This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 2/2] gdb: Fix testsuite issue in gdb.arch/amd64-disp-step-avx.exp


Hi Andrew,

Sounds fine to me in principle.  A couple comments below.

On 03/22/2018 12:57 PM, Andrew Burgess wrote:

> @@ -103,8 +105,10 @@ proc disp_step_func { func } {
>  with_test_prefix "vex2" {
>      # This case writes to the 'xmm0' register.  Confirm the register's
>      # value is what we believe it is before the AVX instruction runs.
> -    gdb_test "p /x \$xmm0.uint128" " = 0x0" \
> -	"xmm0 has expected value before"
> +    for {set i 0 } { $i < 16 } { incr i } {
> +	gdb_test "p /x \$xmm${i}.uint128" " = 0x0" \
> +	    "xmm${i} has expected value before"
> +    }

This leaves a slight disconnect between the comment and the
code.  I.e., someone reading the comment may wonder why
we check more than xmm0?

Also, should we test xmm1-15 are still 0 after, too, for
completeness?

>  
>      disp_step_func "test_rip_vex2"
>  
> 

Thanks,
Pedro Alves


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]