[00/30] dmaengine: add COMPILE_TEST to drivers

Commit Message

Arnd Bergmann Sept. 16, 2016, 3:09 p.m. UTC

On Friday, September 16, 2016 7:50:35 PM CEST Vinod Koul wrote:
> On Fri, Sep 16, 2016 at 09:40:55AM +0200, Arnd Bergmann wrote:
> > On Thursday, September 15, 2016 11:03:39 PM CEST Vinod Koul wrote:
> > > We recently had a conversation about me cribbing that some drivers being
> > > hard to compile [1]. Arnd suggested [2] to add COMPILE_TEST to these so
> > > right now. With this I have one ARM config which compiles all drivers.
> > > 
> > > I used multi_v7 default config and checked the drivers which are not
> > > compiling and starting fixing them. So most of them were pretty fine, few
> > > got warning mostly on bad usage of print specifiers with dma_addr_t and
> > > size_t being major culprits there and few sparse warnings.
> > > 
> > > While compile testing most of the driver compiled fine on others (i have
> > > used ppc, arm, x86_64, i386, mips) expect the two where we explicitly depends
> > > on ARM, one is omap and second one is tegra. Jz one had issues but I moved
> > > header out of arm, so it worked fine.
> > > They have some API dependency, which we should resolve later..
> > > 
> > > Please test...
> > > 
> > 
> > Very nice series. If you have a git tree with these patches, I
> > can add the branch to my randconfig testing to see if I get
> > any build regression on 32-bit ARM.
> 
> I had pushed it my kernel.org tree where Fenguuang's bot found a numbe of
> issues, this series is present at:
> 
> git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/slave-dma.git test/compile-fixes
> 
> Please let me know if you find something.
> 


This is what I needed on top

I got this warning for the drivers that select DMA_OF, but it's CONFIG_OF
is always enabled for those platforms, and that makes DMA_OF enabled as well.
Build testing works fine without DMA_OF, so we don't need the 'select'

warning: (DMA_SUN4I && MOXART_DMA && STM32_DMA) selects DMA_OF which has unmet direct dependencies (DMADEVICES && OF)

Please fold this into your patches as appropriate.

	Arnd

Signed-off-by: Arnd Bergmann <[e-mail]>


--
To unsubscribe from this list: send the line "unsubscribe dmaengine" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index 00302dfcde95..713c13189926 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -151,7 +151,6 @@  config DMA_SUN4I
 	depends on MACH_SUN4I || MACH_SUN5I || MACH_SUN7I
 	default (MACH_SUN4I || MACH_SUN5I || MACH_SUN7I)
 	select DMA_ENGINE
-	select DMA_OF
 	select DMA_VIRTUAL_CHANNELS
 	help
 	  Enable support for the DMA controller present in the sun4i,
@@ -307,7 +306,7 @@  config MMP_TDMA
 	bool "MMP Two-Channel DMA support" if COMPILE_TEST && !ARCH_MMP
 	default ARCH_MMP
 	select DMA_ENGINE
-	select MMP_SRAM
+	select MMP_SRAM if ARCH_MMP
 	help
 	  Support the MMP Two-Channel DMA engine.
 	  This engine used for MMP Audio DMA and pxa910 SQU.
@@ -317,7 +316,6 @@  config MOXART_DMA
 	tristate "MOXART DMA support"
 	depends on ARCH_MOXART
 	select DMA_ENGINE
-	select DMA_OF
 	select DMA_VIRTUAL_CHANNELS
 	help
 	  Enable support for the MOXA ART SoC DMA controller.
@@ -442,7 +440,6 @@  config STM32_DMA
 	bool "STMicroelectronics STM32 DMA support" if COMPILE_TEST && !ARCH_STM32
 	default ARCH_STM32
 	select DMA_ENGINE
-	select DMA_OF
 	select DMA_VIRTUAL_CHANNELS
 	help
 	  Enable support for the on-chip DMA controller on STMicroelectronics
diff --git a/include/linux/platform_data/dma-mmp_tdma.h b/include/linux/platform_data/dma-mmp_tdma.h
index 0c72886030ef..422d4504dbac 100644
--- a/include/linux/platform_data/dma-mmp_tdma.h
+++ b/include/linux/platform_data/dma-mmp_tdma.h
@@ -28,7 +28,7 @@  struct sram_platdata {
 	int granularity;
 };
 
-#ifdef CONFIG_ARM
+#ifdef CONFIG_MMP_SRAM
 extern struct gen_pool *sram_get_gpool(char *pool_name);
 #else
 static inline struct gen_pool *sram_get_gpool(char *pool_name)