[lvc-project] [PATCH v3] clk: mvebu: Prevent division by zero in clk_double_div_recalc_rate()

[Fedor Pchelkin]

On Wed, 09. Oct 14:23, Andrew Lunn wrote:
> > > I would say, let is divide by 0, so there is an obvious kernel stack
> > > trace and hopefully a report of the issue. It can then be investigated
> > > in a way we can then find out what the hardware actually is doing.
> > 
> > Is it worth adding some kind of WARN assertions? Or actually just leave it
> > for now as is?
> 
> What actually happens on a / 0 on ARM? I assume it triggers an
> exception, which will give a stack trace? If so a WARN adds no value.

Oh, I see. I should have better said "adding WARN assertions and bailing
out with a default value if they are violated". Thus avoiding to have a
division by zero exception. Non panic_on_warn systems would at least
survive in this case but still have a valuable trace.

Somehow more importantly, it would state in the codebase that the condition
is very-very unexpected and most probably won't ever happen but not 100%
sure because it depends on hardware behavior (as I perceive reading the
current thread).

That said, if adding such WARN-bail-out pattern seems unnecessary and just
wasteful in this situation, I don't think we have any options other than
keeping the code as is.