Date: Tue, 18 Oct 2016 12:38:37 +0200
From: Christoph Hellwig <hch@lst.de>
To: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Christoph Hellwig <hch@lst.de>, akpm@linux-foundation.org,
        joelaf@google.com, jszhang@marvell.com, joaodias@google.com,
        linux-mm@kvack.org, linux-rt-users@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/6] mm: mark all calls into the vmalloc subsystem as
        potentially sleeping
Message-ID: <20161018103837.GA8945@lst.de>
References: <1476773771-11470-1-git-send-email-hch@lst.de> <1476773771-11470-3-git-send-email-hch@lst.de> <20161018103359.GM29072@nuc-i3427.alporthouse.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161018103359.GM29072@nuc-i3427.alporthouse.com>
User-Agent: Mutt/1.5.17 (2007-11-01)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 617
Lines: 13

On Tue, Oct 18, 2016 at 11:33:59AM +0100, Chris Wilson wrote:
> On Tue, Oct 18, 2016 at 08:56:07AM +0200, Christoph Hellwig wrote:
> > This is how everyone seems to already use them, but let's make that
> > explicit.
> 
> mm/page_alloc.c: alloc_large_system_hash() is perhaps the exception to
> the rule.

While alloc_large_system_hash passes GFP_ATOMIC it still is called
from context where it can sleep - I think it just abuses GFP_ATOMIC
so that it gets an "early" failure.  For which GFP_ATOMIC isn't
exactly a good choice as it dips into additional reserves, GFP_NOWAIT
would have probably been a better choice.