Re: [Intel-gfx] [PATCH 2/4] drm/i915: Adding Panel Filter function for DP
On Wed, Aug 19, 2015 at 09:11:11AM +, Zhang, Xiong Y wrote: On Fri, Aug 14, 2015 at 07:28:44PM +0300, Ville Syrjälä wrote: On Fri, Aug 14, 2015 at 05:12:57AM +, Zhang, Xiong Y wrote: On Mon, Aug 10, 2015 at 03:26:09PM +0800, Xiong Zhang wrote: Only internal eDP, LVDS, DVI screen could set scalling mode, some customers need to set scalling mode for external DP, HDMI, VGA screen. Let's fulfill this. bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90989 Signed-off-by: Xiong Zhang xiong.y.zh...@intel.com --- drivers/gpu/drm/i915/intel_dp.c | 63 - 1 file changed, 44 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index f1b9f93..2da334b 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -207,7 +207,13 @@ intel_dp_mode_valid(struct drm_connector *connector, int target_clock = mode-clock; int max_rate, mode_rate, max_lanes, max_link_clock; - if (is_edp(intel_dp) fixed_mode) { + if (mode-clock 1) + return MODE_CLOCK_LOW; + + if (mode-flags DRM_MODE_FLAG_DBLCLK) + return MODE_H_ILLEGAL; + + if (!intel_panel_scale_none(intel_connector-panel)) { if (mode-hdisplay fixed_mode-hdisplay) return MODE_PANEL; @@ -226,12 +232,6 @@ intel_dp_mode_valid(struct drm_connector *connector, if (mode_rate max_rate) return MODE_CLOCK_HIGH; - if (mode-clock 1) - return MODE_CLOCK_LOW; - - if (mode-flags DRM_MODE_FLAG_DBLCLK) - return MODE_H_ILLEGAL; - return MODE_OK; } @@ -1378,7 +1378,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, pipe_config-has_drrs = false; pipe_config-has_audio = intel_dp-has_audio port != PORT_A; - if (is_edp(intel_dp) intel_connector-panel.fixed_mode) { + if (!intel_panel_scale_none(intel_connector-panel)) { intel_fixed_panel_mode(intel_connector-panel.fixed_mode, adjusted_mode); @@ -4592,6 +4592,23 @@ static int intel_dp_get_modes(struct drm_connector *connector) edid = intel_connector-detect_edid; if (edid) { int ret = intel_connector_update_modes(connector, edid); + + if (ret intel_connector-panel.fixed_mode == NULL) { + /* init fixed mode as preferred mode for DP */ + struct drm_display_mode *fixed_mode = NULL; + struct drm_display_mode *scan; + + list_for_each_entry(scan, connector-probed_modes, head) { + if (scan-type DRM_MODE_TYPE_PREFERRED) + fixed_mode = drm_mode_duplicate(connector-dev, + scan); + } + + if (fixed_mode) + intel_panel_init(intel_connector-panel, +fixed_mode, NULL); + } How are we supposed to get rid of a stale fixed_mode when some other display gets plugged in? [Zhang, Xiong Y] Thanks so much for your good question. Yes, we should clear the stale fitting_mode and fixed_mode when display is disconnect in intel_dp_hpd_pulse() function. Also what would happen if the preferred mode can't be supported due to some source limitation? [Zhang, Xiong Y] In this case, which mode should be selected as fixed_mode ? At the very least we should make sure it's a mode we can use. As you said maybe kernel isn't the right place to do such decision. There are a lot of options how we could pick the mode. Eg. might want to pick the next largest mode, and if there is none try to pick the largest smaller mode (since pfit can't downscale by much). Also should we try to pick an intelaced mode if the user requested one etc. Lots of open questions how this policy should be handled. Would be easier to punt it all to userspace, which would also avoid the kernel policy doing the wrong thing when userspace knows what it wants. In general I'm not entirely happy with having this kind of policy in the kernel. I'd much prefer if we could get crtc size and border properties done so that userspace could set up the scaling any which way it chooses. [Zhang, Xiong Y] Could you give more detail about your
Re: [Intel-gfx] [PATCH 2/4] drm/i915: Adding Panel Filter function for DP
