blob: e8b664cae9c3f5aff050eccc340ca8b8cf38987e (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
|
<think>
Okay, so I'm looking at this bug report about QEMU on a LoongArch host where the LSX instructions are giving wrong results. Let's break it down.
The user describes that when they run a test program using QEMU in loongarch64-linux-user mode, some of the vector elements (like index 4,5,6,7) show incorrect values like ffffffff or 0 instead of all zeros as expected when running natively on LoongArch or using QEMU on an x86_64 host.
The steps involve building a Linux user-mode QEMU, compiling test code statically with GCC, and then executing it via QEMU. The results show discrepancies in certain vector elements which suggests there might be an issue with how the LSX (Loongson Extension) instructions are being handled by the emulator.
LSX instructions are part of LoongArch's SIMD (Single Instruction Multiple Data) extensions, used for parallel processing. If these aren't emulated correctly, it would cause incorrect results in vector operations.
The classification is between 'user-mode' and 'system-mode'. User-mode deals with executing userland applications, which matches the use of QEMU's linux-user mode here. System-mode involves higher privileges, devices, etc., which doesn't seem to be the issue here since the problem arises from running a user application.
So, this bug is related to how user-mode QEMU handles vector instructions on LoongArch.
</think>
user
|