[LITMUS^RT] Running LITMUS-RT on ARM64

Andrii Anisov andrii_anisov at epam.com
Mon Sep 18 15:42:25 CEST 2017


Hello Björn,


Sorry for a delayed answer. I was on vacations last week.


On 15.09.17 01:07, Björn Brandenburg wrote:
> Hi Andrii,
>
> a quick heads-up and a question...
>
> 1) I had some prior patches pending (which I forgot about) that already used up the -A option, so I had to resolve the option clash. The calibration loop is now triggered with -a0.
I have no specific preferences on options naming. All, more or less 
suitable, letters were already used, so I just took first available in 
alphabetical order.

> 2) Do you have some guidance for how to use your new calibration and workload code?
There is a kind of guidance in a commit message [1]. Also I do a slight 
manual adjustment of number of cycles for minimal error during 
calibration. E.g. if `rtspin -A`, in different runs, gives me values 
between 193000 and 195000, I take 195000 for verification and see 
calibration error less than 0.1%.

> When I ran the calibration loop and then plugged the obtained value for 1ms into -a, I saw calibration errors in excess of 50% when using -l.  Is this expected?
No, this is not as expected.
Could you please check if you have loop() function not inlined in your 
rtspin binary. I.e. with `${CROSS_COMPILE}objdump -t rtspin | grep loop`.

[1] 
https://github.com/LITMUS-RT/liblitmus/commit/ce8aeb80262d60cdfe762931f6c71e2b5ab22ef7

-- 

*Andrii Anisov*





More information about the litmus-dev mailing list