On Wednesday 19 August 2015 09:11:11 Zhang, Xiong Y wrote: On Fri, Aug 14, 2015 at 07:28:44PM +0300, Ville Syrjälä wrote: Another thing we could do with this approach is expose the pipe PF-ID mode when the fixed mode is interlaced and the user mode is progressive. And if both are interlaced, or there's just an interlaced used mode w/o a fixed mode, we'd keep using IF-ID like we do today. [Zhang, Xiong Y] I checked the B spec, there isn't PF-ID / IF-ID control bit in PIPE_CONF since BDW. I never see a monitor supporting interlaced mode. Does a monitor could support both Interlaced and progressive mode ? Most of the computer monitors I see support interlaced modes on HDMI; I usually see CEA-861 VIC 5 (1920x1080i60) and VIC 20 (1920x1080i50) supported on HDMI by computer monitors. -- Simon Farnsworth Software Engineer ONELAN Ltd http://www.onelan.com ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] [PATCH 2/4] drm/i915: Adding Panel Filter function for DP
On Fri, Aug 14, 2015 at 07:28:44PM +0300, Ville Syrjälä wrote: On Fri, Aug 14, 2015 at 05:12:57AM +, Zhang, Xiong Y wrote: On Mon, Aug 10, 2015 at 03:26:09PM +0800, Xiong Zhang wrote: Only internal eDP, LVDS, DVI screen could set scalling mode, some customers need to set scalling mode for external DP, HDMI, VGA screen. Let's fulfill this. bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90989 Signed-off-by: Xiong Zhang xiong.y.zh...@intel.com --- drivers/gpu/drm/i915/intel_dp.c | 63 - 1 file changed, 44 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index f1b9f93..2da334b 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -207,7 +207,13 @@ intel_dp_mode_valid(struct drm_connector *connector, int target_clock = mode-clock; int max_rate, mode_rate, max_lanes, max_link_clock; - if (is_edp(intel_dp) fixed_mode) { + if (mode-clock 1) + return MODE_CLOCK_LOW; + + if (mode-flags DRM_MODE_FLAG_DBLCLK) + return MODE_H_ILLEGAL; + + if (!intel_panel_scale_none(intel_connector-panel)) { if (mode-hdisplay fixed_mode-hdisplay) return MODE_PANEL; @@ -226,12 +232,6 @@ intel_dp_mode_valid(struct drm_connector *connector, if (mode_rate max_rate) return MODE_CLOCK_HIGH; - if (mode-clock 1) - return MODE_CLOCK_LOW; - - if (mode-flags DRM_MODE_FLAG_DBLCLK) - return MODE_H_ILLEGAL; - return MODE_OK; } @@ -1378,7 +1378,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, pipe_config-has_drrs = false; pipe_config-has_audio = intel_dp-has_audio port != PORT_A; - if (is_edp(intel_dp) intel_connector-panel.fixed_mode) { + if (!intel_panel_scale_none(intel_connector-panel)) { intel_fixed_panel_mode(intel_connector-panel.fixed_mode, adjusted_mode); @@ -4592,6 +4592,23 @@ static int intel_dp_get_modes(struct drm_connector *connector) edid = intel_connector-detect_edid; if (edid) { int ret = intel_connector_update_modes(connector, edid); + + if (ret intel_connector-panel.fixed_mode == NULL) { + /* init fixed mode as preferred mode for DP */ + struct drm_display_mode *fixed_mode = NULL; + struct drm_display_mode *scan; + + list_for_each_entry(scan, connector-probed_modes, head) { + if (scan-type DRM_MODE_TYPE_PREFERRED) + fixed_mode = drm_mode_duplicate(connector-dev, + scan); + } + + if (fixed_mode) + intel_panel_init(intel_connector-panel, + fixed_mode, NULL); + } How are we supposed to get rid of a stale fixed_mode when some other display gets plugged in? [Zhang, Xiong Y] Thanks so much for your good question. Yes, we should clear the stale fitting_mode and fixed_mode when display is disconnect in intel_dp_hpd_pulse() function. Also what would happen if the preferred mode can't be supported due to some source limitation? [Zhang, Xiong Y] In this case, which mode should be selected as fixed_mode ? At the very least we should make sure it's a mode we can use. As you said maybe kernel isn't the right place to do such decision. There are a lot of options how we could pick the mode. Eg. might want to pick the next largest mode, and if there is none try to