[Libvir] RFC PATCH - Initial NodeGetCellsFreeMemory patch

[beth kon]

Richard W.M. Jones wrote:

>
> +    if (startCell > nbCells - 1)
> +        return -1;
>
> Surely, the condition should be `startCell + maxCells > nbCells'?
>
Actually I will discuss this now since I plan to leave it as is.

First I'll make sure terms are clear. startCell is the first cell to 
gather freeMems info on. maxCells is the maximum number of entries 
available in freeMems (otherwise the buffer will be overrun). nbCells is 
the total number of cells on the node.

I assume the caller could legally specify a maxCells value that is 
greater than nbCells. It just means they provided a bigger buffer than 
necessary. My test here is just verifying that the startCell provided is 
not out of range of the number of cells that exist on the node. freeMems 
entries will be returned for up to maxCells cells, starting from 
startCell, ending at either nbCells, or maxCells number of entries, 
whichever comes first. Does that make sense?

> Rich.
>
>


-- 
Elizabeth Kon (Beth)
IBM Linux Technology Center
Open Hypervisor Team
email: eak at us.ibm.com