Re: [PATCH] media: ipu-bridge: Add Himax HM1092 IR sensor
Sakari Ailus <[email protected]>
| Newsgroups | org.kernel.vger.linux-media |
|---|---|
| Organization | Intel Finland Oy - BIC 0357606-4 - c/o Alberga Business Park, 6 krs, Bertel Jungin Aukio 5, 02600 Espoo |
| Message-ID | <[email protected]> |
Hi Jake, On Wed, Jul 29, 2026 at 01:46:41PM -0400, Jake S wrote: > From: Jake Steinman <[email protected]> > > The HM1092 is a 648x368 mono IR sensor used for face unlock on recent > Intel IPU6/IPU7 laptops, sitting behind a Synaptics or Intel vision > bridge. Without an entry here ipu-bridge builds the sensor endpoint with > no link-frequencies property, the driver's frequency validation rejects > every mode, and the sensor never binds. > > Two frequencies are listed because the sensor is driven from different > external clocks depending on the platform, and the PLL registers are > programmed to match: > > 0x030d (pre-divider), 0x030f (multiplier), link_freq = EXTCLK * m / p > > Dell, IPU7 EXTCLK 19.2 MHz, p=10, m=94 -> 180,480,000 Hz > ASUS, Qualcomm EXTCLK 24.0 MHz, p=12, m=90 -> 180,000,000 Hz Please avoid commas in numbers unless there are non-integer parts in them. > > Both are the CSI-2 DDR clock, so the per-lane bit rate is twice the > value, 360.96 and 360 Mbit/s respectively, over one data lane. > > Worth stating explicitly because published figures for this part have > been wrong: V4L2_CID_LINK_FREQ is the DDR clock, while the vendor mode > descriptors publish the per-lane BIT rate. Taking the descriptor value > verbatim gives 360,960,000, isys then configures the D-PHY for 721 Mbit/s > against a sensor transmitting 361, the clock lane toggles, and no frame > is ever framed. Halving it to the DDR clock is what produced the first > captured frames from this sensor on Linux. > > The two platform values differ by 0.27%, which is within what the > sensor's PLL tolerates, but they are listed separately rather than > merged so each platform matches exactly what its firmware programs. Is Documentation/process/coding-assistants.rst relevant for this? A lot of this could be on a cover letter (or after '---') but the commit messages should be more concise than this. > > Depends on the HM1092 sensor driver, currently in review. > > Signed-off-by: Jake Steinman <[email protected]> > --- > drivers/media/pci/intel/ipu-bridge.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/drivers/media/pci/intel/ipu-bridge.c b/drivers/media/pci/intel/ipu-bridge.c > --- a/drivers/media/pci/intel/ipu-bridge.c > +++ b/drivers/media/pci/intel/ipu-bridge.c > @@ -50,6 +50,8 @@ > * Please keep the list sorted by ACPI HID. > */ > static const struct ipu_sensor_config ipu_supported_sensors[] = { > + /* Himax HM1092 */ > + IPU_SENSOR_CONFIG("HIMX1092", 2, 180000000, 180480000), > /* Himax HM11B1 */ > IPU_SENSOR_CONFIG("HIMX11B1", 1, 384000000), > /* Himax HM2170 */ -- Regards, Sakari Ailus