Re: [PATCH 0/6] IOMMU, vtd and iotlb flush rework (v8)

[Keir Fraser <keir@xen.org>]

On 16/11/2011 19:25, "Jean Guyader" <jean.guyader@eu.citrix.com> wrote:

> 
> In one of my previous email I detailed a bug I was seeing when passing
> through a Intel GPU on a guest that has more that 4G or RAM.
> 
> Allen suggested that I go for the Plan B but after a discussion with Tim
> we agreed that Plan B was way to disruptive in term of code change.
> 
> This patch series implements Plan A.

I'll check these in when they get an Ack from Jan.

 -- Keir

> http://xen.1045712.n5.nabble.com/VTD-Intel-iommu-IOTLB-flush-really-slow-td495
> 2866.html
> 
> Changes between v7 and v8:
>         - Rebase on new add_to_physmap from Andres Lagar-Cavilla
>         - Only do copy_to_guest on the new XENMAPSPACE
> (XENMAPSPACE_gmfn_range).
>         - Convert the xatp argument to a pointer in xenmem_add_to_physmap.
>         - Modify xenmem_add_to_physmap so it doesn't touchs the input (xatp is
> const now).
> 
> Changes between v6 and v7:
>        - xenmem_add_to_physmap_once can't take a pointer on xatp because
>          it modifies .idx and .gpfn.
>        - Cancel hypercall continuation in the compat layer if the
> copy_to_guest
>          failed.
> 
> Changes between v5 and v6:
>       - Rework the patch queue to make it more readable.
>       - Modify xatp in place in xenmem_add_to_physmap
>       - Only check for preemption if we are not at the last iteration
>       - Copy xatp guest handler back to the guest only in case of continuation
>       - Add continuation only when dealing with the new xenmem space
>         (XENMAPSPACE_gmfn_range).
> 
> Changes between v4 and v5:
>       - Fix hypercall continuation for add_to_physmap in compat mode.
> 
> Changes between v3 and v4:
>       - Move the loop for gmfn_range from arch_memory_op to
> xenmem_add_to_physmap.
>       - Add a comment to comment to explain the purpose of
> iommu_dont_flush_iotlb.
> 
> Changes between v2 and v3:
>       - Check for the presence iotlb_flush_all callback before calling it.
> 
> Changes between v1 and v2:
>       - Move size in struct xen_add_to_physmap in padding between .domid and
> .space.
>       - Store iommu_dont_flush per cpu
>       - Change the code in hvmloader to relocate by batch of 64K, .size is now
> 16 bits.
> 
> Jean Guyader (6):
>   vtd: Refactor iotlb flush code
>   iommu: Introduce iommu_flush and iommu_flush_all.
>   add_to_physmap: Move the code for XENMEM_add_to_physmap
>   mm: New XENMEM space, XENMAPSPACE_gmfn_range
>   hvmloader: Change memory relocation loop when overlap with PCI hole
>   Introduce per cpu flag (iommu_dont_flush_iotlb) to avoid unnecessary
>     iotlb flush
> 
>  tools/firmware/hvmloader/pci.c      |   20 +++-
>  xen/arch/x86/mm.c                   |  212
> +++++++++++++++++++++++------------
>  xen/arch/x86/x86_64/compat/mm.c     |   18 +++
>  xen/drivers/passthrough/iommu.c     |   25 ++++
>  xen/drivers/passthrough/vtd/iommu.c |  100 ++++++++++-------
>  xen/include/public/memory.h         |    4 +
>  xen/include/xen/iommu.h             |   17 +++
>  7 files changed, 277 insertions(+), 119 deletions(-)
>