linux-kernel - Re: [PATCH v3 2/4] memblock tests: add verbose output to memblock tests

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:   Thu, 23 Jun 2022 08:11:58 -0500
From:   Mike Rapoport <rppt@...nel.org>
To:     Rebecca Mckeever <remckee0@...il.com>
Cc:     "Huang, Shaoqin" <shaoqin.huang@...el.com>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org, David Hildenbrand <david@...hat.com>
Subject: Re: [PATCH v3 2/4] memblock tests: add verbose output to memblock
 tests

On Thu, Jun 23, 2022 at 02:57:30AM -0500, Rebecca Mckeever wrote:
> On Thu, Jun 23, 2022 at 12:04:33AM -0500, Mike Rapoport wrote:
> > On Wed, Jun 22, 2022 at 11:56:30PM -0500, Rebecca Mckeever wrote:
> > > On Wed, Jun 22, 2022 at 11:05:27PM -0500, Mike Rapoport wrote:
> > > > On Thu, Jun 23, 2022 at 09:29:05AM +0800, Huang, Shaoqin wrote:
> > > > > 
> > > > > 
> > > > > On 6/23/2022 8:45 AM, Rebecca Mckeever wrote:
> > > > > > On Wed, Jun 22, 2022 at 06:32:04PM +0800, Huang, Shaoqin wrote:
> > > > > > > Just test it and everything works fine. And I think there are some thing can
> > > > > > > improve:
> > > > > > > 
> > > > > > >      The prefix_push() and prefix_pop() are used in so many functions and
> > > > > > > almost of them just put the prefix_push(__func__) begin in the head and the
> > > > > > > prefix_pop() in the end.
> > > > > > >      May be you can define some macro that when you output something and
> > > > > > > automatically push the __func__ as prefix. And when leave the function,
> > > > > > > automatically pop it. And only in some special place, you call it manually.
> > > > > > > 
> > > > > > Thank you for your review. I'm not sure how you would automatically push
> > > > > > __func__ since you have to be inside the function to access that
> > > > > > variable. Let me know if you have any suggestions. I am thinking about
> > > > > > adding another function in common.c that just calls test_pass() followed
> > > > > > by prefix_pop() since those are called together so often.
> > > > > 
> > > > > Just like:
> > > > > #define test_pass_macro()               \
> > > > >          do {                            \
> > > > >                  prefix_push(__func__);  \
> > > > >                  test_pass();            \
> > > > >                  prefix_pop();           \
> > > > >          } while (0)
> > > > 
> > > > This will not print the name of the failing test, e.g. instead of 
> > > > 
> > > > not ok 28 : memblock_alloc: alloc_bottom_up_disjoint_check: failed
> > > > 
> > > > with Rebecca's implementation it'll print
> > > > 
> > > > not ok 28 : memblock_alloc: failed
> > > > 
> > > Oh yeah, prefix_push() needs to be called before the asserts.
> > > 
> > > > How about
> > > > 
> > > > #define PREFIX_PUSH() 	prefix_push(__func__)?
> > > >  
> > > Good idea. What about 
> > > 
> > > #define TEST_PASS() do { \
> > > 	test_pass(); \
> > > 	prefix_pop(); \
> > > } while (0)
> > > 
> > > ? Or would it be better to make a function?
> > 
> > static inline function would be better.
> > 
> Would there be any advantage to defining a different version for each
> side of #ifdef VERBOSE? 

No, a single version will do. For !VERBOSE builds it will be optimized out
anyway.
 
> 
> Thanks,
> Rebecca

-- 
Sincerely yours,
Mike.