On Thu, 2015-10-08 at 14:56 +0100, Tvrtko Ursulin wrote:
> Hi,
>
> On 08/10/15 07:24, ankitprasad.r.sha...@intel.com wrote:
> > From: Ankitprasad Sharma <ankitprasad.r.sha...@intel.com>
> >
> > This patch adds support for extending the pread/pwrite functionality
> > for objects not backed by shmem. The access will be made through
> > gtt interface.
> > This will cover prime objects as well as stolen memory backed objects
> > but for userptr objects it is still forbidden.
>
> Where is the part which forbids it for userptr objects?
In version 5, updated the patch handle pwrite/pread for all non-shmem
backed objects, including userptr objects
Will update the Commit message
>
> > v2: Drop locks around slow_user_access, prefault the pages before
> > access (Chris)
> >
> > v3: Rebased to the latest drm-intel-nightly (Ankit)
> >
> > v4: Moved page base & offset calculations outside the copy loop,
> > corrected data types for size and offset variables, corrected if-else
> > braces format (Tvrtko/kerneldocs)
> >
> > v5: Enabled pread/pwrite for all non-shmem backed objects including
> > without tiling restrictions (Ankit)
> >
> > v6: Using pwrite_fast for non-shmem backed objects as well (Chris)
> >
> > Testcase: igt/gem_stolen
> >
> > Signed-off-by: Ankitprasad Sharma <ankitprasad.r.sha...@intel.com>
> > ---
> > drivers/gpu/drm/i915/i915_gem.c | 125
> > +++++++++++++++++++++++++++++++++-------
> > 1 file changed, 104 insertions(+), 21 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/i915_gem.c
> > b/drivers/gpu/drm/i915/i915_gem.c
> > index 91a2e97..2c94e22 100644
> > --- a/drivers/gpu/drm/i915/i915_gem.c
> > +++ b/drivers/gpu/drm/i915/i915_gem.c
> > @@ -614,6 +614,99 @@ shmem_pread_slow(struct page *page, int
> > shmem_page_offset, int page_length,
> > return ret ? - EFAULT : 0;
> > }
> >
> > +static inline uint64_t
> > +slow_user_access(struct io_mapping *mapping,
> > + uint64_t page_base, int page_offset,
> > + char __user *user_data,
> > + int length, bool pwrite)
> > +{
> > + void __iomem *vaddr_inatomic;
> > + void *vaddr;
> > + uint64_t unwritten;
> > +
> > + vaddr_inatomic = io_mapping_map_wc(mapping, page_base);
> > + /* We can use the cpu mem copy function because this is X86. */
> > + vaddr = (void __force *)vaddr_inatomic + page_offset;
> > + if (pwrite)
> > + unwritten = __copy_from_user(vaddr, user_data, length);
> > + else
> > + unwritten = __copy_to_user(user_data, vaddr, length);
> > +
> > + io_mapping_unmap(vaddr_inatomic);
> > + return unwritten;
> > +}
> > +
> > +static int
> > +i915_gem_gtt_pread(struct drm_device *dev,
> > + struct drm_i915_gem_object *obj, uint64_t size,
> > + uint64_t data_offset, uint64_t data_ptr)
> > +{
> > + struct drm_i915_private *dev_priv = dev->dev_private;
> > + char __user *user_data;
> > + uint64_t remain;
> > + uint64_t offset, page_base;
> > + int page_offset, page_length, ret = 0;
> > +
> > + ret = i915_gem_obj_ggtt_pin(obj, 0, PIN_MAPPABLE);
> > + if (ret)
> > + goto out;
> > +
> > + ret = i915_gem_object_set_to_gtt_domain(obj, false);
> > + if (ret)
> > + goto out_unpin;
> > +
> > + ret = i915_gem_object_put_fence(obj);
> > + if (ret)
> > + goto out_unpin;
> > +
> > + user_data = to_user_ptr(data_ptr);
> > + remain = size;
> > + offset = i915_gem_obj_ggtt_offset(obj) + data_offset;
> > +
> > + mutex_unlock(&dev->struct_mutex);
> > + if (likely(!i915.prefault_disable))
> > + ret = fault_in_multipages_writeable(user_data, remain);
> > +
> > + /*
> > + * page_offset = offset within page
> > + * page_base = page offset within aperture
