BUG: Wrong values in TPM2_PT enum

[SharkOverBite]

I'm not sure whether all the constant were generated manually or with a script, but there's a systematic error due to imporeprly combing Hex and Decimal numbers:

from /usr/include/tss2/tss2_tpm2_types.h:

#define TPM2_PT_VENDOR_STRING_1          ((TPM2_PT) (TPM2_PT_FIXED + 6)) 
#define TPM2_PT_VENDOR_STRING_2          ((TPM2_PT) (TPM2_PT_FIXED + 7))
#define TPM2_PT_VENDOR_STRING_3          ((TPM2_PT) (TPM2_PT_FIXED + 8)) 
#define TPM2_PT_VENDOR_STRING_4          ((TPM2_PT) (TPM2_PT_FIXED + 9)) 
#define TPM2_PT_VENDOR_TPM_TYPE          ((TPM2_PT) (TPM2_PT_FIXED + 10))  // <--- in hex this would be 0x10a, not 0x110
#define TPM2_PT_FIRMWARE_VERSION_1       ((TPM2_PT) (TPM2_PT_FIXED + 11))  // 0x10b, not 0x110
#define TPM2_PT_FIRMWARE_VERSION_2       ((TPM2_PT) (TPM2_PT_FIXED + 12)) // etc'

Vs.

tpmstream/src/tpmstream/spec/structures/constants.py

Lines 344 to 350 in f5084ec

	VENDOR_STRING_1 = 0x00000106
	VENDOR_STRING_2 = 0x00000107
	VENDOR_STRING_3 = 0x00000108
	VENDOR_STRING_4 = 0x00000109
	VENDOR_TPM_TYPE = 0x00000110
	FIRMWARE_VERSION_1 = 0x00000111
	FIRMWARE_VERSION_2 = 0x00000112

I discovered this through unexplained decoding errors in example streams. Depending on how this happend, it's possible the same issue also exists in other tpmstream data structures.

[joholl]

Thank you! You are absolutely right, that is a bug. I checked all of constants.py and could not find another issue like this.

The spec was written mostly manually. Trust me, I tried generating but the pain was even bigger.

[SharkOverBite]

Thanks. You might also monitor web tpmstream logs (if it sees meaningful amount of traffic) for suspicious looking decoding errors. It might flush out other bugs.

Thanks a lot for this project. I don't suppose It'll ever reach 10k stars or anything but it's been incredibly useful to me.

[joholl]

@SharkOverBite I'm glad it helps and thanks for the idea!