From: Roberto P. <pad...@gm...> - 2009-04-09 19:06:38
|
I found a mistake in (my own) ucs.h definition file, about the XT2 crystal drive strength; here is the patch against the previously attached file. R# 2009/4/9 Roberto Padovani <pad...@gm...>: > 2009/4/8 Sergey A. Borshch <sb...@us...>: >> Roberto Padovani wrote: >>> Hi all, >>> >>> I cannot find the x5xx family definitions for UCS module.....am I >>> missing them, or are they missing at all ? >> Yes, it's not contributed yet. >>> If no one has written them, I suppose I have to write and contribute >>> them myself, right ? >> Exactly. >> >>[..] > > Ok, here is a patch to the msp430x54xx.h and the ucs.h > The bit definitions are complete but I typed only the word oriented > definitions, without the _H and _L replicas, and there is no > structured packed type because.....I honestly did not understand it > very well...:-/ > > Can someone have a look at register UCSCTL8? It is somewhat > particular: there are some bits that are reserved in the TI family > guide (slau208), but which MUST always be written with 1s to have the > clocks running. > I made the choice to define: > > /* UCSCTL8 */ > #define MODOSCREQEN (1<<3) /* MODOSC Clock Request Enable */ > #define UCSCTL8_MASK (0x0317) /* Bit 0,1,2,4,8,9 MUST always be > written as 1 !!!!! > > /* Aliases by mspgcc */ > #define MODOSCREQEN_0 (UCSCTL8_MASK) /* Correctly Clear MODOSC: > UCSCTL8 = MODOSCREQEN_0 */ > #define MODOSCREQEN_1 ((UCSCTL8_MASK)||MODOSCREQEN) /* Correctly Set > MODOSC : UCSCTL8 = MODOSCREQEN_1 */ > > so that by using the MODOSCREQEN_0 or MODOSCREQEN_1 everything is > fine....but is someone uses the MODOSCREQEN directly, he has to pay > attention not to forget the bit mask. > > During the next days I'll try everything on an actual board and see > what I can do for the other new peripherals. > > R# > |