Re: [PATCH] Fix aarch64 ftrace JIT condition testcase

[Pierre Langlois <pierre.langlois@foss.arm.com>]

Hi Antoine,

On 12/04/16 16:13, Antoine Tremblay wrote:
> This patch fixes the following failure:
> FAIL: gdb.trace/trace-condition.exp: ftrace: -(21 << 1) == -42: check 10
> frames were collected.
> 
> This was due to aarch64_emit_sub using the wrong order in its operands, so the
> operation would end up being 42 - 0 rather than 0 - 42.

Ooops, thanks for the fix!  I was a little confused how I could have
missed this, it turns out I had forgotten to had parentheses in
`-(21 << 1) == -42' at the time, so `emit_sub' was not tested here.

I aimed at testing `emit_sub' with the following test case which is
clearly not good enough:

  gdb.trace/trace-condition.exp: ftrace: 21 - 21 == 0

Would you be OK with changing it to "42 - 21 == 21" or something?

Thanks again!
Pierre

> 
> This patch also fixes the order of aarch64_emit_add for clarity.
> 
> Tested on aarch64-native-extended-gdbserver.
> 
> Note: trace-condition.exp was broken a bit so I had to modify it to run
> the test. A fix is coming for that in another patch.
> 
> gdb/gdbserver/ChangeLog:
> 
> 	* linux-aarch64-low.c (aarch64_emit_add): Switch x1 and x0.
> 	(aarch64_emit_sub): Likewise.
> ---
>  gdb/gdbserver/linux-aarch64-low.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/gdb/gdbserver/linux-aarch64-low.c b/gdb/gdbserver/linux-aarch64-low.c
> index 12fe2e6..d237bde 100644
> --- a/gdb/gdbserver/linux-aarch64-low.c
> +++ b/gdb/gdbserver/linux-aarch64-low.c
> @@ -2258,7 +2258,7 @@ aarch64_emit_add (void)
>    uint32_t *p = buf;
>  
>    p += emit_pop (p, x1);
> -  p += emit_add (p, x0, x0, register_operand (x1));
> +  p += emit_add (p, x0, x1, register_operand (x0));
>  
>    emit_ops_insns (buf, p - buf);
>  }
> @@ -2272,7 +2272,7 @@ aarch64_emit_sub (void)
>    uint32_t *p = buf;
>  
>    p += emit_pop (p, x1);
> -  p += emit_sub (p, x0, x0, register_operand (x1));
> +  p += emit_sub (p, x0, x1, register_operand (x0));
>  
>    emit_ops_insns (buf, p - buf);
>  }
>