Re: [Xen-devel] [PATCH v3 07/13] xen/arm: compile and initialize vmap

[Ian Campbell <Ian.Campbell@xxxxxxxxxx>]

On Fri, 2013-04-26 at 13:02 +0100, Stefano Stabellini wrote:
> On Fri, 26 Apr 2013, Ian Campbell wrote:
> > > > Isn't it a bug for the third to be already mapped? that suggests
> > > > something is overwriting the mapping, does vmap do that?
> > > 
> > > I don't know, but I thought that this function should be able to handle
> > > that case.
> > 
> > At the vmap layer I think it must always be a bug for it to be trying to
> > replace a mapping, the API simply doesn't allow for the possibility.
> 
> Empirical tests give evidence to the contrary: vmap does replace
> mappings.

Empirical test == map, unmap, map, where the second map uses the same
address space as the first?

vunmap is:
        void vunmap(const void *va)
        {
            unsigned long addr = (unsigned long)va;
        
            destroy_xen_mappings(addr, addr + PAGE_SIZE * vm_size(va));
            vm_free(va);
        }
are you sure our destroy_xen_mappings isn't buggy?

Ian.


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel