linux-kernel - Re: [PACTH v1] mm, proc: Implement /proc/<pid>/totmaps

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] [day] [month] [year] [list]
Message-ID: <20160813123952.GA30408@pc.thejh.net>
Date:	Sat, 13 Aug 2016 14:39:53 +0200
From:	Jann Horn <jann@...jh.net>
To:	Robert Foss <robert.foss@...labora.com>
Cc:	Sonny Rao <sonnyrao@...omium.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Kees Cook <keescook@...omium.org>,
	Al Viro <viro@...iv.linux.org.uk>,
	Cyrill Gorcunov <gorcunov@...nvz.org>,
	John Stultz <john.stultz@...aro.org>,
	Robin Humble <plaguedbypenguins@...il.com>,
	Mateusz Guzik <mguzik@...hat.com>,
	Alexey Dobriyan <adobriyan@...il.com>,
	Janis Danisevskis <jdanis@...gle.com>, calvinowens@...com,
	Michal Hocko <mhocko@...e.com>,
	Konstantin Khlebnikov <koct9i@...il.com>,
	Vlastimil Babka <vbabka@...e.cz>,
	Naoya Horiguchi <n-horiguchi@...jp.nec.com>,
	"Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
	ldufour@...ux.vnet.ibm.com, Johannes Weiner <hannes@...xchg.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Ben Zhang <benzh@...omium.org>,
	Bryan Freed <bfreed@...omium.org>,
	Filipe Brandenburger <filbranden@...omium.org>,
	linux-api@...r.kernel.org
Subject: Re: [PACTH v1] mm, proc: Implement /proc/<pid>/totmaps

On Fri, Aug 12, 2016 at 12:28:11PM -0400, Robert Foss wrote:
> 
> 
> On 2016-08-10 02:05 PM, Jann Horn wrote:
> >On Wed, Aug 10, 2016 at 10:45:51AM -0700, Sonny Rao wrote:
> >>On Wed, Aug 10, 2016 at 10:37 AM, Jann Horn <jann@...jh.net> wrote:
> >>>On Wed, Aug 10, 2016 at 10:23:53AM -0700, Sonny Rao wrote:
> >>>>On Tue, Aug 9, 2016 at 2:01 PM, Robert Foss <robert.foss@...labora.com> wrote:
> >>>>>
> >>>>>
> >>>>>On 2016-08-09 03:24 PM, Jann Horn wrote:
> >>>>>>
> >>>>>>On Tue, Aug 09, 2016 at 12:05:43PM -0400, robert.foss@...labora.com wrote:
> >>>>>>>
> >>>>>>>From: Sonny Rao <sonnyrao@...omium.org>
> >>>>>>>
> >>>>>>>This is based on earlier work by Thiago Goncales. It implements a new
> >>>>>>>per process proc file which summarizes the contents of the smaps file
> >>>>>>>but doesn't display any addresses.  It gives more detailed information
> >>>>>>>than statm like the PSS (proprotional set size).  It differs from the
> >>>>>>>original implementation in that it doesn't use the full blown set of
> >>>>>>>seq operations, uses a different termination condition, and doesn't
> >>>>>>>displayed "Locked" as that was broken on the original implemenation.
> >>>>>>>
> >>>>>>>This new proc file provides information faster than parsing the
> >>>>>>>potentially
> >>>>>>>huge smaps file.
> >>>>>>>
> >>>>>>>Signed-off-by: Sonny Rao <sonnyrao@...omium.org>
> >>>>>>>
> >>>>>>>Tested-by: Robert Foss <robert.foss@...labora.com>
> >>>>>>>Signed-off-by: Robert Foss <robert.foss@...labora.com>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>>+static int totmaps_proc_show(struct seq_file *m, void *data)
> >>>>>>>+{
> >>>>>>>+       struct proc_maps_private *priv = m->private;
> >>>>>>>+       struct mm_struct *mm;
> >>>>>>>+       struct vm_area_struct *vma;
> >>>>>>>+       struct mem_size_stats *mss_sum = priv->mss;
> >>>>>>>+
> >>>>>>>+       /* reference to priv->task already taken */
> >>>>>>>+       /* but need to get the mm here because */
> >>>>>>>+       /* task could be in the process of exiting */
> >>>>>>
> >>>>>>
> >>>>>>Can you please elaborate on this? My understanding here is that you
> >>>>>>intend for the caller to be able to repeatedly read the same totmaps
> >>>>>>file with pread() and still see updated information after the target
> >>>>>>process has called execve() and be able to detect process death
> >>>>>>(instead of simply seeing stale values). Is that accurate?
> >>>>>>
> >>>>>>I would prefer it if you could grab a reference to the mm_struct
> >>>>>>directly at open time.
> >>>>>
> >>>>>
> >>>>>Sonny, do you know more about the above comment?
> >>>>
> >>>>I think right now the file gets re-opened every time, but the mode
> >>>>where the file is opened once and repeatedly read is interesting
> >>>>because it avoids having to open the file again and again.
> >>>>
> >>>>I guess you could end up with a wierd situation where you don't read
> >>>>the entire contents of the file in open call to read() and you might
> >>>>get inconsistent data across the different statistics?
> >>>
> >>>If the file is read in two chunks, totmaps_proc_show is only called
> >>>once. The patch specifies seq_read as read handler. Have a look at its
> >>>definition. As long as you don't read from the same seq file in
> >>>parallel or seek around in it, simple sequential reads will not
> >>>re-invoke the show() method for data that has already been formatted.
> >>>For partially consumed data, the kernel buffers the rest until someone
> >>>reads it or seeks to another offset.
> >>
> >>Ok that's good.  If the consumer were using pread() though, would that
> >>look like a seek?
> >
> >Only if the consumer uses pread() with an offset that is not the same as
> >the end offset of the previous read.
> >
> >So if you tried to use the same file from multiple threads in parallel,
> >you might still have issues, but as long as you don't do that, it should
> >be fine.
> >
> >I guess it might make sense to document this behavior somewhere - maybe
> >the proc.5 manpage?
> >
> 
> I'll add a note about limitations for parallel read. The overall
> documentation for this feature should live in the proc.5 manpage as well?

Yes, I think so.

+Cc linux-api@...r.kernel.org

Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)