Dear all,
I have a tested code flashed into the DSK Flash, I power cycled the DSK and
it executed as I expected. I can connect with it and download symbols for
debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
through my XDS510PP+ when my program is flashed and running. Anybody know
why is it so? I have reduced my data rate to half and it is still the same.
Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
the same issue?
Background:
The XDS510PP+ is tested and manage to work with the DSK when it is flashed
with post.hex or blink.hex
The driver used is: sdgo6400_11.dvr
Target: DSK6416, 600MHz
CCS: V3.00.00
Flashburn: V2.8 (version 5.90.0.110
About my code:
The code uses two edma channels, one doing a ping pong transferring from
emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
edma generate HW interrupts (that initiates SW interrupt) which I use to
preprocess the data in ping pong buffer and dump it into a input buffer and
post-process the data from an output buffer. These are done with qdma.
Then, I have a Task that repeatedly process the data in the input buffer
and place it at the output buffer. I also have a periodic interrupt that
generates patterns of led display. The input sampling rate is now 500ksps.
Problem re-connecting to target thru XDS510PP+ when my code is flashed and running in DSK, please help
Started by ●July 4, 2005
Reply by ●July 5, 20052005-07-05
Are you using a GEL file when you connect in either [or both cases]??
If you are, try connecting without one [if you have a complete working
program in flash, a GEL file should not be required unless it is to support
a debug feature.
mikedunn
Koay Teong Beng <k...@arl.nus.edu.sg> wrote:
Koay Teong Beng <k...@arl.nus.edu.sg> wrote:
Dear all,
I have a tested code flashed into the DSK Flash, I power cycled the DSK and
it executed as I expected. I can connect with it and download symbols for
debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
through my XDS510PP+ when my program is flashed and running. Anybody know
why is it so? I have reduced my data rate to half and it is still the same.
Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
the same issue?
Background:
The XDS510PP+ is tested and manage to work with the DSK when it is flashed
with post.hex or blink.hex
The driver used is: sdgo6400_11.dvr
Target: DSK6416, 600MHz
CCS: V3.00.00
Flashburn: V2.8 (version 5.90.0.110
About my code:
The code uses two edma channels, one doing a ping pong transferring from
emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
edma generate HW interrupts (that initiates SW interrupt) which I use to
preprocess the data in ping pong buffer and dump it into a input buffer and
post-process the data from an output buffer. These are done with qdma.
Then, I have a Task that repeatedly process the data in the input buffer
and place it at the output buffer. I also have a periodic interrupt that
generates patterns of led display. The input sampling rate is now 500ksps.
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/c6x/
<*> To unsubscribe from this group, send an email to:
c...@yahoogroups.com
<*
Reply by ●July 5, 20052005-07-05
Mike,
I didn't use startup GEL file in both cases.
Peripheral initializations are all done in my code.
Mike and all,
I start to think along the line that the DSP is too occupied to service the
JTAG. Since my main data processing work is running in a while loop in a
Task, I tried a few things described below but doesn't help, any more ideas?
1) I added a small code into the Task containing my data process algorithm.
The code sense an external pin and run IDL_run() instead of my processing
algorithm upon a pullup to the external pin.
2) I added a small code that jump to another while loop containing only
IDL_run() upon a pullup to an external pin.
3) I break out from the while loop in the Task upon a pullup to an external
pin.
4) I break out from the while loop in the Task and disable all the
interrupts (except PRD function) and EDMA events upon a pullup to an
external pin.
I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
me to reconnect my XDS510PP+, but doesn't work too. :(
Thanks for taking interest.
Best,
Teong Beng
At 01:06 PM 7/5/2005, Mike Dunn wrote:
>
>Are you using a GEL file when you connect in either [or both cases]??
>
>If you are, try connecting without one [if you have a complete working
>program in flash, a GEL file should not be required unless it is to
>support a debug feature.
>
>mikedunn
>
>Koay Teong Beng <koay@koay...> wrote:
>Dear all,
>
>I have a tested code flashed into the DSK Flash, I power cycled the DSK and
>it executed as I expected. I can connect with it and download symbols for
>debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
>through my XDS510PP+ when my program is flashed and running. Anybody know
>why is it so? I have reduced my data rate to half and it is still the same.
>
>Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
>the same issue?
>
>Background:
>The XDS510PP+ is tested and manage to work with the DSK when it is flashed
>with post.hex or blink.hex
>The driver used is: sdgo6400_11.dvr
>
>Target: DSK6416, 600MHz
>CCS: V3.00.00
>Flashburn: V2.8 (version 5.90.0.110
>About my code:
>The code uses two edma channels, one doing a ping pong transferring from
>emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
>edma generate HW interrupts (that initiates SW interrupt) which I use to
>preprocess the data in ping pong buffer and dump it into a input buffer and
>post-process the data from an output buffer. These are done with qdma.
>Then, I have a Task that repeatedly process the data in the input buffer
>and place it at the output buffer. I also have a periodic interrupt that
>generates patterns of led display. The input sampling rate is now 500ksps. >
>
> * >----------
I didn't use startup GEL file in both cases.
Peripheral initializations are all done in my code.
Mike and all,
I start to think along the line that the DSP is too occupied to service the
JTAG. Since my main data processing work is running in a while loop in a
Task, I tried a few things described below but doesn't help, any more ideas?
1) I added a small code into the Task containing my data process algorithm.
The code sense an external pin and run IDL_run() instead of my processing
algorithm upon a pullup to the external pin.
2) I added a small code that jump to another while loop containing only
IDL_run() upon a pullup to an external pin.
3) I break out from the while loop in the Task upon a pullup to an external
pin.
4) I break out from the while loop in the Task and disable all the
interrupts (except PRD function) and EDMA events upon a pullup to an
external pin.
I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
me to reconnect my XDS510PP+, but doesn't work too. :(
Thanks for taking interest.
Best,
Teong Beng
At 01:06 PM 7/5/2005, Mike Dunn wrote:
>
>Are you using a GEL file when you connect in either [or both cases]??
>
>If you are, try connecting without one [if you have a complete working
>program in flash, a GEL file should not be required unless it is to
>support a debug feature.
>
>mikedunn
>
>Koay Teong Beng <koay@koay...> wrote:
>Dear all,
>
>I have a tested code flashed into the DSK Flash, I power cycled the DSK and
>it executed as I expected. I can connect with it and download symbols for
>debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
>through my XDS510PP+ when my program is flashed and running. Anybody know
>why is it so? I have reduced my data rate to half and it is still the same.
>
>Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
>the same issue?
>
>Background:
>The XDS510PP+ is tested and manage to work with the DSK when it is flashed
>with post.hex or blink.hex
>The driver used is: sdgo6400_11.dvr
>
>Target: DSK6416, 600MHz
>CCS: V3.00.00
>Flashburn: V2.8 (version 5.90.0.110
>About my code:
>The code uses two edma channels, one doing a ping pong transferring from
>emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
>edma generate HW interrupts (that initiates SW interrupt) which I use to
>preprocess the data in ping pong buffer and dump it into a input buffer and
>post-process the data from an output buffer. These are done with qdma.
>Then, I have a Task that repeatedly process the data in the input buffer
>and place it at the output buffer. I also have a periodic interrupt that
>generates patterns of led display. The input sampling rate is now 500ksps. >
>
> * >----------
Reply by ●July 5, 20052005-07-05
It sounds like you are on the right track. The emulator does need
free cycles to connect. if your application has too much EDMA cycle usage,
the emulator may not be able to connect [it must perform some DMA cycles to
connect].
mikedunn
Koay Teong Beng <k...@arl.nus.edu.sg> wrote:
Koay Teong Beng <k...@arl.nus.edu.sg> wrote:
Mike,
I didn't use startup GEL file in both cases.
Peripheral initializations are all done in my code.
Mike and all,
I start to think along the line that the DSP is too occupied to service the
JTAG. Since my main data processing work is running in a while loop in a
Task, I tried a few things described below but doesn't help, any more ideas?
1) I added a small code into the Task containing my data process algorithm.
The code sense an external pin and run IDL_run() instead of my processing
algorithm upon a pullup to the external pin.
2) I added a small code that jump to another while loop containing only
IDL_run() upon a pullup to an external pin.
3) I break out from the while loop in the Task upon a pullup to an external
pin.
4) I break out from the while loop in the Task and disable all the
interrupts (except PRD function) and EDMA events upon a pullup to an
external pin.
I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
me to reconnect my XDS510PP+, but doesn't work too. :(
Thanks for taking interest.
Best,
Teong Beng
At 01:06 PM 7/5/2005, Mike Dunn wrote:
>
>Are you using a GEL file when you connect in either [or both cases]??
>
>If you are, try connecting without one [if you have a complete working
>program in flash, a GEL file should not be required unless it is to
>support a debug feature.
>
>mikedunn
>
>Koay Teong Beng wrote:
>Dear all,
>
>I have a tested code flashed into the DSK Flash, I power cycled the DSK and
>it executed as I expected. I can connect with it and download symbols for
>debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
>through my XDS510PP+ when my program is flashed and running. Anybody know
>why is it so? I have reduced my data rate to half and it is still the same.
>
>Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
>the same issue?
>
>Background:
>The XDS510PP+ is tested and manage to work with the DSK when it is flashed
>with post.hex or blink.hex
>The driver used is: sdgo6400_11.dvr
>
>Target: DSK6416, 600MHz
>CCS: V3.00.00
>Flashburn: V2.8 (version 5.90.0.110
>About my code:
>The code uses two edma channels, one doing a ping pong transferring from
>emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
>edma generate HW interrupts (that initiates SW interrupt) which I use to
>preprocess the data in ping pong buffer and dump it into a input buffer and
>post-process the data from an output buffer. These are done with qdma.
>Then, I have a Task that repeatedly process the data in the input buffer
>and place it at the output buffer. I also have a periodic interrupt that
>generates patterns of led display. The input sampling rate is now 500ksps.>
>
> *>----------
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/c6x/
<*> To unsubscribe from this group, send an email to:
c...@yahoogroups.com
<*
Reply by ●July 6, 20052005-07-06
Thanks Mike,
I had another few tests along this direction but still no luck. They
include not doing everything else after my self test routine in main(),
which should goto DSP/BIOS idle straight.
There is another observation, if I boot from POST and then download the
.out file of the exact same code (not flashing) into DSK, I am able to
connect to the DSK using XDS510PP+. Now that make me confuse: If it is
purely due to the CPU is too tied up to service my emulator, I should face
the same problem whether the code is being downloaded or is flashed, YET I
can connect with XDS510PP+ if I downloaded the code but not if I flash it.
What is the different between the two process except that there is a boot
loader to load my code from Flash into ISRAM? Anyway, I am using the same
boot.asm used by the POST.pjt in CCS example, which I have no problem
connecting with XDS510PP+.
Thanks in advance,
Teong Beng At 11:52 PM 7/5/2005, Mike Dunn wrote:
>
>It sounds like you are on the right track. The emulator does need free
>cycles to connect. if your application has too much EDMA cycle usage, the
>emulator may not be able to connect [it must perform some DMA cycles to
>connect].
>
>mikedunn
>
>Koay Teong Beng <koay@koay...> wrote:
>Mike,
>
>I didn't use startup GEL file in both cases.
>Peripheral initializations are all done in my code.
>
>Mike and all,
>I start to think along the line that the DSP is too occupied to service the
>JTAG. Since my main data processing work is running in a while loop in a
>Task, I tried a few things described below but doesn't help, any more ideas?
>1) I added a small code into the Task containing my data process algorithm.
>The code sense an external pin and run IDL_run() instead of my processing
>algorithm upon a pullup to the external pin.
>2) I added a small code that jump to another while loop containing only
>IDL_run() upon a pullup to an external pin.
>3) I break out from the while loop in the Task upon a pullup to an external
>pin.
>4) I break out from the while loop in the Task and disable all the
>interrupts (except PRD function) and EDMA events upon a pullup to an
>external pin.
>I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
>me to reconnect my XDS510PP+, but doesn't work too. :(
>
>Thanks for taking interest.
>
>Best,
>Teong Beng
>
>At 01:06 PM 7/5/2005, Mike Dunn wrote:
> >
> >Are you using a GEL file when you connect in either [or both cases]??
> >
> >If you are, try connecting without one [if you have a complete working
> >program in flash, a GEL file should not be required unless it is to
> >support a debug feature.
> >
> >mikedunn
> >
> >Koay Teong Beng wrote:
> >Dear all,
> >
> >I have a tested code flashed into the DSK Flash, I power cycled the DSK and
> >it executed as I expected. I can connect with it and download symbols for
> >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> >through my XDS510PP+ when my program is flashed and running. Anybody know
> >why is it so? I have reduced my data rate to half and it is still the same.
> >
> >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
> >the same issue?
> >
> >Background:
> >The XDS510PP+ is tested and manage to work with the DSK when it is flashed
> >with post.hex or blink.hex
> >The driver used is: sdgo6400_11.dvr
> >
> >Target: DSK6416, 600MHz
> >CCS: V3.00.00
> >Flashburn: V2.8 (version 5.90.0.110
> >About my code:
> >The code uses two edma channels, one doing a ping pong transferring from
> >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> >edma generate HW interrupts (that initiates SW interrupt) which I use to
> >preprocess the data in ping pong buffer and dump it into a input buffer and
> >post-process the data from an output buffer. These are done with qdma.
> >Then, I have a Task that repeatedly process the data in the input buffer
> >and place it at the output buffer. I also have a periodic interrupt that
> >generates patterns of led display. The input sampling rate is now 500ksps.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > *
> >
> >
> >---------- >
>
> * >----------
I had another few tests along this direction but still no luck. They
include not doing everything else after my self test routine in main(),
which should goto DSP/BIOS idle straight.
There is another observation, if I boot from POST and then download the
.out file of the exact same code (not flashing) into DSK, I am able to
connect to the DSK using XDS510PP+. Now that make me confuse: If it is
purely due to the CPU is too tied up to service my emulator, I should face
the same problem whether the code is being downloaded or is flashed, YET I
can connect with XDS510PP+ if I downloaded the code but not if I flash it.
What is the different between the two process except that there is a boot
loader to load my code from Flash into ISRAM? Anyway, I am using the same
boot.asm used by the POST.pjt in CCS example, which I have no problem
connecting with XDS510PP+.
Thanks in advance,
Teong Beng At 11:52 PM 7/5/2005, Mike Dunn wrote:
>
>It sounds like you are on the right track. The emulator does need free
>cycles to connect. if your application has too much EDMA cycle usage, the
>emulator may not be able to connect [it must perform some DMA cycles to
>connect].
>
>mikedunn
>
>Koay Teong Beng <koay@koay...> wrote:
>Mike,
>
>I didn't use startup GEL file in both cases.
>Peripheral initializations are all done in my code.
>
>Mike and all,
>I start to think along the line that the DSP is too occupied to service the
>JTAG. Since my main data processing work is running in a while loop in a
>Task, I tried a few things described below but doesn't help, any more ideas?
>1) I added a small code into the Task containing my data process algorithm.
>The code sense an external pin and run IDL_run() instead of my processing
>algorithm upon a pullup to the external pin.
>2) I added a small code that jump to another while loop containing only
>IDL_run() upon a pullup to an external pin.
>3) I break out from the while loop in the Task upon a pullup to an external
>pin.
>4) I break out from the while loop in the Task and disable all the
>interrupts (except PRD function) and EDMA events upon a pullup to an
>external pin.
>I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
>me to reconnect my XDS510PP+, but doesn't work too. :(
>
>Thanks for taking interest.
>
>Best,
>Teong Beng
>
>At 01:06 PM 7/5/2005, Mike Dunn wrote:
> >
> >Are you using a GEL file when you connect in either [or both cases]??
> >
> >If you are, try connecting without one [if you have a complete working
> >program in flash, a GEL file should not be required unless it is to
> >support a debug feature.
> >
> >mikedunn
> >
> >Koay Teong Beng wrote:
> >Dear all,
> >
> >I have a tested code flashed into the DSK Flash, I power cycled the DSK and
> >it executed as I expected. I can connect with it and download symbols for
> >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> >through my XDS510PP+ when my program is flashed and running. Anybody know
> >why is it so? I have reduced my data rate to half and it is still the same.
> >
> >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
> >the same issue?
> >
> >Background:
> >The XDS510PP+ is tested and manage to work with the DSK when it is flashed
> >with post.hex or blink.hex
> >The driver used is: sdgo6400_11.dvr
> >
> >Target: DSK6416, 600MHz
> >CCS: V3.00.00
> >Flashburn: V2.8 (version 5.90.0.110
> >About my code:
> >The code uses two edma channels, one doing a ping pong transferring from
> >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> >edma generate HW interrupts (that initiates SW interrupt) which I use to
> >preprocess the data in ping pong buffer and dump it into a input buffer and
> >post-process the data from an output buffer. These are done with qdma.
> >Then, I have a Task that repeatedly process the data in the input buffer
> >and place it at the output buffer. I also have a periodic interrupt that
> >generates patterns of led display. The input sampling rate is now 500ksps.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > *
> >
> >
> >---------- >
>
> * >----------
Reply by ●July 6, 20052005-07-06
Teong Beng,
Your problem seems a bit unusual... have you tried holding down the
reset button and letting it go immediately after you click "connect"??
Please keep us posted with anything that you find out.
mikedunn
Koay Teong Beng <k...@arl.nus.edu.sg> wrote:
Koay Teong Beng <k...@arl.nus.edu.sg> wrote:
Thanks Mike,
I had another few tests along this direction but still no luck. They
include not doing everything else after my self test routine in main(),
which should goto DSP/BIOS idle straight.
There is another observation, if I boot from POST and then download the
.out file of the exact same code (not flashing) into DSK, I am able to
connect to the DSK using XDS510PP+. Now that make me confuse: If it is
purely due to the CPU is too tied up to service my emulator, I should face
the same problem whether the code is being downloaded or is flashed, YET I
can connect with XDS510PP+ if I downloaded the code but not if I flash it.
What is the different between the two process except that there is a boot
loader to load my code from Flash into ISRAM? Anyway, I am using the same
boot.asm used by the POST.pjt in CCS example, which I have no problem
connecting with XDS510PP+.
Thanks in advance,
Teong BengAt 11:52 PM 7/5/2005, Mike Dunn wrote:
>
>It sounds like you are on the right track. The emulator does need free
>cycles to connect. if your application has too much EDMA cycle usage, the
>emulator may not be able to connect [it must perform some DMA cycles to
>connect].
>
>mikedunn
>
>Koay Teong Beng wrote:
>Mike,
>
>I didn't use startup GEL file in both cases.
