erebion/coreboot
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

drivers/pc80/tpm: Fix tis_readresponse()

Browse source 
TPM_RDRESP_NEED_DELAY was introduced in 2018 in CB:25322 after
observing errors with SLB9635 by Infineon. It has been confirmed
also SLB9670 and SLB9672 require a fix or delay here.

Presumably, prior to CB:4388 SLB9635 did not have this problem,
as this particular TPM shipped with samsung/lumpy Chromebook since
2011. In CB:4388 the code changed from polling the status register
(+burst_count) using a 32bit read to separated 8bit reads.

So far, experiments on samsung/lumpy and SLB9635 indicate that
it would be sufficient to add a single tpm_read_status() call
to see TIS_STS_DATA_AVAILABLE as set at the time of evaluating
the loop exit condition.

Change-Id: If5c3e93c7946ebf8226f7bba47b38253f6920c61
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Co-authored-by: Bill XIE <persmule@hardenedlinux.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76315
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Bill XIE <persmule@hardenedlinux.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
This commit is contained in:
Kyösti Mälkki 2023-07-06 06:20:31 +03:00 • committed by Martin L Roth
commit 71c8dac9ca
1 changed files with 8 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

16
src/drivers/pc80/tpm/tis.c
View file

@ -286,6 +286,14 @@ static inline tpm_result_t tis_wait_valid_data(int locality)
static inline int tis_has_valid_data(int locality)
{
const u8 has_data = TIS_STS_DATA_AVAILABLE | TIS_STS_VALID;
/*
* Certain TPMs require a small delay here, as they have set
* TIS_STS_VALID first and TIS_STS_DATA_AVAILABLE few clocks later.
*/
if ((tpm_read_status(locality) & has_data) == has_data)
return 1;
return (tpm_read_status(locality) & has_data) == has_data;
}
@ -635,14 +643,6 @@ static tpm_result_t tis_readresponse(u8 *buffer, size_t *len)
if (offset == expected_count)
break; /* We got all we need */
/*
* Certain TPMs seem to need some delay between tis_wait_valid()
* and tis_has_valid_data(), or some race-condition-related
* issue will occur.
*/
if (CONFIG(TPM_RDRESP_NEED_DELAY))
udelay(10);
} while (tis_has_valid_data(locality));
/* * Make sure we indeed read all there was. */