[dpdk-dev] [PATCH] vfio: fix sPAPR IOMMU DMA window size

[Jonas Pfefferle1]

"Ananyev, Konstantin" <konstantin.ananyev at intel.com> wrote on 08/08/2017
10:27:28 AM:

> From: "Ananyev, Konstantin" <konstantin.ananyev at intel.com>
> To: Alexey Kardashevskiy <aik at ozlabs.ru>, Jonas Pfefferle
> <jpf at zurich.ibm.com>, "Burakov, Anatoly" <anatoly.burakov at intel.com>
> Cc: "dev at dpdk.org" <dev at dpdk.org>
> Date: 08/08/2017 10:27 AM
> Subject: RE: [dpdk-dev] [PATCH] vfio: fix sPAPR IOMMU DMA window size
>
>
>
> > -----Original Message-----
> > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Alexey
Kardashevskiy
> > Sent: Tuesday, August 8, 2017 10:38 AM
> > To: Jonas Pfefferle <jpf at zurich.ibm.com>; Burakov, Anatoly
> <anatoly.burakov at intel.com>
> > Cc: dev at dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH] vfio: fix sPAPR IOMMU DMA window size
> >
> > On 08/08/17 01:11, Jonas Pfefferle wrote:
> > > DMA window size needs to be big enough to span all memory segment's
> > > physical addresses. We do not need multiple levels of IOMMU tables
> > > as we already span ~70TB of physical memory with 16MB hugepages.
> > >
> > > Signed-off-by: Jonas Pfefferle <jpf at zurich.ibm.com>
> > > ---
> > >  lib/librte_eal/linuxapp/eal/eal_vfio.c | 25 +++++++++++++++++++++
+---
> > >  1 file changed, 22 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/lib/librte_eal/linuxapp/eal/eal_vfio.c b/lib/
> librte_eal/linuxapp/eal/eal_vfio.c
> > > index 946df7e..8502216 100644
> > > --- a/lib/librte_eal/linuxapp/eal/eal_vfio.c
> > > +++ b/lib/librte_eal/linuxapp/eal/eal_vfio.c
> > > @@ -722,6 +722,18 @@ vfio_type1_dma_map(int vfio_container_fd)
> > >     return 0;
> > >  }
> > >
> > > +static uint64_t
> > > +roundup_next_pow2(uint64_t n)
> > > +{
> > > +   uint32_t i;
> > > +
> > > +   n--;
> > > +   for (i = 1; i < sizeof(n) * CHAR_BIT; i += i)
> > > +      n |= n >> i;
> > > +
> > > +   return ++n;
> > > +}
> > > +
> >
> > wow :)
> >
> > QEMU does it using __builtin_ctzll() (used below for the page_shift)
> > without a loop:
> >
> > https://git.qemu.org/gitweb.cgi?p=qemu.git;a=blob;f=include/qemu/host-
> >
>
utils.h;h=95cf4f4163e50457cdf808263065ca5ef3f935da;hb=f22ab6cb0c47bd2a2785b7d58130949bd7d8d9af#l382

> >
> >
> > Anyway, seems working.
>
> As I remember, there already exists rte_align64pow2().
> Konstantin

Thanks. Fixed it.

>
> >
> >
> > Reviewed-by: Alexey Kardashevskiy <aik at ozlabs.ru>
> >
> >
> >
> >
> > >  static int
> > >  vfio_spapr_dma_map(int vfio_container_fd)
> > >  {
> > > @@ -759,10 +771,12 @@ vfio_spapr_dma_map(int vfio_container_fd)
> > >        return -1;
> > >     }
> > >
> > > -   /* calculate window size based on number of hugepages configured
*/
> > > -   create.window_size = rte_eal_get_physmem_size();
> > > +   /* physicaly pages are sorted descending i.e. ms
> [0].phys_addr is max */
> > > +   /* create DMA window from 0 to max(phys_addr + len) */
> > > +   /* sPAPR requires window size to be a power of 2 */
> > > +   create.window_size = roundup_next_pow2(ms[0].phys_addr + ms
[0].len);
> > >     create.page_shift = __builtin_ctzll(ms->hugepage_sz);
> > > -   create.levels = 2;
> > > +   create.levels = 1;
> > >
> > >     ret = ioctl(vfio_container_fd, VFIO_IOMMU_SPAPR_TCE_CREATE,
&create);
> > >     if (ret) {
> > > @@ -771,6 +785,11 @@ vfio_spapr_dma_map(int vfio_container_fd)
> > >        return -1;
> > >     }
> > >
> > > +   if (create.start_addr != 0) {
> > > +      RTE_LOG(ERR, EAL, "  DMA window start address != 0\n");
> > > +      return -1;
> > > +   }
> > > +
> > >     /* map all DPDK segments for DMA. use 1:1 PA to IOVA mapping */
> > >     for (i = 0; i < RTE_MAX_MEMSEG; i++) {
> > >        struct vfio_iommu_type1_dma_map dma_map;
> > >
> >
> >
> > --
> > Alexey