On 10/28/22 00:18, Ilya Leoshkevich wrote:
in one of the wasmtime tests (host=x86_64, guest=s390x). GDB shows that the root cause is actually this:Thread 181 "wasi_tokio::pat" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7ffff6c54640 (LWP 168352)] 0x0000555555626736 in do_tb_phys_invalidate (tb=tb@entry=0x7fffea4b8500 <code_gen_buffer+38503635>, rm_from_page_list=rm_from_page_list@entry=true) at qemu/accel/tcg/translate-all.c:1192 1192 if (qatomic_read(&jc->array[h].tb) == tb) { (gdb) bt #0 0x0000555555626736 in do_tb_phys_invalidate (tb=tb@entry=0x7fffea4b8500 <code_gen_buffer+38503635>, rm_from_page_list=rm_from_page_list@entry=true) at qemu/accel/tcg/translate-all.c:1192 #1 0x0000555555626b98 in tb_phys_invalidate__locked (tb=0x7fffea4b8500 <code_gen_buffer+38503635>) at qemu/accel/tcg/translate-all.c:1211 #2 tb_invalidate_phys_page_range__locked (p=<optimized out>, start=start@entry=836716683264, end=end@entry=836716687360, retaddr=0, pages=0x0) at qemu/accel/tcg/translate-all.c:1678 #3 0x0000555555626dfb in tb_invalidate_phys_range (start=836716683264, start@entry=836716584960, end=end@entry=836716982272) at qemu/accel/tcg/translate-all.c:1753 #4 0x0000555555639e43 in target_munmap (start=start@entry=836716584960, len=len@entry=397312) at qemu/linux-user/mmap.c:769 Let me know if you need more information, I can try to extract a minimal reproducer.
A reproducer would be helpful. r~
