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, 12 Oct 2021 09:37:26 +0000
From:   SeongJae Park <sj@...nel.org>
To:     Xin Hao <xhao@...ux.alibaba.com>
Cc:     SeongJae Park <sj@...nel.org>, sjpark@...zon.de,
        akpm@...ux-foundation.org, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm/damon/dbgfs: add region_stat interface

On Tue, 12 Oct 2021 15:57:17 +0800 Xin Hao <xhao@...ux.alibaba.com> wrote:

> 
> 在 2021/10/12 下午3:11, SeongJae Park 写道:
> > Hello Xin, thank you for this patch!
> >
> > On Tue, 12 Oct 2021 13:49:48 +0800 Xin Hao <xhao@...ux.alibaba.com> wrote:
> >
> >> Using damon-dbgfs has brought great convenience to user-mode
> >> operation damon, but sometimes if i want to be able to view
> >> the division of task regions, nr_access values etc,but i found
> >> that it is impossible to view directly through the dbgfs interface,
> >> so there i add a interface "region_stat", it displays like this.
> >>
> >>   # cat region_stat
> >>   last_aggregation=120.87s
> >>   target_id=5148
> >>   nr_regions=10
> >>   400000-258c000(34352 KiB): 1
> >>   258c000-4719000(34356 KiB): 0
> >>   4719000-abbf000(103064 KiB): 0
> >>   abbf000-c4d4000(25684 KiB): 11
> >>   c4d4000-ff5c000(59936 KiB): 15
> >>   ff5c000-152f9000(85620 KiB): 20
> >>   152f9000-1599e000(6804 KiB): 10
> >>   1599e000-19573000(61268 KiB): 0
> >>   19573000-1f92c000(102116 KiB): 0
> >>   1f92c000-22a4c000(50304 KiB): 0
> > I think similar information could also be collected via the 'damon_aggregated'
> > tracepoint[1], which is merged in the mainline, or 'DAMOS_STAT'[2], which is
> > merged in -mm.  The recording feature[3] could also be used, though it would
> > take some time before it is merged in the mainline.  Have you considered using
> > those but found some problem?
> 
> Yes, i know we can use damon_aggregated tracepoint, but i think, add a 
> "region_stat" will be more
> 
> convenient and intuitive, especially when we use damon-dbgfs interface.

Thanks for the answer.

For more conveniend and intuitive interfaces, I recommend DAMON user-sapce
tool[1] rather than the debugfs interface.  You can also implement your own
user-space tool on top of the debugfs interface.  For example, the reference
implementation[1] implements 'record' feature on top of the tracepoint.  The
feature can provide below information, which is quite similar to what you want.
For more information on this, please refer to the document[2].

    # damo record $(pidof $workload)
    # damo report raw
    base_time_absolute: 8 m 59.809 s

    monitoring_start:                0 ns
    monitoring_end:            104.599 ms
    monitoring_duration:       104.599 ms
    target_id: 18446623438842320000
    nr_regions: 3
    563ebaa00000-563ebc99e000(  31.617 MiB):        1
    7f938d7e1000-7f938ddfc000(   6.105 MiB):        0
    7fff66b0a000-7fff66bb2000( 672.000 KiB):        0

    monitoring_start:          104.599 ms
    monitoring_end:            208.590 ms
    monitoring_duration:       103.991 ms
    target_id: 18446623438842320000
    nr_regions: 4
    563ebaa00000-563ebc99e000(  31.617 MiB):        1
    7f938d7e1000-7f938d9b5000(   1.828 MiB):        0
    7f938d9b5000-7f938ddfc000(   4.277 MiB):        0
    7fff66b0a000-7fff66bb2000( 672.000 KiB):        5

If this is still not so useful for you, could you please elaborate more on your
detailed requirements and use cases?

[1] https://github.com/awslabs/damo
[2] https://github.com/awslabs/damo/blob/next/USAGE.md#raw


Thanks,
SJ

> 
> >
> > [1] https://git.kernel.org/torvalds/c/2fcb93629ad8
> > [2] https://lore.kernel.org/linux-mm/20211001125604.29660-6-sj@kernel.org/
> > [3] https://lore.kernel.org/linux-mm/20211008094509.16179-1-sj@kernel.org/
> >
> >
> > Thanks,
> > SJ
> >
> > [...]
> 
> -- 
> Best Regards!
> Xin Hao

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