From efc2924e733631a64c0afed8dbba2741d186998c Mon Sep 17 00:00:00 2001 From: Keith Packard Date: Mon, 6 Jun 2011 17:12:49 -0700 Subject: drm/i915: Call intel_enable_plane from i9xx_crtc_mode_set (again) This change got placed in the ironlake path instead of the 9xx path during a recent code shuffle. Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/intel_display.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 81a9059b6a94..aa43e7be6053 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -4687,6 +4687,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, I915_WRITE(DSPCNTR(plane), dspcntr); POSTING_READ(DSPCNTR(plane)); + intel_enable_plane(dev_priv, plane, pipe); ret = intel_pipe_set_base(crtc, x, y, old_fb); @@ -5217,8 +5218,6 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc, I915_WRITE(DSPCNTR(plane), dspcntr); POSTING_READ(DSPCNTR(plane)); - if (!HAS_PCH_SPLIT(dev)) - intel_enable_plane(dev_priv, plane, pipe); ret = intel_pipe_set_base(crtc, x, y, old_fb); -- cgit v1.2.3 From ec6a890dfed7dd245beba5e5bcdfcffbd934c284 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 21 Jun 2011 18:37:59 +0100 Subject: drm/i915: Apply HWSTAM workaround for BSD ring on SandyBridge ...we need to apply exactly the same workaround for missing interrupts from BSD as for the BLT ring, apparently. See also commit 498e720b96379d8ee9c294950a01534a73defcf3 (drm/i915: Fix gen6 (SNB) missed BLT ring interrupts). Reported-and-tested-by: nkalkhof@web.de Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38529 Signed-off-by: Chris Wilson Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/i915_irq.c | 1 + drivers/gpu/drm/i915/i915_reg.h | 1 + 2 files changed, 2 insertions(+) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 9e34a1abeb61..ae2b49969b99 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1749,6 +1749,7 @@ void ironlake_irq_preinstall(struct drm_device *dev) * happens. */ I915_WRITE(GEN6_BLITTER_HWSTAM, ~GEN6_BLITTER_USER_INTERRUPT); + I915_WRITE(GEN6_BSD_HWSTAM, ~GEN6_BSD_USER_INTERRUPT); } /* XXX hotplug from PCH */ diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 2f967af8e62e..5d5def756c9e 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -531,6 +531,7 @@ #define GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_ENABLE 0 #define GEN6_BSD_SLEEP_PSMI_CONTROL_IDLE_INDICATOR (1 << 3) +#define GEN6_BSD_HWSTAM 0x12098 #define GEN6_BSD_IMR 0x120a8 #define GEN6_BSD_USER_INTERRUPT (1 << 12) -- cgit v1.2.3 From e92d03bff9a0d0bcbb812c9b1290ca96c9338d45 Mon Sep 17 00:00:00 2001 From: Eric Anholt Date: Tue, 14 Jun 2011 16:43:09 -0700 Subject: Revert "drm/i915: Kill GTT mappings when moving from GTT domain" This reverts commit 4a684a4117abd756291969336af454e8a958802f. Userland has always been required to set the object's domain to GTT before using it through a GTT mapping, it's not something that the kernel is supposed to enforce. (The pagefault support is so that we can handle multiple mappings without userland having to pin across them, not so that userland can use GTT after GPU domains without telling the kernel). Fixes 19.2% +/- 0.8% (n=6) performance regression in cairo-gl firefox-talos-gfx on my T420 latop. Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/i915_gem.c | 10 ++++------ drivers/gpu/drm/i915/i915_gem_execbuffer.c | 4 ---- 2 files changed, 4 insertions(+), 10 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 94c84d744100..c6389de53161 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1219,11 +1219,11 @@ int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf) ret = i915_gem_object_bind_to_gtt(obj, 0, true); if (ret) goto unlock; - } - ret = i915_gem_object_set_to_gtt_domain(obj, write); - if (ret) - goto unlock; + ret = i915_gem_object_set_to_gtt_domain(obj, write); + if (ret) + goto unlock; + } if (obj->tiling_mode == I915_TILING_NONE) ret = i915_gem_object_put_fence(obj); @@ -2926,8 +2926,6 @@ i915_gem_object_flush_gtt_write_domain(struct drm_i915_gem_object *obj) */ wmb(); - i915_gem_release_mmap(obj); - old_write_domain = obj->base.write_domain; obj->base.write_domain = 0; diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 20a4cc5b818f..4934cf84c320 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -187,10 +187,6 @@ i915_gem_object_set_to_gpu_domain(struct drm_i915_gem_object *obj, if ((flush_domains | invalidate_domains) & I915_GEM_DOMAIN_CPU) i915_gem_clflush_object(obj); - /* blow away mappings if mapped through GTT */ - if ((flush_domains | invalidate_domains) & I915_GEM_DOMAIN_GTT) - i915_gem_release_mmap(obj); - if (obj->base.pending_write_domain) cd->flips |= atomic_read(&obj->pending_flip); -- cgit v1.2.3 From 483f1798998ede75b2575cc1813c3d89ba64a34e Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Wed, 22 Jun 2011 09:55:01 -0700 Subject: drm/i915: save/resume forcewake lock fixes The lock must be held for the saving and restoring of VGA state. Signed-off-by: Ben Widawsky CC: Alexander Zhaunerchyk CC: Andrey Rahmatullin Reviewed-by: Chris Wilson Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/i915_suspend.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c index 60a94d2b5264..e8152d23d5b6 100644 --- a/drivers/gpu/drm/i915/i915_suspend.c +++ b/drivers/gpu/drm/i915/i915_suspend.c @@ -678,6 +678,7 @@ void i915_save_display(struct drm_device *dev) } /* VGA state */ + mutex_lock(&dev->struct_mutex); dev_priv->saveVGA0 = I915_READ(VGA0); dev_priv->saveVGA1 = I915_READ(VGA1); dev_priv->saveVGA_PD = I915_READ(VGA_PD); @@ -687,6 +688,7 @@ void i915_save_display(struct drm_device *dev) dev_priv->saveVGACNTRL = I915_READ(VGACNTRL); i915_save_vga(dev); + mutex_unlock(&dev->struct_mutex); } void i915_restore_display(struct drm_device *dev) @@ -780,6 +782,8 @@ void i915_restore_display(struct drm_device *dev) I915_WRITE(CPU_VGACNTRL, dev_priv->saveVGACNTRL); else I915_WRITE(VGACNTRL, dev_priv->saveVGACNTRL); + + mutex_lock(&dev->struct_mutex); I915_WRITE(VGA0, dev_priv->saveVGA0); I915_WRITE(VGA1, dev_priv->saveVGA1); I915_WRITE(VGA_PD, dev_priv->saveVGA_PD); @@ -787,6 +791,7 @@ void i915_restore_display(struct drm_device *dev) udelay(150); i915_restore_vga(dev); + mutex_unlock(&dev->struct_mutex); } int i915_save_state(struct drm_device *dev) -- cgit v1.2.3 From d698a34da7c5626c3c16a6311c6290522e902e7f Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 23 Jun 2011 00:49:29 -0400 Subject: drm/radeon/kms: fix num_banks tiling config for fusion The field is encoded: 0 = 4 banks 1 = 8 banks 2 = 16 banks Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/evergreen.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index 445af7981637..12d2fdc52414 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -2013,9 +2013,9 @@ static void evergreen_gpu_init(struct radeon_device *rdev) rdev->config.evergreen.tile_config |= (3 << 0); break; } - /* num banks is 8 on all fusion asics */ + /* num banks is 8 on all fusion asics. 0 = 4, 1 = 8, 2 = 16 */ if (rdev->flags & RADEON_IS_IGP) - rdev->config.evergreen.tile_config |= 8 << 4; + rdev->config.evergreen.tile_config |= 1 << 4; else rdev->config.evergreen.tile_config |= ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) << 4; -- cgit v1.2.3 From ee4017f4ac8163737793cc64df535cd246792887 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 23 Jun 2011 12:19:32 -0400 Subject: drm/radeon/kms: handle special cases for vddc A voltage value of 0xff01 requires that the driver look up the max voltage for the board based using the atom SetVoltage command table. Setting the proper voltage should fix stability on some newer asics. Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/radeon.h | 1 + drivers/gpu/drm/radeon/radeon_atombios.c | 36 ++++++++++++++++++++++++++++++++ 2 files changed, 37 insertions(+) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 27f45579e64b..ef0e0e016914 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -179,6 +179,7 @@ void radeon_pm_resume(struct radeon_device *rdev); void radeon_combios_get_power_modes(struct radeon_device *rdev); void radeon_atombios_get_power_modes(struct radeon_device *rdev); void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type); +int radeon_atom_get_max_vddc(struct radeon_device *rdev, u16 *voltage); void rs690_pm_info(struct radeon_device *rdev); extern int rv6xx_get_temp(struct radeon_device *rdev); extern int rv770_get_temp(struct radeon_device *rdev); diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c index 1e725d9f767f..bf2b61584cdb 100644 --- a/drivers/gpu/drm/radeon/radeon_atombios.c +++ b/drivers/gpu/drm/radeon/radeon_atombios.c @@ -2320,6 +2320,14 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev, le16_to_cpu(clock_info->r600.usVDDC); } + /* patch