WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-ia64-devel

Re: [Xen-ia64-devel] [PATCH 2/2] remove xencomm page size limit(linux si

To: Alex Williamson <alex.williamson@xxxxxx>
Subject: Re: [Xen-ia64-devel] [PATCH 2/2] remove xencomm page size limit(linux side)
From: Isaku Yamahata <yamahata@xxxxxxxxxxxxx>
Date: Wed, 1 Aug 2007 15:39:38 +0900
Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx, xen-ppc-devel <xen-ppc-devel@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Tue, 31 Jul 2007 23:37:04 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <1185913843.6802.79.camel@bling>
List-help: <mailto:xen-ia64-devel-request@lists.xensource.com?subject=help>
List-id: Discussion of the ia64 port of Xen <xen-ia64-devel.lists.xensource.com>
List-post: <mailto:xen-ia64-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ia64-devel>, <mailto:xen-ia64-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ia64-devel>, <mailto:xen-ia64-devel-request@lists.xensource.com?subject=unsubscribe>
References: <20070731061019.GB6039%yamahata@xxxxxxxxxxxxx> <20070731061234.GC6039%yamahata@xxxxxxxxxxxxx> <1185913843.6802.79.camel@bling>
Sender: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Mutt/1.4.2.1i
On Tue, Jul 31, 2007 at 02:30:43PM -0600, Alex Williamson wrote:
>    The 32bit restriction on this one is unfortunate, especially if we
> want to also apply this to the common xencomm code.  IIRC, ppc supports
> 32bit guests.  Maybe we need to keep a page allocation interface for
> 32bit and a smaller granularity allocation for 64bit?  I think a ~100GB
> limit would be more than acceptable for a 32bit guest.  Thanks,

Here is the patch for common code.
"if (sizeof(*desc) > sizeof(void*))" is somewhat ugly, but it should work.
I did only compile test. The patch needs review and test by PPC developer
before commit.


# HG changeset patch
# User yamahata@xxxxxxxxxxxxx
# Date 1185948239 -32400
# Node ID 99bc011645da5b4a10bc5aa48f806f2993313a5e
# Parent  14380a7179e6a0d6759454132fa1f2135bd3b8e3
remove xencomm page size limit.
Currently xencomm has page size limit so that a domain with many memory
(e.g. 100GB~) can't be created.

Now that xencomm of xen side accepts struct xencomm_desc whose address array
crosses page boundary. Thus it isn't necessary to allocate single page
not to cross page boundary. We can allocate exact sized memory.
Note that struct xencomm_desc can't cross page boundary and slab allocator
returns sizeof(void*) aligned pointer.
Where sizeof(*desc) > sizeof(void*), e.g. 32 bit environment,
the slab allocator return pointer doesn't gurantee that
struct xencomm_desc doesn't cross page boundary. So we fall back to
page allocator.
PATCHNAME: remove_xencomm_page_size_limit_common_code

Signed-off-by: Isaku Yamahata <yamahata@xxxxxxxxxxxxx>

diff -r 14380a7179e6 -r 99bc011645da drivers/xen/core/xencomm.c
--- a/drivers/xen/core/xencomm.c        Mon Jul 30 11:38:32 2007 +0900
+++ b/drivers/xen/core/xencomm.c        Wed Aug 01 15:03:59 2007 +0900
@@ -65,25 +65,54 @@ static int xencomm_init(struct xencomm_d
        return 0;
 }
 
-/* XXX use slab allocator */
-static struct xencomm_desc *xencomm_alloc(gfp_t gfp_mask)
-{
-       struct xencomm_desc *desc;
-
-       desc = (struct xencomm_desc *)__get_free_page(gfp_mask);
-       if (desc == NULL)
-               return NULL;
-
-       desc->nr_addrs = (PAGE_SIZE - sizeof(struct xencomm_desc)) /
+static struct xencomm_desc *xencomm_alloc(gfp_t gfp_mask,
+                                         void *buffer, unsigned long bytes)
+{
+       struct xencomm_desc *desc;
+       unsigned long buffer_ulong = (unsigned long)buffer;
+       unsigned long start = buffer_ulong & PAGE_MASK;
+       unsigned long end = (buffer_ulong + bytes) | ~PAGE_MASK;
+       unsigned long nr_addrs = (end - start + 1) >> PAGE_SHIFT;
+       unsigned long size = sizeof(*desc) +
+               sizeof(desc->address[0]) * nr_addrs;
+
+       /*
+        * slab allocator returns at least sizeof(void*) aligned pointer.
+        * When sizeof(*desc) > sizeof(void*), struct xencomm_desc might
+        * cross page boundary.
+        */
+       if (sizeof(*desc) > sizeof(void*)) {
+               unsigned long order = get_order(size);
+               desc = (struct xencomm_desc *)__get_free_pages(gfp_mask,
+                                                              order);
+               if (desc == NULL)
+                       return NULL;
+
+               desc->nr_addrs =
+                       ((PAGE_SIZE << order) - sizeof(struct xencomm_desc)) /
                        sizeof(*desc->address);
-
+       } else {
+               desc = kmalloc(size, gfp_mask);
+               if (desc == NULL)
+                       return NULL;
+
+               desc->nr_addrs = nr_addrs;
+       }
        return desc;
 }
 
 void xencomm_free(void *desc)
 {
-       if (desc && !((ulong)desc & XENCOMM_INLINE_FLAG))
-               free_page((unsigned long)__va(desc));
+       if (desc && !((ulong)desc & XENCOMM_INLINE_FLAG)) {
+               struct xencomm_desc *desc__ = (struct xencomm_desc*)desc;
+               if (sizeof(*desc__) > sizeof(void*)) {
+                       unsigned long size = sizeof(*desc__) +
+                               sizeof(desc__->address[0]) * desc__->nr_addrs;
+                       unsigned long order = get_order(size);
+                       free_pages((unsigned long)__va(desc), order);
+               } else
+                       kfree(__va(desc));
+       }
 }
 
 static int xencomm_create(void *buffer, unsigned long bytes, struct 
xencomm_desc **ret, gfp_t gfp_mask)
@@ -102,7 +131,7 @@ static int xencomm_create(void *buffer, 
 
        BUG_ON(buffer == NULL); /* 'bytes' is non-zero */
 
-       desc = xencomm_alloc(gfp_mask);
+       desc = xencomm_alloc(gfp_mask, buffer, bytes);
        if (!desc) {
                printk("%s failure\n", "xencomm_alloc");
                return -ENOMEM;

-- 
yamahata

Attachment: 165_99bc011645da_remove_xencomm_page_size_limit_common_code.patch
Description: Text Data

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel