Hi, 14 серпня 2013 о 16:54 +0200 Karel Gardas написав(-ла): > What makes me more nervous about supporting older chips is addition: > > rcc &= SYSCONRCC_MOSCDIS; > syscon->rcc = rcc; Ugh, wait, you are clearing all bits but SYSCONRCC_MOSCDIS in RCC? That could not be right. Did you mean to do
rcc &= ~SYSCONRCC_MOSCDIS;
??
>
> ^ w/o this my code is not running at all, i.e. it probably runs on
> default 12 MHz. I'm curious how older chips worked without this.
The switching to PLL procedure for lm3s3749 does not mention MOSCDIS
bit. In fact, it is only mentioned in the register description, but not
anywhere else in the datasheet. I guess it should be safe to clear it,
but I cannot test this on the real hardware till next week.
> I've also added various waits as recommended and as used in TI's own
> StellarisWare library. This is needed as in chip spec., itself,
> there are no such details provided... :-(
In lm3s3749 datasheet there are some things like this spread all over
the datasheet:
There must be a delay of 3 system clocks after the μDMA
module clock is enabled before any μDMA module registers are
accessed.
No word about any longer delays. I guess you can also remove
delay_3_clocks() from syscon.c and use your new function there.
Regards,
Eugeniy Meshcheryakov
signature.asc
Description: Digital signature
_______________________________________________ rtems-devel mailing list rtems-devel@rtems.org http://www.rtems.org/mailman/listinfo/rtems-devel
