Not able to get BOOTME Message - DM6467 Processor

If the pins are tied such that the DM6467 part is in UART boot mode, you should see one 'BOOTME' message (some other parts produce many BOOTME messages) from UART0.  Even if you tied the pins backwards (NAND boot mode), if there is nothing on NAND, it should fall back to UART boot and display similar message.  This message is generated by RBL which is the very first thing that runs, therefore if you are not seeing this message I would suggest that either UART port is not connected properly, power is not supplied to the part, or maybe clock is missing. 

Hi

The DM648 bootloader doc had a section on debugging boot loading. It specified a registers that you can look at with a JTAG emulator to see what your boot pins are. We somehow damaged our fast boot pin and would never had figured out what was wrong unless we looked at this register.

http://e2e.ti.com/support/dsp/tms320c6000_high_performance_dsps/f/112/p/11564/45101.aspx#45101

Hopefully there is a like register for the DM6467.

Good luck

Check BOOTMODE[3:0] bits in BOOTCFG register (0x01C40014) and make sure boot pins are latched correctly (to UART boot).

Hi , Still we are not able to get the BOOTME Msg, i read the BOOTCFG register. it shows the value as Zero.

We Face another problem, we directly applying the 3.3V VPIF_DOUT(3) pin amd making other pins to 0, but still there is no change. while doing this, i have not able to conenct the CCS studio and Emulator, what would be the problem.

Could anyone faced this type of issue, what are the things i can test to get the bootme message. is there any other way to enable to UART0 mode,In parallel i have laso tried the loiopback tset code and there is no response form the board.there is no signals available on the UART pins.

Could anyone pls comment.

Thanks,

Balaji S

a BOOTCFG value of zero would indicate the hardware is in emulation boot mode, which would not agree with a BTMODE[3:0] = 1000 value (UART0 boot mode).  If this is true (there are doubts since you are suggesting you are tying the BTMODE[3:0] pins to 1000 ), then this would explain why you may not be seeing a BOOTME message as emulation boot mode does not default to UART boot mode (some of the other boot modes such as NAND boot does default to UART0 boot if nothing is found in NAND).

How are you reading BOOTCFG register; doesn't the DM6467 part need to be booted (HW initialized) for the register values to be meaningful?

This sounds like a hardware issue; I would suggest that if you have the proper power sequencing applied, tied the BTMODE[3:0] pins correctly and provide proper clock to the part, you should see a BOOTME message without doing any software configuration (RBL is all you need at this point). 

Hi Juan Gonzales,

Thanks for the response.

We were reading BOOTCFG value as zero, when we configured BTMODE[3:0] resitors as 1000.

We did one more experiment, we configured BTMODE[3:0] as 0111 and we powered up the board. This time we could read the BOOTCFG register as NAND Flash mode.  We repeated this quite a few times and every time we read the correct BOOTCFG register. This rules out the possibilty of having any issue with power sequencing (also we verifed the power supply voltages 1.2 V, 1.8 V and 3.3 V on a digital storage oscilloscope).

Request you to provide solution why we are not able to read correct BOOTCFG register value in UART0 Mode (BTMODE[3:] = 1000). Here we physically verified the voltages (at the processor test points) for the configuration pins (VPIF_DOUT[7:0]).

Anticipating an early reply from your side.

Regards,

Balaji S

Balaji,

We never saw this issue before. To work UART boot mode, you should see BTMODE[3:0] =1000b in BOOTCFG register. From your desciption, it looks like, some how,  BTMODE[3] pin is struck low. All four boot pins BTMODE[3:0] should behave the same. I don't see any reason for BTMODE [2:0] work and BTMODE[3] doesn't work.

Did you try on mulitple boards? Do all boards behave the same? I am trying to rule out option of bad silicon or bad board.

Also refer to DM6467 EVM schematic at the following link for the connection of BTMODE[3:0] pins.

http://support.spectrumdigital.com/boards/evmdm6467

Regards, Srirami.

Hi,

still we didnt get the bootme message in the serial port, we suspect that vpif_dout(3) pin. i have successfully loaded the spi and i2c eeprom, if i make the u-boot settings to spi or i2c, will i get ubl messges in the uart port.

thanks,

balaji s

Hi,

Thanks for the response.

We are able to boot DM6467 in the following modes:

1) Emulation boot

2) HPI-16

3) HPI-32

4) EMIFA direct boot

With this we are assuming that there is no issue with configuration resistor setting(s) of the DM6467.

