Hi Daniel,

El 14/04/2017 a las 19:43, Daniel Schwierzeck escribió:

+cc Simon

Am 13.04.2017 um 17:52 schrieb Álvaro Fernández Rojas:

...

Commit 740d5d3 added two new features but only one feature name, which results in NULL prints when device_id feature is selected. Fix this by not printing features without a corresponding name.

Before: HG556a # cpu detail -1: cpu@0 BCM6358A1 ID = 0, freq = 300 MHz: L1 cache, MMU, NULL Device ID 0x2a010 -1: cpu@1 BCM6358A1 ID = 1, freq = 300 MHz: L1 cache, MMU, NULL Device ID 0x2a010 After: HG556a # cpu detail -1: cpu@0 BCM6358A1 ID = 0, freq = 300 MHz: L1 cache, MMU Device ID 0x2a010 -1: cpu@1 BCM6358A1 ID = 1, freq = 300 MHz: L1 cache, MMU Device ID 0x2a010

Signed-off-by: Álvaro Fernández Rojas noltari@gmail.com

cmd/cpu.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/cmd/cpu.c b/cmd/cpu.c index bc4dc5c..14053d2 100644 --- a/cmd/cpu.c +++ b/cmd/cpu.c @@ -52,7 +52,8 @@ static int print_cpu_list(bool detail) print_freq(info.cpu_freq, ""); first = true; for (i = 0; i < CPU_FEAT_COUNT; i++) {

• 	if (info.features & (1 << i)) {
  

• 	if (info.features & (1 << i) &&
  

• 	    cpu_feature_name[i] != NULL) {
  	printf("%s%s", first ? ": " : ", ",
  	       cpu_feature_name[i]);
  	first = false;
  
  

wouldn't be adding the missing feature name the better fix?

The thing is I wouldn't print Microcode feature either, because it's printed right below if it's really avaiable and I don't see the point of printing it twice (first just a description and then the actual device id too, and the same goes for microcode). If I fix it like you suggest we will have the following: HG556a # cpu detail -1: cpu@0 BCM6358A1 ID = 0, freq = 300 MHz: L1 cache, MMU, *Device ID* *Device ID 0x2a010* -1: cpu@1 BCM6358A1 ID = 1, freq = 300 MHz: L1 cache, MMU, *Device ID* *Device ID 0x2a010*