>Peripheral initializations are all done in my code.
>
>Mike and all,
>I start to think along the line that the DSP is too occupied to service the
>JTAG. Since my main data processing work is running in a while loop in a
>Task, I tried a few things described below but doesn't help, any more ideas?
>1) I added a small code into the Task containing my data process algorithm.
>The code sense an external pin and run IDL_run() instead of my processing
>algorithm upon a pullup to the external pin.
>2) I added a small code that jump to another while loop containing only
>IDL_run() upon a pullup to an external pin.
>3) I break out from the while loop in the Task upon a pullup to an external
>pin.
>4) I break out from the while loop in the Task and disable all the
>interrupts (except PRD function) and EDMA events upon a pullup to an
>external pin.
>I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
>me to reconnect my XDS510PP+, but doesn't work too. :(
>
>Thanks for taking interest.
>
>Best,
>Teong Beng
>
>At 01:06 PM 7/5/2005, Mike Dunn wrote:
> >
> >Are you using a GEL file when you connect in either [or both cases]??
> >
> >If you are, try connecting without one [if you have a complete working
> >program in flash, a GEL file should not be required unless it is to
> >support a debug feature.
> >
> >mikedunn
> >
> >Koay Teong Beng wrote:
> >Dear all,
> >
> >I have a tested code flashed into the DSK Flash, I power cycled the DSK and
> >it executed as I expected. I can connect with it and download symbols for
> >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> >through my XDS510PP+ when my program is flashed and running. Anybody know
> >why is it so? I have reduced my data rate to half and it is still the same.
> >
> >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
> >the same issue?
> >
> >Background:
> >The XDS510PP+ is tested and manage to work with the DSK when it is flashed
> >with post.hex or blink.hex
> >The driver used is: sdgo6400_11.dvr
> >
> >Target: DSK6416, 600MHz
> >CCS: V3.00.00
> >Flashburn: V2.8 (version 5.90.0.110
> >About my code:
> >The code uses two edma channels, one doing a ping pong transferring from
> >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> >edma generate HW interrupts (that initiates SW interrupt) which I use to
> >preprocess the data in ping pong buffer and dump it into a input buffer and
> >post-process the data from an output buffer. These are done with qdma.
> >Then, I have a Task that repeatedly process the data in the input buffer
> >and place it at the output buffer. I also have a periodic interrupt that
> >generates patterns of led display. The input sampling rate is now 500ksps.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > *
> >
> >
> >---------->
>
> *>----------
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/c6x/
<*> To unsubscribe from this group, send an email to:
c...@yahoogroups.com
<*
Reply by ●July 7, 20052005-07-07
Mike,
Thank you! Holding the reset and release while connect works to reconnect
to it with XDS510PP+, and I manage to flash new codes into the DSK, even my
prototype target!!
Also, disabling RTDX in DSP/BIOS in my target code enables me to connect to
it without having to do the above. Thanks to TI that suggested this. RTDX
engine must have been doing something that prevented the connection, and
your solution give me time to connect.
I can perform flashburn into my target a few times through XDS510PP+, until
I have the POST burned in my target. Then I have problem connecting the
Flashburn.
My process:
1) connect XDS510PP+ to target
2) power XDS510PP+
3) power on target
4) launch CCS
5) connect CCS to target (success)
6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
Flashburn shows that it fail to
connect)
Now step 6) works a few times until I have POST flashed into it. :( haven't
succeed since then.
There is an error message when I try to do 'reset CPU' from CCS,
Failed Software reset: Error 0x00000022/-1147 Error during: memory,
execution, it appears that the target is being held in
reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
I am now working on this issue now. Any idea? Best,Teong Beng At 09:13 PM 7/6/2005, Mike Dunn wrote:
>Teong Beng,
>
>Your problem seems a bit unusual... have you tried holding down the reset
>button and letting it go immediately after you click "connect"??
>
>Please keep us posted with anything that you find out.
>
>mikedunn
>
>Koay Teong Beng <koay@koay...> wrote:
>Thanks Mike,
>I had another few tests along this direction but still no luck. They
>include not doing everything else after my self test routine in main(),
>which should goto DSP/BIOS idle straight.
>There is another observation, if I boot from POST and then download the
>.out file of the exact same code (not flashing) into DSK, I am able to
>connect to the DSK using XDS510PP+. Now that make me confuse: If it is
>purely due to the CPU is too tied up to service my emulator, I should face
>the same problem whether the code is being downloaded or is flashed, YET I
>can connect with XDS510PP+ if I downloaded the code but not if I flash it.
>What is the different between the two process except that there is a boot
>loader to load my code from Flash into ISRAM? Anyway, I am using the same
>boot.asm used by the POST.pjt in CCS example, which I have no problem
>connecting with XDS510PP+.
>Thanks in advance,
>Teong Beng
>
>At 11:52 PM 7/5/2005, Mike Dunn wrote:
> >
> >It sounds like you are on the right track. The emulator does need free
> >cycles to connect. if your application has too much EDMA cycle usage, the
> >emulator may not be able to connect [it must perform some DMA cycles to
> >connect].
> >
> >mikedunn
> >
> >Koay Teong Beng wrote:
> >Mike,
> >
> >I didn't use startup GEL file in both cases.
> >Peripheral initializations are all done in my code.
> >
> >Mike and all,
> >I start to think along the line that the DSP is too occupied to service the
> >JTAG. Since my main data processing work is running in a while loop in a
> >Task, I tried a few things described below but doesn't help, any more
> ideas?
> >1) I added a small code into the Task containing my data process algorithm.
> >The code sense an external pin and run IDL_run() instead of my processing
> >algorithm upon a pullup to the external pin.
> >2) I added a small code that jump to another while loop containing only
> >IDL_run() upon a pullup to an external pin.
> >3) I break out from the while loop in the Task upon a pullup to an external
> >pin.
> >4) I break out from the while loop in the Task and disable all the
> >interrupts (except PRD function) and EDMA events upon a pullup to an
> >external pin.
> >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
> >me to reconnect my XDS510PP+, but doesn't work too. :(
> >
> >Thanks for taking interest.
> >
> >Best,
> >Teong Beng
> >
> >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > >
> > >Are you using a GEL file when you connect in either [or both cases]??
> > >
> > >If you are, try connecting without one [if you have a complete working
> > >program in flash, a GEL file should not be required unless it is to
> > >support a debug feature.
> > >
> > >mikedunn
> > >
> > >Koay Teong Beng wrote:
> > >Dear all,
> > >
> > >I have a tested code flashed into the DSK Flash, I power cycled the
> DSK and
> > >it executed as I expected. I can connect with it and download symbols for
> > >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> > >through my XDS510PP+ when my program is flashed and running. Anybody know
> > >why is it so? I have reduced my data rate to half and it is still the
> same.
> > >
> > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
> > >the same issue?
> > >
> > >Background:
> > >The XDS510PP+ is tested and manage to work with the DSK when it is
> flashed
> > >with post.hex or blink.hex
> > >The driver used is: sdgo6400_11.dvr
> > >
> > >Target: DSK6416, 600MHz
> > >CCS: V3.00.00
> > >Flashburn: V2.8 (version 5.90.0.110
> > >About my code:
> > >The code uses two edma channels, one doing a ping pong transferring from
> > >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> > >edma generate HW interrupts (that initiates SW interrupt) which I use to
> > >preprocess the data in ping pong buffer and dump it into a input
> buffer and
> > >post-process the data from an output buffer. These are done with qdma.
> > >Then, I have a Task that repeatedly process the data in the input buffer
> > >and place it at the output buffer. I also have a periodic interrupt that
> > >generates patterns of led display. The input sampling rate is now
> 500ksps.
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > *
> > >
> > >
> > >----------
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > *
> >
> >
> >---------- >
Thank you! Holding the reset and release while connect works to reconnect
to it with XDS510PP+, and I manage to flash new codes into the DSK, even my
prototype target!!
Also, disabling RTDX in DSP/BIOS in my target code enables me to connect to
it without having to do the above. Thanks to TI that suggested this. RTDX
engine must have been doing something that prevented the connection, and
your solution give me time to connect.
I can perform flashburn into my target a few times through XDS510PP+, until
I have the POST burned in my target. Then I have problem connecting the
Flashburn.
My process:
1) connect XDS510PP+ to target
2) power XDS510PP+
3) power on target
4) launch CCS
5) connect CCS to target (success)
6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
Flashburn shows that it fail to
connect)
Now step 6) works a few times until I have POST flashed into it. :( haven't
succeed since then.
There is an error message when I try to do 'reset CPU' from CCS,
Failed Software reset: Error 0x00000022/-1147 Error during: memory,
execution, it appears that the target is being held in
reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
I am now working on this issue now. Any idea? Best,Teong Beng At 09:13 PM 7/6/2005, Mike Dunn wrote:
>Teong Beng,
>
>Your problem seems a bit unusual... have you tried holding down the reset
>button and letting it go immediately after you click "connect"??
>
>Please keep us posted with anything that you find out.
>
>mikedunn
>
>Koay Teong Beng <koay@koay...> wrote:
>Thanks Mike,
>I had another few tests along this direction but still no luck. They
>include not doing everything else after my self test routine in main(),
>which should goto DSP/BIOS idle straight.
