+struct fdi_m_n {
+ u32 tu;
+ u32 gmch_m;
+ u32 gmch_n;
+ u32 link_m;
+ u32 link_n;
+};
+
+static void
+fdi_reduce_ratio(u32 *num, u32 *den)
+{
+ while (*num > 0xffffff || *den > 0xffffff) {
+ *num >>= 1;
+ *den >>= 1;
+ }
+}
+
+#define DATA_N 0x800000
+#define LINK_N 0x80000
+
+static void
+igdng_compute_m_n(int bytes_per_pixel, int nlanes,
+ int pixel_clock, int link_clock,
+ struct fdi_m_n *m_n)
+{
+ u64 temp;
+
+ m_n->tu = 64; /* default size */
+
+ temp = (u64) DATA_N * pixel_clock;
+ temp = div_u64(temp, link_clock);
+ m_n->gmch_m = div_u64(temp * bytes_per_pixel, nlanes);
+ m_n->gmch_n = DATA_N;
+ fdi_reduce_ratio(&m_n->gmch_m, &m_n->gmch_n);
+
+ temp = (u64) LINK_N * pixel_clock;
+ m_n->link_m = div_u64(temp, link_clock);
+ m_n->link_n = LINK_N;
+ fdi_reduce_ratio(&m_n->link_m, &m_n->link_n);
+}
+
+
+struct intel_watermark_params {
+ unsigned long fifo_size;
+ unsigned long max_wm;
+ unsigned long default_wm;
+ unsigned long guard_size;
+ unsigned long cacheline_size;
+};
+
+/* IGD has different values for various configs */
+static struct intel_watermark_params igd_display_wm = {
+ IGD_DISPLAY_FIFO,
+ IGD_MAX_WM,
+ IGD_DFT_WM,
+ IGD_GUARD_WM,
+ IGD_FIFO_LINE_SIZE
+};
+static struct intel_watermark_params igd_display_hplloff_wm = {
+ IGD_DISPLAY_FIFO,
+ IGD_MAX_WM,
+ IGD_DFT_HPLLOFF_WM,
+ IGD_GUARD_WM,
+ IGD_FIFO_LINE_SIZE
+};
+static struct intel_watermark_params igd_cursor_wm = {
+ IGD_CURSOR_FIFO,
+ IGD_CURSOR_MAX_WM,
+ IGD_CURSOR_DFT_WM,
+ IGD_CURSOR_GUARD_WM,
+ IGD_FIFO_LINE_SIZE,
+};
+static struct intel_watermark_params igd_cursor_hplloff_wm = {
+ IGD_CURSOR_FIFO,
+ IGD_CURSOR_MAX_WM,
+ IGD_CURSOR_DFT_WM,
+ IGD_CURSOR_GUARD_WM,
+ IGD_FIFO_LINE_SIZE
+};
+static struct intel_watermark_params i945_wm_info = {
+ I945_FIFO_SIZE,
+ I915_MAX_WM,
+ 1,
+ 2,
+ I915_FIFO_LINE_SIZE
+};
+static struct intel_watermark_params i915_wm_info = {
+ I915_FIFO_SIZE,
+ I915_MAX_WM,
+ 1,
+ 2,
+ I915_FIFO_LINE_SIZE
+};
+static struct intel_watermark_params i855_wm_info = {
+ I855GM_FIFO_SIZE,
+ I915_MAX_WM,
+ 1,
+ 2,
+ I830_FIFO_LINE_SIZE
+};
+static struct intel_watermark_params i830_wm_info = {
+ I830_FIFO_SIZE,
+ I915_MAX_WM,
+ 1,
+ 2,
+ I830_FIFO_LINE_SIZE
+};
+
+/**
+ * intel_calculate_wm - calculate watermark level
+ * @clock_in_khz: pixel clock
+ * @wm: chip FIFO params
+ * @pixel_size: display pixel size
+ * @latency_ns: memory latency for the platform
+ *
+ * Calculate the watermark level (the level at which the display plane will
+ * start fetching from memory again). Each chip has a different display
+ * FIFO size and allocation, so the caller needs to figure that out and pass
+ * in the correct intel_watermark_params structure.
