Date: Mon, 13 Feb 2012 13:55:31 -0800 (PST)
From: David Rientjes <rientjes@google.com>
To: Dan Smith <danms@us.ibm.com>
cc: akpm@linux-foundation.org, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] Ensure that walk_page_range()'s start and end are
 page-aligned
In-Reply-To: <87zkcm23az.fsf@caffeine.danplanet.com>
Message-ID: <alpine.DEB.2.00.1202131350500.17296@chino.kir.corp.google.com>
References: <1328902796-30389-1-git-send-email-danms@us.ibm.com> <alpine.DEB.2.00.1202130211400.4324@chino.kir.corp.google.com> <87zkcm23az.fsf@caffeine.danplanet.com>
User-Agent: Alpine 2.00 (DEB 1167 2008-08-23)
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1143
Lines: 24

On Mon, 13 Feb 2012, Dan Smith wrote:

> DR> It doesn't "ensure" anything without CONFIG_DEBUG_VM enabled, which
> DR> isn't the default.
> 
> Are you proposing a change in verbiage or a stronger check? A
> VM_BUG_ON() seemed on par with other checks, such as the one in
> get_user_pages_fast().
> 

That's not a precedent, there's a big difference between the performance 
of gup_fast(), where we can't spare an additional compare and branch, and 
walk_page_range().  VM_BUG_ON() is typically used in situations where a 
debug kernel has been built, including CONFIG_DEBUG_VM, and the check 
helps to isolate a problem that would be otherwise difficult to find.  If 
that fits the criteria, fine, but it doesn't "ensure" walk_page_range() 
always has start and end addresses that are page aligned, so the changelog 
needs to be modified to describe why an error in this path wouldn't be 
evident.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/