- 05 Jul, 2021 4 commits
-
-
Sandrine Bailleux authored
- In tpm_record_measurement(): The platform layer is responsible for providing an exhaustive list of images to measure. If it doesn't then this should be treated as a programming error, as documented in [1]. Thus, turn the error test into an assertion. [1] https://trustedfirmware-a.readthedocs.io/en/latest/process/coding-guidelines.html#using-assert-to-check-for-programming-errors Change-Id: I002309c2ebdf2d348a7d12a8f7f9e82465046b8e Signed-off-by:
Sandrine Bailleux <sandrine.bailleux@arm.com>
-
Sandrine Bailleux authored
- In add_event2(): Turn the first error condition checking whether there is room for an extra event2 data structure into an assertion. The platform layer is responsible for choosing an appropriate event log buffer size based on the number of measurements it expects. If this assertion fires, the platform macro EVENT_LOG_SIZE should be adjusted and the firmware recompiled. Call this assumption out in the function documentation. Also remove the second error condition check, which is a subset of the first one and thus is redundant. As a result of these changes, add_event2() can no longer fail. Thus, change its return type from int to void. Also, the 'size_of_event' local variable is now unused in release builds so remove it and move its value into the assertion. Change-Id: I113fc141de59708b20435a0c7126255561ab7786 Signed-off-by:
Sandrine Bailleux <sandrine.bailleux@arm.com>
-
Sandrine Bailleux authored
- In event_log_init(): Throughout the function, we are incrementing a pointer by some fixed amounts of bytes (corresponding to the size of some data structure or to some constant number of bytes), there is no variable-size increments in the picture. Thus it seems pointless to verify that the pointer has indeed been incremented by this fixed amount of bytes afterwards. For this reason, remove these checks altogether. As a result, the start_ptr local variable is now unused so remove it as well. Change-Id: I612e2278cd3a63d1417427e45d81e285503f5efe Signed-off-by:
Sandrine Bailleux <sandrine.bailleux@arm.com>
-
Sandrine Bailleux authored
Change-Id: I5b84a28ed254a7c7bb95c18fa999592a4e3f6d90 Signed-off-by:
Sandrine Bailleux <sandrine.bailleux@arm.com>
-
- 01 Jul, 2021 1 commit
-
-
Sandrine Bailleux authored
Weak definitions are confusing and should be avoided if possible. Thus, turn plat_get_measured_boot_data() into a strong definition that platforms must provide (if they need measured boot). We could have moved the old weak implementation under plat/common as a sane, default implementation that platforms may pull in if it suits them. However, this implementation right now simply measures BL2, which is not enough to get a complete measured boot flow, so this patch just removes it. This change only affects the Arm FVP platform, as no other upstream platform implements measured boot at the moment. Change-Id: If8680a39ae0ef1044ee981315439d5e0c8461229 Signed-off-by:
Sandrine Bailleux <sandrine.bailleux@arm.com>
-
- 28 Sep, 2020 1 commit
-
-
Alexei Fedorov authored
This patch fixes MISRA C-2012 Pointers and Arrays Rule 18.4 defects reported by Coverity scan: "misra_c_2012_rule_18_4_violation: Using arithmetic on pointer " Change-Id: I06753b28467c473e346b9871c1657284fc43a3f3 Signed-off-by:
Alexei Fedorov <Alexei.Fedorov@arm.com>
-
- 21 Jul, 2020 1 commit
-
-
Alexei Fedorov authored
This patch adds support for Event Log generation required for Measured Boot functionality. Change-Id: I34f05a33565e6659e78499d62cc6fb00b7d6c2dc Signed-off-by:
Alexei Fedorov <Alexei.Fedorov@arm.com>
-