lkml.org 
[lkml]   [2022]   [Dec]   [14]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH 3/4] arm64: dts: qcom: sa8540p-ride: add qup1_i2c15 and qup2_i2c18 nodes
From


On 12/14/2022 6:00 PM, Brian Masney wrote:
> On Tue, Dec 13, 2022 at 03:48:27PM +0100, Konrad Dybcio wrote:
>>> + qup1_i2c15_default: qup1-i2c15-state {
>>> + mux-pins {
>>> + pins = "gpio36", "gpio37";
>>> + function = "qup15";
>>> + };
>>> +
>>> + config-pins {
>>> + pins = "gpio36", "gpio37";
>>> + drive-strength = <0x02>;
>>> + bias-pull-up;
>>> + };
>>> + };
>>
>> You can drop mux/config-pins and have the pin properties live directly
>> under the qup1-i2cN-state node.
>
> Hi Konrad (and Shazad below),
>
> I need to enable 5 i2c buses (0, 1, 12, 15, 18) on this board. I tried
> the following combinations with the pin mapping configuration and the
> only one that seems to work reliably for me is what I originally had.
>
> With the following, only 2 out of the 5 buses are detected. There's no
> i2c mesages in dmesg.
>
> i2c0_default: i2c0-default-state {
> pins = "gpio135", "gpio136";
> function = "qup15";
> };
>
> Next, I added a drive-strength and bias-pull-up. All 5 buses are
> detected. One bus throws read errors when I probe it with i2cdetect, two
> others 'i2cdetect -a -y $BUSNUM' takes ~5 seconds to run, and the

This I have also observed on downstream as well, where scanning all
addresses takes some amount of time near to 5-6 seconds.

> remaining two are fast.
>
> i2c0_default: i2c0-default-state {
> pins = "gpio135", "gpio136";
> function = "qup15";
> drive-strength = <2>;
> bias-pull-up;
> };
>

This is the default config we should use.

> This is the style where i2cdetect seems to be happy for all 5 buses and
> is fast:
>
> i2c0_default: i2c0-default-state {
> mux-pins {
> pins = "gpio135", "gpio136";
> function = "qup0";
> };
>
> config-pins {
> pins = "gpio135", "gpio136";
> drive-strength = <2>;
> bias-pull-up;
> };
> };
>
>
> Shazad: 'i2cdetect -a -y $BUSNUM) shows that all 5 buses have the same
> addresses listening. Is that expected? That seems a bit odd to me.
>

Brian, even I haven't checked with all enabled, let me check this on
other projects and with downstream as well and get back to you.

-Shazad

> [root@localhost ~]# i2cdetect -a -y 0
> Warning: Can't use SMBus Quick Write command, will skip some addresses
> 0 1 2 3 4 5 6 7 8 9 a b c d e f
> 00:
> 10:
> 20:
> 30: -- -- -- -- -- -- -- --
> 40:
> 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
> 60:
> 70:
>
> I triple checked that I have the QUP pins defined correctly for the 5
> buses. I checked them against what's in the downstream kernel and I also
> checked them against what's in upstream's
> drivers/pinctrl/qcom/pinctrl-sc8280xp.c. This is the pin mapping that I
> have:
>
> i2c0: gpio135, gpio136
> i2c1: gpio158, gpio159
> i2c12: gpio0, gpio1
> i2c15: gpio36, gpio37
> i2c18: gpio66, gpio67
>
> Brian
>

\
 
 \ /
  Last update: 2022-12-14 16:37    [W:0.103 / U:1.404 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site