On Mon, 7 Mar 2022 at 21:38, Vivek Kasireddy vivek.kasireddy@intel.com wrote:

On platforms capable of allowing 8K (7680 x 4320) modes, pinning 2 or more framebuffers/scanout buffers results in only one that is mappable/ fenceable. Therefore, pageflipping between these 2 FBs where only one is mappable/fenceable creates latencies large enough to miss alternate vblanks thereby producing less optimal framerate.

This mainly happens because when i915_gem_object_pin_to_display_plane() is called to pin one of the FB objs, the associated vma is identified as misplaced and therefore i915_vma_unbind() is called which unbinds and evicts it. This misplaced vma gets subseqently pinned only when i915_gem_object_ggtt_pin_ww() is called without PIN_MAPPABLE. This results in a latency of ~10ms and happens every other vblank/repaint cycle. Therefore, to fix this issue, we try to see if there is space to map at-least two objects of a given size and return early if there isn't. This would ensure that we do not try with PIN_MAPPABLE for any objects that are too big to map thereby preventing unncessary unbind.

Testcase: Running Weston and weston-simple-egl on an Alderlake_S (ADLS) platform with a 8K@60 mode results in only ~40 FPS. Since upstream Weston submits a frame ~7ms before the next vblank, the latencies seen between atomic commit and flip event are 7, 24 (7 + 16.66), 7, 24..... suggesting that it misses the vblank every other frame.

Here is the ftrace snippet that shows the source of the ~10ms latency: i915_gem_object_pin_to_display_plane() { 0.102 us | i915_gem_object_set_cache_level(); i915_gem_object_ggtt_pin_ww() { 0.390 us | i915_vma_instance(); 0.178 us | i915_vma_misplaced(); i915_vma_unbind() { __i915_active_wait() { 0.082 us | i915_active_acquire_if_busy(); 0.475 us | } intel_runtime_pm_get() { 0.087 us | intel_runtime_pm_acquire(); 0.259 us | } __i915_active_wait() { 0.085 us | i915_active_acquire_if_busy(); 0.240 us | } __i915_vma_evict() { ggtt_unbind_vma() { gen8_ggtt_clear_range() { 10507.255 us | } 10507.689 us | } 10508.516 us | }

v2: Instead of using bigjoiner checks, determine whether a scanout buffer is too big by checking to see if it is possible to map two of them into the ggtt.

v3 (Ville):

• Count how many fb objects can be fit into the available holes instead of checking for a hole twice the object size.
• Take alignment constraints into account.
• Limit this large scanout buffer check to >= Gen 11 platforms.

v4:

• Remove existing heuristic that checks just for size. (Ville)
• Return early if we find space to map at-least two objects. (Tvrtko)
• Slightly update the commit message.

v5: (Tvrtko)

• Rename the function to indicate that the object may be too big to map into the aperture.
• Account for guard pages while calculating the total size required for the object.
• Do not subject all objects to the heuristic check and instead consider objects only of a certain size.
• Do the hole walk using the rbtree.
• Preserve the existing PIN_NONBLOCK logic.
• Drop the PIN_MAPPABLE check while pinning the VMA.

v6: (Tvrtko)

• Return 0 on success and the specific error code on failure to preserve the existing behavior.

v7: (Ville)

• Drop the HAS_GMCH(i915), DISPLAY_VER(i915) < 11 and size < ggtt->mappable_end / 4 checks.
• Drop the redundant check that is based on previous heuristic.

v8:

• Make sure that we are holding the mutex associated with ggtt vm as we traverse the hole nodes.

v9: (Tvrtko)

• Use mutex_lock_interruptible_nested() instead of mutex_lock().

Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Cc: Tvrtko Ursulin tvrtko.ursulin@linux.intel.com Cc: Manasi Navare manasi.d.navare@intel.com Reviewed-by: Tvrtko Ursulin tvrtko.ursulin@intel.com Signed-off-by: Vivek Kasireddy vivek.kasireddy@intel.com

---

drivers/gpu/drm/i915/i915_gem.c | 128 +++++++++++++++++++++++--------- 1 file changed, 94 insertions(+), 34 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 9747924cc57b..e0d731b3f215 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -49,6 +49,7 @@ #include "gem/i915_gem_pm.h" #include "gem/i915_gem_region.h" #include "gem/i915_gem_userptr.h" +#include "gem/i915_gem_tiling.h" #include "gt/intel_engine_user.h" #include "gt/intel_gt.h" #include "gt/intel_gt_pm.h" @@ -882,6 +883,96 @@ static void discard_ggtt_vma(struct i915_vma *vma) spin_unlock(&obj->vma.lock); }

+static int +i915_gem_object_fits_in_aperture(struct drm_i915_gem_object *obj,

•                            u64 alignment, u64 flags)
  

Tvrtko asked me to ack the first patch, but then I looked at this and started wondering.

Conceptually this doesn't pass the smell test. What if we have multiple per-crtc buffers? Multiple planes on the same crtc? What if the app does triple buffer? You'll be forever busy tuning this heuristics, which can't fundamentally be fixed I think. The old "half of mappable" heuristic isn't really better, but at least it was dead simple.

Imo what we need here is a change in approach: 1. Check whether the useable view for scanout exists already. If yes, use that. This should avoid the constant unbinding stalls. 2. Try to in buffer to mappabley, but without evicting anything (so not the non-blocking thing) 3. Pin the buffer with the most lenient approach

Even the non-blocking interim stage is dangerous, since it'll just result in other buffers (e.g. when triple-buffering) getting unbound and we're back to the same stall. Note that this could have an impact on cpu rendering compositors, where we might end up relying a lot more partial views. But as long as we are a tad more aggressive (i.e. the non-blocking binding) in the mmap path that should work out to keep everything balanced, since usually you render first before you display anything. And so the buffer should end up in the ideal place.

I'd try to first skip the 2. step since I think it'll require a bit of work, and frankly I don't think we care about the potential fallout. -Daniel

+{

•   struct drm_i915_private *i915 = to_i915(obj->base.dev);
  

•   struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
  

•   struct drm_mm_node *hole;
  

•   u64 hole_start, hole_end, start, end;
  

•   u64 fence_size, fence_alignment;
  

•   unsigned int count = 0;
  

•   int err = 0;
  

•   /*
  

•    * If the required space is larger than the available
  

•    * aperture, we will not able to find a slot for the
  

•    * object and unbinding the object now will be in
  

•    * vain. Worse, doing so may cause us to ping-pong
  

•    * the object in and out of the Global GTT and
  

•    * waste a lot of cycles under the mutex.
  

•    */
  

•   if (obj->base.size > ggtt->mappable_end)
  

•           return -E2BIG;
  

•   /*
  

•    * If NONBLOCK is set the caller is optimistically
  

•    * trying to cache the full object within the mappable
  

•    * aperture, and *must* have a fallback in place for
  

•    * situations where we cannot bind the object. We
  

•    * can be a little more lax here and use the fallback
  

•    * more often to avoid costly migrations of ourselves
  

•    * and other objects within the aperture.
  

•    */
  

•   if (!(flags & PIN_NONBLOCK))
  

•           return 0;
  

•   /*
  

•    * Other objects such as batchbuffers are fairly small compared
  

•    * to FBs and are unlikely to exahust the aperture space.
  

•    * Therefore, return early if this obj is not an FB.
  

•    */
  

•   if (!i915_gem_object_is_framebuffer(obj))
  

•           return 0;
  

•   fence_size = i915_gem_fence_size(i915, obj->base.size,
  

•                                    i915_gem_object_get_tiling(obj),
  

•                                    i915_gem_object_get_stride(obj));
  

•   if (i915_vm_has_cache_coloring(&ggtt->vm))
  

•           fence_size += 2 * I915_GTT_PAGE_SIZE;
  

•   fence_alignment = i915_gem_fence_alignment(i915, obj->base.size,
  

•                                              i915_gem_object_get_tiling(obj),
  

•                                              i915_gem_object_get_stride(obj));
  

•   alignment = max_t(u64, alignment, fence_alignment);
  

•   err = mutex_lock_interruptible_nested(&ggtt->vm.mutex, 0);
  

•   if (err)
  

•           return err;
  

•   /*
  

•    * Assuming this object is a large scanout buffer, we try to find
  

•    * out if there is room to map at-least two of them. There could
  

•    * be space available to map one but to be consistent, we try to
  

•    * avoid mapping/fencing any of them.
  

•    */
  

•   drm_mm_for_each_suitable_hole(hole, &ggtt->vm.mm, 0, ggtt->mappable_end,
  

•                                 fence_size, DRM_MM_INSERT_LOW) {
  

•           hole_start = drm_mm_hole_node_start(hole);
  

•           hole_end = hole_start + hole->hole_size;
  

•           do {
  

•                   start = round_up(hole_start, alignment);
  

•                   end = min_t(u64, hole_end, ggtt->mappable_end);
  

•                   if (range_overflows(start, fence_size, end))
  

•                           break;
  

•                   if (++count >= 2) {
  

•                           mutex_unlock(&ggtt->vm.mutex);
  

•                           return 0;
  

•                   }
  

•                   hole_start = start + fence_size;
  

•           } while (1);
  

•   }
  

•   mutex_unlock(&ggtt->vm.mutex);
  

•   return -ENOSPC;
  

+}

struct i915_vma * i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object *obj, struct i915_gem_ww_ctx *ww, @@ -897,36 +988,9 @@ i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object *obj,

    if (flags & PIN_MAPPABLE &&
        (!view || view->type == I915_GGTT_VIEW_NORMAL)) {

•           /*
  

•            * If the required space is larger than the available
  

•            * aperture, we will not able to find a slot for the
  

•            * object and unbinding the object now will be in
  

•            * vain. Worse, doing so may cause us to ping-pong
  

•            * the object in and out of the Global GTT and
  

•            * waste a lot of cycles under the mutex.
  

•            */
  

•           if (obj->base.size > ggtt->mappable_end)
  

•                   return ERR_PTR(-E2BIG);
  

•           /*
  

•            * If NONBLOCK is set the caller is optimistically
  

•            * trying to cache the full object within the mappable
  

•            * aperture, and *must* have a fallback in place for
  

•            * situations where we cannot bind the object. We
  

•            * can be a little more lax here and use the fallback
  

•            * more often to avoid costly migrations of ourselves
  

•            * and other objects within the aperture.
  

•            *
  

•            * Half-the-aperture is used as a simple heuristic.
  

•            * More interesting would to do search for a free
  

•            * block prior to making the commitment to unbind.
  

•            * That caters for the self-harm case, and with a
  

•            * little more heuristics (e.g. NOFAULT, NOEVICT)
  

•            * we could try to minimise harm to others.
  

•            */
  

•           if (flags & PIN_NONBLOCK &&
  

•               obj->base.size > ggtt->mappable_end / 2)
  

•                   return ERR_PTR(-ENOSPC);
  

•           ret = i915_gem_object_fits_in_aperture(obj, alignment, flags);
  

•           if (ret)
  

•                   return ERR_PTR(ret);
    }
  
  

new_vma: @@ -938,10 +1002,6 @@ i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object *obj, if (flags & PIN_NONBLOCK) { if (i915_vma_is_pinned(vma) || i915_vma_is_active(vma)) return ERR_PTR(-ENOSPC);

•                   if (flags & PIN_MAPPABLE &&
  

•                       vma->fence_size > ggtt->mappable_end / 2)
  

•                           return ERR_PTR(-ENOSPC);
            }
  
            if (i915_vma_is_pinned(vma) || i915_vma_is_active(vma)) {
  

-- 2.35.1

Hi Tvrtko, Daniel,

On 11/03/2022 09:39, Daniel Vetter wrote:

...

On Mon, 7 Mar 2022 at 21:38, Vivek Kasireddy vivek.kasireddy@intel.com wrote:

...

On platforms capable of allowing 8K (7680 x 4320) modes, pinning 2 or more framebuffers/scanout buffers results in only one that is mappable/ fenceable. Therefore, pageflipping between these 2 FBs where only one is mappable/fenceable creates latencies large enough to miss alternate vblanks thereby producing less optimal framerate.

This mainly happens because when i915_gem_object_pin_to_display_plane() is called to pin one of the FB objs, the associated vma is identified as misplaced and therefore i915_vma_unbind() is called which unbinds and evicts it. This misplaced vma gets subseqently pinned only when i915_gem_object_ggtt_pin_ww() is called without PIN_MAPPABLE. This results in a latency of ~10ms and happens every other vblank/repaint cycle. Therefore, to fix this issue, we try to see if there is space to map at-least two objects of a given size and return early if there isn't. This would ensure that we do not try with PIN_MAPPABLE for any objects that are too big to map thereby preventing unncessary unbind.

Testcase: Running Weston and weston-simple-egl on an Alderlake_S (ADLS) platform with a 8K@60 mode results in only ~40 FPS. Since upstream Weston submits a frame ~7ms before the next vblank, the latencies seen between atomic commit and flip event are 7, 24 (7 + 16.66), 7, 24..... suggesting that it misses the vblank every other frame.

Here is the ftrace snippet that shows the source of the ~10ms latency: i915_gem_object_pin_to_display_plane() { 0.102 us | i915_gem_object_set_cache_level(); i915_gem_object_ggtt_pin_ww() { 0.390 us | i915_vma_instance(); 0.178 us | i915_vma_misplaced(); i915_vma_unbind() { __i915_active_wait() { 0.082 us | i915_active_acquire_if_busy(); 0.475 us | } intel_runtime_pm_get() { 0.087 us | intel_runtime_pm_acquire(); 0.259 us | } __i915_active_wait() { 0.085 us | i915_active_acquire_if_busy(); 0.240 us | } __i915_vma_evict() { ggtt_unbind_vma() { gen8_ggtt_clear_range() { 10507.255 us | } 10507.689 us | } 10508.516 us | }

v2: Instead of using bigjoiner checks, determine whether a scanout buffer is too big by checking to see if it is possible to map two of them into the ggtt.

v3 (Ville):

• Count how many fb objects can be fit into the available holes instead of checking for a hole twice the object size.
• Take alignment constraints into account.
• Limit this large scanout buffer check to >= Gen 11 platforms.

v4:

• Remove existing heuristic that checks just for size. (Ville)
• Return early if we find space to map at-least two objects. (Tvrtko)
• Slightly update the commit message.

v5: (Tvrtko)

• Rename the function to indicate that the object may be too big to map into the aperture.
• Account for guard pages while calculating the total size required for the object.
• Do not subject all objects to the heuristic check and instead consider objects only of a certain size.
• Do the hole walk using the rbtree.
• Preserve the existing PIN_NONBLOCK logic.
• Drop the PIN_MAPPABLE check while pinning the VMA.

v6: (Tvrtko)

• Return 0 on success and the specific error code on failure to preserve the existing behavior.

v7: (Ville)

• Drop the HAS_GMCH(i915), DISPLAY_VER(i915) < 11 and size < ggtt->mappable_end / 4 checks.
• Drop the redundant check that is based on previous heuristic.

v8:

• Make sure that we are holding the mutex associated with ggtt vm as we traverse the hole nodes.

v9: (Tvrtko)

• Use mutex_lock_interruptible_nested() instead of mutex_lock().

Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Cc: Tvrtko Ursulin tvrtko.ursulin@linux.intel.com Cc: Manasi Navare manasi.d.navare@intel.com Reviewed-by: Tvrtko Ursulin tvrtko.ursulin@intel.com Signed-off-by: Vivek Kasireddy vivek.kasireddy@intel.com

---

drivers/gpu/drm/i915/i915_gem.c | 128 +++++++++++++++++++++++--------- 1 file changed, 94 insertions(+), 34 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 9747924cc57b..e0d731b3f215 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -49,6 +49,7 @@ #include "gem/i915_gem_pm.h" #include "gem/i915_gem_region.h" #include "gem/i915_gem_userptr.h" +#include "gem/i915_gem_tiling.h" #include "gt/intel_engine_user.h" #include "gt/intel_gt.h" #include "gt/intel_gt_pm.h" @@ -882,6 +883,96 @@ static void discard_ggtt_vma(struct i915_vma *vma) spin_unlock(&obj->vma.lock); }

+static int +i915_gem_object_fits_in_aperture(struct drm_i915_gem_object *obj,

•                            u64 alignment, u64 flags)
  

Tvrtko asked me to ack the first patch, but then I looked at this and started wondering.

Conceptually this doesn't pass the smell test. What if we have multiple per-crtc buffers? Multiple planes on the same crtc? What if the app does triple buffer? You'll be forever busy tuning this heuristics, which can't fundamentally be fixed I think. The old "half of mappable" heuristic isn't really better, but at least it was dead simple.

Imo what we need here is a change in approach:

1. Check whether the useable view for scanout exists already. If yes,

use that. This should avoid the constant unbinding stalls. 2. Try to in buffer to mappabley, but without evicting anything (so not the non-blocking thing) 3. Pin the buffer with the most lenient approach

Even the non-blocking interim stage is dangerous, since it'll just result in other buffers (e.g. when triple-buffering) getting unbound and we're back to the same stall. Note that this could have an impact on cpu rendering compositors, where we might end up relying a lot more partial views. But as long as we are a tad more aggressive (i.e. the non-blocking binding) in the mmap path that should work out to keep everything balanced, since usually you render first before you display anything. And so the buffer should end up in the ideal place.

I'd try to first skip the 2. step since I think it'll require a bit of work, and frankly I don't think we care about the potential fallout.

To be sure I understand, you propose to stop trying to pin mappable by default. Ie. stop respecting this comment from i915_gem_object_pin_to_display_plane:

/* * As the user may map the buffer once pinned in the display plane * (e.g. libkms for the bootup splash), we have to ensure that we * always use map_and_fenceable for all scanout buffers. However, * it may simply be too big to fit into mappable, in which case * put it anyway and hope that userspace can cope (but always first * try to preserve the existing ABI). */

[Kasireddy, Vivek] Digging further, this is what the commit message that added the above comment says: commit 2efb813d5388e18255c54afac77bd91acd586908 Author: Chris Wilson chris@chris-wilson.co.uk Date: Thu Aug 18 17:17:06 2016 +0100

drm/i915: Fallback to using unmappable memory for scanout

The existing ABI says that scanouts are pinned into the mappable region so that legacy clients (e.g. old Xorg or plymouthd) can write directly into the scanout through a GTT mapping. However if the surface does not fit into the mappable region, we are better off just trying to fit it anywhere and hoping for the best. (Any userspace that is capable of using ginormous scanouts is also likely not to rely on pure GTT updates.) With the partial vma fault support, we are no longer restricted to only using scanouts that we can pin (though it is still preferred for performance reasons and for powersaving features like FBC).

By a quick look, for this case it appears we would end up creating partial views for CPU access (since the normal mapping would be busy/unpinnable). Worst case for this is to create a bunch of 1MiB VMAs so something to check would be how long those persist in memory before they get released. Or perhaps the bootup splash use case is not common these days?

[Kasireddy, Vivek] AFAIK, Plymouth is still the default bootup splash service on Fedora, Ubuntu and most other distributions. And, I took a quick look at it and IIUC, it (Plymouth's drm plugin) seems to create a dumb FB, mmap and update it via the dirty_fb ioctl. This would not to be a problem on ADL-S where there is space in mappable for one 8K FB.

Given this, do you think it would work if we just preserve the existing behavior and tweak the heuristic introduced in this patch to look for space in aperture for only one FB instead of two? Or, is there no good option for solving this issue other than to create 1MB VMAs?

Thanks, Vivek

Regards,

Tvrtko

...

-Daniel

...

+{

•   struct drm_i915_private *i915 = to_i915(obj->base.dev);
  

•   struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
  

•   struct drm_mm_node *hole;
  

•   u64 hole_start, hole_end, start, end;
  

•   u64 fence_size, fence_alignment;
  

•   unsigned int count = 0;
  

•   int err = 0;
  

•   /*
  

•    * If the required space is larger than the available
  

•    * aperture, we will not able to find a slot for the
  

•    * object and unbinding the object now will be in
  

•    * vain. Worse, doing so may cause us to ping-pong
  

•    * the object in and out of the Global GTT and
  

•    * waste a lot of cycles under the mutex.
  

•    */
  

•   if (obj->base.size > ggtt->mappable_end)
  

•           return -E2BIG;
  

•   /*
  

•    * If NONBLOCK is set the caller is optimistically
  

•    * trying to cache the full object within the mappable
  

•    * aperture, and *must* have a fallback in place for
  

•    * situations where we cannot bind the object. We
  

•    * can be a little more lax here and use the fallback
  

•    * more often to avoid costly migrations of ourselves
  

•    * and other objects within the aperture.
  

•    */
  

•   if (!(flags & PIN_NONBLOCK))
  

•           return 0;
  

•   /*
  

•    * Other objects such as batchbuffers are fairly small compared
  

•    * to FBs and are unlikely to exahust the aperture space.
  

•    * Therefore, return early if this obj is not an FB.
  

•    */
  

•   if (!i915_gem_object_is_framebuffer(obj))
  

•           return 0;
  

•   fence_size = i915_gem_fence_size(i915, obj->base.size,
  

•                                    i915_gem_object_get_tiling(obj),
  

•                                    i915_gem_object_get_stride(obj));
  

•   if (i915_vm_has_cache_coloring(&ggtt->vm))
  

•           fence_size += 2 * I915_GTT_PAGE_SIZE;
  

•   fence_alignment = i915_gem_fence_alignment(i915, obj->base.size,
  

•                                              i915_gem_object_get_tiling(obj),
  

•                                              i915_gem_object_get_stride(obj));
  

•   alignment = max_t(u64, alignment, fence_alignment);
  

•   err = mutex_lock_interruptible_nested(&ggtt->vm.mutex, 0);
  

•   if (err)
  

•           return err;
  

•   /*
  

•    * Assuming this object is a large scanout buffer, we try to find
  

•    * out if there is room to map at-least two of them. There could
  

•    * be space available to map one but to be consistent, we try to
  

•    * avoid mapping/fencing any of them.
  

•    */
  

•   drm_mm_for_each_suitable_hole(hole, &ggtt->vm.mm, 0, ggtt->mappable_end,
  

•                                 fence_size, DRM_MM_INSERT_LOW) {
  

•           hole_start = drm_mm_hole_node_start(hole);
  

•           hole_end = hole_start + hole->hole_size;
  

•           do {
  

•                   start = round_up(hole_start, alignment);
  

•                   end = min_t(u64, hole_end, ggtt->mappable_end);
  

•                   if (range_overflows(start, fence_size, end))
  

•                           break;
  

•                   if (++count >= 2) {
  

•                           mutex_unlock(&ggtt->vm.mutex);
  

•                           return 0;
  

•                   }
  

•                   hole_start = start + fence_size;
  

•           } while (1);
  

•   }
  

•   mutex_unlock(&ggtt->vm.mutex);
  

•   return -ENOSPC;
  

+}

• struct i915_vma * i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object *obj, struct i915_gem_ww_ctx *ww,

@@ -897,36 +988,9 @@ i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object

*obj,

...

...

     if (flags & PIN_MAPPABLE &&
         (!view || view->type == I915_GGTT_VIEW_NORMAL)) {

•           /*
  

•            * If the required space is larger than the available
  

•            * aperture, we will not able to find a slot for the
  

•            * object and unbinding the object now will be in
  

•            * vain. Worse, doing so may cause us to ping-pong
  

•            * the object in and out of the Global GTT and
  

•            * waste a lot of cycles under the mutex.
  

•            */
  

•           if (obj->base.size > ggtt->mappable_end)
  

•                   return ERR_PTR(-E2BIG);
  

•           /*
  

•            * If NONBLOCK is set the caller is optimistically
  

•            * trying to cache the full object within the mappable
  

•            * aperture, and *must* have a fallback in place for
  

•            * situations where we cannot bind the object. We
  

•            * can be a little more lax here and use the fallback
  

•            * more often to avoid costly migrations of ourselves
  

•            * and other objects within the aperture.
  

•            *
  

•            * Half-the-aperture is used as a simple heuristic.
  

•            * More interesting would to do search for a free
  

•            * block prior to making the commitment to unbind.
  

•            * That caters for the self-harm case, and with a
  

•            * little more heuristics (e.g. NOFAULT, NOEVICT)
  

•            * we could try to minimise harm to others.
  

•            */
  

•           if (flags & PIN_NONBLOCK &&
  

•               obj->base.size > ggtt->mappable_end / 2)
  

•                   return ERR_PTR(-ENOSPC);
  

•           ret = i915_gem_object_fits_in_aperture(obj, alignment, flags);
  

•           if (ret)
  

•                   return ERR_PTR(ret);
     }
  
  

  new_vma:

@@ -938,10 +1002,6 @@ i915_gem_object_ggtt_pin_ww(struct

drm_i915_gem_object *obj,

...

...

             if (flags & PIN_NONBLOCK) {
                     if (i915_vma_is_pinned(vma) || i915_vma_is_active(vma))
                             return ERR_PTR(-ENOSPC);

•                   if (flags & PIN_MAPPABLE &&
  

•                       vma->fence_size > ggtt->mappable_end / 2)
  

•                           return ERR_PTR(-ENOSPC);
             }
  
             if (i915_vma_is_pinned(vma) || i915_vma_is_active(vma)) {
  

-- 2.35.1

Hi Tvrtko,

On 15/03/2022 07:28, Kasireddy, Vivek wrote:

...

Hi Tvrtko, Daniel,

...

On 11/03/2022 09:39, Daniel Vetter wrote:

...

On Mon, 7 Mar 2022 at 21:38, Vivek Kasireddy vivek.kasireddy@intel.com wrote:

...

On platforms capable of allowing 8K (7680 x 4320) modes, pinning 2 or more framebuffers/scanout buffers results in only one that is mappable/ fenceable. Therefore, pageflipping between these 2 FBs where only one is mappable/fenceable creates latencies large enough to miss alternate vblanks thereby producing less optimal framerate.

This mainly happens because when i915_gem_object_pin_to_display_plane() is called to pin one of the FB objs, the associated vma is identified as misplaced and therefore i915_vma_unbind() is called which unbinds and evicts it. This misplaced vma gets subseqently pinned only when i915_gem_object_ggtt_pin_ww() is called without PIN_MAPPABLE. This results in a latency of ~10ms and happens every other vblank/repaint cycle. Therefore, to fix this issue, we try to see if there is space to map at-least two objects of a given size and return early if there isn't. This would ensure that we do not try with PIN_MAPPABLE for any objects that are too big to map thereby preventing unncessary unbind.

