Skip to contentSkip to conversation

selftests/vDSO: fix clang build errors and warnings

PATCH Superseded John Hubbard 4 participants 8 messages May 27, 2024 – Jun 14, 2024

Subsystems (2)

v2 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 selftests/vDSO: fix clang build errors and warnings

When building with clang, via:

make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and allows these tests to run and pass.

1. Fix linker error (undefined reference to memcpy) by providing a local version of memcpy.

2. clang complains about using this form:

    if (g = h & 0xf0000000)

...so factor out the assignment into a separate step.

3. The code is passing a signed const char* to elf_hash(), which expects a const unsigned char *. There are several callers, so fix this at the source by allowing the function to accept a signed argument, and then converting to unsigned operations, once inside the function.

4. clang doesn't have __attribute__((externally_visible)) and generates a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem to require that attribute in order to build, run and pass tests here, so remove it.

[1] https://lore.kernel.org/all/20240329-selftests-libmk-llvm-rfc-v1-1-2f9ed7d1c49f@valentinobst.de/

---

Changes since the first version:

1) Rebased onto Linux 6.10-rc1

thanks,
John Hubbard

 tools/testing/selftests/vDSO/parse_vdso.c      | 16 +++++++++++-----
 .../selftests/vDSO/vdso_standalone_test_x86.c  | 18 ++++++++++++++++--
 2 files changed, 27 insertions(+), 7 deletions(-)
tools/testing/selftests/vDSO/parse_vdso.c +11 -5
index 413f75620a35..4ae417372e9e 100644
@@ -55,14 +55,20 @@ static struct vdso_info
 	ELF(Verdef) *verdef;
 } vdso_info;
 
-/* Straight from the ELF specification. */
-static unsigned long elf_hash(const unsigned char *name)
+/*
+ * Straight from the ELF specification...and then tweaked slightly, in order to
+ * avoid a few clang warnings.
+ */
+static unsigned long elf_hash(const char *name)
 {
 	unsigned long h = 0, g;
-	while (*name)
+	const unsigned char *uch_name = (const unsigned char *)name;
+
+	while (*uch_name)
 	{
-		h = (h << 4) + *name++;
-		if (g = h & 0xf0000000)
+		h = (h << 4) + *uch_name++;
+		g = h & 0xf0000000;
+		if (g)
 			h ^= g >> 24;
 		h &= ~g;
 	}
tools/testing/selftests/vDSO/vdso_standalone_test_x86.c +16 -2
index 8a44ff973ee1..27f6fdf11969 100644
@@ -18,7 +18,7 @@
 
 #include "parse_vdso.h"
 
-/* We need a libc functions... */
+/* We need some libc functions... */
 int strcmp(const char *a, const char *b)
 {
 	/* This implementation is buggy: it never returns -1. */
@@ -34,6 +34,20 @@ int strcmp(const char *a, const char *b)
 	return 0;
 }
 
+/*
+ * The clang build needs this, although gcc does not.
+ * Stolen from lib/string.c.
+ */
+void *memcpy(void *dest, const void *src, size_t count)
+{
+	char *tmp = dest;
+	const char *s = src;
+
+	while (count--)
+		*tmp++ = *s++;
+	return dest;
+}
+
 /* ...and two syscalls.  This is x86-specific. */
 static inline long x86_syscall3(long nr, long a0, long a1, long a2)
 {
@@ -70,7 +84,7 @@ void to_base10(char *lastdig, time_t n)
 	}
 }
 
-__attribute__((externally_visible)) void c_main(void **stack)
+void c_main(void **stack)
 {
 	/* Parse the stack */
 	long argc = (long)*stack;

base-commit: 2bfcfd584ff5ccc8bb7acde19b42570414bf880b

Signature · 1 line
2.45.1
Reviewed-by Tested-by on v2 1/1
#1 lore raw
Quoted 21 lines · John Hubbard

When building with clang, via:

make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and allows these tests to run and pass.

1. Fix linker error (undefined reference to memcpy) by providing a local version of memcpy.

2. clang complains about using this form:

    if (g = h & 0xf0000000)

...so factor out the assignment into a separate step.

3. The code is passing a signed const char* to elf_hash(), which expects a const unsigned char *. There are several callers, so fix this at the source by allowing the function to accept a signed argument, and then converting to unsigned operations, once inside the function.

Quoted 4 lines

4. clang doesn't have __attribute__((externally_visible)) and generates a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem to require that attribute in order to build, run and pass tests here, so remove it.

Just checked with GCC 5.1, it builds fine without any errors.

Quoted 4 lines

[1] https://lore.kernel.org/all/20240329-selftests-libmk-llvm-rfc-v1-1-2f9ed7d1c49f@valentinobst.de/

LGTM

Quoted 88 lines

---

Changes since the first version:

1) Rebased onto Linux 6.10-rc1

thanks,
John Hubbard

 tools/testing/selftests/vDSO/parse_vdso.c      | 16 +++++++++++-----
 .../selftests/vDSO/vdso_standalone_test_x86.c  | 18 ++++++++++++++++--
 2 files changed, 27 insertions(+), 7 deletions(-)
tools/testing/selftests/vDSO/parse_vdso.c +11 -5
index 413f75620a35..4ae417372e9e 100644
@@ -55,14 +55,20 @@ static struct vdso_info
 	ELF(Verdef) *verdef;
 } vdso_info;
 
-/* Straight from the ELF specification. */
-static unsigned long elf_hash(const unsigned char *name)
+/*
+ * Straight from the ELF specification...and then tweaked slightly, in order to
+ * avoid a few clang warnings.
+ */
+static unsigned long elf_hash(const char *name)
 {
 	unsigned long h = 0, g;
-	while (*name)
+	const unsigned char *uch_name = (const unsigned char *)name;
+
+	while (*uch_name)
 	{
-		h = (h << 4) + *name++;
-		if (g = h & 0xf0000000)
+		h = (h << 4) + *uch_name++;
+		g = h & 0xf0000000;
+		if (g)
 			h ^= g >> 24;
 		h &= ~g;
 	}
tools/testing/selftests/vDSO/vdso_standalone_test_x86.c +16 -2
index 8a44ff973ee1..27f6fdf11969 100644
@@ -18,7 +18,7 @@
 
 #include "parse_vdso.h"
 
-/* We need a libc functions... */
+/* We need some libc functions... */
 int strcmp(const char *a, const char *b)
 {
 	/* This implementation is buggy: it never returns -1. */
@@ -34,6 +34,20 @@ int strcmp(const char *a, const char *b)
 	return 0;
 }
 
+/*
+ * The clang build needs this, although gcc does not.
+ * Stolen from lib/string.c.
+ */
+void *memcpy(void *dest, const void *src, size_t count)
+{
+	char *tmp = dest;
+	const char *s = src;
+
+	while (count--)
+		*tmp++ = *s++;
+	return dest;
+}
+
 /* ...and two syscalls.  This is x86-specific. */
 static inline long x86_syscall3(long nr, long a0, long a1, long a2)
 {
@@ -70,7 +84,7 @@ void to_base10(char *lastdig, time_t n)
 	}
 }
 
-__attribute__((externally_visible)) void c_main(void **stack)
+void c_main(void **stack)
 {
 	/* Parse the stack */
 	long argc = (long)*stack;

base-commit: 2bfcfd584ff5ccc8bb7acde19b42570414bf880b

Signature · 2 lines
BR,
Muhammad Usama Anjum
#2 lore raw
Quoted 29 lines · Muhammad Usama Anjum
Quoted 21 lines · John Hubbard

When building with clang, via:

make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and allows these tests to run and pass.

1. Fix linker error (undefined reference to memcpy) by providing a local version of memcpy.

2. clang complains about using this form:

     if (g = h & 0xf0000000)

...so factor out the assignment into a separate step.

3. The code is passing a signed const char* to elf_hash(), which expects a const unsigned char *. There are several callers, so fix this at the source by allowing the function to accept a signed argument, and then converting to unsigned operations, once inside the function.

Quoted 4 lines

4. clang doesn't have __attribute__((externally_visible)) and generates a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem to require that attribute in order to build, run and pass tests here, so remove it.

Just checked with GCC 5.1, it builds fine without any errors.

Quoted 1 line
Quoted 7 lines
Quoted 3 lines

[1] https://lore.kernel.org/all/20240329-selftests-libmk-llvm-rfc-v1-1-2f9ed7d1c49f@valentinobst.de/

LGTM

Thanks for the review and testing, Muhammad!

I'm also adding Andy Lutomirski to this thread, in case Shuah would like the vDSO maintainer to ack or review. (scripts/get_maintainer.pl is letting me down, I will need to run "git blame" more I guess.

thanks,

Signature · 2 lines
John Hubbard
NVIDIA
Quoted 89 lines
Quoted 88 lines

---

Changes since the first version:

1) Rebased onto Linux 6.10-rc1

thanks,
John Hubbard

  tools/testing/selftests/vDSO/parse_vdso.c      | 16 +++++++++++-----
  .../selftests/vDSO/vdso_standalone_test_x86.c  | 18 ++++++++++++++++--
  2 files changed, 27 insertions(+), 7 deletions(-)
tools/testing/selftests/vDSO/parse_vdso.c +11 -5
index 413f75620a35..4ae417372e9e 100644
@@ -55,14 +55,20 @@ static struct vdso_info
  	ELF(Verdef) *verdef;
  } vdso_info;
 
-/* Straight from the ELF specification. */
-static unsigned long elf_hash(const unsigned char *name)
+/*
+ * Straight from the ELF specification...and then tweaked slightly, in order to
+ * avoid a few clang warnings.
+ */
+static unsigned long elf_hash(const char *name)
  {
  	unsigned long h = 0, g;
-	while (*name)
+	const unsigned char *uch_name = (const unsigned char *)name;
+
+	while (*uch_name)
  	{
-		h = (h << 4) + *name++;
-		if (g = h & 0xf0000000)
+		h = (h << 4) + *uch_name++;
+		g = h & 0xf0000000;
+		if (g)
  			h ^= g >> 24;
  		h &= ~g;
  	}
tools/testing/selftests/vDSO/vdso_standalone_test_x86.c +16 -2
index 8a44ff973ee1..27f6fdf11969 100644
@@ -18,7 +18,7 @@
 
  #include "parse_vdso.h"
 
-/* We need a libc functions... */
+/* We need some libc functions... */
  int strcmp(const char *a, const char *b)
  {
  	/* This implementation is buggy: it never returns -1. */
@@ -34,6 +34,20 @@ int strcmp(const char *a, const char *b)
  	return 0;
  }
 
+/*
+ * The clang build needs this, although gcc does not.
+ * Stolen from lib/string.c.
+ */
+void *memcpy(void *dest, const void *src, size_t count)
+{
+	char *tmp = dest;
+	const char *s = src;
+
+	while (count--)
+		*tmp++ = *s++;
+	return dest;
+}
+
  /* ...and two syscalls.  This is x86-specific. */
  static inline long x86_syscall3(long nr, long a0, long a1, long a2)
  {
@@ -70,7 +84,7 @@ void to_base10(char *lastdig, time_t n)
  	}
  }
 
-__attribute__((externally_visible)) void c_main(void **stack)
+void c_main(void **stack)
  {
  	/* Parse the stack */
  	long argc = (long)*stack;

base-commit: 2bfcfd584ff5ccc8bb7acde19b42570414bf880b

Reviewed-by on v2 1/1
#3 lore raw
In reply to John Hubbard
Quoted 110 lines

When building with clang, via:

make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and allows these tests to run and pass.

1. Fix linker error (undefined reference to memcpy) by providing a local version of memcpy.

2. clang complains about using this form:

    if (g = h & 0xf0000000)

...so factor out the assignment into a separate step.

3. The code is passing a signed const char* to elf_hash(), which expects a const unsigned char *. There are several callers, so fix this at the source by allowing the function to accept a signed argument, and then converting to unsigned operations, once inside the function.

4. clang doesn't have __attribute__((externally_visible)) and generates a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem to require that attribute in order to build, run and pass tests here, so remove it.

