* [PATCH 1/2] Add multi-page ring support to xen-blkback.
2010-01-12 17:44 Add multi-page shared ring support to xen-blkfront/back Paul Durrant
@ 2010-01-12 17:44 ` Paul Durrant
2010-01-12 17:44 ` [PATCH 2/2] Add multi-page ring support to xen-blkfront Paul Durrant
2010-01-12 17:51 ` Add multi-page shared ring support to xen-blkfront/back Paul Durrant
1 sibling, 1 reply; 9+ messages in thread
From: Paul Durrant @ 2010-01-12 17:44 UTC (permalink / raw)
To: xen-devel; +Cc: Paul Durrant
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
drivers/xen/blkback/common.h | 10 +++-
drivers/xen/blkback/interface.c | 113 +++++++++++++++++++++++++++-----------
drivers/xen/blkback/xenbus.c | 100 ++++++++++++++++++++++++++++++-----
3 files changed, 174 insertions(+), 49 deletions(-)
diff --git a/drivers/xen/blkback/common.h b/drivers/xen/blkback/common.h
index d12d4d5..6404a4a 100644
--- a/drivers/xen/blkback/common.h
+++ b/drivers/xen/blkback/common.h
@@ -58,6 +58,9 @@ struct vbd {
struct backend_info;
+#define BLKIF_MAX_RING_PAGE_ORDER 2
+#define BLKIF_MAX_RING_PAGES (1<<BLKIF_MAX_RING_PAGE_ORDER)
+
typedef struct blkif_st {
/* Unique identifier for this interface. */
domid_t domid;
@@ -92,14 +95,15 @@ typedef struct blkif_st {
wait_queue_head_t waiting_to_free;
- grant_handle_t shmem_handle;
- grant_ref_t shmem_ref;
+ unsigned int nr_shared_pages;
+ grant_handle_t shmem_handle[BLKIF_MAX_RING_PAGES];
} blkif_t;
blkif_t *blkif_alloc(domid_t domid);
void blkif_disconnect(blkif_t *blkif);
void blkif_free(blkif_t *blkif);
-int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn);
+int blkif_map(blkif_t *blkif, unsigned long shared_pages[],
+ unsigned int nr_shared_pages, unsigned int evtchn);
#define blkif_get(_b) (atomic_inc(&(_b)->refcnt))
#define blkif_put(_b) \
diff --git a/drivers/xen/blkback/interface.c b/drivers/xen/blkback/interface.c
index e397a41..8951543 100644
--- a/drivers/xen/blkback/interface.c
+++ b/drivers/xen/blkback/interface.c
@@ -56,50 +56,93 @@ blkif_t *blkif_alloc(domid_t domid)
return blkif;
}
-static int map_frontend_page(blkif_t *blkif, unsigned long shared_page)
-{
- struct gnttab_map_grant_ref op;
-
- gnttab_set_map_op(&op, (unsigned long)blkif->blk_ring_area->addr,
- GNTMAP_host_map, shared_page, blkif->domid);
+#define INVALID_GRANT_HANDLE ((grant_handle_t)~0U)
- if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
- BUG();
-
- if (op.status) {
- DPRINTK(" Grant table operation failure !\n");
- return op.status;
+static void unmap_frontend_pages(blkif_t *blkif)
+{
+ struct vm_struct *area = blkif->blk_ring_area;
+ struct gnttab_unmap_grant_ref op[BLKIF_MAX_RING_PAGES];
+ unsigned int i;
+ unsigned int j;
+
+ j = 0;
+ for (i = 0; i < blkif->nr_shared_pages; i++) {
+ unsigned long addr = (unsigned long)area->addr +
+ (i * PAGE_SIZE);
+
+ if (blkif->shmem_handle[i] != INVALID_GRANT_HANDLE) {
+ gnttab_set_unmap_op(&op[j++], addr,
+ GNTMAP_host_map,
+ blkif->shmem_handle[i]);
+
+ blkif->shmem_handle[i] = INVALID_GRANT_HANDLE;
+ }
}
- blkif->shmem_ref = shared_page;
- blkif->shmem_handle = op.handle;
+ blkif->nr_shared_pages = 0;
- return 0;
+ if (j != 0) {
+ if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref,
+ op, j))
+ BUG();
+ }
}
-static void unmap_frontend_page(blkif_t *blkif)
+static int map_frontend_pages(blkif_t *blkif, unsigned long shared_pages[],
+ unsigned int nr_shared_pages)
{
- struct gnttab_unmap_grant_ref op;
+ struct vm_struct *area = blkif->blk_ring_area;
+ struct gnttab_map_grant_ref op[BLKIF_MAX_RING_PAGES];
+ unsigned int i;
+ int status = 0;
- gnttab_set_unmap_op(&op, (unsigned long)blkif->blk_ring_area->addr,
- GNTMAP_host_map, blkif->shmem_handle);
+ for (i = 0; i < nr_shared_pages; i++) {
+ unsigned long addr = (unsigned long)area->addr +
+ (i * PAGE_SIZE);
+
+ gnttab_set_map_op(&op[i], addr, GNTMAP_host_map,
+ shared_pages[i], blkif->domid);
+ }
- if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+ if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, op,
+ nr_shared_pages))
BUG();
+
+ for (i = 0; i < nr_shared_pages; i++) {
+ status = op[i].status;
+ if (status != 0) {
+ blkif->shmem_handle[i] = INVALID_GRANT_HANDLE;
+ continue;
+ }
+
+ blkif->shmem_handle[i] = op[i].handle;
+ }
+
+ blkif->nr_shared_pages = nr_shared_pages;
+
+ if (status != 0) {
+ DPRINTK(" Grant table operation failure !\n");
+ unmap_frontend_pages(blkif);
+ }
+
+ return status;
}
-int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
+int blkif_map(blkif_t *blkif, unsigned long shared_pages[],
+ unsigned int nr_shared_pages, unsigned int evtchn)
{
+ unsigned long size = nr_shared_pages * PAGE_SIZE;
int err;
/* Already connected through? */
if (blkif->irq)
return 0;
- if ( (blkif->blk_ring_area = alloc_vm_area(PAGE_SIZE)) == NULL )
+ blkif->blk_ring_area = alloc_vm_area(size);
+ if (blkif->blk_ring_area == NULL)
return -ENOMEM;
- err = map_frontend_page(blkif, shared_page);
+ err = map_frontend_pages(blkif, shared_pages, nr_shared_pages);
if (err) {
free_vm_area(blkif->blk_ring_area);
return err;
@@ -110,21 +153,21 @@ int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
{
struct blkif_sring *sring;
sring = (struct blkif_sring *)blkif->blk_ring_area->addr;
- BACK_RING_INIT(&blkif->blk_rings.native, sring, PAGE_SIZE);
+ BACK_RING_INIT(&blkif->blk_rings.native, sring, size);
break;
}
case BLKIF_PROTOCOL_X86_32:
{
- struct blkif_x86_32_sring *sring_x86_32;
- sring_x86_32 = (struct blkif_x86_32_sring *)blkif->blk_ring_area->addr;
- BACK_RING_INIT(&blkif->blk_rings.x86_32, sring_x86_32, PAGE_SIZE);
+ struct blkif_x86_32_sring *sring;
+ sring = (struct blkif_x86_32_sring *)blkif->blk_ring_area->addr;
+ BACK_RING_INIT(&blkif->blk_rings.x86_32, sring, size);
break;
}
case BLKIF_PROTOCOL_X86_64:
{
- struct blkif_x86_64_sring *sring_x86_64;
- sring_x86_64 = (struct blkif_x86_64_sring *)blkif->blk_ring_area->addr;
- BACK_RING_INIT(&blkif->blk_rings.x86_64, sring_x86_64, PAGE_SIZE);
+ struct blkif_x86_64_sring *sring;
+ sring = (struct blkif_x86_64_sring *)blkif->blk_ring_area->addr;
+ BACK_RING_INIT(&blkif->blk_rings.x86_64, sring, size);
break;
}
default:
@@ -132,14 +175,17 @@ int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
}
err = bind_interdomain_evtchn_to_irqhandler(
- blkif->domid, evtchn, blkif_be_int, 0, "blkif-backend", blkif);
+ blkif->domid, evtchn, blkif_be_int, 0, "blkif-backend",
+ blkif);
if (err < 0)
{
- unmap_frontend_page(blkif);
+ unmap_frontend_pages(blkif);
free_vm_area(blkif->blk_ring_area);
+ blkif->blk_ring_area = NULL;
blkif->blk_rings.common.sring = NULL;
return err;
}
+
blkif->irq = err;
return 0;
@@ -162,8 +208,9 @@ void blkif_disconnect(blkif_t *blkif)
}
if (blkif->blk_rings.common.sring) {
- unmap_frontend_page(blkif);
+ unmap_frontend_pages(blkif);
free_vm_area(blkif->blk_ring_area);
+ blkif->blk_ring_area = NULL;
blkif->blk_rings.common.sring = NULL;
}
}
diff --git a/drivers/xen/blkback/xenbus.c b/drivers/xen/blkback/xenbus.c
index 04c0a12..4ee10b8 100644
--- a/drivers/xen/blkback/xenbus.c
+++ b/drivers/xen/blkback/xenbus.c
@@ -42,6 +42,11 @@ static int connect_ring(struct backend_info *);
static void backend_changed(struct xenbus_watch *, const char **,
unsigned int);
+/* Order of maximum shared ring size advertised to the front end. */
+static int blkif_max_ring_page_order;
+module_param_named(max_ring_page_order, blkif_max_ring_page_order, int, 0);
+MODULE_PARM_DESC(max_ring_page_order, "Order of maximum VM shared ring size");
+
static int blkback_name(blkif_t *blkif, char *buf)
{
char *devpath, *devname;
@@ -243,6 +248,11 @@ static int blkback_probe(struct xenbus_device *dev,
if (err)
goto fail;
+ err = xenbus_printf(XBT_NIL, dev->nodename, "max-ring-page-order",
+ "%u", blkif_max_ring_page_order);
+ if (err)
+ goto fail;
+
err = xenbus_switch_state(dev, XenbusStateInitWait);
if (err)
goto fail;
@@ -469,22 +479,82 @@ again:
static int connect_ring(struct backend_info *be)
{
struct xenbus_device *dev = be->dev;
- unsigned long ring_ref;
unsigned int evtchn;
+ unsigned int ring_order;