Testcase: Running Weston and weston-simple-egl on an Alderlake_S (ADLS) platform with a 8K@60 mode results in only ~40 FPS. Since upstream Weston submits a frame ~7ms before the next vblank, the latencies seen between atomic commit and flip event are 7, 24 (7 + 16.66), 7, 24..... suggesting that it misses the vblank every other frame.

Here is the ftrace snippet that shows the source of the ~10ms latency: i915_gem_object_pin_to_display_plane() { 0.102 us | i915_gem_object_set_cache_level(); i915_gem_object_ggtt_pin_ww() { 0.390 us | i915_vma_instance(); 0.178 us | i915_vma_misplaced(); i915_vma_unbind() { __i915_active_wait() { 0.082 us | i915_active_acquire_if_busy(); 0.475 us | } intel_runtime_pm_get() { 0.087 us | intel_runtime_pm_acquire(); 0.259 us | } __i915_active_wait() { 0.085 us | i915_active_acquire_if_busy(); 0.240 us | } __i915_vma_evict() { ggtt_unbind_vma() { gen8_ggtt_clear_range() { 10507.255 us | } 10507.689 us | } 10508.516 us | }

v2: Instead of using bigjoiner checks, determine whether a scanout buffer is too big by checking to see if it is possible to map two of them into the ggtt.

v3 (Ville):

• Count how many fb objects can be fit into the available holes instead of checking for a hole twice the object size.
• Take alignment constraints into account.
• Limit this large scanout buffer check to >= Gen 11 platforms.

v4:

• Remove existing heuristic that checks just for size. (Ville)
• Return early if we find space to map at-least two objects. (Tvrtko)
• Slightly update the commit message.

v5: (Tvrtko)

• Rename the function to indicate that the object may be too big to map into the aperture.
• Account for guard pages while calculating the total size required for the object.
• Do not subject all objects to the heuristic check and instead consider objects only of a certain size.
• Do the hole walk using the rbtree.
• Preserve the existing PIN_NONBLOCK logic.
• Drop the PIN_MAPPABLE check while pinning the VMA.

v6: (Tvrtko)

• Return 0 on success and the specific error code on failure to preserve the existing behavior.

v7: (Ville)

• Drop the HAS_GMCH(i915), DISPLAY_VER(i915) < 11 and size < ggtt->mappable_end / 4 checks.
• Drop the redundant check that is based on previous heuristic.

v8:

• Make sure that we are holding the mutex associated with ggtt vm as we traverse the hole nodes.

v9: (Tvrtko)

• Use mutex_lock_interruptible_nested() instead of mutex_lock().

Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Cc: Tvrtko Ursulin tvrtko.ursulin@linux.intel.com Cc: Manasi Navare manasi.d.navare@intel.com Reviewed-by: Tvrtko Ursulin tvrtko.ursulin@intel.com Signed-off-by: Vivek Kasireddy vivek.kasireddy@intel.com

---

drivers/gpu/drm/i915/i915_gem.c | 128 +++++++++++++++++++++++--------- 1 file changed, 94 insertions(+), 34 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 9747924cc57b..e0d731b3f215 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -49,6 +49,7 @@ #include "gem/i915_gem_pm.h" #include "gem/i915_gem_region.h" #include "gem/i915_gem_userptr.h" +#include "gem/i915_gem_tiling.h" #include "gt/intel_engine_user.h" #include "gt/intel_gt.h" #include "gt/intel_gt_pm.h" @@ -882,6 +883,96 @@ static void discard_ggtt_vma(struct i915_vma *vma) spin_unlock(&obj->vma.lock); }

+static int +i915_gem_object_fits_in_aperture(struct drm_i915_gem_object *obj,

•                            u64 alignment, u64 flags)
  

Tvrtko asked me to ack the first patch, but then I looked at this and started wondering.

Conceptually this doesn't pass the smell test. What if we have multiple per-crtc buffers? Multiple planes on the same crtc? What if the app does triple buffer? You'll be forever busy tuning this heuristics, which can't fundamentally be fixed I think. The old "half of mappable" heuristic isn't really better, but at least it was dead simple.

Imo what we need here is a change in approach:

1. Check whether the useable view for scanout exists already. If yes,

use that. This should avoid the constant unbinding stalls. 2. Try to in buffer to mappabley, but without evicting anything (so not the non-blocking thing) 3. Pin the buffer with the most lenient approach

Even the non-blocking interim stage is dangerous, since it'll just result in other buffers (e.g. when triple-buffering) getting unbound and we're back to the same stall. Note that this could have an impact on cpu rendering compositors, where we might end up relying a lot more partial views. But as long as we are a tad more aggressive (i.e. the non-blocking binding) in the mmap path that should work out to keep everything balanced, since usually you render first before you display anything. And so the buffer should end up in the ideal place.

I'd try to first skip the 2. step since I think it'll require a bit of work, and frankly I don't think we care about the potential fallout.

To be sure I understand, you propose to stop trying to pin mappable by default. Ie. stop respecting this comment from i915_gem_object_pin_to_display_plane:

/* * As the user may map the buffer once pinned in the display plane * (e.g. libkms for the bootup splash), we have to ensure that we * always use map_and_fenceable for all scanout buffers. However, * it may simply be too big to fit into mappable, in which case * put it anyway and hope that userspace can cope (but always first * try to preserve the existing ABI). */

[Kasireddy, Vivek] Digging further, this is what the commit message that added the above comment says: commit 2efb813d5388e18255c54afac77bd91acd586908 Author: Chris Wilson chris@chris-wilson.co.uk Date: Thu Aug 18 17:17:06 2016 +0100

 drm/i915: Fallback to using unmappable memory for scanout

 The existing ABI says that scanouts are pinned into the mappable region
 so that legacy clients (e.g. old Xorg or plymouthd) can write directly
 into the scanout through a GTT mapping. However if the surface does not
 fit into the mappable region, we are better off just trying to fit it
 anywhere and hoping for the best. (Any userspace that is capable of
 using ginormous scanouts is also likely not to rely on pure GTT
 updates.) With the partial vma fault support, we are no longer
 restricted to only using scanouts that we can pin (though it is still
 preferred for performance reasons and for powersaving features like
 FBC).

...

By a quick look, for this case it appears we would end up creating partial views for

CPU

...

...

access (since the normal mapping would be busy/unpinnable). Worst case for this is to create a bunch of 1MiB VMAs so something to check would be how long those persist

in

...

...

memory before they get released. Or perhaps the bootup splash use case is not common these days?

[Kasireddy, Vivek] AFAIK, Plymouth is still the default bootup splash service on Fedora, Ubuntu and most other distributions. And, I took a quick look at it and IIUC, it

(Plymouth's

...

drm plugin) seems to create a dumb FB, mmap and update it via the dirty_fb ioctl. This would not to be a problem on ADL-S where there is space in mappable for one 8K FB.

FBC is a good point - correct me if I am wrong, but if we dropped trying to map in aperture by default it looks like we would lose it and that would be a significant power regression. In which case it doesn't seem like that would be an option.

[Kasireddy, Vivek] Ok, makes sense.

Which I think leaves us with _some_ heuristics in any case.

1. N-holes heuristics.

2. Don't ever try PIN_MAPPABLE for framebuffers larger than some

percentage of aperture.

Could this solve the 8k issue, most of the time, maybe? Could the current "aperture / 2" test be expressed generically in some terms? Like "(aperture - 10% (or some absolute value)) / 2" to account for non-fb objects? I forgot what you said the relationship between aperture size and 8k fb size was.

3. Don't evict for PIN_MAPPABLE mismatches when

i915_gem_object_ggtt_pin_ww->i915_vma_misplaced is called on behalf of i915_gem_object_pin_to_display_plane. Assumption being if we ended up with a non-mappable fb to start with, we must not try to re-bind it or we risk ping-pong latencies.

The last would I guess need to distinguish between PIN_MAPPABLE passed in versus opportunistically added by i915_gem_object_pin_to_display_plane.

How intrusive would it be to implement this option I am not sure without trying myself.

[Kasireddy, Vivek] I suspect I might be missing something, but could it not be as simple as below: @@ -940,7 +940,8 @@ i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object *obj, return ERR_PTR(-ENOSPC);

if (flags & PIN_MAPPABLE && - vma->fence_size > ggtt->mappable_end / 2) + (vma->fence_size > ggtt->mappable_end / 2 || + !i915_vma_is_map_and_fenceable(vma))) return ERR_PTR(-ENOSPC); }

...

Given this, do you think it would work if we just preserve the existing behavior and tweak the heuristic introduced in this patch to look for space in aperture for only one FB instead of two? Or, is there no good option for solving this issue other than to create 1MB VMAs?

I did not get how having one hole would solve the issue. Wouldn't it still hit the re-bind ping-pong? Or there isn't even a single hole for 8k fb typically?

