On Wed, Apr 01, 2020 at 10:35:23PM +0200, Tom Warren wrote:
-----Original Message----- From: Thierry Reding treding@nvidia.com Sent: Wednesday, April 1, 2020 8:20 AM To: Tom Warren TWarren@nvidia.com Cc: Peter Robinson pbrobinson@gmail.com; tomcwarren3959@gmail.com; u-boot@lists.denx.de; Stephen Warren swarren@nvidia.com; Jonathan Hunter jonathanh@nvidia.com; Vishruth Jain vishruthj@nvidia.com Subject: Re: [PATCH 3/3] ARM: tegra: p2371-2180: add I2C nodes to DT
On Wed, Apr 01, 2020 at 02:03:09AM +0200, Tom Warren wrote:
-----Original Message----- From: Peter Robinson pbrobinson@gmail.com Sent: Tuesday, March 31, 2020 3:54 AM To: tomcwarren3959@gmail.com Cc: u-boot@lists.denx.de; Stephen Warren swarren@nvidia.com; Thierry Reding treding@nvidia.com; Jonathan Hunter jonathanh@nvidia.com; Tom Warren TWarren@nvidia.com; Vishruth Jain vishruthj@nvidia.com Subject: Re: [PATCH 3/3] ARM: tegra: p2371-2180: add I2C nodes to DT
External email: Use caution opening links or attachments
From: Stephen Warren swarren@nvidia.com
This adds to the DT the I2C controllers that connect to the board ID EEPROM, camera board EEPROM, etc. With this change, you can now probe all I2C devices on a TX1 board.
Signed-off-by: Tom Warren twarren@nvidia.com
arch/arm/dts/tegra210-p2371-2180.dts | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)
diff --git a/arch/arm/dts/tegra210-p2371-2180.dts b/arch/arm/dts/tegra210-p2371-2180.dts index c2f497c..d982b5f 100644 --- a/arch/arm/dts/tegra210-p2371-2180.dts +++ b/arch/arm/dts/tegra210-p2371-2180.dts @@ -12,6 +12,9 @@
aliases { i2c0 = "/i2c@7000d000";
i2c2 = "/i2c@7000c400";i2c3 = "/i2c@7000c500";i2c5 = "/i2c@546c0c00";I don't think this is correct, it doesn't show up in U-Boot with the "i2c bus" command where the others do, looking in the tegra210.dtsi it looks like it should be i2c@546c0000? [Tom] That I2C address is working in downstream (L4T) TX1 U-Boot. The VI_I2C controller is a little weird, it's normal I2C registers are offset from base by 0xC00. A different driver is needed, but I haven't posted it yet upstream. I should probably drop if from the DTS for now until I post the VI_I2C driver.
I think the problem here is that the upstream U-Boot device tree doesn't contain an i2c@546c0c00 node. Instead it has i2c@546c0000, which we also have in the upstream kernel. My recollection is that that's also the address listed in the Tegra210 system address map of the TRM and there are some registers before the regular I2C interface at offset 0xc00.
I've been carrying a patch against the upstream Linux I2C controller driver to special-case the VI/I2C to always add that 0xc00 offset when accessing registers, which allows us to reuse the existing driver and at the same time keeps all registers mapped so we can also access the VI/I2C specific registers.
My recollection is that the U-Boot driver is fairly similar to the Linux driver, so I suspect something similar could be done.
Thierry [Tom] Thanks, Thierry. That's my recollection, too, about the VI_I2C 0xC00 offset. I'll take a look at what we have in L4T U-Boot for T210 and address it in a set of patches for upstream soon. That I2C controller isn't used for anything on any Jetson board except on TX1, I believe, where it allows U-Boot to talk to the camera add-in board to read the board ID. But we've moved all that out to CBoot (board ID EEPROM reading), so there isn't a pressing need for it in U-Boot anymore, IIRC.
We've had internal discussions about potentially using the ID EEPROMs to dynamically modify the kernel DTB at runtime (possibly from U-Boot) in order to support things like different add-in boards.
For example, we currently enable the DSI output in Linux for Jetson TX1. However, the Jetson TX1 developer kits don't actually ship with that DSI display (I don't think they even come with the display add-in card), so this can lead to confusion. The idea was to have U-Boot probe ID EEPROMs from several sources to determine what to add to the device tree, so that the kernel DT *source* would only contain the standard developer kit hardware, but U-Boot (or something else perhaps) could add in extra nodes for a display module, or camera add-in board, etc.
Note that this is just a vague idea at this time and nothing concrete has been done to implement this, yet.
Thierry