Search Results

I am using a J-Trace Pro with a J-Trace Pro, with Ozone. When I look at the timeline view, it looks like the history only goes back to 10M instructions. On a 100MHz cpu, that's not much time! Is this a hard limitation? My understanding was that streaming trace should give me "unlimited" history, or as much as my PC memory/hard drive can store. Then a similar question is whether or not there is a history horizon on what is going into "Code Profile" output, particularly the "Load" column. I am dea…

Does anyone over there at Segger have an implementation for RTT that is known to work with TI-RTOS? The documentation for TI-RTOS is a total hellstorm, so if someone at Segger has already solved this, that would be *super* useful. I am just now trying to put together a project where a CC13xx is used as a coprocessor, and having RTT would be VERY helpful, due to there only being a single UART on the CC1310. However, when I try to build RTT out-of-the-box, I get an unresolved symbol error for OS_G…

Could we get a proper progress bar for flash write in OSX in Ozone?I am frequently writing fairly large images (750K-1MB), but in OSX there is no progress, and it just looks like an application hang. It is a small feature, but would be really nice to have.

The use case is to flash the SPI flash via the existing microcontroller. The SPI bus isn't sufficiently available on the board (without building a flying probe system, which I don't know how to do) to use JFlashSPI. The micro is an ST Cortex M4, so should be well supported. Since I saw at least one mention of the MX25 on the wiki, the hope was that the flashing command set or RAMCode already existed somewhere in the JLink libraries. I am still not quite certain what the setup should look like to…

I have a situation wherein I need to flash a specific image to a Macronix MX25 flash that is already on my boards. So far, I have been using custom programs with assets converted to C binary blobs and compiled in (yes, seriously), to then be flashed to the external flash upon execution. It would be great if I could use J-Flash to accomplish this, especially as it gets more complex. Looking through the J-Flash manual, it looks like there is some support for tasks like this. But it looks like it i…

How can I get access to the OS awareness Plugin SDK? I saw it mentioned here: segger.com/products/developmen…echnology/rtos-awareness/ I would like to implement awareness for ThreadX. Or at least try to. I am using ETM trace on a J-Trace Pro, and it is pretty great, but anything beyond a short timeline doesn't mean much due to incorrect unstacking. I emailed info@segger.com, but got no response. Is the SDK available?

I am using Ozone + J-Trace Pro on a target that is running ThreadX, and I am seeing some pretty crazy results on the Timeline view. Unfortunately, the only environment I know of with a ThreadX awareness plugin is IAR, which this project is not set up for. Image 1 shows what I am seeing at some places; note the deep stacking of the same function (a while loop waiting on a queue). Is it expected that the stack unwinding for Timeline view will not be able to decipher the context switches? Of course…

That is an OK solution. It does work, but it's a bit roundabout, so I have ended up just using the RTT output and copying from the console (JLinkRTTClient) anyway, but in already formatted output to paste into Matlab.

That does not help with my use case, unfortunately, since I want to pause the debugger at some breakpoint and then look at the value of a 4096-element array in memory. Of course, I can do that in the Watch window, but it's still not a graph, and unfortunately, there does not even appear to be an "Export" for the Watch values. Consider that another feature request. For now, I am just printing out the array and copy-pasting it into a CSV file, and plotting in Matlab.

So, the Data Graph feature in Ozone is great. I used it extensively to verify some LED and display effects that went into a product, ensuring that the fades and holds were all correct for time and value. BUT: How about the ability to graph an array of data on an X-Y graph from an array? With X being the ordinal index and Y being the data. That would be GREAT. Some IDE's (CCS and some Eclipse based ones) support that.

When using Ozone, I commonly encounter situations where pointers to dynamic memory are used, but are picked up as just 1-element arrays by the Ozone debugger since there is no compile-time info available, so that only the [0] element can be dereferenced in the "Watched Data" windows. Is there support in Expressions used by Ozone support a way to look at the pointer as an arbitrary type, like casting (sort of) to a char[128] or similar?

Is there an option to disable local echo for the command-line version of RTTClient on OSX? I check the manual and such and I didn't see one ... but I am well aware that sometimes things are so obvious that they get glossed over. If that option doesn't exist, is there any chance it could get added? And is there any chance of a GUI for OSX, like there is on Windows? I am using Ozone now, but the RTT terminal that is there isn't too useful since it is line-oriented, rather than character by charact…

Yeah, that is an OK option. There are occasional changes to the NVM that have to be reflashed ... This seems to work, though: C Source Code (4 lines) C Source Code (7 lines) so then the NVM/bootloader loads will be trivial loads (only change if needed), and then the device will reset and run as normal? Understanding that you need to disconnect and reconnect to reflect any changes into the flash.

Interesting ... that's a little more complicated than I thought. The intent really was just to, on connect: loader bootloader & NVM bin's, then reset the chip and run as normal. I even have an ELF for the bootloader, although I expect to never do any debugging inside it.

I tried to follow the example given in the Ozone user manual for setting up a multi-image download, but I am having some problems. The project consists of 3 ELF files: bootloader, flash area used as NVM, and then the main program. I have all 3 as both elf and bin files. The main program elf is selected as the elf file to load in the new project creator. Following "6.2.3.1 Writing a Multi-Image Download Routine" in the manual, I added the section to my project file: C Source Code (8 lines) And th…

I am currently evaluating SystemView, with the v2.50 application running on OSX. Unfortunately, I am experiencing a reproducible crash situation after ~40s of live recording. I am flashing & running the target with Ozone, and then initiating recording with the SystemView application. Running with a J-Trace Pro probe on the Segger Cortex-M Trace Reference board (STM32F407 micro). All the default SystemView config options are set, other than changing the SYSVIEW_RAM_BASE to 0x20000000. Crash log i…

OK, so I now have a J-trace Pro and it is going to be used for full, real 4-pin ETM trace over ethernet on an upcoming project that is properly set up for it (micro with trace pins, and routed to a proper connector). But I would also like to figure out what is possible for older projects that were not or could not be set up for full trace. I know, these questions come up a lot given how many debug technologies there are now. One scenario is for an STM32F1 chip where SWD+SWO is available, but not…

So, I am very interested in using SystemView on an existing project that I have here. But, the RTOS in use in ThreadX, and moving to FreeRTOS is not presently tractable. This is further complicated by ThreadX being usually a binary distribution, as I understand it and in the SDK that I am using that incorporates ThreadX, there are only headers and archive files, so patching as in the example included in the SystemView download would not be possible. But maybe there are enough hook functions avai…

Interesting ... if I do Project.SetRootPath("..."), where "...." is the absolute path to the root directory where the makefile build is invoked, then all the source files auto-populate. By adding that line into the OnProjectLoad function manually, then all the right settings seem to persist when I quit out and reload the project. Are there any relative paths or root path information present in the ELF file? All I can see from objdump and readelf are the file names, but without any paths. Even wh…

I am just trying out Ozone after seeing it at Arm TechCon today. I have a large makefile based project that I would like to bring some debug capability to. I have done some very limited debugging via a roundabout method using TrueStudio, which I believe just automates the interface over gdb-server and wraps a gui around it. But I would like to get set up with full streaming trace. The ELF file output loads with no problem in Ozone, but there are hundreds of source and header files which are not …