pick the largest smaller mode (since pfit can't downscale by much). Also should we try to pick an intelaced mode if the user requested one etc. Lots of open questions how this policy should be handled. Would be easier to punt it all to userspace, which would also avoid the kernel policy doing the wrong thing when userspace knows what it wants. In general I'm not entirely happy with having this kind of policy in the kernel. I'd much prefer if we could get crtc size and border properties done so that userspace could set up the scaling any which way it chooses. [Zhang, Xiong Y] Could you give more detail about your preference ? The idea would be to expose some sort of crtc size properties that would provide pipe_src_{w,h}, and the
Re: [Intel-gfx] [PATCH 2/4] drm/i915: Adding Panel Filter function for DP
On Fri, Aug 14, 2015 at 07:28:44PM +0300, Ville Syrjälä wrote: On Fri, Aug 14, 2015 at 05:12:57AM +, Zhang, Xiong Y wrote: On Mon, Aug 10, 2015 at 03:26:09PM +0800, Xiong Zhang wrote: Only internal eDP, LVDS, DVI screen could set scalling mode, some customers need to set scalling mode for external DP, HDMI, VGA screen. Let's fulfill this. bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90989 Signed-off-by: Xiong Zhang xiong.y.zh...@intel.com --- drivers/gpu/drm/i915/intel_dp.c | 63 - 1 file changed, 44 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index f1b9f93..2da334b 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -207,7 +207,13 @@ intel_dp_mode_valid(struct drm_connector *connector, int target_clock = mode-clock; int max_rate, mode_rate, max_lanes, max_link_clock; - if (is_edp(intel_dp) fixed_mode) { + if (mode-clock 1) + return MODE_CLOCK_LOW; + + if (mode-flags DRM_MODE_FLAG_DBLCLK) + return MODE_H_ILLEGAL; + + if (!intel_panel_scale_none(intel_connector-panel)) { if (mode-hdisplay fixed_mode-hdisplay) return MODE_PANEL; @@ -226,12 +232,6 @@ intel_dp_mode_valid(struct drm_connector *connector, if (mode_rate max_rate) return MODE_CLOCK_HIGH; - if (mode-clock 1) - return MODE_CLOCK_LOW; - - if (mode-flags DRM_MODE_FLAG_DBLCLK) - return MODE_H_ILLEGAL; - return MODE_OK; } @@ -1378,7 +1378,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, pipe_config-has_drrs = false; pipe_config-has_audio = intel_dp-has_audio port != PORT_A; - if (is_edp(intel_dp) intel_connector-panel.fixed_mode) { + if (!intel_panel_scale_none(intel_connector-panel)) { intel_fixed_panel_mode(intel_connector-panel.fixed_mode, adjusted_mode); @@ -4592,6 +4592,23 @@ static int intel_dp_get_modes(struct drm_connector *connector) edid = intel_connector-detect_edid; if (edid) { int ret = intel_connector_update_modes(connector, edid); + + if (ret intel_connector-panel.fixed_mode == NULL) { + /* init fixed mode as preferred mode for DP */ + struct drm_display_mode *fixed_mode = NULL; + struct drm_display_mode *scan; + + list_for_each_entry(scan, connector-probed_modes, head) { + if (scan-type DRM_MODE_TYPE_PREFERRED) + fixed_mode = drm_mode_duplicate(connector-dev, + scan); + } + + if (fixed_mode) + intel_panel_init(intel_connector-panel, +fixed_mode, NULL); + } How are we supposed to get rid of a stale fixed_mode when some other display gets plugged in? [Zhang, Xiong Y] Thanks so much for your good question. Yes, we should clear the stale fitting_mode and fixed_mode when display is disconnect in intel_dp_hpd_pulse() function. Also what would happen if the preferred mode can't be supported due to some source limitation? [Zhang, Xiong Y] In this case, which mode should be selected as fixed_mode ? At the very least we should make sure it's a mode we can use. As you said maybe kernel isn't the right place to do such decision. There are a lot of options how we could pick the mode. Eg. might want to pick the next largest mode, and if there is none try to pick the largest smaller mode (since pfit can't downscale by much). Also should we try to