From: "Lisovskiy, Stanislav" <stanislav.lisovskiy@intel.com>
To: Luca Coelho <luciano.coelho@intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [Intel-gfx] [PATCH v7 1/2] drm/i915/mtl: limit second scaler vertical scaling in ver >= 14
Date: Mon, 9 Jan 2023 09:06:05 +0200 [thread overview]
Message-ID: <Y7u83SCf5CGbNZ6v@intel.com> (raw)
In-Reply-To: <20221223130509.43245-2-luciano.coelho@intel.com>
On Fri, Dec 23, 2022 at 03:05:08PM +0200, Luca Coelho wrote:
> In newer hardware versions (i.e. display version >= 14), the second
> scaler doesn't support vertical scaling.
>
> The current implementation of the scaling limits is simplified and
> only occurs when the planes are created, so we don't know which scaler
> is being used.
>
> In order to handle separate scaling limits for horizontal and vertical
> scaling, and different limits per scaler, split the checks in two
> phases. We first do a simple check during plane creation and use the
> best-case scenario (because we don't know the scaler that may be used
> at a later point) and then do a more specific check when the scalers
> are actually being set up.
>
> Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
> ---
>
> In v2:
> * fix DRM_PLANE_NO_SCALING renamed macros;
>
> In v3:
> * No changes.
>
> In v4:
> * Got rid of the changes in the general planes max scale code;
> * Added a couple of FIXMEs;
> * Made intel_atomic_setup_scaler() return an int with errors;
>
> In v5:
> * Just resent with a cover letter.
>
> In v6:
> * Now the correct version again (same as v4).
>
> In v7:
> * Constify a couple of local variables;
> * Return -EINVAL, instead of -EOPNOTSUPP and -EBUSY;
> * Add another FIXME;
> * Remove unnecessary undoing of change in error cases.
>
>
> drivers/gpu/drm/i915/display/intel_atomic.c | 85 ++++++++++++++++++---
> 1 file changed, 75 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_atomic.c b/drivers/gpu/drm/i915/display/intel_atomic.c
> index 6621aa245caf..a9a3f3715279 100644
> --- a/drivers/gpu/drm/i915/display/intel_atomic.c
> +++ b/drivers/gpu/drm/i915/display/intel_atomic.c
> @@ -41,6 +41,7 @@
> #include "intel_global_state.h"
> #include "intel_hdcp.h"
> #include "intel_psr.h"
> +#include "intel_fb.h"
> #include "skl_universal_plane.h"
>
> /**
> @@ -310,11 +311,11 @@ intel_crtc_destroy_state(struct drm_crtc *crtc,
> kfree(crtc_state);
> }
>
> -static void intel_atomic_setup_scaler(struct intel_crtc_scaler_state *scaler_state,
> - int num_scalers_need, struct intel_crtc *intel_crtc,
> - const char *name, int idx,
> - struct intel_plane_state *plane_state,
> - int *scaler_id)
> +static int intel_atomic_setup_scaler(struct intel_crtc_scaler_state *scaler_state,
> + int num_scalers_need, struct intel_crtc *intel_crtc,
> + const char *name, int idx,
> + struct intel_plane_state *plane_state,
> + int *scaler_id)
> {
> struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
> int j;
> @@ -334,7 +335,7 @@ static void intel_atomic_setup_scaler(struct intel_crtc_scaler_state *scaler_sta
>
> if (drm_WARN(&dev_priv->drm, *scaler_id < 0,
> "Cannot find scaler for %s:%d\n", name, idx))
> - return;
> + return -EINVAL;
As I understand that change is a bit irrelevant to the patch topic,
ideally it should be reflected in the commit message, that we are doing
this and most importantly why.
However I'm not going to be picky here, as it is a small thing, just
as a side note.
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
>
> /* set scaler mode */
> if (plane_state && plane_state->hw.fb &&
> @@ -375,9 +376,71 @@ static void intel_atomic_setup_scaler(struct intel_crtc_scaler_state *scaler_sta
> mode = SKL_PS_SCALER_MODE_DYN;
> }
>
> + /*
> + * FIXME: we should also check the scaler factors for pfit, so
> + * this shouldn't be tied directly to planes.
> + */
> + if (plane_state && plane_state->hw.fb) {
> + const struct drm_framebuffer *fb = plane_state->hw.fb;
> + const struct drm_rect *src = &plane_state->uapi.src;
> + const struct drm_rect *dst = &plane_state->uapi.dst;
> + int hscale, vscale, max_vscale, max_hscale;
> +
> + /*
> + * FIXME: When two scalers are needed, but only one of
> + * them needs to downscale, we should make sure that
> + * the one that needs downscaling support is assigned
> + * as the first scaler, so we don't reject downscaling
> + * unnecessarily.
