Hi Heinrich
On Mon, 8 Nov 2021 at 17:55, Heinrich Schuchardt xypron.glpk@gmx.de wrote:
On 11/7/21 22:33, Ilias Apalodimas wrote:
With the upcoming TPM2 API, some of the function names are part of the new header file. So switch conflicting driver defined function names and defines.
Signed-off-by: Ilias Apalodimas ilias.apalodimas@linaro.org
drivers/tpm/tpm_tis_infineon.c | 34 +++++++++++++++++----------------- 1 file changed, 17 insertions(+), 17 deletions(-)
diff --git a/drivers/tpm/tpm_tis_infineon.c b/drivers/tpm/tpm_tis_infineon.c index f414e5657db9..aa66e931bada 100644 --- a/drivers/tpm/tpm_tis_infineon.c +++ b/drivers/tpm/tpm_tis_infineon.c @@ -50,10 +50,10 @@ static const char * const chip_name[] = { [UNKNOWN] = "unknown/fallback to slb9635", };
-#define TPM_ACCESS(l) (0x0000 | ((l) << 4)) -#define TPM_STS(l) (0x0001 | ((l) << 4)) -#define TPM_DATA_FIFO(l) (0x0005 | ((l) << 4)) -#define TPM_DID_VID(l) (0x0006 | ((l) << 4)) +#define TPM1_ACCESS(l) (0x0000 | ((l) << 4)) +#define TPM1_STS(l) (0x0001 | ((l) << 4)) +#define TPM1_DATA_FIFO(l) (0x0005 | ((l) << 4)) +#define TPM1_DID_VID(l) (0x0006 | ((l) << 4))
To what does '1' relate here?
- Do you mean TPMv1?
Yea that was the intention
- Or are the constants Infenion specific?
I think so. I can't find that definition in any of the TCG specs [1] [2] (search for 0f04 to get to the description table).
If the constants are TPMv1 specific, they should be in a header file. If the constants are Infineon specific we could put _INFINEON_ into the constant name.
Ok I'll swap 1 with INFINEON
Linux has resolved the conflict by putting the TPMv2 TIS constants into drivers/char/tpm/tpm_tis_core.h which is not included by drivers/char/tpm/tpm_i2c_infineon.c.
Yea and they still define them differently into that infineon TPM.
[1] https://trustedcomputinggroup.org/wp-content/uploads/TCG_PCClientTPMInterfac... [2] https://trustedcomputinggroup.org/wp-content/uploads/PC-Client-Specific-Plat...
Thanks /Ilias
Best regards
Heinrich
[...]