Skip to contentSkip to conversation

build: work around secondary expansion limitation with some Make versions

PATCH In review Nicholas Piggin 2 participants 5 messages Sep 8, 2025

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/2 build: work around secondary expansion limitation with some Make versions

GNU Make 4.2.1 as shipped in Ubuntu 20.04 has a problem with secondary expansion and variable names containing the '/' character. Make 4.3 and 4.4 don't have the problem.

Instead of using the variable name from riscv/sbi-deps and matching it from the riscv/sbi.* target name, name the variable sbi-deps and match it by stripping the riscv/ directory name off the riscv/sbi.* target name.

---
 riscv/Makefile | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
riscv/Makefile +7 -7
index beaeaefa..64720c38 100644
@@ -18,12 +18,12 @@ tests += $(TEST_DIR)/isa-dbltrp.$(exe)
 
 all: $(tests)
 
-$(TEST_DIR)/sbi-deps += $(TEST_DIR)/sbi-asm.o
-$(TEST_DIR)/sbi-deps += $(TEST_DIR)/sbi-dbtr.o
-$(TEST_DIR)/sbi-deps += $(TEST_DIR)/sbi-fwft.o
-$(TEST_DIR)/sbi-deps += $(TEST_DIR)/sbi-sse.o
+sbi-deps += $(TEST_DIR)/sbi-asm.o
+sbi-deps += $(TEST_DIR)/sbi-dbtr.o
+sbi-deps += $(TEST_DIR)/sbi-fwft.o
+sbi-deps += $(TEST_DIR)/sbi-sse.o
 
-all_deps += $($(TEST_DIR)/sbi-deps)
+all_deps += $(sbi-deps)
 
 # When built for EFI sieve needs extra memory, run with e.g. '-m 256' on QEMU
 $(TEST_DIR)/sieve.$(exe): AUXFLAGS = 0x1
@@ -113,7 +113,7 @@ cflatobjs += lib/efi.o
 .PRECIOUS: %.so
 
 %.so: EFI_LDFLAGS += -defsym=EFI_SUBSYSTEM=0xa --no-undefined
-%.so: %.o $(FLATLIBS) $(SRCDIR)/riscv/efi/elf_riscv64_efi.lds $(cstart.o) %.aux.o $$($$*-deps)
+%.so: %.o $(FLATLIBS) $(SRCDIR)/riscv/efi/elf_riscv64_efi.lds $(cstart.o) %.aux.o $$($$(notdir $$*)-deps)
 	$(LD) $(EFI_LDFLAGS) -o $@ -T $(SRCDIR)/riscv/efi/elf_riscv64_efi.lds \
 		$(filter %.o, $^) $(FLATLIBS) $(EFI_LIBS)
 
@@ -129,7 +129,7 @@ cflatobjs += lib/efi.o
 		-O binary $^ $@
 else
 %.elf: LDFLAGS += -pie -n -z notext
-%.elf: %.o $(FLATLIBS) $(SRCDIR)/riscv/flat.lds $(cstart.o) %.aux.o $$($$*-deps)
+%.elf: %.o $(FLATLIBS) $(SRCDIR)/riscv/flat.lds $(cstart.o) %.aux.o $$($$(notdir $$*)-deps)
 	$(LD) $(LDFLAGS) -o $@ -T $(SRCDIR)/riscv/flat.lds \
 		$(filter %.o, $^) $(FLATLIBS)
 	@chmod a-x $@
Signature · 1 line
2.51.0
#1 lore raw

PATCH 2/2 shellcheck: suppress SC2327,2328 false positives

Shellcheck warnings SC2327,SC2328 complain that a command substitution will be empty if the output is redirected, which is a valid warning but shellcheck is not smart enough to see when output is redirected into a command that outputs what the command substitution wanted.

Add comments and shellcheck directives to these cases.

---
 scripts/arch-run.bash | 9 +++++++++
 scripts/runtime.bash  | 6 ++++++
 2 files changed, 15 insertions(+)
scripts/arch-run.bash +9 -0
index 58e4f93f..9c089f88 100644
@@ -9,6 +9,11 @@ run_test ()
 
