One action for the PFC storm is to set a zero buffer profile on the PG under PFC storm.
The zero buffer won't be removed until the PFC storm has gone.
If the user wants to modify the buffer profile for the PG, the bufferorch will return "task_need_retry".
General speaking it doesn't matter unless that:
- the system can't be warm rebooted until the PFC storm has gone.
- the "task_need_retry" will block the update of the entire BUFFER_PG table from being programmed to ASIC.
In this sense, we need a better solution.

The new solution will:
- record the new profile user wants to apply during PFC storm as the "pending profile" for that PG and return "task_success" if the PG is under PFC storm.
- apply the pending profile once the PG is unlocked.
- the latest pending profile will take effect in case user tries updating the profile for more than 1 times.

Signed-off-by: Stephen Sun <stephens@nvidia.com>

Signed-off-by: Stephen Sun <stephens@nvidia.com>

Signed-off-by: Stephen Sun <stephens@nvidia.com>

Signed-off-by: Stephen Sun <stephens@nvidia.com>