+ *
+ * As the pixel clock runs, the FIFO will be drained at a rate that depends
+ * on the pixel size. When it reaches the watermark level, it'll start
+ * fetching FIFO line sized based chunks from memory until the FIFO fills
+ * past the watermark point. If the FIFO drains completely, a FIFO underrun
+ * will occur, and a display engine hang could result.
+ */
+static unsigned long intel_calculate_wm(unsigned long clock_in_khz,
+ struct intel_watermark_params *wm,
+ int pixel_size,
+ unsigned long latency_ns)
+{
+ long entries_required, wm_size;
+
+ entries_required = (clock_in_khz * pixel_size * latency_ns) / 1000000;
+ entries_required /= wm->cacheline_size;
+
+ DRM_DEBUG("FIFO entries required for mode: %d\n", entries_required);
+
+ wm_size = wm->fifo_size - (entries_required + wm->guard_size);
+
+ DRM_DEBUG("FIFO watermark level: %d\n", wm_size);
+
+ /* Don't promote wm_size to unsigned... */
+ if (wm_size > (long)wm->max_wm)
+ wm_size = wm->max_wm;
+ if (wm_size <= 0)
+ wm_size = wm->default_wm;
+ return wm_size;
+}
+
+struct cxsr_latency {
+ int is_desktop;
+ unsigned long fsb_freq;
+ unsigned long mem_freq;
+ unsigned long display_sr;
+ unsigned long display_hpll_disable;
+ unsigned long cursor_sr;
+ unsigned long cursor_hpll_disable;
+};
+
+static struct cxsr_latency cxsr_latency_table[] = {
+ {1, 800, 400, 3382, 33382, 3983, 33983}, /* DDR2-400 SC */
+ {1, 800, 667, 3354, 33354, 3807, 33807}, /* DDR2-667 SC */
+ {1, 800, 800, 3347, 33347, 3763, 33763}, /* DDR2-800 SC */
+
+ {1, 667, 400, 3400, 33400, 4021, 34021}, /* DDR2-400 SC */
+ {1, 667, 667, 3372, 33372, 3845, 33845}, /* DDR2-667 SC */
+ {1, 667, 800, 3386, 33386, 3822, 33822}, /* DDR2-800 SC */
+
+ {1, 400, 400, 3472, 33472, 4173, 34173}, /* DDR2-400 SC */
+ {1, 400, 667, 3443, 33443, 3996, 33996}, /* DDR2-667 SC */
+ {1, 400, 800, 3430, 33430, 3946, 33946}, /* DDR2-800 SC */
+
+ {0, 800, 400, 3438, 33438, 4065, 34065}, /* DDR2-400 SC */
+ {0, 800, 667, 3410, 33410, 3889, 33889}, /* DDR2-667 SC */
+ {0, 800, 800, 3403, 33403, 3845, 33845}, /* DDR2-800 SC */
+
+ {0, 667, 400, 3456, 33456, 4103, 34106}, /* DDR2-400 SC */
+ {0, 667, 667, 3428, 33428, 3927, 33927}, /* DDR2-667 SC */
+ {0, 667, 800, 3443, 33443, 3905, 33905}, /* DDR2-800 SC */
+
+ {0, 400, 400, 3528, 33528, 4255, 34255}, /* DDR2-400 SC */
+ {0, 400, 667, 3500, 33500, 4079, 34079}, /* DDR2-667 SC */
+ {0, 400, 800, 3487, 33487, 4029, 34029}, /* DDR2-800 SC */
+};
+
+static struct cxsr_latency *intel_get_cxsr_latency(int is_desktop, int fsb,
+ int mem)
+{
+ int i;
+ struct cxsr_latency *latency;
+
+ if (fsb == 0 || mem == 0)
+ return NULL;
+
+ for (i = 0; i < ARRAY_SIZE(cxsr_latency_table); i++) {
+ latency = &cxsr_latency_table[i];