 	# stdout to {stdout}, stderr to $errors and stderr
 	exec {stdout}>&1
+	# SC complains that redirection without tee takes output way from
+	# command substitution, but that is what we want here (stderr output
+	# does go to command substitution because tee is used, but stdout does
+	# not).
+	# shellcheck disable=SC2327,SC2328
 	errors=$("${@}" $INITRD </dev/null 2> >(tee /dev/stderr) > /dev/fd/$stdout)
 	ret=$?
 	exec {stdout}>&-
@@ -23,6 +28,10 @@ run_test_status ()
 	local stdout ret
 
 	exec {stdout}>&1
+	# SC complains that redirection without tee takes output way from
+	# command substitution, but that is what we want here (tee is used
+	# inside the parenthesis).
+	# shellcheck disable=SC2327,SC2328
 	lines=$(run_test "$@" > >(tee /dev/fd/$stdout))
 	ret=$?
 	exec {stdout}>&-
scripts/runtime.bash +6 -0
index 289e52bb..12ac0f38 100644
@@ -190,6 +190,12 @@ function run()
     # qemu_params/extra_params in the config file may contain backticks that
     # need to be expanded, so use eval to start qemu.  Use "> >(foo)" instead of
     # a pipe to preserve the exit status.
+    #
+    # SC complains that redirection without tee takes output way from command
+    # substitution, but that is what we want here (tee is used inside the
+    # parenthesis and output piped to extract_summary which is captured by
+    # command substitution).
+    # shellcheck disable=SC2327,SC2328
     summary=$(eval "$cmdline" 2> >(RUNTIME_log_stderr $testname) \
                              > >(tee >(RUNTIME_log_stdout $testname $kernel) | extract_summary))
     ret=$?
Signature · 1 line
2.51.0
#2 lore raw

Re: [PATCH 2/2] shellcheck: suppress SC2327,2328 false positives

Quoted 18 lines · Nicholas Piggin

Shellcheck warnings SC2327,SC2328 complain that a command substitution will be empty if the output is redirected, which is a valid warning but shellcheck is not smart enough to see when output is redirected into a command that outputs what the command substitution wanted.

Add comments and shellcheck directives to these cases.

---
 scripts/arch-run.bash | 9 +++++++++
 scripts/runtime.bash  | 6 ++++++
 2 files changed, 15 insertions(+)
scripts/arch-run.bash +0 -0
index 58e4f93f..9c089f88 100644

scripts/arch-run.bash run_test () · line 12

Quoted 4 lines
scripts/arch-run.bash +1 -0
@@ -9,6 +9,11 @@ run_test ()
 
 	# stdout to {stdout}, stderr to $errors and stderr
 	exec {stdout}>&1
+	# SC complains that redirection without tee takes output way from

s/way/away/

Same comment for the other uses below.

Quoted 37 lines
scripts/arch-run.bash +8 -0
+	# command substitution, but that is what we want here (stderr output
+	# does go to command substitution because tee is used, but stdout does
+	# not).
+	# shellcheck disable=SC2327,SC2328
 	errors=$("${@}" $INITRD </dev/null 2> >(tee /dev/stderr) > /dev/fd/$stdout)
 	ret=$?
 	exec {stdout}>&-
@@ -23,6 +28,10 @@ run_test_status ()
 	local stdout ret
 
 	exec {stdout}>&1
+	# SC complains that redirection without tee takes output way from
+	# command substitution, but that is what we want here (tee is used
+	# inside the parenthesis).
+	# shellcheck disable=SC2327,SC2328
 	lines=$(run_test "$@" > >(tee /dev/fd/$stdout))
 	ret=$?
 	exec {stdout}>&-
scripts/runtime.bash +6 -0
index 289e52bb..12ac0f38 100644
@@ -190,6 +190,12 @@ function run()
     # qemu_params/extra_params in the config file may contain backticks that
     # need to be expanded, so use eval to start qemu.  Use "> >(foo)" instead of
     # a pipe to preserve the exit status.
+    #
+    # SC complains that redirection without tee takes output way from command
+    # substitution, but that is what we want here (tee is used inside the
+    # parenthesis and output piped to extract_summary which is captured by
+    # command substitution).
+    # shellcheck disable=SC2327,SC2328
     summary=$(eval "$cmdline" 2> >(RUNTIME_log_stderr $testname) \
                              > >(tee >(RUNTIME_log_stdout $testname $kernel) | extract_summary))
     ret=$?
Signature · 1 line
2.51.0

Changed way to away while merging.

Thanks,
drew

#3 lore raw
In reply to Nicholas Piggin

Missing the kvm-unit-tests prefix and, for this patch, the riscv prefix.

Quoted 9 lines · Nicholas Piggin

GNU Make 4.2.1 as shipped in Ubuntu 20.04 has a problem with secondary expansion and variable names containing the '/' character. Make 4.3 and 4.4 don't have the problem.

Instead of using the variable name from riscv/sbi-deps and matching it from the riscv/sbi.* target name, name the variable sbi-deps and match it by stripping the riscv/ directory name off the riscv/sbi.* target name.

Quoted 6 lines
---
 riscv/Makefile | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

Added the riscv prefix while merging.

Thanks,
drew

#4 lore raw
Quoted 1 line · Andrew Jones

Missing the kvm-unit-tests prefix and, for this patch, the riscv prefix.

Hey, sorry about that, been a while between kvm-unit-tests patches, thanks for fixing it all up.

Thanks,
Nick