * @domid is allowed to access bytes [@page_off,@page_off+@length)
+     * in frame @frame.
+     */
+    struct {
+	struct grant_entry_header hdr;
+	uint16_t page_off;
+	uint16_t length;
+	uint64_t frame;
+    } sub_page;
+
+    /*
+     * If the grant is GTF_transitive, @domid is allowed to use the
+     * grant @gref in domain @trans_domid, as if it was the local
+     * domain.  Obviously, the transitive access must be compatible
+     * with the original grant.
+     */
+    struct {
+	struct grant_entry_header hdr;
+	domid_t trans_domid;
+	uint16_t pad0;
+	grant_ref_t gref;
+    } transitive;
+
+    uint32_t __spacer[4]; /* Pad to a power of two */
+};
+
+typedef uint16_t grant_status_t;
+
+/***********************************
+ * GRANT TABLE QUERIES AND USES
  */
-typedef uint32_t grant_ref_t;
 
 /*
  * Handle to track a mapping created via a grant reference.
@@ -322,6 +406,79 @@ struct gnttab_query_size {
 DEFINE_GUEST_HANDLE_STRUCT(gnttab_query_size);
 
 /*
+ * GNTTABOP_unmap_and_replace: Destroy one or more grant-reference mappings
+ * tracked by <handle> but atomically replace the page table entry with one
+ * pointing to the machine address under <new_addr>.  <new_addr> will be
+ * redirected to the null entry.
+ * NOTES:
+ *  1. The call may fail in an undefined manner if either mapping is not
+ *     tracked by <handle>.
+ *  2. After executing a batch of unmaps, it is guaranteed that no stale
+ *     mappings will remain in the device or host TLBs.
+ */
+#define GNTTABOP_unmap_and_replace    7
+struct gnttab_unmap_and_replace {
+    /* IN parameters. */
+    uint64_t host_addr;
+    uint64_t new_addr;
+    grant_handle_t handle;
+    /* OUT parameters. */
+    int16_t  status;              /* GNTST_* */
+};
+DEFINE_GUEST_HANDLE_STRUCT(gnttab_unmap_and_replace);
+
+/*
+ * GNTTABOP_set_version: Request a particular version of the grant
+ * table shared table structure.  This operation can only be performed
+ * once in any given domain.  It must be performed before any grants
+ * are activated; otherwise, the domain will be stuck with version 1.
+ * The only defined versions are 1 and 2.
+ */
+#define GNTTABOP_set_version          8
+struct gnttab_set_version {
+    /* IN parameters */
+    uint32_t version;
+};
+DEFINE_GUEST_HANDLE_STRUCT(gnttab_set_version);
+
+/*
+ * GNTTABOP_get_status_frames: Get the list of frames used to store grant
+ * status for <dom>. In grant format version 2, the status is separated
+ * from the other shared grant fields to allow more efficient synchronization
+ * using barriers instead of atomic cmpexch operations.
+ * <nr_frames> specify the size of vector <frame_list>.
+ * The frame addresses are returned in the <frame_list>.
+ * Only <nr_frames> addresses are returned, even if the table is larger.
+ * NOTES:
+ *  1. <dom> may be specified as DOMID_SELF.
+ *  2. Only a sufficiently-privileged domain may specify <dom> != DOMID_SELF.
+ */
+#define GNTTABOP_get_status_frames     9
+struct gnttab_get_status_frames {
+    /* IN parameters. */
+    uint32_t nr_frames;
+    domid_t  dom;
+    /* OUT parameters. */
+    int16_t  status;              /* GNTST_* */
+    GUEST_HANDLE(uint64_t) frame_list;
+};
+DEFINE_GUEST_HANDLE_STRUCT(gnttab_get_status_frames);
+
+/*
+ * GNTTABOP_get_version: Get the grant table version which is in
+ * effect for domain <dom>.
+ */
+#define GNTTABOP_get_version          10
+struct gnttab_get_version {
+    /* IN parameters */
+    domid_t dom;
+    uint16_t pad;
+    /* OUT parameters */
+    uint32_t version;
+};
+DEFINE_GUEST_HANDLE_STRUCT(gnttab_get_version);
+
+/*
  * Bitfield values for update_pin_status.flags.
  */
  /* Map the grant entry for access by I/O devices. */
diff --git a/include/xen/interface/xen.h b/include/xen/interface/xen.h
index 6a6e914..a890804 100644
--- a/include/xen/interface/xen.h
+++ b/include/xen/interface/xen.h
@@ -523,6 +523,8 @@ struct tmem_op {
 	} u;
 };
 
+DEFINE_GUEST_HANDLE(u64);
+
 #else /* __ASSEMBLY__ */
 
 /* In assembly code we cannot use C numeric constant suffixes. */
-- 
1.7.6.4

[PATCH 2/4] xen/granttable: Refactor some code

[annie.li]

From: Annie Li <annie.li@oracle.com>

Signed-off-by: Annie Li <annie.li@oracle.com>
---
 drivers/xen/grant-table.c |   15 ++++++++++-----
 1 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c
index 18355a5..0518d04 100644
--- a/drivers/xen/grant-table.c
+++ b/drivers/xen/grant-table.c
@@ -257,15 +257,17 @@ EXPORT_SYMBOL_GPL(gnttab_query_foreign_access);
 static int gnttab_end_foreign_access_ref_v1(grant_ref_t ref, int readonly)
 {
 	u16 flags, nflags;
+	u16 *pflags;
 
-	nflags = gnttab_shared.v1[ref].flags;
+	pflags = &gnttab_shared.v1[ref].flags;
+	nflags = *pflags;
 	do {
 		flags = nflags;
 		if (flags & (GTF_reading|GTF_writing)) {
 			printk(KERN_ALERT "WARNING: g.e. still in use!\n");
 			return 0;
 		}
-	} while ((nflags = sync_cmpxchg(&gnttab_shared.v1[ref].flags, flags, 0)) != flags);
+	} while ((nflags = sync_cmpxchg(pflags, flags, 0)) != flags);
 
 	return 1;
 }
@@ -316,20 +318,23 @@ static unsigned long gnttab_end_foreign_transfer_ref_v1(grant_ref_t ref)
 {
 	unsigned long frame;
 	u16           flags;
+	u16          *pflags;
+
+	pflags = &gnttab_shared.v1[ref].flags;
 
 	/*
 	 * If a transfer is not even yet started, try to reclaim the grant
 	 * reference and return failure (== 0).
 	 */
-	while (!((flags = gnttab_shared.v1[ref].flags) & GTF_transfer_committed)) {
-		if (sync_cmpxchg(&gnttab_shared.v1[ref].flags, flags, 0) == flags)
+	while (!((flags = *pflags) & GTF_transfer_committed)) {
+		if (sync_cmpxchg(pflags, flags, 0) == flags)
 			return 0;
 		cpu_relax();
 	}
 
 	/* If a transfer is in progress then wait until it is completed. */
 	while (!(flags & GTF_transfer_completed)) {
-		flags = gnttab_shared.v1[ref].flags;
+		flags = *pflags;
 		cpu_relax();
 	}
 
-- 
1.7.6.4

[PATCH 3/4] xen/granttable: Grant tables V2 implementation

[annie.li]

From: Annie Li <annie.li@oracle.com>

Receiver-side copying of packets is based on this implementation, it gives
better performance and better CPU accounting. It totally supports three types:
full-page, sub-page and transitive grants.

However this patch does not cover sub-page and transitive grants, it mainly
focus on Full-page part and implements grant table V2 interfaces corresponding
to what already exists in grant table V1, such as: grant table V2
initialization, mapping, releasing and exported interfaces.