pick an intelaced mode if the user requested one etc. Lots of open questions how this policy should be handled. Would be easier to punt it all to userspace, which would also avoid the kernel policy doing the wrong thing when userspace knows what it wants. In general I'm not entirely happy with having this kind of policy in the kernel. I'd much prefer if we could get crtc size and border properties done so that userspace could set up the scaling any which way it chooses. [Zhang, Xiong Y] Could you give more detail about your preference ? The idea would be to expose some sort of crtc size properties that would provide pipe_src_{w,h}, and the mode would provide
Re: [Intel-gfx] [PATCH 2/4] drm/i915: Adding Panel Filter function for DP
On Fri, Aug 14, 2015 at 05:12:57AM +, Zhang, Xiong Y wrote: On Mon, Aug 10, 2015 at 03:26:09PM +0800, Xiong Zhang wrote: Only internal eDP, LVDS, DVI screen could set scalling mode, some customers need to set scalling mode for external DP, HDMI, VGA screen. Let's fulfill this. bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90989 Signed-off-by: Xiong Zhang xiong.y.zh...@intel.com --- drivers/gpu/drm/i915/intel_dp.c | 63 - 1 file changed, 44 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index f1b9f93..2da334b 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -207,7 +207,13 @@ intel_dp_mode_valid(struct drm_connector *connector, int target_clock = mode-clock; int max_rate, mode_rate, max_lanes, max_link_clock; - if (is_edp(intel_dp) fixed_mode) { + if (mode-clock 1) + return MODE_CLOCK_LOW; + + if (mode-flags DRM_MODE_FLAG_DBLCLK) + return MODE_H_ILLEGAL; + + if (!intel_panel_scale_none(intel_connector-panel)) { if (mode-hdisplay fixed_mode-hdisplay) return MODE_PANEL; @@ -226,12 +232,6 @@ intel_dp_mode_valid(struct drm_connector *connector, if (mode_rate max_rate) return MODE_CLOCK_HIGH; - if (mode-clock 1) - return MODE_CLOCK_LOW; - - if (mode-flags DRM_MODE_FLAG_DBLCLK) - return MODE_H_ILLEGAL; - return MODE_OK; } @@ -1378,7 +1378,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, pipe_config-has_drrs = false; pipe_config-has_audio = intel_dp-has_audio port != PORT_A; - if (is_edp(intel_dp) intel_connector-panel.fixed_mode) { + if (!intel_panel_scale_none(intel_connector-panel)) { intel_fixed_panel_mode(intel_connector-panel.fixed_mode, adjusted_mode); @@ -4592,6 +4592,23 @@ static int intel_dp_get_modes(struct drm_connector *connector) edid = intel_connector-detect_edid; if (edid) { int ret = intel_connector_update_modes(connector, edid); + + if (ret intel_connector-panel.fixed_mode == NULL) { + /* init fixed mode as preferred mode for DP */ + struct drm_display_mode *fixed_mode = NULL; + struct drm_display_mode *scan; + + list_for_each_entry(scan, connector-probed_modes, head) { + if (scan-type DRM_MODE_TYPE_PREFERRED) + fixed_mode = drm_mode_duplicate(connector-dev, + scan); + } + + if (fixed_mode) + intel_panel_init(intel_connector-panel, + fixed_mode, NULL); + } How are we supposed to get rid of a stale fixed_mode when some other display gets plugged in? [Zhang, Xiong Y] Thanks so much for your good question. Yes, we should clear the stale fitting_mode and fixed_mode when display is disconnect in intel_dp_hpd_pulse() function. Also what would happen if the preferred mode can't be supported due to some source limitation? [Zhang, Xiong Y] In this case, which mode should be selected as fixed_mode ? At the very least we should make sure it's a mode we can use. As you said maybe kernel isn't the right place to do such decision. There are a lot of options how we could pick the mode. Eg. might want to pick the next largest mode, and if there is none try to pick the largest smaller mode (since pfit can't downscale by much). Also should we try to pick an intelaced mode if the user requested one etc. Lots of open questions how this policy should be handled. Would be easier to