linux-kernel - Re: [PATCH v5 10/10] mmap locking API: rename mmap_sem to mmap

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Wed, 22 Apr 2020 18:59:17 -0700
From:   Matthew Wilcox <willy@...radead.org>
To:     Michel Lespinasse <walken@...gle.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        linux-mm <linux-mm@...ck.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Laurent Dufour <ldufour@...ux.ibm.com>,
        Vlastimil Babka <vbabka@...e.cz>,
        Liam Howlett <Liam.Howlett@...cle.com>,
        Jerome Glisse <jglisse@...hat.com>,
        Davidlohr Bueso <dave@...olabs.net>,
        David Rientjes <rientjes@...gle.com>,
        Hugh Dickins <hughd@...gle.com>, Ying Han <yinghan@...gle.com>,
        Jason Gunthorpe <jgg@...pe.ca>,
        Daniel Jordan <daniel.m.jordan@...cle.com>
Subject: Re: [PATCH v5 10/10] mmap locking API: rename mmap_sem to mmap_lock

On Wed, Apr 22, 2020 at 03:54:32PM -0700, Michel Lespinasse wrote:
> On Tue, Apr 21, 2020 at 6:58 PM Matthew Wilcox <willy@...radead.org> wrote:
> >
> > On Tue, Apr 21, 2020 at 05:14:22PM -0700, Michel Lespinasse wrote:
> > > Rename the mmap_sem field to mmap_lock. Any new uses of this lock
> >
> > Shouldn't some of these be folded into the previous patch?
> 
> So, I didn't do it because previous patch only handled rwsem_is_locked
> call sites. I leaned towards adding as few new API functions as
> possible until we figure out exactly what is required.
> 
> That said, I agree it seems reasonable to split mmap_assert_locked()
> into mmap_assert_read_locked() and mmap_assert_write_locked(), and
> convert the lockdep asserts to use these instead.

Just add mmap_assert_write_locked() -- some of these places can be called
with the rwsem held for either read or write; it doesn't matter which.
Others need it held for write.  There aren't any places (that I'm aware
of) that need to assert that it's held for read, and not held for write.

> I'm not sure we need to do it right away though; we are at least not
> losing any test coverage with the existing version of the patchset...

It seems like a better way to remove users of the term 'mmap_sem' than
just converting them to use the new 'mmap_lock'.