[PATCH] arm64: defconfig: enable Generic on-chip SRAM driver

[PATCH] arm64: defconfig: enable Generic on-chip SRAM driver

[Sudeep Holla]

Most of the ARM SoCs have small SRAM memory used for various purposes
that require faster access to data compared to normal DDR.

This patch enables the Generic on-chip SRAM driver on ARM64.

Cc: Kevin Hilman <khilman@kernel.org>
Cc: Olof Johansson <olof@lixom.net>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
---
 arch/arm64/configs/defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index d9e914aec081..bf0b0d258246 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -101,6 +101,7 @@ CONFIG_MTD_M25P80=y
 CONFIG_MTD_SPI_NOR=y
 CONFIG_BLK_DEV_LOOP=y
 CONFIG_VIRTIO_BLK=y
+CONFIG_SRAM=y
 # CONFIG_SCSI_PROC_FS is not set
 CONFIG_BLK_DEV_SD=y
 # CONFIG_SCSI_LOWLEVEL is not set
--
2.7.4

[PATCH] arm64: defconfig: enable Generic on-chip SRAM driver

[Jon Medhurst (Tixy)]

On Mon, 2016-06-20 at 16:41 +0100, Sudeep Holla wrote:
> Most of the ARM SoCs have small SRAM memory used for various purposes
> that require faster access to data compared to normal DDR.
> 
> This patch enables the Generic on-chip SRAM driver on ARM64.

Would this cause things in the kernel to automatically start using SRAM
found in device-tree, or would each use need to be added explicitly?
I'm just wondering if we might start seeing new bugs caused by firmware
using SRAM for special purposes but that memory not being properly
reserved or hidden from the kernel.

-- 
Tixy

[PATCH] arm64: defconfig: enable Generic on-chip SRAM driver

[Sudeep Holla]

On 20/06/16 17:34, Jon Medhurst (Tixy) wrote:
> On Mon, 2016-06-20 at 16:41 +0100, Sudeep Holla wrote:
>> Most of the ARM SoCs have small SRAM memory used for various purposes
>> that require faster access to data compared to normal DDR.
>>
>> This patch enables the Generic on-chip SRAM driver on ARM64.
>
> Would this cause things in the kernel to automatically start using SRAM
> found in device-tree, or would each use need to be added explicitly?
> I'm just wondering if we might start seeing new bugs caused by firmware
> using SRAM for special purposes but that memory not being properly
> reserved or hidden from the kernel.
>

No only the unreserved part of SRAM will be used by generic allocator
and the users/drivers needing this have to explicitly request for the
same(my understanding based on some usage I have seen on ARM32
platforms). The reserved part is not touched by the generic allocator.

-- 
Regards,
Sudeep