Re: [PATCH] clk: vc5: Abort clock configuration without upstream clock

[Stephen Boyd <sboyd@kernel.org>]

Quoting Marek Vasut (2018-12-16 09:14:29)
> On 12/16/2018 08:19 AM, Laurent Pinchart wrote:
> > Hi Marek,
> > 
> > Thank you for the patch.
> > 
> > On Saturday, 15 December 2018 02:55:19 EET Marek Vasut wrote:
> >> In case the upstream clock are not set, which can happen in case the
> >> VC5 has no valid upstream clock, the $src variable is used uninited
> >> by regmap_update_bits(). Check for this condition and return -EINVAL
> >> in such case.
> > 
> > Note that the probe() function will fail in this case, so vc5_mux_set_parent() 
> > won't be reached.
> > 
> >> Note that in case the VC5 has no valid upstream clock, the VC5 can
> >> not operate correctly. That is a hardware property of the VC5. The
> >> internal oscilator present in some VC5 models is also considered
> >> upstream clock.
> >>
> >> Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
> >> Cc: Alexey Firago <alexey_firago@mentor.com>
> >> Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> >> Cc: Stephen Boyd <sboyd@kernel.org>
> >> Cc: linux-renesas-soc@vger.kernel.org
> >> ---
> >> NOTE: This is an updated version of:
> >>       https://patchwork.kernel.org/patch/10731699/
> >> ---
> >>  drivers/clk/clk-versaclock5.c | 4 +++-
> >>  1 file changed, 3 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/clk/clk-versaclock5.c b/drivers/clk/clk-versaclock5.c
> >> index 5b393e711e94..b10801506518 100644
> >> --- a/drivers/clk/clk-versaclock5.c
> >> +++ b/drivers/clk/clk-versaclock5.c
> >> @@ -262,8 +262,10 @@ static int vc5_mux_set_parent(struct clk_hw *hw, u8
> >> index)
> >>
> >>              if (vc5->clk_mux_ins == VC5_MUX_IN_XIN)
> >>                      src = VC5_PRIM_SRC_SHDN_EN_XTAL;
> >> -            if (vc5->clk_mux_ins == VC5_MUX_IN_CLKIN)
> >> +            else if (vc5->clk_mux_ins == VC5_MUX_IN_CLKIN)
> >>                      src = VC5_PRIM_SRC_SHDN_EN_CLKIN;
> >> +            else
> >> +                    return -EINVAL;
> >>      }
> > 
> > I'd rather go for Stephen's approach if the goal is just to silence a warning 
> > for a condition that can't happen in practice.
> 
> Sure, probe will fail, but it's safer to handle the possibility that
> probe() is broken and this code is reached by properly handling the
> failure instead of doing something obviously wrong (like configuring the
> hardware with value 0).

I'm fine with this approach. Laurent?