On Wed, 2015-06-03 at 09:37 -0500, Suravee Suthikulanit wrote:

On 5/28/2015 9:38 PM, Mark Salter wrote:

quoted

On Wed, 2015-05-20 at 17:09 -0500, Suravee Suthikulpanit wrote:

quoted

quoted

Fromhttp://www.uefi.org/sites/default/files/resources/ACPI_6.0.pdf,
section 6.2.17 _CCA states that ARM platforms require ACPI _CCA
object to be specified for DMA-cabpable devices. Therefore, this patch
specifies ACPI_CCA_REQUIRED in arm64 Kconfig.

In addition, to handle the case when _CCA is missing, arm64 would assign
dummy_dma_ops to disable DMA capability of the device.

Acked-by: Catalin Marinas<catalin.marinas@arm.com>
Signed-off-by: Mark Salter<redacted>
Signed-off-by: Suravee Suthikulpanit<Suravee.Suthikulpanit@amd.com>
---
 arch/arm64/Kconfig                   |  1 +
 arch/arm64/include/asm/dma-mapping.h | 18 ++++++-
 arch/arm64/mm/dma-mapping.c          | 92 ++++++++++++++++++++++++++++++++++++
 3 files changed, 109 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 4269dba..95307b4 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig

@@ -1,5 +1,6 @@
 config ARM64
 	def_bool y
+	select ACPI_CCA_REQUIRED if ACPI
 	select ACPI_GENERIC_GSI if ACPI
 	select ACPI_REDUCED_HARDWARE_ONLY if ACPI
 	select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE

diff --git a/arch/arm64/include/asm/dma-mapping.h b/arch/arm64/include/asm/dma-mapping.h
index 9437e3d..f0d6d0b 100644
--- a/arch/arm64/include/asm/dma-mapping.h
+++ b/arch/arm64/include/asm/dma-mapping.h

@@ -18,6 +18,7 @@

 #ifdef __KERNEL__

+#include <linux/acpi.h>
 #include <linux/types.h>
 #include <linux/vmalloc.h>

^^^ This hunk causes build issues with a couple of drivers:

drivers/scsi/megaraid/megaraid_sas_fp.c:69:0: warning: "FALSE" redefined [enabled by default]
  #define FALSE 0
  ^
In file included from include/acpi/acpi.h:58:0,
                  from include/linux/acpi.h:37,
                  from ./arch/arm64/include/asm/dma-mapping.h:21,
                  from include/linux/dma-mapping.h:86,
                  from ./arch/arm64/include/asm/pci.h:7,
                  from include/linux/pci.h:1460,
                  from drivers/scsi/megaraid/megaraid_sas_fp.c:37:
include/acpi/actypes.h:433:0: note: this is the location of the previous definition
  #define FALSE                           (1 == 0)
  ^


In file included from include/acpi/acpi.h:58:0,
                  from include/linux/acpi.h:37,
                  from ./arch/arm64/include/asm/dma-mapping.h:21,
                  from include/linux/dma-mapping.h:86,
                  from include/scsi/scsi_cmnd.h:4,
                  from drivers/scsi/ufs/ufshcd.h:60,
                  from drivers/scsi/ufs/ufshcd.c:43:
include/acpi/actypes.h:433:41: error: expected identifier before ‘(’ token
  #define FALSE                           (1 == 0)
                                          ^
drivers/scsi/ufs/unipro.h:203:2: note: in expansion of macro ‘FALSE’
   FALSE = 0,
   ^

This happens because the ACPI definitions of TRUE and FALSE conflict
with local definitions in megaraid and enum declaration in ufs.

Mark,

Thanks for pointing this out. Although, I would think that the 
megaraid_sas_fp.c should have had the #ifndef to check before defining 
the TRUE and FALSE as following.

#ifndef TRUE
#define TRUE 1
#endif
#ifndef FALSE
#define FALSE 0
#endif

This seems to be what other drivers are also doing. If this is okay, I 
can send out a fix-up patch for the megaraid driver.

Yeah, or #undef them if defined so megaraid defines them as desired.
And #undef if defined would work for unipro.h as well.