diff --git a/firmware/config/boards/common_make.sh b/firmware/config/boards/common_make.sh
old mode 100644
new mode 100755
index 86f592e3b1..e5d10aa70d
--- a/firmware/config/boards/common_make.sh
+++ b/firmware/config/boards/common_make.sh
@@ -8,16 +8,25 @@ set -e
 
 SCRIPT_NAME="common_make.sh"
 echo "Entering $SCRIPT_NAME with board $1 and CPU $2"
+BOARD_DIR=$(pwd)
+echo "Board dir is $BOARD_DIR"
 
-cd ../../..
+# Back out to the firmware root, relative to this script's location, as it may be called
+# from outside the firmware tree.
+FW_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )/../..
+FW_DIR=$(readlink -f $FW_DIR)
+echo "FW dir is $FW_DIR"
+cd $FW_DIR
 
 mkdir -p .dep
-make -j6 -r PROJECT_BOARD=$PROJECT_BOARD PROJECT_CPU=$PROJECT_CPU
+echo "Calling make for the main firmware..."
+make -j6 -r PROJECT_BOARD=$PROJECT_BOARD PROJECT_CPU=$PROJECT_CPU BOARD_DIR=$BOARD_DIR
 [ -e build/fome.hex ] || { echo "FAILED to compile by $SCRIPT_NAME with $PROJECT_BOARD $DEBUG_LEVEL_OPT and $EXTRA_PARAMS"; exit 1; }
 if [ "$USE_OPENBLT" = "yes" ]; then
   # TODO: why is this rm necessary?
   rm -f pch/pch.h.gch/*
-  cd bootloader; make PROJECT_BOARD=$PROJECT_BOARD PROJECT_CPU=$PROJECT_CPU -j12; cd ..
+  echo "Calling make for the bootloader..."
+  cd bootloader; make -j6 PROJECT_BOARD=$PROJECT_BOARD PROJECT_CPU=$PROJECT_CPU BOARD_DIR=$BOARD_DIR; cd ..
   [ -e bootloader/blbuild/fome_bl.hex ] || { echo "FAILED to compile OpenBLT by $SCRIPT_NAME with $PROJECT_BOARD"; exit 1; }
 fi