punt it all to userspace, which would also avoid the kernel policy doing the wrong thing when userspace knows what it wants. In general I'm not entirely happy with having this kind of policy in the kernel. I'd much prefer if we could get crtc size and border properties done so that userspace could set up the scaling any which way it chooses. [Zhang, Xiong Y] Could you give more detail about your preference ? The idea would be to expose some sort of crtc size properties that would provide pipe_src_{w,h}, and the mode would provide the actual display timings as it does today. And if we add some kind of border properties to control the output size of the pfit (since the mode doesn't have that information) userspace could do whatever it pleased. I think the last attempt just sort of died out during review: http://lists.freedesktop.org/archives/intel-gfx/2014-August/050732.html -- Ville Syrjälä Intel OTC ___ Intel-gfx
Re: [Intel-gfx] [PATCH 2/4] drm/i915: Adding Panel Filter function for DP
On Mon, Aug 10, 2015 at 03:26:09PM +0800, Xiong Zhang wrote: Only internal eDP, LVDS, DVI screen could set scalling mode, some customers need to set scalling mode for external DP, HDMI, VGA screen. Let's fulfill this. bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90989 Signed-off-by: Xiong Zhang xiong.y.zh...@intel.com --- drivers/gpu/drm/i915/intel_dp.c | 63 - 1 file changed, 44 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index f1b9f93..2da334b 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -207,7 +207,13 @@ intel_dp_mode_valid(struct drm_connector *connector, int target_clock = mode-clock; int max_rate, mode_rate, max_lanes, max_link_clock; - if (is_edp(intel_dp) fixed_mode) { + if (mode-clock 1) + return MODE_CLOCK_LOW; + + if (mode-flags DRM_MODE_FLAG_DBLCLK) + return MODE_H_ILLEGAL; + + if (!intel_panel_scale_none(intel_connector-panel)) { if (mode-hdisplay fixed_mode-hdisplay) return MODE_PANEL; @@ -226,12 +232,6 @@ intel_dp_mode_valid(struct drm_connector *connector, if (mode_rate max_rate) return MODE_CLOCK_HIGH; - if (mode-clock 1) - return MODE_CLOCK_LOW; - - if (mode-flags DRM_MODE_FLAG_DBLCLK) - return MODE_H_ILLEGAL; - return MODE_OK; } @@ -1378,7 +1378,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, pipe_config-has_drrs = false; pipe_config-has_audio = intel_dp-has_audio port != PORT_A; - if (is_edp(intel_dp) intel_connector-panel.fixed_mode) { + if (!intel_panel_scale_none(intel_connector-panel)) { intel_fixed_panel_mode(intel_connector-panel.fixed_mode, adjusted_mode); @@ -4592,6 +4592,23 @@ static int intel_dp_get_modes(struct drm_connector *connector) edid = intel_connector-detect_edid; if (edid) { int ret = intel_connector_update_modes(connector, edid); + + if (ret intel_connector-panel.fixed_mode == NULL) { + /* init fixed mode as preferred mode for DP */ + struct drm_display_mode *fixed_mode = NULL; + struct drm_display_mode *scan; + + list_for_each_entry(scan, connector-probed_modes, head) { + if (scan-type DRM_MODE_TYPE_PREFERRED) + fixed_mode = drm_mode_duplicate(connector-dev, + scan); + } + + if (fixed_mode) + intel_panel_init(intel_connector-panel, +fixed_mode, NULL); + } How are we supposed to get rid of a stale fixed_mode when some other display gets plugged in? [Zhang, Xiong Y] Thanks so much for your good question. Yes, we should clear the stale fitting_mode and fixed_mode when display is disconnect in intel_dp_hpd_pulse() function. Also what would happen if the preferred mode can't be supported due to some source limitation? [Zhang, Xiong Y] In this case, which mode should be selected as fixed_mode ? As you said maybe kernel isn't the right place to do such decision. In general I'm not entirely happy with having this kind of policy in the kernel. I'd much prefer if we could get crtc size and border properties done so that userspace could set up the scaling any which way it chooses. [Zhang, Xiong Y] Could you give more detail about your preference ? thanks ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] [PATCH 2/4] drm/i915: Adding Panel Filter function for DP
