[PATCH v3 0/7] Fix the reported SGX selftest makefile issues

[PATCH v3 0/7] Fix the reported SGX selftest makefile issues

[Jarkko Sakkinen]

Cedric Xing reported various issues in the selftest makefiles, which are
hopefully taken care by these patches.

As the last patch this patch set contains a patch that implements the
subdirectory traversal the for the x86 selftest directory tree. This
patch is meant to be a separate patch in the main SGX patch set so that
the bindings can be reviewed separately.

v3: 1,$s/SUB_DIRS/SUBDIRS/g
v2: Added patch to implement the traversal to the subdirectories
    selftest.

Jarkko Sakkinen (7):
  selftests/x86/sgx: Fix objcopy call in the Makefile
  selftests/x86/sgx: Do not generate object files
  selftests/x86/sgx: Add -I$(OUTPUT) when compiling encl_piggy.S
  selftests/x86/sgx: Fix EXTRA_CLEAN
  selftests/x86/sgx: Fix target in the Makefile
  selftests/x86: Revert SGX changes
  selftests/x86: Recurse into subdirectories

 tools/testing/selftests/x86/Makefile     | 54 +++++++++++++++++++-----
 tools/testing/selftests/x86/sgx/Makefile | 33 ++++++++-------
 tools/testing/selftests/x86/sgx/encl.lds |  1 +
 3 files changed, 63 insertions(+), 25 deletions(-)

-- 
2.20.1

[PATCH v3 1/7] selftests/x86/sgx: Fix objcopy call in the Makefile

[Jarkko Sakkinen]

Use $(OBJCOPY) to point to the correct objcopy binary. Do not discard
.got.plt because we need to detect libcalls and fail the compilation if
they exist.

Cc: Sean Christopherson <sean.j.christopherson@intel.com>
Reported-by: Cedric Xing <cedric.xing@intel.com>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
---
 tools/testing/selftests/x86/sgx/Makefile | 6 +++++-
 tools/testing/selftests/x86/sgx/encl.lds | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/x86/sgx/Makefile b/tools/testing/selftests/x86/sgx/Makefile
index 10136b73096b..fdd67478c52b 100644
--- a/tools/testing/selftests/x86/sgx/Makefile
+++ b/tools/testing/selftests/x86/sgx/Makefile
@@ -2,6 +2,10 @@ top_srcdir = ../../../../..
 
 include ../../lib.mk
 
+ifndef OBJCOPY
+OBJCOPY := $(CROSS_COMPILE)objcopy
+endif
+
 HOST_CFLAGS := -Wall -Werror -g $(INCLUDES) -fPIC -z noexecstack
 ENCL_CFLAGS := -Wall -Werror -static -nostdlib -nostartfiles -fPIC \
 	       -fno-stack-protector -mrdrnd $(INCLUDES)
@@ -23,7 +27,7 @@ $(OUTPUT)/encl_piggy.o: $(OUTPUT)/encl.bin $(OUTPUT)/encl.ss
 	$(CC) $(HOST_CFLAGS) -c encl_piggy.S -o $@
 
 $(OUTPUT)/encl.bin: $(OUTPUT)/encl.elf $(OUTPUT)/sgxsign
-	objcopy --remove-section=.got.plt -O binary $< $@
+	$(OBJCOPY) -O binary $< $@
 
 $(OUTPUT)/encl.elf: $(OUTPUT)/encl.o $(OUTPUT)/encl_bootstrap.o
 	$(CC) $(ENCL_CFLAGS) -T encl.lds -o $@ $^
diff --git a/tools/testing/selftests/x86/sgx/encl.lds b/tools/testing/selftests/x86/sgx/encl.lds
index 2ee01ac3ec79..9a56d3064104 100644
--- a/tools/testing/selftests/x86/sgx/encl.lds
+++ b/tools/testing/selftests/x86/sgx/encl.lds
@@ -31,3 +31,4 @@ ASSERT(!DEFINED(.altinstructions), "ALTERNATIVES are not supported in enclaves")
 ASSERT(!DEFINED(.altinstr_replacement), "ALTERNATIVES are not supported in enclaves")
 ASSERT(!DEFINED(.discard.retpoline_safe), "RETPOLINE ALTERNATIVES are not supported in enclaves")
 ASSERT(!DEFINED(.discard.nospec), "RETPOLINE ALTERNATIVES are not supported in enclaves")
+ASSERT(!DEFINED(.got.plt), "Libcalls are not supported in enclaves")
-- 
2.20.1

