On Sun, Jun 02, 2019 at 07:49:59PM +0530, Vivek Pernamitta wrote: > When user space application request for change in spi clock > using ioctl, current value is taken back-up and new value is > assigned to spi->max_speed_hz, then spi_setup() function is > called with new value. If spi_setup() function fails, it needs > reverting to old spi_max_speed value only in failure condition. The current behaviour is intentional, we're not modifying max_speed_hz so we can continue to enforce the kernel limit no matter what userspace asked for - actually looking at the code we should probably have an explicit error check in there to reject attempts to raise it here, not sure why we're not doing that already.
Attachment:
signature.asc
Description: PGP signature