+ unsigned long ring_ref[BLKIF_MAX_RING_PAGES];
char protocol[64] = "";
int err;
DPRINTK("%s", dev->otherend);
- err = xenbus_gather(XBT_NIL, dev->otherend, "ring-ref", "%lu", &ring_ref,
- "event-channel", "%u", &evtchn, NULL);
- if (err) {
- xenbus_dev_fatal(dev, err,
- "reading %s/ring-ref and event-channel",
+ err = xenbus_scanf(XBT_NIL, dev->otherend, "event-channel", "%u",
+ &evtchn);
+ if (err != 1) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err, "reading %s/event-channel",
dev->otherend);
return err;
}
+ printk(KERN_INFO "blkback: event-channel %u\n", evtchn);
+
+ err = xenbus_scanf(XBT_NIL, dev->otherend, "ring-page-order", "%u",
+ &ring_order);
+ if (err != 1) {
+ DPRINTK("%s: using single page handshake", dev->otherend);
+
+ ring_order = 0;
+
+ err = xenbus_scanf(XBT_NIL, dev->otherend, "ring-ref",
+ "%lu", &ring_ref[0]);
+ if (err != 1) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err, "reading %s/ring-ref",
+ dev->otherend);
+ return err;
+ }
+
+ printk(KERN_INFO "blkback: ring-ref %lu\n", ring_ref[0]);
+ } else {
+ unsigned int i;
+
+ if (ring_order > blkif_max_ring_page_order) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err,
+ "%s/ring-page-order too big",
+ dev->otherend);
+ return err;
+ }
+
+ DPRINTK("%s: using %u page(s)", dev->otherend,
+ (1 << ring_order));
+
+ for (i = 0; i < (1u << ring_order); i++) {
+ char ring_ref_name[10];
+
+ snprintf(ring_ref_name, sizeof(ring_ref_name),
+ "ring-ref%1u", i);
+ err = xenbus_scanf(XBT_NIL, dev->otherend,
+ ring_ref_name, "%lu",
+ &ring_ref[i]);
+ if (err != 1) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err,
+ "reading %s/%s",
+ dev->otherend,
+ ring_ref_name);
+ return err;
+ }
+
+ printk(KERN_INFO "blkback: ring-ref%u %lu\n", i,
+ ring_ref[i]);
+ }
+ }
+
be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
"%63s", protocol, NULL);
@@ -497,18 +567,19 @@ static int connect_ring(struct backend_info *be)
else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_64))
be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_64;
else {
+ err = -EINVAL;
+
xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol);
- return -1;
+ return err;
}
- printk(KERN_INFO
- "blkback: ring-ref %ld, event-channel %d, protocol %d (%s)\n",
- ring_ref, evtchn, be->blkif->blk_protocol, protocol);
+
+ printk(KERN_INFO "blkback: protocol %d (%s)\n",
+ be->blkif->blk_protocol, protocol);
/* Map the shared frame, irq etc. */
- err = blkif_map(be->blkif, ring_ref, evtchn);
+ err = blkif_map(be->blkif, ring_ref, (1u << ring_order), evtchn);
if (err) {
- xenbus_dev_fatal(dev, err, "mapping ring-ref %lu port %u",
- ring_ref, evtchn);
+ xenbus_dev_fatal(dev, err, "mapping ring-refs and evtchn");
return err;
}
@@ -537,5 +608,8 @@ static struct xenbus_driver blkback = {
int blkif_xenbus_init(void)
{
+ if (blkif_max_ring_page_order > BLKIF_MAX_RING_PAGE_ORDER)
+ blkif_max_ring_page_order = BLKIF_MAX_RING_PAGE_ORDER;
+
return xenbus_register_backend(&blkback);
}
--
1.5.6.5
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: Add multi-page shared ring support to xen-blkfront/back
2010-01-12 17:44 Add multi-page shared ring support to xen-blkfront/back Paul Durrant
2010-01-12 17:44 ` [PATCH 1/2] Add multi-page ring support to xen-blkback Paul Durrant
@ 2010-01-12 17:51 ` Paul Durrant
2010-01-12 20:58 ` Add multi-page shared ring support toxen-blkfront/back James Harper
` (2 more replies)
1 sibling, 3 replies; 9+ messages in thread
From: Paul Durrant @ 2010-01-12 17:51 UTC (permalink / raw)
To: xen-devel@lists.xensource.com
[-- Attachment #1: Type: text/plain, Size: 1009 bytes --]
Paul Durrant wrote:
> These patches add support for multi-page shared rings to xen-blkfront/back.
> Shared rings must always be a power of 2 in size and so the size is controlled
> by the new 'max_ring_page_order' module parameter of xen-blkback. This defaults
> to 0 so operationally there should be no difference with these patches in place.
> The maximum value of 'max_ring_page_order' is currently limited to 2
> (i.e. 4 pages).
