C6747 Bootpins and bootcfg is not in actual state

Either something is driving boot0 low or else you have a solder bridge or something causing the issue.

boot0 pin is driving high i saw this by oscilloscope

Is this on your own custom hardware?

Is the issue present on more than one board?

Where did you make your measurement?  For example, is there a via where you can make the measurement as close to the ball as possible?

When did you make your measurement?  Are you able to trigger off the reset pin such that you can see the state of the pin as reset is deasserted?

That looks weird, I just tried reading the bootcfg register (viewing memory address 0x01C14020 at CCS) in the EVM - and in the EVM it is consistent with the pins.

This is my custom board, no such problem in the EVM. I have two identical custom board and one board work fine but another have a problem with bootpins (bootcfg).
I measure bootpin signals right on the via of the ball and the DSP is in reset.

The pin gets latched on the rising edge of reset so you should get a capture of the pin state at that moment.  You should probably x-ray the board to see what is wrong with the connection.

In the two custom board bootpins state on the rising edge of reset are identical. But, i repeat, in the one board bootcfg is real in the other board bootcfg looks crazy.

That's very odd!  Try swapping the devices to see if the trouble follows the device or follows the board.

I have to agree with Brad. Capturing the value of the boot pins is a very simple process, so I do not know what could go wrong with it (unless the pins aren't actually pulled the proper way). From what I can see there is likely an issue with the connection of those pins.

Are you able to modify the direction of the boot pins to see if the register values changes with it?

Great suggestion, Tim!  Going further with Tim's suggestion you should use that pin as a general purpose output to verify that you see proper the proper corresponding "high" and "low" on the via you are observing.  You can also test as a general purpose input by applying a voltage to that via and seeing if the corresponding level is shown in the input register.

Not Great suggestion.

I try this:

- config bootpins as outputs and no problem with high and low level on the pins. toogled by the timer interrupt

- config bootpins as inputs and capture value, this values are boot0=1 boot1=0 boot2=1 boot3=0 boot7=0 but in the bootcfg register values are boot0=0 boot1=0 boot2=1 boot3=0 boot7=0.

Look likes my TMX320C6747 just a crazy DSP. I observe that TI has no solution for my problem and it disturb.

Cant swap devices from board to board because i got only 2 devices and afraid to damage this dsp.

The only other thing I would recommend inspecting would be to verify your power supplies and clock are stable and within the recommended range when the DSP is released from reset.  If that doesn't solve your issue then you should email DSP Technical Support from here:

http://support.ti.com/

Reference this thread and they can assist you in having failure analysis performed and/or getting a new DSP.

Brad