[1] https://lore.kernel.org/all/20240329-selftests-libmk-llvm-rfc-v1-1-2f9ed7d1c49f@valentinobst.de/

---

Changes since the first version:

1) Rebased onto Linux 6.10-rc1

thanks,
John Hubbard

 tools/testing/selftests/vDSO/parse_vdso.c      | 16 +++++++++++-----
 .../selftests/vDSO/vdso_standalone_test_x86.c  | 18 ++++++++++++++++--
 2 files changed, 27 insertions(+), 7 deletions(-)

base-commit: 2bfcfd584ff5ccc8bb7acde19b42570414bf880b

tools/testing/selftests/vDSO/parse_vdso.c +11 -5
index 413f75620a35..4ae417372e9e 100644
@@ -55,14 +55,20 @@ static struct vdso_info
 	ELF(Verdef) *verdef;
 } vdso_info;
 
-/* Straight from the ELF specification. */
-static unsigned long elf_hash(const unsigned char *name)
+/*
+ * Straight from the ELF specification...and then tweaked slightly, in order to
+ * avoid a few clang warnings.
+ */
+static unsigned long elf_hash(const char *name)
 {
 	unsigned long h = 0, g;
-	while (*name)
+	const unsigned char *uch_name = (const unsigned char *)name;
+
+	while (*uch_name)
 	{
-		h = (h << 4) + *name++;
-		if (g = h & 0xf0000000)
+		h = (h << 4) + *uch_name++;
+		g = h & 0xf0000000;
+		if (g)
 			h ^= g >> 24;
 		h &= ~g;
 	}
tools/testing/selftests/vDSO/vdso_standalone_test_x86.c +15 -1
index 8a44ff973ee1..27f6fdf11969 100644
@@ -18,7 +18,7 @@
 
 #include "parse_vdso.h"
 
-/* We need a libc functions... */
+/* We need some libc functions... */
 int strcmp(const char *a, const char *b)
 {
 	/* This implementation is buggy: it never returns -1. */
@@ -34,6 +34,20 @@ int strcmp(const char *a, const char *b)
 	return 0;
 }
 
+/*
+ * The clang build needs this, although gcc does not.
+ * Stolen from lib/string.c.
+ */
+void *memcpy(void *dest, const void *src, size_t count)
+{
+	char *tmp = dest;
+	const char *s = src;
+
+	while (count--)
+		*tmp++ = *s++;
+	return dest;
+}
+
 /* ...and two syscalls.  This is x86-specific. */
 static inline long x86_syscall3(long nr, long a0, long a1, long a2)
 {

tools/testing/selftests/vDSO/vdso_standalone_test_x86.c void to_base10(char *lastdig, time_t n) · line 90

Quoted 8 lines
tools/testing/selftests/vDSO/vdso_standalone_test_x86.c +1 -1
@@ -70,7 +84,7 @@ void to_base10(char *lastdig, time_t n)
 	}
 }
 
-__attribute__((externally_visible)) void c_main(void **stack)
+void c_main(void **stack)
 {
 	/* Parse the stack */
 	long argc = (long)*stack;

These work for me for compiling with clang.

Reviewed-by on v2 1/1
#4 lore raw
In reply to John Hubbard
Quoted 6 lines · John Hubbard

When building with clang, via:

make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and allows these tests to run and pass.

It might be best to split the 4 _different_ fixes into separate patches.

Quoted 9 lines

1. Fix linker error (undefined reference to memcpy) by providing a local version of memcpy.

2. clang complains about using this form:

    if (g = h & 0xf0000000)

...so factor out the assignment into a separate step.

There has been multiple attempts to fix this. I can see these two:
https://lore.kernel.org/all/[email protected]/
https://lore.kernel.org/all/[email protected]/

... I guess we somehow missed those?

Quoted 6 lines

3. The code is passing a signed const char* to elf_hash(), which expects a const unsigned char *. There are several callers, so fix this at the source by allowing the function to accept a signed argument, and then converting to unsigned operations, once inside the function.

There is also a v4 fix for this item that was sent out here:
https://lore.kernel.org/all/[email protected]/

Quoted 6 lines

4. clang doesn't have __attribute__((externally_visible)) and generates a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem to require that attribute in order to build, run and pass tests here, so remove it.

