linux-kernel - Re: [patch] mm, thp: add new background defrag option

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:   Tue, 10 Jan 2017 15:52:31 -0800 (PST)
From:   David Rientjes <rientjes@...gle.com>
To:     Vlastimil Babka <vbabka@...e.cz>
cc:     Hugh Dickins <hughd@...gle.com>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Michal Hocko <mhocko@...nel.org>,
        Jonathan Corbet <corbet@....net>,
        "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [patch] mm, thp: add new background defrag option

On Tue, 10 Jan 2017, Vlastimil Babka wrote:

> > I get very confused by the /sys/kernel/mm/transparent_hugepage/defrag
> > versus enabled flags, and this may be a terrible, even more confusing,
> > idea: but I've been surprised and sad to see defrag with a "defer"
> > option, but poor enabled without one; and it has crossed my mind that
> > perhaps the peculiar "madvise+defer" syntax in defrag might rather be
> > handled by "madvise" in defrag with "defer" in enabled?  Or something
> > like that: 4 x 4 possibilities instead of 5 x 3.
> 
> But would all the possibilities make sense? For example, if I saw
> "defer" in enabled, my first expectation would be that it would only use
> khugepaged, and no THP page faults at all - possibly including madvised
> regions.
> 

And this is why I've tried to minimize the config requirements and respect 
userspace decisions to do MADV_HUGEPAGE, MADV_NOHUGEPAGE, or set/clear 
PR_SET_THP_DISABLE because all these system-wide options combined with 
userspace syscalls truly seems unmaintainable and waay too confusing to 
correctly describe.  Owell, I am fine with introducing 
yet-another-defrag-mode if it lets us move in a direction that supports 
our usecase.