On Fri Aug 2, 2024 at 6:19 PM EEST, Jett Rink wrote:

Could I get some feedback on this patch please? Is there something I
am not doing correctly?

Sorry, I just came from holidays.

-Jett

On Thu, Jul 18, 2024 at 2:24 PM Jett Rink [off-list ref] wrote:

quoted

Accept another DID:VID for the next generation Google TPM. This TPM
has the same Ti50 firmware and fulfills the same interface.

Signed-off-by: Jett Rink <redacted>
---

Changes in v2:
Patchset 2 applies cleanly

 drivers/char/tpm/tpm_tis_i2c_cr50.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/drivers/char/tpm/tpm_tis_i2c_cr50.c b/drivers/char/tpm/tpm_tis_i2c_cr50.c
index adf22992138e..b50005ccfc5e 100644
--- a/drivers/char/tpm/tpm_tis_i2c_cr50.c
+++ b/drivers/char/tpm/tpm_tis_i2c_cr50.c

@@ -31,7 +31,8 @@
 #define TPM_CR50_TIMEOUT_SHORT_MS      2               /* Short timeout during transactions */
 #define TPM_CR50_TIMEOUT_NOIRQ_MS      20              /* Timeout for TPM ready without IRQ */
 #define TPM_CR50_I2C_DID_VID           0x00281ae0L     /* Device and vendor ID reg value */
-#define TPM_TI50_I2C_DID_VID           0x504a6666L     /* Device and vendor ID reg value */
+#define TPM_TI50_DT_I2C_DID_VID                0x504a6666L     /* Device and vendor ID reg value */
+#define TPM_TI50_OT_I2C_DID_VID                0x50666666L     /* Device and vendor ID reg value */
 #define TPM_CR50_I2C_MAX_RETRIES       3               /* Max retries due to I2C errors */
 #define TPM_CR50_I2C_RETRY_DELAY_LO    55              /* Min usecs between retries on I2C */
 #define TPM_CR50_I2C_RETRY_DELAY_HI    65              /* Max usecs between retries on I2C */

@@ -741,14 +742,18 @@ static int tpm_cr50_i2c_probe(struct i2c_client *client)
        }

        vendor = le32_to_cpup((__le32 *)buf);
-       if (vendor != TPM_CR50_I2C_DID_VID && vendor != TPM_TI50_I2C_DID_VID) {
+       if (vendor != TPM_CR50_I2C_DID_VID &&
+           vendor != TPM_TI50_DT_I2C_DID_VID &&
+           vendor != TPM_TI50_OT_I2C_DID_VID) {
                dev_err(dev, "Vendor ID did not match! ID was %08x\n", vendor);
                tpm_cr50_release_locality(chip, true);
                return -ENODEV;
        }

        dev_info(dev, "%s TPM 2.0 (i2c 0x%02x irq %d id 0x%x)\n",
-                vendor == TPM_TI50_I2C_DID_VID ? "ti50" : "cr50",
+                vendor == TPM_CR50_I2C_DID_VID    ? "cr50" :
+                vendor == TPM_TI50_DT_I2C_DID_VID ? "ti50 DT" :
+                                                    "ti50 OT",

Whenever possible ternary operator should be avoided, unless the use
case super trivial: this complexity brings us zero measurable benefit
and unnecessarily obfuscates code.

I'd suggest to simply add a helper:

/*
 * Maps VID to a name.
 */
const char *tpm_cr50_vid_to_name(u32 vendor)
{
	switch (vendor) {
	case TPM_CR50_I2C_DID_VID:
		return "cr50";
	case TPM_TI50_DT_I2C_DID_VID:
		return "ti50 DT";
	case TPM_TI50_OT_I2C_DID_VID:
		return "ti50 OT";
	default:	
		break;
	}

	tpm_cr50_release_locality(chip, true);
	return NULL;
}

The code then transforms to:

	vendor = le32_to_cpup((__le32 *)buf);

	name = tpm2_cr50_vid_to_name(vendor);
	if (name == NULL) {
		dev_err(dev, "Vendor ID did not match! ID was %08x\n", vendor);
		return -ENODEV;
	}

	dev_info(dev, "%s TPM 2.0 (i2c 0x%02x irq %d id 0x%x)\n", name.
		 client->addr, client->addr, client->irq, vendor >> 16);

[and add suggested-by]

quoted

                 client->addr, client->irq, vendor >> 16);
        return tpm_chip_register(chip);
 }
--
2.45.2.1089.g2a221341d9-goog

BR, Jarkko