Gregor-
> I have it !
>
> The problem is in tccAllocTable. In CSL 2.31.00.2 the 1st entry in
> this table is set to 1, but in CSL 2.31.00.16 it is set to 0.
> MCBS UART calls EDMA_intAlloc. It returns the 1st free TCC code. But
> the TCC code means also interrupt number. In case of new CSL, TCC
> number 0 is returned, BUT interrupt 0 is dedicated for host-to-DSP
> interrupt.
> Thus if JTAGjet is connected to the CPU, EDMA with TCC = 0 is not
> working !
> The solution - I forced UART not to use TCC = 0.
>
> I have question - is it possible to detect presence of JTAGjet ?
> Because if it is not connected, UART can use TCC = 0.
That's a great catch, nice work to figure that out. I can't see yet
why
presence/absence of JTAGjet makes a difference. Is the "host-to-DSP
interrupt" you mention associated with HPI? (If so it would be called
DSPINT). Or is it associated with external interrupts (i.e. EINT4, EINT5,
etc) ?
-Jeff
> Note:
> I must admit I'm quite confused by interrupts <-> EDMA channels <->
> EDMA events <-> EDMA TCC. Can anybody explain shortly the relation
> betweem theese things?
Re: MCBSP UART not working after CSL upgrade ( sprc090 installation )
Started by ●November 3, 2006
Reply by ●November 6, 20062006-11-06
Hi Gregor. I am implementing a UART on DSK 6211 in CCS2.0, Had you
tried the same program that is given in spra633b,pdf from the texas
instruments. Please reply, this information is very useful for me.
On 03/11/06, Jeff Brower wrote:
> Gregor-
>
> > I have it !
> >
> > The problem is in tccAllocTable. In CSL 2.31.00.2 the 1st entry in
> > this table is set to 1, but in CSL 2.31.00.16 it is set to 0.
> > MCBS UART calls EDMA_intAlloc. It returns the 1st free TCC code. But
> > the TCC code means also interrupt number. In case of new CSL, TCC
> > number 0 is returned, BUT interrupt 0 is dedicated for host-to-DSP
> > interrupt.
> > Thus if JTAGjet is connected to the CPU, EDMA with TCC = 0 is not
> > working !
> > The solution - I forced UART not to use TCC = 0.
> >
> > I have question - is it possible to detect presence of JTAGjet ?
> > Because if it is not connected, UART can use TCC = 0.
>
> That's a great catch, nice work to figure that out. I can't see yet why
> presence/absence of JTAGjet makes a difference. Is the "host-to-DSP
> interrupt" you mention associated with HPI? (If so it would be called
> DSPINT). Or is it associated with external interrupts (i.e. EINT4, EINT5,
> etc) ?
>
> -Jeff
>
> > Note:
> > I must admit I'm quite confused by interrupts <-> EDMA channels <->
> > EDMA events <-> EDMA TCC. Can anybody explain shortly the relation
> > betweem theese things?
--
World Belongs to few enthusiasts who keeps kool
tried the same program that is given in spra633b,pdf from the texas
instruments. Please reply, this information is very useful for me.
On 03/11/06, Jeff Brower wrote:
> Gregor-
>
> > I have it !
> >
> > The problem is in tccAllocTable. In CSL 2.31.00.2 the 1st entry in
> > this table is set to 1, but in CSL 2.31.00.16 it is set to 0.
> > MCBS UART calls EDMA_intAlloc. It returns the 1st free TCC code. But
> > the TCC code means also interrupt number. In case of new CSL, TCC
> > number 0 is returned, BUT interrupt 0 is dedicated for host-to-DSP
> > interrupt.
> > Thus if JTAGjet is connected to the CPU, EDMA with TCC = 0 is not
> > working !
> > The solution - I forced UART not to use TCC = 0.
> >
> > I have question - is it possible to detect presence of JTAGjet ?
> > Because if it is not connected, UART can use TCC = 0.
>
> That's a great catch, nice work to figure that out. I can't see yet why
> presence/absence of JTAGjet makes a difference. Is the "host-to-DSP
> interrupt" you mention associated with HPI? (If so it would be called
> DSPINT). Or is it associated with external interrupts (i.e. EINT4, EINT5,
> etc) ?
>
> -Jeff
>
> > Note:
> > I must admit I'm quite confused by interrupts <-> EDMA channels <->
> > EDMA events <-> EDMA TCC. Can anybody explain shortly the relation
> > betweem theese things?
--
World Belongs to few enthusiasts who keeps kool
