Re: [RFC PATCH net-next 3/9] ethtool: Add an interface for flashing transceiver modules' firmware

[Jakub Kicinski]

On Tue, 23 Jan 2024 13:34:18 +0000 Danielle Ratson wrote:
> > > +The firmware update process can take several minutes to complete.
> > > +Therefore, during the update process notifications are emitted from
> > > +the kernel to user space updating it about the status and progress.  
> > 
> > We should state more explicitly that the op just starts the process, and does
> > not block. Looks like cable test already uses _ACT as a suffix, is it based on
> > some standard? Doesn't seem all that intuitive to me (or at least less intuitive
> > than calling it _START...)  
> 
> From Documentation/networking/ethtool-netlink.rst:
> "
> List of message types
> =====================
> 
> All constants identifying message types use ``ETHTOOL_CMD_`` prefix and suffix
> according to message purpose:
> 
> ==============    ======================================
> ``_GET``          userspace request to retrieve data
> ``_SET``          userspace request to set data
> ``_ACT``          userspace request to perform an action
> ``_GET_REPLY``    kernel reply to a ``GET`` request
> ``_SET_REPLY``    kernel reply to a ``SET`` request
> ``_ACT_REPLY``    kernel reply to an ``ACT`` request
> ``_NTF``          kernel notification
> ==============    ======================================
> "
> 
> So, it looks suitable to me.

True, didn't see that. It's fine as a distinction of "doing something"
vs "setting configuration" but it doesn't express the fact that the
action is async. AFAIU cable test is also async, so that's fine.
We'll worry about it when some tries to add _ACT which isn't async.. 🤷️