This is my reasoning:
If because of what Novak said to Matt about unreleased 4.0.0 resulted in
(A) Matt knew being below 4.0.0 was not vulnerable
+
(B) Matt knew his guide was not vulnerable ie dice...
review https://github.com/Coldcard/firmware/blob/2021-01-14T1617-v3.2.2/stm32/COLDCARD/vcp_lockdown.c
vs.
https://github.com/Coldcard/firmware/commit/dac940cc893ace370738c0c4ce3f763036089a8f#diff-b14a0bb021c4ebd7f34d21aae7f7374e894c99008d99a225f14da4cf2fdc7c7e
In pre-4.0.0 firmware, the ckcc_vcp_enabled is set to false.
The 4.0.1 patch on top of 4.0.0 is:
- //XXX/bool ckcc_vcp_enabled = false;
- bool ckcc_vcp_enabled = true;
+ extern bool ckcc_vcp_enabled;
So pre 4.0.0, USB issue does not exist because this value was set to false. In 4.0.1, this is changed from "true" to "false", closing the issue that only appeared in the update from pre-4.
Does that make sense? there is a bug that didn't exist pre-4.0.0 that was introduced in the refactor, that isn't the entropy bug. Matt's guides were for air gap usage, so you would no be impacted by this bug if you followed the guide.
vs.
https://github.com/Coldcard/firmware/commit/dac940cc893ace370738c0c4ce3f763036089a8f#diff-b14a0bb021c4ebd7f34d21aae7f7374e894c99008d99a225f14da4cf2fdc7c7e
In pre-4.0.0 firmware, the ckcc_vcp_enabled is set to false.
The 4.0.1 patch on top of 4.0.0 is:
- //XXX/bool ckcc_vcp_enabled = false;
- bool ckcc_vcp_enabled = true;
+ extern bool ckcc_vcp_enabled;
So pre 4.0.0, USB issue does not exist because this value was set to false. In 4.0.1, this is changed from "true" to "false", closing the issue that only appeared in the update from pre-4.
Does that make sense? there is a bug that didn't exist pre-4.0.0 that was introduced in the refactor, that isn't the entropy bug. Matt's guides were for air gap usage, so you would no be impacted by this bug if you followed the guide.