> xen-blkfront will always use the largest shared ring size supported by the back
> end. To override this and continue to use a single page ring, as before, set the
> 'allow_multi_page_ring' module parameter to 0.
>
I don't know why git-send-email seems to have sent my patch series into
a black hole. Here they are as attachments this time.
Paul
--
===============================
Paul Durrant, Software Engineer
Citrix Systems (R&D) Ltd.
First Floor, Building 101
Cambridge Science Park
Milton Road
Cambridge CB4 0FY
United Kingdom
===============================
[-- Attachment #2: 0001-Add-multi-page-ring-support-to-xen-blkback.patch --]
[-- Type: text/x-diff, Size: 11421 bytes --]
>From 6aea96d968f885fcdd252a17577a71b1742fac68 Mon Sep 17 00:00:00 2001
From: Paul Durrant <paul.durrant@citrix.com>
Date: Tue, 12 Jan 2010 17:25:25 +0000
Subject: [PATCH 1/2] Add multi-page ring support to xen-blkback.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
drivers/xen/blkback/common.h | 10 +++-
drivers/xen/blkback/interface.c | 113 +++++++++++++++++++++++++++-----------
drivers/xen/blkback/xenbus.c | 100 ++++++++++++++++++++++++++++++-----
3 files changed, 174 insertions(+), 49 deletions(-)
diff --git a/drivers/xen/blkback/common.h b/drivers/xen/blkback/common.h
index d12d4d5..6404a4a 100644
--- a/drivers/xen/blkback/common.h
+++ b/drivers/xen/blkback/common.h
@@ -58,6 +58,9 @@ struct vbd {
struct backend_info;
+#define BLKIF_MAX_RING_PAGE_ORDER 2
+#define BLKIF_MAX_RING_PAGES (1<<BLKIF_MAX_RING_PAGE_ORDER)
+
typedef struct blkif_st {
/* Unique identifier for this interface. */
domid_t domid;
@@ -92,14 +95,15 @@ typedef struct blkif_st {
wait_queue_head_t waiting_to_free;
- grant_handle_t shmem_handle;
- grant_ref_t shmem_ref;
+ unsigned int nr_shared_pages;
+ grant_handle_t shmem_handle[BLKIF_MAX_RING_PAGES];
} blkif_t;
blkif_t *blkif_alloc(domid_t domid);
void blkif_disconnect(blkif_t *blkif);
void blkif_free(blkif_t *blkif);
-int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn);
+int blkif_map(blkif_t *blkif, unsigned long shared_pages[],
+ unsigned int nr_shared_pages, unsigned int evtchn);
#define blkif_get(_b) (atomic_inc(&(_b)->refcnt))
#define blkif_put(_b) \
diff --git a/drivers/xen/blkback/interface.c b/drivers/xen/blkback/interface.c
index e397a41..8951543 100644
--- a/drivers/xen/blkback/interface.c
+++ b/drivers/xen/blkback/interface.c
@@ -56,50 +56,93 @@ blkif_t *blkif_alloc(domid_t domid)
return blkif;
}
-static int map_frontend_page(blkif_t *blkif, unsigned long shared_page)
-{
- struct gnttab_map_grant_ref op;
-
- gnttab_set_map_op(&op, (unsigned long)blkif->blk_ring_area->addr,
- GNTMAP_host_map, shared_page, blkif->domid);
+#define INVALID_GRANT_HANDLE ((grant_handle_t)~0U)
- if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
- BUG();
-
- if (op.status) {
- DPRINTK(" Grant table operation failure !\n");
- return op.status;
+static void unmap_frontend_pages(blkif_t *blkif)
+{
+ struct vm_struct *area = blkif->blk_ring_area;
+ struct gnttab_unmap_grant_ref op[BLKIF_MAX_RING_PAGES];
+ unsigned int i;
+ unsigned int j;
+
+ j = 0;
+ for (i = 0; i < blkif->nr_shared_pages; i++) {
+ unsigned long addr = (unsigned long)area->addr +
+ (i * PAGE_SIZE);
+
+ if (blkif->shmem_handle[i] != INVALID_GRANT_HANDLE) {
+ gnttab_set_unmap_op(&op[j++], addr,
+ GNTMAP_host_map,
+ blkif->shmem_handle[i]);
+
+ blkif->shmem_handle[i] = INVALID_GRANT_HANDLE;
+ }
}
- blkif->shmem_ref = shared_page;
- blkif->shmem_handle = op.handle;
+ blkif->nr_shared_pages = 0;
- return 0;
+ if (j != 0) {
+ if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref,
+ op, j))
+ BUG();
+ }
}
-static void unmap_frontend_page(blkif_t *blkif)
+static int map_frontend_pages(blkif_t *blkif, unsigned long shared_pages[],
+ unsigned int nr_shared_pages)
{
- struct gnttab_unmap_grant_ref op;
+ struct vm_struct *area = blkif->blk_ring_area;
+ struct gnttab_map_grant_ref op[BLKIF_MAX_RING_PAGES];
+ unsigned int i;
+ int status = 0;
- gnttab_set_unmap_op(&op, (unsigned long)blkif->blk_ring_area->addr,
- GNTMAP_host_map, blkif->shmem_handle);
+ for (i = 0; i < nr_shared_pages; i++) {
+ unsigned long addr = (unsigned long)area->addr +
+ (i * PAGE_SIZE);
