linux-kernel - Re: [PATCH] lib: test module for find_*

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: <20171114100730.fhqvupfa5oxmutfl@yury-thinkpad>
Date:   Tue, 14 Nov 2017 13:07:30 +0300
From:   Yury Norov <ynorov@...iumnetworks.com>
To:     Alexey Dobriyan <adobriyan@...il.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        linux-kernel@...r.kernel.org, Clement Courbet <courbet@...gle.com>,
        Matthew Wilcox <mawilcox@...rosoft.com>,
        Rasmus Villemoes <linux@...musvillemoes.dk>
Subject: Re: [PATCH] lib: test module for find_*_bit() functions

Hi Alexey, Andrew,

Thanks for comments.

On Fri, Nov 10, 2017 at 12:45:18PM +0200, Alexey Dobriyan wrote:
> On 11/10/17, Andrew Morton <akpm@...ux-foundation.org> wrote:
> > On Thu,  9 Nov 2017 17:07:14 +0300 Yury Norov <ynorov@...iumnetworks.com>
> > wrote:
> >
> >> find_bit functions are widely used in the kernel, including hot paths.
> >> This module tests performance of that functions in 2 typical scenarios:
> >> randomly filled bitmap with relatively equal distribution of set and
> >> cleared bits, and sparse bitmap which has 1 set bit for 500 cleared bits.
> >>
> >> ...
> >>
> >> +config TEST_FIND_BIT
> >
> > Well.  It doesn't actually "test" the code.  It measures its performance ;)
> 
> Yes!
> 
> Yyra, you can grab CONFIG_BENCHMARK_* namespace :-)
 
There's no CONFIG_BENCHMARK_* namespace actually. The 'CONFIG_*_BENCHMARK' is
referenced only 3 times in linux sources - CONFIG_RING_BUFFER_BENCHMARK,
CONFIG_TRACEPOINT_BENCHMARK and CONFIG_GUP_BENCHMARK, so I simply didn't know
about it. Some other tests like lib/rbtree_test.c also measure performance and
use TEST namespace, but if you think it's better, I don't object to change it.
 
> Another thing:
> 
> > +
> > +       return 0;
> > +}
> > +module_init(find_bit_test);
> > +
> > +static void __exit test_find_bit_cleanup(void)
> > +{
> > +}
> > +module_exit(test_find_bit_cleanup);
> 
> module exit hook is entirely unnecessary as you can return -E from init hook.
> See lib/test-kstrtox.c

Ack. 

I thought to send v3, but the patch is already in next tree, so I'll
send fix in separated patch. OK?

Yury