Re: [PATCH] i2c: amd_mp2: handle num is 0 input for i2c_amd_xfer

From: Wolfram Sang
Date: Mon Sep 07 2020 - 10:34:21 EST


On Fri, Sep 04, 2020 at 11:06:47AM -0700, trix@xxxxxxxxxx wrote:
> From: Tom Rix <trix@xxxxxxxxxx>
>
> clang static analyzer reports this problem
>
> i2c-amd-mp2-plat.c:174:9: warning: Branch condition evaluates
> to a garbage value
> return err ? err : num;
> ^~~
>
> err is not initialized, it depends on the being set in the
> transfer loop which will not happen if num is 0. Surveying
> other master_xfer() implementations show all handle a 0 num.
>
> Because returning 0 is expected, initialize err to 0.

Well, it is not expected. The core does:

2019 if (WARN_ON(!msgs || num < 1))
2020 return -EINVAL;

Dunno if we should apply the patch nonetheless or add a comment that num
is guaranteed to be at least 1.

>
> Signed-off-by: Tom Rix <trix@xxxxxxxxxx>
> ---
> drivers/i2c/busses/i2c-amd-mp2-plat.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/i2c/busses/i2c-amd-mp2-plat.c b/drivers/i2c/busses/i2c-amd-mp2-plat.c
> index 17df9e8845b6..506433bc0ff2 100644
> --- a/drivers/i2c/busses/i2c-amd-mp2-plat.c
> +++ b/drivers/i2c/busses/i2c-amd-mp2-plat.c
> @@ -155,7 +155,7 @@ static int i2c_amd_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
> struct amd_i2c_dev *i2c_dev = i2c_get_adapdata(adap);
> int i;
> struct i2c_msg *pmsg;
> - int err;
> + int err = 0;
>
> /* the adapter might have been deleted while waiting for the bus lock */
> if (unlikely(!i2c_dev->common.mp2_dev))
> --
> 2.18.1
>

Attachment: signature.asc
Description: PGP signature