[PATCH v3 2/7] selftests/x86/sgx: Do not generate object files

[Jarkko Sakkinen]

Invoke GCC in a way that it generates and links the binary in a single
step because we do not need the side products. Additionally, add a
missing dependency to encls.lds (i.e. make did not do anything if
encls.lds was changed).

Reported-by: Cedric Xing <cedric.xing@intel.com>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
---
 tools/testing/selftests/x86/sgx/Makefile | 10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/x86/sgx/Makefile b/tools/testing/selftests/x86/sgx/Makefile
index fdd67478c52b..b96aff1a7fba 100644
--- a/tools/testing/selftests/x86/sgx/Makefile
+++ b/tools/testing/selftests/x86/sgx/Makefile
@@ -29,14 +29,8 @@ $(OUTPUT)/encl_piggy.o: $(OUTPUT)/encl.bin $(OUTPUT)/encl.ss
 $(OUTPUT)/encl.bin: $(OUTPUT)/encl.elf $(OUTPUT)/sgxsign
 	$(OBJCOPY) -O binary $< $@
 
-$(OUTPUT)/encl.elf: $(OUTPUT)/encl.o $(OUTPUT)/encl_bootstrap.o
-	$(CC) $(ENCL_CFLAGS) -T encl.lds -o $@ $^
-
-$(OUTPUT)/encl.o: encl.c
-	$(CC) $(ENCL_CFLAGS) -c $< -o $@
-
-$(OUTPUT)/encl_bootstrap.o: encl_bootstrap.S
-	$(CC) $(ENCL_CFLAGS) -c $< -o $@
+$(OUTPUT)/encl.elf: encl.lds encl.c encl_bootstrap.S
+	$(CC) $(ENCL_CFLAGS) -T $^ -o $@
 
 $(OUTPUT)/encl.ss: $(OUTPUT)/encl.bin  $(OUTPUT)/sgxsign
 	$(OUTPUT)/sgxsign signing_key.pem $(OUTPUT)/encl.bin $(OUTPUT)/encl.ss
-- 
2.20.1

[PATCH v3 3/7] selftests/x86/sgx: Add -I$(OUTPUT) when compiling encl_piggy.S

[Jarkko Sakkinen]

We need to the build output directory to the include path when compiling
encl_piggy.S because otherwise the compiler can't reliably find encl.bin
and encl.ss.

Reported-by: Cedric Xing <cedric.xing@intel.com>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
---
 tools/testing/selftests/x86/sgx/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/x86/sgx/Makefile b/tools/testing/selftests/x86/sgx/Makefile
index b96aff1a7fba..0b8813cc0b7f 100644
--- a/tools/testing/selftests/x86/sgx/Makefile
+++ b/tools/testing/selftests/x86/sgx/Makefile
@@ -24,7 +24,7 @@ $(OUTPUT)/sgx_call.o: sgx_call.S
 	$(CC) $(HOST_CFLAGS) -c $< -o $@
 
 $(OUTPUT)/encl_piggy.o: $(OUTPUT)/encl.bin $(OUTPUT)/encl.ss
-	$(CC) $(HOST_CFLAGS) -c encl_piggy.S -o $@
+	$(CC) $(HOST_CFLAGS) -I$(OUTPUT) -c encl_piggy.S -o $@
 
 $(OUTPUT)/encl.bin: $(OUTPUT)/encl.elf $(OUTPUT)/sgxsign
 	$(OBJCOPY) -O binary $< $@
-- 
2.20.1

[PATCH v3 4/7] selftests/x86/sgx: Fix EXTRA_CLEAN

[Jarkko Sakkinen]

