netdev - Re: [PATCH 2/5] accounting unit and variable

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: <473B852B.3010107@redhat.com>
Date:	Wed, 14 Nov 2007 18:30:51 -0500
From:	Hideo AOKI <haoki@...hat.com>
To:	David Miller <davem@...emloft.net>, herbert@...dor.apana.org.au
CC:	Hideo AOKI <haoki@...hat.com>, netdev@...r.kernel.org,
	satoshi.oshima.fk@...achi.com, andi@...stfloor.org,
	shemminger@...ux-foundation.org, johnpol@....mipt.ru,
	yoshfuji@...ux-ipv6.org, yumiko.sugita.yf@...achi.com
Subject: Re: [PATCH 2/5] accounting unit and variable

Hideo AOKI wrote:
> David Miller wrote:
>> From: Hideo AOKI <haoki@...hat.com>
>> Date: Tue, 13 Nov 2007 22:27:13 -0500
>>
>>> Herbert Xu wrote:
>>>> On Mon, Oct 29, 2007 at 05:23:10PM -0400, Hideo AOKI wrote:
>>>>>  
>>>>> +#define SK_DATAGRAM_MEM_QUANTUM ((int)PAGE_SIZE)
>>>>> +
>>>>> +static inline int sk_datagram_pages(int amt)
>>>>> +{
>>>>> +    return DIV_ROUND_UP(amt, SK_DATAGRAM_MEM_QUANTUM);
>>>>> +}
>>>> Does this really have to be int? Unsigned would let the compiler
>>>> optimise this to a simple shift.
>>> Thank you for the comment.
>>>
>>> This inline function is used to calculate the first argument of 
>>> atomic_add()
>>> and atomic_sub(). Since the argument is int, I believe that using int is
>>> better than using unsigned int.
>>
>> If you know the values will always be positive, as you will know here,
>> it is OK to us unsigned int here and avoids the unacceptable expensive
>> divide instruction.
>>
>> Please fix this.
> 
> Thanks for your comments. I finally understood.
> 
> I'll fix it and resubmit the patch as soon as possible.

Let me propose the following code to use a shift instruction instead
of a divide instruction. I confirmed that the code could remove a
divide instruction, however, I would like to have comment on this
implementation.

+#define SK_DATAGRAM_MEM_QUANTUM ((unsigned int)PAGE_SIZE)
+
+static inline int sk_datagram_pages(int amt)
+{
+	/* Cast to unsigned as an optimization, since amt is always positive. */
+	return DIV_ROUND_UP((unsigned int)amt, SK_DATAGRAM_MEM_QUANTUM);
+}
+

Please let me know if there is proper coding style.

I'm re-testing the whole patch set right now. If there is no problem,
I'll resubmit new patch set, which includes this fix, tomorrow.

Many thanks,
Hideo

-- 
Hitachi Computer Products (America) Inc.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html