>There is another observation, if I boot from POST and then download the
>.out file of the exact same code (not flashing) into DSK, I am able to
>connect to the DSK using XDS510PP+. Now that make me confuse: If it is
>purely due to the CPU is too tied up to service my emulator, I should face
>the same problem whether the code is being downloaded or is flashed, YET I
>can connect with XDS510PP+ if I downloaded the code but not if I flash it.
>What is the different between the two process except that there is a boot
>loader to load my code from Flash into ISRAM? Anyway, I am using the same
>boot.asm used by the POST.pjt in CCS example, which I have no problem
>connecting with XDS510PP+.
>Thanks in advance,
>Teong Beng
>
>At 11:52 PM 7/5/2005, Mike Dunn wrote:
> >
> >It sounds like you are on the right track. The emulator does need free
> >cycles to connect. if your application has too much EDMA cycle usage, the
> >emulator may not be able to connect [it must perform some DMA cycles to
> >connect].
> >
> >mikedunn
> >
> >Koay Teong Beng wrote:
> >Mike,
> >
> >I didn't use startup GEL file in both cases.
> >Peripheral initializations are all done in my code.
> >
> >Mike and all,
> >I start to think along the line that the DSP is too occupied to service the
> >JTAG. Since my main data processing work is running in a while loop in a
> >Task, I tried a few things described below but doesn't help, any more
> ideas?
> >1) I added a small code into the Task containing my data process algorithm.
> >The code sense an external pin and run IDL_run() instead of my processing
> >algorithm upon a pullup to the external pin.
> >2) I added a small code that jump to another while loop containing only
> >IDL_run() upon a pullup to an external pin.
> >3) I break out from the while loop in the Task upon a pullup to an external
> >pin.
> >4) I break out from the while loop in the Task and disable all the
> >interrupts (except PRD function) and EDMA events upon a pullup to an
> >external pin.
> >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and allow
> >me to reconnect my XDS510PP+, but doesn't work too. :(
> >
> >Thanks for taking interest.
> >
> >Best,
> >Teong Beng
> >
> >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > >
> > >Are you using a GEL file when you connect in either [or both cases]??
> > >
> > >If you are, try connecting without one [if you have a complete working
> > >program in flash, a GEL file should not be required unless it is to
> > >support a debug feature.
> > >
> > >mikedunn
> > >
> > >Koay Teong Beng wrote:
> > >Dear all,
> > >
> > >I have a tested code flashed into the DSK Flash, I power cycled the
> DSK and
> > >it executed as I expected. I can connect with it and download symbols for
> > >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> > >through my XDS510PP+ when my program is flashed and running. Anybody know
> > >why is it so? I have reduced my data rate to half and it is still the
> same.
> > >
> > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody face
> > >the same issue?
> > >
> > >Background:
> > >The XDS510PP+ is tested and manage to work with the DSK when it is
> flashed
> > >with post.hex or blink.hex
> > >The driver used is: sdgo6400_11.dvr
> > >
> > >Target: DSK6416, 600MHz
> > >CCS: V3.00.00
> > >Flashburn: V2.8 (version 5.90.0.110
> > >About my code:
> > >The code uses two edma channels, one doing a ping pong transferring from
> > >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> > >edma generate HW interrupts (that initiates SW interrupt) which I use to
> > >preprocess the data in ping pong buffer and dump it into a input
> buffer and
> > >post-process the data from an output buffer. These are done with qdma.
> > >Then, I have a Task that repeatedly process the data in the input buffer
> > >and place it at the output buffer. I also have a periodic interrupt that
> > >generates patterns of led display. The input sampling rate is now
> 500ksps.
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > *
> > >
> > >
> > >----------
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > *
> >
> >
> >---------- >
Reply by ●July 8, 20052005-07-08
More update,
I can now download different programs into the DSK (and into my prototype -
which is based on the DSK6416 excluding USB emulator) and execute them
through XDS510PP+. But I am still not able to connect to Flashburn, it just
stay disconnected.
I also verified from the disassembly window that the FBTC program is
downloaded and executed, after I launch the cdd configuration. Just that
the Flashburn program indicate that it is not connected and the flash
burning/erasing buttons are disabled.
I tried to use Flashburn from command line, also unsuccessful.
Yet, after a number of trials, I sometimes manage to connect Flashburn with
DSK, sometimes not. But never to my target. I start to think that it could
related to electrical quality. Any input?
The failed software reset issue still remains.
Any input are very welcomed.
Thanks in advance,
Teong Beng
At 07:51 PM 7/7/2005, Koay Teong Beng wrote:
>Mike,
>
>Thank you! Holding the reset and release while connect works to reconnect
>to it with XDS510PP+, and I manage to flash new codes into the DSK, even my
>prototype target!!
>Also, disabling RTDX in DSP/BIOS in my target code enables me to connect to
>it without having to do the above. Thanks to TI that suggested this. RTDX
>engine must have been doing something that prevented the connection, and
>your solution give me time to connect.
>
>I can perform flashburn into my target a few times through XDS510PP+, until
>I have the POST burned in my target. Then I have problem connecting the
>Flashburn.
>
>My process:
>1) connect XDS510PP+ to target
>2) power XDS510PP+
>3) power on target
>4) launch CCS
>5) connect CCS to target (success)
>6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
>Flashburn shows that it fail to
>connect)
>
>Now step 6) works a few times until I have POST flashed into it. :( haven't
>succeed since then.
>
>There is an error message when I try to do 'reset CPU' from CCS,
>Failed Software reset: Error 0x00000022/-1147 Error during: memory,
>execution, it appears that the target is being held in
>reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
>
>I am now working on this issue now. Any idea? >Best,Teong Beng >At 09:13 PM 7/6/2005, Mike Dunn wrote:
> >Teong Beng,
> >
> >Your problem seems a bit unusual... have you tried holding down the reset
> >button and letting it go immediately after you click "connect"??
> >
> >Please keep us posted with anything that you find out.
> >
> >mikedunn
> >
> >Koay Teong Beng <koay@koay...> wrote:
> >Thanks Mike,
> >I had another few tests along this direction but still no luck. They
> >include not doing everything else after my self test routine in main(),
> >which should goto DSP/BIOS idle straight.
> >There is another observation, if I boot from POST and then download the
> >.out file of the exact same code (not flashing) into DSK, I am able to
> >connect to the DSK using XDS510PP+. Now that make me confuse: If it is
> >purely due to the CPU is too tied up to service my emulator, I should face
> >the same problem whether the code is being downloaded or is flashed, YET I
> >can connect with XDS510PP+ if I downloaded the code but not if I flash it.
> >What is the different between the two process except that there is a boot
> >loader to load my code from Flash into ISRAM? Anyway, I am using the same
> >boot.asm used by the POST.pjt in CCS example, which I have no problem
> >connecting with XDS510PP+.
> >Thanks in advance,
> >Teong Beng
> >
> >At 11:52 PM 7/5/2005, Mike Dunn wrote:
> > >
> > >It sounds like you are on the right track. The emulator does need free
> > >cycles to connect. if your application has too much EDMA cycle usage, the
> > >emulator may not be able to connect [it must perform some DMA cycles to
> > >connect].
> > >
> > >mikedunn
> > >
> > >Koay Teong Beng wrote:
> > >Mike,
> > >
> > >I didn't use startup GEL file in both cases.
> > >Peripheral initializations are all done in my code.
> > >
> > >Mike and all,
> > >I start to think along the line that the DSP is too occupied to
> service the
> > >JTAG. Since my main data processing work is running in a while loop in a
> > >Task, I tried a few things described below but doesn't help, any more
> > ideas?
> > >1) I added a small code into the Task containing my data process
> algorithm.
> > >The code sense an external pin and run IDL_run() instead of my processing
> > >algorithm upon a pullup to the external pin.
> > >2) I added a small code that jump to another while loop containing only
> > >IDL_run() upon a pullup to an external pin.
> > >3) I break out from the while loop in the Task upon a pullup to an
> external
> > >pin.
> > >4) I break out from the while loop in the Task and disable all the
> > >interrupts (except PRD function) and EDMA events upon a pullup to an
> > >external pin.
> > >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and
> allow
> > >me to reconnect my XDS510PP+, but doesn't work too. :(
> > >
> > >Thanks for taking interest.
> > >
> > >Best,
> > >Teong Beng
> > >
> > >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > > >
> > > >Are you using a GEL file when you connect in either [or both cases]??
> > > >
> > > >If you are, try connecting without one [if you have a complete working
> > > >program in flash, a GEL file should not be required unless it is to
> > > >support a debug feature.
> > > >
> > > >mikedunn
> > > >
> > > >Koay Teong Beng wrote:
> > > >Dear all,
> > > >
> > > >I have a tested code flashed into the DSK Flash, I power cycled the
> > DSK and
> > > >it executed as I expected. I can connect with it and download
> symbols for
> > > >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> > > >through my XDS510PP+ when my program is flashed and running. Anybody
> know
> > > >why is it so? I have reduced my data rate to half and it is still the
> > same.
> > > >
> > > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody
> face
> > > >the same issue?
> > > >
> > > >Background:
> > > >The XDS510PP+ is tested and manage to work with the DSK when it is
> > flashed
> > > >with post.hex or blink.hex
> > > >The driver used is: sdgo6400_11.dvr
> > > >
> > > >Target: DSK6416, 600MHz
> > > >CCS: V3.00.00
> > > >Flashburn: V2.8 (version 5.90.0.110
> > > >About my code:
> > > >The code uses two edma channels, one doing a ping pong transferring from
> > > >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> > > >edma generate HW interrupts (that initiates SW interrupt) which I use to
> > > >preprocess the data in ping pong buffer and dump it into a input
> > buffer and
> > > >post-process the data from an output buffer. These are done with qdma.
