[PATCH] cpu: microcode: remove failures when kernel does not have newer version (LP: #1322534)

[Colin Ian King]

On 17/06/14 00:12, Alex Hung wrote:
> Signed-off-by: Alex Hung <alex.hung at canonical.com>
> ---
>  src/cpu/microcode/microcode.c | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/src/cpu/microcode/microcode.c b/src/cpu/microcode/microcode.c
> index 016b938..d01c922 100644
> --- a/src/cpu/microcode/microcode.c
> +++ b/src/cpu/microcode/microcode.c
> @@ -230,13 +230,11 @@ static int microcode_test1(fwts_framework *fw)
>  				}
>  
>  				/*
> -				 * We found the old revision but not a
> -				 * new revsion, failed
> +				 * Kernel does not have newer version than BIOS
>  				 */
>  				if (info->new_revision == UNKNOWN) {
>  					failed++;
> -					fwts_failed(fw, LOG_LEVEL_MEDIUM, "MicrocodeNotUpdated",
> -						"The kernel did not report that CPU %d has had a microcode update. "
> +					fwts_log_info(fw, "The kernel did not report that CPU %d has had a microcode update. "
>  						"The current firmware is revision 0x%x and probably has not been updated.",
>  						cpu, info->old_revision);
>  					continue;
> 
OK, Looks like a sane way to cater for the fact that the the microcode
has not been update. The failure was too harsh.   However,  this is no
longer a call to fwts_failed, and yet we still do a failed++ which I
think will lead to the final test summary not reporting a failure (since
fwts_failed is now not called) or a pass (since a fwts_passed is not
called either).

What is needed is a fwts_skipped() to be reported at the end of the test
if there have been no failures and no passes *AND* we triggered the skip
for this particular microcode issue.

Colin