[Kasireddy, Vivek] IIUC, Mesa gives Weston a max of 4 backbuffers but it almost always uses only 2 except when it needs to share the FB -- with a plugin such as "remoting" for desktop streaming. Given the common use-case, lets assume there are two 8K FBs: FB1 and FB2 FB1 is mappable/fenceable and therefore not misplaced. FB2 is NOT mappable and hence identified as misplaced (because it fails the check (flags & PIN_MAPPABLE && !i915_vma_is_map_and_fenceable(vma))

As you suggest in 3) above the goal is to ensure that FB2 does not get evicted when we try to pin with PIN_MAPABBLE -- after it gets identified as misplaced. Or, alternatively, when we pin with PIN_MAPABBLE, we could just check to see if there is space in aperture for only FB2 (N = 1) and return early -- before even getting to i915_vma_misplaced(). As you can see, we avoid the ping-pong issue in both these cases.

The current version of this patch -- when running Weston -- puts both FB1 and FB2 (N = 2) outside of aperture although there may be space for FB1. I don't think this makes sense anymore given Plymouth's single-buffer use-case that uses dirtyfb ioctl.

Thanks, Vivek

Regards,

Tvrtko

Hi Tvrtko,

On 16/03/2022 07:37, Kasireddy, Vivek wrote:

...

Hi Tvrtko,

...

On 15/03/2022 07:28, Kasireddy, Vivek wrote:

...

Hi Tvrtko, Daniel,

...

On 11/03/2022 09:39, Daniel Vetter wrote:

...

On Mon, 7 Mar 2022 at 21:38, Vivek Kasireddy vivek.kasireddy@intel.com

wrote:

...

...

...

...

...

> > On platforms capable of allowing 8K (7680 x 4320) modes, pinning 2 or > more framebuffers/scanout buffers results in only one that is mappable/ > fenceable. Therefore, pageflipping between these 2 FBs where only one > is mappable/fenceable creates latencies large enough to miss alternate > vblanks thereby producing less optimal framerate. > > This mainly happens because when i915_gem_object_pin_to_display_plane() > is called to pin one of the FB objs, the associated vma is identified > as misplaced and therefore i915_vma_unbind() is called which unbinds and > evicts it. This misplaced vma gets subseqently pinned only when > i915_gem_object_ggtt_pin_ww() is called without PIN_MAPPABLE. This > results in a latency of ~10ms and happens every other vblank/repaint cycle. > Therefore, to fix this issue, we try to see if there is space to map > at-least two objects of a given size and return early if there isn't. This > would ensure that we do not try with PIN_MAPPABLE for any objects that > are too big to map thereby preventing unncessary unbind. > > Testcase: > Running Weston and weston-simple-egl on an Alderlake_S (ADLS) platform > with a 8K@60 mode results in only ~40 FPS. Since upstream Weston submits > a frame ~7ms before the next vblank, the latencies seen between atomic > commit and flip event are 7, 24 (7 + 16.66), 7, 24..... suggesting that > it misses the vblank every other frame. > > Here is the ftrace snippet that shows the source of the ~10ms latency: > i915_gem_object_pin_to_display_plane() { > 0.102 us | i915_gem_object_set_cache_level(); > i915_gem_object_ggtt_pin_ww() { > 0.390 us | i915_vma_instance(); > 0.178 us | i915_vma_misplaced(); > i915_vma_unbind() { > __i915_active_wait() { > 0.082 us | i915_active_acquire_if_busy(); > 0.475 us | } > intel_runtime_pm_get() { > 0.087 us | intel_runtime_pm_acquire(); > 0.259 us | } > __i915_active_wait() { > 0.085 us | i915_active_acquire_if_busy(); > 0.240 us | } > __i915_vma_evict() { > ggtt_unbind_vma() { > gen8_ggtt_clear_range() { > 10507.255 us | } > 10507.689 us | } > 10508.516 us | } > > v2: Instead of using bigjoiner checks, determine whether a scanout > buffer is too big by checking to see if it is possible to map > two of them into the ggtt. > > v3 (Ville): > - Count how many fb objects can be fit into the available holes > instead of checking for a hole twice the object size. > - Take alignment constraints into account. > - Limit this large scanout buffer check to >= Gen 11 platforms. > > v4: > - Remove existing heuristic that checks just for size. (Ville) > - Return early if we find space to map at-least two objects. (Tvrtko) > - Slightly update the commit message. > > v5: (Tvrtko) > - Rename the function to indicate that the object may be too big to > map into the aperture. > - Account for guard pages while calculating the total size required > for the object. > - Do not subject all objects to the heuristic check and instead > consider objects only of a certain size. > - Do the hole walk using the rbtree. > - Preserve the existing PIN_NONBLOCK logic. > - Drop the PIN_MAPPABLE check while pinning the VMA. > > v6: (Tvrtko) > - Return 0 on success and the specific error code on failure to > preserve the existing behavior. > > v7: (Ville) > - Drop the HAS_GMCH(i915), DISPLAY_VER(i915) < 11 and > size < ggtt->mappable_end / 4 checks. > - Drop the redundant check that is based on previous heuristic. > > v8: > - Make sure that we are holding the mutex associated with ggtt vm > as we traverse the hole nodes. > > v9: (Tvrtko) > - Use mutex_lock_interruptible_nested() instead of mutex_lock(). > > Cc: Ville Syrjälä ville.syrjala@linux.intel.com > Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com > Cc: Tvrtko Ursulin tvrtko.ursulin@linux.intel.com > Cc: Manasi Navare manasi.d.navare@intel.com > Reviewed-by: Tvrtko Ursulin tvrtko.ursulin@intel.com > Signed-off-by: Vivek Kasireddy vivek.kasireddy@intel.com > --- > drivers/gpu/drm/i915/i915_gem.c | 128 +++++++++++++++++++++++--------

...

...

...

...

...

> 1 file changed, 94 insertions(+), 34 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > index 9747924cc57b..e0d731b3f215 100644 > --- a/drivers/gpu/drm/i915/i915_gem.c > +++ b/drivers/gpu/drm/i915/i915_gem.c > @@ -49,6 +49,7 @@ > #include "gem/i915_gem_pm.h" > #include "gem/i915_gem_region.h" > #include "gem/i915_gem_userptr.h" > +#include "gem/i915_gem_tiling.h" > #include "gt/intel_engine_user.h" > #include "gt/intel_gt.h" > #include "gt/intel_gt_pm.h" > @@ -882,6 +883,96 @@ static void discard_ggtt_vma(struct i915_vma *vma) > spin_unlock(&obj->vma.lock); > } > > +static int > +i915_gem_object_fits_in_aperture(struct drm_i915_gem_object *obj, > + u64 alignment, u64 flags)

Tvrtko asked me to ack the first patch, but then I looked at this and started wondering.

Conceptually this doesn't pass the smell test. What if we have multiple per-crtc buffers? Multiple planes on the same crtc? What if the app does triple buffer? You'll be forever busy tuning this heuristics, which can't fundamentally be fixed I think. The old "half of mappable" heuristic isn't really better, but at least it was dead simple.

Imo what we need here is a change in approach:

1. Check whether the useable view for scanout exists already. If yes,

use that. This should avoid the constant unbinding stalls. 2. Try to in buffer to mappabley, but without evicting anything (so not the non-blocking thing) 3. Pin the buffer with the most lenient approach

Even the non-blocking interim stage is dangerous, since it'll just result in other buffers (e.g. when triple-buffering) getting unbound and we're back to the same stall. Note that this could have an impact on cpu rendering compositors, where we might end up relying a lot more partial views. But as long as we are a tad more aggressive (i.e. the non-blocking binding) in the mmap path that should work out to keep everything balanced, since usually you render first before you display anything. And so the buffer should end up in the ideal place.

I'd try to first skip the 2. step since I think it'll require a bit of work, and frankly I don't think we care about the potential fallout.

To be sure I understand, you propose to stop trying to pin mappable by default. Ie.

stop

...

...

...

...

respecting this comment from i915_gem_object_pin_to_display_plane:

/* * As the user may map the buffer once pinned in the display plane * (e.g. libkms for the bootup splash), we have to ensure that we * always use map_and_fenceable for all scanout buffers. However, * it may simply be too big to fit into mappable, in which case * put it anyway and hope that userspace can cope (but always first * try to preserve the existing ABI). */

[Kasireddy, Vivek] Digging further, this is what the commit message that added the above comment says: commit 2efb813d5388e18255c54afac77bd91acd586908 Author: Chris Wilson chris@chris-wilson.co.uk Date: Thu Aug 18 17:17:06 2016 +0100

  drm/i915: Fallback to using unmappable memory for scanout

  The existing ABI says that scanouts are pinned into the mappable region
  so that legacy clients (e.g. old Xorg or plymouthd) can write directly
  into the scanout through a GTT mapping. However if the surface does not
  fit into the mappable region, we are better off just trying to fit it
  anywhere and hoping for the best. (Any userspace that is capable of
  using ginormous scanouts is also likely not to rely on pure GTT
  updates.) With the partial vma fault support, we are no longer
  restricted to only using scanouts that we can pin (though it is still
  preferred for performance reasons and for powersaving features like
  FBC).

...

By a quick look, for this case it appears we would end up creating partial views for

CPU

...

...

access (since the normal mapping would be busy/unpinnable). Worst case for this is

to

...

...

...

...

create a bunch of 1MiB VMAs so something to check would be how long those

persist

...

...

in

...

...

memory before they get released. Or perhaps the bootup splash use case is not

common

...

...

...

...

these days?

[Kasireddy, Vivek] AFAIK, Plymouth is still the default bootup splash service on

Fedora,

...

...

...

Ubuntu and most other distributions. And, I took a quick look at it and IIUC, it

(Plymouth's

...

drm plugin) seems to create a dumb FB, mmap and update it via the dirty_fb ioctl.

This

...

...

...

would not to be a problem on ADL-S where there is space in mappable for one 8K FB.

FBC is a good point - correct me if I am wrong, but if we dropped trying to map in aperture by default it looks like we would lose it and that would be a significant power regression. In which case it doesn't seem like that would be an option.

[Kasireddy, Vivek] Ok, makes sense.

...

Which I think leaves us with _some_ heuristics in any case.

1. N-holes heuristics.

2. Don't ever try PIN_MAPPABLE for framebuffers larger than some

percentage of aperture.

Could this solve the 8k issue, most of the time, maybe? Could the current "aperture / 2" test be expressed generically in some terms? Like "(aperture - 10% (or some absolute value)) / 2" to account for non-fb objects? I forgot what you said the relationship between aperture size and 8k fb size was.

3. Don't evict for PIN_MAPPABLE mismatches when

i915_gem_object_ggtt_pin_ww->i915_vma_misplaced is called on behalf of i915_gem_object_pin_to_display_plane. Assumption being if we ended up with a non-mappable fb to start with, we must not try to re-bind it or we risk ping-pong latencies.

The last would I guess need to distinguish between PIN_MAPPABLE passed in versus opportunistically added by i915_gem_object_pin_to_display_plane.

How intrusive would it be to implement this option I am not sure without trying myself.

[Kasireddy, Vivek] I suspect I might be missing something, but could it not be as simple as below: @@ -940,7 +940,8 @@ i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object

*obj,

...

                             return ERR_PTR(-ENOSPC);

                     if (flags & PIN_MAPPABLE &&

•                       vma->fence_size > ggtt->mappable_end / 2)
  

•                       (vma->fence_size > ggtt->mappable_end / 2 ||
  

•                       !i915_vma_is_map_and_fenceable(vma)))
                                 return ERR_PTR(-ENOSPC);
             }
  

Looks like this would work...

...

...

...

Given this, do you think it would work if we just preserve the existing behavior and tweak the heuristic introduced in this patch to look for space in aperture for only one FB instead of two? Or, is there no good option for solving this issue other than to create 1MB VMAs?

I did not get how having one hole would solve the issue. Wouldn't it still hit the re-bind ping-pong? Or there isn't even a single hole for 8k fb typically?

[Kasireddy, Vivek] IIUC, Mesa gives Weston a max of 4 backbuffers but it almost always uses only 2 except when it needs to share the FB -- with a plugin such as "remoting" for desktop streaming. Given the common use-case, lets assume there are two 8K FBs: FB1 and FB2 FB1 is mappable/fenceable and therefore not misplaced. FB2 is NOT mappable and hence identified as misplaced (because it fails the check (flags & PIN_MAPPABLE && !i915_vma_is_map_and_fenceable(vma))

As you suggest in 3) above the goal is to ensure that FB2 does not get evicted when we try to pin with PIN_MAPABBLE -- after it gets identified as misplaced. Or, alternatively, when we pin with PIN_MAPABBLE, we could just check to see if there is space in aperture for only FB2 (N = 1) and return early -- before even getting to i915_vma_misplaced(). As you can see, we avoid the ping-pong issue in both these cases.

... got it, yes, it seems both approaches works for this use case.

Not sure that I have a preference between the two approaches at this point.

Both would be behind a "PIN_MAPPABLE && PIN_NONBLOCK" check, so both would only apply to opportunistic PIN_MAPPABLE attempts. That is, any caller who only passes PIN_MAPPABLE would be unaffected which is what we want.

The extra i915_vma_is_map_and_fenceable check I guess is simpler and self-contained. I assume you have a test setup and can try it out to check it really works?

[Kasireddy, Vivek] Yes, it works; my testcase just involves running Weston with a mode of 8K@60 on ADL-S and checking the FPS of the sample client weston-simple-egl. With the fix included, the perf improves to 59 FPS from 40 FPS. I'll send out a new patch for review soon.

Oh, btw, do you think it is now pointless to merge the drm/mm patch that adds the iterator given that we'd no longer have the i915 patch that uses it anymore?

Thanks, Vivek

...

The current version of this patch -- when running Weston -- puts both FB1 and FB2 (N = 2) outside of aperture although there may be space for FB1. I don't think this makes sense anymore given Plymouth's single-buffer use-case that uses dirtyfb ioctl.

Yes agreed, it sounds preferable to preserve the current behaviour there.

Regards,

Tvrtko

On Tue, Mar 15, 2022 at 09:45:20AM +0000, Tvrtko Ursulin wrote:

On 15/03/2022 07:28, Kasireddy, Vivek wrote:

...

Hi Tvrtko, Daniel,

...

On 11/03/2022 09:39, Daniel Vetter wrote:

...

On Mon, 7 Mar 2022 at 21:38, Vivek Kasireddy vivek.kasireddy@intel.com wrote:

...

On platforms capable of allowing 8K (7680 x 4320) modes, pinning 2 or more framebuffers/scanout buffers results in only one that is mappable/ fenceable. Therefore, pageflipping between these 2 FBs where only one is mappable/fenceable creates latencies large enough to miss alternate vblanks thereby producing less optimal framerate.

This mainly happens because when i915_gem_object_pin_to_display_plane() is called to pin one of the FB objs, the associated vma is identified as misplaced and therefore i915_vma_unbind() is called which unbinds and evicts it. This misplaced vma gets subseqently pinned only when i915_gem_object_ggtt_pin_ww() is called without PIN_MAPPABLE. This results in a latency of ~10ms and happens every other vblank/repaint cycle. Therefore, to fix this issue, we try to see if there is space to map at-least two objects of a given size and return early if there isn't. This would ensure that we do not try with PIN_MAPPABLE for any objects that are too big to map thereby preventing unncessary unbind.

Testcase: Running Weston and weston-simple-egl on an Alderlake_S (ADLS) platform with a 8K@60 mode results in only ~40 FPS. Since upstream Weston submits a frame ~7ms before the next vblank, the latencies seen between atomic commit and flip event are 7, 24 (7 + 16.66), 7, 24..... suggesting that it misses the vblank every other frame.

Here is the ftrace snippet that shows the source of the ~10ms latency: i915_gem_object_pin_to_display_plane() { 0.102 us | i915_gem_object_set_cache_level(); i915_gem_object_ggtt_pin_ww() { 0.390 us | i915_vma_instance(); 0.178 us | i915_vma_misplaced(); i915_vma_unbind() { __i915_active_wait() { 0.082 us | i915_active_acquire_if_busy(); 0.475 us | } intel_runtime_pm_get() { 0.087 us | intel_runtime_pm_acquire(); 0.259 us | } __i915_active_wait() { 0.085 us | i915_active_acquire_if_busy(); 0.240 us | } __i915_vma_evict() { ggtt_unbind_vma() { gen8_ggtt_clear_range() { 10507.255 us | } 10507.689 us | } 10508.516 us | }

v2: Instead of using bigjoiner checks, determine whether a scanout buffer is too big by checking to see if it is possible to map two of them into the ggtt.

v3 (Ville):

• Count how many fb objects can be fit into the available holes instead of checking for a hole twice the object size.
• Take alignment constraints into account.
• Limit this large scanout buffer check to >= Gen 11 platforms.

v4:

• Remove existing heuristic that checks just for size. (Ville)
• Return early if we find space to map at-least two objects. (Tvrtko)
• Slightly update the commit message.

v5: (Tvrtko)

• Rename the function to indicate that the object may be too big to map into the aperture.
• Account for guard pages while calculating the total size required for the object.
• Do not subject all objects to the heuristic check and instead consider objects only of a certain size.
• Do the hole walk using the rbtree.
• Preserve the existing PIN_NONBLOCK logic.
• Drop the PIN_MAPPABLE check while pinning the VMA.

v6: (Tvrtko)

• Return 0 on success and the specific error code on failure to preserve the existing behavior.

v7: (Ville)

• Drop the HAS_GMCH(i915), DISPLAY_VER(i915) < 11 and size < ggtt->mappable_end / 4 checks.
• Drop the redundant check that is based on previous heuristic.

v8:

• Make sure that we are holding the mutex associated with ggtt vm as we traverse the hole nodes.

v9: (Tvrtko)

• Use mutex_lock_interruptible_nested() instead of mutex_lock().

