[PATCH] Add API for flushing Anon pages

Currently, get_user_pages() returns fully coherent pages to the kernel for
anything other than anonymous pages.  This is a problem for things like
fuse and the SCSI generic ioctl SG_IO which can potentially wish to do DMA
to anonymous pages passed in by users.

The fix is to add a new memory management API: flush_anon_page() which
is used in get_user_pages() to make anonymous pages coherent.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Documentation/cachetlb.txt

@@ -362,6 +362,15 @@ maps this page at its virtual address.
 	likely that you will need to flush the instruction cache
 	for copy_to_user_page().
 
+  void flush_anon_page(struct page *page, unsigned long vmaddr)
+  	When the kernel needs to access the contents of an anonymous
+	page, it calls this function (currently only
+	get_user_pages()).  Note: flush_dcache_page() deliberately
+	doesn't work for an anonymous page.  The default
+	implementation is a nop (and should remain so for all coherent
+	architectures).  For incoherent architectures, it should flush
+	the cache of the page at vmaddr in the current user process.
+
   void flush_icache_range(unsigned long start, unsigned long end)
   	When the kernel stores into addresses that it will execute
 	out of (eg when loading modules), this function is called.

include/linux/highmem.h

@@ -7,6 +7,12 @@
 
 #include <asm/cacheflush.h>
 
+#ifndef ARCH_HAS_FLUSH_ANON_PAGE
+static inline void flush_anon_page(struct page *page, unsigned long vmaddr)
+{
+}
+#endif
+
 #ifdef CONFIG_HIGHMEM
 
 #include <asm/highmem.h>

mm/memory.c

@@ -1071,6 +1071,8 @@ int get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
 			}
 			if (pages) {
 				pages[i] = page;
+
+				flush_anon_page(page, start);
 				flush_dcache_page(page);
 			}
 			if (vmas)