linux-kernel - Re: [PATCH v5 1/2] mm: Allow the page cache to allocate large pages

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190903191819.GD14028@dhcp22.suse.cz>
Date:   Tue, 3 Sep 2019 21:18:19 +0200
From:   Michal Hocko <mhocko@...nel.org>
To:     Matthew Wilcox <willy@...radead.org>
Cc:     William Kucharski <william.kucharski@...cle.com>,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        linux-fsdevel@...r.kernel.org,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Song Liu <songliubraving@...com>,
        Bob Kasten <robert.a.kasten@...el.com>,
        Mike Kravetz <mike.kravetz@...cle.com>,
        Chad Mynhier <chad.mynhier@...cle.com>,
        "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
        Johannes Weiner <jweiner@...com>
Subject: Re: [PATCH v5 1/2] mm: Allow the page cache to allocate large pages

On Tue 03-09-19 09:28:31, Matthew Wilcox wrote:
> On Tue, Sep 03, 2019 at 02:19:52PM +0200, Michal Hocko wrote:
> > On Tue 03-09-19 05:11:55, Matthew Wilcox wrote:
> > > On Tue, Sep 03, 2019 at 01:57:48PM +0200, Michal Hocko wrote:
> > > > On Mon 02-09-19 03:23:40, William Kucharski wrote:
> > > > > Add an 'order' argument to __page_cache_alloc() and
> > > > > do_read_cache_page(). Ensure the allocated pages are compound pages.
> > > > 
> > > > Why do we need to touch all the existing callers and change them to use
> > > > order 0 when none is actually converted to a different order? This just
> > > > seem to add a lot of code churn without a good reason. If anything I
> > > > would simply add __page_cache_alloc_order and make __page_cache_alloc
> > > > call it with order 0 argument.
> > > 
> > > Patch 2/2 uses a non-zero order.
> > 
> > It is a new caller and it can use a new function right?
> > 
> > > I agree it's a lot of churn without
> > > good reason; that's why I tried to add GFP_ORDER flags a few months ago.
> > > Unfortunately, you didn't like that approach either.
> > 
> > Is there any future plan that all/most __page_cache_alloc will get a
> > non-zero order argument?
> 
> I'm not sure about "most".  It will certainly become more common, as
> far as I can tell.

I would personally still go with  __page_cache_alloc_order way, but this
is up to you and other fs people what suits best. I was just surprised
to see a lot of code churn when it was not really used in the second
patch. That's why I brought it up. 

> > > > Also is it so much to ask callers to provide __GFP_COMP explicitly?
> > > 
> > > Yes, it's an unreasonable burden on the callers.
> > 
> > Care to exaplain why? __GFP_COMP tends to be used in the kernel quite
> > extensively.
> 
> Most of the places which call this function get their gfp_t from
> mapping->gfp_mask.  If we only want to allocate a single page, we
> must not set __GFP_COMP.  If we want to allocate a large page, we must
> set __GFP_COMP.  Rather than require individual filesystems to concern
> themselves with this wart of the GFP interface, we can solve it in the
> page cache.

Fair enough.

-- 
Michal Hocko
SUSE Labs