Re: [PATCH v2 00/15] Intel IPU6 and IPU6 input system drivers

From: Hans de Goede <hdegoede@redhat.com>
To: Bingbu Cao <bingbu.cao@linux.intel.com>,
	Hans de Goede <hans@hansg.org>,
	bingbu.cao@intel.com, linux-media@vger.kernel.org,
	sakari.ailus@linux.intel.com, laurent.pinchart@ideasonboard.com
Cc: andriy.shevchenko@linux.intel.com, ilpo.jarvinen@linux.intel.com,
	andreaskleist@gmail.com, claus.stovgaard@gmail.com,
	tfiga@chromium.org, senozhatsky@chromium.org,
	tomi.valkeinen@ideasonboard.com, tian.shu.qiu@intel.com,
	hongju.wang@intel.com
Subject: Re: [PATCH v2 00/15] Intel IPU6 and IPU6 input system drivers
Date: Mon, 8 Jan 2024 15:23:59 +0100	[thread overview]
Message-ID: <445d7cbf-8d68-4b55-a252-7187973d63c4@redhat.com> (raw)
In-Reply-To: <2d3a85fe-dfb9-52e1-fe1b-e0b9a5ac14ab@linux.intel.com>

Hi Bingbu,

On 1/8/24 05:07, Bingbu Cao wrote:
> Hans,
> 
> On 11/8/23 7:59 PM, Hans de Goede wrote:
>> Hi Bingbu,
>>
>> On 10/24/23 13:29, bingbu.cao@intel.com wrote:
>>> From: Bingbu Cao <bingbu.cao@intel.com>
>>>
>>> This patch series adds a driver for Intel IPU6 input system.
>>> IPU6 is the sixth generation of Imaging Processing Unit, it is a PCI
>>> device which can be found in some Intel Client Platforms. User can use
>>> IPU6 to capture images from MIPI camera sensors.
>>>
>>> IPU6 has its own firmware which exposes ABIs to driver, and communicates
>>> with CSE to do firmware authentication. IPU6 has its MMU hardware, so
>>> the driver sets up a page table to allow IPU6 DMA to access the system
>>> memory.
>>>
>>> IPU6 input system driver uses MC and V4L2 sub-device APIs besides V4L2.
>>
>> I have been testing this on a TigerLake system, a Dell Latitude 9420
>> with ov01a1s and the packed 10 bit bayer pixel fmt: V4L2_PIX_FMT_SGRBG10P,
>> which libcamera together with (WIP) software debayer support picks
>> by default does not work. There are many many CSI errors in dmesg
>> and only the first 10 or so lines of the image show.
>>
>> Disabling the packed format by removing it from ipu6_isys_pfmts[],
>> making libcamera pick the unpacked (every 10 bits per pixel data
>> stored in a 16 bit word in output buffer) fixes this.
>>
>> Are the packed bayer formats supposed to work on Tiger Lake, or
>> is this a known issue which Intel's own userspace stack avoids
>> by always picking the unpacked format ?
> 
> I just tested the packed bayer on my device (ov01a10) and did not
> observe the problems like yours, could you share the isys kernel log?

Was this on a device with a Tiger Lake CPU / plain ipu6 (not ipu6ep) ?

Regards,

Hans