12.01.2021 15:58, Dmitry Osipenko пишет:
Assert hardware reset before clocks are enabled and then de-assert it after clocks are enabled. This brings hardware into a predictable state and removes relying on implicit de-assertion of resets which is done by the clk driver.
Tested-by: Peter Geis pgwipeout@gmail.com Tested-by: Nicolas Chauvet kwizart@gmail.com Signed-off-by: Dmitry Osipenko digetx@gmail.com
sound/soc/tegra/tegra30_ahub.c | 33 ++++++++++++++++----------------- sound/soc/tegra/tegra30_ahub.h | 1 + 2 files changed, 17 insertions(+), 17 deletions(-)
diff --git a/sound/soc/tegra/tegra30_ahub.c b/sound/soc/tegra/tegra30_ahub.c index 4dbb58f7ea36..246cf6a373a1 100644 --- a/sound/soc/tegra/tegra30_ahub.c +++ b/sound/soc/tegra/tegra30_ahub.c @@ -65,10 +65,20 @@ static int tegra30_ahub_runtime_resume(struct device *dev) { int ret;
ret = reset_control_assert(ahub->reset);
if (ret)
return ret;ret = clk_bulk_prepare_enable(ahub->nclocks, ahub->clocks); if (ret) return ret;
ret = reset_control_reset(ahub->reset);
if (ret) {
clk_bulk_disable_unprepare(ahub->nclocks, ahub->clocks);return ret;}
regcache_cache_only(ahub->regmap_apbif, false); regcache_cache_only(ahub->regmap_ahub, false);
I just realized that this is incorrect version of the patch which misses the regcache syncing after the h/w reset. I'll make a v2.