+
+ gnttab_set_map_op(&op[i], addr, GNTMAP_host_map,
+ shared_pages[i], blkif->domid);
+ }
- if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+ if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, op,
+ nr_shared_pages))
BUG();
+
+ for (i = 0; i < nr_shared_pages; i++) {
+ status = op[i].status;
+ if (status != 0) {
+ blkif->shmem_handle[i] = INVALID_GRANT_HANDLE;
+ continue;
+ }
+
+ blkif->shmem_handle[i] = op[i].handle;
+ }
+
+ blkif->nr_shared_pages = nr_shared_pages;
+
+ if (status != 0) {
+ DPRINTK(" Grant table operation failure !\n");
+ unmap_frontend_pages(blkif);
+ }
+
+ return status;
}
-int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
+int blkif_map(blkif_t *blkif, unsigned long shared_pages[],
+ unsigned int nr_shared_pages, unsigned int evtchn)
{
+ unsigned long size = nr_shared_pages * PAGE_SIZE;
int err;
/* Already connected through? */
if (blkif->irq)
return 0;
- if ( (blkif->blk_ring_area = alloc_vm_area(PAGE_SIZE)) == NULL )
+ blkif->blk_ring_area = alloc_vm_area(size);
+ if (blkif->blk_ring_area == NULL)
return -ENOMEM;
- err = map_frontend_page(blkif, shared_page);
+ err = map_frontend_pages(blkif, shared_pages, nr_shared_pages);
if (err) {
free_vm_area(blkif->blk_ring_area);
return err;
@@ -110,21 +153,21 @@ int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
{
struct blkif_sring *sring;
sring = (struct blkif_sring *)blkif->blk_ring_area->addr;
- BACK_RING_INIT(&blkif->blk_rings.native, sring, PAGE_SIZE);
+ BACK_RING_INIT(&blkif->blk_rings.native, sring, size);
break;
}
case BLKIF_PROTOCOL_X86_32:
{
- struct blkif_x86_32_sring *sring_x86_32;
- sring_x86_32 = (struct blkif_x86_32_sring *)blkif->blk_ring_area->addr;
- BACK_RING_INIT(&blkif->blk_rings.x86_32, sring_x86_32, PAGE_SIZE);
+ struct blkif_x86_32_sring *sring;
+ sring = (struct blkif_x86_32_sring *)blkif->blk_ring_area->addr;
+ BACK_RING_INIT(&blkif->blk_rings.x86_32, sring, size);
break;
}
case BLKIF_PROTOCOL_X86_64:
{
- struct blkif_x86_64_sring *sring_x86_64;
- sring_x86_64 = (struct blkif_x86_64_sring *)blkif->blk_ring_area->addr;
- BACK_RING_INIT(&blkif->blk_rings.x86_64, sring_x86_64, PAGE_SIZE);
+ struct blkif_x86_64_sring *sring;
+ sring = (struct blkif_x86_64_sring *)blkif->blk_ring_area->addr;
+ BACK_RING_INIT(&blkif->blk_rings.x86_64, sring, size);
break;
}
default:
@@ -132,14 +175,17 @@ int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
}
err = bind_interdomain_evtchn_to_irqhandler(
- blkif->domid, evtchn, blkif_be_int, 0, "blkif-backend", blkif);
+ blkif->domid, evtchn, blkif_be_int, 0, "blkif-backend",
+ blkif);
if (err < 0)
{
- unmap_frontend_page(blkif);
+ unmap_frontend_pages(blkif);
free_vm_area(blkif->blk_ring_area);
+ blkif->blk_ring_area = NULL;
blkif->blk_rings.common.sring = NULL;
return err;
}
+
blkif->irq = err;
return 0;
@@ -162,8 +208,9 @@ void blkif_disconnect(blkif_t *blkif)
}
if (blkif->blk_rings.common.sring) {
- unmap_frontend_page(blkif);
+ unmap_frontend_pages(blkif);
free_vm_area(blkif->blk_ring_area);
+ blkif->blk_ring_area = NULL;
blkif->blk_rings.common.sring = NULL;
}
}
diff --git a/drivers/xen/blkback/xenbus.c b/drivers/xen/blkback/xenbus.c
index 04c0a12..4ee10b8 100644
--- a/drivers/xen/blkback/xenbus.c
+++ b/drivers/xen/blkback/xenbus.c
@@ -42,6 +42,11 @@ static int connect_ring(struct backend_info *);
static void backend_changed(struct xenbus_watch *, const char **,
unsigned int);
+/* Order of maximum shared ring size advertised to the front end. */
+static int blkif_max_ring_page_order;
+module_param_named(max_ring_page_order, blkif_max_ring_page_order, int, 0);
+MODULE_PARM_DESC(max_ring_page_order, "Order of maximum VM shared ring size");
+
static int blkback_name(blkif_t *blkif, char *buf)
{
char *devpath, *devname;
@@ -243,6 +248,11 @@ static int blkback_probe(struct xenbus_device *dev,
if (err)
goto fail;
+ err = xenbus_printf(XBT_NIL, dev->nodename, "max-ring-page-order",
+ "%u", blkif_max_ring_page_order);
+ if (err)
+ goto fail;
+
err = xenbus_switch_state(dev, XenbusStateInitWait);
if (err)
goto fail;
@@ -469,22 +479,82 @@ again:
static int connect_ring(struct backend_info *be)
{
struct xenbus_device *dev = be->dev;
- unsigned long ring_ref;
unsigned int evtchn;
