From: "Ville Syrjälä" <ville.syrjala@linux.intel.com>
To: "Murthy, Arun R" <arun.r.murthy@intel.com>
Cc: "intel-gfx@lists.freedesktop.org" <intel-gfx@lists.freedesktop.org>
Subject: Re: [PATCH 1/6] drm/i915: Align PLANE_SURF to 16k on ADL for async flips
Date: Fri, 19 Apr 2024 19:08:09 +0300 [thread overview]
Message-ID: <ZiKW6YK6mVf9rHIW@intel.com> (raw)
In-Reply-To: <IA0PR11MB7307385AC116A560D0E6466FBA0D2@IA0PR11MB7307.namprd11.prod.outlook.com>
On Fri, Apr 19, 2024 at 04:20:40AM +0000, Murthy, Arun R wrote:
>
> > -----Original Message-----
> > From: Intel-gfx <intel-gfx-bounces@lists.freedesktop.org> On Behalf Of Ville
> > Syrjala
> > Sent: Wednesday, March 20, 2024 9:34 PM
> > To: intel-gfx@lists.freedesktop.org
> > Subject: [PATCH 1/6] drm/i915: Align PLANE_SURF to 16k on ADL for async flips
> >
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> >
> > On ADL async flips apparently generate DMAR and GGTT faults (with
> > accompanying visual glitches) unless PLANE_SURF is aligned to at least 16k.
> > Bump up the alignment to 16k.
>
> I don’t find any such restriction in the spec. Can you please add link to the spec?
I don't think it's documented, hence the FIXME.
>
> Thanks and Regards,
> Arun R Murthy
> -------------------
> >
> > TODO: analyze things better to figure out what is really
> > going on here
> >
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > ---
> > drivers/gpu/drm/i915/display/intel_dpt.c | 7 ++++---
> > drivers/gpu/drm/i915/display/intel_dpt.h | 3 ++-
> > drivers/gpu/drm/i915/display/intel_fb.c | 17 ++++++++++++++++-
> > drivers/gpu/drm/i915/display/intel_fb_pin.c | 10 +++++-----
> > 4 files changed, 27 insertions(+), 10 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_dpt.c
> > b/drivers/gpu/drm/i915/display/intel_dpt.c
> > index b29bceff73f2..786d3f2e94c7 100644
> > --- a/drivers/gpu/drm/i915/display/intel_dpt.c
> > +++ b/drivers/gpu/drm/i915/display/intel_dpt.c
> > @@ -121,7 +121,8 @@ static void dpt_cleanup(struct i915_address_space
> > *vm)
> > i915_gem_object_put(dpt->obj);
> > }
> >
> > -struct i915_vma *intel_dpt_pin(struct i915_address_space *vm)
> > +struct i915_vma *intel_dpt_pin(struct i915_address_space *vm,
> > + unsigned int alignment)
> > {
> > struct drm_i915_private *i915 = vm->i915;
> > struct i915_dpt *dpt = i915_vm_to_dpt(vm); @@ -143,8 +144,8 @@
> > struct i915_vma *intel_dpt_pin(struct i915_address_space *vm)
> > if (err)
> > continue;
> >
> > - vma = i915_gem_object_ggtt_pin_ww(dpt->obj, &ww, NULL, 0,
> > 4096,
> > - pin_flags);
> > + vma = i915_gem_object_ggtt_pin_ww(dpt->obj, &ww, NULL, 0,
> > + alignment, pin_flags);
> > if (IS_ERR(vma)) {
> > err = PTR_ERR(vma);
> > continue;
> > diff --git a/drivers/gpu/drm/i915/display/intel_dpt.h
> > b/drivers/gpu/drm/i915/display/intel_dpt.h
> > index e18a9f767b11..f467578a4950 100644
> > --- a/drivers/gpu/drm/i915/display/intel_dpt.h
> > +++ b/drivers/gpu/drm/i915/display/intel_dpt.h
> > @@ -13,7 +13,8 @@ struct i915_vma;
> > struct intel_framebuffer;
> >
> > void intel_dpt_destroy(struct i915_address_space *vm); -struct i915_vma
> > *intel_dpt_pin(struct i915_address_space *vm);
> > +struct i915_vma *intel_dpt_pin(struct i915_address_space *vm,
> > + unsigned int alignment);
> > void intel_dpt_unpin(struct i915_address_space *vm); void
> > intel_dpt_suspend(struct drm_i915_private *i915); void
> > intel_dpt_resume(struct drm_i915_private *i915); diff --git
> > a/drivers/gpu/drm/i915/display/intel_fb.c
> > b/drivers/gpu/drm/i915/display/intel_fb.c
> > index 3ea6470d6d92..58ead05fba6f 100644
> > --- a/drivers/gpu/drm/i915/display/intel_fb.c
> > +++ b/drivers/gpu/drm/i915/display/intel_fb.c
> > @@ -805,8 +805,23 @@ unsigned int intel_surf_alignment(const struct
> > drm_framebuffer *fb, {
> > struct drm_i915_private *dev_priv = to_i915(fb->dev);
> >
> > - if (intel_fb_uses_dpt(fb))
> > + if (intel_fb_uses_dpt(fb)) {
> > + /* AUX_DIST needs only 4K alignment */
> > + if (intel_fb_is_ccs_aux_plane(fb, color_plane))
> > + return 512 * 4096;
> > +
> > + /*
> > + * FIXME ADL sees GGTT/DMAR faults with async
> > + * flips unless we align to 16k at least.
