[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160314121823.GE22912@ulmo.nvidia.com>
Date: Mon, 14 Mar 2016 13:18:24 +0100
From: Thierry Reding <treding@...dia.com>
To: Andy Yan <andy.yan@...nel-upstream.org>
CC: <arnd@...db.de>, <javier@....samsung.com>, <kgene@...nel.org>,
<k.kozlowski@...sung.com>, <geert+renesas@...der.be>,
<linux-kernel@...r.kernel.org>, <public_timo.s@...entcreek.de>,
<linux@....linux.org.uk>, <alexandre.belloni@...e-electrons.com>,
<sjoerd.simons@...labora.co.uk>, <olof@...om.net>,
<linux-arm-kernel@...ts.infradead.org>
Subject: Re: [PATCH 1/2] ARM: multi_v7_defconfig: enable CONFIG_PWM_ROCKCHIP=y
On Sat, Mar 12, 2016 at 08:50:29PM +0800, Andy Yan wrote:
> All the current rk3066 based boards use pwm modulate vdd_logic
> voltage, the board may hang when the system boot and run to a higher
> frequency without pwm enabled, so enable it as builtin.
I don't understand the rationale here. Surely if the system can't safely
switch to a higher frequency because the regulator isn't available, then
it should be prevented from switching, shouldn't it?
Making this change alone will not prevent users from switching this back
to =m on their own. So either you need to add code to prevent a switch
to an unsupported frequency in the absence of the regulator, or you need
to change the PWM_ROCKCHIP Kconfig option to be bool instead, though the
latter is of course a somewhat ugly hack, whereas the former would be a
proper fix.
Thierry
Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)
Powered by blists - more mailing lists