Merge branch 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel into drm-fixes

* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel:
  drm/i915: fix corruptions on i8xx due to relaxed fencing
  drm/i915: skip FDI & PCH enabling for DP_A
  agp/intel: Experiment with a 855GM GWB bit
  drm/i915: don't enable FDI & transcoder interrupts after all
  drm/i915: Ignore a hung GPU when flushing the framebuffer prior to a switch

5 files changed, 126 insertions, 46 deletions

diff --git a/drivers/char/agp/intel-agp.h b/drivers/char/agp/intel-agp.h
index c195bfeade11..5feebe2800e9 100644
--- a/drivers/char/agp/intel-agp.h
+++ b/drivers/char/agp/intel-agp.h
@@ -130,6 +130,7 @@
 #define INTEL_GMCH_GMS_STOLEN_352M      (0xd << 4)
 
 #define I915_IFPADDR    0x60
+#define I830_HIC        0x70
 
 /* Intel 965G registers */
 #define I965_MSAC 0x62
diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index fab3d3265adb..0d09b537bb9a 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -21,6 +21,7 @@
 #include <linux/kernel.h>
 #include <linux/pagemap.h>
 #include <linux/agp_backend.h>
+#include <linux/delay.h>
 #include <asm/smp.h>
 #include "agp.h"
 #include "intel-agp.h"
@@ -70,12 +71,8 @@ static struct _intel_private {
         u32 __iomem *gtt;               /* I915G */
         bool clear_fake_agp; /* on first access via agp, fill with scratch */
         int num_dcache_entries;
-        union {
-                void __iomem *i9xx_flush_page;
-                void *i8xx_flush_page;
-        };
+        void __iomem *i9xx_flush_page;
         char *i81x_gtt_table;
-        struct page *i8xx_page;
         struct resource ifp_resource;
         int resource_valid;
         struct page *scratch_page;
@@ -722,28 +719,6 @@ static int intel_fake_agp_fetch_size(void)
 
 static void i830_cleanup(void)
 {
-        if (intel_private.i8xx_flush_page) {
-                kunmap(intel_private.i8xx_flush_page);
-                intel_private.i8xx_flush_page = NULL;
-        }
-
-        __free_page(intel_private.i8xx_page);
-        intel_private.i8xx_page = NULL;
-}
-
-static void intel_i830_setup_flush(void)
-{
-        /* return if we've already set the flush mechanism up */
-        if (intel_private.i8xx_page)
-                return;
-
-        intel_private.i8xx_page = alloc_page(GFP_KERNEL);
-        if (!intel_private.i8xx_page)
-                return;
-
-        intel_private.i8xx_flush_page = kmap(intel_private.i8xx_page);
-        if (!intel_private.i8xx_flush_page)
-                i830_cleanup();
 }
 
 /* The chipset_flush interface needs to get data that has already been
@@ -758,14 +733,27 @@ static void intel_i830_setup_flush(void)
  */
 static void i830_chipset_flush(void)
 {
-        unsigned int *pg = intel_private.i8xx_flush_page;
+        unsigned long timeout = jiffies + msecs_to_jiffies(1000);
+
+        /* Forcibly evict everything from the CPU write buffers.
+         * clflush appears to be insufficient.
+         */
+        wbinvd_on_all_cpus();
+
+        /* Now we've only seen documents for this magic bit on 855GM,
+         * we hope it exists for the other gen2 chipsets...
+         *
+         * Also works as advertised on my 845G.
+         */
+        writel(readl(intel_private.registers+I830_HIC) | (1<<31),
+               intel_private.registers+I830_HIC);
 
-        memset(pg, 0, 1024);
+        while (readl(intel_private.registers+I830_HIC) & (1<<31)) {
+                if (time_after(jiffies, timeout))
+                        break;
 
-        if (cpu_has_clflush)
-                clflush_cache_range(pg, 1024);
-        else if (wbinvd_on_all_cpus() != 0)
-                printk(KERN_ERR "Timed out waiting for cache flush.\n");
+                udelay(50);
+        }
 }
 
 static void i830_write_entry(dma_addr_t addr, unsigned int entry,
@@ -849,8 +837,6 @@ static int i830_setup(void)
 
         intel_private.gtt_bus_addr = reg_addr + I810_PTE_BASE;
 
-        intel_i830_setup_flush();
-
         return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 22a32b9932c5..79a04fde69b5 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -184,7 +184,7 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 static bool
 i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
 {
-        int tile_width;
+        int tile_width, tile_height;
 
         /* Linear is always fine */
         if (tiling_mode == I915_TILING_NONE)
@@ -215,6 +215,20 @@ i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode)
                 }
         }
 
+        if (IS_GEN2(dev) ||
+            (tiling_mode == I915_TILING_Y && HAS_128_BYTE_Y_TILING(dev)))
+                tile_height = 32;
+        else
+                tile_height = 8;
+        /* i8xx is strange: It has 2 interleaved rows of tiles, so needs an even
+         * number of tile rows. */
+        if (IS_GEN2(dev))
+                tile_height *= 2;
+
+        /* Size needs to be aligned to a full tile row */
+        if (size & (tile_height * stride - 1))
+                return false;
+
         /* 965+ just needs multiples of tile width */
         if (INTEL_INFO(dev)->gen >= 4) {
                 if (stride & (tile_width - 1))
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 97f946dcc1aa..8a9e08bf1cf7 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -316,6 +316,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
         struct drm_mode_config *mode_config = &dev->mode_config;
         struct intel_encoder *encoder;
 
+        DRM_DEBUG_KMS("running encoder hotplug functions\n");
+
         list_for_each_entry(encoder, &mode_config->encoder_list, base.head)
                 if (encoder->hot_plug)
                         encoder->hot_plug(encoder);
@@ -1649,9 +1651,7 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
         } else {
                 hotplug_mask = SDE_CRT_HOTPLUG | SDE_PORTB_HOTPLUG |
                                SDE_PORTC_HOTPLUG | SDE_PORTD_HOTPLUG;
-                hotplug_mask |= SDE_AUX_MASK | SDE_FDI_MASK | SDE_TRANS_MASK;
-                I915_WRITE(FDI_RXA_IMR, 0);
-                I915_WRITE(FDI_RXB_IMR, 0);
+                hotplug_mask |= SDE_AUX_MASK;
         }
 
         dev_priv->pch_irq_mask = ~hotplug_mask;
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 3b006536b3d2..9ca1bb2554fc 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1630,19 +1630,19 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
                 struct drm_i915_gem_object *obj = to_intel_framebuffer(old_fb)->obj;
 
                 wait_event(dev_priv->pending_flip_queue,
+                           atomic_read(&dev_priv->mm.wedged) ||
                            atomic_read(&obj->pending_flip) == 0);
 
                 /* Big Hammer, we also need to ensure that any pending
                  * MI_WAIT_FOR_EVENT inside a user batch buffer on the
                  * current scanout is retired before unpinning the old
                  * framebuffer.
+                 *
+                 * This should only fail upon a hung GPU, in which case we
+                 * can safely continue.
                  */
                 ret = i915_gem_object_flush_gpu(obj, false);
-                if (ret) {
-                        i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj);
-                        mutex_unlock(&dev->struct_mutex);
-                        return ret;
-                }
+                (void) ret;
         }
 
         ret = intel_pipe_set_base_atomic(crtc, crtc->fb, x, y,
@@ -2045,6 +2045,31 @@ static void intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
                    atomic_read(&obj->pending_flip) == 0);
 }
 
+static bool intel_crtc_driving_pch(struct drm_crtc *crtc)
+{
+        struct drm_device *dev = crtc->dev;
+        struct drm_mode_config *mode_config = &dev->mode_config;
+        struct intel_encoder *encoder;
+
+        /*
+         * If there's a non-PCH eDP on this crtc, it must be DP_A, and that
+         * must be driven by its own crtc; no sharing is possible.
+         */
+        list_for_each_entry(encoder, &mode_config->encoder_list, base.head) {
+                if (encoder->base.crtc != crtc)
+                        continue;
+
+                switch (encoder->type) {
+                case INTEL_OUTPUT_EDP:
+                        if (!intel_encoder_is_pch_edp(&encoder->base))
+                                return false;
+                        continue;
+                }
+        }
+
+        return true;
+}
+
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
 {
         struct drm_device *dev = crtc->dev;
@@ -2053,6 +2078,7 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
         int pipe = intel_crtc->pipe;
         int plane = intel_crtc->plane;
         u32 reg, temp;
+        bool is_pch_port = false;
 
         if (intel_crtc->active)
                 return;
@@ -2066,7 +2092,56 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
                         I915_WRITE(PCH_LVDS, temp | LVDS_PORT_EN);
         }
 
-        ironlake_fdi_enable(crtc);
+        is_pch_port = intel_crtc_driving_pch(crtc);
+
+        if (is_pch_port)
+                ironlake_fdi_enable(crtc);
+        else {
+                /* disable CPU FDI tx and PCH FDI rx */
+                reg = FDI_TX_CTL(pipe);
+                temp = I915_READ(reg);
+                I915_WRITE(reg, temp & ~FDI_TX_ENABLE);
+                POSTING_READ(reg);
+
+                reg = FDI_RX_CTL(pipe);
+                temp = I915_READ(reg);
+                temp &= ~(0x7 << 16);
+                temp |= (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) << 11;
+                I915_WRITE(reg, temp & ~FDI_RX_ENABLE);
+
+                POSTING_READ(reg);
+                udelay(100);
+
+                /* Ironlake workaround, disable clock pointer after downing FDI */
+                if (HAS_PCH_IBX(dev))
+                        I915_WRITE(FDI_RX_CHICKEN(pipe),
+                                   I915_READ(FDI_RX_CHICKEN(pipe) &
+                                             ~FDI_RX_PHASE_SYNC_POINTER_ENABLE));
+
+                /* still set train pattern 1 */
+                reg = FDI_TX_CTL(pipe);
+                temp = I915_READ(reg);
+                temp &= ~FDI_LINK_TRAIN_NONE;
+                temp |= FDI_LINK_TRAIN_PATTERN_1;
+                I915_WRITE(reg, temp);
+
+                reg = FDI_RX_CTL(pipe);
+                temp = I915_READ(reg);
+                if (HAS_PCH_CPT(dev)) {
+                        temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+                        temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+                } else {
+                        temp &= ~FDI_LINK_TRAIN_NONE;
+                        temp |= FDI_LINK_TRAIN_PATTERN_1;
+                }
+                /* BPC in FDI rx is consistent with that in PIPECONF */
+                temp &= ~(0x07 << 16);
+                temp |= (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) << 11;
+                I915_WRITE(reg, temp);
+
+                POSTING_READ(reg);
+                udelay(100);
+        }
 
         /* Enable panel fitting for LVDS */
         if (dev_priv->pch_pf_size &&
@@ -2100,6 +2175,10 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
                 intel_flush_display_plane(dev, plane);
         }
 
+        /* Skip the PCH stuff if possible */
+        if (!is_pch_port)
+                goto done;
+
         /* For PCH output, training FDI link */
         if (IS_GEN6(dev))
                 gen6_fdi_link_train(crtc);
@@ -2184,7 +2263,7 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
         I915_WRITE(reg, temp | TRANS_ENABLE);
         if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100))
                 DRM_ERROR("failed to enable transcoder %d\n", pipe);
-
+done:
         intel_crtc_load_lut(crtc);
         intel_update_fbc(dev);
         intel_crtc_update_cursor(crtc, true);