Skip to contentSkip to conversation

tpm: Disable TPM2_TCG_HMAC by default

PATCH In review Jarkko Sakkinen 3 participants 5 messages Aug 25, 2025 – Sep 18, 2025

Subsystems (4)
lore
Share & tools
lore

Jarkko Sakkinen

author maintainer
#0 lore raw

PATCH 1/1 tpm: Disable TPM2_TCG_HMAC by default

After reading all the feedback, right now disabling the TPM2_TCG_HMAC is the right call.

Other views discussed:

A. Having a kernel command-line parameter or refining the feature
otherwise. This goes to the area of improvements. E.g., one example is my own idea where the null key specific code would be replaced with a persistent handle parameter (which can be *unambigously* defined as part of attestation process when done correctly).

B. Removing the code. I don't buy this because that is same as saying
that HMAC encryption cannot work at all (if really nitpicking) in any form. Also I disagree on the view that the feature could not be refined to something more reasoable.

Also, both A and B are worst options in terms of backporting.

Thuss, this is the best possible choice.

Cc [email protected] # v6.10+
Fixes d2add27cf2b8 ("tpm: Add NULL primary creation")

---
PS. I did not post this last week because that would have been most likely the most counter-productive action to taken. It's better sometimes to take a bit of time to think (which can be seen that I've given also more reasonable weight to my own eaerlier proposals).

I also accept further changes, if there is e.g., inconsistency with TCG_TPM_HMAC setting or similar (obviously).

---
 drivers/char/tpm/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/tpm/Kconfig +1 -1
index dddd702b2454..3e4684f6b4af 100644
@@ -29,7 +29,7 @@ if TCG_TPM
 
 config TCG_TPM2_HMAC
 	bool "Use HMAC and encrypted transactions on the TPM bus"
-	default X86_64
+	default n
 	select CRYPTO_ECDH
 	select CRYPTO_LIB_AESCFB
 	select CRYPTO_LIB_SHA256
Signature · 1 line
2.39.5
Reviewed-by on v1 1/1
#1 lore raw
Quoted 12 lines · Jarkko Sakkinen

After reading all the feedback, right now disabling the TPM2_TCG_HMAC is the right call.

Other views discussed:

A. Having a kernel command-line parameter or refining the feature
otherwise. This goes to the area of improvements. E.g., one example is my own idea where the null key specific code would be replaced with a persistent handle parameter (which can be *unambigously* defined as part of attestation process when done correctly).

Quoted 8 lines

B. Removing the code. I don't buy this because that is same as saying
that HMAC encryption cannot work at all (if really nitpicking) in any form. Also I disagree on the view that the feature could not be refined to something more reasoable.

Also, both A and B are worst options in terms of backporting.

Thus, this is the best possible choice.

I think this is reasonable; it's adding runtime overhead and not adding enough benefit to be the default upstream.

Quoted 27 lines
Cc [email protected] # v6.10+
Fixes d2add27cf2b8 ("tpm: Add NULL primary creation")

---
PS. I did not post this last week because that would have been most likely the most counter-productive action to taken. It's better sometimes to take a bit of time to think (which can be seen that I've given also more reasonable weight to my own eaerlier proposals).

I also accept further changes, if there is e.g., inconsistency with TCG_TPM_HMAC setting or similar (obviously).
---
drivers/char/tpm/Kconfig | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/tpm/Kconfig +0 -0
index dddd702b2454..3e4684f6b4af 100644
@@ -29,7 +29,7 @@ if TCG_TPM
 

config TCG_TPM2_HMAC
bool "Use HMAC and encrypted transactions on the TPM bus"

drivers/char/tpm/Kconfig +1 -1
-	default X86_64
+	default n
Quoted 5 lines

select CRYPTO_ECDH
select CRYPTO_LIB_AESCFB
select CRYPTO_LIB_SHA256

Signature · 1 line
2.39.5

J.

Signature · 4 lines
] https://www.earth.li/~noodles/ []  Is this real - that's the first   [
]  PGP/GPG Key @ the.earth.li    []    thing I think every morning.    [
] via keyserver, web or email.   []                                    [
] RSA: 4096/0x94FA372B2DA8B985   []                                    [
author maintainer
#2 lore raw
Quoted 18 lines · Jonathan McDowell
Quoted 17 lines · Jarkko Sakkinen

After reading all the feedback, right now disabling the TPM2_TCG_HMAC is the right call.

Other views discussed:

A. Having a kernel command-line parameter or refining the feature
otherwise. This goes to the area of improvements. E.g., one example is my own idea where the null key specific code would be replaced with a persistent handle parameter (which can be *unambigously* defined as part of attestation process when done correctly).

B. Removing the code. I don't buy this because that is same as saying
that HMAC encryption cannot work at all (if really nitpicking) in any form. Also I disagree on the view that the feature could not be refined to something more reasoable.

Quoted 6 lines
Quoted 3 lines

Also, both A and B are worst options in terms of backporting.

Thus, this is the best possible choice.

I think this is reasonable; it's adding runtime overhead and not adding enough benefit to be the default upstream.

Yes, I think this is a balanced change. I agree what you say and at the same time this gives more space to refine it something usable. Right now it is much harder to tackle those issue, as it is part of the default config. By looking at things from this angle, the change is also benefical for the feature itself (in the long run).

Quoted 1 line

Thank you! I appreciate this and will append this to the commit.

BR, Jarkko

Reviewed-by
#3 lore raw

Agreed, the feature needs some work in order to provide meaningful security value, and disabling it by default facilitates that work.

Quoted 38 lines · Jarkko Sakkinen
Quoted 18 lines · Jonathan McDowell
Quoted 17 lines · Jarkko Sakkinen

After reading all the feedback, right now disabling the TPM2_TCG_HMAC is the right call.

Other views discussed:

A. Having a kernel command-line parameter or refining the feature
otherwise. This goes to the area of improvements. E.g., one example is my own idea where the null key specific code would be replaced with a persistent handle parameter (which can be *unambigously* defined as part of attestation process when done correctly).

B. Removing the code. I don't buy this because that is same as saying
that HMAC encryption cannot work at all (if really nitpicking) in any form. Also I disagree on the view that the feature could not be refined to something more reasoable.

Quoted 6 lines
Quoted 3 lines

Also, both A and B are worst options in terms of backporting.

Thus, this is the best possible choice.

I think this is reasonable; it's adding runtime overhead and not adding enough benefit to be the default upstream.

Yes, I think this is a balanced change. I agree what you say and at the same time this gives more space to refine it something usable. Right now it is much harder to tackle those issue, as it is part of the default config. By looking at things from this angle, the change is also benefical for the feature itself (in the long run).

Quoted 1 line

Thank you! I appreciate this and will append this to the commit.

BR, Jarkko

author maintainer
#4 lore raw
Quoted 4 lines · Chris Fenner

Agreed, the feature needs some work in order to provide meaningful security value, and disabling it by default facilitates that work.

Thanks!

BR, Jarkko