up vddc if necessary */ + if (rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage == 0xff01) { + u16 vddc; + + if (radeon_atom_get_max_vddc(rdev, &vddc) == 0) + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = vddc; + } + if (rdev->flags & RADEON_IS_IGP) { /* skip invalid modes */ if (rdev->pm.power_state[state_index].clock_info[mode_index].sclk == 0) @@ -2630,7 +2638,35 @@ void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 v atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } +int radeon_atom_get_max_vddc(struct radeon_device *rdev, + u16 *voltage) +{ + union set_voltage args; + int index = GetIndexIntoMasterTable(COMMAND, SetVoltage); + u8 frev, crev; + + if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev)) + return -EINVAL; + + switch (crev) { + case 1: + return -EINVAL; + case 2: + args.v2.ucVoltageType = SET_VOLTAGE_GET_MAX_VOLTAGE; + args.v2.ucVoltageMode = 0; + args.v2.usVoltageLevel = 0; + + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + + *voltage = le16_to_cpu(args.v2.usVoltageLevel); + break; + default: + DRM_ERROR("Unknown table version %d, %d\n", frev, crev); + return -EINVAL; + } + return 0; +} void radeon_atom_initialize_bios_scratch_regs(struct drm_device *dev) { -- cgit v1.2.3 From 8c9f3aaf8e174ca914889ab7a916586f8fd1e641 Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Thu, 16 Jun 2011 09:19:13 -0700 Subject: drm/i915: split page flip queueing into per-chipset functions This makes things a little clearer and prevents us from running old code on a new chipset that may not be supported. Signed-off-by: Jesse Barnes Reviewied-by: Ben Widawsky Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/i915_drv.h | 3 + drivers/gpu/drm/i915/intel_display.c | 269 ++++++++++++++++++++++++----------- 2 files changed, 189 insertions(+), 83 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index f63ee162f124..eddabf68e97a 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -211,6 +211,9 @@ struct drm_i915_display_funcs { void (*fdi_link_train)(struct drm_crtc *crtc); void (*init_clock_gating)(struct drm_device *dev); void (*init_pch_clock_gating)(struct drm_device *dev); + int (*queue_flip)(struct drm_device *dev, struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj); /* clock updates for mode set */ /* cursor updates */ /* render clock increase/decrease */ diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index aa43e7be6053..0f4ed2a575a4 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -6261,6 +6261,164 @@ void intel_prepare_page_flip(struct drm_device *dev, int plane) spin_unlock_irqrestore(&dev->event_lock, flags); } +static int intel_gen2_queue_flip(struct drm_device *dev, + struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + unsigned long offset; + u32 flip_mask; + int ret; + + ret = intel_pin_and_fence_fb_obj(dev, obj, LP_RING(dev_priv)); + if (ret) + goto out; + + /* Offset into the new buffer for cases of shared fbs between CRTCs */ + offset = crtc->y * fb->pitch + crtc->x * fb->bits_per_pixel/8; + + ret = BEGIN_LP_RING(6); + if (ret) + goto out; + + /* Can't queue multiple flips, so wait for the previous + * one to finish before executing the next. + */ + if (intel_crtc->plane) + flip_mask = MI_WAIT_FOR_PLANE_B_FLIP; + else + flip_mask = MI_WAIT_FOR_PLANE_A_FLIP; + OUT_RING(MI_WAIT_FOR_EVENT | flip_mask); + OUT_RING(MI_NOOP); + OUT_RING(MI_DISPLAY_FLIP | + MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); + OUT_RING(fb->pitch); + OUT_RING(obj->gtt_offset + offset); + OUT_RING(MI_NOOP); + ADVANCE_LP_RING(); +out: + return ret; +} + +static int intel_gen3_queue_flip(struct drm_device *dev, + struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + unsigned long offset; + u32 flip_mask; + int ret; + + ret = intel_pin_and_fence_fb_obj(dev, obj, LP_RING(dev_priv)); + if (ret) + goto out; + + /* Offset into the new buffer for cases of shared fbs between CRTCs */ + offset = crtc->y * fb->pitch + crtc->x * fb->bits_per_pixel/8; + + ret = BEGIN_LP_RING(6); + if (ret) + goto out; + + if (intel_crtc->plane) + flip_mask = MI_WAIT_FOR_PLANE_B_FLIP; + else + flip_mask = MI_WAIT_FOR_PLANE_A_FLIP; + OUT_RING(MI_WAIT_FOR_EVENT | flip_mask); + OUT_RING(MI_NOOP); + OUT_RING(MI_DISPLAY_FLIP_I915 | + MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); + OUT_RING(fb->pitch); + OUT_RING(obj->gtt_offset + offset); + OUT_RING(MI_NOOP); + + ADVANCE_LP_RING(); +out: + return ret; +} + +static int intel_gen4_queue_flip(struct drm_device *dev, + struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + uint32_t pf, pipesrc; + int ret; + + ret = intel_pin_and_fence_fb_obj(dev, obj, LP_RING(dev_priv)); + if (ret) + goto out; + + ret = BEGIN_LP_RING(4); + if (ret) + goto out; + + /* i965+ uses the linear or tiled offsets from the + * Display Registers (which do not change across a page-flip) + * so we need only reprogram the base address. + */ + OUT_RING(MI_DISPLAY_FLIP | + MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); + OUT_RING(fb->pitch); + OUT_RING(obj->gtt_offset | obj->tiling_mode); + + /* XXX Enabling the panel-fitter across page-flip is so far + * untested on non-native modes, so ignore it for now. + * pf = I915_READ(pipe == 0 ? PFA_CTL_1 : PFB_CTL_1) & PF_ENABLE; + */ + pf = 0; + pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff; + OUT_RING(pf | pipesrc); + ADVANCE_LP_RING(); +out: + return ret; +} + +static int intel_gen6_queue_flip(struct drm_device *dev, + struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + uint32_t pf, pipesrc; + int ret; + + ret = intel_pin_and_fence_fb_obj(dev, obj, LP_RING(dev_priv)); + if (ret) + goto out; + + ret = BEGIN_LP_RING(4); + if (ret) + goto out; + + OUT_RING(MI_DISPLAY_FLIP | + MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); + OUT_RING(fb->pitch | obj->tiling_mode); + OUT_RING(obj->gtt_offset); + + pf = I915_READ(PF_CTL(intel_crtc->pipe)) & PF_ENABLE; + pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff; + OUT_RING(pf | pipesrc); + ADVANCE_LP_RING(); +out: + return ret; +} + +static int intel_default_queue_flip(struct drm_device *dev, + struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj) +{ + return -ENODEV; +} + static int intel_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb, struct drm_pending_vblank_event *event) @@ -6271,9 +6429,7 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, struct drm_i915_gem_object *obj; struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct intel_unpin_work *work; - unsigned long flags, offset; - int pipe = intel_crtc->pipe; - u32 pf, pipesrc; + unsigned long flags; int ret; work = kzalloc(sizeof *work, GFP_KERNEL); @@ -6302,9 +6458,6 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, obj = intel_fb->obj; mutex_lock(&dev->struct_mutex); - ret = intel_pin_and_fence_fb_obj(dev, obj, LP_RING(dev_priv)); - if (ret) - goto cleanup_work; /* Reference the objects for the scheduled work. */ drm_gem_object_reference(&work->old_fb_obj->base); @@ -6316,91 +6469,18 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, if (ret) goto cleanup_objs; - if (IS_GEN3(dev) || IS_GEN2(dev)) { - u32 flip_mask; - - /* Can't queue multiple flips, so wait for the previous - * one to finish before executing the next. - */ - ret = BEGIN_LP_RING(2); - if (ret) - goto cleanup_objs; - - if (intel_crtc->plane) - flip_mask = MI_WAIT_FOR_PLANE_B_FLIP; - else - flip_mask = MI_WAIT_FOR_PLANE_A_FLIP; - OUT_RING(MI_WAIT_FOR_EVENT | flip_mask); - OUT_RING(MI_NOOP); - ADVANCE_LP_RING(); - } - work->pending_flip_obj = obj; work->enable_stall_check = true; - /* Offset into the new buffer for cases of shared fbs between CRTCs */ - offset = crtc->y * fb->pitch + crtc->x * fb->bits_per_pixel/8; - - ret = BEGIN_LP_RING(4); - if (ret) - goto cleanup_objs; - /* Block clients from rendering to the new back buffer until * the flip occurs and the object is no longer visible. */ atomic_add(1 << intel_crtc->plane, &work->old_fb_obj->pending_flip); - switch (INTEL_INFO(dev)->gen) { - case 2: - OUT_RING(MI_DISPLAY_FLIP | - MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch); - OUT_RING(obj->gtt_offset + offset); - OUT_RING(MI_NOOP); - break; - - case 3: - OUT_RING(MI_DISPLAY_FLIP_I915 | - MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch); - OUT_RING(obj->gtt_offset + offset); - OUT_RING(MI_NOOP); - break; - - case 4: - case 5: - /* i965+ uses the linear or tiled offsets from the - * Display Registers (which do not change across a page-flip) - * so we need only reprogram the base address. - */ - OUT_RING(MI_DISPLAY_FLIP | - MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch); - OUT_RING(obj->gtt_offset | obj->tiling_mode); - - /* XXX Enabling the panel-fitter across page-flip is so far - * untested on non-native modes, so ignore it for now. - * pf = I915_READ(pipe == 0 ? PFA_CTL_1 : PFB_CTL_1) & PF_ENABLE; - */ - pf = 0; - pipesrc = I915_READ(PIPESRC(pipe)) & 0x0fff0fff; - OUT_RING(pf | pipesrc); - break; - - case 6: - case 7: - OUT_RING(MI_DISPLAY_FLIP | - MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch | obj->tiling_mode); - OUT_RING(obj->gtt_offset); - - pf = I915_READ(PF_CTL(pipe)) & PF_ENABLE; - pipesrc = I915_READ(PIPESRC(pipe)) & 0x0fff0fff; - OUT_RING(pf | pipesrc); - break; - } - ADVANCE_LP_RING(); + ret = dev_priv->display.queue_flip(dev, crtc, fb, obj); + if (ret) + goto cleanup_pending; mutex_unlock(&dev->struct_mutex); @@ -6408,10 +6488,11 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, return 0; +cleanup_pending: + atomic_sub(1 << intel_crtc->plane, &work->old_fb_obj->pending_flip); cleanup_objs: drm_gem_object_unreference(&work->old_fb_obj->base); drm_gem_object_unreference(&obj->base); -cleanup_work: mutex_unlock(&dev->struct_mutex); spin_lock_irqsave(&dev->event_lock, flags); @@ -7656,6 +7737,28 @@ static void intel_init_display(struct drm_device *dev) else dev_priv->display.get_fifo_size = i830_get_fifo_size; } + + /* Default just returns -ENODEV to indicate unsupported */ + dev_priv->display.queue_flip = intel_default_queue_flip; + + switch (INTEL_INFO(dev)->gen) { + case 2: + dev_priv->display.queue_flip = intel_gen2_queue_flip; + break; + + case 3: + dev_priv->display.queue_flip = intel_gen3_queue_flip; + break; + + case 4: + case 5: + dev_priv->display.queue_flip = intel_gen4_queue_flip; + break; + + case 6: + dev_priv->display.queue_flip = intel_gen6_queue_flip; + break; + } } /* -- cgit v1.2.3 From 7c9017e5b77118439952fe8dc22809bae4fae4b6 Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Thu, 16 Jun 2011 12:18:54 -0700 Subject: drm/i915: add Ivy Bridge page flip support Use the blit ring for submitting flips since the render ring doesn't generate flip complete interrupts. Fixes bugs: https://bugs.freedesktop.org/show_bug.cgi?id=38362 https://bugs.freedesktop.org/show_bug.cgi?id=38392 https://bugs.freedesktop.org/show_bug.cgi?id=38393 Signed-off-by: Jesse Barnes Reviewed-by: Daniel Vetter Reviewed-by: Kenneth Graunke Reviewed-by: Ben Widawsky Tested-by: Jian J Zhao Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/intel_display.c | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 0f4ed2a575a4..21b6f93fe919 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -6411,6 +6411,39 @@ out: return ret; } +/* + * On gen7 we currently use the blit ring because (in early silicon at least) + * the render ring doesn't give us interrpts for page flip completion, which + * means clients will hang after the first flip is queued. Fortunately the + * blit ring generates interrupts properly, so use it instead. + */ +static int intel_gen7_queue_flip(struct drm_device *dev, + struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + struct intel_ring_buffer *ring = &dev_priv->ring[BCS]; + int ret; + + ret = intel_pin_and_fence_fb_obj(dev, obj, ring); + if (ret) + goto out; + + ret = intel_ring_begin(ring, 4); + if (ret) + goto out; + + intel_ring_emit(ring, MI_DISPLAY_FLIP_I915 | (intel_crtc->plane << 19)); + intel_ring_emit(ring, (fb->pitch | obj->tiling_mode)); + intel_ring_emit(ring, (obj->gtt_offset)); + intel_ring_emit(ring, (MI_NOOP)); + intel_ring_advance(ring); +out: + return ret; +} + static int intel_default_queue_flip(struct drm_device *dev, struct drm_crtc *crtc, struct drm_framebuffer *fb, @@ -7758,6 +7791,9 @@ static void intel_init_display(struct drm_device *dev) case 6: dev_priv->display.queue_flip = intel_gen6_queue_flip; break; + case 7: + dev_priv->display.queue_flip = intel_gen7_queue_flip; + break; } } -- cgit v1.2.3 From 3142b651ad2232cf0e375c291ee4b893c8559df5 Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Mon, 27 Jun 2011 16:18:17 -0700 Subject: drm/ttm: use shmem_read_mapping_page Soon tmpfs will stop supporting ->readpage and read_mapping_page(): once "tmpfs: add shmem_read_mapping_page_gfp" has been applied, this patch can be applied to ease the transition. ttm_tt_swapin() and ttm_tt_swapout() use shmem_read_mapping_page() in place of read_mapping_page(), since their swap_space has been created with shmem_file_setup(). Signed-off-by: Hugh Dickins Cc: Christoph Hellwig Cc: Thomas Hellstrom Cc: Dave Airlie Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/gpu/drm/ttm/ttm_tt.