[1] https://lore.kernel.org/all/20240329-selftests-libmk-llvm-rfc-v1-1-2f9ed7d1c49f@valentinobst.de/

What is this about? Left over from v1 maybe?

Quoted 3 lines

---

I would prefer to pick up the fixes from folks who sent out the patches first but I'm fine either way.

#5 lore raw
Quoted 9 lines · Carlos Llamas
Quoted 6 lines · John Hubbard

When building with clang, via:

make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and allows these tests to run and pass.

It might be best to split the 4 _different_ fixes into separate patches.

If necessary, that can be done. It's sufficiently difficult to attract attention for selftests (see below) that I'm reluctant to make it even harder to get it all fixed, though.

Quoted 9 lines
Quoted 9 lines

1. Fix linker error (undefined reference to memcpy) by providing a local version of memcpy.

2. clang complains about using this form:

     if (g = h & 0xf0000000)
Quoted 7 lines
Quoted 1 line

...so factor out the assignment into a separate step.

There has been multiple attempts to fix this. I can see these two:
https://lore.kernel.org/all/[email protected]/
https://lore.kernel.org/all/[email protected]/

... I guess we somehow missed those?

I guess so. :)

Quoted 10 lines
Quoted 7 lines

3. The code is passing a signed const char* to elf_hash(), which expects a const unsigned char *. There are several callers, so fix this at the source by allowing the function to accept a signed argument, and then converting to unsigned operations, once inside the function.

There is also a v4 fix for this item that was sent out here:
https://lore.kernel.org/all/[email protected]/

No idea why these fixes are not getting picked up.

Quoted 9 lines
Quoted 7 lines

4. clang doesn't have __attribute__((externally_visible)) and generates a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem to require that attribute in order to build, run and pass tests here, so remove it.

[1] https://lore.kernel.org/all/20240329-selftests-libmk-llvm-rfc-v1-1-2f9ed7d1c49f@valentinobst.de/

What is this about? Left over from v1 maybe?

Maybe.

Quoted 7 lines
Quoted 4 lines

---

I would prefer to pick up the fixes from folks who sent out the patches first but I'm fine either way.

I'm perfectly fine with dropping this and letting the other patches go in, but *something* needs to go in.

Quoted 2 lines

Thanks for the review! I have no earthly idea what will happen next. I'd like to hear at least something from the maintainers about their intentions here.

thanks,

Signature · 2 lines
John Hubbard
NVIDIA
#6 lore raw
Quoted 2 lines · John Hubbard
Quoted 1 line · Carlos Llamas

On Mon, May 27, 2024 at 02:16:22PM -0700, John Hubbard wrote:

...

Quoted 6 lines
Quoted 1 line

Thanks for the review! I have no earthly idea what will happen next. I'd like to hear at least something from the maintainers about their intentions here.

In fact, I have two more patches for vDSO (these are separate issues, for the Makefile), and I just noticed that there are several accumulated Reviewed-by and Tested-by tags on this patch here. So I think the way forward is this:

I'll post a v3, with three patches for vDSO selftests, and the latest tags. And let's see how that fares.

thanks,

Signature · 2 lines
John Hubbard
NVIDIA
#7 lore raw
Quoted 11 lines · John Hubbard
Quoted 2 lines · John Hubbard
Quoted 1 line · Carlos Llamas

On Mon, May 27, 2024 at 02:16:22PM -0700, John Hubbard wrote:

...

Quoted 6 lines
Quoted 1 line

Thanks for the review! I have no earthly idea what will happen next. I'd like to hear at least something from the maintainers about their intentions here.

Quoted 8 lines

In fact, I have two more patches for vDSO (these are separate issues, for the Makefile), and I just noticed that there are several accumulated Reviewed-by and Tested-by tags on this patch here. So I think the way forward is this:

I'll post a v3, with three patches for vDSO selftests, and the latest tags. And let's see how that fares.

Yeap, a resend would have been good. This time it might be best to bring Andy and Thomas in the loop too.

Cc Andy Lutomirski <[email protected]>
Cc Thomas Gleixner <[email protected]>
Quoted 5 lines

thanks,

Signature · 2 lines
John Hubbard
NVIDIA