From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750940AbWCUTpu (ORCPT ); Tue, 21 Mar 2006 14:45:50 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751497AbWCUTpu (ORCPT ); Tue, 21 Mar 2006 14:45:50 -0500 Received: from stat9.steeleye.com ([209.192.50.41]:30893 "EHLO hancock.sc.steeleye.com") by vger.kernel.org with ESMTP id S1750940AbWCUTpt (ORCPT ); Tue, 21 Mar 2006 14:45:49 -0500 Subject: [PATCH 2/2] add flush_kernel_dcache_page() API From: James Bottomley To: Linus Torvalds , Andrew Morton Cc: linux-kernel Content-Type: text/plain Date: Tue, 21 Mar 2006 13:45:41 -0600 Message-Id: <1142970341.3428.15.camel@mulgrave.il.steeleye.com> Mime-Version: 1.0 X-Mailer: Evolution 2.2.3 (2.2.3-2.fc4) Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org We have a problem in a lot of emulated storage in that it takes a page from get_user_pages() and does something like kmap_atomic(page) modify page kunmap_atomic(page) However, nothing has flushed the kernel cache view of the page before the kunmap. We need a lightweight API to do this, so this new API would specifically be for flushing the kernel cache view of a user page which the kernel has modified. The driver would need to add flush_kernel_dcache_page(page) before the final kunmap. Signed-off-by: James Bottomley Index: BUILD-2.6/Documentation/cachetlb.txt =================================================================== --- BUILD-2.6.orig/Documentation/cachetlb.txt 2006-03-21 13:37:45.000000000 -0600 +++ BUILD-2.6/Documentation/cachetlb.txt 2006-03-21 13:37:54.000000000 -0600 @@ -371,6 +371,18 @@ architectures). For incoherent architectures, it should flush the cache of the page at vmaddr in the current user process. + void flush_kernel_dcache_page(struct page *page) + When the kernel needs to modify a user page is has obtained + with kmap, it calls this function after all modifications are + complete (but before kunmapping it) to bring the underlying + page up to date. It is assumed here that the user has no + incoherent cached copies (i.e. the original page was obtained + from a mechanism like get_user_pages()). The default + implementation is a nop and should remain so on all coherent + architectures. On incoherent architectures, this should flush + the kernel cache for page (using page_address(page)). + + 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. Index: BUILD-2.6/include/linux/highmem.h =================================================================== --- BUILD-2.6.orig/include/linux/highmem.h 2006-03-21 13:37:19.000000000 -0600 +++ BUILD-2.6/include/linux/highmem.h 2006-03-21 13:37:54.000000000 -0600 @@ -13,6 +13,12 @@ } #endif +#ifndef ARCH_HAS_FLUSH_KERNEL_DCACHE_PAGE +static inline void flush_kernel_dcache_page(struct page *page) +{ +} +#endif + #ifdef CONFIG_HIGHMEM #include