Each guest can only supports one type of grant table type, every entry in grant
table should be the same version. It is necessary to set V1 or V2 version before
initializing the grant table.

Grant table exported interfaces of V2 are same with those of V1, Xen is
responsible to judge what grant table version guests are using in every grant
operation.

V2 fulfills the same role of V1, and it is totally backwards compitable with V1.
If dom0 support grant table V2, the guests runing on it can run with either V1
or V2.

Signed-off-by: Annie Li <annie.li@oracle.com>
---
 arch/x86/xen/grant-table.c |   37 +++++++++-
 drivers/xen/grant-table.c  |  171 ++++++++++++++++++++++++++++++++++++++++++-
 include/xen/grant_table.h  |    6 +-
 3 files changed, 207 insertions(+), 7 deletions(-)

diff --git a/arch/x86/xen/grant-table.c b/arch/x86/xen/grant-table.c
index c6ab2e7..510fca0 100644
--- a/arch/x86/xen/grant-table.c
+++ b/arch/x86/xen/grant-table.c
@@ -54,6 +54,20 @@ static int map_pte_fn(pte_t *pte, struct page *pmd_page,
 	return 0;
 }
 
+/*
+ * This function is used to map shared frames to store grant status. It is
+ * different from map_pte_fn above, the frames type here is uint64_t.
+ */
+static int map_pte_fn_status(pte_t *pte, struct page *pmd_page,
+			     unsigned long addr, void *data)
+{
+	uint64_t **frames = (uint64_t **)data;
+
+	set_pte_at(&init_mm, addr, pte, mfn_pte((*frames)[0], PAGE_KERNEL));
+	(*frames)++;
+	return 0;
+}
+
 static int unmap_pte_fn(pte_t *pte, struct page *pmd_page,
 			unsigned long addr, void *data)
 {
@@ -83,7 +97,28 @@ int arch_gnttab_map_shared(unsigned long *frames, unsigned long nr_gframes,
 	return rc;
 }
 
-void arch_gnttab_unmap_shared(void *shared, unsigned long nr_gframes)
+int arch_gnttab_map_status(uint64_t *frames, unsigned long nr_gframes,
+			   unsigned long max_nr_gframes,
+			   grant_status_t **__shared)
+{
+	int rc;
+	grant_status_t *shared = *__shared;
+
+	if (shared == NULL) {
+		struct vm_struct *area =
+			alloc_vm_area(PAGE_SIZE * max_nr_gframes);
+		BUG_ON(area == NULL);
+		shared = area->addr;
+		*__shared = shared;
+	}
+
+	rc = apply_to_page_range(&init_mm, (unsigned long)shared,
+				 PAGE_SIZE * nr_gframes,
+				 map_pte_fn_status, &frames);
+	return rc;
+}
+
+void arch_gnttab_unmap(void *shared, unsigned long nr_gframes)
 {
 	apply_to_page_range(&init_mm, (unsigned long)shared,
 			    PAGE_SIZE * nr_gframes, unmap_pte_fn, NULL);
diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c
index 0518d04..77a060c 100644
--- a/drivers/xen/grant-table.c
+++ b/drivers/xen/grant-table.c
@@ -44,6 +44,7 @@
 #include <xen/page.h>
 #include <xen/grant_table.h>
 #include <xen/interface/memory.h>
+#include <xen/hvc-console.h>
 #include <asm/xen/hypercall.h>
 
 #include <asm/pgtable.h>
@@ -53,7 +54,10 @@
 /* External tools reserve first few grant table entries. */
 #define NR_RESERVED_ENTRIES 8
 #define GNTTAB_LIST_END 0xffffffff
-#define GREFS_PER_GRANT_FRAME (PAGE_SIZE / sizeof(struct grant_entry_v1))
+#define GREFS_PER_GRANT_FRAME \
+(grant_table_version == 1 ?                      \
+ (PAGE_SIZE / sizeof(struct grant_entry_v1)) :   \
+ (PAGE_SIZE / sizeof(union grant_entry_v2)))
 
 static grant_ref_t **gnttab_list;
 static unsigned int nr_grant_frames;
@@ -66,6 +70,7 @@ EXPORT_SYMBOL_GPL(xen_hvm_resume_frames);
 
 static union {
 	struct grant_entry_v1 *v1;
+	union grant_entry_v2 *v2;
 	void *addr;
 } gnttab_shared;
 
