Re: [Qemu-devel] [PATCH v3 12/12] vhost: iommu: cache static mapping if there is

[Peter Xu]

On Thu, May 11, 2017 at 04:35:21PM +0800, Jason Wang wrote:
> 
> 
> On 2017年05月10日 16:01, Peter Xu wrote:
> >This patch pre-heat vhost iotlb cache when passthrough mode enabled.
> >
> >Sometimes, even if user specified iommu_platform for vhost devices,
> >IOMMU might still be disabled. One case is passthrough mode in VT-d
> >implementation. We can detect this by observing iommu_list. If it's
> >empty, it means IOMMU translation is disabled, then we can actually
> >pre-heat the translation (it'll be static mapping then) by first
> >invalidating all IOTLB, then cache existing memory ranges into vhost
> >backend iotlb using 1:1 mapping.
> >
> >Signed-off-by: Peter Xu <address@hidden>
> >---
> >  hw/virtio/trace-events |  4 ++++
> >  hw/virtio/vhost.c      | 49 
> > +++++++++++++++++++++++++++++++++++++++++++++++++
> >  2 files changed, 53 insertions(+)
> >
> >diff --git a/hw/virtio/trace-events b/hw/virtio/trace-events
> >index 1f7a7c1..54dcbb3 100644
> >--- a/hw/virtio/trace-events
> >+++ b/hw/virtio/trace-events
> >@@ -24,3 +24,7 @@ virtio_balloon_handle_output(const char *name, uint64_t 
> >gpa) "section name: %s g
> >  virtio_balloon_get_config(uint32_t num_pages, uint32_t actual) "num_pages: 
> > %d actual: %d"
> >  virtio_balloon_set_config(uint32_t actual, uint32_t oldactual) "actual: %d 
> > oldactual: %d"
> >  virtio_balloon_to_target(uint64_t target, uint32_t num_pages) "balloon 
> > target: %"PRIx64" num_pages: %d"
> >+
> >+# hw/virtio/vhost.c
> >+vhost_iommu_commit(void) ""
> >+vhost_iommu_static_preheat(void) ""
> >diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
> >index 0001e60..1c92e62 100644
> >--- a/hw/virtio/vhost.c
> >+++ b/hw/virtio/vhost.c
> >@@ -27,6 +27,7 @@
> >  #include "hw/virtio/virtio-access.h"
> >  #include "migration/migration.h"
> >  #include "sysemu/dma.h"
> >+#include "trace.h"
> >  /* enabled until disconnected backend stabilizes */
> >  #define _VHOST_DEBUG 1
> >@@ -730,6 +731,11 @@ static void vhost_iommu_unmap_notify(IOMMUNotifier *n, 
> >IOMMUTLBEntry *iotlb)
> >      }
> >  }
> >+static bool vhost_iommu_mr_enabled(struct vhost_dev *dev)
> >+{
> >+    return !QLIST_EMPTY(&dev->iommu_list);
> >+}
> >+
> >  static void vhost_iommu_region_add(MemoryListener *listener,
> >                                     MemoryRegionSection *section)
> >  {
> >@@ -782,6 +788,48 @@ static void vhost_iommu_region_del(MemoryListener 
> >*listener,
> >      }
> >  }
> >+static void vhost_iommu_commit(MemoryListener *listener)
> >+{
> >+    struct vhost_dev *dev = container_of(listener, struct vhost_dev,
> >+                                         iommu_listener);
> >+    struct vhost_memory_region *r;
> >+    int i;
> >+
> >+    trace_vhost_iommu_commit();
> >+
> >+    if (!vhost_iommu_mr_enabled(dev)) {
> >+        /*
> >+        * This means iommu_platform is enabled, however iommu memory
> >+        * region is disabled, e.g., when device passthrough is setup.
> >+        * Then, no translation is needed any more.
> >+        *
> >+        * Let's first invalidate the whole IOTLB, then pre-heat the
> >+        * static mapping by looping over vhost memory ranges.
> >+        */
> >+
> >+        if (dev->vhost_ops->vhost_invalidate_device_iotlb(dev, 0,
> >+                                                          UINT64_MAX-1)) {
> >+            error_report("%s: flush existing IOTLB failed", __func__);
> >+            return;
> >+        }
> >+
> >+        for (i = 0; i < dev->mem->nregions; i++) {
> >+            r = &dev->mem->regions[i];
> >+            /* Vhost regions are writable RAM, so IOMMU_RW suites. */
> >+            if (dev->vhost_ops->vhost_update_device_iotlb(dev,
> >+                                                          
> >r->guest_phys_addr,
> >+                                                          r->userspace_addr,
> >+                                                          r->memory_size,
> >+                                                          IOMMU_RW)) {
> >+                error_report("%s: pre-heat static mapping failed", 
> >__func__);
> >+                return;
> >+            }
> >+        }
> >+
> >+        trace_vhost_iommu_static_preheat();
> >+    }
> >+}
> 
> Looks like vfio does the map in region_add(), if we can have different types
> of memory regions (e.g some were under an IOMMU but others were not), do we
> need to switch to do this in vhost_iommu_region_add() ?

Currently this is only a pre-heat of cache only if IOMMU is totally
disabled (!vhost_iommu_mr_enabled(dev) means no IOMMU memory regions).
This patch won't be activated without this condition, so for the cases
(non-x86 platforms) where there are some IOMMU regions, it'll be just
automatically disabled. And, I don't really quite sure whether we
should cache non-IOMMU regions when there are some IOMMU regions... So
imho we can keep this until one day we really want to support some
non-x86 platforms for vhost-dmar, then we can work on top. Thanks,

-- 
Peter Xu