From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from e24smtp04.br.ibm.com (e24smtp04.br.ibm.com [32.104.18.25]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "e24smtp04.br.ibm.com", Issuer "GeoTrust SSL CA" (not verified)) by ozlabs.org (Postfix) with ESMTPS id 289CA2C0095 for ; Fri, 2 Aug 2013 22:10:04 +1000 (EST) Received: from /spool/local by e24smtp04.br.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Fri, 2 Aug 2013 09:09:58 -0300 Received: from d24relay03.br.ibm.com (d24relay03.br.ibm.com [9.13.184.25]) by d24dlp02.br.ibm.com (Postfix) with ESMTP id 8EEAE1DC0066 for ; Fri, 2 Aug 2013 08:09:55 -0400 (EDT) Received: from d24av03.br.ibm.com (d24av03.br.ibm.com [9.8.31.95]) by d24relay03.br.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id r72C8Xkj22478908 for ; Fri, 2 Aug 2013 09:08:33 -0300 Received: from d24av03.br.ibm.com (localhost [127.0.0.1]) by d24av03.br.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id r72C9tSc010912 for ; Fri, 2 Aug 2013 09:09:55 -0300 From: Marcelo Cerri To: benh@kernel.crashing.org Subject: [PATCH v2 1/2] drivers/crypto/nx: fix physical addresses added to sg lists Date: Fri, 2 Aug 2013 09:09:51 -0300 Message-Id: <1375445392-16237-2-git-send-email-mhcerri@linux.vnet.ibm.com> In-Reply-To: <1375445392-16237-1-git-send-email-mhcerri@linux.vnet.ibm.com> References: <1375445392-16237-1-git-send-email-mhcerri@linux.vnet.ibm.com> Cc: Marcelo Cerri , linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org, linux-crypto@vger.kernel.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , The co-processor receives data to be hashed through scatter/gather lists pointing to physical addresses. When a vmalloc'ed data is given, the driver must calculate the physical address to each page of the data. However the current version of it just calculates the physical address once and keeps incrementing it even when a page boundary is crossed. This patch fixes this behaviour. Reviewed-by: Fionnuala Gunter Reviewed-by: Joel Schopp Reviewed-by: Joy Latten Signed-off-by: Marcelo Cerri --- drivers/crypto/nx/nx.c | 22 +++++++++++++++++++--- 1 file changed, 19 insertions(+), 3 deletions(-) diff --git a/drivers/crypto/nx/nx.c b/drivers/crypto/nx/nx.c index bbdab6e..ad07dc6 100644 --- a/drivers/crypto/nx/nx.c +++ b/drivers/crypto/nx/nx.c @@ -114,13 +114,29 @@ struct nx_sg *nx_build_sg_list(struct nx_sg *sg_head, * have been described (or @sgmax elements have been written), the * loop ends. min_t is used to ensure @end_addr falls on the same page * as sg_addr, if not, we need to create another nx_sg element for the - * data on the next page */ + * data on the next page. + * + * Also when using vmalloc'ed data, every time that a system page + * boundary is crossed the physical address needs to be re-calculated. + */ for (sg = sg_head; sg_len < len; sg++) { + u64 next_page; + sg->addr = sg_addr; - sg_addr = min_t(u64, NX_PAGE_NUM(sg_addr + NX_PAGE_SIZE), end_addr); - sg->len = sg_addr - sg->addr; + sg_addr = min_t(u64, NX_PAGE_NUM(sg_addr + NX_PAGE_SIZE), + end_addr); + + next_page = (sg->addr & PAGE_MASK) + PAGE_SIZE; + sg->len = min_t(u64, sg_addr, next_page) - sg->addr; sg_len += sg->len; + if (sg_addr >= next_page && + is_vmalloc_addr(start_addr + sg_len)) { + sg_addr = page_to_phys(vmalloc_to_page( + start_addr + sg_len)); + end_addr = sg_addr + len - sg_len; + } + if ((sg - sg_head) == sgmax) { pr_err("nx: scatter/gather list overflow, pid: %d\n", current->pid); -- 1.7.12