Re: [PATCH] loosing time with CPU counter timer

[Juan Quintela <quintela@mandrakesoft.com>]

>>>>> "jun" == Jun Sun <jsun@mvista.com> writes:

Hi
        only aesthetical fixes:

jun> diff -Nru linux/arch/mips/ddb5xxx/ddb5477/setup.c.orig linux/arch/mips/ddb5xxx/ddb5477/setup.c
jun> --- linux/arch/mips/ddb5xxx/ddb5477/setup.c.orig	Mon Apr 14 15:28:57 2003
jun> +++ linux/arch/mips/ddb5xxx/ddb5477/setup.c	Mon Apr 14 16:08:35 2003
jun> @@ -330,6 +348,16 @@
jun> * high-level timer interrupt service routines.  This function
jun> * is set as irqaction->handler and is invoked through do_IRQ.
jun> */
jun> +static inline int 
jun> +64bit_compare(unsigned hi1, unsigned lo1, unsigned hi2, unsigned lo2)

Please, use "unsigned int" /me notices that file is not consistent, in
some places it uses "unsigned" and in other "unsigned int".

jun> +{
jun> +	if (hi1 == hi2) {
jun> +		return lo1 - lo2;
jun> +	} else {
jun> +		return hi1 - hi2;
jun> +	}
jun> +}

Ralf tax on braces apply here.

jun> +		/* check to see if we have missed a timer interrupt */
jun> +		if (64bit_compare(timerhi, timerlo, expirehi, expirelo) < 0) {
jun> +			unsigned int old_expirelo=expirelo;
jun> +			expirelo += cycles_per_jiffy;
jun> +			expirehi += (expirelo < old_expirelo);

Tax on parens :)

Later, Juan.

-- 
In theory, practice and theory are the same, but in practice they 
are different -- Larry McVoy