> > > >Then, I have a Task that repeatedly process the data in the input buffer
> > > >and place it at the output buffer. I also have a periodic interrupt that
> > > >generates patterns of led display. The input sampling rate is now
> > 500ksps.
I can now download different programs into the DSK (and into my prototype -
which is based on the DSK6416 excluding USB emulator) and execute them
through XDS510PP+. But I am still not able to connect to Flashburn, it just
stay disconnected.
I also verified from the disassembly window that the FBTC program is
downloaded and executed, after I launch the cdd configuration. Just that
the Flashburn program indicate that it is not connected and the flash
burning/erasing buttons are disabled.
I tried to use Flashburn from command line, also unsuccessful.
Yet, after a number of trials, I sometimes manage to connect Flashburn with
DSK, sometimes not. But never to my target. I start to think that it could
related to electrical quality. Any input?
The failed software reset issue still remains.
Any input are very welcomed.
Thanks in advance,
Teong Beng
At 07:51 PM 7/7/2005, Koay Teong Beng wrote:
>Mike,
>
>Thank you! Holding the reset and release while connect works to reconnect
>to it with XDS510PP+, and I manage to flash new codes into the DSK, even my
>prototype target!!
>Also, disabling RTDX in DSP/BIOS in my target code enables me to connect to
>it without having to do the above. Thanks to TI that suggested this. RTDX
>engine must have been doing something that prevented the connection, and
>your solution give me time to connect.
>
>I can perform flashburn into my target a few times through XDS510PP+, until
>I have the POST burned in my target. Then I have problem connecting the
>Flashburn.
>
>My process:
>1) connect XDS510PP+ to target
>2) power XDS510PP+
>3) power on target
>4) launch CCS
>5) connect CCS to target (success)
>6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
>Flashburn shows that it fail to
>connect)
>
>Now step 6) works a few times until I have POST flashed into it. :( haven't
>succeed since then.
>
>There is an error message when I try to do 'reset CPU' from CCS,
>Failed Software reset: Error 0x00000022/-1147 Error during: memory,
>execution, it appears that the target is being held in
>reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
>
>I am now working on this issue now. Any idea? >Best,Teong Beng >At 09:13 PM 7/6/2005, Mike Dunn wrote:
> >Teong Beng,
> >
> >Your problem seems a bit unusual... have you tried holding down the reset
> >button and letting it go immediately after you click "connect"??
> >
> >Please keep us posted with anything that you find out.
> >
> >mikedunn
> >
> >Koay Teong Beng <koay@koay...> wrote:
> >Thanks Mike,
> >I had another few tests along this direction but still no luck. They
> >include not doing everything else after my self test routine in main(),
> >which should goto DSP/BIOS idle straight.
> >There is another observation, if I boot from POST and then download the
> >.out file of the exact same code (not flashing) into DSK, I am able to
> >connect to the DSK using XDS510PP+. Now that make me confuse: If it is
> >purely due to the CPU is too tied up to service my emulator, I should face
> >the same problem whether the code is being downloaded or is flashed, YET I
> >can connect with XDS510PP+ if I downloaded the code but not if I flash it.
> >What is the different between the two process except that there is a boot
> >loader to load my code from Flash into ISRAM? Anyway, I am using the same
> >boot.asm used by the POST.pjt in CCS example, which I have no problem
> >connecting with XDS510PP+.
> >Thanks in advance,
> >Teong Beng
> >
> >At 11:52 PM 7/5/2005, Mike Dunn wrote:
> > >
> > >It sounds like you are on the right track. The emulator does need free
> > >cycles to connect. if your application has too much EDMA cycle usage, the
> > >emulator may not be able to connect [it must perform some DMA cycles to
> > >connect].
> > >
> > >mikedunn
> > >
> > >Koay Teong Beng wrote:
> > >Mike,
> > >
> > >I didn't use startup GEL file in both cases.
> > >Peripheral initializations are all done in my code.
> > >
> > >Mike and all,
> > >I start to think along the line that the DSP is too occupied to
> service the
> > >JTAG. Since my main data processing work is running in a while loop in a
> > >Task, I tried a few things described below but doesn't help, any more
> > ideas?
> > >1) I added a small code into the Task containing my data process
> algorithm.
> > >The code sense an external pin and run IDL_run() instead of my processing
> > >algorithm upon a pullup to the external pin.
> > >2) I added a small code that jump to another while loop containing only
> > >IDL_run() upon a pullup to an external pin.
> > >3) I break out from the while loop in the Task upon a pullup to an
> external
> > >pin.
> > >4) I break out from the while loop in the Task and disable all the
> > >interrupts (except PRD function) and EDMA events upon a pullup to an
> > >external pin.
> > >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and
> allow
> > >me to reconnect my XDS510PP+, but doesn't work too. :(
> > >
> > >Thanks for taking interest.
> > >
> > >Best,
> > >Teong Beng
> > >
> > >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > > >
> > > >Are you using a GEL file when you connect in either [or both cases]??
> > > >
> > > >If you are, try connecting without one [if you have a complete working
> > > >program in flash, a GEL file should not be required unless it is to
> > > >support a debug feature.
> > > >
> > > >mikedunn
> > > >
> > > >Koay Teong Beng wrote:
> > > >Dear all,
> > > >
> > > >I have a tested code flashed into the DSK Flash, I power cycled the
> > DSK and
> > > >it executed as I expected. I can connect with it and download
> symbols for
> > > >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> > > >through my XDS510PP+ when my program is flashed and running. Anybody
> know
> > > >why is it so? I have reduced my data rate to half and it is still the
> > same.
> > > >
> > > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody
> face
> > > >the same issue?
> > > >
> > > >Background:
> > > >The XDS510PP+ is tested and manage to work with the DSK when it is
> > flashed
> > > >with post.hex or blink.hex
> > > >The driver used is: sdgo6400_11.dvr
> > > >
> > > >Target: DSK6416, 600MHz
> > > >CCS: V3.00.00
> > > >Flashburn: V2.8 (version 5.90.0.110
> > > >About my code:
> > > >The code uses two edma channels, one doing a ping pong transferring from
> > > >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> > > >edma generate HW interrupts (that initiates SW interrupt) which I use to
> > > >preprocess the data in ping pong buffer and dump it into a input
> > buffer and
> > > >post-process the data from an output buffer. These are done with qdma.
> > > >Then, I have a Task that repeatedly process the data in the input buffer
> > > >and place it at the output buffer. I also have a periodic interrupt that
> > > >generates patterns of led display. The input sampling rate is now
> > 500ksps.
Reply by ●July 8, 20052005-07-08
Koay Teong Beng-
> I can now download different programs into the DSK (and into my prototype -
> which is based on the DSK6416 excluding USB emulator) and execute them
> through XDS510PP+. But I am still not able to connect to Flashburn, it just
> stay disconnected.
>
> I also verified from the disassembly window that the FBTC program is
> downloaded and executed, after I launch the cdd configuration. Just that
> the Flashburn program indicate that it is not connected and the flash
> burning/erasing buttons are disabled.
> I tried to use Flashburn from command line, also unsuccessful.
>
> Yet, after a number of trials, I sometimes manage to connect Flashburn with
> DSK, sometimes not. But never to my target. I start to think that it could
> related to electrical quality. Any input?
If you can connect XDS510PP+ to your prototype board to download code, then there is
no electrical problem. JTAG for function A is same as JTAG for function B.
If there is any difference between your prototype EMIF/memory map and DSK C6416
board, then you have to modify low-level Flash device-driver in Flashburn code.
-Jeff
> At 07:51 PM 7/7/2005, Koay Teong Beng wrote:
> >Mike,
> >
> >Thank you! Holding the reset and release while connect works to reconnect
> >to it with XDS510PP+, and I manage to flash new codes into the DSK, even my
> >prototype target!!
> >Also, disabling RTDX in DSP/BIOS in my target code enables me to connect to
> >it without having to do the above. Thanks to TI that suggested this. RTDX
> >engine must have been doing something that prevented the connection, and
> >your solution give me time to connect.
> >
> >I can perform flashburn into my target a few times through XDS510PP+, until
> >I have the POST burned in my target. Then I have problem connecting the
> >Flashburn.
> >
> >My process:
> >1) connect XDS510PP+ to target
> >2) power XDS510PP+
> >3) power on target
> >4) launch CCS
> >5) connect CCS to target (success)
> >6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
> >Flashburn shows that it fail to
> >connect)
> >
> >Now step 6) works a few times until I have POST flashed into it. :( haven't
> >succeed since then.
> >
> >There is an error message when I try to do 'reset CPU' from CCS,
> >Failed Software reset: Error 0x00000022/-1147 Error during: memory,
> >execution, it appears that the target is being held in
> >reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
> >
> >I am now working on this issue now. Any idea?
> >
> >
> >Best,Teong Beng
> >
> >
> >At 09:13 PM 7/6/2005, Mike Dunn wrote:
> > >Teong Beng,
> > >
> > >Your problem seems a bit unusual... have you tried holding down the reset
> > >button and letting it go immediately after you click "connect"??
> > >
> > >Please keep us posted with anything that you find out.
> > >
> > >mikedunn
> > >
> > >Koay Teong Beng <koay@koay...> wrote:
> > >Thanks Mike,
> > >I had another few tests along this direction but still no luck. They
> > >include not doing everything else after my self test routine in main(),
> > >which should goto DSP/BIOS idle straight.
> > >There is another observation, if I boot from POST and then download the
> > >.out file of the exact same code (not flashing) into DSK, I am able to
> > >connect to the DSK using XDS510PP+. Now that make me confuse: If it is
> > >purely due to the CPU is too tied up to service my emulator, I should face
> > >the same problem whether the code is being downloaded or is flashed, YET I
> > >can connect with XDS510PP+ if I downloaded the code but not if I flash it.
> > >What is the different between the two process except that there is a boot
> > >loader to load my code from Flash into ISRAM? Anyway, I am using the same
> > >boot.asm used by the POST.pjt in CCS example, which I have no problem
> > >connecting with XDS510PP+.
> > >Thanks in advance,
> > >Teong Beng
> > >
> > >At 11:52 PM 7/5/2005, Mike Dunn wrote:
> > > >
> > > >It sounds like you are on the right track. The emulator does need free
> > > >cycles to connect. if your application has too much EDMA cycle usage, the
> > > >emulator may not be able to connect [it must perform some DMA cycles to
> > > >connect].
> > > >
> > > >mikedunn
> > > >
> > > >Koay Teong Beng wrote:
> > > >Mike,
> > > >
> > > >I didn't use startup GEL file in both cases.
> > > >Peripheral initializations are all done in my code.
> > > >
> > > >Mike and all,
> > > >I start to think along the line that the DSP is too occupied to
> > service the
> > > >JTAG. Since my main data processing work is running in a while loop in a
> > > >Task, I tried a few things described below but doesn't help, any more
> > > ideas?
> > > >1) I added a small code into the Task containing my data process
> > algorithm.
> > > >The code sense an external pin and run IDL_run() instead of my processing
> > > >algorithm upon a pullup to the external pin.
> > > >2) I added a small code that jump to another while loop containing only
> > > >IDL_run() upon a pullup to an external pin.
> > > >3) I break out from the while loop in the Task upon a pullup to an
> > external
> > > >pin.
> > > >4) I break out from the while loop in the Task and disable all the
> > > >interrupts (except PRD function) and EDMA events upon a pullup to an
> > > >external pin.
> > > >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and
> > allow
> > > >me to reconnect my XDS510PP+, but doesn't work too. :(
> > > >
> > > >Thanks for taking interest.
> > > >
> > > >Best,
> > > >Teong Beng
> > > >
> > > >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > > > >
> > > > >Are you using a GEL file when you connect in either [or both cases]??
> > > > >
> > > > >If you are, try connecting without one [if you have a complete working
> > > > >program in flash, a GEL file should not be required unless it is to
> > > > >support a debug feature.
> > > > >
> > > > >mikedunn
> > > > >
> > > > >Koay Teong Beng wrote:
> > > > >Dear all,
> > > > >
> > > > >I have a tested code flashed into the DSK Flash, I power cycled the
> > > DSK and
> > > > >it executed as I expected. I can connect with it and download
> > symbols for
> > > > >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> > > > >through my XDS510PP+ when my program is flashed and running. Anybody
> > know
> > > > >why is it so? I have reduced my data rate to half and it is still the
> > > same.
> > > > >
> > > > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody
> > face
> > > > >the same issue?
> > > > >
> > > > >Background:
> > > > >The XDS510PP+ is tested and manage to work with the DSK when it is
> > > flashed
> > > > >with post.hex or blink.hex
> > > > >The driver used is: sdgo6400_11.dvr
> > > > >
> > > > >Target: DSK6416, 600MHz
> > > > >CCS: V3.00.00
> > > > >Flashburn: V2.8 (version 5.90.0.110
> > > > >About my code:
> > > > >The code uses two edma channels, one doing a ping pong transferring from
> > > > >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> > > > >edma generate HW interrupts (that initiates SW interrupt) which I use to
> > > > >preprocess the data in ping pong buffer and dump it into a input
> > > buffer and
> > > > >post-process the data from an output buffer. These are done with qdma.
> > > > >Then, I have a Task that repeatedly process the data in the input buffer
> > > > >and place it at the output buffer. I also have a periodic interrupt that
> > > > >generates patterns of led display. The input sampling rate is now
> > > 500ksps. >
> I can now download different programs into the DSK (and into my prototype -
> which is based on the DSK6416 excluding USB emulator) and execute them
> through XDS510PP+. But I am still not able to connect to Flashburn, it just
> stay disconnected.
>
> I also verified from the disassembly window that the FBTC program is
> downloaded and executed, after I launch the cdd configuration. Just that
> the Flashburn program indicate that it is not connected and the flash
> burning/erasing buttons are disabled.
> I tried to use Flashburn from command line, also unsuccessful.
>
> Yet, after a number of trials, I sometimes manage to connect Flashburn with
> DSK, sometimes not. But never to my target. I start to think that it could
> related to electrical quality. Any input?
If you can connect XDS510PP+ to your prototype board to download code, then there is
no electrical problem. JTAG for function A is same as JTAG for function B.
If there is any difference between your prototype EMIF/memory map and DSK C6416
board, then you have to modify low-level Flash device-driver in Flashburn code.
-Jeff
> At 07:51 PM 7/7/2005, Koay Teong Beng wrote:
> >Mike,
> >
> >Thank you! Holding the reset and release while connect works to reconnect
> >to it with XDS510PP+, and I manage to flash new codes into the DSK, even my
> >prototype target!!
> >Also, disabling RTDX in DSP/BIOS in my target code enables me to connect to
> >it without having to do the above. Thanks to TI that suggested this. RTDX
> >engine must have been doing something that prevented the connection, and
> >your solution give me time to connect.
> >
> >I can perform flashburn into my target a few times through XDS510PP+, until
> >I have the POST burned in my target. Then I have problem connecting the
> >Flashburn.
> >
> >My process:
> >1) connect XDS510PP+ to target
> >2) power XDS510PP+
> >3) power on target
> >4) launch CCS
> >5) connect CCS to target (success)
> >6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
> >Flashburn shows that it fail to
> >connect)
> >
> >Now step 6) works a few times until I have POST flashed into it. :( haven't
> >succeed since then.
> >
> >There is an error message when I try to do 'reset CPU' from CCS,
> >Failed Software reset: Error 0x00000022/-1147 Error during: memory,
> >execution, it appears that the target is being held in
> >reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
> >
> >I am now working on this issue now. Any idea?
> >
> >
> >Best,Teong Beng
> >
> >
> >At 09:13 PM 7/6/2005, Mike Dunn wrote:
> > >Teong Beng,
> > >
> > >Your problem seems a bit unusual... have you tried holding down the reset
> > >button and letting it go immediately after you click "connect"??
> > >
> > >Please keep us posted with anything that you find out.
> > >
> > >mikedunn
> > >
> > >Koay Teong Beng <koay@koay...> wrote:
> > >Thanks Mike,
> > >I had another few tests along this direction but still no luck. They
> > >include not doing everything else after my self test routine in main(),
> > >which should goto DSP/BIOS idle straight.
> > >There is another observation, if I boot from POST and then download the
> > >.out file of the exact same code (not flashing) into DSK, I am able to
> > >connect to the DSK using XDS510PP+. Now that make me confuse: If it is
> > >purely due to the CPU is too tied up to service my emulator, I should face
> > >the same problem whether the code is being downloaded or is flashed, YET I
> > >can connect with XDS510PP+ if I downloaded the code but not if I flash it.
> > >What is the different between the two process except that there is a boot
> > >loader to load my code from Flash into ISRAM? Anyway, I am using the same
> > >boot.asm used by the POST.pjt in CCS example, which I have no problem
> > >connecting with XDS510PP+.
> > >Thanks in advance,
> > >Teong Beng
> > >
> > >At 11:52 PM 7/5/2005, Mike Dunn wrote:
> > > >
> > > >It sounds like you are on the right track. The emulator does need free
> > > >cycles to connect. if your application has too much EDMA cycle usage, the
> > > >emulator may not be able to connect [it must perform some DMA cycles to
> > > >connect].
> > > >
> > > >mikedunn
> > > >
> > > >Koay Teong Beng wrote:
> > > >Mike,
> > > >
> > > >I didn't use startup GEL file in both cases.
> > > >Peripheral initializations are all done in my code.
> > > >
> > > >Mike and all,
> > > >I start to think along the line that the DSP is too occupied to
> > service the
> > > >JTAG. Since my main data processing work is running in a while loop in a
> > > >Task, I tried a few things described below but doesn't help, any more
> > > ideas?
> > > >1) I added a small code into the Task containing my data process
> > algorithm.
> > > >The code sense an external pin and run IDL_run() instead of my processing
> > > >algorithm upon a pullup to the external pin.
> > > >2) I added a small code that jump to another while loop containing only
> > > >IDL_run() upon a pullup to an external pin.
> > > >3) I break out from the while loop in the Task upon a pullup to an
> > external
> > > >pin.
> > > >4) I break out from the while loop in the Task and disable all the
> > > >interrupts (except PRD function) and EDMA events upon a pullup to an
> > > >external pin.
> > > >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and
> > allow
> > > >me to reconnect my XDS510PP+, but doesn't work too. :(
> > > >
> > > >Thanks for taking interest.
> > > >
> > > >Best,
> > > >Teong Beng
> > > >
> > > >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > > > >
> > > > >Are you using a GEL file when you connect in either [or both cases]??