+ if (is_desktop == latency->is_desktop &&
+ fsb == latency->fsb_freq && mem == latency->mem_freq)
+ break;
+ }
+ if (i >= ARRAY_SIZE(cxsr_latency_table)) {
+ DRM_DEBUG("Unknown FSB/MEM found, disable CxSR\n");
+ return NULL;
+ }
+ return latency;
+}
+
+static void igd_disable_cxsr(struct drm_device *dev)
+{
+ struct drm_i915_private *dev_priv = dev->dev_private;
+ u32 reg;
+
+ /* deactivate cxsr */
+ reg = I915_READ(DSPFW3);
+ reg &= ~(IGD_SELF_REFRESH_EN);
+ I915_WRITE(DSPFW3, reg);
+ DRM_INFO("Big FIFO is disabled\n");
+}
+
+static void igd_enable_cxsr(struct drm_device *dev, unsigned long clock,
+ int pixel_size)
+{
+ struct drm_i915_private *dev_priv = dev->dev_private;
+ u32 reg;
+ unsigned long wm;
+ struct cxsr_latency *latency;
+
+ latency = intel_get_cxsr_latency(IS_IGDG(dev), dev_priv->fsb_freq,
+ dev_priv->mem_freq);
+ if (!latency) {
+ DRM_DEBUG("Unknown FSB/MEM found, disable CxSR\n");
+ igd_disable_cxsr(dev);
+ return;
+ }
+
+ /* Display SR */
+ wm = intel_calculate_wm(clock, &igd_display_wm, pixel_size,
+ latency->display_sr);
+ reg = I915_READ(DSPFW1);
+ reg &= 0x7fffff;
+ reg |= wm << 23;
+ I915_WRITE(DSPFW1, reg);
+ DRM_DEBUG("DSPFW1 register is %x\n", reg);
+
+ /* cursor SR */
+ wm = intel_calculate_wm(clock, &igd_cursor_wm, pixel_size,
+ latency->cursor_sr);
+ reg = I915_READ(DSPFW3);
+ reg &= ~(0x3f << 24);
+ reg |= (wm & 0x3f) << 24;
+ I915_WRITE(DSPFW3, reg);
+
+ /* Display HPLL off SR */
+ wm = intel_calculate_wm(clock, &igd_display_hplloff_wm,
+ latency->display_hpll_disable, I915_FIFO_LINE_SIZE);
+ reg = I915_READ(DSPFW3);
+ reg &= 0xfffffe00;
+ reg |= wm & 0x1ff;
+ I915_WRITE(DSPFW3, reg);
+
+ /* cursor HPLL off SR */
+ wm = intel_calculate_wm(clock, &igd_cursor_hplloff_wm, pixel_size,
+ latency->cursor_hpll_disable);
+ reg = I915_READ(DSPFW3);
+ reg &= ~(0x3f << 16);
+ reg |= (wm & 0x3f) << 16;
+ I915_WRITE(DSPFW3, reg);
+ DRM_DEBUG("DSPFW3 register is %x\n", reg);
+
+ /* activate cxsr */
+ reg = I915_READ(DSPFW3);
+ reg |= IGD_SELF_REFRESH_EN;
+ I915_WRITE(DSPFW3, reg);
+
+ DRM_INFO("Big FIFO is enabled\n");
+
+ return;
+}
+
+/*
+ * Latency for FIFO fetches is dependent on several factors:
+ * - memory configuration (speed, channels)
+ * - chipset
+ * - current MCH state
+ * It can be fairly high in some situations, so here we assume a fairly
+ * pessimal value. It's a tradeoff between extra memory fetches (if we
+ * set this value too high, the FIFO will fetch frequently to stay full)
+ * and power consumption (set it too low to save power and we might see
+ * FIFO underruns and display "flicker").
+ *
+ * A value of 5us seems to be a good balance; safe for very low end
+ * platforms but not overly aggressive on lower latency configs.
