powerpc/atomic: Use YZ constraints for DS-form instructions

[PATCH] powerpc/atomic: Use YZ constraints for DS-form instructions

Posted by Michael Ellerman 2 months, 2 weeks ago

The 'ld' and 'std' instructions require a 4-byte aligned displacement
because they are DS-form instructions. But the "m" asm constraint
doesn't enforce that.

That can lead to build errors if the compiler chooses a non-aligned
displacement, as seen with GCC 14:

  /tmp/ccuSzwiR.s: Assembler messages:
  /tmp/ccuSzwiR.s:2579: Error: operand out of domain (39 is not a multiple of 4)
  make[5]: *** [scripts/Makefile.build:229: net/core/page_pool.o] Error 1

Dumping the generated assembler shows:

  ld 8,39(8)       # MEM[(const struct atomic64_t *)_29].counter, t

Use the YZ constraints to tell the compiler either to generate a DS-form
displacement, or use an X-form instruction, either of which prevents the
build error.

See commit 2d43cc701b96 ("powerpc/uaccess: Fix build errors seen with
GCC 13/14") for more details on the constraint letters.

Fixes: 9f0cbea0d8cc ("[POWERPC] Implement atomic{, 64}_{read, write}() without volatile")
Cc: stable@vger.kernel.org # v2.6.24+
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Closes: https://lore.kernel.org/all/20240913125302.0a06b4c7@canb.auug.org.au
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 arch/powerpc/include/asm/asm-compat.h | 6 ++++++
 arch/powerpc/include/asm/atomic.h     | 5 +++--
 arch/powerpc/include/asm/uaccess.h    | 7 +------
 3 files changed, 10 insertions(+), 8 deletions(-)

diff --git a/arch/powerpc/include/asm/asm-compat.h b/arch/powerpc/include/asm/asm-compat.h
index 2bc53c646ccd..83848b534cb1 100644
--- a/arch/powerpc/include/asm/asm-compat.h
+++ b/arch/powerpc/include/asm/asm-compat.h
@@ -39,6 +39,12 @@
 #define STDX_BE	stringify_in_c(stdbrx)
 #endif
 
+#ifdef CONFIG_CC_IS_CLANG
+#define DS_FORM_CONSTRAINT "Z<>"
+#else
+#define DS_FORM_CONSTRAINT "YZ<>"
+#endif
+
 #else /* 32-bit */
 
 /* operations for longs and pointers */
diff --git a/arch/powerpc/include/asm/atomic.h b/arch/powerpc/include/asm/atomic.h
index 5bf6a4d49268..d1ea554c33ed 100644
--- a/arch/powerpc/include/asm/atomic.h
+++ b/arch/powerpc/include/asm/atomic.h
@@ -11,6 +11,7 @@
 #include <asm/cmpxchg.h>
 #include <asm/barrier.h>
 #include <asm/asm-const.h>
+#include <asm/asm-compat.h>
 
 /*
  * Since *_return_relaxed and {cmp}xchg_relaxed are implemented with
@@ -197,7 +198,7 @@ static __inline__ s64 arch_atomic64_read(const atomic64_t *v)
 	if (IS_ENABLED(CONFIG_PPC_KERNEL_PREFIXED))
 		__asm__ __volatile__("ld %0,0(%1)" : "=r"(t) : "b"(&v->counter));
 	else
-		__asm__ __volatile__("ld%U1%X1 %0,%1" : "=r"(t) : "m<>"(v->counter));
+		__asm__ __volatile__("ld%U1%X1 %0,%1" : "=r"(t) : DS_FORM_CONSTRAINT (v->counter));
 
 	return t;
 }
@@ -208,7 +209,7 @@ static __inline__ void arch_atomic64_set(atomic64_t *v, s64 i)
 	if (IS_ENABLED(CONFIG_PPC_KERNEL_PREFIXED))
 		__asm__ __volatile__("std %1,0(%2)" : "=m"(v->counter) : "r"(i), "b"(&v->counter));
 	else
-		__asm__ __volatile__("std%U0%X0 %1,%0" : "=m<>"(v->counter) : "r"(i));
+		__asm__ __volatile__("std%U0%X0 %1,%0" : "=" DS_FORM_CONSTRAINT (v->counter) : "r"(i));
 }
 