Cc: Ville Syrjälä ville.syrjala@linux.intel.com Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Cc: Tvrtko Ursulin tvrtko.ursulin@linux.intel.com Cc: Manasi Navare manasi.d.navare@intel.com Reviewed-by: Tvrtko Ursulin tvrtko.ursulin@intel.com Signed-off-by: Vivek Kasireddy vivek.kasireddy@intel.com

---

drivers/gpu/drm/i915/i915_gem.c | 128 +++++++++++++++++++++++--------- 1 file changed, 94 insertions(+), 34 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 9747924cc57b..e0d731b3f215 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -49,6 +49,7 @@ #include "gem/i915_gem_pm.h" #include "gem/i915_gem_region.h" #include "gem/i915_gem_userptr.h" +#include "gem/i915_gem_tiling.h" #include "gt/intel_engine_user.h" #include "gt/intel_gt.h" #include "gt/intel_gt_pm.h" @@ -882,6 +883,96 @@ static void discard_ggtt_vma(struct i915_vma *vma) spin_unlock(&obj->vma.lock); }

+static int +i915_gem_object_fits_in_aperture(struct drm_i915_gem_object *obj,

•                            u64 alignment, u64 flags)
  

Tvrtko asked me to ack the first patch, but then I looked at this and started wondering.

Conceptually this doesn't pass the smell test. What if we have multiple per-crtc buffers? Multiple planes on the same crtc? What if the app does triple buffer? You'll be forever busy tuning this heuristics, which can't fundamentally be fixed I think. The old "half of mappable" heuristic isn't really better, but at least it was dead simple.

Imo what we need here is a change in approach:

1. Check whether the useable view for scanout exists already. If yes,

use that. This should avoid the constant unbinding stalls. 2. Try to in buffer to mappabley, but without evicting anything (so not the non-blocking thing) 3. Pin the buffer with the most lenient approach

Even the non-blocking interim stage is dangerous, since it'll just result in other buffers (e.g. when triple-buffering) getting unbound and we're back to the same stall. Note that this could have an impact on cpu rendering compositors, where we might end up relying a lot more partial views. But as long as we are a tad more aggressive (i.e. the non-blocking binding) in the mmap path that should work out to keep everything balanced, since usually you render first before you display anything. And so the buffer should end up in the ideal place.

I'd try to first skip the 2. step since I think it'll require a bit of work, and frankly I don't think we care about the potential fallout.

To be sure I understand, you propose to stop trying to pin mappable by default. Ie. stop respecting this comment from i915_gem_object_pin_to_display_plane:

/* * As the user may map the buffer once pinned in the display plane * (e.g. libkms for the bootup splash), we have to ensure that we * always use map_and_fenceable for all scanout buffers. However, * it may simply be too big to fit into mappable, in which case * put it anyway and hope that userspace can cope (but always first * try to preserve the existing ABI). */

[Kasireddy, Vivek] Digging further, this is what the commit message that added the above comment says: commit 2efb813d5388e18255c54afac77bd91acd586908 Author: Chris Wilson chris@chris-wilson.co.uk Date: Thu Aug 18 17:17:06 2016 +0100

 drm/i915: Fallback to using unmappable memory for scanout

 The existing ABI says that scanouts are pinned into the mappable region
 so that legacy clients (e.g. old Xorg or plymouthd) can write directly
 into the scanout through a GTT mapping. However if the surface does not
 fit into the mappable region, we are better off just trying to fit it
 anywhere and hoping for the best. (Any userspace that is capable of
 using ginormous scanouts is also likely not to rely on pure GTT
 updates.) With the partial vma fault support, we are no longer
 restricted to only using scanouts that we can pin (though it is still
 preferred for performance reasons and for powersaving features like
 FBC).

...

By a quick look, for this case it appears we would end up creating partial views for CPU access (since the normal mapping would be busy/unpinnable). Worst case for this is to create a bunch of 1MiB VMAs so something to check would be how long those persist in memory before they get released. Or perhaps the bootup splash use case is not common these days?

[Kasireddy, Vivek] AFAIK, Plymouth is still the default bootup splash service on Fedora, Ubuntu and most other distributions. And, I took a quick look at it and IIUC, it (Plymouth's drm plugin) seems to create a dumb FB, mmap and update it via the dirty_fb ioctl. This would not to be a problem on ADL-S where there is space in mappable for one 8K FB.

FBC is a good point - correct me if I am wrong, but if we dropped trying to map in aperture by default it looks like we would lose it and that would be a significant power regression. In which case it doesn't seem like that would be an option.

FBC fence is only required for frontbuffer hw tracking, which is another thing that's somewhere between "meh" and "we should just sunset set it right away". I think that work has even been done.

So I wouldn't worry about this.

If you are worried, then I'd check with display folks whether we need a platform based cut-off for this heuristics.

Which I think leaves us with _some_ heuristics in any case.

1. N-holes heuristics.

2. Don't ever try PIN_MAPPABLE for framebuffers larger than some percentage

of aperture.

Could this solve the 8k issue, most of the time, maybe? Could the current "aperture / 2" test be expressed generically in some terms? Like "(aperture

• 10% (or some absolute value)) / 2" to account for non-fb objects? I forgot

what you said the relationship between aperture size and 8k fb size was.

3. Don't evict for PIN_MAPPABLE mismatches when

i915_gem_object_ggtt_pin_ww->i915_vma_misplaced is called on behalf of i915_gem_object_pin_to_display_plane. Assumption being if we ended up with a non-mappable fb to start with, we must not try to re-bind it or we risk ping-pong latencies.

The last would I guess need to distinguish between PIN_MAPPABLE passed in versus opportunistically added by i915_gem_object_pin_to_display_plane.

How intrusive would it be to implement this option I am not sure without trying myself.

This won't work, see my initial mail. All you need is triple buffering (or multiple per-crtc buffers that flip)

1. fb A gets pinned as mappable 2. fb B gets pinned as mappable, fb A is unpinned 3. fb C gets pinned as mappable, we don't have space and end up evicting fb A

Repeat, and you have exactly the same old eviction loop as with two buffers. Not good.

Therefore for this to work we don't just need to make sure that we don't move our own buffer, but also that we don't move any other buffer.

The downside of that is that if a buffer is ever misplaced as mappable, we never fix up that mistake (at least not until the application entirely destroys all the involved fb and bo). I think that's acceptable, but definitely deserves a comment.

Cheers, Daniel

...

Given this, do you think it would work if we just preserve the existing behavior and tweak the heuristic introduced in this patch to look for space in aperture for only one FB instead of two? Or, is there no good option for solving this issue other than to create 1MB VMAs?

I did not get how having one hole would solve the issue. Wouldn't it still hit the re-bind ping-pong? Or there isn't even a single hole for 8k fb typically?

Regards,

Tvrtko

On Thu, Mar 17, 2022 at 10:04:36AM +0000, Tvrtko Ursulin wrote:

On 17/03/2022 09:47, Daniel Vetter wrote:

...

On Tue, Mar 15, 2022 at 09:45:20AM +0000, Tvrtko Ursulin wrote:

...

On 15/03/2022 07:28, Kasireddy, Vivek wrote:

...

Hi Tvrtko, Daniel,

...

On 11/03/2022 09:39, Daniel Vetter wrote:

...