+ */
+const static int latency_ns = 5000;
+
+static int intel_get_fifo_size(struct drm_device *dev, int plane)
+{
+ struct drm_i915_private *dev_priv = dev->dev_private;
+ uint32_t dsparb = I915_READ(DSPARB);
+ int size;
+
+ if (IS_I9XX(dev)) {
+ if (plane == 0)
+ size = dsparb & 0x7f;
+ else
+ size = ((dsparb >> DSPARB_CSTART_SHIFT) & 0x7f) -
+ (dsparb & 0x7f);
+ } else if (IS_I85X(dev)) {
+ if (plane == 0)
+ size = dsparb & 0x1ff;
+ else
+ size = ((dsparb >> DSPARB_BEND_SHIFT) & 0x1ff) -
+ (dsparb & 0x1ff);
+ size >>= 1; /* Convert to cachelines */
+ } else if (IS_845G(dev)) {
+ size = dsparb & 0x7f;
+ size >>= 2; /* Convert to cachelines */
+ } else {
+ size = dsparb & 0x7f;
+ size >>= 1; /* Convert to cachelines */
+ }
+
+ DRM_DEBUG("FIFO size - (0x%08x) %s: %d\n", dsparb, plane ? "B" : "A",
+ size);
+
+ return size;
+}
+
+static void g4x_update_wm(struct drm_device *dev)
+{
+ struct drm_i915_private *dev_priv = dev->dev_private;
+ u32 fw_blc_self = I915_READ(FW_BLC_SELF);
+
+ if (i915_powersave)
+ fw_blc_self |= FW_BLC_SELF_EN;
+ else
+ fw_blc_self &= ~FW_BLC_SELF_EN;
+ I915_WRITE(FW_BLC_SELF, fw_blc_self);
+}
+
+static void i965_update_wm(struct drm_device *dev)
+{
+ struct drm_i915_private *dev_priv = dev->dev_private;
+
+ DRM_DEBUG("Setting FIFO watermarks - A: 8, B: 8, C: 8, SR 8\n");
+
+ /* 965 has limitations... */
+ I915_WRITE(DSPFW1, (8 << 16) | (8 << 8) | (8 << 0));
+ I915_WRITE(DSPFW2, (8 << 8) | (8 << 0));
+}
+
+static void i9xx_update_wm(struct drm_device *dev, int planea_clock,
+ int planeb_clock, int sr_hdisplay, int pixel_size)
+{
+ struct drm_i915_private *dev_priv = dev->dev_private;
+ uint32_t fwater_lo;
+ uint32_t fwater_hi;
+ int total_size, cacheline_size, cwm, srwm = 1;
+ int planea_wm, planeb_wm;
+ struct intel_watermark_params planea_params, planeb_params;
+ unsigned long line_time_us;
+ int sr_clock, sr_entries = 0;
+
+ /* Create copies of the base settings for each pipe */
+ if (IS_I965GM(dev) || IS_I945GM(dev))
+ planea_params = planeb_params = i945_wm_info;
+ else if (IS_I9XX(dev))
+ planea_params = planeb_params = i915_wm_info;
+ else
+ planea_params = planeb_params = i855_wm_info;
+
+ /* Grab a couple of global values before we overwrite them */
+ total_size = planea_params.fifo_size;
+ cacheline_size = planea_params.cacheline_size;
+
+ /* Update per-plane FIFO sizes */
+ planea_params.fifo_size = intel_get_fifo_size(dev, 0);
+ planeb_params.fifo_size = intel_get_fifo_size(dev, 1);
+
+ planea_wm = intel_calculate_wm(planea_clock, &planea_params,
+ pixel_size, latency_ns);
+ planeb_wm = intel_calculate_wm(planeb_clock, &planeb_params,
+ pixel_size, latency_ns);
+ DRM_DEBUG("FIFO watermarks - A: %d, B: %d\n", planea_wm, planeb_wm);
+
+ /*
+ * Overlay gets an aggressive default since video jitter is bad.
