Bug in ALL function: Applying the wrong filters on other columns when removing filter on one column!

I have here a sample table:

I've created a measure as follows:

SumQuantity = CALCULATE( SUM(Data[Quantity]), ALL(Data[Color]) )

I've created two slicers for Brand and Order and a visual table with these columns: Brand, Color, SumQuantity. With this table, I expect to see the same SumQuantity of different colors products with the same Brand in all circumstances. But when I choose Brand Dell & HP and Order 1 & 2 on the slicers, the result is shown as follows:

Notice the SumQuantity of HP-Blue products is 13 while HP-Black/HP-Red is 20; the SumQuantity of Dell-Black is 6 while Dell-Blue is 18. 

When I filters Brand HP and Order 1 & 2 in Excel, the sum quantity is 20 so the SumQuantity of HP-Black/HP-Red rows are as expected while HP-Blue is not. I've tried to replicate the result of HP-Blue and found out Power BI is working like so:

- Applying all filters and contexts (Brand HP and Color Blue)

- With that, it finds that only Order 2 has HP-Blue products. Then it removes filter on column Color with filters Order 2 and Brand HP applied, and calculates the result SumQuantity is 13.

When I use the same measure replaced ALL funtion with ALLSELECTED, the result is as expected:

SumQuantity_AllSelected = CALCULATE(SUM(Data[Quantity]), ALLSELECTED(Data[Color]))

So I guess the behavior of the ALL function is not intended.

I hope Power BI team can confirm if it's a bug and fix it if it is. Thank you.