Re: [PATCH 12/14] swiotlb: move global variables into a new io_tlb_mem structure

[Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>]

On Wed, Mar 17, 2021 at 06:57:42PM +0100, Christoph Hellwig wrote:
> On Wed, Mar 17, 2021 at 01:51:56PM -0400, Konrad Rzeszutek Wilk wrote:
> > On Wed, Mar 17, 2021 at 02:53:27PM +0100, Christoph Hellwig wrote:
> > > On Wed, Mar 17, 2021 at 01:42:07PM +0000, Konrad Rzeszutek Wilk wrote:
> > > > > -     alloc_size = PAGE_ALIGN(io_tlb_nslabs * sizeof(size_t));
> > > > > -     io_tlb_alloc_size = memblock_alloc(alloc_size, PAGE_SIZE);
> > > > > -     if (!io_tlb_alloc_size)
> > > > > -             panic("%s: Failed to allocate %zu bytes align=0x%lx\n",
> > > > > -                   __func__, alloc_size, PAGE_SIZE);
> > > > 
> > > > Shouldn't this be converted to:
> > > >         mem->alloc_size = memblock_alloc(alloc_size, PAGE_SIZE);
> > > >         if (...)
> > > > 
> > > > Seems that it got lost in the search and replace?
> > > 
> > > Yes, I messed that up during the rebase.  That being said it magically
> > > gets fixed in the next patch..
> > 
> > Yes. However if someone does a bisection they are going to be mighty unhappy
> > with you.
> 
> Sure, I was planning on fixing it anyway.  Just waiting for feedback
> on the rest of the patches before doing a respin.

I put the patches up-to this one on :

 git://git.kernel.org/pub/scm/linux/kernel/git/konrad/swiotlb 
devel/for-linus-5.13

Would you be OK rebasing on top of that and sending the patches?

Juergen, would you be OK testing that branch on your Xen setup?