diff options
author | Tony Lindgren <tony@atomide.com> | 2014-04-10 16:47:18 -0700 |
---|---|---|
committer | Lee Jones <lee.jones@linaro.org> | 2014-04-23 15:31:05 +0100 |
commit | a613b739b8c08eab811e677810045cc0522fc3e6 (patch) | |
tree | c3d8142c39c53333af788264ef38019fa304334d /drivers/mfd/twl4030-power.c | |
parent | 455c6fdbd219161bd09b1165f11699d6d73de11c (diff) | |
download | op-kernel-dev-a613b739b8c08eab811e677810045cc0522fc3e6.zip op-kernel-dev-a613b739b8c08eab811e677810045cc0522fc3e6.tar.gz |
mfd: twl-core: Fix idle mode signaling for omaps when booted with device tree
I noticed a regression where the omap sys_clkreq signal will never
trigger for omap3 when booted with device tree while it triggers
when booted in legacy mode. This means voltage scaling does not
do anything when booted with device tree.
Turns out the reason is we fail to initialize the SmartReflex
enable bit in twl4030 with the following error:
twl: not initialized
And that happens because we are wrongly tinkering with the twl4030
registers in arch/arm/mach-omap2/omap_twl.c before the driver is
initialized. Looking at the the SmartReflex bit enable code in
omap_twl.c, we need to always set it.
So let's fix the issue by always enabling the twl4030 SmartReflex
bit in the drivers/mfd/twl-core.c probe, and drop the related
code in omap_twl.c.
Note that we still have some twl4030 tinkering left in omap_twl.c
for the twl6030 case, but that's a different patch.
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Diffstat (limited to 'drivers/mfd/twl4030-power.c')
0 files changed, 0 insertions, 0 deletions