[PATCH] crypto: blake2b - Fix clang optimization for ARMv7-M

[PATCH] crypto: blake2b - Fix clang optimization for ARMv7-M

[Arnd Bergmann]

When building for ARMv7-M, clang-9 or higher tries to unroll some loops,
which ends up confusing the register allocator to the point of generating
rather bad code and using more than the warning limit for stack frames:

warning: stack frame size of 1200 bytes in function 'blake2b_compress' [-Wframe-larger-than=]

Forcing it to not unroll the final loop avoids this problem.

Fixes: 91d689337fe8 ("crypto: blake2b - add blake2b generic implementation")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 crypto/blake2b_generic.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/crypto/blake2b_generic.c b/crypto/blake2b_generic.c
index 1d262374fa4e..0ffd8d92e308 100644
--- a/crypto/blake2b_generic.c
+++ b/crypto/blake2b_generic.c
@@ -129,7 +129,9 @@ static void blake2b_compress(struct blake2b_state *S,
 	ROUND(9);
 	ROUND(10);
 	ROUND(11);
-
+#ifdef CONFIG_CC_IS_CLANG
+#pragma nounroll /* https://bugs.llvm.org/show_bug.cgi?id=45803 */
+#endif
 	for (i = 0; i < 8; ++i)
 		S->h[i] = S->h[i] ^ v[i] ^ v[i + 8];
 }
-- 
2.26.0

Re: [PATCH] crypto: blake2b - Fix clang optimization for ARMv7-M

[Nathan Chancellor]

On Tue, May 05, 2020 at 03:53:45PM +0200, Arnd Bergmann wrote:
> When building for ARMv7-M, clang-9 or higher tries to unroll some loops,
> which ends up confusing the register allocator to the point of generating
> rather bad code and using more than the warning limit for stack frames:
> 
> warning: stack frame size of 1200 bytes in function 'blake2b_compress' [-Wframe-larger-than=]
> 
> Forcing it to not unroll the final loop avoids this problem.
> 
> Fixes: 91d689337fe8 ("crypto: blake2b - add blake2b generic implementation")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
>  crypto/blake2b_generic.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/crypto/blake2b_generic.c b/crypto/blake2b_generic.c
> index 1d262374fa4e..0ffd8d92e308 100644
> --- a/crypto/blake2b_generic.c
> +++ b/crypto/blake2b_generic.c
> @@ -129,7 +129,9 @@ static void blake2b_compress(struct blake2b_state *S,
>  	ROUND(9);
>  	ROUND(10);
>  	ROUND(11);
> -
> +#ifdef CONFIG_CC_IS_CLANG

Given your comment in the bug:

"The code is written to assume no loops are unrolled"

Does it make sense to make this unconditional and take compiler
heuristics out of it?

> +#pragma nounroll /* https://bugs.llvm.org/show_bug.cgi?id=45803 */
> +#endif
>  	for (i = 0; i < 8; ++i)
>  		S->h[i] = S->h[i] ^ v[i] ^ v[i + 8];
>  }
> -- 
> 2.26.0
> 

Re: [PATCH] crypto: blake2b - Fix clang optimization for ARMv7-M

[Arnd Bergmann]

On Wed, May 6, 2020 at 7:12 AM Nathan Chancellor
<natechancellor@gmail.com> wrote:
> > -
> > +#ifdef CONFIG_CC_IS_CLANG
>
> Given your comment in the bug:
>
> "The code is written to assume no loops are unrolled"
>
> Does it make sense to make this unconditional and take compiler
> heuristics out of it?
>
> > +#pragma nounroll /* https://bugs.llvm.org/show_bug.cgi?id=45803 */
> > +#endif
> >       for (i = 0; i < 8; ++i)
> >               S->h[i] = S->h[i] ^ v[i] ^ v[i + 8];

No, that would not work, as gcc does not support this pragma.

        Arnd

Re: [PATCH] crypto: blake2b - Fix clang optimization for ARMv7-M

[Nathan Chancellor]

On Fri, May 08, 2020 at 11:31:07PM +0200, Arnd Bergmann wrote:
> On Wed, May 6, 2020 at 7:12 AM Nathan Chancellor
> <natechancellor@gmail.com> wrote:
> > > -
> > > +#ifdef CONFIG_CC_IS_CLANG
> >
> > Given your comment in the bug:
> >
> > "The code is written to assume no loops are unrolled"
> >
> > Does it make sense to make this unconditional and take compiler
> > heuristics out of it?
> >
> > > +#pragma nounroll /* https://bugs.llvm.org/show_bug.cgi?id=45803 */
> > > +#endif
> > >       for (i = 0; i < 8; ++i)
> > >               S->h[i] = S->h[i] ^ v[i] ^ v[i + 8];
> 
> No, that would not work, as gcc does not support this pragma.
> 
>         Arnd

Ah fair enough.

Reviewed-by: Nathan Chancellor <natechancellor@gmail.com>

Re: [PATCH] crypto: blake2b - Fix clang optimization for ARMv7-M

[Herbert Xu]

On Tue, May 05, 2020 at 03:53:45PM +0200, Arnd Bergmann wrote:
> When building for ARMv7-M, clang-9 or higher tries to unroll some loops,
> which ends up confusing the register allocator to the point of generating
> rather bad code and using more than the warning limit for stack frames:
> 
> warning: stack frame size of 1200 bytes in function 'blake2b_compress' [-Wframe-larger-than=]
> 
> Forcing it to not unroll the final loop avoids this problem.
> 
> Fixes: 91d689337fe8 ("crypto: blake2b - add blake2b generic implementation")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
>  crypto/blake2b_generic.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)

Patch applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt