Re: [PATCH 1/2] crypto: drivers - avoid memcpy size warning

[Herbert Xu]

On Mon, Aug 07, 2023 at 02:04:05PM +0200, Arnd Bergmann wrote:
>
> See https://pastebin.com/raw/ip3tfpJF for a config that triggers this
> on x86 with the chelsio and atmel drivers. The bcm driver is only
> available on arm64, so you won't hit that one here. I also
> see this with allmodconfig, as well as defconfig after enabling
> CONFIG_FORTIFY_SOURCE and the three crypto drivers.

OK I can reproduce this now:

In file included from ../include/linux/string.h:254,
                 from ../arch/x86/include/asm/page_32.h:18,
                 from ../arch/x86/include/asm/page.h:14,
                 from ../arch/x86/include/asm/processor.h:20,
                 from ../arch/x86/include/asm/timex.h:5,
                 from ../include/linux/timex.h:67,
                 from ../include/linux/time32.h:13,
                 from ../include/linux/time.h:60,
                 from ../include/linux/stat.h:19,
                 from ../include/linux/module.h:13,
                 from ../drivers/crypto/atmel-sha.c:15:
../drivers/crypto/atmel-sha.c: In function ‘atmel_sha_hmac_compute_ipad_hash’:
../include/linux/fortify-string.h:57:33: error: ‘__builtin_memcpy’ accessing 129 or more bytes at offsets 304 and 176 overlaps 1 or more bytes at offset 304 [-Werror=restrict]
   57 | #define __underlying_memcpy     __builtin_memcpy
      |                                 ^
../include/linux/fortify-string.h:648:9: note: in expansion of macro ‘__underlying_memcpy’
  648 |         __underlying_##op(p, q, __fortify_size);                        \
      |         ^~~~~~~~~~~~~
../include/linux/fortify-string.h:693:26: note: in expansion of macro ‘__fortify_memcpy_chk’
  693 | #define memcpy(p, q, s)  __fortify_memcpy_chk(p, q, s,                  \
      |                          ^~~~~~~~~~~~~~~~~~~~
../drivers/crypto/atmel-sha.c:1773:9: note: in expansion of macro ‘memcpy’
 1773 |         memcpy(hmac->opad, hmac->ipad, bs);
      |         ^~~~~~
../include/linux/fortify-string.h:57:33: error: ‘__builtin_memcpy’ accessing 129 or more bytes at offsets 304 and 176 overlaps 1 or more bytes at offset 304 [-Werror=restrict]
   57 | #define __underlying_memcpy     __builtin_memcpy
      |                                 ^
../include/linux/fortify-string.h:648:9: note: in expansion of macro ‘__underlying_memcpy’
  648 |         __underlying_##op(p, q, __fortify_size);                        \
      |         ^~~~~~~~~~~~~
../include/linux/fortify-string.h:693:26: note: in expansion of macro ‘__fortify_memcpy_chk’
  693 | #define memcpy(p, q, s)  __fortify_memcpy_chk(p, q, s,                  \
      |                          ^~~~~~~~~~~~~~~~~~~~
../drivers/crypto/atmel-sha.c:1773:9: note: in expansion of macro ‘memcpy’
 1773 |         memcpy(hmac->opad, hmac->ipad, bs);
      |         ^~~~~~
cc1: all warnings being treated as errors

But why are we turning these warnings on if they're giving completely
bogus false positives like this?

	struct atmel_sha_hmac_ctx {
		struct atmel_sha_ctx	base;

		struct atmel_sha_hmac_key	hkey;
		u32			ipad[SHA512_BLOCK_SIZE / sizeof(u32)];
		u32			opad[SHA512_BLOCK_SIZE / sizeof(u32)];
		atmel_sha_fn_t		resume;
	};

	struct atmel_sha_hmac_ctx *hmac = crypto_ahash_ctx(tfm);
	size_t bs = ctx->block_size;

	memcpy(hmac->opad, hmac->ipad, bs);

The block_size is set by the algorithm, you can easily grep for
it in atmel-sha.c and the biggest one there is SHA512_BLOCK_SIZE,
which is how big hmac->ipad/hmac->opad are.

So logically this code is perfectly fine.

There is no way for the compiler to know how big ctx->block_size is.
So why do we expect it to make deductions on how big bs can be?

This warning looks broken to me.

It looks like there is already a solution to this though.  Just use
unsafe_memcpy and be done with it.

Cheers,
-- 
Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt