From mboxrd@z Thu Jan  1 00:00:00 1970
From: Corey Minyard <cminyard@mvista.com>
Date: Mon, 17 Jan 2005 22:41:15 +0000
Subject: [KJ] Re: [PATCH 9/21] char/ipmi_si_intf: replace schedule_timeout()
Message-Id: <41EC3F0B.5090605@mvista.com>
List-Id: <kernel-janitors.vger.kernel.org>
References: <41EC332A.8020603@mvista.com>
In-Reply-To: <41EC332A.8020603@mvista.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: kernel-janitors@vger.kernel.org

Ok, I was unaware of the difference between msleep and mdelay (I was 
actually reading "mdelay").

Actually, in the code I want to wait as small a time as possible without 
spinning.  The operations being waited for generally happen in ~500us, 
which is far too long to spin, but 10ms would be non-optimal if a faster 
increment was available.  So I think it is still best as it is.

Thanks,

-Corey

Nishanth Aravamudan wrote:

>On Mon, Jan 17, 2005 at 03:53:12PM -0600, Corey Minyard wrote:
>  
>
>>On my previous question, to be more clear, why was this change 
>>necessary?  In this particular instance, it doesn't matter how long it 
>>sleeps, it just can't spin waiting for something to happen.  msleep() 
>>spins, right?  That would be very bad in this code.  Even if not, the 
>>particular timing is not important.
>>    
>>
>
>msleep() does not spin. The name itself is intended to indicate this. Delays
>(mdelay(), udelay()) spin by looping tightly. Sleeps (msleep(), ssleep(), 
>msleep_interruptible()) use schedule_timeout() to give up the CPU for a certain
>amount of time.
>
>Given this last comment, there is a clear benefit (IMO) to use actual time
>units for sleeps (msleep() and co. use milliseconds or seconds) as opposed to
>jiffy-relative units (as in schedule_timeout() where you are requesting a delay
>in jiffies, which varies from arch to arch).
>
>This becomes very important with dynamic HZ, for instance. Clearly if HZ
>changes, then the delay caused by schedule_timeout(1) will change, which is not
>necessarily desired and may be confusing the user. A time specified in real
>time units will be adjusted with the change in HZ.
>
>If it is not important how long the driver sleeps, then I believe msleep()
>should be preferred in effectively all cases (depending on whether wait-queue
>events or signals may be important early triggers, of course). Milliseconds
>indicate a clear delay, independent of HZ's value. Jiffy delays have a clear
>reliance on the value of HZ.
>
>I am open to other suggestions, but I think these are good reasons. Another
>basic one is that it will lead to consistency across the board. HZ is not a
>measure of time, it should not be used as a measure of time. Instead,
>milliseconds should be used and, when/if the facility has been added,
>microseconds and nanoseconds.
>
>Thanks,
>Nish
>  
>

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
http://lists.osdl.org/mailman/listinfo/kernel-janitors