> I second this opinion. The logic analyzer does in my mind > classify as a GPIO debugging feature. Surely someone > debugging anything connected to GPIO, such as a key or > MMC card detect or whatever could use this feature to see > what is going on on that line. Okay, with that picture I can see where your argument is coming from. However, making it a gpiolib debugging feature will surely raise expectations. And it is not only the non-equi-distant sampling. The script trying to isolate a CPU core tries really hard but is still hacky IMO. It has to disable the RCU stall detector and will likely interfere with your CPUSET configuration if you have one. As I always said, it is last resort debugging. As I write this, I start to wonder if this should be really upstream or if I just keep it as a branch in my repo. Maybe it is just too hackish to be supported mainline?
Attachment:
signature.asc
Description: PGP signature