@@ -120,6 +125,9 @@ struct gnttab_ops {
 
 static struct gnttab_ops *gnttab_interface;
 
+/*This reflects status of grant entries, so act as a global value*/
+static grant_status_t *grstatus;
+
 static int grant_table_version;
 
 static struct gnttab_free_callback *gnttab_free_callback_list;
@@ -127,6 +135,7 @@ static struct gnttab_free_callback *gnttab_free_callback_list;
 static int gnttab_expand(unsigned int req_entries);
 
 #define RPP (PAGE_SIZE / sizeof(grant_ref_t))
+#define SPP (PAGE_SIZE / sizeof(grant_status_t))
 
 static inline grant_ref_t *__gnttab_entry(grant_ref_t entry)
 {
@@ -199,6 +208,7 @@ static void put_free_entry(grant_ref_t ref)
 }
 
 /*
+ * Following applies to gnttab_update_entry_v1 and gnttab_update_entry_v2.
  * Introducing a valid entry into the grant table:
  *  1. Write ent->domid.
  *  2. Write ent->frame:
@@ -217,6 +227,15 @@ static void gnttab_update_entry_v1(grant_ref_t ref, domid_t domid,
 	gnttab_shared.v1[ref].flags = flags;
 }
 
+static void gnttab_update_entry_v2(grant_ref_t ref, domid_t domid,
+				   unsigned long frame, unsigned flags)
+{
+	gnttab_shared.v2[ref].hdr.domid = domid;
+	gnttab_shared.v2[ref].full_page.frame = frame;
+	wmb();
+	gnttab_shared.v2[ref].hdr.flags = GTF_permit_access | flags;
+}
+
 /*
  * Public grant-issuing interface functions
  */
@@ -248,6 +267,11 @@ static int gnttab_query_foreign_access_v1(grant_ref_t ref)
 	return gnttab_shared.v1[ref].flags & (GTF_reading|GTF_writing);
 }
 
+static int gnttab_query_foreign_access_v2(grant_ref_t ref)
+{
+	return grstatus[ref] & (GTF_reading|GTF_writing);
+}
+
 int gnttab_query_foreign_access(grant_ref_t ref)
 {
 	return gnttab_interface->query_foreign_access(ref);
@@ -272,6 +296,29 @@ static int gnttab_end_foreign_access_ref_v1(grant_ref_t ref, int readonly)
 	return 1;
 }
 
+static int gnttab_end_foreign_access_ref_v2(grant_ref_t ref, int readonly)
+{
+	gnttab_shared.v2[ref].hdr.flags = 0;
+	mb();
+	if (grstatus[ref] & (GTF_reading|GTF_writing)) {
+		return 0;
+	} else {
+		/* The read of grstatus needs to have acquire
+		semantics.  On x86, reads already have
+		that, and we just need to protect against
+		compiler reorderings.  On other
+		architectures we may need a full
+		barrier. */
+#ifdef CONFIG_X86
+		barrier();
+#else
+		mb();
+#endif
+		}
+
+	return 1;
+}
+
 int gnttab_end_foreign_access_ref(grant_ref_t ref, int readonly)
 {
 	return gnttab_interface->end_foreign_access_ref(ref, readonly);
@@ -345,6 +392,37 @@ static unsigned long gnttab_end_foreign_transfer_ref_v1(grant_ref_t ref)
 	return frame;
 }
 
