I have an Atmel SAM4S Xplained Pro, which I program and debug via a USB cable and Atmel's on-board EDBG chip (in other words, the way God intended). At the moment, if I "Stop Debugging" through Atmel Studio (v7.0), the code seemingly continues running (because I continue to get output in the terminal, which is application specific output).
Does anyone know how to stop this from happening, i.e., how to make "Stop" mean "Stop"? It's doing my head in!
Having contacted Atmel, the answer is that the debugger does not hold the MCU reset after debugging is stopped, so the code will continue to run. This is a known issue and is on their backlog to resolve.
Related
First time debugging from PC to Mobile Phone work fine, when try again get error.
Any solution how fix this when try more debugging.
Add -debugport=54321 to Run->Parameters Parameters. By the way, the "inconsistent certificates" message is because the apk on the device was compiled with a different version of Delphi than the one you are using right now. You will have to uninstall it from the phone first. You can also use -cleaninstall in the Run->Parameters Parameters field.
There is no easy and fast solution.
You may need to:
Restart IDE
Kill adb.exe
Restart Windows
Restart device
Sometimes you will be able to debug again after doing only some of the above steps, however often you will need to do all of them.
Since every try takes plenty of time, I found out that complete Windows restart + device restart is the fastest thing to do.
With Delphi 10.3.3, it is enough to restart the device. After restarting the device problem goes away. It seems after the cancelation of code execution of the project debugged, the debugging service on Android continues to run. At this point, when new debugging started, IDE tries to run a new service with the same id of previous running service.
Symptom: When I select "program flash" menu from the IDE, it starts working, but fails on a random point. Sometimes it erases flash successfully, sometimes fails earlier (see logs below).
History: It was working for weeks. When I uploaded my program, it was running, printing stuff to the UART port. Then, a nice day, flashing has stopped. I've tried, but it printed various errors. Next day it was working again (without any fix), but only for 2-3 days, then stopped working forever.
Diagnosis: In Vivado, I can see the device, it looks healthy. I checked .bat scripts, they set the environment and finally calls rdi_zynq_flash.exe, which is responsible for the whole flashing process (using server on localhost). From this point, I have no chance to know what is really happening under the hood. I was playing around with the XSCT tool, but I have no idea what should I check.
Configuration:
Board: Xilinx Zynq 7000, ZC702
IDE: Vivaldo 2020.1, on Windows 10
program type: standalone (not Linux), only ARM code, no FPGA used
programming mode: QSPI (switch 4 is on, others are off)
The errors are:
unrecognized JEDEC id bytes: 10, 01, 00
unrecognized JEDEC id bytes: 68, 03, 00
Failed to read boot mode register
ERROR: Flash Operation Failed
zynq_qspi_irq_poll: Timeout
See full logs for more details.
Solved automatically. I don't know why, I moved some cables, turned switches on and off. Maybe, it was a hardware issue.
If you have better answer, please, leave here.
So we have Crystal Reports running on a server in a multi user environment. Crystal is used to send documents directly to printers on the local network. Today we noticed the server running at full load and multiple instances of rundll32 hogging all the CPU cycles.
On further investigation it seems that the rundll32 instances are Crystal Print jobs which have hung or otherwise not finished.
I have not been able to reproduce the error yet but does anyone have any suggestions as to what could be causing the rundll32 process to hang like that, and how to avoid it?
Killing the processes manually is not ideal but it's the only way to get the users back going again.
Turns out this was a bug in the printer driver when the 'Printer Notifications' options was enabled. The printer was low on toner, and was issuing a message to the server, which was then using 100% of the cpu.
I disabled the option in the driver and haven't had that issue since.
I'm in an unusual spot. The built in screen on the laptop I'm using right now is broken so bad I can barely see anything besides random flickering lights on it. Because of that, I got and plugged in a monitor to my laptop. This moniter works fine once Windows 7 has already loaded.
The issue is: I'm trying to open the BIOS/boot menu to live boot off of my USB, but I can't see anything. How do I make it so that I get visual output on my monitor early enough to open the BIOS/boot menu?
If it's a Dell laptop, try FN&F1 or F1
Dell Forum
After running into this issue for the 10 billionth time and nearly whipping my mac air out the window in frustration, I've decided to break down and ask if anyone else has run into this...
Environment: running the latest version of XCode (as of this question: 4.2, build 4C199) for an iOS 4.2+ application (universal app).
It appears that GDB randomly decides to completely hang while stepping through lines of code. This has happened to me now on multiple codebases, and in various locations throughout each codebase. The stepping process hangs on an arbitrary point and requires that the running iOS simulator or on-device process be stopped and restarted. Typically, after restarting my debugging session I am able to get to a different arbitrary point in stepping into/thru -- sometimes even enough to actually debug my code entirely (gasp).
I tried wiping my drive, re-installing a fresh copy of Mac OS X 10.6.8 and the aforementioned XCode... No difference.
I tried switching debuggers to LLDB; this resolved the hanging issue, but I'm not as big a fan of LLDB and prefer to use my environment as close to stock as possible.
Any thoughts?
If GDB is still hanging after reinstalling your OS and your SDK, it may be a design issue with your particular code. Does this happen with other projects?
If it makes you feel any better, Apple is moving to LLDB as the stock debugger for Xcode, probably due to issues like your aforementioned issue.