Skip to content

Commit

Permalink
drm/i915/kvmgt: Support setting dma map for huge pages
Browse files Browse the repository at this point in the history 
To support huge gtt, we need to support huge pages in kvmgt first.
This patch adds a 'size' param to the intel_gvt_mpt::dma_map_guest_page
API and implements it in kvmgt.

v2: rebase.

Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
  • Loading branch information
Changbin Du authored and Zhenyu Wang committed Jul 9, 2018
1 parent eb3a353 commit 79e542f
Show file tree
Hide file tree
Showing 4 changed files with 102 additions and 39 deletions.
6 changes: 3 additions & 3 deletions drivers/gpu/drm/i915/gvt/gtt.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1106,7 +1106,7 @@ static int split_64KB_gtt_entry(struct intel_vgpu *vgpu,

for (i = 0; i < GTT_64K_PTE_STRIDE; i++) {
ret = intel_gvt_hypervisor_dma_map_guest_page(vgpu,
start_gfn + i, &dma_addr);
start_gfn + i, PAGE_SIZE, &dma_addr);
if (ret)
return ret;

Expand Down Expand Up @@ -1152,7 +1152,7 @@ static int ppgtt_populate_shadow_entry(struct intel_vgpu *vgpu,
};

/* direct shadow */
ret = intel_gvt_hypervisor_dma_map_guest_page(vgpu, gfn, &dma_addr);
ret = intel_gvt_hypervisor_dma_map_guest_page(vgpu, gfn, PAGE_SIZE, &dma_addr);
if (ret)
return -ENXIO;

Expand Down Expand Up @@ -2080,7 +2080,7 @@ static int emulate_ggtt_mmio_write(struct intel_vgpu *vgpu, unsigned int off,
}

ret = intel_gvt_hypervisor_dma_map_guest_page(vgpu, gfn,
&dma_addr);
PAGE_SIZE, &dma_addr);
if (ret) {
gvt_vgpu_err("fail to populate guest ggtt entry\n");
/* guest driver may read/write the entry when partial
Expand Down
2 changes: 1 addition & 1 deletion drivers/gpu/drm/i915/gvt/hypercall.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,7 @@ struct intel_gvt_mpt {
unsigned long (*gfn_to_mfn)(unsigned long handle, unsigned long gfn);

int (*dma_map_guest_page)(unsigned long handle, unsigned long gfn,
dma_addr_t *dma_addr);
unsigned long size, dma_addr_t *dma_addr);
void (*dma_unmap_guest_page)(unsigned long handle, dma_addr_t dma_addr);

int (*map_gfn_to_mfn)(unsigned long handle, unsigned long gfn,
Expand Down
126 changes: 94 additions & 32 deletions drivers/gpu/drm/i915/gvt/kvmgt.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -94,6 +94,7 @@ struct gvt_dma {
struct rb_node dma_addr_node;
gfn_t gfn;
dma_addr_t dma_addr;
unsigned long size;
struct kref ref;
};

Expand All @@ -106,45 +107,103 @@ static int kvmgt_guest_init(struct mdev_device *mdev);
static void intel_vgpu_release_work(struct work_struct *work);
static bool kvmgt_guest_exit(struct kvmgt_guest_info *info);

static void gvt_unpin_guest_page(struct intel_vgpu *vgpu, unsigned long gfn,
unsigned long size)
{
int total_pages;
int npage;
int ret;

total_pages = roundup(size, PAGE_SIZE) / PAGE_SIZE;

for (npage = 0; npage < total_pages; npage++) {
unsigned long cur_gfn = gfn + npage;

ret = vfio_unpin_pages(mdev_dev(vgpu->vdev.mdev), &cur_gfn, 1);
WARN_ON(ret != 1);
}
}

/* Pin a normal or compound guest page for dma. */
static int gvt_pin_guest_page(struct intel_vgpu *vgpu, unsigned long gfn,
unsigned long size, struct page **page)
{
unsigned long base_pfn = 0;
int total_pages;
int npage;
int ret;

total_pages = roundup(size, PAGE_SIZE) / PAGE_SIZE;
/*
* We pin the pages one-by-one to avoid allocating a big arrary
* on stack to hold pfns.
*/
for (npage = 0; npage < total_pages; npage++) {
unsigned long cur_gfn = gfn + npage;
unsigned long pfn;

ret = vfio_pin_pages(mdev_dev(vgpu->vdev.mdev), &cur_gfn, 1,
IOMMU_READ | IOMMU_WRITE, &pfn);
if (ret != 1) {
gvt_vgpu_err("vfio_pin_pages failed for gfn 0x%lx, ret %d\n",
cur_gfn, ret);
goto err;
}

if (!pfn_valid(pfn)) {
gvt_vgpu_err("pfn 0x%lx is not mem backed\n", pfn);
npage++;
ret = -EFAULT;
goto err;
}

if (npage == 0)
base_pfn = pfn;
else if (base_pfn + npage != pfn) {
gvt_vgpu_err("The pages are not continuous\n");
ret = -EINVAL;
npage++;
goto err;
}
}

*page = pfn_to_page(base_pfn);
return 0;
err:
gvt_unpin_guest_page(vgpu, gfn, npage * PAGE_SIZE);
return ret;
}

static int gvt_dma_map_page(struct intel_vgpu *vgpu, unsigned long gfn,
dma_addr_t *dma_addr)
dma_addr_t *dma_addr, unsigned long size)
{
struct device *dev = &vgpu->gvt->dev_priv->drm.pdev->dev;
struct page *page;
unsigned long pfn;
struct page *page = NULL;
int ret;

/* Pin the page first. */
ret = vfio_pin_pages(mdev_dev(vgpu->vdev.mdev), &gfn, 1,
IOMMU_READ | IOMMU_WRITE, &pfn);
if (ret != 1) {
gvt_vgpu_err("vfio_pin_pages failed for gfn 0x%lx: %d\n",
gfn, ret);
return -EINVAL;
}
ret = gvt_pin_guest_page(vgpu, gfn, size, &page);
if (ret)
return ret;

/* Setup DMA mapping. */
page = pfn_to_page(pfn);
*dma_addr = dma_map_page(dev, page, 0, PAGE_SIZE,
PCI_DMA_BIDIRECTIONAL);
if (dma_mapping_error(dev, *dma_addr)) {
gvt_vgpu_err("DMA mapping failed for gfn 0x%lx\n", gfn);
vfio_unpin_pages(mdev_dev(vgpu->vdev.mdev), &gfn, 1);
return -ENOMEM;
*dma_addr = dma_map_page(dev, page, 0, size, PCI_DMA_BIDIRECTIONAL);
ret = dma_mapping_error(dev, *dma_addr);
if (ret) {
gvt_vgpu_err("DMA mapping failed for pfn 0x%lx, ret %d\n",
page_to_pfn(page), ret);
gvt_unpin_guest_page(vgpu, gfn, size);
}

return 0;
return ret;
}

static void gvt_dma_unmap_page(struct intel_vgpu *vgpu, unsigned long gfn,
dma_addr_t dma_addr)
dma_addr_t dma_addr, unsigned long size)
{
struct device *dev = &vgpu->gvt->dev_priv->drm.pdev->dev;
int ret;

dma_unmap_page(dev, dma_addr, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
ret = vfio_unpin_pages(mdev_dev(vgpu->vdev.mdev), &gfn, 1);
WARN_ON(ret != 1);
dma_unmap_page(dev, dma_addr, size, PCI_DMA_BIDIRECTIONAL);
gvt_unpin_guest_page(vgpu, gfn, size);
}

static struct gvt_dma *__gvt_cache_find_dma_addr(struct intel_vgpu *vgpu,
Expand Down Expand Up @@ -185,7 +244,7 @@ static struct gvt_dma *__gvt_cache_find_gfn(struct intel_vgpu *vgpu, gfn_t gfn)
}

static int __gvt_cache_add(struct intel_vgpu *vgpu, gfn_t gfn,
dma_addr_t dma_addr)
dma_addr_t dma_addr, unsigned long size)
{
struct gvt_dma *new, *itr;
struct rb_node **link, *parent = NULL;
Expand All @@ -197,6 +256,7 @@ static int __gvt_cache_add(struct intel_vgpu *vgpu, gfn_t gfn,
new->vgpu = vgpu;
new->gfn = gfn;
new->dma_addr = dma_addr;
new->size = size;
kref_init(&new->ref);