+static unsigned long gnttab_end_foreign_transfer_ref_v2(grant_ref_t ref)
+{
+	unsigned long frame;
+	u16           flags;
+	u16          *pflags;
+
+	pflags = &gnttab_shared.v2[ref].hdr.flags;
+
+	/*
+	 * If a transfer is not even yet started, try to reclaim the grant
+	 * reference and return failure (== 0).
+	 */
+	while (!((flags = *pflags) & GTF_transfer_committed)) {
+		if (sync_cmpxchg(pflags, flags, 0) == flags)
+			return 0;
+		cpu_relax();
+	}
+
+	/* If a transfer is in progress then wait until it is completed. */
+	while (!(flags & GTF_transfer_completed)) {
+		flags = *pflags;
+		cpu_relax();
+	}
+
+	rmb();  /* Read the frame number /after/ reading completion status. */
+	frame = gnttab_shared.v2[ref].full_page.frame;
+	BUG_ON(frame == 0);
+
+	return frame;
+}
+
 unsigned long gnttab_end_foreign_transfer_ref(grant_ref_t ref)
 {
 	return gnttab_interface->end_foreign_transfer_ref(ref);
@@ -592,6 +670,11 @@ int gnttab_unmap_refs(struct gnttab_unmap_grant_ref *unmap_ops,
 }
 EXPORT_SYMBOL_GPL(gnttab_unmap_refs);
 
+static unsigned nr_status_frames(unsigned nr_grant_frames)
+{
+	return (nr_grant_frames * GREFS_PER_GRANT_FRAME + SPP - 1) / SPP;
+}
+
 static int gnttab_map_frames_v1(unsigned long *frames, unsigned int nr_gframes)
 {
 	int rc;
@@ -606,7 +689,56 @@ static int gnttab_map_frames_v1(unsigned long *frames, unsigned int nr_gframes)
 
 static void gnttab_unmap_frames_v1(void)
 {
-	arch_gnttab_unmap_shared(gnttab_shared.addr, nr_grant_frames);
+	arch_gnttab_unmap(gnttab_shared.addr, nr_grant_frames);
+}
+
+static int gnttab_map_frames_v2(unsigned long *frames, unsigned int nr_gframes)
+{
+	uint64_t *sframes;
+	unsigned int nr_sframes;
+	struct gnttab_get_status_frames getframes;
+	int rc;
+
+	nr_sframes = nr_status_frames(nr_gframes);
+
+	/* No need for kzalloc as it is initialized in following hypercall
+	 * GNTTABOP_get_status_frames.
+	 */
+	sframes = kmalloc(nr_sframes  * sizeof(uint64_t), GFP_ATOMIC);
+	if (!sframes)
+		return -ENOMEM;
+
+	getframes.dom        = DOMID_SELF;
+	getframes.nr_frames  = nr_sframes;
+	set_xen_guest_handle(getframes.frame_list, sframes);
+
+	rc = HYPERVISOR_grant_table_op(GNTTABOP_get_status_frames,
+				       &getframes, 1);
+	if (rc == -ENOSYS) {
+		kfree(sframes);
+		return -ENOSYS;
+	}
+
+	BUG_ON(rc || getframes.status);
+
+	rc = arch_gnttab_map_status(sframes, nr_sframes,
+				    nr_status_frames(gnttab_max_grant_frames()),
+				    &grstatus);
+	BUG_ON(rc);
+	kfree(sframes);
+
+	rc = arch_gnttab_map_shared(frames, nr_gframes,
+				    gnttab_max_grant_frames(),
+				    &gnttab_shared.addr);
+	BUG_ON(rc);
+
+	return 0;
+}
+
+static void gnttab_unmap_frames_v2(void)
+{
+	arch_gnttab_unmap(gnttab_shared.addr, nr_grant_frames);
+	arch_gnttab_unmap(grstatus, nr_status_frames(nr_grant_frames));
 }
 
 static int gnttab_map(unsigned int start_idx, unsigned int end_idx)
@@ -640,6 +772,9 @@ static int gnttab_map(unsigned int start_idx, unsigned int end_idx)
 		return rc;
 	}
 
+	/* No need for kzalloc as it is initialized in following hypercall
+	 * GNTTABOP_setup_table.
+	 */
 	frames = kmalloc(nr_gframes * sizeof(unsigned long), GFP_ATOMIC);
 	if (!frames)
 		return -ENOMEM;
@@ -672,10 +807,38 @@ static struct gnttab_ops gnttab_v1_ops = {
 	.query_foreign_access		= gnttab_query_foreign_access_v1,
 };
 
+static struct gnttab_ops gnttab_v2_ops = {
+	.map_frames			= gnttab_map_frames_v2,
+	.unmap_frames			= gnttab_unmap_frames_v2,
+	.update_entry			= gnttab_update_entry_v2,
+	.end_foreign_access_ref		= gnttab_end_foreign_access_ref_v2,
+	.end_foreign_transfer_ref	= gnttab_end_foreign_transfer_ref_v2,
+	.query_foreign_access		= gnttab_query_foreign_access_v2,
+};
+
 static void gnttab_request_version(void)
 {
-	grant_table_version = 1;
-	gnttab_interface = &gnttab_v1_ops;
+	int rc;
+	struct gnttab_set_version gsv;
+
+	gsv.version = 2;
+	rc = HYPERVISOR_grant_table_op(GNTTABOP_set_version, &gsv, 1);
+	if (rc == 0) {
+		grant_table_version = 2;
+		gnttab_interface = &gnttab_v2_ops;
+	} else if (grant_table_version == 2) {
+		/*
+		 * If we've already used version 2 features,
+		 * but then suddenly discover that they're not
+		 * available (e.g. migrating to an older
+		 * version of Xen), almost unbounded badness
+		 * can happen.
+		 */
+		panic("we need grant tables version 2, but only version 1 is available");
+	} else {
+		grant_table_version = 1;
+		gnttab_interface = &gnttab_v1_ops;
+	}
 	printk(KERN_INFO "Grant tables using version %d layout.\n",
 		grant_table_version);
 }
diff --git a/include/xen/grant_table.h b/include/xen/grant_table.h
index c7a40f8..5494c40 100644
--- a/include/xen/grant_table.h
+++ b/include/xen/grant_table.h
@@ -146,8 +146,10 @@ gnttab_set_unmap_op(struct gnttab_unmap_grant_ref *unmap, phys_addr_t addr,
 int arch_gnttab_map_shared(unsigned long *frames, unsigned long nr_gframes,
 			   unsigned long max_nr_gframes,
 			   void **__shared);
-void arch_gnttab_unmap_shared(void *shared,
-			      unsigned long nr_gframes);
+int arch_gnttab_map_status(uint64_t *frames, unsigned long nr_gframes,
+			   unsigned long max_nr_gframes,
+			   grant_status_t **__shared);
+void arch_gnttab_unmap(void *shared, unsigned long nr_gframes);
 
 extern unsigned long xen_hvm_resume_frames;
 unsigned int gnttab_max_grant_frames(void);
-- 
1.7.6.4

Re: [PATCH 3/4] xen/granttable: Grant tables V2 implementation

[Ian Campbell]

On Tue, 2011-11-22 at 01:59 +0000, annie.li@oracle.com wrote:
[...]
> +#ifdef CONFIG_X86
> +		barrier();
> +#else
> +		mb();
> +#endif
> +		}

Niggle: Indentation of this final } is off.

Otherwise all 4 patches:
Acked-by: Ian Campbell <ian.campbell@citrix.com>

[...]

