Re: [PATCH v2 09/13] RISC-V: crypto: add Zvknha/b accelerated SHA224/256 implementations

[Eric Biggers]

On Tue, Nov 28, 2023 at 03:16:53PM +0800, Jerry Shih wrote:
> On Nov 28, 2023, at 12:12, Eric Biggers <ebiggers@xxxxxxxxxx> wrote:
> > On Mon, Nov 27, 2023 at 03:06:59PM +0800, Jerry Shih wrote:
> >> +/*
> >> + * sha256 using zvkb and zvknha/b vector crypto extension
> >> + *
> >> + * This asm function will just take the first 256-bit as the sha256 state from
> >> + * the pointer to `struct sha256_state`.
> >> + */
> >> +asmlinkage void
> >> +sha256_block_data_order_zvkb_zvknha_or_zvknhb(struct sha256_state *digest,
> >> +					      const u8 *data, int num_blks);
> > 
> > The SHA-2 and SM3 assembly functions are potentially being called using indirect
> > calls, depending on whether the compiler optimizes out the indirect call that
> > exists in the code or not.  These assembly functions also are not defined using
> > SYM_TYPED_FUNC_START.  This is not compatible with Control Flow Integrity
> > (CONFIG_CFI_CLANG); these indirect calls might generate CFI failures.
> > 
> > I recommend using wrapper functions to avoid this issue, like what is done in
> > arch/arm64/crypto/sha2-ce-glue.c.
> > 
> > - Eric
> 
> Here is the previous review comment for the assembly function wrapper:
> > > +asmlinkage void sha256_block_data_order_zvbb_zvknha(u32 *digest, const void *data,
> > > +					unsigned int num_blks);
> > > +
> > > +static void __sha256_block_data_order(struct sha256_state *sst, u8 const *src,
> > > +				      int blocks)
> > > +{
> > > +	sha256_block_data_order_zvbb_zvknha(sst->state, src, blocks);
> > > +}
> > Having a double-underscored function wrap around a non-underscored one like this
> > isn't conventional for Linux kernel code.  IIRC some of the other crypto code
> > happens to do this, but it really is supposed to be the other way around.
> > 
> > I think you should just declare the assembly function to take a 'struct
> > sha256_state', with a comment mentioning that only the 'u32 state[8]' at the
> > beginning is actually used.  That's what arch/x86/crypto/sha256_ssse3_glue.c
> > does, for example.  Then, __sha256_block_data_order() would be unneeded.
> 
> Do you mean that we need the wrapper functions back for both SHA-* and SM3?
> If yes, we also don't need to check the state offset like:
> 	BUILD_BUG_ON(offsetof(struct sha256_state, state) != 0);
> 
> Could we just use the `SYM_TYPED_FUNC_START` in asm directly without the
> wrappers?

Sorry, I forgot that I had recommended against wrapper functions earlier.  I
didn't realize that SYM_TYPED_FUNC_START was missing.  Yes, you can also do it
without wrapper functions if you add SYM_TYPED_FUNC_START to the assembly.

- Eric