+ unsigned int ring_order;
+ unsigned long ring_ref[BLKIF_MAX_RING_PAGES];
char protocol[64] = "";
int err;
DPRINTK("%s", dev->otherend);
- err = xenbus_gather(XBT_NIL, dev->otherend, "ring-ref", "%lu", &ring_ref,
- "event-channel", "%u", &evtchn, NULL);
- if (err) {
- xenbus_dev_fatal(dev, err,
- "reading %s/ring-ref and event-channel",
+ err = xenbus_scanf(XBT_NIL, dev->otherend, "event-channel", "%u",
+ &evtchn);
+ if (err != 1) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err, "reading %s/event-channel",
dev->otherend);
return err;
}
+ printk(KERN_INFO "blkback: event-channel %u\n", evtchn);
+
+ err = xenbus_scanf(XBT_NIL, dev->otherend, "ring-page-order", "%u",
+ &ring_order);
+ if (err != 1) {
+ DPRINTK("%s: using single page handshake", dev->otherend);
+
+ ring_order = 0;
+
+ err = xenbus_scanf(XBT_NIL, dev->otherend, "ring-ref",
+ "%lu", &ring_ref[0]);
+ if (err != 1) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err, "reading %s/ring-ref",
+ dev->otherend);
+ return err;
+ }
+
+ printk(KERN_INFO "blkback: ring-ref %lu\n", ring_ref[0]);
+ } else {
+ unsigned int i;
+
+ if (ring_order > blkif_max_ring_page_order) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err,
+ "%s/ring-page-order too big",
+ dev->otherend);
+ return err;
+ }
+
+ DPRINTK("%s: using %u page(s)", dev->otherend,
+ (1 << ring_order));
+
+ for (i = 0; i < (1u << ring_order); i++) {
+ char ring_ref_name[10];
+
+ snprintf(ring_ref_name, sizeof(ring_ref_name),
+ "ring-ref%1u", i);
+ err = xenbus_scanf(XBT_NIL, dev->otherend,
+ ring_ref_name, "%lu",
+ &ring_ref[i]);
+ if (err != 1) {
+ err = -EINVAL;
+
+ xenbus_dev_fatal(dev, err,
+ "reading %s/%s",
+ dev->otherend,
+ ring_ref_name);
+ return err;
+ }
+
+ printk(KERN_INFO "blkback: ring-ref%u %lu\n", i,
+ ring_ref[i]);
+ }
+ }
+
be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
"%63s", protocol, NULL);
@@ -497,18 +567,19 @@ static int connect_ring(struct backend_info *be)
else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_64))
be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_64;
else {
+ err = -EINVAL;
+
xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol);
- return -1;
+ return err;
}
- printk(KERN_INFO
- "blkback: ring-ref %ld, event-channel %d, protocol %d (%s)\n",
- ring_ref, evtchn, be->blkif->blk_protocol, protocol);
+
+ printk(KERN_INFO "blkback: protocol %d (%s)\n",
+ be->blkif->blk_protocol, protocol);
/* Map the shared frame, irq etc. */
- err = blkif_map(be->blkif, ring_ref, evtchn);
+ err = blkif_map(be->blkif, ring_ref, (1u << ring_order), evtchn);
if (err) {
- xenbus_dev_fatal(dev, err, "mapping ring-ref %lu port %u",
- ring_ref, evtchn);
+ xenbus_dev_fatal(dev, err, "mapping ring-refs and evtchn");
return err;
}
@@ -537,5 +608,8 @@ static struct xenbus_driver blkback = {
int blkif_xenbus_init(void)
{
+ if (blkif_max_ring_page_order > BLKIF_MAX_RING_PAGE_ORDER)
+ blkif_max_ring_page_order = BLKIF_MAX_RING_PAGE_ORDER;
+
return xenbus_register_backend(&blkback);
}
--
1.5.6.5
[-- Attachment #3: 0002-Add-multi-page-ring-support-to-xen-blkfront.patch --]
[-- Type: text/x-diff, Size: 8425 bytes --]
>From fe4d68a612ae740cef39ae0e3ce714c40e375cb9 Mon Sep 17 00:00:00 2001
From: Paul Durrant <paul.durrant@citrix.com>
Date: Tue, 12 Jan 2010 17:32:36 +0000
Subject: [PATCH 2/2] Add multi-page ring support to xen-blkfront
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
---
drivers/block/xen-blkfront.c | 137 ++++++++++++++++++++++++++++++++----------
1 files changed, 105 insertions(+), 32 deletions(-)
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 837b992..1aaa5a4 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -67,7 +67,25 @@ struct blk_shadow {
static struct block_device_operations xlvbd_block_fops;
-#define BLK_RING_SIZE __RING_SIZE((struct blkif_sring *)0, PAGE_SIZE)
+#define BLK_RING_SIZE(_order) \
+ __RING_SIZE((struct blkif_sring *)0, PAGE_SIZE<<(_order))
+
+#define BLKIF_MAX_RING_PAGE_ORDER 2
+#define MAX_BLK_RING_SIZE BLK_RING_SIZE(BLKIF_MAX_RING_PAGE_ORDER)
+
+#define BLKIF_MAX_RING_PAGES (1<<BLKIF_MAX_RING_PAGE_ORDER)
+
+/*
+ * Allow multi page shared ring to be used if the capability is advertised by
+ * the back end.