1. Add missing build products to EXTRA_CLEAN.
2. Remove non-existent build products from EXTRA_CLEAN.

Reported-by: Cedric Xing <cedric.xing@intel.com>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
---
 tools/testing/selftests/x86/sgx/Makefile | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/tools/testing/selftests/x86/sgx/Makefile b/tools/testing/selftests/x86/sgx/Makefile
index 0b8813cc0b7f..0f1c50c347d7 100644
--- a/tools/testing/selftests/x86/sgx/Makefile
+++ b/tools/testing/selftests/x86/sgx/Makefile
@@ -38,9 +38,14 @@ $(OUTPUT)/encl.ss: $(OUTPUT)/encl.bin  $(OUTPUT)/sgxsign
 $(OUTPUT)/sgxsign: sgxsign.c
 	$(CC) -o $@ $< -lcrypto
 
-EXTRA_CLEAN := $(OUTPUT)/sgx-selftest $(OUTPUT)/sgx-selftest.o \
-	       $(OUTPUT)/sgx_call.o $(OUTPUT)/encl.bin $(OUTPUT)/encl.ss \
-	       $(OUTPUT)/encl.elf $(OUTPUT)/encl.o $(OUTPUT)/encl_bootstrap.o \
-	       $(OUTPUT)/sgxsign
+EXTRA_CLEAN := \
+	$(OUTPUT)/encl.bin \
+	$(OUTPUT)/encl_piggy.o \
+	$(OUTPUT)/encl.elf \
+	$(OUTPUT)/encl.ss \
+	$(OUTPUT)/sgx_call.o \
+	$(OUTPUT)/sgxsign \
+	$(OUTPUT)/test_sgx \
+	$(OUTPUT)/test_sgx.o \
 
 .PHONY: clean
-- 
2.20.1

[PATCH v3 5/7] selftests/x86/sgx: Fix target in the Makefile

[Jarkko Sakkinen]

The target should be "all", not "all_64".

Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
---
 tools/testing/selftests/x86/sgx/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/x86/sgx/Makefile b/tools/testing/selftests/x86/sgx/Makefile
index 0f1c50c347d7..4310a5b6ecc7 100644
--- a/tools/testing/selftests/x86/sgx/Makefile
+++ b/tools/testing/selftests/x86/sgx/Makefile
@@ -11,7 +11,7 @@ ENCL_CFLAGS := -Wall -Werror -static -nostdlib -nostartfiles -fPIC \
 	       -fno-stack-protector -mrdrnd $(INCLUDES)
 
 TEST_CUSTOM_PROGS := $(OUTPUT)/test_sgx
-all_64: $(TEST_CUSTOM_PROGS)
+all: $(TEST_CUSTOM_PROGS)
 
 $(TEST_CUSTOM_PROGS): $(OUTPUT)/main.o $(OUTPUT)/sgx_call.o \
 		      $(OUTPUT)/encl_piggy.o
-- 
2.20.1

[PATCH v3 6/7] selftests/x86: Revert SGX changes

[Jarkko Sakkinen]

The SGX changes are incorrect in the makefile but the more important
reason to revert the SGX changes is that we can split the binding
to a separate commit in our main patch set.

There is a design choice to be made whether to bound the selftest by
recursing from the x86 makefile or alternatively have it added to
TARGETS (i.e. TARGET += "x86/sgx"). Thus, the change that implements
the binding must a be a separate commit.

Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
---
 tools/testing/selftests/x86/Makefile | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
index a1831406fd01..fa07d526fe39 100644
--- a/tools/testing/selftests/x86/Makefile
+++ b/tools/testing/selftests/x86/Makefile
@@ -1,7 +1,4 @@
 # SPDX-License-Identifier: GPL-2.0
-
-SUBDIRS_64 := sgx
-
 all:
 
 include ../lib.mk
@@ -71,13 +68,6 @@ all_32: $(BINARIES_32)
 
 all_64: $(BINARIES_64)
 
-all_64: $(SUBDIRS_64)
-	@for DIR in $(SUBDIRS_64); do			\
-		BUILD_TARGET=$(OUTPUT)/$$DIR;		\
-		mkdir $$BUILD_TARGET  -p;		\
-		make OUTPUT=$$BUILD_TARGET -C $$DIR $@;	\
-	done
-
 EXTRA_CLEAN := $(BINARIES_32) $(BINARIES_64)
 
 $(BINARIES_32): $(OUTPUT)/%_32: %.c
-- 
2.20.1

[PATCH v3 7/7] selftests/x86: Recurse into subdirectories

[Jarkko Sakkinen]

Recurse into a list of subdirectories defined by SUBDIRS when running
x86 selftests. Override run_tests, install, emit_tests and clean
targets to implement this behaviour.

A possible alternative would be to add "x86/sgx" to TARGETS. However,
this would be problematic because detecting 64-bit build would have
to duplicated.

The implementation is derived from the makefiles of powerpc and sparc64
selftests.

Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
---
 tools/testing/selftests/x86/Makefile | 44 ++++++++++++++++++++++++++++
 1 file changed, 44 insertions(+)

diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
index fa07d526fe39..80571bac8ed5 100644
--- a/tools/testing/selftests/x86/Makefile
+++ b/tools/testing/selftests/x86/Makefile
@@ -10,6 +10,8 @@ CAN_BUILD_I386 := $(shell ./check_cc.sh $(CC) trivial_32bit_program.c -m32)
 CAN_BUILD_X86_64 := $(shell ./check_cc.sh $(CC) trivial_64bit_program.c)
 CAN_BUILD_WITH_NOPIE := $(shell ./check_cc.sh $(CC) trivial_program.c -no-pie)
 
+SUBDIRS := sgx
+
 TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs syscall_nt test_mremap_vdso \
 			check_initial_reg_state sigreturn iopl mpx-mini-test ioperm \
 			protection_keys test_vdso test_vsyscall mov_ss_trap \
@@ -59,6 +61,48 @@ endif
 
 ifeq ($(CAN_BUILD_X86_64),1)
 all: all_64
+	@for DIR in $(SUBDIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$DIR; \
+		mkdir $$BUILD_TARGET  -p; \
+		make OUTPUT=$$BUILD_TARGET -C $$DIR $@; \
+	done
+
+DEFAULT_RUN_TESTS := $(RUN_TESTS)
+override define RUN_TESTS
+	$(DEFAULT_RUN_TESTS)
+	@for TARGET in $(SUBDIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
+		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET run_tests; \
+	done;
+endef
+
+DEFAULT_INSTALL_RULE := $(INSTALL_RULE)
+override define INSTALL_RULE
+	$(DEFAULT_INSTALL_RULE)
+	@for TARGET in $(SUBDIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
+		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET install; \
+	done;
+endef
+
+DEFAULT_EMIT_TESTS := $(EMIT_TESTS)
+override define EMIT_TESTS
+	$(DEFAULT_EMIT_TESTS)
+	@for TARGET in $(SUBDIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
+		$(MAKE) OUTPUT=$$BUILD_TARGET -s -C $$TARGET emit_tests; \
+	done;
+endef
+
+DEFAULT_CLEAN := $(CLEAN)
+override define CLEAN
+	$(DEFAULT_CLEAN)
+	@for TARGET in $(SUBDIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
+		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET clean; \
+	done;
+endef
+
 TEST_PROGS += $(BINARIES_64)
 EXTRA_CFLAGS += -DCAN_BUILD_64
 $(foreach t,$(TARGETS_C_64BIT_ALL),$(eval $(call gen-target-rule-64,$(t))))
-- 
2.20.1

Re: [PATCH v3 7/7] selftests/x86: Recurse into subdirectories

[Sean Christopherson]

On Mon, Aug 19, 2019 at 04:28:30PM +0300, Jarkko Sakkinen wrote:
> Recurse into a list of subdirectories defined by SUBDIRS when running
> x86 selftests. Override run_tests, install, emit_tests and clean
> targets to implement this behaviour.

The code looks good (which doesn't say much, my Makefile knowledge is
garbage), but the result is a bit odd.  x86/sgx doesn't implement
emit_tests (or at least test_sgx doesn't show up), while the rest of the
x86 tests don't support install.

