Re: [PATCH] memstick: r592: ignore kfifo_out() return code again

[Nick Desaulniers <ndesaulniers@google.com>]

On Wed, Apr 21, 2021 at 6:52 AM Arnd Bergmann <arnd@kernel.org> wrote:
>
> From: Arnd Bergmann <arnd@arndb.de>
>
> A minor cleanup to address a clang warning removed an assigned
> but unused local variable, but this now caused a gcc warning as
> kfifo_out() is annotated to require checking its return code:
>
> In file included from drivers/memstick/host/r592.h:13,
>                  from drivers/memstick/host/r592.c:21:
> drivers/memstick/host/r592.c: In function 'r592_flush_fifo_write':
> include/linux/kfifo.h:588:1: error: ignoring return value of '__kfifo_uint_must_check_helper' declared with attribute 'warn_unused_result' [-Werror=unused-result]
>   588 | __kfifo_uint_must_check_helper( \
>       | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   589 | ({ \
>       | ~~~~
>   590 |         typeof((fifo) + 1) __tmp = (fifo); \
>       |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   591 |         typeof(__tmp->ptr) __buf = (buf); \
>       |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   592 |         unsigned long __n = (n); \
>       |         ~~~~~~~~~~~~~~~~~~~~~~~~~~
>   593 |         const size_t __recsize = sizeof(*__tmp->rectype); \
>       |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   594 |         struct __kfifo *__kfifo = &__tmp->kfifo; \
>       |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   595 |         (__recsize) ?\
>       |         ~~~~~~~~~~~~~~
>   596 |         __kfifo_out_r(__kfifo, __buf, __n, __recsize) : \
>       |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   597 |         __kfifo_out(__kfifo, __buf, __n); \
>       |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   598 | }) \
>       | ~~~~
>   599 | )
>       | ~
> drivers/memstick/host/r592.c:367:9: note: in expansion of macro 'kfifo_out'
>   367 |         kfifo_out(&dev->pio_fifo, buffer, 4);
>       |         ^~~~~~~~~
>
> The value was never checked here, and the purpose of the function
> is only to flush the contents, so restore the old behavior but
> add a cast to void and a comment, which hopefully warns with neither
> gcc nor clang now.
>
> If anyone has an idea for how to fix it without ignoring the return
> code, that is probably better.

Should r592_flush_fifo_write be made to return an int, then callers of
r592_flush_fifo_write percolate up their return code?
r592_transfer_fifo_pio() seems to only return 0, but its callers are
doing return code checking.

>
> Fixes: 4b00ed3c5072 ("memstick: r592: remove unused variable")
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> ---
>  drivers/memstick/host/r592.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/memstick/host/r592.c b/drivers/memstick/host/r592.c
> index 026fadaa1d5d..615a83782e55 100644
> --- a/drivers/memstick/host/r592.c
> +++ b/drivers/memstick/host/r592.c
> @@ -359,12 +359,15 @@ static void r592_write_fifo_pio(struct r592_device *dev,
>  /* Flushes the temporary FIFO used to make aligned DWORD writes */
>  static void r592_flush_fifo_write(struct r592_device *dev)
>  {
> +       int ret;
>         u8 buffer[4] = { 0 };
>
>         if (kfifo_is_empty(&dev->pio_fifo))
>                 return;
>
> -       kfifo_out(&dev->pio_fifo, buffer, 4);
> +       ret = kfifo_out(&dev->pio_fifo, buffer, 4);
> +       /* intentionally ignore __must_check return code */
> +       (void)ret;
>         r592_write_reg_raw_be(dev, R592_FIFO_PIO, *(u32 *)buffer);
>  }
>
> --
> 2.29.2
>


-- 
Thanks,
~Nick Desaulniers