Re: [PATCH] Add Broadcom Brahma-B15 CPU implementor and part number

On 05/05/2015 04:19 PM, Florian Fainelli wrote:
> On 05/05/15 12:17, William Cohen wrote:
>> On 05/05/2015 02:25 PM, Florian Fainelli wrote:
>>> Broadcom's Brahma-B15 CPU features a PMU which is 100% software
>>> compatible with those found on Cortex-A15 designs.  Add the required
>>> code to recognize the implementor id (0x42) and the Brahma-B15 part
>>> number (0x00f) and return "arm/armv7-ca15" so we can utilize the PMU.
>>>
>>> Signed-off-by: Florian Fainelli <[email protected]>
>>> ---
>>>  libop/op_cpu_type.c | 5 +++++
>>>  1 file changed, 5 insertions(+)
>>>
>>> diff --git a/libop/op_cpu_type.c b/libop/op_cpu_type.c
>>> index 362489c40989..3034c146f7bb 100644
>>> --- a/libop/op_cpu_type.c
>>> +++ b/libop/op_cpu_type.c
>>> @@ -415,6 +415,11 @@ static op_cpu _get_arm_cpu_type(void)
>>>  		case 0xd03:
>>>  			return op_get_cpu_number("arm/armv8-ca53");
>>>  		}
>>> +	} else if (vendorid == 0x42) {  /* Broadcom Corporation */
>>> +		switch (cpuid) {
>>> +		case 0x00f:
>>> +			return op_get_cpu_number("arm/armv7-ca15");
>>> +		}
>>>  	} else if (vendorid == 0x50) {	/* Applied Micro Circuits Corporation */
>>>  		switch (cpuid) {
>>>  		case 0x000:
>>>
>>
>>
>> Hi Florian,
>>
>> Thanks for the patch.  Do you have a URL for documentation that describes the Broadcom Brahma-B15 cpu?  It would be really nice to include that URL in the writeup so people can verify that the numbers are correct.  Are there any patches needed in the upstream kernel to allow the kernel to recognize the Broadcom Brahma-B15?
> 
> Unfortunately I cannot find public literature published by Broadcom
> which describes the CPU in greater details. If you want, I could
> reference some document names that Broadcom customers have access to,
> but they are not disclosed to the public, which is mildly useful.
> 
> Linux received its share of changes to support that processor a while
> ago, here are some links to our submission to do that:
> 
> http://markmail.org/message/m5uxatvbeighqyg7
> https://lwn.net/Articles/588384/
> http://comments.gmane.org/gmane.linux.ports.arm.kernel/398196
> 
> PMU in B15-based chips is discovered via Device Tree, where we use an
> armv7 compatible string.
> 

Hi Florian,

I came across the first URL patches and the oprofile patch looks consistent with that.

The patch looks fine and verifies that it builds on a tegra2-based trimslice.  Has it been tested on a Broadcom Brahma-B15 machine?

-Will

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y