On Tue, 11 May 2021, Christoph Hellwig wrote:
> On Tue, May 11, 2021 at 09:47:33AM -0700, Stefano Stabellini wrote:
> > That's a much better plan. It is also not super urgent, so maybe for now
> > we could add an explicit check for io_tlb_default_mem != NULL at the
> > beginning of xen_swiotlb_init? So that at least we can fail explicitly
> > or ignore it explicitly rather than by accident.
>
> Fine with me. Do you want to take over from here and test and submit
> your version?
I can do that. Can I add your signed-off-by for your original fix?
On Tue, May 11, 2021 at 09:51:20AM -0700, Stefano Stabellini wrote:
> On Tue, 11 May 2021, Christoph Hellwig wrote:
> > On Tue, May 11, 2021 at 09:47:33AM -0700, Stefano Stabellini wrote:
> > > That's a much better plan. It is also not super urgent, so maybe for now
> > > we could add an explicit check for io_tlb_default_mem != NULL at the
> > > beginning of xen_swiotlb_init? So that at least we can fail explicitly
> > > or ignore it explicitly rather than by accident.
> >
> > Fine with me. Do you want to take over from here and test and submit
> > your version?
>
> I can do that. Can I add your signed-off-by for your original fix?
Sure:
Signed-off-by: Christoph Hellwig <[email protected]>