> > + * Figure out what's going on here...
> > + */
> > + if (IS_ALDERLAKE_P(dev_priv) &&
> > + !intel_fb_is_ccs_modifier(fb->modifier) &&
> > + HAS_ASYNC_FLIPS(dev_priv))
> > + return 512 * 16 * 1024;
> > +
> > return 512 * 4096;
> > + }
> >
> > /* AUX_DIST needs only 4K alignment */
> > if (intel_fb_is_ccs_aux_plane(fb, color_plane)) diff --git
> > a/drivers/gpu/drm/i915/display/intel_fb_pin.c
> > b/drivers/gpu/drm/i915/display/intel_fb_pin.c
> > index 7b42aef37d2f..c28ae99ebe6a 100644
> > --- a/drivers/gpu/drm/i915/display/intel_fb_pin.c
> > +++ b/drivers/gpu/drm/i915/display/intel_fb_pin.c
> > @@ -19,6 +19,7 @@
> > static struct i915_vma *
> > intel_pin_fb_obj_dpt(struct drm_framebuffer *fb,
> > const struct i915_gtt_view *view,
> > + unsigned int alignment,
> > bool uses_fence,
> > unsigned long *out_flags,
> > struct i915_address_space *vm)
> > @@ -28,7 +29,6 @@ intel_pin_fb_obj_dpt(struct drm_framebuffer *fb,
> > struct drm_i915_gem_object *obj = intel_fb_obj(fb);
> > struct i915_gem_ww_ctx ww;
> > struct i915_vma *vma;
> > - u32 alignment;
> > int ret;
> >
> > /*
> > @@ -41,8 +41,6 @@ intel_pin_fb_obj_dpt(struct drm_framebuffer *fb,
> > if (WARN_ON(!i915_gem_object_is_framebuffer(obj)))
> > return ERR_PTR(-EINVAL);
> >
> > - alignment = 4096 * 512;
> > -
> > atomic_inc(&dev_priv->gpu_error.pending_fb_pin);
> >
> > for_i915_gem_ww(&ww, ret, true) {
> > @@ -257,14 +255,16 @@ int intel_plane_pin_fb(struct intel_plane_state
> > *plane_state)
> > plane_state->ggtt_vma = vma;
> > } else {
> > struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
> > + unsigned int alignment = intel_surf_alignment(fb, 0);
> >
> > - vma = intel_dpt_pin(intel_fb->dpt_vm);
> > + vma = intel_dpt_pin(intel_fb->dpt_vm, alignment / 512);
> > if (IS_ERR(vma))
> > return PTR_ERR(vma);
> >
> > plane_state->ggtt_vma = vma;
> >
> > - vma = intel_pin_fb_obj_dpt(fb, &plane_state->view.gtt, false,
> > + vma = intel_pin_fb_obj_dpt(fb, &plane_state->view.gtt,
> > + alignment, false,
> > &plane_state->flags, intel_fb-
> > >dpt_vm);
> > if (IS_ERR(vma)) {
> > intel_dpt_unpin(intel_fb->dpt_vm);
> > --
> > 2.43.2
>
--
Ville Syrjälä
Intel
next prev parent reply other threads:[~2024-04-19 16:08 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-03-20 16:04 [PATCH 0/6] drm/i915: Allow the first async flip to change modifier Ville Syrjala
2024-03-20 16:04 ` [PATCH 1/6] drm/i915: Align PLANE_SURF to 16k on ADL for async flips Ville Syrjala
2024-04-19 4:20 ` Murthy, Arun R
2024-04-19 16:08 ` Ville Syrjälä [this message]
2024-04-23 3:36 ` Murthy, Arun R
2024-03-20 16:04 ` [PATCH 2/6] drm/i915: Reject async flips if we need to change DDB/watermarks Ville Syrjala
2024-04-19 4:27 ` Murthy, Arun R
2024-04-19 16:25 ` Ville Syrjälä
2024-04-23 3:45 ` Murthy, Arun R
2024-04-19 6:31 ` Kulkarni, Vandita
2024-03-20 16:04 ` [PATCH 3/6] drm/i915: Allow the initial async flip to change modifier Ville Syrjala
2024-04-18 16:11 ` Kulkarni, Vandita
2024-03-20 16:04 ` [PATCH 4/6] drm/i915: Eliminate extra frame from skl-glk sync->async flip change Ville Syrjala
2024-04-19 6:39 ` Murthy, Arun R
2024-04-19 16:41 ` Ville Syrjälä
2024-04-23 3:47 ` Murthy, Arun R
2024-03-20 16:04 ` [PATCH 5/6] drm/i915: s/need_async_flip_disable_wa/need_async_flip_toggle_wa/ Ville Syrjala
2024-04-19 6:41 ` Murthy, Arun R
2024-03-20 16:04 ` [PATCH 6/6] drm/i915: Extract ilk_must_disable_lp_wm() Ville Syrjala
2024-04-19 6:54 ` Murthy, Arun R
2024-03-21 1:05 ` ✗ Fi.CI.SPARSE: warning for drm/i915: Allow the first async flip to change modifier Patchwork
2024-03-21 1:18 ` ✓ Fi.CI.BAT: success " Patchwork
2024-03-21 11:43 ` ✗ Fi.CI.IGT: failure " Patchwork
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ZiKW6YK6mVf9rHIW@intel.com \
--to=ville.syrjala@linux.intel.com \
--cc=arun.r.murthy@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox