What's the expected behaviour of the Verify function for an open flash-loader binary? The examples at the wiki don't have this function, just its signature. I tried a few different things, but I'm not having any luck. Any tips or suggestions for getting this thing working? Edit: I should note that erasing and flashing do appear to be working correctly, it's just that verification always fails.
Thanks Erik. I'll try the new release as soon as possible. jbrandmeyer - I just handrolled a flashing stub using the F021 API and the J-link semihosting support via GDB. I provided example code to Segger, hopefully that helped to move things along.
Hi Erik, Thanks for the reply. I am passing the big endianness flags to all the different pieces of the chain. For what it's worth, I've tried emailing and calling as well, but didn't get any responses there either. What's the recommended bug reporting/support path? -- Carl
No matter what vector I try to use, "monitor semihosting enable" always sets the vector at 0x08. Am I doing something wrong? Is it supported only for some targets? Help!
FWIW, the normal ARM SWI encodes as 0xef123456, which is why the endian error in the semihosting implementation of JLinkGDBServer freaks out; because of the big-endianness, it reads that word and extracts the wrong three bytes in the wrong order.
Jlink & JLinkGDBserver appear to have endianness bugs related to semihosting. I have a big-endian Cortex-R5, and I have to tell the J-Link that the SWI is "0x3412EF" to make it work. Any chance of a bug fix?
I've got a J-Link up and working great with debugging on a TMS570LC4357ZWT part. It doesn't seem that there's flashing support though? It's listed on the "supported parts" page, but as having no flash, which isn't the case - this part has 4 MB of on-board flash. Is there support coming? What's going on here? If it's important: Source Code (2 lines) Source Code (2 lines) I've tried both 'load' in GDB as well as 'loadbin' with the Commander.
Log attached - this failure occurs for me 100% of the time when quitting the attached GDB session. After the error, subsequent attempts to attach hang forever; only unplugging & repluggin the JLink brings it back to life. Known bug? Workaround? Suggestions welcome! Source Code (66 lines)
Quote from SEGGER - Alex: “ Now I am confused... If you want to debug in parallel, you will have GDB active, so no problem. If you do not want to debug in parallel, use the SWOViewer, since it will be the only process accessing J-Link. Currently, I do not see the problem.” I am using GDB to flash, so if I want debug output without running the full-fledged debugger, I'm stuck killing and restarting SWOViewer and JLinkGDBServer a hundred times a day. If it worked as I'd have expected, I wouldn't h…