Re: [PATCH v2] clk: exynos4: Add CLK_GET_RATE_NOCACHE flag for theExynos4x12 ISP clocks

[Sylwester Nawrocki]

On 07/26/2013 10:30 PM, Mike Turquette wrote:
> Quoting Sylwester Nawrocki (2013-07-25 14:07:05)
> > From: Sylwester Nawrocki<s.nawrocki@xxxxxxxxxxx>
> >
> > The ISP clock registers belong to the ISP power domain and may change
> > their values if this power domain is switched off/on. Add
> > CLK_GET_RATE_NOCACHE flags to ensure we do not rely on invalid cached
> > data when setting or getting frequency of those clocks.
> >
> > Signed-off-by: Sylwester Nawrocki<s.nawrocki@xxxxxxxxxxx>
> > Signed-off-by: Kyungmin Park<kyungmin.park@xxxxxxxxxxx>
>
> Thanks for the fix. I've taken into clk-fixes. Is there a specific
> regression this fixes (besides just having wrong clock rates). I can
> amend the changelog if you have an example (e.g. device X explodes).

Hmm, yes, this fixes a pretty serious problem. There is a companion
commit [1] already in Linus' tree, more details can be found there.

Perhaps something like this could be added:

"Otherwise the FIMC-IS Cortex-A5 core and AXI bus clocks have
incorrect frequencies, which breaks the ISP operation and starting
the video pipeline fails with timeouts reported by the FIMC-IS
firmware.

See related commit 722a860ecb29aa34ec6f7d7f32b949209e8
"[media] exynos4-is: Fix FIMC-IS clocks initialization" for more
details."

[1] 
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=722a860ecb29aa34ec6f7d7f32b949209e86a2f3

Thanks,
Sylwester
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/