diff options
Diffstat (limited to 'results/classifier/deepseek-r1:14b/output/hypervisor/2895')
| -rw-r--r-- | results/classifier/deepseek-r1:14b/output/hypervisor/2895 | 31 |
1 files changed, 31 insertions, 0 deletions
diff --git a/results/classifier/deepseek-r1:14b/output/hypervisor/2895 b/results/classifier/deepseek-r1:14b/output/hypervisor/2895 new file mode 100644 index 000000000..1475d5400 --- /dev/null +++ b/results/classifier/deepseek-r1:14b/output/hypervisor/2895 @@ -0,0 +1,31 @@ + +qemu-system-aarch64: Error: r = HV_BAD_ARGUMENT (0xfae94003, at ../target/arm/hvf/hvf.c:2259) +Description of problem: +When I launch a Linux guest in QEMU with `-accel hvf` + `-gdb "tcp::1234"`, and I try to debug the kernel with `lldb`, QEMU crashes with the following report: + +``` +qemu-system-aarch64: Error: r = HV_BAD_ARGUMENT (0xfae94003, at ../target/arm/hvf/hvf.c:2259) +``` +Steps to reproduce: +1. Run QEMU with `-accel hvf` + `-gdb "tcp::1234"` and a custom kernel (`-kernel Image`) +2. Try to attach using `lldb vmlinux` + `(lldb) gdb-remote 1234` +Additional information: +Debugging QEMU I see the crash is due to the `cpu->accel->fd` file descriptor being `0`: + +``` +warning: qemu-system-aarch64 was compiled with optimization - stepping may behave oddly; variables may not be available. +frame #4: 0x00000001003dd24c qemu-system-aarch64`hvf_arch_update_guest_debug [inlined] hvf_put_guest_debug_registers(cpu=0x0000000158118000) at hvf.c:2259:9 [opt] + 2256 for (i = 0; i < max_hw_bps; i++) { + 2257 r = hv_vcpu_set_sys_reg(cpu->accel->fd, dbgbcr_regs[i], + 2258 env->cp15.dbgbcr[i]); +-> 2259 assert_hvf_ok(r); + 2260 r = hv_vcpu_set_sys_reg(cpu->accel->fd, dbgbvr_regs[i], + 2261 env->cp15.dbgbvr[i]); + 2262 assert_hvf_ok(r); +(lldb) print cpu->accel->fd +(hvf_vcpuid) 0 +(lldb) print dbgbcr_regs[i] +(const uint16_t) 32773 +(lldb) print env->cp15.dbgbcr[i] +(uint64_t) 480 +``` |