CSL for 6455 still Supported?

Brad Griffis said:

Unfortunately I think it's unlikely that you'll see another release of the 6455 CSL.

I thought you'd say that. Again I'm disappointed by the support offered for the 6455. We're designing the 6455 into a new product but it appears TI no longer support it (no PSP, no DSP/BIOS v6, etc).

Brad Griffis said:

Personally I found (functional layer) CSL 3.x to be buggy and hard to understand (at least for 6455).

That's exactly what I was thinking. I'm disappointed by the actual software provided for the 6455 but I do like the forums and the support you and others are providing here.

Brad Griffis said:

I recommend using the underlying register level CSL.  A good example of what I'm talking about can be found here in the "CSL 3.x Register Layer Examples".

I think this is effectively what I've been doing although I not sure the that 'CSL_F' macros "bring a great deal to the party". Personally I think the 'CSL_F' macros are a step to far and it would be just as clear to use normal 'and' and 'or' operations in the examples but as with all programming issues you will get a different opinion from every programmer you ask.

Thanks,

Matt

MattB said:

We're designing the 6455 into a new product but it appears TI no longer support it (no PSP, no DSP/BIOS v6, etc).

I would agree with you in general that software support for this device has been minimal.  Note, however, that it continues to be fully supported by DSP/BIOS which in my mind is by far the best software product TI produces.  I actually find BIOS 6 a bit difficult to use right now (though admittedly it's been about 6 months since I gave it a hard look).  In my mind BIOS 5 remains the "gold standard" in TI software.  FWIW, BIOS 6 supports c6455.  Here's are the supported devices in the current release.  I think that as BIOS 6 matures a bit it will be terrific.  I think it's solid/reliable right now, just not as refined in terms of ease of use like you'd see with BIOS 5.

MattB said:

I recommend using the underlying register level CSL.  A good example of what I'm talking about can be found here in the "CSL 3.x Register Layer Examples".

I think this is effectively what I've been doing although I not sure the that 'CSL_F' macros "bring a great deal to the party". Personally I think the 'CSL_F' macros are a step to far and it would be just as clear to use normal 'and' and 'or' operations in the examples but as with all programming issues you will get a different opinion from every programmer you ask.

[/quote]

If you're looking to use named bitfields rather than a bunch of hex digits OR'd together then I think the macros are useful.  Admittedly it's a bit of work to get the right names for everything in order to take advantage of them.  So it might actually take a little bit longer to utilize them the first time around, but I think when it comes to maintenance of the code they are very helpful.  My two cents...

Brad Griffis said:

FWIW, BIOS 6 supports c6455.  Here's are the supported devices in the current release.

Yeah, you right of course. BIOS 6 supports the 6455 because there is no PSP for the 6455. I forgot that the main reason we've decided to stick with BIOS 5 is that we're also using an OMAPL-137 which has a PSP that only works with BIOS 5.

I'm a bit worried that BIOS 5 won't get any attention now, there won't be any improvements and various issues I've seen won't be fixed (see http://e2e.ti.com/forums/p/11440/50624.aspx#50624 and http://e2e.ti.com/forums/t/13075.aspx).

Brad Griffis said:

If you're looking to use named bitfields rather than a bunch of hex digits OR'd together then I think the macros are useful.

I meant use macros (or possibly consts) to define the bitfields (which I think the CSL does) but then just use them directly - but it's personal taste I think.

Thanks,

Matt

MattB said:

I'm a bit worried that BIOS 5 won't get any attention now, there won't be any improvements and various issues I've seen won't be fixed (see http://e2e.ti.com/forums/p/11440/50624.aspx#50624 and http://e2e.ti.com/forums/t/13075.aspx).

That's not the case at all.  BIOS 5 still gets substantial development resources and continues to have regular releases, bug fixes, enhancements, etc.  Download the latest from here:

http://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/index.html

I think some of your RTA issues stem from the fact that the BIOS team had to "start over" to integrate RTA support into CCS 4 and so it isn't quite as stable/mature as CCS 3.  Since CCS 4 has been released it has already come a long way as you saw in your update to CCS 4.0.2.  Note that you should also grab the latest version of BIOS from the link I referenced above and some of the fixes are contained in BIOS, not in CCS.  You should also grab the latest xdctools to get fixes for the "ROV" viewer (the replacement to Kernel Object Viewer KOV in CCS 3.3).

It sounds like you're past many of the RTA issues, but I thought I'd mention this wiki article just in case:

http://tiexpressdsp.com/wiki/index.php?title=DSP/BIOS_Real-time_Analysis_%28RTA%29_troubleshooting

Brad Griffis said:

That's not the case at all.  BIOS 5 still gets substantial development resources and continues to have regular releases, bug fixes, enhancements, etc. 

I'm glad to hear that! :-)

Brad Griffis said:

Download the latest from here:

http://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/index.html

That's a good tip, I hadn't been watching this page. I'm a few DSP/BIOS versions behind so I'll do an update now!

Thanks,

Matt

One other thing -- after you install the latest version of BIOS you need to go into your project options and select the new version.

Brad Griffis said:

That's not the case at all.  BIOS 5 still gets substantial development resources and continues to have regular releases, bug fixes, enhancements, etc.  Download the latest from here:

http://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/index.html

I had been looking at this page http://software-dl.ti.com/dsps/dsps_registered_sw/sdo_sb/targetcontent/bios/index.html which doesn't show the later versions. Hence my confusion!

Matt