Re: [PATCH] misc: sram: extend usage of reserved partitions

[Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>]

Hello Greg,

On 10.08.2015 13:47, Vladimir Zapolskiy wrote:
> Hi Greg,
> 
> On 10.08.2015 06:27, Greg Kroah-Hartman wrote:
>> On Mon, Aug 10, 2015 at 12:40:02AM +0300, Vladimir Zapolskiy wrote:
>>> This change adds functionality to operate on reserved SRAM partitions
>>> described in device tree file. Two partition properties are added,
>>> "pool" and "export", the first one allows to share a specific partition
>>> for usage by a kernel consumer in the same manner as it is done for
>>> the whole SRAM device, and "export" property provides access to some
>>> SRAM area from userspace over sysfs interface. Practically it is
>>> possible to specify both properties for an SRAM partition, however
>>> simultaneous access from a kernel consumer and from userspace is not
>>> serialized, but still the combination may be useful for debugging
>>> purpose.
>>>
>>> Signed-off-by: Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>
>>> ---
>>> To test SRAM partition export to userspace it might be sufficient
>>> to add an exported subnode, write data to created binary attribute file,
>>> do soft reboot and read the contents of the same binary attribute
>>> file. Aware of any potential use of SRAM by a bootloader etc.
>>>
>>> To give an idea how to use "pool" partition feature here is an example
>>> for Freescale iMX6Q SabreAuto with CODA driver as a user:
>>>
>>> arch/arm/boot/dts/imx6q.dtsi @@ -82,6 +82,10 @@
>>>  			compatible = "mmio-sram";
>>>  			reg = <0x00900000 0x40000>;
>>>  			clocks = <&clks IMX6QDL_CLK_OCRAM>;
>>> +
>>> +			#address-cells = <1>;
>>> +			#size-cells = <1>;
>>> +			ranges = <0 0x00900000 0x40000>;
>>>  		};
>>>  
>>>  		aips-bus@02000000 { /* AIPS1 */
>>> arch/arm/boot/dts/imx6qdl-sabreauto.dtsi @@ -490,3 +490,18 @@
>>>  				0x0000c000 0x1404a38e 0x00000000>;
>>>  	};
>>>  };
>>> +
>>> +&ocram {
>>> +	#address-cells = <1>;
>>> +	#size-cells = <1>;
>>> +
>>> +	vpu_iram: vpu_iram {
>>> +		reg = <0x00008000 0x00021000>;
>>> +		pool;
>>> +	};
>>> +};
>>> +
>>> +&vpu {
>>> +	iram = <&vpu_iram>;
>>> +};
>>>
>>
>> Putting a diff in a patch changelog is tricky, do you really need it?
> 
> you are right, it is tricky, that's why this is actually not a valid
> diff, I tested on my environment and the email can be applied with git-am.

on the second glance this quasi-diff is located outside of the commit
message, it is in the section under --- line and before the actual
change. git-am digests it.

>> And why send 2 copies of this patch, what changed?
>>
>> confused,
>>
> 
> sorry for confusion, occasionally I gave two same files to
> git-send-email. The versions are absolutely identical, please ignore one
> of them.

Please let me know, if you like to see any updates to the change. The
updated genalloc interfaces used by the change are found in v4.3.0-rc1.

--
With best wishes,
Vladimir