RE: [ECOS] Problems with Thread-start up delays

["Chris Sekula" <chriss at turnpikeglobal dot com>]

Johnathan:
 
Thank you for your comments. Increasing the priority of the test thread
solved the problem. It didn't occur to me that the main thread had to
finish its timeslice before the test thread would be scheduled. I'm
wondering whether I should also call cyg_thread_yield() after resuming
the test thread in my application to ensure that the main thread gives
up its timeslice under all conditions, or does this function only have
an effect if there are threads of equal priority?
 
Thanks,
 
CHRIS.

	-----Original Message----- 
	From: Jonathan Larmour 
	Sent: Wed 31/10/2001 4:55 PM 
	To: Chris Sekula 
	Cc: ecos-discuss@sources.redhat.com 
	Subject: Re: [ECOS] Problems with Thread-start up delays
	
	

	Chris Sekula wrote:
	>
	> The results of this test suggest to me that there is delay of
	> approximately 40 or 50 milliseconds associated with resuming a
thread.
	> However, this disagrees with what has been published in the
eCos
	> documentaion. Have I missed something in the interpretation of
my
	> results or the code itself? Can anyone offer an explanation
for what I
	> have observed?
	
	Easy :-). 50 milliseconds is..... the default between
timeslices! In your
	test case you gave the threads equal priority, but the newly
resumed thread
	won't run until it gets timesliced.
	
	Two things to make a more accurate test: you should set the test
thread to
	a higher priority, and make it read the high definition clock
using
	HAL_CLOCK_READ (although you have to watch out for wrapping) and
write that
	into a global variable (otherwise you have to worry about the
time taken to
	exit TestThread and reschedule MainProc).
	
	Have a look at the code in the kernel tm_basic test as that's
where the
	measurements in the documentation came from.
	
	Jifl
	--
	Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44
(1223) 271062
	Maybe this world is another planet's Hell -Aldous Huxley ||
Opinions==mine