lkml.org 
[lkml]   [1998]   [Sep]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Today Linus redesigns the networking driver interface (was Re: tulip driver in ...)
On Wed, 23 Sep 1998, Richard Gooch wrote:

>Andrea Arcangeli writes:
>> On Wed, 23 Sep 1998, Richard Gooch wrote:
>>
>> >BTW: if you're idling, that doesn't mean that BH's are called before
>> >the next timer tick. AFAIK do_bottom_half() is only called on return
>> >from syscall/interrupt and when rescheduling.
>> >Waiting for the next timer tick would be horrible for latency,
>> >though.
>>
>> If you queue the handler in the timer_bh OK, you are sure that the
>> bh will be marked (in the worst case after 1/HZ sec). All other bh
>> has not timeout so if you don' t mark it and there are not other
>> code that mark it _incidentally_ the bh handler will never run. This
>> is what I understand from the code.
>
>My point is that 1/HZ is a *long* time. 10 ms for packet processing is
>awful.

We _was_ talking about no timeout for a bh hander (Linus' s email) -> so
we was talking about bh handers not queued in the timer_bh (what I tried
to say in my last email) -> so the point of the last 2/3 email is sure not
the frequency of the timer interrupt.

And btw you are the one worried by sheduling latency so you really don' t
want an high HZ.

Andrea[s] Arcangeli


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:44    [from the cache]
©2003-2014 Jasper Spaans. Advertise on this site