Change FIT_KERNEL_COMP_ALG to lzma This change is applied to all aarch64 platforms. Reduce compressed kernel size for Google BMC. Tested: ``` ┌────────────────────┬─────────────┬─────────────┬─────────────────┐ │ Metric │ Prod (Gzip) │ Prod (LZMA) │ Reduction │ ├────────────────────┼─────────────┼─────────────┼─────────────────┤ │ Uncompressed Image │ 11.28 MiB │ 11.28 MiB │ - │ │ Compressed Kernel │ 4.86 MiB │ 3.63 MiB │ ~1.23 MiB (25%) │ │ Full FIT Image │ 6.07 MiB │ 4.84 MiB │ ~1.23 MiB │ └────────────────────┴─────────────┴─────────────┴─────────────────┘ ``` Tested: - platform11: https://fusion2.corp.google.com/ab1d2f1f-08da-3352-b24b-328bc8d2e2a0 - platform11: https://fusion2.corp.google.com/19c963e7-860e-3a12-844d-cf7e84fe7492 - platform5: https://fusion2.corp.google.com/78a00bbd-ba49-3a62-bf28-63f9a550f01a - platform15: https://fusion2.corp.google.com/4fc481c8-15fa-3382-93fd-7b3f3401f745 - platform17: https://fusion2.corp.google.com/d449fef1-c631-353b-a241-0d21b2a7cd5f - platform18: https://fusion2.corp.google.com/2939bcb2-59cf-3c16-999f-412d77dc24fb Google-Bug-Id: 512539812 Change-Id: I8ddc80c57adc5c831cb9cfa675733e1dbb0d5780 Signed-off-by: Willy Tu <wltu@google.com>
diff --git a/conf/include/gbmc-staging.inc b/conf/include/gbmc-staging.inc new file mode 100644 index 0000000..1d64aec --- /dev/null +++ b/conf/include/gbmc-staging.inc
@@ -0,0 +1,3 @@ +# Staging kernel compression algorithm for 64-bit gBMC +FIT_KERNEL_COMP_ALG:gbmc:aarch64 = "lzma" +FIT_KERNEL_COMP_ALG_EXTENSION:gbmc:aarch64 = ".lzma"
diff --git a/conf/layer.conf b/conf/layer.conf index f2a12bc..c37f231 100644 --- a/conf/layer.conf +++ b/conf/layer.conf
@@ -33,3 +33,5 @@ # Enable hardware ASAN if GBMC_SAN_MODE enables it. OVERRIDES:append = "${@':gbmc_hwasan' if (d.getVar('GBMC_SAN_MODE', False) or '' ).endswith('hwasan') else ''}" + +require conf/include/gbmc-staging.inc