Hi Grigory,

When I ported the delay functions to OMAP/925 from the 920 tree there were (and still are) a few bugs which showed themselves when the Innovator port was first done. (I use a different board and didn't run into the bugs until more recently). The person which worked with the Innovator choose to put in hard delays instead of taking the time to find the bugs (probably didn't have time). At some point I also ran into the bugs and was able to allot time to fix them and posted a patch.

Now that the bugs are hopefully fixed the old code should be removed. I don't have an Innovator handy at the moment to validate and submit a patch but I think it should be done.

One side note is I also changed the period of the clock in my local include/configs/xxxx.h file such that it was less likely to overflow in normal use. What is there by default is ok, just not optimal.

Regards,

Richard W.

-----Original Message----- From: Grigory A. [mailto:Ryhor@tut.by] Sent: Monday, January 05, 2004 7:16 PM To: Woodruff, Richard Cc: u-boot-users@lists.sourceforge.net Subject: Re: [U-Boot-Users] OMAP1510 Innovator kit booting questions

Hello Richard,

WR> In cpu/arm925/interrupts.c. Remove the hard coded delay loops which are WR> wrapped with INNOVATOROMAP1510. This enables the code which will use ARM WR> timers (which are driven by the 12MHz reference / by a factor in the header WR> file). This should provide constant time irrespective to your cpu dpll WR> speed. This is what I use in (though not with innovator).

WR> This will fix some problems, and it "might" fix your problem. It will WR> probably make your boot speed correct, but it still sounds like some clock WR> is not set properly. I have a bit more fixed up dpll setup code which I WR> will get around to sending in sometime. WR> Regards, WR> Richard W.

Yes - it is work - now all is correct. I've removed (commented) these wrapped lines and seconds came to normal duration. What was a reason to put this delay loops?

Best regards, Grigory mailto:Ryhor@tut.by