Messages in this thread | | | Subject | Re: [PATCHv3 2/4] ARM: Allow machines to override __delay() | From | Daniel Walker <> | Date | Wed, 08 Dec 2010 12:26:32 -0800 |
| |
On Wed, 2010-12-08 at 12:24 -0800, Stephen Boyd wrote: > On 12/08/2010 11:22 AM, Daniel Walker wrote: > > On Tue, 2010-12-07 at 20:38 -0800, Stephen Boyd wrote: > >> + > >> +static void (*delay_fn)(unsigned long) = delay_loop; > >> + > >> +void set_delay_fn(void (*fn)(unsigned long)) > >> +{ > >> + delay_fn = fn; > >> +} > >> + > >> +/* > >> + * loops = usecs * HZ * loops_per_jiffy / 1000000 > >> + */ > >> +void __delay(unsigned long loops) > >> +{ > >> + delay_fn(loops); > >> +} > >> EXPORT_SYMBOL(__delay); > > Didn't we already go over this part ? Why are aren't you putting these > > in a header file ? > > Last time we saw that inlining set_delay_fn() actually increased the > text size of the kernel. I know it sounds wrong, but its probably due to > that compiler behavior Russell posted about last month on arm-lkml.
The last time I thought you said there was no size change ? My tests showed a decrease ..
What compiler optimizations are you talking about ?
Daniel
--
Sent by a consultant of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
| |