From mboxrd@z Thu Jan 1 00:00:00 1970 From: arnd@arndb.de (Arnd Bergmann) Date: Wed, 25 Nov 2015 22:57:33 +0100 Subject: [PATCH 1/7] clk: mmp: stop using platform headers In-Reply-To: <20151125213222.GF11298@codeaurora.org> References: <1448465875-435039-1-git-send-email-arnd@arndb.de> <3508030.Q62gfHvXvx@wuerfel> <20151125213222.GF11298@codeaurora.org> Message-ID: <3842139.m8b6S96gUd@wuerfel> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wednesday 25 November 2015 13:32:22 Stephen Boyd wrote: > On 11/25, Arnd Bergmann wrote: > > On Wednesday 25 November 2015 11:21:14 Stephen Boyd wrote: > > > On 11/25, Arnd Bergmann wrote: > > > > diff --git a/drivers/clk/mmp/clk-mmp2.c b/drivers/clk/mmp/clk-mmp2.c > > > > index 09d2832fbd78..38931dbd1eff 100644 > > > > --- a/drivers/clk/mmp/clk-mmp2.c > > > > +++ b/drivers/clk/mmp/clk-mmp2.c > > > > @@ -9,6 +9,7 @@ > > > > * warranty of any kind, whether express or implied. > > > > */ > > > > > > > > +#include > > > > > > Why are we adding this include? I don't see any clk consumer API > > > usage being added. > > > > drivers/clk/mmp/clk-pxa910.c: In function 'pxa910_clk_init': > > drivers/clk/mmp/clk-pxa910.c:166:2: error: implicit declaration of function 'clk_set_rate' [-Werror=implicit-function-declaration] > > clk_set_rate(uart_pll, 14745600); > > ^ > > drivers/clk/mmp/clk-pxa910.c:209:2: error: implicit declaration of function 'clk_set_parent' [-Werror=implicit-function-declaration] > > clk_set_parent(clk, uart_pll); > > ^ > > > > Is there anything I need to change here? > > > > Hmph. I seems this broke when we removed the clk.h include from > clk-provider.h, but nobody has noticed for months. Can you make a > different commit to add the header? Otherwise we're sneaking in > that include to get this file to compile again. Ok, I hadn't noticed at all that it was broken in mainline already, as the three mmp defconfigs do not use drivers/clk/ at all but instead rely on arch/arm/mach-mmp/clock*.c instead, which has a separate implementation. With the move to multiplatform, the choice goes away and we use the common clock implementation unconditionally. Arnd