AFAICT the Makefile is itself weird; it's definitely different than other
selftests Makfiles.  I don't see any reason to delay this patch, but it
feels like something here needs to be cleaned up.

> A possible alternative would be to add "x86/sgx" to TARGETS. However,
> this would be problematic because detecting 64-bit build would have
> to duplicated.
> 
> The implementation is derived from the makefiles of powerpc and sparc64
> selftests.
> 
> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
> ---
>  tools/testing/selftests/x86/Makefile | 44 ++++++++++++++++++++++++++++
>  1 file changed, 44 insertions(+)
> 
> diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
> index fa07d526fe39..80571bac8ed5 100644
> --- a/tools/testing/selftests/x86/Makefile
> +++ b/tools/testing/selftests/x86/Makefile
> @@ -10,6 +10,8 @@ CAN_BUILD_I386 := $(shell ./check_cc.sh $(CC) trivial_32bit_program.c -m32)
>  CAN_BUILD_X86_64 := $(shell ./check_cc.sh $(CC) trivial_64bit_program.c)
>  CAN_BUILD_WITH_NOPIE := $(shell ./check_cc.sh $(CC) trivial_program.c -no-pie)
>  
> +SUBDIRS := sgx
> +
>  TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs syscall_nt test_mremap_vdso \
>  			check_initial_reg_state sigreturn iopl mpx-mini-test ioperm \
>  			protection_keys test_vdso test_vsyscall mov_ss_trap \
> @@ -59,6 +61,48 @@ endif
>  
>  ifeq ($(CAN_BUILD_X86_64),1)
>  all: all_64
> +	@for DIR in $(SUBDIRS); do \
> +		BUILD_TARGET=$(OUTPUT)/$$DIR; \
> +		mkdir $$BUILD_TARGET  -p; \
> +		make OUTPUT=$$BUILD_TARGET -C $$DIR $@; \
> +	done
> +
> +DEFAULT_RUN_TESTS := $(RUN_TESTS)
> +override define RUN_TESTS
> +	$(DEFAULT_RUN_TESTS)
> +	@for TARGET in $(SUBDIRS); do \
> +		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
> +		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET run_tests; \
> +	done;
> +endef
> +
> +DEFAULT_INSTALL_RULE := $(INSTALL_RULE)
> +override define INSTALL_RULE
> +	$(DEFAULT_INSTALL_RULE)
> +	@for TARGET in $(SUBDIRS); do \
> +		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
> +		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET install; \
> +	done;
> +endef
> +
> +DEFAULT_EMIT_TESTS := $(EMIT_TESTS)
> +override define EMIT_TESTS
> +	$(DEFAULT_EMIT_TESTS)
> +	@for TARGET in $(SUBDIRS); do \
> +		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
> +		$(MAKE) OUTPUT=$$BUILD_TARGET -s -C $$TARGET emit_tests; \
> +	done;
> +endef
> +
> +DEFAULT_CLEAN := $(CLEAN)
> +override define CLEAN
> +	$(DEFAULT_CLEAN)
> +	@for TARGET in $(SUBDIRS); do \
> +		BUILD_TARGET=$(OUTPUT)/$$TARGET; \
> +		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET clean; \
> +	done;
> +endef
> +
>  TEST_PROGS += $(BINARIES_64)
>  EXTRA_CFLAGS += -DCAN_BUILD_64
>  $(foreach t,$(TARGETS_C_64BIT_ALL),$(eval $(call gen-target-rule-64,$(t))))
> -- 
> 2.20.1
> 

Re: [PATCH v3 7/7] selftests/x86: Recurse into subdirectories

[Jarkko Sakkinen]

On Wed, 2019-08-21 at 20:42 -0700, Sean Christopherson wrote:
> On Mon, Aug 19, 2019 at 04:28:30PM +0300, Jarkko Sakkinen wrote:
> > Recurse into a list of subdirectories defined by SUBDIRS when running
> > x86 selftests. Override run_tests, install, emit_tests and clean
> > targets to implement this behaviour.
> 
> The code looks good (which doesn't say much, my Makefile knowledge is
> garbage), but the result is a bit odd.  x86/sgx doesn't implement
> emit_tests (or at least test_sgx doesn't show up), while the rest of the
> x86 tests don't support install.
> 
> AFAICT the Makefile is itself weird; it's definitely different than other
> selftests Makfiles.  I don't see any reason to delay this patch, but it
> feels like something here needs to be cleaned up.

