On Wed, Nov 6, 2019 at 2:09 AM <[email protected]> wrote:
> Hi All
>
> I have to following problem. I'm trying to setup a custom cape with and
> SPI TPM based on SLB9670VQ2.0 chip.
>
> What I've done so far is:
> 1 - building and deploying a custom kernel with SPI TPM 2.0 support
> enabled.
> 2 - Writing a dtbo to configure the right pins in their right modes for
> the TPM and to instruct the u-boot to load the specific kernel objects.
>
> After many problems and a lot of research, I managed to to build the DTBO
> activate it by including
> uboot_overlay_addr5=/lib/firmware/TPM-SLB9670-00A0.dtbo in the
> /boot/uEnv.txt and making sure the DTBO is loaded and does it's thing.
>
> Now, the DTBO IS loading and causes the respective modules to be loaded:
>
> ivan@beaglebone:~$ lsmod
> Module Size Used by
> 8021q 32768 0
> garp 16384 1 8021q
> stp 16384 1 garp
> mrp 20480 1 8021q
> llc 16384 2 garp,stp
> usb_f_acm 16384 2
> u_serial 20480 3 usb_f_acm
> usb_f_ecm 20480 2
> usb_f_rndis 32768 4
> u_ether 20480 2 usb_f_ecm,usb_f_rndis
> libcomposite 61440 16 usb_f_ecm,usb_f_acm,usb_f_rndis
> iptable_nat 16384 0
> nf_nat_ipv4 16384 1 iptable_nat
> nf_nat 36864 1 nf_nat_ipv4
> nf_conntrack 155648 2 nf_nat_ipv4,nf_nat
> nf_defrag_ipv6 20480 1 nf_conntrack
> nf_defrag_ipv4 16384 1 nf_conntrack
> iptable_mangle 16384 0
> iptable_filter 16384 0
>
>
>
> *tpm_tis_spi 16384 0tpm_tis_core 24576 1
> tpm_tis_spitpm 57344 2
> tpm_tis_spi,tpm_tis_core*uio_pdrv_genirq
> 16384 0
> uio 20480 1 uio_pdrv_genirq
> ip_tables 24576 3 iptable_mangle,iptable_filter,iptable_nat
> x_tables 32768 3 iptable_mangle,ip_tables,iptable_filter
>
> While researching, I was led to believe that the modules would create the
> "all-important" device file in the /dev directory, i.e. /dev/tpm0
> But that doesn't happen?!
>
Can you also provide the dmesg log output? Since the module is loaded, I'd
be inclined to think it would create whatever interfaces it would
provide...if the probe was successful.
> I tried modifying the dts file for my dtbo in every way I could imagine.
> It doesn't work.
>
> This is the complete dts file for my TPM dtbo:
>
> /*
> * Device Tree overlay for Infineon SLx 9670
> */
>
> #include <dt-bindings/board/am335x-bbw-bbb-base.h>
> #include <dt-bindings/interrupt-controller/irq.h>
> #include <dt-bindings/pinctrl/am33xx.h>
> #include <dt-bindings/gpio/gpio.h>
>
> /dts-v1/;
> /plugin/;
>
> / {
> compatible = "ti,beaglebone", "ti,beaglebone-black";
>
> /* Identification */
> part-number = "TPM-SLB9670";
> version = "00A0";
>
> exclusive-use = "P9.21", // Used as SPI0 MISO and is the Default +/
> "P9.18", // Used as SPI0 MOSI and is the Default +/
> "P9.22", // Used as SPI0 SCKL and is the Default +/
> "P9.17", // Used as SPI0 CS0 +/
>
> "P9.15", // GPIO1_16 aka GPIO_48 used as input to
> handle the IRQ from the TPM.
> "P9.13", // GPIO0_31 aka GPIO_31 used as output to
> drive the reset pin of the TPM
>
> "spi0",
> "i2c0",
> "i2c2";
>
> fragment@0 // Releasing the spi0 pins from their current
> assignments. (Tried without this fragment no difference)
> {
> target = <&ocp>;
>
> __overlay__
> {
> // P9_21_pinmux { status = "disabled"; };
> // P9_18_pinmux { status = "disabled"; };
> // P9_22_pinmux { status = "disabled"; };
> // P9_17_pinmux { status = "disabled"; };
>
Odd. I'd think you'd need to disable the pinmuxes if you want the default
to be SPI mode.
Check with 'sudo /opt/scripts/device/bone/show-pins.pl'. I really need to
create a flag that removes the color printing on that utility and put it in
the standard path. Anyway, strip the color info and provide the output for
all the relevant pins.
>
> P9_15_pinmux { status = "disabled"; };
> P9_13_pinmux { status = "disabled"; };
> };
> };
>
> fragment@1 // Disabling both i2c interfaces since there are some
> pin conflicts, in the off chance they moght interfere (Tried without those
> two fragments - No difference)
> {
> target = <&i2c0>;
>
> __overlay__
> {
> status = "disabled";
>
Isn't this the on-board i2c? Disabling that would be bad.
> };
> };
>
> fragment@2
> {
> target = <&i2c2>;
>
> __overlay__
> {
> status = "disabled";
> };
> };
>
> fragment@3
> {
> target = <&am33xx_pinmux>;
>
> __overlay__
> {
> pb_spi0_pins: pinmux_pb_spi0_pins // I checked if
> a got the right pins several times, tried with or without pullups no
> difference
> { // My only doubt
> is did I get the inputs anoutputs right?
> pinctrl-single,pins = < BONE_P9_21
> (PIN_INPUT_PULLUP | MUX_MODE0) // Used as SPI0 MISO +/
> BONE_P9_18
> (PIN_OUTPUT_PULLUP | MUX_MODE0) // Used as SPI0 MOSI +/
> BONE_P9_22
> (PIN_OUTPUT_PULLUP | MUX_MODE0) // Used as SPI0 SCKL +/
> BONE_P9_17
> (PIN_OUTPUT | MUX_MODE0)>; // Used as SPI0 CS0 (This pin has a physical
> pullup resistor) +/
>
> // BONE_P9_15
> (PIN_INPUT_PULLUP | MUX_MODE7) // Used as GPIO_48 for IRQ to TPM
> // BONE_P9_13
> (PIN_OUTPUT_PULLUP | MUX_MODE7)>; // Used as GPIO_31 for RST to TPM
>
};
> };
> };
>
> fragment@4 // Every example I looked at has the spi configuring
> fragmet as separate form the actual TPM config fragmet even though they can
> be one...
> { // Here I assign the actual pins to the SPI interface
> and disable both channels since I want the TPM to be the active device
> rther than having spidev in /dev
>
Just remember that there are 2 drivers, one for the SPI controller and one
for the TPM.
> target = <&spi0>;
>
> __overlay__
> {
> pinctrl-names = "default";
> pinctrl-0 = <&pb_spi0_pins>;
> status = "okay";
>
> #address-cells = <1>;
> #size-cells = <0>;
>
> channel@0
> {
> compatible = "spidev";
> reg = <0x0>;
> status = "disabled";
>
> spi-max-frequency = <32000000>;
> };
>
> channel@1
> {
> compatible = "spidev";
> reg = <0x1>;
> status = "disabled";
>
> spi-max-frequency = <32000000>;
> };
>
disabling the spidev drivers shouldn't need any of the other arguments.
Also, I think you'd want to be targeting the spidev symbols to overwrite
those values. Can you confirm the spidev drivers aren't being loaded?
> };
> };
>
> fragment@5 // The main fragment. Configuring the
> TPM. I did it as per the example given by those who wrote the SPI kernel
> modules.
> { // That's how I managed to get the dtbo
> to load the actual modules. All my previous versions of this fragment
> failed to do that.
> target = <&spi0>; // I did try wtihout the gpio interupt
> stuf but it didn't work either.
>
> __overlay__
> {
> #address-cells = <1>;
> #size-cells = <0>;
>
> tpm_tis@0
> {
> compatible = "tcg,tpm_tis-spi"; //
> tcg,tpm_spi_tis,infineon,slb9670
> status = "okay";
>
> #gpio-cells = <2>;
> #interrupt-cells = <2>;
>
why?
>
> reg = <0x0>;
> spi-max-frequency = <32000000>;
> num-cs = <1>;
>
Is this needed?
> interrupt-parent = <&gpio1>;
> interrupts = < 16 2>;
>
I don't see this in the bindings:
https://github.com/torvalds/linux/blob/master/Documentation/devicetree/bindings/security/tpm/tpm_tis_spi.txt
> gpios = <&gpio1 16 GPIO_ACTIVE_LOW>,
> <&gpio0 31 GPIO_ACTIVE_LOW>;
>
I don't see this in the bindings either. Where are you pulling the
instructions from?
> };
> };
> };
> };
>
> dmesg gives no relevant information.
>
are you sure?
>
> Thanks in advance and sorsy for long post :)
>
> --
> For more options, visit http://beagleboard.org/discuss
> ---
> You received this message because you are subscribed to the Google Groups
> "BeagleBoard" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/beagleboard/f26111dd-7119-47a2-9205-2658acd90a47%40googlegroups.com
> <https://groups.google.com/d/msgid/beagleboard/f26111dd-7119-47a2-9205-2658acd90a47%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
--
https://beagleboard.org/about - a 501c3 non-profit educating around open
hardware computing
--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups
"BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/beagleboard/CA%2BT6QPnRMgMra%2B76uPkG_PNznBBtoCzv0b3rwC2nzY1QwVPoKQ%40mail.gmail.com.
