Re: [PATCH v7 2/5] io_uring/bpf-ops: implement loop_step with BPF struct_ops

[Pavel Begunkov <asml.silence@gmail.com>]

On 2/17/26 05:24, kernel test robot wrote:
> io_uring/bpf-ops.c:45 bpf_io_btf_struct_access() warn: always true condition '(off >= $expr_0x7fb415932650(30)) => (s32min-s32max >= 0)'

Got me confused at first, but it's the first field => offsetof() == 0,
and after type promotions it compares an unsigned type with 0.

I think smatch is too strict, and it's better the current way,
i.e. handling all fileds in the same way, but I'll respin to
silence smatch.

> 
> vim +45 io_uring/bpf-ops.c
> 
>      37	
>      38	static int bpf_io_btf_struct_access(struct bpf_verifier_log *log,
>      39					    const struct bpf_reg_state *reg, int off,
>      40					    int size)
>      41	{
>      42		const struct btf_type *t = btf_type_by_id(reg->btf, reg->btf_id);
>      43	
>      44		if (t == loop_params_type) {
>    > 45			if (off >= offsetof(struct iou_loop_params, cq_wait_idx) &&
>      46			    off + size <= offsetofend(struct iou_loop_params, cq_wait_idx))
>      47				return SCALAR_VALUE;
>      48		}
>      49	
>      50		return -EACCES;
>      51	}
>      52	
> 

-- 
Pavel Begunkov