Lets fix emit_tests before sending v22.

/Jarkko

Re: [PATCH v3 7/7] selftests/x86: Recurse into subdirectories

[Jarkko Sakkinen]

On Thu, 2019-08-22 at 17:59 +0300, Jarkko Sakkinen wrote:
> On Wed, 2019-08-21 at 20:42 -0700, Sean Christopherson wrote:
> > On Mon, Aug 19, 2019 at 04:28:30PM +0300, Jarkko Sakkinen wrote:
> > > Recurse into a list of subdirectories defined by SUBDIRS when running
> > > x86 selftests. Override run_tests, install, emit_tests and clean
> > > targets to implement this behaviour.
> > 
> > The code looks good (which doesn't say much, my Makefile knowledge is
> > garbage), but the result is a bit odd.  x86/sgx doesn't implement
> > emit_tests (or at least test_sgx doesn't show up), while the rest of the
> > x86 tests don't support install.
> > 
> > AFAICT the Makefile is itself weird; it's definitely different than other
> > selftests Makfiles.  I don't see any reason to delay this patch, but it
> > feels like something here needs to be cleaned up.
> 
> Lets fix emit_tests before sending v22.

Selftest patches have been squahed. Further improvements are welcome.
Thanks for the review!

/Jarkko

Re: [PATCH v3 7/7] selftests/x86: Recurse into subdirectories

[Jarkko Sakkinen]

On Thu, 2019-08-22 at 18:02 +0300, Jarkko Sakkinen wrote:
> On Thu, 2019-08-22 at 17:59 +0300, Jarkko Sakkinen wrote:
> > On Wed, 2019-08-21 at 20:42 -0700, Sean Christopherson wrote:
> > > On Mon, Aug 19, 2019 at 04:28:30PM +0300, Jarkko Sakkinen wrote:
> > > > Recurse into a list of subdirectories defined by SUBDIRS when running
> > > > x86 selftests. Override run_tests, install, emit_tests and clean
> > > > targets to implement this behaviour.
> > > 
> > > The code looks good (which doesn't say much, my Makefile knowledge is
> > > garbage), but the result is a bit odd.  x86/sgx doesn't implement
> > > emit_tests (or at least test_sgx doesn't show up), while the rest of the
> > > x86 tests don't support install.
> > > 
> > > AFAICT the Makefile is itself weird; it's definitely different than other
> > > selftests Makfiles.  I don't see any reason to delay this patch, but it
> > > feels like something here needs to be cleaned up.
> > 
> > Lets fix emit_tests before sending v22.
> 
> Selftest patches have been squahed. Further improvements are welcome.
> Thanks for the review!

At least my use case is satisfied with these fixes:

$ ls output/target/usr/lib/kselftests/x86 
test_sgx

I can build initramfs now with the selftest bundled. Oddly enough none of the
other x86 selftests get included.

/Jarkko

Re: [PATCH v3 7/7] selftests/x86: Recurse into subdirectories

[Jarkko Sakkinen]

On Thu, 2019-08-22 at 19:13 +0300, Jarkko Sakkinen wrote:
> On Thu, 2019-08-22 at 18:02 +0300, Jarkko Sakkinen wrote:
> > On Thu, 2019-08-22 at 17:59 +0300, Jarkko Sakkinen wrote:
> > > On Wed, 2019-08-21 at 20:42 -0700, Sean Christopherson wrote:
> > > > On Mon, Aug 19, 2019 at 04:28:30PM +0300, Jarkko Sakkinen wrote:
> > > > > Recurse into a list of subdirectories defined by SUBDIRS when running
> > > > > x86 selftests. Override run_tests, install, emit_tests and clean
> > > > > targets to implement this behaviour.
> > > > 
> > > > The code looks good (which doesn't say much, my Makefile knowledge is
> > > > garbage), but the result is a bit odd.  x86/sgx doesn't implement
> > > > emit_tests (or at least test_sgx doesn't show up), while the rest of the
> > > > x86 tests don't support install.
> > > > 
> > > > AFAICT the Makefile is itself weird; it's definitely different than other
> > > > selftests Makfiles.  I don't see any reason to delay this patch, but it
> > > > feels like something here needs to be cleaned up.
> > > 
> > > Lets fix emit_tests before sending v22.
> > 
> > Selftest patches have been squahed. Further improvements are welcome.
> > Thanks for the review!
> 
> At least my use case is satisfied with these fixes:
> 
> $ ls output/target/usr/lib/kselftests/x86 
> test_sgx
> 
> I can build initramfs now with the selftest bundled. Oddly enough none of the
> other x86 selftests get included.

# ls /usr/lib/kselftests/x86
test_sgx
# /usr/lib/kselftests/x86/test_sgx 
Binary size 24576 (0x6000), SIGSTRUCT size 1808
Loading the enclave.
Input: 0x1122334455667788
Output: 0x1122334455667788

This is a big improvement for me :-) Do not have to upload selftest anymore
with minicom or scp.

/Jarkko

Re: [PATCH v3 0/7] Fix the reported SGX selftest makefile issues

[Sean Christopherson]

On Mon, Aug 19, 2019 at 04:28:23PM +0300, Jarkko Sakkinen wrote:
> Cedric Xing reported various issues in the selftest makefiles, which are
> hopefully taken care by these patches.
> 
> As the last patch this patch set contains a patch that implements the
> subdirectory traversal the for the x86 selftest directory tree. This
> patch is meant to be a separate patch in the main SGX patch set so that
> the bindings can be reviewed separately.

Doesn't have to be this series, but the build artifacts should be added
to .gitignore.

Untracked files:
  (use "git add <file>..." to include in what will be committed)

        sgx/encl.ss
        sgx/sgxsign
        sgx/test_sgx

> 
> v3: 1,$s/SUB_DIRS/SUBDIRS/g
> v2: Added patch to implement the traversal to the subdirectories
>     selftest.
> 
> Jarkko Sakkinen (7):
>   selftests/x86/sgx: Fix objcopy call in the Makefile
>   selftests/x86/sgx: Do not generate object files
>   selftests/x86/sgx: Add -I$(OUTPUT) when compiling encl_piggy.S
>   selftests/x86/sgx: Fix EXTRA_CLEAN
>   selftests/x86/sgx: Fix target in the Makefile
>   selftests/x86: Revert SGX changes
>   selftests/x86: Recurse into subdirectories
> 
>  tools/testing/selftests/x86/Makefile     | 54 +++++++++++++++++++-----
>  tools/testing/selftests/x86/sgx/Makefile | 33 ++++++++-------
>  tools/testing/selftests/x86/sgx/encl.lds |  1 +
>  3 files changed, 63 insertions(+), 25 deletions(-)
> 
> -- 
> 2.20.1
> 

Re: [PATCH v3 0/7] Fix the reported SGX selftest makefile issues

[Jarkko Sakkinen]

On Wed, 2019-08-21 at 20:31 -0700, Sean Christopherson wrote:
> On Mon, Aug 19, 2019 at 04:28:23PM +0300, Jarkko Sakkinen wrote:
> > Cedric Xing reported various issues in the selftest makefiles, which are
> > hopefully taken care by these patches.
> > 
> > As the last patch this patch set contains a patch that implements the
> > subdirectory traversal the for the x86 selftest directory tree. This
> > patch is meant to be a separate patch in the main SGX patch set so that
> > the bindings can be reviewed separately.
> 
> Doesn't have to be this series, but the build artifacts should be added
> to .gitignore.
> 
> Untracked files:
>   (use "git add <file>..." to include in what will be committed)
> 
>         sgx/encl.ss
>         sgx/sgxsign
>         sgx/test_sgx

Oh, good catch. Lets do it as a separate patch after I've squashed
everything.

/Jarkko