linux-kernel - Re: [PATCH] mm: remove unnecessary __get_user_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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <20161025234631.GA5946@lucifer>
Date:   Wed, 26 Oct 2016 00:46:31 +0100
From:   Lorenzo Stoakes <lstoakes@...il.com>
To:     linux-mm@...ck.org
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Michal Hocko <mhocko@...nel.org>, Jan Kara <jack@...e.cz>,
        Hugh Dickins <hughd@...gle.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Rik van Riel <riel@...hat.com>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Radim Krčmář <rkrcmar@...hat.com>,
        kvm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm: remove unnecessary __get_user_pages_unlocked() calls

The holdout for unexporting __get_user_pages_unlocked() is its invocation in
mm/process_vm_access.c: process_vm_rw_single_vec(), as this definitely _does_
seem to invoke VM_FAULT_RETRY behaviour which get_user_pages_remote() will not
trigger if we were to replace it with the latter.

I'm not sure how to proceed in this case - get_user_pages_remote() invocations
assume mmap_sem is held so can't offer VM_FAULT_RETRY behaviour as the lock
can't be assumed to be safe to release, and get_user_pages_unlocked() assumes
tsk, mm are set to current, current->mm respectively so we can't use that here
either.

Is it important to retain VM_FAULT_RETRY behaviour here, does it matter? If it
isn't so important then we can just go ahead and replace with
get_user_pages_remote() and unexport away.

Of course the whole idea of unexporting __get_user_pages_unlocked() might be
bogus so let me know in that case also :)