On Mon, Aug 10, 2015 at 03:26:09PM +0800, Xiong Zhang wrote: Only internal eDP, LVDS, DVI screen could set scalling mode, some customers need to set scalling mode for external DP, HDMI, VGA screen. Let's fulfill this. bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90989 Signed-off-by: Xiong Zhang xiong.y.zh...@intel.com --- drivers/gpu/drm/i915/intel_dp.c | 63 - 1 file changed, 44 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index f1b9f93..2da334b 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -207,7 +207,13 @@ intel_dp_mode_valid(struct drm_connector *connector, int target_clock = mode-clock; int max_rate, mode_rate, max_lanes, max_link_clock; - if (is_edp(intel_dp) fixed_mode) { + if (mode-clock 1) + return MODE_CLOCK_LOW; + + if (mode-flags DRM_MODE_FLAG_DBLCLK) + return MODE_H_ILLEGAL; + + if (!intel_panel_scale_none(intel_connector-panel)) { if (mode-hdisplay fixed_mode-hdisplay) return MODE_PANEL; @@ -226,12 +232,6 @@ intel_dp_mode_valid(struct drm_connector *connector, if (mode_rate max_rate) return MODE_CLOCK_HIGH; - if (mode-clock 1) - return MODE_CLOCK_LOW; - - if (mode-flags DRM_MODE_FLAG_DBLCLK) - return MODE_H_ILLEGAL; - return MODE_OK; } @@ -1378,7 +1378,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, pipe_config-has_drrs = false; pipe_config-has_audio = intel_dp-has_audio port != PORT_A; - if (is_edp(intel_dp) intel_connector-panel.fixed_mode) { + if (!intel_panel_scale_none(intel_connector-panel)) { intel_fixed_panel_mode(intel_connector-panel.fixed_mode, adjusted_mode); @@ -4592,6 +4592,23 @@ static int intel_dp_get_modes(struct drm_connector *connector) edid = intel_connector-detect_edid; if (edid) { int ret = intel_connector_update_modes(connector, edid); + + if (ret intel_connector-panel.fixed_mode == NULL) { + /* init fixed mode as preferred mode for DP */ + struct drm_display_mode *fixed_mode = NULL; + struct drm_display_mode *scan; + + list_for_each_entry(scan, connector-probed_modes, head) { + if (scan-type DRM_MODE_TYPE_PREFERRED) + fixed_mode = drm_mode_duplicate(connector-dev, + scan); + } + + if (fixed_mode) + intel_panel_init(intel_connector-panel, + fixed_mode, NULL); + } How are we supposed to get rid of a stale fixed_mode when some other display gets plugged in? Also what would happen if the preferred mode can't be supported due to some source limitation? In general I'm not entirely happy with having this kind of policy in the kernel. I'd much prefer if we could get crtc size and border properties done so that userspace could set up the scaling any which way it chooses. + if (ret) return ret; } @@ -4688,15 +4705,14 @@ intel_dp_set_property(struct drm_connector *connector, goto done; } - if (is_edp(intel_dp) - property == connector-dev-mode_config.scaling_mode_property) { - if (val == DRM_MODE_SCALE_NONE) { - DRM_DEBUG_KMS(no scaling not supported\n); + if (property == connector-dev-mode_config.scaling_mode_property) { + if (is_edp(intel_dp) val == DRM_MODE_SCALE_NONE) { + DRM_DEBUG_KMS(eDP: no scaling not supported\n); return -EINVAL; } if (intel_connector-panel.fitting_mode == val) { - /* the eDP scaling property is not changed */ + /* the connector scaling property is not changed */ return 0; } intel_connector-panel.fitting_mode = val; @@ -4989,13 +5005,22 @@ intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connect intel_attach_broadcast_rgb_property(connector); intel_dp-color_range_auto = true; - if (is_edp(intel_dp)) { + /* Each pipe has panel filter since Ironlake. */ + if (INTEL_INFO(connector-dev)-gen = 5) { drm_mode_create_scaling_mode_property(connector-dev); - drm_object_attach_property( - connector-base, - connector-dev-mode_config.scaling_mode_property, -