Re: [PATCH] crypto: serpent - mark __serpent_setkey_sbox noinline

[Eric Biggers <ebiggers@kernel.org>]

Hi Arnd,

On Mon, Jun 17, 2019 at 01:59:08PM +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>
> ---
>  crypto/serpent_generic.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/crypto/serpent_generic.c b/crypto/serpent_generic.c
> index e57757904677..183661faf420 100644
> --- a/crypto/serpent_generic.c
> +++ b/crypto/serpent_generic.c
> @@ -225,7 +225,7 @@
>  	x4 ^= x2;					\
>  	})
>  
> -static void __serpent_setkey_sbox(u32 r0, u32 r1, u32 r2, u32 r3, u32 r4, u32 *k)
> +static void noinline __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
> 

A few nits: 'static noinline void' seems to be the much more common order,
presumably to match 'static inline void'.  Line breaking at 80 characters would
also be nice.  Finally, a brief comment explaining the purpose of 'noinline'
might be helpful for people working with this code later.

- Eric