> > > > >
> > > > >If you are, try connecting without one [if you have a complete working
> > > > >program in flash, a GEL file should not be required unless it is to
> > > > >support a debug feature.
> > > > >
> > > > >mikedunn
> > > > >
> > > > >Koay Teong Beng wrote:
> > > > >Dear all,
> > > > >
> > > > >I have a tested code flashed into the DSK Flash, I power cycled the
> > > DSK and
> > > > >it executed as I expected. I can connect with it and download
> > symbols for
> > > > >debugging through the biult-in USB emulator. BUT I CAN NOT connect to it
> > > > >through my XDS510PP+ when my program is flashed and running. Anybody
> > know
> > > > >why is it so? I have reduced my data rate to half and it is still the
> > > same.
> > > > >
> > > > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody
> > face
> > > > >the same issue?
> > > > >
> > > > >Background:
> > > > >The XDS510PP+ is tested and manage to work with the DSK when it is
> > > flashed
> > > > >with post.hex or blink.hex
> > > > >The driver used is: sdgo6400_11.dvr
> > > > >
> > > > >Target: DSK6416, 600MHz
> > > > >CCS: V3.00.00
> > > > >Flashburn: V2.8 (version 5.90.0.110
> > > > >About my code:
> > > > >The code uses two edma channels, one doing a ping pong transferring from
> > > > >emifa to ISRAM and the other doing a ping-pong from ISRAM to McBSP. Both
> > > > >edma generate HW interrupts (that initiates SW interrupt) which I use to
> > > > >preprocess the data in ping pong buffer and dump it into a input
> > > buffer and
> > > > >post-process the data from an output buffer. These are done with qdma.
> > > > >Then, I have a Task that repeatedly process the data in the input buffer
> > > > >and place it at the output buffer. I also have a periodic interrupt that
> > > > >generates patterns of led display. The input sampling rate is now
> > > 500ksps. >
Reply by ●July 11, 20052005-07-11
Jeff and all,
My target is a first version prototype, so we follow as close as possible
to the DSK. It has the same memory map as the DSK. We followed the
reference schematics and gerber given by the DSK. We tried to follow the
BOM list given by the DSK too, down to everything except the component lot
and the passive components.
Have been trying out Flashburn V3 beta, it shows that the Flashburn has
been connected. But, somehow the Erase Flash, Burn Flash etc, buttons are
still disabled.
Still working on it...
Best,
Teong Beng
At 02:00 AM 7/9/2005, Jeff Brower wrote:
>Koay Teong Beng-
>
> > I can now download different programs into the DSK (and into my prototype -
> > which is based on the DSK6416 excluding USB emulator) and execute them
> > through XDS510PP+. But I am still not able to connect to Flashburn, it just
> > stay disconnected.
> >
> > I also verified from the disassembly window that the FBTC program is
> > downloaded and executed, after I launch the cdd configuration. Just that
> > the Flashburn program indicate that it is not connected and the flash
> > burning/erasing buttons are disabled.
> > I tried to use Flashburn from command line, also unsuccessful.
> >
> > Yet, after a number of trials, I sometimes manage to connect Flashburn with
> > DSK, sometimes not. But never to my target. I start to think that it could
> > related to electrical quality. Any input?
>
>If you can connect XDS510PP+ to your prototype board to download code,
>then there is
>no electrical problem. JTAG for function A is same as JTAG for function B.
>
>If there is any difference between your prototype EMIF/memory map and DSK
>C6416
>board, then you have to modify low-level Flash device-driver in Flashburn
>code.
>
>-Jeff
>
> > At 07:51 PM 7/7/2005, Koay Teong Beng wrote:
> > >Mike,
> > >
> > >Thank you! Holding the reset and release while connect works to reconnect
> > >to it with XDS510PP+, and I manage to flash new codes into the DSK,
> even my
> > >prototype target!!
> > >Also, disabling RTDX in DSP/BIOS in my target code enables me to
> connect to
> > >it without having to do the above. Thanks to TI that suggested this. RTDX
> > >engine must have been doing something that prevented the connection, and
> > >your solution give me time to connect.
> > >
> > >I can perform flashburn into my target a few times through XDS510PP+,
> until
> > >I have the POST burned in my target. Then I have problem connecting the
> > >Flashburn.
> > >
> > >My process:
> > >1) connect XDS510PP+ to target
> > >2) power XDS510PP+
> > >3) power on target
> > >4) launch CCS
> > >5) connect CCS to target (success)
> > >6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
> > >Flashburn shows that it fail to
> > >connect)
> > >
> > >Now step 6) works a few times until I have POST flashed into it. :(
> haven't
> > >succeed since then.
> > >
> > >There is an error message when I try to do 'reset CPU' from CCS,
> > >Failed Software reset: Error 0x00000022/-1147 Error during: memory,
> > >execution, it appears that the target is being held in
> > >reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
> > >
> > >I am now working on this issue now. Any idea?
> > >
> > >
> > >Best,Teong Beng
> > >
> > >
> > >At 09:13 PM 7/6/2005, Mike Dunn wrote:
> > > >Teong Beng,
> > > >
> > > >Your problem seems a bit unusual... have you tried holding down the
> reset
> > > >button and letting it go immediately after you click "connect"??
> > > >
> > > >Please keep us posted with anything that you find out.
> > > >
> > > >mikedunn
> > > >
> > > >Koay Teong Beng <koay@koay...> wrote:
> > > >Thanks Mike,
> > > >I had another few tests along this direction but still no luck. They
> > > >include not doing everything else after my self test routine in main(),
> > > >which should goto DSP/BIOS idle straight.
> > > >There is another observation, if I boot from POST and then download the
> > > >.out file of the exact same code (not flashing) into DSK, I am able to
> > > >connect to the DSK using XDS510PP+. Now that make me confuse: If it is
> > > >purely due to the CPU is too tied up to service my emulator, I
> should face
> > > >the same problem whether the code is being downloaded or is flashed,
> YET I
> > > >can connect with XDS510PP+ if I downloaded the code but not if I
> flash it.
> > > >What is the different between the two process except that there is a
> boot
> > > >loader to load my code from Flash into ISRAM? Anyway, I am using the
> same
> > > >boot.asm used by the POST.pjt in CCS example, which I have no problem
> > > >connecting with XDS510PP+.
> > > >Thanks in advance,
> > > >Teong Beng
> > > >
> > > >At 11:52 PM 7/5/2005, Mike Dunn wrote:
> > > > >
> > > > >It sounds like you are on the right track. The emulator does need free
> > > > >cycles to connect. if your application has too much EDMA cycle
> usage, the
> > > > >emulator may not be able to connect [it must perform some DMA
> cycles to
> > > > >connect].
> > > > >
> > > > >mikedunn
> > > > >
> > > > >Koay Teong Beng wrote:
> > > > >Mike,
> > > > >
> > > > >I didn't use startup GEL file in both cases.
> > > > >Peripheral initializations are all done in my code.
> > > > >
> > > > >Mike and all,
> > > > >I start to think along the line that the DSP is too occupied to
> > > service the
> > > > >JTAG. Since my main data processing work is running in a while
> loop in a
> > > > >Task, I tried a few things described below but doesn't help, any more
> > > > ideas?
> > > > >1) I added a small code into the Task containing my data process
> > > algorithm.
> > > > >The code sense an external pin and run IDL_run() instead of my
> processing
> > > > >algorithm upon a pullup to the external pin.
> > > > >2) I added a small code that jump to another while loop containing
> only
> > > > >IDL_run() upon a pullup to an external pin.
> > > > >3) I break out from the while loop in the Task upon a pullup to an
> > > external
> > > > >pin.
> > > > >4) I break out from the while loop in the Task and disable all the
> > > > >interrupts (except PRD function) and EDMA events upon a pullup to an
> > > > >external pin.
> > > > >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and
> > > allow
> > > > >me to reconnect my XDS510PP+, but doesn't work too. :(
> > > > >
> > > > >Thanks for taking interest.
> > > > >
> > > > >Best,
> > > > >Teong Beng
> > > > >
> > > > >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > > > > >
> > > > > >Are you using a GEL file when you connect in either [or both
> cases]??
> > > > > >
> > > > > >If you are, try connecting without one [if you have a complete
> working
> > > > > >program in flash, a GEL file should not be required unless it is to
> > > > > >support a debug feature.
> > > > > >
> > > > > >mikedunn
> > > > > >
> > > > > >Koay Teong Beng wrote:
> > > > > >Dear all,
> > > > > >
> > > > > >I have a tested code flashed into the DSK Flash, I power cycled the
> > > > DSK and
> > > > > >it executed as I expected. I can connect with it and download
> > > symbols for
> > > > > >debugging through the biult-in USB emulator. BUT I CAN NOT
> connect to it
> > > > > >through my XDS510PP+ when my program is flashed and running. Anybody
> > > know
> > > > > >why is it so? I have reduced my data rate to half and it is
> still the
> > > > same.
> > > > > >
> > > > > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody
> > > face
> > > > > >the same issue?
