On 4/2/2025 4:22 AM, Henry Martin wrote:
Hi Arend, Thank you for your thorough review and catching the return type mismatch. Upon further investigation, I’ve confirmed that this issue was flagged by static analysis but appears to be a false positive, as all call sites already handle NULL checks appropriately. I appreciate your time and insight—please let me know if you’d like me to drop this patch or revise it differently.
If I look at the code I think the driver probe will eventually fail when the board_type is not available although USB devices seem to be the exception here. For PCIe and SDIO the board_type seems required so we could bail out in brcmf_get_module_param() when there is no board_type found, ie. returning NULL iso settings. I think I found another issue for SDIO. Upon failure it may end up with sdiodev->settings being ERR_PTR() so not NULL. This is not properly handled in the remove path.
So drop the patch and I will see if I can incorporate the musings above in some driver patches.
Regards, Arend
