linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline
@ 2019-06-18 11:19 Arnd Bergmann
  2019-06-18 18:05 ` Eric Biggers
  2019-06-28  4:19 ` Herbert Xu
  0 siblings, 2 replies; 5+ messages in thread
From: Arnd Bergmann @ 2019-06-18 11:19 UTC (permalink / raw)
  To: Herbert Xu, David S. Miller
  Cc: Arnd Bergmann, Eric Biggers, linux-crypto, linux-kernel

The same bug that gcc hit in the past is apparently now showing
up with clang, which decides to inline __serpent_setkey_sbox:

crypto/serpent_generic.c:268:5: error: stack frame size of 2112 bytes in function '__serpent_setkey' [-Werror,-Wframe-larger-than=]

Marking it 'noinline' reduces the stack usage from 2112 bytes to
192 and 96 bytes, respectively, and seems to generate more
useful object code.

Fixes: c871c10e4ea7 ("crypto: serpent - improve __serpent_setkey with UBSAN")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
v2: style improvements suggested by Eric Biggers
---
 crypto/serpent_generic.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/crypto/serpent_generic.c b/crypto/serpent_generic.c
index e57757904677..56fa665a4f01 100644
--- a/crypto/serpent_generic.c
+++ b/crypto/serpent_generic.c
@@ -225,7 +225,13 @@
 	x4 ^= x2;					\
 	})
 
-static void __serpent_setkey_sbox(u32 r0, u32 r1, u32 r2, u32 r3, u32 r4, u32 *k)
+/*
+ * both gcc and clang have misoptimized this function in the past,
+ * producing horrible object code from spilling temporary variables
+ * on the stack. Forcing this part out of line avoids that.
+ */
+static noinline void __serpent_setkey_sbox(u32 r0, u32 r1, u32 r2,
+					   u32 r3, u32 r4, u32 *k)
 {
 	k += 100;
 	S3(r3, r4, r0, r1, r2); store_and_load_keys(r1, r2, r4, r3, 28, 24);
-- 
2.20.0


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline
  2019-06-18 11:19 [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline Arnd Bergmann
@ 2019-06-18 18:05 ` Eric Biggers
  2019-06-28  4:19 ` Herbert Xu
  1 sibling, 0 replies; 5+ messages in thread
From: Eric Biggers @ 2019-06-18 18:05 UTC (permalink / raw)
  To: Arnd Bergmann; +Cc: Herbert Xu, David S. Miller, linux-crypto, linux-kernel

On Tue, Jun 18, 2019 at 01:19:42PM +0200, Arnd Bergmann wrote:
> The same bug that gcc hit in the past is apparently now showing
> up with clang, which decides to inline __serpent_setkey_sbox:
> 
> crypto/serpent_generic.c:268:5: error: stack frame size of 2112 bytes in function '__serpent_setkey' [-Werror,-Wframe-larger-than=]
> 
> Marking it 'noinline' reduces the stack usage from 2112 bytes to
> 192 and 96 bytes, respectively, and seems to generate more
> useful object code.
> 
> Fixes: c871c10e4ea7 ("crypto: serpent - improve __serpent_setkey with UBSAN")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> v2: style improvements suggested by Eric Biggers
> ---
>  crypto/serpent_generic.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/crypto/serpent_generic.c b/crypto/serpent_generic.c
> index e57757904677..56fa665a4f01 100644
> --- a/crypto/serpent_generic.c
> +++ b/crypto/serpent_generic.c
> @@ -225,7 +225,13 @@
>  	x4 ^= x2;					\
>  	})
>  
> -static void __serpent_setkey_sbox(u32 r0, u32 r1, u32 r2, u32 r3, u32 r4, u32 *k)
> +/*
> + * both gcc and clang have misoptimized this function in the past,
> + * producing horrible object code from spilling temporary variables
> + * on the stack. Forcing this part out of line avoids that.
> + */
> +static noinline void __serpent_setkey_sbox(u32 r0, u32 r1, u32 r2,
> +					   u32 r3, u32 r4, u32 *k)
>  {
>  	k += 100;
>  	S3(r3, r4, r0, r1, r2); store_and_load_keys(r1, r2, r4, r3, 28, 24);
> -- 
> 2.20.0
> 

Reviewed-by: Eric Biggers <ebiggers@kernel.org>

- Eric

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline
  2019-06-18 11:19 [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline Arnd Bergmann
  2019-06-18 18:05 ` Eric Biggers
@ 2019-06-28  4:19 ` Herbert Xu
  2019-06-28 17:04   ` Eric Biggers
  1 sibling, 1 reply; 5+ messages in thread
From: Herbert Xu @ 2019-06-28  4:19 UTC (permalink / raw)
  To: Arnd Bergmann; +Cc: David S. Miller, Eric Biggers, linux-crypto, linux-kernel

On Tue, Jun 18, 2019 at 01:19:42PM +0200, Arnd Bergmann wrote:
> The same bug that gcc hit in the past is apparently now showing
> up with clang, which decides to inline __serpent_setkey_sbox:
> 
> crypto/serpent_generic.c:268:5: error: stack frame size of 2112 bytes in function '__serpent_setkey' [-Werror,-Wframe-larger-than=]
> 
> Marking it 'noinline' reduces the stack usage from 2112 bytes to
> 192 and 96 bytes, respectively, and seems to generate more
> useful object code.
> 
> Fixes: c871c10e4ea7 ("crypto: serpent - improve __serpent_setkey with UBSAN")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
> v2: style improvements suggested by Eric Biggers
> ---
>  crypto/serpent_generic.c | 8 +++++++-
>  1 file changed, 7 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

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline
  2019-06-28  4:19 ` Herbert Xu
@ 2019-06-28 17:04   ` Eric Biggers
  2019-06-29  1:15     ` Herbert Xu
  0 siblings, 1 reply; 5+ messages in thread
From: Eric Biggers @ 2019-06-28 17:04 UTC (permalink / raw)
  To: Herbert Xu; +Cc: Arnd Bergmann, David S. Miller, linux-crypto, linux-kernel

On Fri, Jun 28, 2019 at 12:19:12PM +0800, Herbert Xu wrote:
> On Tue, Jun 18, 2019 at 01:19:42PM +0200, Arnd Bergmann wrote:
> > The same bug that gcc hit in the past is apparently now showing
> > up with clang, which decides to inline __serpent_setkey_sbox:
> > 
> > crypto/serpent_generic.c:268:5: error: stack frame size of 2112 bytes in function '__serpent_setkey' [-Werror,-Wframe-larger-than=]
> > 
> > Marking it 'noinline' reduces the stack usage from 2112 bytes to
> > 192 and 96 bytes, respectively, and seems to generate more
> > useful object code.
> > 
> > Fixes: c871c10e4ea7 ("crypto: serpent - improve __serpent_setkey with UBSAN")
> > Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > ---
> > v2: style improvements suggested by Eric Biggers
> > ---
> >  crypto/serpent_generic.c | 8 +++++++-
> >  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> Patch applied.  Thanks.
> -- 

Hi Herbert, seems you forgot to push?  I don't see this in cryptodev.

- Eric

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline
  2019-06-28 17:04   ` Eric Biggers
@ 2019-06-29  1:15     ` Herbert Xu
  0 siblings, 0 replies; 5+ messages in thread
From: Herbert Xu @ 2019-06-29  1:15 UTC (permalink / raw)
  To: Eric Biggers; +Cc: Arnd Bergmann, David S. Miller, linux-crypto, linux-kernel

On Fri, Jun 28, 2019 at 10:04:38AM -0700, Eric Biggers wrote:
>
> Hi Herbert, seems you forgot to push?  I don't see this in cryptodev.

Yes you're right.  It should be there now.

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

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2019-06-29  1:16 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-06-18 11:19 [PATCH] [v2] crypto: serpent - mark __serpent_setkey_sbox noinline Arnd Bergmann
2019-06-18 18:05 ` Eric Biggers
2019-06-28  4:19 ` Herbert Xu
2019-06-28 17:04   ` Eric Biggers
2019-06-29  1:15     ` Herbert Xu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).