> > + */
> > + page_offset = offset_in_page(offset);
> > + page_base = offset & PAGE_MASK;
> > +
> > + while (remain > 0) {
> > + /* page_length = bytes to copy for this page */
> > + page_length = remain;
> > + if ((page_offset + remain) > PAGE_SIZE)
> > + page_length = PAGE_SIZE - page_offset;
> > +
> > + /* This is a slow read/write as it tries to read from
> > + * and write to user memory which may result into page
> > + * faults
> > + */
> > + ret = slow_user_access(dev_priv->gtt.mappable, page_base,
> > + page_offset, user_data,
> > + page_length, false);
> > +
> > + if (ret) {
> > + ret = -EFAULT;
> > + break;
> > + }
> > +
> > + remain -= page_length;
> > + user_data += page_length;
> > + page_base += page_length;
> > + page_offset = 0;
> > + }
> > +
> > + mutex_lock(&dev->struct_mutex);
> > +
> > +out_unpin:
> > + i915_gem_object_ggtt_unpin(obj);
> > +out:
> > + return ret;
> > +}
> > +
> > static int
> > i915_gem_shmem_pread(struct drm_device *dev,
> > struct drm_i915_gem_object *obj,
> > @@ -737,17 +830,14 @@ i915_gem_pread_ioctl(struct drm_device *dev, void
> > *data,
> > goto out;
> > }
> >
> > - /* prime objects have no backing filp to GEM pread/pwrite
> > - * pages from.
> > - */
> > - if (!obj->base.filp) {
> > - ret = -EINVAL;
> > - goto out;
> > - }
> > -
> > trace_i915_gem_object_pread(obj, args->offset, args->size);
> >
> > - ret = i915_gem_shmem_pread(dev, obj, args, file);
> > + /* pread for non shmem backed objects */
> > + if (!obj->base.filp && obj->tiling_mode == I915_TILING_NONE)
> > + ret = i915_gem_gtt_pread(dev, obj, args->size,
> > + args->offset, args->data_ptr);
> > + else
> > + ret = i915_gem_shmem_pread(dev, obj, args, file);
> >
> > out:
> > drm_gem_object_unreference(&obj->base);
> > @@ -795,7 +885,7 @@ i915_gem_gtt_pwrite_fast(struct drm_device *dev,
> > char __user *user_data;
> > int page_offset, page_length, ret;
> >
> > - ret = i915_gem_obj_ggtt_pin(obj, 0, PIN_MAPPABLE | PIN_NONBLOCK);
> > + ret = i915_gem_obj_ggtt_pin(obj, 0, PIN_MAPPABLE);
>
> Why is this needed?
This was Chris' suggestion. This change can go as a separate patch, if
needed. I do not think pwrite/pread has any dependency on this.
Need Chris to respond on this.
>
> > if (ret)
> > goto out;
> >
> > @@ -1090,14 +1180,6 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void
> > *data,
> > goto out;
> > }
> >
> > - /* prime objects have no backing filp to GEM pread/pwrite
> > - * pages from.
> > - */
> > - if (!obj->base.filp) {
> > - ret = -EINVAL;
> > - goto out;
> > - }
> > -
> > trace_i915_gem_object_pwrite(obj, args->offset, args->size);
> >
> > ret = -EFAULT;
> > @@ -1108,8 +1190,9 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void
> > *data,
> > * perspective, requiring manual detiling by the client.
> > */
> > if (obj->tiling_mode == I915_TILING_NONE &&
> > - obj->base.write_domain != I915_GEM_DOMAIN_CPU &&
> > - cpu_write_needs_clflush(obj)) {
> > + (!obj->base.filp ||
> > + (obj->base.write_domain != I915_GEM_DOMAIN_CPU &&
> > + cpu_write_needs_clflush(obj)))) {
> > ret = i915_gem_gtt_pwrite_fast(dev, obj, args, file);
>
> So the pwrite path will fail if a page fault happens, as opposed to the
> pread path which makes an effort to handle it. What is the reason for
> this asymmetry in the API? Or I am missing something?
I had earlier implemented the pwrite and pread maintaining the symmetry
in the API. After couple of revisions we landed on this implementation.
Need Chris to respond on this.
Thanks,
Ankit
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
