From 87afd448b186c885d67a08b7417cd46253b6a9d6 Mon Sep 17 00:00:00 2001 From: Eli Cohen Date: Mon, 23 Jun 2008 09:29:58 -0700 Subject: IB/mthca: Clear ICM pages before handing to FW Current memfree FW has a bug which in some cases, assumes that ICM pages passed to it are cleared. This patch uses __GFP_ZERO to allocate all ICM pages passed to the FW. Once firmware with a fix is released, we can make the workaround conditional on firmware version. This fixes the bug reported by Arthur Kepner here: http://lists.openfabrics.org/pipermail/general/2008-May/050026.html Cc: Signed-off-by: Eli Cohen [ Rewritten to be a one-liner using __GFP_ZERO instead of vmap()ing ICM memory and memset()ing it to 0. - Roland ] Signed-off-by: Roland Dreier --- drivers/infiniband/hw/mthca/mthca_memfree.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/infiniband/hw/mthca/mthca_memfree.c b/drivers/infiniband/hw/mthca/mthca_memfree.c index b224079d4e1..d5862e5d99a 100644 --- a/drivers/infiniband/hw/mthca/mthca_memfree.c +++ b/drivers/infiniband/hw/mthca/mthca_memfree.c @@ -109,7 +109,11 @@ static int mthca_alloc_icm_pages(struct scatterlist *mem, int order, gfp_t gfp_m { struct page *page; - page = alloc_pages(gfp_mask, order); + /* + * Use __GFP_ZERO because buggy firmware assumes ICM pages are + * cleared, and subtle failures are seen if they aren't. + */ + page = alloc_pages(gfp_mask | __GFP_ZERO, order); if (!page) return -ENOMEM; -- cgit v1.2.3