17.12.2020 21:06, Dmitry Osipenko пишет: > The GR3D1 hardware unit needs to pulse hardware reset after removing power > clamp, otherwise reset won't be deasserted. Hence give reset a pulse after > removing the clamp. This stayed unnoticed previously because power > management wasn't supported by the 3D driver until recently and all power > gates are usually ungated after bootloader by default. > > Signed-off-by: Dmitry Osipenko <digetx@xxxxxxxxx> > --- > drivers/soc/tegra/pmc.c | 8 ++++++++ > 1 file changed, 8 insertions(+) > > diff --git a/drivers/soc/tegra/pmc.c b/drivers/soc/tegra/pmc.c > index fd2ba3c59178..985373ce52b1 100644 > --- a/drivers/soc/tegra/pmc.c > +++ b/drivers/soc/tegra/pmc.c > @@ -654,6 +654,14 @@ static int tegra_powergate_power_up(struct tegra_powergate *pg, > > usleep_range(10, 20); > > + /* > + * Some hardware blocks may need a 0->1->0 reset pulse in order > + * to propagate the reset, Tegra30 3D1 is one example. > + */ > + err = reset_control_reset(pg->reset); > + if (err) > + goto powergate_off; > + > if (pg->pmc->soc->needs_mbist_war) > err = tegra210_clk_handle_mbist_war(pg->id); > if (err) > After some more testing, I found that 3D1 is reset properly if its memory client is placed in reset before power domain is turned off. The memory client resetting is completely missed in v2. I'll improve it in v3 and then this PMC workaround-patch won't be needed anymore.
