From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759945Ab2D0JQr (ORCPT ); Fri, 27 Apr 2012 05:16:47 -0400 Received: from mx1.redhat.com ([209.132.183.28]:44117 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754254Ab2D0JQq (ORCPT ); Fri, 27 Apr 2012 05:16:46 -0400 Subject: Re: [RFC] vmalloc: add warning in __vmalloc From: Steven Whitehouse To: Minchan Kim Cc: Andrew Morton , linux-mm@kvack.org, linux-kernel@vger.kernel.org, npiggin@gmail.com, kamezawa.hiroyu@jp.fujitsu.com, kosaki.motohiro@gmail.com, rientjes@google.com, Neil Brown , Artem Bityutskiy , David Woodhouse , "Theodore Ts'o" , Adrian Hunter , "David S. Miller" , James Morris , Alexander Viro , Sage Weil In-Reply-To: <1335516144-3486-1-git-send-email-minchan@kernel.org> References: <1335516144-3486-1-git-send-email-minchan@kernel.org> Content-Type: text/plain; charset="UTF-8" Organization: Red Hat UK Ltd Date: Fri, 27 Apr 2012 10:16:00 +0100 Message-ID: <1335518161.2686.2.camel@menhir> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi, On Fri, 2012-04-27 at 17:42 +0900, Minchan Kim wrote: > Now there are several places to use __vmalloc with GFP_ATOMIC, > GFP_NOIO, GFP_NOFS but unfortunately __vmalloc calls map_vm_area > which calls alloc_pages with GFP_KERNEL to allocate page tables. > It means it's possible to happen deadlock. > I don't know why it doesn't have reported until now. > > Firstly, I tried passing gfp_t to lower functions to support __vmalloc > with such flags but other mm guys don't want and decided that > all of caller should be fixed. > > http://marc.info/?l=linux-kernel&m=133517143616544&w=2 > > To begin with, let's listen other's opinion whether they can fix it > by other approach without calling __vmalloc with such flags. > > So this patch adds warning to detect and to be fixed hopely. > I Cced related maintainers. > If I miss someone, please Cced them. > That seems ok to me. GFS2 only uses it as a back up in case the kmalloc call fails, and I suspect that we could easily eliminate it entirely since I doubt that it does actually ever fail in reality. If it were to fail then that is handled correctly anyway, Steve. > side-note: > I added WARN_ON instead of WARN_ONCE to detect all of callers > and each WARN_ON for each flag to detect to use any flag easily. > After we fix all of caller or reduce such caller, we can merge > a warning with WARN_ONCE. > > Cc: Neil Brown > Cc: Artem Bityutskiy > Cc: David Woodhouse > Cc: "Theodore Ts'o" > Cc: Adrian Hunter > Cc: Steven Whitehouse > Cc: "David S. Miller" > Cc: James Morris > Cc: Alexander Viro > Cc: Sage Weil > Signed-off-by: Minchan Kim > --- > mm/vmalloc.c | 9 +++++++++ > 1 file changed, 9 insertions(+) > > diff --git a/mm/vmalloc.c b/mm/vmalloc.c > index 94dff88..36beccb 100644 > --- a/mm/vmalloc.c > +++ b/mm/vmalloc.c > @@ -1700,6 +1700,15 @@ static void *__vmalloc_node(unsigned long size, unsigned long align, > gfp_t gfp_mask, pgprot_t prot, > int node, void *caller) > { > + /* > + * This function calls map_vm_area so that it allocates > + * page table with GFP_KERNEL so caller should avoid using > + * GFP_NOIO, GFP_NOFS and !__GFP_WAIT. > + */ > + WARN_ON(!(gfp_mask & __GFP_WAIT)); > + WARN_ON(!(gfp_mask & __GFP_IO)); > + WARN_ON(!(gfp_mask & __GFP_FS)); > + > return __vmalloc_node_range(size, align, VMALLOC_START, VMALLOC_END, > gfp_mask, prot, node, caller); > }