dpcm_init_runtime_hw() (= A) is used from dpcm_set_fe_runtime() (= B)
with for_each_rtd_cpu_dais() loop (= C), and it checks formats (= D).

(A)	static void dpcm_init_runtime_hw(...)
	{
		...
 ^		if (runtime->hw.formats)
 | (D1)			runtime->hw.formats &= stream->formats;
(D)		else
 | (D2)			runtime->hw.formats = stream->formats;
 v	}

(B)	static void dpcm_set_fe_runtime(...)
	{
		...
(C)		for_each_rtd_cpu_dais(rtd, i, cpu_dai) {
			...
(A)			dpcm_init_runtime_hw(...);
		}
	}

If this for_each_rtd_cpu_dais() loop (= C) calls
dpcm_init_runtime_hw() (= A) multiple times, this means it is Multi-CPU.

If we focus to format operation at (D), using mask (= D1) is understandable
because it restricts unsupported format.
But, enabling format when zero format case (= D2) is very strange,
because it might enables unsupported format.

This runtime->hw.formats is initialized by ULLONG_MAX at soc_pcm_hw_init(),
thus becoming zero format means it can't use such format.
And doing this strange format operation is only here.

This patch removes strange format operation (= D2), and use standard
soc_pcm_hw_update_format() for it.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87sg5pvshq.wl-kuninori.morimoto.gx@renesas.comSigned-off-by: Mark Brown <broonie@kernel.org>