linux-kernel - Re: [PATCH] mm: get rid of odd jump label in find_mergeable_anon

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: <b4d8cbf4-c553-2594-9d92-6d7e58eb246c@nvidia.com>
Date:   Fri, 15 Nov 2019 15:38:30 -0800
From:   John Hubbard <jhubbard@...dia.com>
To:     David Hildenbrand <david@...hat.com>,
        linmiaohe <linmiaohe@...wei.com>, <akpm@...ux-foundation.org>,
        <richardw.yang@...ux.intel.com>, <sfr@...b.auug.org.au>,
        <rppt@...ux.ibm.com>, <jannh@...gle.com>, <steve.capper@....com>,
        <catalin.marinas@....com>, <aarcange@...hat.com>,
        <chenjianhong2@...wei.com>, <walken@...gle.com>,
        <dave.hansen@...ux.intel.com>, <tiny.windzz@...il.com>
CC:     <linux-mm@...ck.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] mm: get rid of odd jump label in find_mergeable_anon_vma

On 11/15/19 4:58 AM, David Hildenbrand wrote:
> On 15.11.19 07:36, linmiaohe wrote:
>> From: Miaohe Lin <linmiaohe@...wei.com>
> 
> I'm pro removing unnecessary jump labels.

Thank you, simpler code is good--all other things being equal. 
The  tradeoff is, as Qian points out: code churn and risk in critical 
code paths.

In this case, I'd claim it's OK to improve this one, because we
can likely get it right by visual inspection, and the pre-existing
code is quite poor. And being in the kernel does not necessarily give
weak code a free pass to remain there and incur maintenance and annoyance
costs until the end of time. :)

However, please spend equal time when you write your commit descriptions,
because that's also very important. Commit logs should also be clear!

> 
> Subject: "mm: get rid of jump labels in find_mergeable_anon_vma()"
> 
>>
>> The odd jump label try_prev and none is not really need

s/need/needed/

> 
> s/odd jump label/jump labels/
> 
> s/is/are/
> 
>> in func find_mergeable_anon_vma, eliminate them to
>> improve readability.
>>
>> Signed-off-by: Miaohe Lin <linmiaohe@...wei.com>
>> ---
>>  mm/mmap.c | 18 +++++++-----------
>>  1 file changed, 7 insertions(+), 11 deletions(-)
>>
>> diff --git a/mm/mmap.c b/mm/mmap.c
>> index 4d4db76a07da..ab980d468a10 100644
>> --- a/mm/mmap.c
>> +++ b/mm/mmap.c
>> @@ -1276,25 +1276,21 @@ static struct anon_vma *reusable_anon_vma(struct vm_area_struct *old, struct vm_
>>   */
>>  struct anon_vma *find_mergeable_anon_vma(struct vm_area_struct *vma)
>>  {
>> -	struct anon_vma *anon_vma;
>> +	struct anon_vma *anon_vma = NULL;
>>  	struct vm_area_struct *near;
>>  
>>  	near = vma->vm_next;
>> -	if (!near)
>> -		goto try_prev;
>> -
>> -	anon_vma = reusable_anon_vma(near, vma, near);
>> +	if (near)
>> +		anon_vma = reusable_anon_vma(near, vma, near);>  	if (anon_vma)
>>  		return anon_vma;
> 
> Let me suggest the following instead:
> 
> /* Try next first */
> near = vma->vm_next;
> if (near) {
> 	anon_vma = reusable_anon_vma(near, vma, near);
> 	if (anon_vma)
> 		return anon_vma;
> }
> /* Try prev next */
> near = vma->vm_prev;
> if (near) {
> 	anon_vma = reusable_anon_vma(near, vma, near);
> 	if (anon_vma)
> 		return anon_vma;
> }
> 

Actually, it can be further simplified, because you don't need
that last "if" statement, because you're returning NULL after this.

So just return anon_vma there. (And adjust the comment block at the 
end, so that it's clear that anon_vma might be null.)


thanks,
-- 
John Hubbard
NVIDIA

>> -try_prev:
>> -	near = vma->vm_prev;
>> -	if (!near)
>> -		goto none;
>>  
>> -	anon_vma = reusable_anon_vma(near, near, vma);
>> +	near = vma->vm_prev;
>> +	if (near)
>> +		anon_vma = reusable_anon_vma(near, near, vma);
>>  	if (anon_vma)
>>  		return anon_vma;
>> -none:
>> +
>>  	/*
>>  	 * There's no absolute need to look only at touching neighbours:
>>  	 * we could search further afield for "compatible" anon_vmas.
>>
> 
>