Re: [Lightning] Segfault on arm when calling jit_disassemble

[Paulo César Pereira de Andrade]

2014-04-04 18:51 GMT-03:00 Domingo Alvarez Duarte <address@hidden>:
> Hello !

  Hi,

> I'm testing lightning on arm and the printf.c example that come in the docs
> folder segfaults when calling jit_disassemble, also there is no bugtracker

  Thanks for reporting this problem. It was a side effect of a change done
long ago when zeroing freed pointers. Just in case, after the correction
I see this on armv7hl:

$ ./printf
generated 52 bytes
# :printf.c:17
        0x2abbe000      stmia   r0!, {r0}
        0x2abbe002      b.n     0x2abbe524
        0x2abbe004      vrhadd.u16      d14, d12, d31
        0x2abbe008      stmdb   sp!, {r4, r5, r6, r7, r8, r9, fp, lr}
        0x2abbe00c      vpush   {d8-d15}
        0x2abbe010      stmdb   sp!, {r0, r1, r2, r3}
        0x2abbe014      mov     fp, sp
        0x2abbe016      sub.w   sp, sp, #0
        0x2abbe01a      mov     r5, r0
        0x2abbe01c      movw    r0, #35604      ; 0x8b14
        0x2abbe020      mov     r1, r5
        0x2abbe022      movw    ip, #34704      ; 0x8790
        0x2abbe026      blx     ip
        0x2abbe028      add.w   sp, fp, #16
        0x2abbe02c      vpop    {d8-d15}
        0x2abbe030      ldmia.w sp!, {r4, r5, r6, r7, r8, r9, fp, pc}

  The correction is to call jit_disassemble before jit_clear_state,
in doc/printf.c, because jit_clear_state actually release memory
used by disassemble printing.

> to resgister bugs ???

  I am sorry that the project is quite small, and only the mailing list
is used to report bugs :-(

  I have a pending new lightning release, and will add this correction
also.

Thanks,
Paulo