Thanks Rafael
↧
Forum Post: RE: Debugger for Jacinto 6 DRA746 and uniflash like flashtool for linux/windows host
↧
Forum Post: RE: Tiva registers do not reset between debug sessions
Hi, I know that there is a reset button on the Tiva c series launchpad. But, is there any way to reset the program without pressing the reset button but to use programming code?
↧
↧
Forum Post: RE: CSS download problem
Hi Zlatko,
ok, glad to hear and all the best with ccs.
Best regards,
Lisa
↧
Forum Post: RE: CCSv6: Changing Path variable in linked resources does not change build variable
Hi Stephen,
I am not sure why they separated out the path for linked resources but I can tell you is a behaviour inherited from Eclipse.
Note both these can be set at the workspace level if this makes is a small amount more convenient. Window->Preferences .. you can do a filter on var for build variables and linked for linked resources.
Best regards,
Lisa
↧
Forum Post: RE: CCSv6 terminal window layout bug
Hi,
This is an open source Eclipse component (part of the Target Management project) and therefore we don't have much power to fix bugs on it.
However, I found out on their project page that release 3.6M4 (a milestone release) has this bug fixed:
In this case, you can potentially add the update site of this project to your CCS and give it a try. To do this, go to the menu Help --> Install New software, then click on the top right link "Available Software Sites", add a new software site and type the parameters shown at the top of the screenshot below (the Work with box). Then select the two components as shown:
I did not do a thorough test on the terminal, but it worked fine with my BeagleBone. Use it at your own risk.
Hope this helps,
Rafael
↧
↧
Forum Post: RE: RE: Difference between Compiler of CCS3.1 and CCS5.5
[quote user="Manish Kumar10"]What is the differnce between the Executables (.obj or .out) generated by the compiler version 5.1.0 and 7.4.4 ?[/quote]
Those compilers were released about 8 years apart. The final executed result will be the same. The runtime conventions are the same (type sizes, which registers are used to pass arguments, etc). But just about everything else will be different.
There is nothing we can act on here. You need to narrow things down to something more concrete.
My first guess, which is probably wrong, is that optimization performed by the compiler has exposed a bug in your source that has been there all along. To test this theory, lower the level of optimization and see if it helps.
Thanks and regards,
-George
↧
Forum Post: RE: How to get something like logic analyzer in CCS?
Hi,
When dealing with hardware your options are:
[quote user="Mohammed Noureldin"]for example I want to check out the systick_current register value after every tick [/quote]
You can insert a breakpoint that updates the entire CCS window - something called a watchpoint. Check section 3.4 of the page below:
http://processors.wiki.ti.com/index.php/GSG:Debugging_projects_v5
Keep in mind that you will be setting a breakpoint to visualize the status of the register, thus interrupting the target and potentially breaking any real-time processing.
[quote user="Mohammed Noureldin"]how long does it take between the ticks to check if everything is go well by timing setup.[/quote]
For that you can use the Profile clock:
http://processors.wiki.ti.com/index.php/Profile_clock_in_CCS
Hope this helps,
Rafael
↧
Forum Post: RE: Eagle models for csd17381F4 and LM5106
Johannes,
Please see the attached .TSM file for the CSD17384F4 TINA-TI model. You can place it using the Insert-->Macro function in TINA-TI.
(Please visit the site to view this file)
Britt
↧
Forum Post: RE: Customizing memory .cmd file
Hi,
The file is completely customizable, but as you guessed it will be constrained to whatever the hardware physically has. In the specific case of this file, its structure matches the hardware present on a C6713 DSK board.
A few highlights that may help you:
- If you are not using the L2 cache, then the L2RAM memory space can be merged with the IRAM. In this case you would comment out the L2RAM line and increase IRAM's length to 0x40000.
- Given the size of your .far section, you have no option other than allocate it in external RAM. If you are using a C6713DSK, you can simply rewrite it to: .far > EMIFCE0
- Be careful when allocating the .text section to EMIFCE1, especially if you are still at the develop/debug phase of your project. The reason is that CCS is unable to write to this memory (as it is an external flash device), therefore I would allocate it to RAM (either internal or external). If you are using the C6713DSK, you can use the Flashburn utility to flash its external Flash device (I think the workshop mentions that).
In general you will want to allocate initialized sections to non-volatile memory and un-initialized sections to RAM. Details can be seen in section 5.3.5 of the C6000 Compiler User's guide (SPRU187).
Hope this helps,
Rafael
↧
↧
Forum Post: RE: Error connecting to the target:
Sagar,
One detail that isn't clear for me: are you trying to connect to both kits connected to the PC at the same time? If so, you would need to follow question 6.1.9 of the XDS100 page.
