Re: [PATCH] infiniband/mthca : Fix userland mapping of mthca infiniband cards in Xen dom0

[Roland Dreier <rdreier-FYB4Gu1CFyUAvxtiuMwx3w@public.gmane.org>]

 > --- kernel-2.6.32.fc12.orig/drivers/infiniband/hw/mthca/mthca_provider.c	2010-12-16 16:39:06.901130993 +0100
 > +++ kernel-2.6.32.fc12.new/drivers/infiniband/hw/mthca/mthca_provider.c	2010-12-16 17:02:07.944127489 +0100
 > @@ -391,6 +391,8 @@ static int mthca_mmap_uar(struct ib_ucon
 >  	if (vma->vm_end - vma->vm_start != PAGE_SIZE)
 >  		return -EINVAL;
 >  
 > +	vma->vm_flags |= VM_IO;
 > +	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
 >  	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 >  
 >  	if (io_remap_pfn_range(vma, vma->vm_start,

This is kind of unappealing -- is there no way to make
io_remap_pfn_range do the right thing under Xen?  It seems unfortunate
to make drivers both set VM_IO and call io_remap_pfn_range.  Or maybe we
should get rid of io_remap_pfn_range() and just have drivers set VM_IO
and call remap_pfn_range() all the time?

There are quite a few drivers under drivers/infiniband/hw that call
io_remap_pfn_range() -- presumably they all need the analogous fix?

Finally as a stylistic thing I would probably prefer to see the
vm_page_prot manipulation written as

	vma->vm_page_prot = pgprot_noncached(vm_get_page_prot(vma->vm_flags));

and if this fix is really the right thing to do, this should probably be
wrapped up in a helper function; presumably every instance of

	vma->vm_page_prot = pgprot_FOO(vma->vm_page_prot)

under drivers/ would need the same fix for Xen.

 - R.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html