On 19/09/2025 1:18, Jakub Kicinski wrote:
On Thu, 18 Sep 2025 22:41:38 +0300 Carolina Jubran wrote:
On 18/09/2025 18:40, Jakub Kicinski wrote:
I understand that the modes should not be exposed.
I don't get why this has anything to do with the number of bins.
Does the FW hardcode that the non-Ethernet modes use bins >=16?
When you say "internal modes that can report more than 16 bins"
it sounds like it uses bins starting from 0, e.g. 0..31.
The FW hardcodes that Ethernet modes report up to 16 bins,
while non-Ethernet modes may report up to 19.
And yes, those modes use bins starting from 0, e.g. 0..18.
Which means that the number of bins doesn't really matter.
You're purely using the bin count as a second order check
to catch the device being in the wrong mode (and I presume
you think that device in the wrong mode should never enter
the function given the WARN_ON_ONCE()).
Please check the mode directly or remove the check completely.
You are right, the check does look like it's combining two different
things. I will add explicit checking for the mode and keep the
WARN_ON_ONCE() to guard against FW changes or potential bugs.
