blob: fe1716e8d51b9c1d0f8fdb27cab54d585f48ed7c (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
|
<think>
Okay, so I'm trying to figure out whether this QEMU issue is user-mode or system-mode related. The user is talking about using semihosting in QEMU-system-arm for their embedded ARM project's unit tests. They mentioned that when they run the tests under QEMU with semihosting, the exit code always returns 0 regardless of what they pass to SYS_EXIT. That's a problem because their CI framework relies on non-zero exit codes to detect test failures.
They looked into the arm-semi.c file and found that in the case for SYS_EXIT, it's ignoring the parameter and just calling exit(0). So they patched it temporarily by using an unsupported BKPT instruction to get a non-zero exit code. But they want a better solution where the exit code actually reflects the value passed to SYS_EXIT.
They suggest modifying the exit behavior based on certain conditions: if the code is 0 or 0x20026, exit with 0; else if it's less than 256, use that as the exit code; otherwise, exit with 1. They also mention that 0x20026 is ADP_Stopped_ApplicationExit, which should be a successful termination.
Now, to classify this bug report. The issue revolves around how semihosting in QEMU handles exit codes. Semihosting typically involves user-mode operations because it's about interacting with the host OS from the guest environment without requiring system-level privileges. This isn't about hardware emulation or device drivers but rather the software interaction between the emulated ARM process and the POSIX environment.
The problem is within the user-space code handling, specifically how the exit codes are managed in the semihosting layer. So it doesn't involve lower-level system components like devices or OS kernels. Therefore, this should be classified under 'user' mode.
</think>
user
|