 #define ATOMIC64_OP(op, asm_op)						\
diff --git a/arch/powerpc/include/asm/uaccess.h b/arch/powerpc/include/asm/uaccess.h
index fd594bf6c6a9..4f5a46a77fa2 100644
--- a/arch/powerpc/include/asm/uaccess.h
+++ b/arch/powerpc/include/asm/uaccess.h
@@ -6,6 +6,7 @@
 #include <asm/page.h>
 #include <asm/extable.h>
 #include <asm/kup.h>
+#include <asm/asm-compat.h>
 
 #ifdef __powerpc64__
 /* We use TASK_SIZE_USER64 as TASK_SIZE is not constant */
@@ -92,12 +93,6 @@ __pu_failed:							\
 		: label)
 #endif
 
-#ifdef CONFIG_CC_IS_CLANG
-#define DS_FORM_CONSTRAINT "Z<>"
-#else
-#define DS_FORM_CONSTRAINT "YZ<>"
-#endif
-
 #ifdef __powerpc64__
 #ifdef CONFIG_PPC_KERNEL_PREFIXED
 #define __put_user_asm2_goto(x, ptr, label)			\
-- 
2.46.0

Re: [PATCH] powerpc/atomic: Use YZ constraints for DS-form instructions

Posted by Michael Ellerman 2 months, 1 week ago

On Mon, 16 Sep 2024 22:05:10 +1000, Michael Ellerman wrote:
> The 'ld' and 'std' instructions require a 4-byte aligned displacement
> because they are DS-form instructions. But the "m" asm constraint
> doesn't enforce that.
> 
> That can lead to build errors if the compiler chooses a non-aligned
> displacement, as seen with GCC 14:
> 
> [...]

Applied to powerpc/next.

[1/1] powerpc/atomic: Use YZ constraints for DS-form instructions
      https://git.kernel.org/powerpc/c/39190ac7cff1fd15135fa8e658030d9646fdb5f2

cheers

Re: [PATCH] powerpc/atomic: Use YZ constraints for DS-form instructions

Posted by Segher Boessenkool 2 months, 1 week ago

Hi!

On Mon, Sep 16, 2024 at 10:05:10PM +1000, Michael Ellerman wrote:
> The 'ld' and 'std' instructions require a 4-byte aligned displacement
> because they are DS-form instructions. But the "m" asm constraint
> doesn't enforce that.
> 
> That can lead to build errors if the compiler chooses a non-aligned
> displacement, as seen with GCC 14:
> 
>   /tmp/ccuSzwiR.s: Assembler messages:
>   /tmp/ccuSzwiR.s:2579: Error: operand out of domain (39 is not a multiple of 4)
>   make[5]: *** [scripts/Makefile.build:229: net/core/page_pool.o] Error 1
> 
> Dumping the generated assembler shows:
> 
>   ld 8,39(8)       # MEM[(const struct atomic64_t *)_29].counter, t
> 
> Use the YZ constraints to tell the compiler either to generate a DS-form
> displacement, or use an X-form instruction, either of which prevents the
> build error.

Great explanation text, a perfect commit!  :-)

> See commit 2d43cc701b96 ("powerpc/uaccess: Fix build errors seen with
> GCC 13/14") for more details on the constraint letters.
> 
> Fixes: 9f0cbea0d8cc ("[POWERPC] Implement atomic{, 64}_{read, write}() without volatile")
> Cc: stable@vger.kernel.org # v2.6.24+
> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> Closes: https://lore.kernel.org/all/20240913125302.0a06b4c7@canb.auug.org.au
> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Reviewed-By: Segher Boessenkool <segher@kernel.crashing.org>


Segher

Re: [PATCH] powerpc/atomic: Use YZ constraints for DS-form instructions

Posted by Michael Ellerman 2 months, 1 week ago

Segher Boessenkool <segher@kernel.crashing.org> writes:
> Hi!
>
> On Mon, Sep 16, 2024 at 10:05:10PM +1000, Michael Ellerman wrote:
>> The 'ld' and 'std' instructions require a 4-byte aligned displacement
>> because they are DS-form instructions. But the "m" asm constraint
>> doesn't enforce that.
>> 
>> That can lead to build errors if the compiler chooses a non-aligned
>> displacement, as seen with GCC 14:
>> 
>>   /tmp/ccuSzwiR.s: Assembler messages:
>>   /tmp/ccuSzwiR.s:2579: Error: operand out of domain (39 is not a multiple of 4)
>>   make[5]: *** [scripts/Makefile.build:229: net/core/page_pool.o] Error 1
>> 
>> Dumping the generated assembler shows:
>> 
>>   ld 8,39(8)       # MEM[(const struct atomic64_t *)_29].counter, t
>> 
>> Use the YZ constraints to tell the compiler either to generate a DS-form
>> displacement, or use an X-form instruction, either of which prevents the
>> build error.
>
> Great explanation text, a perfect commit!  :-)

