| Message ID | 20200427124931.115697-7-amirmizi6@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Add tpm i2c ptp driver | expand |
On Mon, Apr 27, 2020 at 03:49:30PM +0300, amirmizi6@gmail.com wrote: > From: Amir Mizinski <amirmizi6@gmail.com> > > Added a YAML schema to support tpm tis i2c related dt-bindings for the I2c > PTP based physical layer. > > This patch adds the documentation for corresponding device tree bindings of > I2C based Physical TPM. > Refer to the 'I2C Interface Definition' section in > 'TCG PC Client PlatformTPMProfile(PTP) Specification' publication > for specification. Again, DT bindings describe h/w devices, not just a protocol. A device is more than just a protocol interface. There's clocks, power rails, resets, interrupts, firmware, etc. Unless there's something special about TPM chips that makes none of this applicable and no chip will ever have any quirks (or extensions) in their protocol to work-around, then you need compatible string(s) that are specific to the TPM chip. You can have tcg,tpm-tis-i2c as a fallback, but you need specific compatible to handle any quirks. Rob
On 2020-05-05 16:12, Rob Herring wrote: > On Mon, Apr 27, 2020 at 03:49:30PM +0300, amirmizi6@gmail.com wrote: >> From: Amir Mizinski <amirmizi6@gmail.com> >> >> Added a YAML schema to support tpm tis i2c related dt-bindings for the I2c >> PTP based physical layer. >> >> This patch adds the documentation for corresponding device tree bindings of >> I2C based Physical TPM. >> Refer to the 'I2C Interface Definition' section in >> 'TCG PC Client PlatformTPMProfile(PTP) Specification' publication >> for specification. > > Again, DT bindings describe h/w devices, not just a protocol. A device > is more than just a protocol interface. There's clocks, power rails, > resets, interrupts, firmware, etc. > > Unless there's something special about TPM chips that makes none of this > applicable and no chip will ever have any quirks (or extensions) in > their protocol to work-around, then you need compatible string(s) that > are specific to the TPM chip. You can have tcg,tpm-tis-i2c as a > fallback, but you need specific compatible to handle any quirks. > > Rob Hello Rob, currently yes. All TPM chip are implemented according to the TGC specs and should use the same properties for this I2C driver. I can't say for sure that it will be the case in the future. Shouldn't we use the standard "tcg,tpm-tis-i2c" compatible, and if a specific TPM chip will deviate from the specs, the vendor should add an additional compatible string for it? Thank you, Amir
On Wed, May 6, 2020 at 10:20 AM Amir Mizinski <amirmizi6@gmail.com> wrote: > > > On 2020-05-05 16:12, Rob Herring wrote: > > On Mon, Apr 27, 2020 at 03:49:30PM +0300, amirmizi6@gmail.com wrote: > >> From: Amir Mizinski <amirmizi6@gmail.com> > >> > >> Added a YAML schema to support tpm tis i2c related dt-bindings for the I2c > >> PTP based physical layer. > >> > >> This patch adds the documentation for corresponding device tree bindings of > >> I2C based Physical TPM. > >> Refer to the 'I2C Interface Definition' section in > >> 'TCG PC Client PlatformTPMProfile(PTP) Specification' publication > >> for specification. > > > > Again, DT bindings describe h/w devices, not just a protocol. A device > > is more than just a protocol interface. There's clocks, power rails, > > resets, interrupts, firmware, etc. > > > > Unless there's something special about TPM chips that makes none of this > > applicable and no chip will ever have any quirks (or extensions) in > > their protocol to work-around, then you need compatible string(s) that > > are specific to the TPM chip. You can have tcg,tpm-tis-i2c as a > > fallback, but you need specific compatible to handle any quirks. > > > > Rob > > Hello Rob, currently yes. All TPM chip are implemented according to the TGC specs and should use the same properties for this I2C driver. > I can't say for sure that it will be the case in the future. Exactly. That's the issue. If you have just "tcg,tpm-tis-i2c" and need to handle some difference in the future, then you can't without updating the DT. You must be able to handle future issues without updating the DT. > Shouldn't we use the standard "tcg,tpm-tis-i2c" compatible, and if a specific TPM chip will deviate from the specs, the vendor should add an additional compatible string for it? Name something where multiple vendors have implemented a spec and there's no deviation. It simply doesn't exist. How would you know? Does the TPM spec define all the things I listed above outside of just the I2C protocol? Also, what version of the spec is "tcg,tpm-tis-i2c"? Few specs have only 1 version. Rob
diff --git a/Documentation/devicetree/bindings/security/tpm/tpm-tis-i2c.yaml b/Documentation/devicetree/bindings/security/tpm/tpm-tis-i2c.yaml new file mode 100644 index 0000000..2ab8b80 --- /dev/null +++ b/Documentation/devicetree/bindings/security/tpm/tpm-tis-i2c.yaml @@ -0,0 +1,47 @@ +# SPDX-License-Identifier: GPL-2.0 +%YAML 1.2 +--- +$id: http://devicetree.org/schemas/security/tpm/tpm-tis-i2c.yaml# +$schema: http://devicetree.org/meta-schemas/core.yaml# + +title: I2C PTP based TPM Device Tree Bindings + +maintainers: + - Amir Mizinski <amirmizi6@gmail.com> + +description: + Device Tree Bindings for I2C based Trusted Platform Module(TPM). + +properties: + compatible: + contains: + const: tcg,tpm-tis-i2c + + reg: + maxItems: 1 + + interrupt: + maxItems: 1 + + crc-checksum: + $ref: /schemas/types.yaml#/definitions/flag + description: + Set this flag to enable CRC checksum. + +required: + - compatible + - reg + +examples: + - | + i2c { + #address-cells = <1>; + #size-cells = <0>; + + tpm_tis@2e { + compatible = "tcg,tpm-tis-i2c"; + reg = <0x2e>; + crc-checksum; + }; + }; +...