On Mon, 7 Mar 2022 at 21:38, Vivek Kasireddy vivek.kasireddy@intel.com wrote: > > On platforms capable of allowing 8K (7680 x 4320) modes, pinning 2 or > more framebuffers/scanout buffers results in only one that is mappable/ > fenceable. Therefore, pageflipping between these 2 FBs where only one > is mappable/fenceable creates latencies large enough to miss alternate > vblanks thereby producing less optimal framerate. > > This mainly happens because when i915_gem_object_pin_to_display_plane() > is called to pin one of the FB objs, the associated vma is identified > as misplaced and therefore i915_vma_unbind() is called which unbinds and > evicts it. This misplaced vma gets subseqently pinned only when > i915_gem_object_ggtt_pin_ww() is called without PIN_MAPPABLE. This > results in a latency of ~10ms and happens every other vblank/repaint cycle. > Therefore, to fix this issue, we try to see if there is space to map > at-least two objects of a given size and return early if there isn't. This > would ensure that we do not try with PIN_MAPPABLE for any objects that > are too big to map thereby preventing unncessary unbind. > > Testcase: > Running Weston and weston-simple-egl on an Alderlake_S (ADLS) platform > with a 8K@60 mode results in only ~40 FPS. Since upstream Weston submits > a frame ~7ms before the next vblank, the latencies seen between atomic > commit and flip event are 7, 24 (7 + 16.66), 7, 24..... suggesting that > it misses the vblank every other frame. > > Here is the ftrace snippet that shows the source of the ~10ms latency: > i915_gem_object_pin_to_display_plane() { > 0.102 us | i915_gem_object_set_cache_level(); > i915_gem_object_ggtt_pin_ww() { > 0.390 us | i915_vma_instance(); > 0.178 us | i915_vma_misplaced(); > i915_vma_unbind() { > __i915_active_wait() { > 0.082 us | i915_active_acquire_if_busy(); > 0.475 us | } > intel_runtime_pm_get() { > 0.087 us | intel_runtime_pm_acquire(); > 0.259 us | } > __i915_active_wait() { > 0.085 us | i915_active_acquire_if_busy(); > 0.240 us | } > __i915_vma_evict() { > ggtt_unbind_vma() { > gen8_ggtt_clear_range() { > 10507.255 us | } > 10507.689 us | } > 10508.516 us | } > > v2: Instead of using bigjoiner checks, determine whether a scanout > buffer is too big by checking to see if it is possible to map > two of them into the ggtt. > > v3 (Ville): > - Count how many fb objects can be fit into the available holes > instead of checking for a hole twice the object size. > - Take alignment constraints into account. > - Limit this large scanout buffer check to >= Gen 11 platforms. > > v4: > - Remove existing heuristic that checks just for size. (Ville) > - Return early if we find space to map at-least two objects. (Tvrtko) > - Slightly update the commit message. > > v5: (Tvrtko) > - Rename the function to indicate that the object may be too big to > map into the aperture. > - Account for guard pages while calculating the total size required > for the object. > - Do not subject all objects to the heuristic check and instead > consider objects only of a certain size. > - Do the hole walk using the rbtree. > - Preserve the existing PIN_NONBLOCK logic. > - Drop the PIN_MAPPABLE check while pinning the VMA. > > v6: (Tvrtko) > - Return 0 on success and the specific error code on failure to > preserve the existing behavior. > > v7: (Ville) > - Drop the HAS_GMCH(i915), DISPLAY_VER(i915) < 11 and > size < ggtt->mappable_end / 4 checks. > - Drop the redundant check that is based on previous heuristic. > > v8: > - Make sure that we are holding the mutex associated with ggtt vm > as we traverse the hole nodes. > > v9: (Tvrtko) > - Use mutex_lock_interruptible_nested() instead of mutex_lock(). > > Cc: Ville Syrjälä ville.syrjala@linux.intel.com > Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com > Cc: Tvrtko Ursulin tvrtko.ursulin@linux.intel.com > Cc: Manasi Navare manasi.d.navare@intel.com > Reviewed-by: Tvrtko Ursulin tvrtko.ursulin@intel.com > Signed-off-by: Vivek Kasireddy vivek.kasireddy@intel.com > --- > drivers/gpu/drm/i915/i915_gem.c | 128 +++++++++++++++++++++++--------- > 1 file changed, 94 insertions(+), 34 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > index 9747924cc57b..e0d731b3f215 100644 > --- a/drivers/gpu/drm/i915/i915_gem.c > +++ b/drivers/gpu/drm/i915/i915_gem.c > @@ -49,6 +49,7 @@ > #include "gem/i915_gem_pm.h" > #include "gem/i915_gem_region.h" > #include "gem/i915_gem_userptr.h" > +#include "gem/i915_gem_tiling.h" > #include "gt/intel_engine_user.h" > #include "gt/intel_gt.h" > #include "gt/intel_gt_pm.h" > @@ -882,6 +883,96 @@ static void discard_ggtt_vma(struct i915_vma *vma) > spin_unlock(&obj->vma.lock); > } > > +static int > +i915_gem_object_fits_in_aperture(struct drm_i915_gem_object *obj, > + u64 alignment, u64 flags)

Tvrtko asked me to ack the first patch, but then I looked at this and started wondering.

Conceptually this doesn't pass the smell test. What if we have multiple per-crtc buffers? Multiple planes on the same crtc? What if the app does triple buffer? You'll be forever busy tuning this heuristics, which can't fundamentally be fixed I think. The old "half of mappable" heuristic isn't really better, but at least it was dead simple.

Imo what we need here is a change in approach:

1. Check whether the useable view for scanout exists already. If yes,

use that. This should avoid the constant unbinding stalls. 2. Try to in buffer to mappabley, but without evicting anything (so not the non-blocking thing) 3. Pin the buffer with the most lenient approach

Even the non-blocking interim stage is dangerous, since it'll just result in other buffers (e.g. when triple-buffering) getting unbound and we're back to the same stall. Note that this could have an impact on cpu rendering compositors, where we might end up relying a lot more partial views. But as long as we are a tad more aggressive (i.e. the non-blocking binding) in the mmap path that should work out to keep everything balanced, since usually you render first before you display anything. And so the buffer should end up in the ideal place.

I'd try to first skip the 2. step since I think it'll require a bit of work, and frankly I don't think we care about the potential fallout.

To be sure I understand, you propose to stop trying to pin mappable by default. Ie. stop respecting this comment from i915_gem_object_pin_to_display_plane:

/* * As the user may map the buffer once pinned in the display plane * (e.g. libkms for the bootup splash), we have to ensure that we * always use map_and_fenceable for all scanout buffers. However, * it may simply be too big to fit into mappable, in which case * put it anyway and hope that userspace can cope (but always first * try to preserve the existing ABI). */

[Kasireddy, Vivek] Digging further, this is what the commit message that added the above comment says: commit 2efb813d5388e18255c54afac77bd91acd586908 Author: Chris Wilson chris@chris-wilson.co.uk Date: Thu Aug 18 17:17:06 2016 +0100

  drm/i915: Fallback to using unmappable memory for scanout

  The existing ABI says that scanouts are pinned into the mappable region
  so that legacy clients (e.g. old Xorg or plymouthd) can write directly
  into the scanout through a GTT mapping. However if the surface does not
  fit into the mappable region, we are better off just trying to fit it
  anywhere and hoping for the best. (Any userspace that is capable of
  using ginormous scanouts is also likely not to rely on pure GTT
  updates.) With the partial vma fault support, we are no longer
  restricted to only using scanouts that we can pin (though it is still
  preferred for performance reasons and for powersaving features like
  FBC).

...

By a quick look, for this case it appears we would end up creating partial views for CPU access (since the normal mapping would be busy/unpinnable). Worst case for this is to create a bunch of 1MiB VMAs so something to check would be how long those persist in memory before they get released. Or perhaps the bootup splash use case is not common these days?

[Kasireddy, Vivek] AFAIK, Plymouth is still the default bootup splash service on Fedora, Ubuntu and most other distributions. And, I took a quick look at it and IIUC, it (Plymouth's drm plugin) seems to create a dumb FB, mmap and update it via the dirty_fb ioctl. This would not to be a problem on ADL-S where there is space in mappable for one 8K FB.

FBC is a good point - correct me if I am wrong, but if we dropped trying to map in aperture by default it looks like we would lose it and that would be a significant power regression. In which case it doesn't seem like that would be an option.

FBC fence is only required for frontbuffer hw tracking, which is another thing that's somewhere between "meh" and "we should just sunset set it right away". I think that work has even been done.

So I wouldn't worry about this.

If you are worried, then I'd check with display folks whether we need a platform based cut-off for this heuristics.

...

Which I think leaves us with _some_ heuristics in any case.

1. N-holes heuristics.

2. Don't ever try PIN_MAPPABLE for framebuffers larger than some percentage

of aperture.

Could this solve the 8k issue, most of the time, maybe? Could the current "aperture / 2" test be expressed generically in some terms? Like "(aperture

• 10% (or some absolute value)) / 2" to account for non-fb objects? I forgot

what you said the relationship between aperture size and 8k fb size was.

3. Don't evict for PIN_MAPPABLE mismatches when

i915_gem_object_ggtt_pin_ww->i915_vma_misplaced is called on behalf of i915_gem_object_pin_to_display_plane. Assumption being if we ended up with a non-mappable fb to start with, we must not try to re-bind it or we risk ping-pong latencies.

The last would I guess need to distinguish between PIN_MAPPABLE passed in versus opportunistically added by i915_gem_object_pin_to_display_plane.

How intrusive would it be to implement this option I am not sure without trying myself.

This won't work, see my initial mail. All you need is triple buffering (or multiple per-crtc buffers that flip)

I asked for clarifications on your initial email but you went a bit quiet on us, which is why I tried to drive this forward.

Yeah I'm burried on mailing list stuff pretty badly.

...

1. fb A gets pinned as mappable
2. fb B gets pinned as mappable, fb A is unpinned
3. fb C gets pinned as mappable, we don't have space and end up evicting

fb A

Repeat, and you have exactly the same old eviction loop as with two buffers. Not good.

Maybe a misunderstanding of what I wrote above? Idea was specifically not to evict for "opportunistic" PIN_MAPPABLE. Anyway, with the current solution to implement that, this is what would happen (see latest patch):

1. fb A get pinned as mappable
2. fb B gets pinned as mappable, assuming there is space, fb A unpinned
3. fb C, assuming there is no space, does not get pinned as mappable so

nothing is evicted

...

Therefore for this to work we don't just need to make sure that we don't move our own buffer, but also that we don't move any other buffer.

I think we achieved it by failing the "opportunistic" PIN_MAPPABLE attempts for all vmas which weren't already bound mappable in the past.

Ah yeah if that all works like that then I think we're fine. -Daniel

...

The downside of that is that if a buffer is ever misplaced as mappable, we never fix up that mistake (at least not until the application entirely destroys all the involved fb and bo). I think that's acceptable, but definitely deserves a comment.

This is true yes.

Regards,

Tvrtko