[PATCH v3 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI

Krishna Chaitanya Chundru posted 6 patches 1 week ago
drivers/bus/mhi/common.h           |  15 ++++
drivers/bus/mhi/host/Kconfig       |   9 +++
drivers/bus/mhi/host/Makefile      |   2 +
drivers/bus/mhi/host/init.c        |  65 +++++++++++++++
drivers/bus/mhi/host/internal.h    |  35 ++++++++
drivers/bus/mhi/host/main.c        |  14 ++++
drivers/bus/mhi/host/mhi_phc.c     | 161 +++++++++++++++++++++++++++++++++++++
drivers/bus/mhi/host/mhi_phc.h     |  21 +++++
drivers/bus/mhi/host/pci_generic.c |  29 +++++++
include/linux/mhi.h                |  11 +++
10 files changed, 362 insertions(+)
[PATCH v3 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI
Posted by Krishna Chaitanya Chundru 1 week ago
This series adds support for exposing the device-side timestamp provided
by the MHI TSC Timesync capability as a PTP Hardware Clock (PHC).
 
The series introduces an MHI PHC driver that registers a PTP clock and
uses the TSC Timesync capability to retrieve device timestamps. These
timestamps are exposed through the Linux PTP framework, enabling host
and device clock correlation using standard userspace PTP interfaces.
 
To support efficient timestamp retrieval, the series also adds generic
64-bit MHI register access support and capability discovery APIs used by
the PHC driver.

The following diagram illustrates the architecture and data flow:

 +---------------+  +----------------------+  +----------------+
 | Userspace App |  | Kernel PTP framework |  | MHI PHC Driver |
 +---------------+  +----------------------+  +----------------+
                                                       |
                                                       v
                                            +--------------------+
                                            |  Endpoint Device   |
                                            | (Timestamp source) |
                                            +--------------------+

- User space applications use the standard Linux PTP interface.
- The PTP subsystem routes IOCTLs to the MHI PHC driver.
- The PHC driver communicates with the MHI device to get the time data.

Signed-off-by: Krishna Chaitanya Chundru <krishna.chundru@oss.qualcomm.com>
---
Changes in v3:
- Use mutex instead of spinlock & couple of nits (Jie).
- Move all tsc timesync to mhi_phc.c (Mani). Moved the init part also as
  it is easy to control from controller without mhi core dependency.
- Couple of nits & de-register ptp driver in case of syserror and enable
  back in mission mode (Mani)
- use PTP_SYS_OFFSET_EXTENDED to read multiple samples and also support
multiple clock ids by using ptp_read_system_prets & ptp_read_system_postts (Vadim Fedorenko)

- Link to v2: https://lore.kernel.org/r/20260411-tsc_timesync-v2-0-6f25f72987b3@oss.qualcomm.com

Changes in V2:
- Rebases to the latest code.
- Patch 1 (bus: mhi: host: Add support to read MHI capabilities) which
is dependent change was brought it here as the series which this patch
is part need to re-designed and will take time https://lore.kernel.org/all/CAFEp6-0ik4B20cRyid9w0f+UgibGciPof9HCWTJ=uBOPvHG35Q@mail.gmail.com/.
- Link to v1: https://lore.kernel.org/all/20250818-tsc_time_sync-v1-0-2747710693ba@oss.qualcomm.com/

---
Krishna Chaitanya Chundru (5):
      bus: mhi: host: mhi_phc: Add support for PHC over MHI
      bus: mhi: host: Add support for 64bit register reads and writes
      bus: mhi: pci_generic: Add support for 64 bit register read & write
      bus: mhi: host: Export mhi_read_reg
      bus: mhi: host: mhi_phc: Update the Time sync logic to read 64 bit register value

Vivek Pernamitta (1):
      bus: mhi: host: Add support to read MHI capabilities

 drivers/bus/mhi/common.h           |  15 ++++
 drivers/bus/mhi/host/Kconfig       |   9 +++
 drivers/bus/mhi/host/Makefile      |   2 +
 drivers/bus/mhi/host/init.c        |  65 +++++++++++++++
 drivers/bus/mhi/host/internal.h    |  35 ++++++++
 drivers/bus/mhi/host/main.c        |  14 ++++
 drivers/bus/mhi/host/mhi_phc.c     | 161 +++++++++++++++++++++++++++++++++++++
 drivers/bus/mhi/host/mhi_phc.h     |  21 +++++
 drivers/bus/mhi/host/pci_generic.c |  29 +++++++
 include/linux/mhi.h                |  11 +++
 10 files changed, 362 insertions(+)
---
base-commit: 238650ef6c7c7cca08e032527329424c9fbd70e5
change-id: 20260411-tsc_timesync-f877a0394393

Best regards,
--  
Krishna Chaitanya Chundru <krishna.chundru@oss.qualcomm.com>

Re: [PATCH v3 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI
Posted by Jakub Kicinski 1 week ago
On Thu, 17 Sep 2026 12:38:01 +0530 Krishna Chaitanya Chundru wrote:
> Subject: [PATCH v3 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI

The MHI "bus" is the worst abstraction I've seen. You're turning it
into "qualcomms corner of the kernel". The PHC drivers belong in
drivers/ptp/

Carry this until told otherwise:

Nacked-by: Jakub Kicinski <kuba@kernel.org>
Re: [PATCH v3 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI
Posted by Krishna Chaitanya Chundru 6 days, 15 hours ago

On 9/17/2026 8:18 PM, Jakub Kicinski wrote:
> On Thu, 17 Sep 2026 12:38:01 +0530 Krishna Chaitanya Chundru wrote:
>> Subject: [PATCH v3 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI
> The MHI "bus" is the worst abstraction I've seen. You're turning it
> into "qualcomms corner of the kernel". The PHC drivers belong in
> drivers/ptp/
>
> Carry this until told otherwise:
>
> Nacked-by: Jakub Kicinski <kuba@kernel.org>
Thanks for the feedback.

We agree that the PHC driver does not belong under the MHI bus.
In the next revision, we'll move the PHC driver to drivers/ptp 

Since the concern raised by the NAK is being addressed by moving the
driver out of drivers/bus/mhi, I'll drop the Nacked-by tag in the next
revision.

- Krishna Chaitanya.

Re: [PATCH v3 0/6] bus: mhi: host: mhi_phc: Add support for PHC over MHI
Posted by Jakub Kicinski 3 days, 5 hours ago
On Fri, 18 Sep 2026 16:22:16 +0530 Krishna Chaitanya Chundru wrote:
> > Carry this until told otherwise:
> >
> > Nacked-by: Jakub Kicinski <kuba@kernel.org>  
> Thanks for the feedback.
> 
> We agree that the PHC driver does not belong under the MHI bus.
> In the next revision, we'll move the PHC driver to drivers/ptp 
> 
> Since the concern raised by the NAK is being addressed by moving the
> driver out of drivers/bus/mhi, I'll drop the Nacked-by tag in the next
> revision.

Is that what "until told otherwise" means?