/* gfn_cache maps gfn to struct gvt_dma. */
Expand Down Expand Up @@ -254,7 +314,7 @@ static void gvt_cache_destroy(struct intel_vgpu *vgpu)
break;
}
dma = rb_entry(node, struct gvt_dma, gfn_node);
gvt_dma_unmap_page(vgpu, dma->gfn, dma->dma_addr);
gvt_dma_unmap_page(vgpu, dma->gfn, dma->dma_addr, dma->size);
__gvt_cache_remove_entry(vgpu, dma);
mutex_unlock(&vgpu->vdev.cache_lock);
}
Expand Down Expand Up @@ -509,7 +569,8 @@ static int intel_vgpu_iommu_notifier(struct notifier_block *nb,
if (!entry)
continue;

gvt_dma_unmap_page(vgpu, entry->gfn, entry->dma_addr);
gvt_dma_unmap_page(vgpu, entry->gfn, entry->dma_addr,
entry->size);
__gvt_cache_remove_entry(vgpu, entry);
}
mutex_unlock(&vgpu->vdev.cache_lock);
Expand Down Expand Up @@ -1616,7 +1677,7 @@ static unsigned long kvmgt_gfn_to_pfn(unsigned long handle, unsigned long gfn)
}

int kvmgt_dma_map_guest_page(unsigned long handle, unsigned long gfn,
dma_addr_t *dma_addr)
unsigned long size, dma_addr_t *dma_addr)
{
struct kvmgt_guest_info *info;
struct intel_vgpu *vgpu;
Expand All @@ -1633,11 +1694,11 @@ int kvmgt_dma_map_guest_page(unsigned long handle, unsigned long gfn,

entry = __gvt_cache_find_gfn(info->vgpu, gfn);
if (!entry) {
ret = gvt_dma_map_page(vgpu, gfn, dma_addr);
ret = gvt_dma_map_page(vgpu, gfn, dma_addr, size);
if (ret)
goto err_unlock;

ret = __gvt_cache_add(info->vgpu, gfn, *dma_addr);
ret = __gvt_cache_add(info->vgpu, gfn, *dma_addr, size);
if (ret)
goto err_unmap;
} else {
Expand All @@ -1649,7 +1710,7 @@ int kvmgt_dma_map_guest_page(unsigned long handle, unsigned long gfn,
return 0;

err_unmap:
gvt_dma_unmap_page(vgpu, gfn, *dma_addr);
gvt_dma_unmap_page(vgpu, gfn, *dma_addr, size);
err_unlock:
mutex_unlock(&info->vgpu->vdev.cache_lock);
return ret;
Expand All @@ -1659,7 +1720,8 @@ static void __gvt_dma_release(struct kref *ref)
{
struct gvt_dma *entry = container_of(ref, typeof(*entry), ref);

gvt_dma_unmap_page(entry->vgpu, entry->gfn, entry->dma_addr);
gvt_dma_unmap_page(entry->vgpu, entry->gfn, entry->dma_addr,
entry->size);
__gvt_cache_remove_entry(entry->vgpu, entry);
}

Expand Down
7 changes: 4 additions & 3 deletions drivers/gpu/drm/i915/gvt/mpt.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -230,17 +230,18 @@ static inline unsigned long intel_gvt_hypervisor_gfn_to_mfn(
/**
* intel_gvt_hypervisor_dma_map_guest_page - setup dma map for guest page
* @vgpu: a vGPU
* @gpfn: guest pfn
* @gfn: guest pfn
* @size: page size
* @dma_addr: retrieve allocated dma addr
*
* Returns:
* 0 on success, negative error code if failed.
*/
static inline int intel_gvt_hypervisor_dma_map_guest_page(
struct intel_vgpu *vgpu, unsigned long gfn,
struct intel_vgpu *vgpu, unsigned long gfn, unsigned long size,
dma_addr_t *dma_addr)
{
return intel_gvt_host.mpt->dma_map_guest_page(vgpu->handle, gfn,
return intel_gvt_host.mpt->dma_map_guest_page(vgpu->handle, gfn, size,
dma_addr);
}

Expand Down

0 comments on commit 79e542f

Please sign in to comment.