But, we are not able to boot in the remaining boot modes (I2C, NAND Flash, UART0).

We are required t boot through UART0 mode (or) NAND Flash mode (or) I2C mode.

Pl advice on the same.

Regards,

Balaji

Balaji,

Are you able see the correct BTMODE[3:0] bits in BOOTCFG register for these three boot modes (I2C, NAND Flash, UART0)? I know are seeing 0000b (instead of 1000b) for UART0 boot . What about other two boot modes?

Regards, Srirami.

HI Sriram,

Actually am reading the bootcfg register settings value through CCS only, so when i set inti I2C, SPI and NAND boot mode. it is not able to read the read the bootcfg register , it strucks at one point and am getting errrot msg like this.

i have attached the error msg with this, could you able to read this.

Thanks,

Balaji S

Hi,

How to read the BOOTCFG register when the board is not getting connected  to CCS?

balaji s

Balaji,

I am not able to see the attached error message. Re-attach the file (snapshot of the error).

There is no way to see the bootcfg register with out CCS.

My understanding is that you are able to connect to CCS, buy when you try to see BOOTCFG register through CCS memory window, you are getting the error. Correct? If that is correct, what PC (Program Counter) value you see as soon as you connnect the CCS?

Regards, Srirami.

HI Sriram,

When i set it to Nand, I2C, UART and SPI mode in the bootmode, am not able to connect the ccs itself. it hangs,when it tries to read the bootcfg register in the gel file.

we have verified the circuit with evm board schematics, everything looks good, even we have validated max232 transceiver also. is thera any other way to check the processor UART0 pins works well??Because we suspect processor itself. custom board processor is ver1.1 and evm processor is ver 3.0.

thanks,

balaji s

Balaji,

Remove the GEL file before connecting the CCS and see if you can connect the CCS. I guess you can. Once you connect the CCS, let me know the PC (Program Counter) value. Now, try to open the memory window and check the value at address BOOTCFG (0x01c40014).

Coming to the processor differences between 1.1 and 3.0 revisions, a new UART feature is added in 3.0 (See Silicon Errata for more details). But, there is no change in UART0 boot mechanism between 1.0 and 3.0.

Also, I am not sure if the processor on the custom board went bad. That was the reason I was asking you to try on another custom board.

Regards, Srirami.

Balaji,

One more inportant thing to use the UART0 boot. There should be a stable 24-MHz clock on “AUX_MXI/AUX_CLKIN” pin to use UARTs (UART0 boot). Check your board and make sure of good AUX clock.

Regards, Srirami.

Hi,

Sorry for the late response, we had  issue in getting 24 MHZ clock in our circuit. thst the reason we were not able to get the boome msg in the serial port. now we recitified and board works fine in all the boot modes.

we had put UBL into spi eeprom and Uboot into the nand flash using CCS utilities. But we face another problem is that ethernet phyceiver is not up. when we try to load the kernel from the TFTP, we are getting "EMAC link is not detected".

we thought of trying to put kernel into the flash using CCS nand_writer_utilities that came with DVSDK, we are successfully able to write into nand flash, but in the command prompt we are getting "Bad CRC". we had set the BOOTCMD and BOOTARGS correctly.

we are writing the kernel from the 320th page of nand flash.

Could anyone tried this option earlier???

is ther anyother way to put kernel and filesysytem into the nand flash other than tftp?

thanks,

Balaji S

Hi,

We are working on a hardware board design which used TMS320DM6467AUTA processor. Currently we are into the board bring up activity. We were able to boot the processor and load UBL into SPI EEPROM and UBOOT into NAND Flash.

But since two days whenver we are booting the processor in UART0 mode, we are noticing a continous smiley characters on the Hyper terminal. Also whenver we tried to boot in SPI boot mode, we are not able to the boot messages completely / correctly.

We have checked the power up sequence and reset signals to the processor and these are proper as per expectations. Infact we are able to connect to the code composer studio (CCS) and run a few test programs successfully.

Request to provide solution as to why a continous smiley characters are seen on Hyper terminal in UART0 mode and why boot messages are corrupted in SPI mode.

Regards,

Balaji S

1> Since last two days you are seeing smile characters means was it working fine before?

2> Are you seeing Smile characters after BOOTME message?

Please open a seperate post with your above information copied. This post was meant for BOOTME message display issue which is resolved.

Regards, Srirami.