linux-kernel - Re: [PATCH 2/2] mm/filemap: fix mapping_seek_hole

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]

Date:   Wed, 21 Apr 2021 22:55:37 -0700 (PDT)
From:   Hugh Dickins <hughd@...gle.com>
To:     Matthew Wilcox <willy@...radead.org>
cc:     Hugh Dickins <hughd@...gle.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        William Kucharski <william.kucharski@...cle.com>,
        Christoph Hellwig <hch@....de>, Jan Kara <jack@...e.cz>,
        Dave Chinner <dchinner@...hat.com>,
        Johannes Weiner <hannes@...xchg.org>,
        "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
        Yang Shi <yang.shi@...ux.alibaba.com>,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-mm@...ck.org
Subject: Re: [PATCH 2/2] mm/filemap: fix mapping_seek_hole_data on THP &
 32-bit

On Thu, 22 Apr 2021, Matthew Wilcox wrote:
> On Wed, Apr 21, 2021 at 05:39:14PM -0700, Hugh Dickins wrote:
> > No problem on 64-bit without huge pages, but xfstests generic/285
> > and other SEEK_HOLE/SEEK_DATA tests have regressed on huge tmpfs,
> > and on 32-bit architectures, with the new mapping_seek_hole_data().
> > Several different bugs turned out to need fixing.
> > 
> > u64 casts added to stop unfortunate sign-extension when shifting
> > (and let's use shifts throughout, rather than mixed with * and /).
> 
> That confuses me.  loff_t is a signed long long, but it can't be negative
> (... right?)  So how does casting it to an u64 before dividing by
> PAGE_SIZE help?

That is a good question. Sprinkling u64s was the first thing I tried,
and I'd swear that it made a good difference at the time; but perhaps
that was all down to just the one on xas.xa_index << PAGE_SHIFT. Or
is it possible that one of the other bugs led to a negative loff_t,
and the casts got better behaviour out of that? Doubtful.

What I certainly recall from yesterday was leaving out one (which?)
of the casts as unnecessary, and wasting quite a bit of time until I
put it back in. Did I really choose precisely the only one necessary?

Taking most of them out did give me good quick runs just now: I'll
go over them again and try full runs on all machines. You'll think me
crazy, but yesterday's experience leaves me reluctant to change without
full testing - but agree it's not good to leave ignorant magic in.

> 
> > Use round_up() when advancing pos, to stop assuming that pos was
> > already THP-aligned when advancing it by THP-size.  (But I believe
> > this use of round_up() assumes that any THP must be THP-aligned:
> > true while tmpfs enforces that alignment, and is the only fs with
> > FS_THP_SUPPORT; but might need to be generalized in the future?
> > If I try to generalize it right now, I'm sure to get it wrong!)
> 
> No generalisation needed in future.  Folios must be naturally aligned
> within a file.

Thanks for the info: I did search around in your various patch series
from last October, and failed to find a decider there: I imagined that
when you started on compound pages for more efficient I/O, there would
be no necessity to align them (whereas huge pmd mappings of shared
files make the alignment important). Anyway, assuming natural alignment
is easiest - but it's remarkable how few places need to rely on it.

> 
> > @@ -2681,7 +2681,8 @@ loff_t mapping_seek_hole_data(struct add
> >  
> >  	rcu_read_lock();
> >  	while ((page = find_get_entry(&xas, max, XA_PRESENT))) {
> > -		loff_t pos = xas.xa_index * PAGE_SIZE;
> > +		loff_t pos = (u64)xas.xa_index << PAGE_SHIFT;
> > +		unsigned int seek_size;
> 
> I've been preferring size_t for 'number of bytes in a page' because
> I'm sure somebody is going to want a page larger than 2GB in the next
> ten years.

Ah, there I was simply following what the author of seek_page_size()
had chosen, and I think that's the right thing to do in today's tree:
let's see who that author was... hmm, someone called Matthew Wilcox :)