Re: [PATCH v2 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI

From: Manivannan Sadhasivam

Date: Mon Apr 13 2026 - 01:46:45 EST


On Sun, Apr 12, 2026 at 08:09:51AM -0700, Jakub Kicinski wrote:
> On Sat, 11 Apr 2026 13:42:00 +0530 Krishna Chaitanya Chundru wrote:
> > - User space applications use the standard Linux PTP interface.
> > - The PTP subsystem routes IOCTLs to the MHI PHC driver.
> > - The MHI PHC driver communicates with the MHI core to fetch timestamps.
> > - The MHI core interacts with the device to retrieve accurate time data.

This is a misleading statement. Only the 'pci_generic' controller driver interacts
with the device for querying timestamp. MHI bus just acts as a messenger.

>
> Nack, stop adding functionality under the mhi "bus".
> Bus is supposed to be an abstraction into which real drivers plug in.

MHI bus is very similar to the PCI bus. Just like PCI capabilities, MHI also has
capabilities to discover the supported functionalities including timesync. So
for making use of the timesync feature, we need to add some hooks into the MHI
bus layer, but the functionality is added to the separate driver,
'drivers/bus/mhi/host/mhi_phc*'. This is also quite similar to how PCI(e)
features like AER, hotplug are structured.

In this series, timesync API definitions are added mistakenly to the core bus
code, which should be moved to the mhi_phc driver instead.

- Mani

--
மணிவண்ணன் சதாசிவம்