Re: TPM driver breaks S3 suspend

[Daniel Drake]

Hi,

Sorry for the delayed response.

Unfortunately we had to return the affected machine to the vendor
(Acer Veriton X4110G) although given we have a fairly complete
understanding of the problem, perhaps we can find and push a fix
anyway.

Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx> wrote:
> Can you try it on git://git.infradead.org/users/jjs/linux-tpmdd.git ?
> Maybe it is related to the LPC #CLKRUN issues we've had.

I don't see anything relevant in that git tree, and it seems like
a distant issue from the CLKRUN changes that have been made.

To recap, the first suspend works OK, and then the TPM shows shows the
following message on the first resume:
  tpm tpm0: A TPM error(38) occured continue selftest
And error(38) is TPM_ERR_INVALID_POSTINIT which means the TPM is
not in the correct state. Then the following suspend will always
fail with the following messages.
  tpm tpm0: Error (38) sending savestate before suspend
  PM: Device 00:0b failed to suspend: error 38

tpm_crb is the driver here (not tpm2).

This exact problem was also discussed in depth years ago at
https://lists.gt.net/linux/kernel/1358716

The patch that we used to avoid the issue is below. What do you think?

Thanks
Daniel


---
 drivers/char/tpm/tpm-interface.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c
index 1d6729be4cd6..444ae2df4a06 100644
--- a/drivers/char/tpm/tpm-interface.c
+++ b/drivers/char/tpm/tpm-interface.c
@@ -953,6 +953,10 @@ int tpm_do_selftest(struct tpm_chip *chip)
 	loops = jiffies_to_msecs(duration) / delay_msec;
 
 	rc = tpm_continue_selftest(chip);
+	if (rc == TPM_ERR_INVALID_POSTINIT) {
+		chip->flags |= TPM_CHIP_FLAG_ALWAYS_POWERED;
+		dev_info(&chip->dev, "TPM not ready (%d)\n", rc);
+	}
 	/* This may fail if there was no TPM driver during a suspend/resume
 	 * cycle; some may return 10 (BAD_ORDINAL), others 28 (FAILEDSELFTEST)
 	 */
-- 
2.14.1