ppc: 0.806
architecture: 0.793
graphic: 0.722
device: 0.669
boot: 0.664
debug: 0.663
semantic: 0.536
kernel: 0.526
vnc: 0.418
performance: 0.378
mistranslation: 0.298
register: 0.273
user-level: 0.233
virtual: 0.228
socket: 0.203
PID: 0.200
peripherals: 0.169
permissions: 0.151
files: 0.135
risc-v: 0.106
network: 0.094
arm: 0.074
hypervisor: 0.072
assembly: 0.068
TCG: 0.059
VMM: 0.053
x86: 0.018
i386: 0.009
KVM: 0.004

[PPC] SegFault due to Stack Overflow in E500


I am getting a Segmentation Fault while simulating a PowerPC e500. I've tried to debug the problem and I've found that it occurs when you have a 0 value decrementer. The function trace is the following:

1) __cpu_ppc_store_decr (ppc.c) is called with value = 0 and raise_excp=booke_decr_cb;
2) Since value < 3, booke_decr_cb is called;
3) booke_decr_cb then calls booke_update_irq() and cpu_ppc_store_decr();
4) cpu_ppc_store_decr calls __cpu_ppc_store_decr

You're stuck on this infinite cycle until your stack overflows eventually.

Command Line:
qemu-system-ppc -cpu e500v2 -d guest_errors,unimp -m 2048 -M ppce500 -nographic -bios ../cc/share/qem
u/u-boot.e500 -kernel XKYAPP.exe

Platform where the bug occured: Bash ubuntu on Windows;

Revision where the bug was found: e3571ae30cd26d19efd4554c25e32ef64d6a36b3 (16 Set 2016)



Thanks!

Do you know what the DECAR SPR contains at that point in time? I guess it's 0 ... but what does that mean here? Should the decrementer be stopped?

[Expired for QEMU because there has been no activity for 60 days.]