> + */
> +
> + if (DISPLAY_VER(dev_priv) >= 14) {
> + /*
> + * On versions 14 and up, only the first
> + * scaler supports a vertical scaling factor
> + * of more than 1.0, while a horizontal
> + * scaling factor of 3.0 is supported.
> + */
> + max_hscale = 0x30000 - 1;
> + if (*scaler_id == 0)
> + max_vscale = 0x30000 - 1;
> + else
> + max_vscale = 0x10000;
> +
> + } else if (DISPLAY_VER(dev_priv) >= 10 ||
> + !intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier)) {
> + max_hscale = 0x30000 - 1;
> + max_vscale = 0x30000 - 1;
> + } else {
> + max_hscale = 0x20000 - 1;
> + max_vscale = 0x20000 - 1;
> + }
> +
> + /*
> + * FIXME: We should change the if-else block above to
> + * support HQ vs dynamic scaler properly.
> + */
> +
> + /* Check if required scaling is within limits */
> + hscale = drm_rect_calc_hscale(src, dst, 1, max_hscale);
> + vscale = drm_rect_calc_vscale(src, dst, 1, max_vscale);
> +
> + if (hscale < 0 || vscale < 0) {
> + drm_dbg_kms(&dev_priv->drm,
> + "Scaler %d doesn't support required plane scaling\n",
> + *scaler_id);
> + drm_rect_debug_print("src: ", src, true);
> + drm_rect_debug_print("dst: ", dst, false);
> +
> + return -EINVAL;
> + }
> + }
> +
> drm_dbg_kms(&dev_priv->drm, "Attached scaler id %u.%u to %s:%d\n",
> intel_crtc->pipe, *scaler_id, name, idx);
> scaler_state->scalers[*scaler_id].mode = mode;
> +
> + return 0;
> }
>
> /**
> @@ -437,7 +500,7 @@ int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv,
> for (i = 0; i < sizeof(scaler_state->scaler_users) * 8; i++) {
> int *scaler_id;
> const char *name;
> - int idx;
> + int idx, ret;
>
> /* skip if scaler not required */
> if (!(scaler_state->scaler_users & (1 << i)))
> @@ -494,9 +557,11 @@ int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv,
> scaler_id = &plane_state->scaler_id;
> }
>
> - intel_atomic_setup_scaler(scaler_state, num_scalers_need,
> - intel_crtc, name, idx,
> - plane_state, scaler_id);
> + ret = intel_atomic_setup_scaler(scaler_state, num_scalers_need,
> + intel_crtc, name, idx,
> + plane_state, scaler_id);
> + if (ret < 0)
> + return ret;
> }
>
> return 0;
> --
> 2.39.0
>
next prev parent reply other threads:[~2023-01-09 7:06 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-23 13:05 [Intel-gfx] [PATCH v7 0/2] drm/i915/mtl: handle some MTL scaler limitations Luca Coelho
2022-12-23 13:05 ` [Intel-gfx] [PATCH v7 1/2] drm/i915/mtl: limit second scaler vertical scaling in ver >= 14 Luca Coelho
2023-01-09 7:06 ` Lisovskiy, Stanislav [this message]
2023-01-09 12:07 ` Coelho, Luciano
2022-12-23 13:05 ` [Intel-gfx] [PATCH v7 2/2] drm/i915/mtl: update scaler source and destination limits for MTL Luca Coelho
2022-12-23 14:07 ` Nautiyal, Ankit K
2022-12-23 14:12 ` Nautiyal, Ankit K
2022-12-23 18:24 ` Coelho, Luciano
2023-01-09 7:32 ` Lisovskiy, Stanislav
2023-01-10 23:27 ` Sripada, Radhakrishna
2022-12-23 13:22 ` [Intel-gfx] ✗ Fi.CI.SPARSE: warning for drm/i915/mtl: handle some MTL scaler limitations Patchwork
2022-12-23 13:51 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2022-12-23 15:07 ` [Intel-gfx] ✓ Fi.CI.IGT: " Patchwork
2023-01-02 10:17 ` [Intel-gfx] [PATCH v7 0/2] " Coelho, Luciano
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=Y7u83SCf5CGbNZ6v@intel.com \
--to=stanislav.lisovskiy@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=luciano.coelho@intel.com \
/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