If you are having problems connecting to just one of the kits when individually plugged to the PC, the troubleshooting details presented by Gautam are ok and may be indicative of a faulty kit.
By the way, A6D0 is the correct VID for XDS100v1 (yours) and XDS100v2 (section 5.3.2 of the same page above).
Hope this helps,
Rafael
↧
Forum Post: CCS 6 installer does nothing
Hi there,
I've been trying to install CCS on my desktop for a couple of months, no luck. I looked at all similar problems in the forum, but could not find my exact situation. So here it goes:
- Windows 7 Professional SP1, fully updated, running in PT-BR language
- Intel Core i5-2400, 8Gb RAM
AVG anti-virus (properly disabled before attempting to install)
Tried CCS 6.0.0 (both online and full download installations), and now back to the subject I am trying with CCS6.0.1.00040. For what is worth, the file name is _win32.zip, and my machine is 64 bits, does that matter?
Double clicking the installer does nothing, no errors, no messages. But it does create an entry on the Control Panel Uninstall Programs... It does create folders (including on the Start Menu), but these folders are empty.
Where do I look now? I followed the "troubleshooting" from another post, and erased anything related to a possible previous installation/attempt from users/<my_username>/...
The "users" folder in PT-BR windows has an accented name "usuários" (thanks Microsoft). Is that maybe the problem?
PS: Meanwhile, I have been using it quite well on my laptop computer - but I do want to come back to my "big screens"!
Thanks
↧
Forum Post: RE: CCSv5 dependencies not working - header file changes do not trigger build when using the "build all" command. Is this a known bug?
This problem existed since introduction of CCS V4.The problem is that dependencies checking of .h file does not work. If just the .c files are changed, it builds correctly. If .h file are changed it does not, therefore the rebuild all command must be used. I do not use assembler so I do not know if dependency checking on .asm file works correctly. I do not now what is Total Insanity doing, but it is ridiculous to have such a trivial problem for years and not to fix it. I bet TI blames Eclipse for it...
↧
Forum Post: RE: UniFlash 3.1 install problem
Mike,
Were you able to work around the installation issues?
If not, can you try attaching the install logs again? As Gautam mentioned, the original attachment is not accessible for some reason.
Thanks,
Ricky
↧
↧
Forum Post: RE: CC C3x-C4x Only steps through disassembly for main
Frank,
If you right click on main.c and select mixed-mode does it interleave the assembly for you? It seems like just for this one file that source association has been lost. You have already checked that symbolic info is on. You could move the function into a file that is working ok and see if that resolves the problem.
This toolset is really old. It could be a bug in the way we were generating debug info or the way the debugger was reading it. Hopefully moving the function will work around the issue.
Regards,
John
↧
Forum Post: RE: __TI_auto_init() --- in CCS3.3
Do you have the C runtime support library added to your build options? In the linker options for the project you should add the appropriate C runtime library. I dont have CCSv3 installed but they should be in a /lib folder under the compiler installation directory.
John
↧
Forum Post: RE: Voltage vs Power graph generation using GUI Composer.
You only set 8 elements for the Gui_Power_for_graph in the for loop. Shouldn't this be 14 elements? So that it matches the the red line?
↧
Forum Post: RE: CCS ram/ flash bugs C2000 piccolo 28069 out of RAM
The compiler will also block global arrays across data pages, in order to improve DP load performance. To determine whether this blocking could be the reason why your data doesn't fit, try disabling the blocking optimization with the --disable_dp_load_opt (or -md) flag. Performance would likely suffer too much for this to be a practical solution, but it could indicate whether the data page blocking is a factor. You can also look at the memory map file to see how your arrays are laid out across the data pages.
↧
↧
Forum Post: RE: DSS openSession
Increasing the JTAG clock and bypassing the other cores speeded up the process, but is still too long (~5 sec/call).
I will try now the TestServer.js (one examples script on TI installation folder) . I think this will solve my problem with timing, since once the debugger session is opened, the other operations execute pretty fast.
↧
Forum Post: RE: Using __text__ directive with CCS?
While you can rely on symbols predefined by the linker like .text and etext, it is pretty easy to just create your own in the linker command file. Here is one way to create your own symbols named TEXT_START and TEXT_SIZE.
.text > RAM, RUN_START(TEXT_START), RUN_SIZE(TEXT_SIZE)
Referring to these symbols in C is straightforward, but a bit non-intuitive. This post describes it in detail.
Thanks and regards,
-George
↧
Forum Post: RE: gmtime/localtime (time.h functions) work incorrectly in CCSv6
↧