Thanks - I'm sure there's something that could be better, but I do try :)

cheers

Re: [PATCH] powerpc/atomic: Use YZ constraints for DS-form instructions

Posted by Mina Almasry 2 months, 2 weeks ago

On Mon, Sep 16, 2024 at 5:05 AM Michael Ellerman <mpe@ellerman.id.au> wrote:
>
> The 'ld' and 'std' instructions require a 4-byte aligned displacement
> because they are DS-form instructions. But the "m" asm constraint
> doesn't enforce that.
>
> That can lead to build errors if the compiler chooses a non-aligned
> displacement, as seen with GCC 14:
>
>   /tmp/ccuSzwiR.s: Assembler messages:
>   /tmp/ccuSzwiR.s:2579: Error: operand out of domain (39 is not a multiple of 4)
>   make[5]: *** [scripts/Makefile.build:229: net/core/page_pool.o] Error 1
>
> Dumping the generated assembler shows:
>
>   ld 8,39(8)       # MEM[(const struct atomic64_t *)_29].counter, t
>
> Use the YZ constraints to tell the compiler either to generate a DS-form
> displacement, or use an X-form instruction, either of which prevents the
> build error.
>
> See commit 2d43cc701b96 ("powerpc/uaccess: Fix build errors seen with
> GCC 13/14") for more details on the constraint letters.
>
> Fixes: 9f0cbea0d8cc ("[POWERPC] Implement atomic{, 64}_{read, write}() without volatile")
> Cc: stable@vger.kernel.org # v2.6.24+
> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> Closes: https://lore.kernel.org/all/20240913125302.0a06b4c7@canb.auug.org.au
> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

I'm not familiar enough with the code around the changes, but I have
been able to confirm cherry-picking this resolves the build issue I'm
seeing on net-next, so, FWIW,

Tested-by: Mina Almasry <almasrymina@google.com>

-- 
Thanks,
Mina

Re: [PATCH] powerpc/atomic: Use YZ constraints for DS-form instructions

Posted by Michael Ellerman 2 months, 1 week ago

Mina Almasry <almasrymina@google.com> writes:
> On Mon, Sep 16, 2024 at 5:05 AM Michael Ellerman <mpe@ellerman.id.au> wrote:
>>
>> The 'ld' and 'std' instructions require a 4-byte aligned displacement
>> because they are DS-form instructions. But the "m" asm constraint
>> doesn't enforce that.
>>
>> That can lead to build errors if the compiler chooses a non-aligned
>> displacement, as seen with GCC 14:
>>
>>   /tmp/ccuSzwiR.s: Assembler messages:
>>   /tmp/ccuSzwiR.s:2579: Error: operand out of domain (39 is not a multiple of 4)
>>   make[5]: *** [scripts/Makefile.build:229: net/core/page_pool.o] Error 1
>>
>> Dumping the generated assembler shows:
>>
>>   ld 8,39(8)       # MEM[(const struct atomic64_t *)_29].counter, t
>>
>> Use the YZ constraints to tell the compiler either to generate a DS-form
>> displacement, or use an X-form instruction, either of which prevents the
>> build error.
>>
>> See commit 2d43cc701b96 ("powerpc/uaccess: Fix build errors seen with
>> GCC 13/14") for more details on the constraint letters.
>>
>> Fixes: 9f0cbea0d8cc ("[POWERPC] Implement atomic{, 64}_{read, write}() without volatile")
>> Cc: stable@vger.kernel.org # v2.6.24+
>> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
>> Closes: https://lore.kernel.org/all/20240913125302.0a06b4c7@canb.auug.org.au
>> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
>
> I'm not familiar enough with the code around the changes, but I have
> been able to confirm cherry-picking this resolves the build issue I'm
> seeing on net-next, so, FWIW,
>
> Tested-by: Mina Almasry <almasrymina@google.com>

Thanks.

cheers