Re: [PATCH v2 05/21] ARM: tegra: clock: Disable clocks left on by bootloader

From: Olof Johansson
Date: Sun Feb 20 2011 - 23:03:27 EST


On Sun, Feb 20, 2011 at 7:43 PM, Colin Cross <ccross@xxxxxxxxxxx> wrote:
> On Sun, Feb 20, 2011 at 4:40 PM, Olof Johansson <olof@xxxxxxxxx> wrote:
>> Hi,
>>
>> On Sat, Feb 19, 2011 at 2:25 PM, Colin Cross <ccross@xxxxxxxxxxx> wrote:
>>> Adds CONFIG_TEGRA_DISABLE_BOOTLOADER_CLOCKS that iterates
>>> through all clocks, disabling any for which the refcount
>>> is 0 but the clock init detected the bootloader left the
>>> clock on.
>>
>>
>> I would argue that any kernel functionality that relies on clocks
>> being left on my bootloader is buggy, and/or should at least have
>> those clocks enabled in the static clock tables for the boards in
>> question, since it creates an undocumented dependency on firmware.
>>
>> I would prefer if it disabled the clocks by default (after warning),
>> but that there was a runtime way to override while debugging (i.e. add
>> a bootarg 'keep_fwclocks' or similar to keep the firmware-enabled
>> clocks running and just warn about them).
>
> That was the plan, but since many of the drivers are missing I didn't
> want to enable it by default yet.  However, it turns out the problems
> I was seeing at boot from this patch were actually due to a bug in the
> clock code causing the cpu clock to be disabled for clocks that have
> no disable.  I will drop this patch, and post a separate series after
> these that fixes the bug and turns off clocks by default, with a
> runtime argument like you suggested.


Sounds good. Alternatively, this can go in and be revisited with the
above proposed changes; both approaches would be OK with me.


-Olof
--
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/