Re: [PING] [PATCH] Harden gdb.base/step-over-syscall.exp

[Alan Hayward <Alan.Hayward@arm.com>]

> On 22 Jan 2020, at 15:20, Luis Machado <luis.machado@linaro.org> wrote:
> 
> On 1/22/20 11:26 AM, Alan Hayward wrote:
>>> On 22 Jan 2020, at 13:30, Luis Machado <luis.machado@linaro.org> wrote:
>>> 
>>> ping?
>>> 
>>> https://sourceware.org/ml/gdb-patches/2020-01/msg00432.html
>>> 
>>> On 1/15/20 5:36 PM, Luis Machado wrote:
>>>> 
>>>> +
>>>> +    set pc_after_stepi [get_hexadecimal_valueof "\$pc" "0" \
>>>> +			    "pc after stepi with x command"]
>>>> +
>>>> +    if {$next_insn_addr != $pc_after_stepi} {
>>>> +      fail "pc after stepi matches insn addr after syscall"
>>>> +    }
>>>> +
>>>>      return [list $syscall_insn_addr [get_hexadecimal_valueof "\$pc" \
>>>>  					 "0" "pc after stepi"]]
>> Minor nit. In the return, you could just use $next_insn_addr instead of
>> calling get_hexadecimal_valueof again. >
> 
> Do you mean $pc_after_stepi as opposed to $next_insn_addr? $pc_after_stepi is what we get from the $pc register. $next_insn_addr is what we get from using the X command.

Yes, $pc_after_stepi :)

> 
>>>>  }
>>>> @@ -156,8 +192,13 @@ proc step_over_syscall { syscall } {
>>>>  		}
>>>>  	    }
>>>>  -	    gdb_test "continue" "Continuing\\..*Breakpoint \[0-9\]+, .*" \
>>>> -		"continue to syscall insn $syscall"
>>>> +	    # Check if the syscall breakpoint is at the syscall instruction
>>>> +	    # address.  If so, no need to continue, otherwise we will run the
>>>> +	    # inferior to completion.
>>>> +	    if {$syscall_insn_addr != [get_hexadecimal_valueof "\$pc" "0"]} {
>>>> +		gdb_test "continue" "Continuing\\..*Breakpoint \[0-9\]+, .*" \
>>>> +		    "continue to syscall insn $syscall"
>>>> +	    }
>>>>    	    gdb_test_no_output "set displaced-stepping $displaced"