> > > > > >
> > > > > >Background:
> > > > > >The XDS510PP+ is tested and manage to work with the DSK when it is
> > > > flashed
> > > > > >with post.hex or blink.hex
> > > > > >The driver used is: sdgo6400_11.dvr
> > > > > >
> > > > > >Target: DSK6416, 600MHz
> > > > > >CCS: V3.00.00
> > > > > >Flashburn: V2.8 (version 5.90.0.110
> > > > > >About my code:
> > > > > >The code uses two edma channels, one doing a ping pong
> transferring from
> > > > > >emifa to ISRAM and the other doing a ping-pong from ISRAM to
> McBSP. Both
> > > > > >edma generate HW interrupts (that initiates SW interrupt) which
> I use to
> > > > > >preprocess the data in ping pong buffer and dump it into a input
> > > > buffer and
> > > > > >post-process the data from an output buffer. These are done with
> qdma.
> > > > > >Then, I have a Task that repeatedly process the data in the
> input buffer
> > > > > >and place it at the output buffer. I also have a periodic
> interrupt that
> > > > > >generates patterns of led display. The input sampling rate is now
> > > > 500ksps.
> >
> >
> >
> >
> >
> >
My target is a first version prototype, so we follow as close as possible
to the DSK. It has the same memory map as the DSK. We followed the
reference schematics and gerber given by the DSK. We tried to follow the
BOM list given by the DSK too, down to everything except the component lot
and the passive components.
Have been trying out Flashburn V3 beta, it shows that the Flashburn has
been connected. But, somehow the Erase Flash, Burn Flash etc, buttons are
still disabled.
Still working on it...
Best,
Teong Beng
At 02:00 AM 7/9/2005, Jeff Brower wrote:
>Koay Teong Beng-
>
> > I can now download different programs into the DSK (and into my prototype -
> > which is based on the DSK6416 excluding USB emulator) and execute them
> > through XDS510PP+. But I am still not able to connect to Flashburn, it just
> > stay disconnected.
> >
> > I also verified from the disassembly window that the FBTC program is
> > downloaded and executed, after I launch the cdd configuration. Just that
> > the Flashburn program indicate that it is not connected and the flash
> > burning/erasing buttons are disabled.
> > I tried to use Flashburn from command line, also unsuccessful.
> >
> > Yet, after a number of trials, I sometimes manage to connect Flashburn with
> > DSK, sometimes not. But never to my target. I start to think that it could
> > related to electrical quality. Any input?
>
>If you can connect XDS510PP+ to your prototype board to download code,
>then there is
>no electrical problem. JTAG for function A is same as JTAG for function B.
>
>If there is any difference between your prototype EMIF/memory map and DSK
>C6416
>board, then you have to modify low-level Flash device-driver in Flashburn
>code.
>
>-Jeff
>
> > At 07:51 PM 7/7/2005, Koay Teong Beng wrote:
> > >Mike,
> > >
> > >Thank you! Holding the reset and release while connect works to reconnect
> > >to it with XDS510PP+, and I manage to flash new codes into the DSK,
> even my
> > >prototype target!!
> > >Also, disabling RTDX in DSP/BIOS in my target code enables me to
> connect to
> > >it without having to do the above. Thanks to TI that suggested this. RTDX
> > >engine must have been doing something that prevented the connection, and
> > >your solution give me time to connect.
> > >
> > >I can perform flashburn into my target a few times through XDS510PP+,
> until
> > >I have the POST burned in my target. Then I have problem connecting the
> > >Flashburn.
> > >
> > >My process:
> > >1) connect XDS510PP+ to target
> > >2) power XDS510PP+
> > >3) power on target
> > >4) launch CCS
> > >5) connect CCS to target (success)
> > >6) run Flashburn (FBTC downlaoded, CCS shows that program is running, but
> > >Flashburn shows that it fail to
> > >connect)
> > >
> > >Now step 6) works a few times until I have POST flashed into it. :(
> haven't
> > >succeed since then.
> > >
> > >There is an error message when I try to do 'reset CPU' from CCS,
> > >Failed Software reset: Error 0x00000022/-1147 Error during: memory,
> > >execution, it appears that the target is being held in
> > >reset. .... sequence ID:17 Error code: -1147, error class: 0x00000022
> > >
> > >I am now working on this issue now. Any idea?
> > >
> > >
> > >Best,Teong Beng
> > >
> > >
> > >At 09:13 PM 7/6/2005, Mike Dunn wrote:
> > > >Teong Beng,
> > > >
> > > >Your problem seems a bit unusual... have you tried holding down the
> reset
> > > >button and letting it go immediately after you click "connect"??
> > > >
> > > >Please keep us posted with anything that you find out.
> > > >
> > > >mikedunn
> > > >
> > > >Koay Teong Beng <koay@koay...> wrote:
> > > >Thanks Mike,
> > > >I had another few tests along this direction but still no luck. They
> > > >include not doing everything else after my self test routine in main(),
> > > >which should goto DSP/BIOS idle straight.
> > > >There is another observation, if I boot from POST and then download the
> > > >.out file of the exact same code (not flashing) into DSK, I am able to
> > > >connect to the DSK using XDS510PP+. Now that make me confuse: If it is
> > > >purely due to the CPU is too tied up to service my emulator, I
> should face
> > > >the same problem whether the code is being downloaded or is flashed,
> YET I
> > > >can connect with XDS510PP+ if I downloaded the code but not if I
> flash it.
> > > >What is the different between the two process except that there is a
> boot
> > > >loader to load my code from Flash into ISRAM? Anyway, I am using the
> same
> > > >boot.asm used by the POST.pjt in CCS example, which I have no problem
> > > >connecting with XDS510PP+.
> > > >Thanks in advance,
> > > >Teong Beng
> > > >
> > > >At 11:52 PM 7/5/2005, Mike Dunn wrote:
> > > > >
> > > > >It sounds like you are on the right track. The emulator does need free
> > > > >cycles to connect. if your application has too much EDMA cycle
> usage, the
> > > > >emulator may not be able to connect [it must perform some DMA
> cycles to
> > > > >connect].
> > > > >
> > > > >mikedunn
> > > > >
> > > > >Koay Teong Beng wrote:
> > > > >Mike,
> > > > >
> > > > >I didn't use startup GEL file in both cases.
> > > > >Peripheral initializations are all done in my code.
> > > > >
> > > > >Mike and all,
> > > > >I start to think along the line that the DSP is too occupied to
> > > service the
> > > > >JTAG. Since my main data processing work is running in a while
> loop in a
> > > > >Task, I tried a few things described below but doesn't help, any more
> > > > ideas?
> > > > >1) I added a small code into the Task containing my data process
> > > algorithm.
> > > > >The code sense an external pin and run IDL_run() instead of my
> processing
> > > > >algorithm upon a pullup to the external pin.
> > > > >2) I added a small code that jump to another while loop containing
> only
> > > > >IDL_run() upon a pullup to an external pin.
> > > > >3) I break out from the while loop in the Task upon a pullup to an
> > > external
> > > > >pin.
> > > > >4) I break out from the while loop in the Task and disable all the
> > > > >interrupts (except PRD function) and EDMA events upon a pullup to an
> > > > >external pin.
> > > > >I am expecting 3) and 4) will return the handle to DSP/BIOS idle, and
> > > allow
> > > > >me to reconnect my XDS510PP+, but doesn't work too. :(
> > > > >
> > > > >Thanks for taking interest.
> > > > >
> > > > >Best,
> > > > >Teong Beng
> > > > >
> > > > >At 01:06 PM 7/5/2005, Mike Dunn wrote:
> > > > > >
> > > > > >Are you using a GEL file when you connect in either [or both
> cases]??
> > > > > >
> > > > > >If you are, try connecting without one [if you have a complete
> working
> > > > > >program in flash, a GEL file should not be required unless it is to
> > > > > >support a debug feature.
> > > > > >
> > > > > >mikedunn
> > > > > >
> > > > > >Koay Teong Beng wrote:
> > > > > >Dear all,
> > > > > >
> > > > > >I have a tested code flashed into the DSK Flash, I power cycled the
> > > > DSK and
> > > > > >it executed as I expected. I can connect with it and download
> > > symbols for
> > > > > >debugging through the biult-in USB emulator. BUT I CAN NOT
> connect to it
> > > > > >through my XDS510PP+ when my program is flashed and running. Anybody
> > > know
> > > > > >why is it so? I have reduced my data rate to half and it is
> still the
> > > > same.
> > > > > >
> > > > > >Also, the XDS510PP+ seems tends to hold the target in RESET. Anybody
> > > face
> > > > > >the same issue?
> > > > > >
> > > > > >Background:
> > > > > >The XDS510PP+ is tested and manage to work with the DSK when it is
> > > > flashed
> > > > > >with post.hex or blink.hex
> > > > > >The driver used is: sdgo6400_11.dvr
> > > > > >
> > > > > >Target: DSK6416, 600MHz
> > > > > >CCS: V3.00.00
> > > > > >Flashburn: V2.8 (version 5.90.0.110
> > > > > >About my code:
> > > > > >The code uses two edma channels, one doing a ping pong
> transferring from
> > > > > >emifa to ISRAM and the other doing a ping-pong from ISRAM to
> McBSP. Both
> > > > > >edma generate HW interrupts (that initiates SW interrupt) which
> I use to
> > > > > >preprocess the data in ping pong buffer and dump it into a input
> > > > buffer and
> > > > > >post-process the data from an output buffer. These are done with
> qdma.
> > > > > >Then, I have a Task that repeatedly process the data in the
> input buffer
> > > > > >and place it at the output buffer. I also have a periodic
> interrupt that
> > > > > >generates patterns of led display. The input sampling rate is now
> > > > 500ksps.
> >
> >
> >
> >
> >
> >
