TMS570 BSP testing and problem with VFP enabled build

Martin Galvan martin.galvan at tallertechnologies.com
Mon Nov 23 20:31:46 UTC 2015


I'm about to test this on our setup. Just to be sure, does your
startup code perform the register initialization required by the
CCM-R4F? I added that to arm/shared/start/start.S a few months ago.

On Mon, Nov 23, 2015 at 5:26 PM, Pavel Pisa <pisa at cmp.felk.cvut.cz> wrote:
> Hello Martin,
>
> On Monday 23 of November 2015 15:19:40 Martin Galvan wrote:
>> Hi Pavel, I'll give it a look. What issues are you having? Did you see
>> this problem on the RTEMS samples (e.g. HelloWorld, Ticker) as well?
>
> we have checked functionality with ticker and our complete LwIP application.
>
> We observed problem for ticker (we run it from SDRAM) as well.
> Start up messages from Init task are printed. Then three individual ticker
> threads starts but then system locks. We have tried to put hardware
> breakpoints all exception handlers (undef, prefetch, abort, ...) but we have
> not catch anything. CPU goes to error state (ERROR let lights).
> I am suspicious that problem appears when the fist started ticker thread
> is to be scheduled again and VFP context restoration fails for some reason.
> But I need to find time for more investigation later. We have replaced
> our RTEMS loader on HDK boards we pose  to run other projects today.
> The final ERROR state can be result of failed CPU reset due to active POM.
> Actual problem source can be VFP contex switch code or some register
> missing initialization. May it be, the fault is not at restore but during
> save when lockstep founds discrepancy during save to memory.
> May it be, your code sets the registers which our does not.
> But generally it seems that care is taken of d0 to d15 in HalCoGen
> the startup code which we use
>
> https://github.com/hornmich/tms570ls3137-hdk-sdram/blob/master/SDRAM_SCI_configuration/source/sys_core.asm
>
> Testing in your setup would be significant help because it is hint
> if this is problem in our setup/tools or mainline change.
> I need to test code with our previous toolchain version we have used
> before summer as well.
>
> Best wishes,
>
>              Pavel



-- 


Martin Galvan

Software Engineer

Taller Technologies Argentina


San Lorenzo 47, 3rd Floor, Office 5

Córdoba, Argentina

Phone: 54 351 4217888 / +54 351 4218211


More information about the devel mailing list