+ */
+static int allow_multi_page_ring = 1;
+module_param_named(allow_multi_page_ring,
+ allow_multi_page_ring,
+ int,
+ 0);
+MODULE_PARM_DESC(allow_multi_page_ring,
+ "Allow multi page shared ring to be used");
/*
* We have one of these per vbd, whether ide, scsi or 'other'. They
@@ -81,14 +99,15 @@ struct blkfront_info
int vdevice;
blkif_vdev_t handle;
enum blkif_state connected;
- int ring_ref;
+ int ring_ref[BLKIF_MAX_RING_PAGES];
+ int ring_order;
struct blkif_front_ring ring;
struct scatterlist sg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
unsigned int evtchn, irq;
struct request_queue *rq;
struct work_struct work;
struct gnttab_free_callback callback;
- struct blk_shadow shadow[BLK_RING_SIZE];
+ struct blk_shadow shadow[MAX_BLK_RING_SIZE];
unsigned long shadow_free;
int feature_barrier;
int is_ready;
@@ -102,8 +121,6 @@ struct blkfront_info
static DEFINE_SPINLOCK(blkif_io_lock);
-#define MAXIMUM_OUTSTANDING_BLOCK_REQS \
- (BLKIF_MAX_SEGMENTS_PER_REQUEST * BLK_RING_SIZE)
#define GRANT_INVALID_REF 0
#define PARTS_PER_DISK 16
@@ -122,7 +139,7 @@ static DEFINE_SPINLOCK(blkif_io_lock);
static int get_id_from_freelist(struct blkfront_info *info)
{
unsigned long free = info->shadow_free;
- BUG_ON(free >= BLK_RING_SIZE);
+ BUG_ON(free >= MAX_BLK_RING_SIZE);
info->shadow_free = info->shadow[free].req.id;
info->shadow[free].req.id = 0x0fffffee; /* debug */
return free;
@@ -495,6 +512,9 @@ static void blkif_restart_queue(struct work_struct *work)
static void blkif_free(struct blkfront_info *info, int suspend)
{
+ int i;
+ unsigned long sring = (unsigned long)info->ring.sring;
+
/* Prevent new requests being issued until we fix things up. */
spin_lock_irq(&blkif_io_lock);
info->connected = suspend ?
@@ -510,12 +530,14 @@ static void blkif_free(struct blkfront_info *info, int suspend)
flush_scheduled_work();
/* Free resources associated with old device channel. */
- if (info->ring_ref != GRANT_INVALID_REF) {
- gnttab_end_foreign_access(info->ring_ref, 0,
- (unsigned long)info->ring.sring);
- info->ring_ref = GRANT_INVALID_REF;
- info->ring.sring = NULL;
+ for (i = 0; i < (1<<info->ring_order); i++) {
+ if (info->ring_ref[i] != GRANT_INVALID_REF) {
+ gnttab_end_foreign_access(info->ring_ref[i], 0,
+ sring + (i<<PAGE_SHIFT));
+ info->ring_ref[i] = GRANT_INVALID_REF;
+ }
}
+ info->ring.sring = NULL;
if (info->irq)
unbind_from_irqhandler(info->irq, info);
info->evtchn = info->irq = 0;
@@ -605,28 +627,35 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
static int setup_blkring(struct xenbus_device *dev,
struct blkfront_info *info)
{
- struct blkif_sring *sring;
+ int i;
+ unsigned long sring;
int err;
- info->ring_ref = GRANT_INVALID_REF;
+ for (i = 0; i < (1<<info->ring_order); i++)
+ info->ring_ref[i] = GRANT_INVALID_REF;
- sring = (struct blkif_sring *)__get_free_page(GFP_NOIO | __GFP_HIGH);
+ sring = __get_free_pages(GFP_NOIO | __GFP_HIGH, info->ring_order);
if (!sring) {
xenbus_dev_fatal(dev, -ENOMEM, "allocating shared ring");
return -ENOMEM;
}
- SHARED_RING_INIT(sring);
- FRONT_RING_INIT(&info->ring, sring, PAGE_SIZE);
+ SHARED_RING_INIT((struct blkif_sring *)sring);
+ FRONT_RING_INIT(&info->ring, (struct blkif_sring *)sring,
+ PAGE_SIZE<<info->ring_order);
sg_init_table(info->sg, BLKIF_MAX_SEGMENTS_PER_REQUEST);
- err = xenbus_grant_ring(dev, virt_to_mfn(info->ring.sring));
- if (err < 0) {
- free_page((unsigned long)sring);
- info->ring.sring = NULL;
- goto fail;
+ for (i = 0; i < (1<<info->ring_order); i++) {
+ unsigned long addr = sring + (i<<PAGE_SHIFT);
+
+ err = xenbus_grant_ring(dev, virt_to_mfn(addr));
