This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Re: Is clock interrupt = tick ?
- To: Huang Qiang <jameshq at liverpool dot ac dot uk>
- Subject: Re: [ECOS] Is clock interrupt = tick ?
- From: Jonathan Larmour <jlarmour at redhat dot com>
- Date: Thu, 19 Apr 2001 23:03:09 +0100
- Cc: eCos <ecos-discuss at sources dot redhat dot com>
- Organization: Red Hat UK Ltd.
- References: <KIEBICHBADHFCLGCKOPDEEDOCBAA.jameshq@liv.ac.uk>
[ Is clock interrupt = tick? ]
It may depend on your terminology. Most hardware clocks have an internal
counter, often a resetting decrementer which you program to trigger an
interrupt when it hits zero.
It's up to you whether you consider the change in the value of the
decrementer to be a tick, or the clock interrupt to be a tick. My
preference is for the latter, and I think that is definitely the most
common usage. So in short, yes :-).
But note that kernel timeslicing doesn't take place every clock interrupt -
it actually happens according to the eCos CDL option
CYGNUM_KERNEL_SCHED_TIMESLICE_TICKS, which defaults to 5 ticks between
reschedules.
Jifl