Thanks! I knew about the struct/word alignment problem as that kind of thing is typical on lots of processors. In my experience, the C compiler generates extra instructions to handle such misalignments.
This MSP430X branch business is a total surprise. My old and possibly future client is switching from the 430x149 to the x249. How do I know if this toolchain recompile is necessary for a processor? (I guess it's time to read your link...) Thanks for the tip. frank --- On Thu, 7/23/09, Roberto Padovani <[email protected]> wrote: > From: Roberto Padovani <[email protected]> > Subject: Re: [Mspgcc-users] using JTAG debugger - several issues > To: "GCC for MSP430 - http://mspgcc.sf.net"; > <[email protected]> > Date: Thursday, July 23, 2009, 11:50 PM > Besides the spy-bi-wire and jtag > difference, the x5xx MCUs are > supported only by the MSP430X branch of the mspgcc project, > i.e. you > have to recompile the toolchain by yourself. The > peripherals are also > partially supported, i.e. you might have to write some > header files to > define the register map. > Have a look at the mspgcc wiki to see what to do for the > MSP430X > branch: > http://sourceforge.net/apps/mediawiki/mspgcc/index.php?title=MSPGCC_Wiki > > R# > > 2009/7/24 Frank Harvey <[email protected]>: > > Most of my msp430 experience used IAR tools. I'm still > switching over to gcc tools (which i've used on non msp > cpus). > > > > If I saw the .h file for a cpu included in the > distribution, I too would think that cpu is supported. I > think you can bank on that. > > > > yes, there's a time when just the FET and your host > computer are busy updating the FET itself. This process has > nothing to do with the eventual target cpu chip. > > > > In an IAR environment, the FET doesn't try to 'talk' > to the chip until you indirectly 'ask' it to start > something. It will then complain if the cpu chip is unable > to answer (because it's not completely cabled, or it has no > clock (?), or power, etc). You can see what it says under > these different states by disconnecting the flat cable from > the FET to the cpu and asking gdbproxy to do something. Note > what complaints it make. > > > > The USB version of the FET is capable of powering > small boards (slau278u, your reference). We generally > wired our systems so the FET was unable to supply power at > all. A problem with the FET powering the system is 1. it may > not be able to supply enough current for too big a board and > 2. There's a software setting (somewhere) which says what > voltage to give it (1.8 to 3.6). I _think_ the default was > 3.6 but that may have come from the old IAR s/w I used. > > > > Since the board is new and "custom", it may have > problems. I'd start by checking the voltages very carefully > against the documentation. Take into account that the 54xx > uses SpiByWire and not pure JTAG. Check to see that the > target board has the expected power voltages. See that the > msp has the expected clock. > > > > It was not unusual for us (a team of five) to have > problems with FETs. They were a pain when switching between > different msp430 models (430x14x to msp430x12xx for > instance). It would complain about the cpu model being > 'different' if there was a problem with cabling, voltage, > etc etc. > > > > frank > > > > --- On Thu, 7/23/09, José Miguel Catela <[email protected]> > wrote: > > > >> From: José Miguel Catela <[email protected]> > >> Subject: [Mspgcc-users] using JTAG debugger - > several issues > >> To: [email protected] > >> Date: Thursday, July 23, 2009, 12:22 PM > >> Hello, > >> > >> First of all, let me tell you I'm really sorry if > my > >> questions are > >> already answered somewhere, but I've been > searching many > >> forums for > >> three days and I haven't got a clue if I can even > use my > >> hardware. > >> > >> I have a MSP-FET430UIF and a custom development > board > >> connected to it > >> according to TI's document "MSP430 Hardware Tools > User's > >> Guide" > >> (slau278a). > >> The microcontroller is a MSP430F5438 and this is > where my > >> doubts > >> begin: is the new MSP430x5xxx series supported by > mspgcc > >> tool chain or > >> not? I haven't found a decisive answer, and the > most recent > >> version of > >> mspgcc includes a file called "msp430x54xx.h", > which makes > >> me believe > >> it is supported. > >> But the main issue is to know if the FET is > working or not. > >> When I try > >> to start msp430-gdbproxy, I can either get the > message > >> "Could not find > >> device (or device not supported) (4)" followed by > >> "Assertion failed : > >> !msp430_status.is_open, file target_msp430.c, line > 745" (if > >> giveio is > >> installed) or "Could not initialize device > interface (1)" > >> (if giveo is > >> not loaded). And this happens whether the FET is > connected > >> to the USB > >> port or not. I can't also update the firmware, it > just says > >> "The FET > >> tool version does not match this program. Update > required." > >> when I try > >> to update it, what seems useless, because that's > what I'm > >> trying to > >> do. So, my big question is: does the self test, > update or > >> something > >> else works just with the FET plugged in? Or must > a > >> compatible > >> processor be connected to the FET through the 14 > pin > >> interface? > >> A last comment: I've described what happens using > windows > >> but under > >> linux the problems are just the same. > >> > >> I appreciate any help anyone can give me, > >> > >> -- > >> José Miguel Catela > >> > >> > ------------------------------------------------------------------------------ > >> _______________________________________________ > >> Mspgcc-users mailing list > >> [email protected] > >> https://lists.sourceforge.net/lists/listinfo/mspgcc-users > >> > > > > > ------------------------------------------------------------------------------ > > _______________________________________________ > > Mspgcc-users mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/mspgcc-users > > > > ------------------------------------------------------------------------------ > _______________________________________________ > Mspgcc-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/mspgcc-users >