+ if (err < 0) {
+ free_pages(sring, info->ring_order);
+ info->ring.sring = NULL;
+ goto fail;
+ }
+ info->ring_ref[i] = err;
}
- info->ring_ref = err;
err = xenbus_alloc_evtchn(dev, &info->evtchn);
if (err)
@@ -648,6 +677,9 @@ fail:
return err;
}
+#ifndef MIN
+#define MIN(a, b) (((a) < (b)) ? (a) : (b))
+#endif
/* Common code used when first setting up, and when resuming. */
static int talk_to_blkback(struct xenbus_device *dev,
@@ -655,8 +687,24 @@ static int talk_to_blkback(struct xenbus_device *dev,
{
const char *message = NULL;
struct xenbus_transaction xbt;
+ int ring_order;
+ int single_page;
int err;
+ err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
+ "max-ring-page-order", "%u", &ring_order,
+ NULL);
+ if (err || !allow_multi_page_ring) {
+ single_page = 1;
+ info->ring_order = 0;
+ dev_dbg(&dev->dev, "using single-page handshake\n");
+ } else {
+ single_page = 0;
+ info->ring_order = MIN(ring_order, BLKIF_MAX_RING_PAGE_ORDER);
+ dev_dbg(&dev->dev, "using multi-page handshake (%d pages)\n",
+ (1<<info->ring_order));
+ }
+
/* Create shared ring, alloc event channel. */
err = setup_blkring(dev, info);
if (err)
@@ -669,18 +717,43 @@ again:
goto destroy_blkring;
}
- err = xenbus_printf(xbt, dev->nodename,
- "ring-ref", "%u", info->ring_ref);
- if (err) {
- message = "writing ring-ref";
- goto abort_transaction;
+ if (single_page) {
+ err = xenbus_printf(xbt, dev->nodename,
+ "ring-ref", "%u", info->ring_ref[0]);
+ if (err) {
+ message = "writing ring-ref";
+ goto abort_transaction;
+ }
+ } else {
+ int i;
+
+ err = xenbus_printf(xbt, dev->nodename,
+ "ring-page-order", "%u", info->ring_order);
+ if (err) {
+ message = "writing ring-page-order";
+ goto abort_transaction;
+ }
+
+ for (i = 0; i < (1<<info->ring_order); i++) {
+ char node[10];
+
+ snprintf(node, sizeof(node), "ring-ref%u", i);
+ err = xenbus_printf(xbt, dev->nodename, node, "%u",
+ info->ring_ref[i]);
+ if (err) {
+ message = "writing ring-ref";
+ goto abort_transaction;
+ }
+ }
}
+
err = xenbus_printf(xbt, dev->nodename,
"event-channel", "%u", info->evtchn);
if (err) {
message = "writing event-channel";
goto abort_transaction;
}
+
err = xenbus_printf(xbt, dev->nodename, "protocol", "%s",
XEN_IO_PROTO_ABI_NATIVE);
if (err) {
@@ -747,9 +820,9 @@ static int blkfront_probe(struct xenbus_device *dev,
info->connected = BLKIF_STATE_DISCONNECTED;
INIT_WORK(&info->work, blkif_restart_queue);
- for (i = 0; i < BLK_RING_SIZE; i++)
+ for (i = 0; i < MAX_BLK_RING_SIZE; i++)
info->shadow[i].req.id = i+1;
- info->shadow[BLK_RING_SIZE-1].req.id = 0x0fffffff;
+ info->shadow[MAX_BLK_RING_SIZE-1].req.id = 0x0fffffff;
/* Front end dir is a number, which is used as the id. */
info->handle = simple_strtoul(strrchr(dev->nodename, '/')+1, NULL, 0);
@@ -782,13 +855,13 @@ static int blkif_recover(struct blkfront_info *info)
/* Stage 2: Set up free list. */
memset(&info->shadow, 0, sizeof(info->shadow));
- for (i = 0; i < BLK_RING_SIZE; i++)
+ for (i = 0; i < MAX_BLK_RING_SIZE; i++)
info->shadow[i].req.id = i+1;
info->shadow_free = info->ring.req_prod_pvt;
- info->shadow[BLK_RING_SIZE-1].req.id = 0x0fffffff;
+ info->shadow[MAX_BLK_RING_SIZE-1].req.id = 0x0fffffff;
/* Stage 3: Find pending requests and requeue them. */
- for (i = 0; i < BLK_RING_SIZE; i++) {
+ for (i = 0; i < MAX_BLK_RING_SIZE; i++) {
/* Not in use? */
if (copy[i].request == 0)
continue;
--
1.5.6.5
[-- Attachment #4: Type: text/plain, Size: 138 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
^ permalink raw reply related [flat|nested] 9+ messages in thread