Re: [PATCH] Fix fail in gdb.base/interrupt-noterm.exp

[Pedro Alves <palves at redhat dot com>]

On 01/22/2016 04:29 PM, Yao Qi wrote:
> Hi,
> In my testing, I see the following fail intermittently,
> 
> interrupt
> (gdb) PASS: gdb.base/interrupt-noterm.exp: interrupt
> [Inferior 1 (process 13407) exited normally]
> 
> Child exited with status 0
> FAIL: gdb.base/interrupt-noterm.exp: inferior received SIGINT (timeout)
> 
> because the interrupt packet may be sent to GDBserver before the SIGIO
> handler is installed.  The fix in this patch is to let GDB wait
> for 500 ms between "continue &" and "interrupt" to make sure
> SIGIO handler is installed already in GDBserver side.

Can you expand the rationale some more?

E.g., why is this not a gdbserver bug?  Instintively I'd say it is.

Thanks,
Pedro Alves

> 
> gdb/testsuite:
> 
> 2016-01-22  Yao Qi  <yao.qi@linaro.org>
> 
> 	* gdb.base/interrupt-noterm.exp: Add "after 500".
> ---
>  gdb/testsuite/gdb.base/interrupt-noterm.exp | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/gdb/testsuite/gdb.base/interrupt-noterm.exp b/gdb/testsuite/gdb.base/interrupt-noterm.exp
> index 05f6076..9b5bb17 100644
> --- a/gdb/testsuite/gdb.base/interrupt-noterm.exp
> +++ b/gdb/testsuite/gdb.base/interrupt-noterm.exp
> @@ -55,6 +55,9 @@ if { $async_supported < 0 } {
>      return 1
>  }
>  
> +# Wait a while so that GDBserver's SIGIO handler is in place.
> +after 500
> +
>  # With native debugging, and no terminal (emulated by interactive-mode
>  # off, above), GDB had a bug where "interrupt" would send SIGINT to
>  # its own process group, instead of the inferior's.
>