Re: [PATCH 3/4] xen/granttable: Grant tables V2 implementation

[ANNIE LI]

On 2011-11-22 17:29, Ian Campbell wrote:
> On Tue, 2011-11-22 at 01:59 +0000, annie.li@oracle.com wrote:
> [...]
>> +#ifdef CONFIG_X86
>> +		barrier();
>> +#else
>> +		mb();
>> +#endif
>> +		}
> Niggle: Indentation of this final } is off.
Sorry, I did not notice that.
> Otherwise all 4 patches:
> Acked-by: Ian Campbell<ian.campbell@citrix.com>
>
> [...]
Thanks, I will resend those patches after above indentation issue is fixed.

Thanks.

Thanks
Annie

Re: [PATCH 3/4] xen/granttable: Grant tables V2 implementation

[Konrad Rzeszutek Wilk]

On Tue, Nov 22, 2011 at 06:02:12PM +0800, ANNIE LI wrote:
> 
> 
> On 2011-11-22 17:29, Ian Campbell wrote:
> >On Tue, 2011-11-22 at 01:59 +0000, annie.li@oracle.com wrote:
> >[...]
> >>+#ifdef CONFIG_X86
> >>+		barrier();
> >>+#else
> >>+		mb();
> >>+#endif
> >>+		}
> >Niggle: Indentation of this final } is off.
> Sorry, I did not notice that.
> >Otherwise all 4 patches:
> >Acked-by: Ian Campbell<ian.campbell@citrix.com>
> >
> >[...]
> Thanks, I will resend those patches after above indentation issue is fixed.

That is OK. I can modify them.
> 
> Thanks.
> 
> Thanks
> Annie

[PATCH 4/4] xen/granttable: Keep code format clean

[annie.li]

From: Annie Li <annie.li@oracle.com>

Signed-off-by: Annie Li <annie.li@oracle.com>
---
 drivers/xen/grant-table.c |    7 +++----
 include/xen/grant_table.h |    2 +-
 2 files changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c
index 77a060c..b76f419 100644
--- a/drivers/xen/grant-table.c
+++ b/drivers/xen/grant-table.c
@@ -50,7 +50,6 @@
 #include <asm/pgtable.h>
 #include <asm/sync_bitops.h>
 
-
 /* External tools reserve first few grant table entries. */
 #define NR_RESERVED_ENTRIES 8
 #define GNTTAB_LIST_END 0xffffffff
@@ -598,8 +597,8 @@ unsigned int gnttab_max_grant_frames(void)
 EXPORT_SYMBOL_GPL(gnttab_max_grant_frames);
 
 int gnttab_map_refs(struct gnttab_map_grant_ref *map_ops,
-			struct gnttab_map_grant_ref *kmap_ops,
-			struct page **pages, unsigned int count)
+		    struct gnttab_map_grant_ref *kmap_ops,
+		    struct page **pages, unsigned int count)
 {
 	int i, ret;
 	pte_t *pte;
@@ -649,7 +648,7 @@ int gnttab_map_refs(struct gnttab_map_grant_ref *map_ops,
 EXPORT_SYMBOL_GPL(gnttab_map_refs);
 
 int gnttab_unmap_refs(struct gnttab_unmap_grant_ref *unmap_ops,
-		struct page **pages, unsigned int count)
+		      struct page **pages, unsigned int count)
 {
 	int i, ret;
 
diff --git a/include/xen/grant_table.h b/include/xen/grant_table.h
index 5494c40..fea4954 100644
--- a/include/xen/grant_table.h
+++ b/include/xen/grant_table.h
@@ -157,7 +157,7 @@ unsigned int gnttab_max_grant_frames(void);
 #define gnttab_map_vaddr(map) ((void *)(map.host_virt_addr))
 
 int gnttab_map_refs(struct gnttab_map_grant_ref *map_ops,
-			struct gnttab_map_grant_ref *kmap_ops,
+		    struct gnttab_map_grant_ref *kmap_ops,
 		    struct page **pages, unsigned int count);
 int gnttab_unmap_refs(struct gnttab_unmap_grant_ref *unmap_ops,
 		      struct page **pages, unsigned int count);
-- 
1.7.6.4