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c index 90e23e0bfadb..58c271ebc0f7 100644 --- a/drivers/gpu/drm/ttm/ttm_tt.c +++ b/drivers/gpu/drm/ttm/ttm_tt.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include #include @@ -484,7 +485,7 @@ static int ttm_tt_swapin(struct ttm_tt *ttm) swap_space = swap_storage->f_path.dentry->d_inode->i_mapping; for (i = 0; i < ttm->num_pages; ++i) { - from_page = read_mapping_page(swap_space, i, NULL); + from_page = shmem_read_mapping_page(swap_space, i); if (IS_ERR(from_page)) { ret = PTR_ERR(from_page); goto out_err; @@ -557,7 +558,7 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage) from_page = ttm->pages[i]; if (unlikely(from_page == NULL)) continue; - to_page = read_mapping_page(swap_space, i, NULL); + to_page = shmem_read_mapping_page(swap_space, i); if (unlikely(IS_ERR(to_page))) { ret = PTR_ERR(to_page); goto out_err; -- cgit v1.2.3 From 5949eac4d9b5bf936c12cb7ec3a09084c1326834 Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Mon, 27 Jun 2011 16:18:18 -0700 Subject: drm/i915: use shmem_read_mapping_page Soon tmpfs will stop supporting ->readpage and read_cache_page_gfp(): once "tmpfs: add shmem_read_mapping_page_gfp" has been applied, this patch can be applied to ease the transition. Make i915_gem_object_get_pages_gtt() use shmem_read_mapping_page_gfp() in the one place it's needed; elsewhere use shmem_read_mapping_page(), with the mapping's gfp_mask properly initialized. Forget about __GFP_COLD: since tmpfs initializes its pages with memset, asking for a cold page is counter-productive. Include linux/shmem_fs.h also in drm_gem.c: with shmem_file_setup() now declared there too, we shall remove the prototype from linux/mm.h later. Signed-off-by: Hugh Dickins Cc: Christoph Hellwig Cc: Chris Wilson Cc: Keith Packard Cc: Dave Airlie Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/gpu/drm/drm_gem.c | 1 + drivers/gpu/drm/i915/i915_gem.c | 31 ++++++++++++++----------------- 2 files changed, 15 insertions(+), 17 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index 74e4ff578017..4012fe423460 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -34,6 +34,7 @@ #include #include #include +#include #include "drmP.h" /** @file drm_gem.c diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index c6389de53161..fa560ceba667 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -31,6 +31,7 @@ #include "i915_drv.h" #include "i915_trace.h" #include "intel_drv.h" +#include #include #include #include @@ -359,8 +360,7 @@ i915_gem_shmem_pread_fast(struct drm_device *dev, if ((page_offset + remain) > PAGE_SIZE) page_length = PAGE_SIZE - page_offset; - page = read_cache_page_gfp(mapping, offset >> PAGE_SHIFT, - GFP_HIGHUSER | __GFP_RECLAIMABLE); + page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT); if (IS_ERR(page)) return PTR_ERR(page); @@ -463,8 +463,7 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, if ((data_page_offset + page_length) > PAGE_SIZE) page_length = PAGE_SIZE - data_page_offset; - page = read_cache_page_gfp(mapping, offset >> PAGE_SHIFT, - GFP_HIGHUSER | __GFP_RECLAIMABLE); + page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT); if (IS_ERR(page)) { ret = PTR_ERR(page); goto out; @@ -797,8 +796,7 @@ i915_gem_shmem_pwrite_fast(struct drm_device *dev, if ((page_offset + remain) > PAGE_SIZE) page_length = PAGE_SIZE - page_offset; - page = read_cache_page_gfp(mapping, offset >> PAGE_SHIFT, - GFP_HIGHUSER | __GFP_RECLAIMABLE); + page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT); if (IS_ERR(page)) return PTR_ERR(page); @@ -907,8 +905,7 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev, if ((data_page_offset + page_length) > PAGE_SIZE) page_length = PAGE_SIZE - data_page_offset; - page = read_cache_page_gfp(mapping, offset >> PAGE_SHIFT, - GFP_HIGHUSER | __GFP_RECLAIMABLE); + page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT); if (IS_ERR(page)) { ret = PTR_ERR(page); goto out; @@ -1558,12 +1555,10 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj, inode = obj->base.filp->f_path.dentry->d_inode; mapping = inode->i_mapping; + gfpmask |= mapping_gfp_mask(mapping); + for (i = 0; i < page_count; i++) { - page = read_cache_page_gfp(mapping, i, - GFP_HIGHUSER | - __GFP_COLD | - __GFP_RECLAIMABLE | - gfpmask); + page = shmem_read_mapping_page_gfp(mapping, i, gfpmask); if (IS_ERR(page)) goto err_pages; @@ -3565,6 +3560,7 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev, { struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj; + struct address_space *mapping; obj = kzalloc(sizeof(*obj), GFP_KERNEL); if (obj == NULL) @@ -3575,6 +3571,9 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev, return NULL; } + mapping = obj->base.filp->f_path.dentry->d_inode->i_mapping; + mapping_set_gfp_mask(mapping, GFP_HIGHUSER | __GFP_RECLAIMABLE); + i915_gem_info_add_obj(dev_priv, size); obj->base.write_domain = I915_GEM_DOMAIN_CPU; @@ -3950,8 +3949,7 @@ void i915_gem_detach_phys_object(struct drm_device *dev, page_count = obj->base.size / PAGE_SIZE; for (i = 0; i < page_count; i++) { - struct page *page = read_cache_page_gfp(mapping, i, - GFP_HIGHUSER | __GFP_RECLAIMABLE); + struct page *page = shmem_read_mapping_page(mapping, i); if (!IS_ERR(page)) { char *dst = kmap_atomic(page); memcpy(dst, vaddr + i*PAGE_SIZE, PAGE_SIZE); @@ -4012,8 +4010,7 @@ i915_gem_attach_phys_object(struct drm_device *dev, struct page *page; char *dst, *src; - page = read_cache_page_gfp(mapping, i, - GFP_HIGHUSER | __GFP_RECLAIMABLE); + page = shmem_read_mapping_page(mapping, i); if (IS_ERR(page)) return PTR_ERR(page); -- cgit v1.2.3 From e2377fe0b65e3c7577ff6df1701c56ef477d336f Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Mon, 27 Jun 2011 16:18:19 -0700 Subject: drm/i915: use shmem_truncate_range The interface to ->truncate_range is changing very slightly: once "tmpfs: take control of its truncate_range" has been applied, this can be applied. For now there is only a slight inefficiency while this remains unapplied, but it will soon become essential for managing shmem's use of swap. Change i915_gem_object_truncate() to use shmem_truncate_range() directly: which should also spare i915 later change if we switch from inode_operations->truncate_range to file_operations->fallocate. Signed-off-by: Hugh Dickins Cc: Christoph Hellwig Cc: Chris Wilson Cc: Keith Packard Cc: Dave Airlie Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/gpu/drm/i915/i915_gem.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index fa560ceba667..85f713746a1f 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1696,13 +1696,10 @@ i915_gem_object_truncate(struct drm_i915_gem_object *obj) /* Our goal here is to return as much of the memory as * is possible back to the system as we are called from OOM. * To do this we must instruct the shmfs to drop all of its - * backing pages, *now*. Here we mirror the actions taken - * when by shmem_delete_inode() to release the backing store. + * backing pages, *now*. */ inode = obj->base.filp->f_path.dentry->d_inode; - truncate_inode_pages(inode->i_mapping, 0); - if (inode->i_op->truncate_range) - inode->i_op->truncate_range(inode, 0, (loff_t)-1); + shmem_truncate_range(inode, 0, (loff_t)-1); obj->madv = __I915_MADV_PURGED; } -- cgit v1.2.3 From ecbec53b1d00ba582f71b210ed96cafc05ebd189 Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Mon, 27 Jun 2011 16:18:20 -0700 Subject: drm/i915: more struct_mutex locking When auditing the locking in i915_gem.c (for a prospective change which I then abandoned), I noticed two places where struct_mutex is not held across GEM object manipulations that would usually require it. Since one is in initial setup and the other in driver unload, I'm guessing the mutex is not required for either; but post a patch in case it is. Signed-off-by: Hugh Dickins Cc: Chris Wilson Cc: Keith Packard Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/gpu/drm/i915/i915_dma.c | 3 +-- drivers/gpu/drm/i915/intel_overlay.c | 5 +++++ 2 files changed, 6 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index 0239e9974bf2..2b79588541e7 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -2182,9 +2182,8 @@ int i915_driver_unload(struct drm_device *dev) /* Flush any outstanding unpin_work. */ flush_workqueue(dev_priv->wq); - i915_gem_free_all_phys_object(dev); - mutex_lock(&dev->struct_mutex); + i915_gem_free_all_phys_object(dev); i915_gem_cleanup_ringbuffer(dev); mutex_unlock(&dev->struct_mutex); if (I915_HAS_FBC(dev) && i915_powersave) diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index a670c006982e..56a8e2aea19c 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -1416,6 +1416,8 @@ void intel_setup_overlay(struct drm_device *dev) goto out_free; overlay->reg_bo = reg_bo; + mutex_lock(&dev->struct_mutex); + if (OVERLAY_NEEDS_PHYSICAL(dev)) { ret = i915_gem_attach_phys_object(dev, reg_bo, I915_GEM_PHYS_OVERLAY_REGS, @@ -1440,6 +1442,8 @@ void intel_setup_overlay(struct drm_device *dev) } } + mutex_unlock(&dev->struct_mutex); + /* init all values */ overlay->color_key = 0x0101fe; overlay->brightness = -19; @@ -1464,6 +1468,7 @@ out_unpin_bo: i915_gem_object_unpin(reg_bo); out_free_bo: drm_gem_object_unreference(®_bo->base); + mutex_unlock(&dev->struct_mutex); out_free: kfree(overlay); return; -- cgit v1.2.3 From 25732821cb965f00475922ca46e84f78e4bada95 Mon Sep 17 00:00:00 2001 From: Ben Widawsky Date: Fri, 24 Jun 2011 14:31:47 -0700 Subject: drm/i915: forcewake fix after reset The failure is as follows: 1. Userspace gets forcewake lock, lock count >=1 2. GPU hang/reset occurs (forcewake bit is reset) 3. count is now incorrect The failure can only occur when using the forcewake userspace lock. This has the unfortunate consequence of messing up the driver as well as userspace, unless userspace closes the debugfs file, the kernel will never end up waking the GT since the refcount will be > 1. The solution is to try to recover the correct forcewake state based on the refcount. There is a period of time where userspace reads/writes may occur after the reset, before the GT has been forcewaked. The interface was never designed to be a perfect solution for userspace reads/writes, and the kernel portion is fixed by this patch. Suggested-by: Chris Wilson Signed-off-by: Ben Widawsky Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/i915_drv.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 0defd4270594..609358faaa90 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -579,6 +579,9 @@ int i915_reset(struct drm_device *dev, u8 flags) } else switch (INTEL_INFO(dev)->gen) { case 6: ret = gen6_do_reset(dev, flags); + /* If reset with a user forcewake, try to restore */ + if (atomic_read(&dev_priv->forcewake_count)) + __gen6_gt_force_wake_get(dev_priv); break; case 5: ret = ironlake_do_reset(dev, flags); -- cgit v1.2.3 From f01c22fd59aa10a3738ede20fd4b9b6fd1e2eac3 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 28 Jun 2011 11:48:51 +0100 Subject: drm/i915: Use chipset-specific irq installers Konstantin Belousov pointed out that 4697995b98417 replaced the generic i915_driver_irq_*install() functions with chipset specific routines accessible only through driver->irq_*install(). So update the sanity check in i915_request_wait() to match. Signed-off-by: Chris Wilson Signed-off-by: Keith Packard --- drivers/gpu/drm/i915/i915_gem.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index c6389de53161..e81dbe57c773 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2080,8 +2080,8 @@ i915_wait_request(struct intel_ring_buffer *ring, if (!ier) { DRM_ERROR("something (likely vbetool) disabled " "interrupts, re-enabling\n"); - i915_driver_irq_preinstall(ring->dev); - i915_driver_irq_postinstall(ring->dev); + ring->dev->driver->irq_preinstall(ring->dev); + ring->dev->driver->irq_postinstall(ring->dev); } trace_i915_gem_request_wait_begin(ring, seqno); -- cgit v1.2.3 From b271a988eb9c3944c50fb62c21ac61860090d3ba Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 24 Jun 2011 13:15:38 +0000 Subject: drm/radeon/kms: increase rom size for atrm method The vbios rom is >64k on a lot of modern asics. Increase the fetch size for atrm to make sure we don't miss part of a larger rom. Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/radeon_bios.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm') diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c index 1aba85cad1a8..3fc5fa1aefd0 100644 --- a/drivers/gpu/drm/radeon/radeon_bios.c +++ b/drivers/gpu/drm/radeon/radeon_bios.c @@ -104,7 +104,7 @@ static bool radeon_read_bios(struct radeon_device *rdev) static bool radeon_atrm_get_bios(struct radeon_device *rdev) { int ret; - int size = 64 * 1024; + int size = 256 * 1024; int i; if (!radeon_atrm_supported(rdev->pdev)) -- cgit v1.2.3