summaryrefslogtreecommitdiff
path: root/Makefile
diff options
context:
space:
mode:
Diffstat (limited to 'Makefile')
-rw-r--r--Makefile11
1 files changed, 9 insertions, 2 deletions
diff --git a/Makefile b/Makefile
index b5b68b20..10c5c4bf 100644
--- a/Makefile
+++ b/Makefile
@@ -840,11 +840,18 @@ ${TLCL_OBJS}: CFLAGS += -DTPM_BLOCKING_CONTINUESELFTEST
# CFLAGS += -DTPM_MANUAL_SELFTEST
+# NOTE: UNROLL_LOOPS *only* affects SHA256, *not* SHA512. This seems to have
+# been a conscious decision at some point (see b/35501356) but whether it still
+# holds up in all situations on all architectures today might need to be
+# reevaluated. For now, since we currently always use SHA256 for (non-recovery)
+# kernel bodies and don't unroll loops for firmware verification, it's not very
+# relevant in practice. To unroll SHA512, UNROLL_LOOPS_SHA512 would need to be
+# defined.
ifneq ($(filter-out 0,$(UNROLL_LOOPS)),)
-$(info vboot hash algos built with unrolled loops (faster, larger code size))
+$(info vboot SHA256 built with unrolled loops (faster, larger code size))
CFLAGS += -DUNROLL_LOOPS
else
-$(info vboot hash algos built with tight loops (slower, smaller code size))
+$(info vboot SHA256 built with tight loops (slower, smaller code size))
endif
.PHONY: fwlib