+ */
+ cwm = 2;
+
+ /* Calc sr entries for one plane configs */
+ if (HAS_FW_BLC(dev) && sr_hdisplay &&
+ (!planea_clock || !planeb_clock)) {
+ /* self-refresh has much higher latency */
+ const static int sr_latency_ns = 6000;
+
+ sr_clock = planea_clock ? planea_clock : planeb_clock;
+ line_time_us = ((sr_hdisplay * 1000) / sr_clock);
+
+ /* Use ns/us then divide to preserve precision */
+ sr_entries = (((sr_latency_ns / line_time_us) + 1) *
+ pixel_size * sr_hdisplay) / 1000;
+ sr_entries = roundup(sr_entries / cacheline_size, 1);
+ DRM_DEBUG("self-refresh entries: %d\n", sr_entries);
+ srwm = total_size - sr_entries;
+ if (srwm < 0)
+ srwm = 1;
+ I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN | (srwm & 0x3f));
+ }
+
+ DRM_DEBUG("Setting FIFO watermarks - A: %d, B: %d, C: %d, SR %d\n",
+ planea_wm, planeb_wm, cwm, srwm);
+
+ fwater_lo = ((planeb_wm & 0x3f) << 16) | (planea_wm & 0x3f);
+ fwater_hi = (cwm & 0x1f);
+
+ /* Set request length to 8 cachelines per fetch */
+ fwater_lo = fwater_lo | (1 << 24) | (1 << 8);
+ fwater_hi = fwater_hi | (1 << 8);
+
+ I915_WRITE(FW_BLC, fwater_lo);
+ I915_WRITE(FW_BLC2, fwater_hi);
+}
+
+static void i830_update_wm(struct drm_device *dev, int planea_clock,
+ int pixel_size)
+{
+ struct drm_i915_private *dev_priv = dev->dev_private;
+ uint32_t fwater_lo = I915_READ(FW_BLC) & ~0xfff;
+ int planea_wm;
+
+ i830_wm_info.fifo_size = intel_get_fifo_size(dev, 0);
+
+ planea_wm = intel_calculate_wm(planea_clock, &i830_wm_info,
+ pixel_size, latency_ns);
+ fwater_lo |= (3<<8) | planea_wm;
+
+ DRM_DEBUG("Setting FIFO watermarks - A: %d\n", planea_wm);
+
+ I915_WRITE(FW_BLC, fwater_lo);
+}
+
+/**
+ * intel_update_watermarks - update FIFO watermark values based on current modes
+ *
+ * Calculate watermark values for the various WM regs based on current mode
+ * and plane configuration.
+ *
+ * There are several cases to deal with here:
+ * - normal (i.e. non-self-refresh)
+ * - self-refresh (SR) mode
+ * - lines are large relative to FIFO size (buffer can hold up to 2)
+ * - lines are small relative to FIFO size (buffer can hold more than 2
+ * lines), so need to account for TLB latency
+ *
+ * The normal calculation is:
+ * watermark = dotclock * bytes per pixel * latency
+ * where latency is platform & configuration dependent (we assume pessimal
+ * values here).
+ *
+ * The SR calculation is:
+ * watermark = (trunc(latency/line time)+1) * surface width *
+ * bytes per pixel
+ * where
+ * line time = htotal / dotclock
+ * and latency is assumed to be high, as above.
+ *
+ * The final value programmed to the register should always be rounded up,
+ * and include an extra 2 entries to account for clock crossings.
+ *
+ * We don't use the sprite, so we can ignore that. And on Crestline we have
+ * to set the non-SR watermarks to 8.
+ */
+static void intel_update_watermarks(struct drm_device *dev)
+{
+ struct drm_crtc *crtc;
+ struct intel_crtc *intel_crtc;
+ int sr_hdisplay = 0;
+ unsigned long planea_clock = 0, planeb_clock = 0, sr_clock = 0;
+ int enabled = 0, pixel_size = 0;
+
+ /* Get the clock config from both planes */
+ list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+ intel_crtc = to_intel_crtc(crtc);
+ if (crtc->enabled) {
+ enabled++;
+ if (intel_crtc->plane == 0) {
+ DRM_DEBUG("plane A (pipe %d) clock: %d\n",
+ intel_crtc->pipe, crtc->mode.clock);
+ planea_clock = crtc->mode.clock;
+ } else {
+ DRM_DEBUG("plane B (pipe %d) clock: %d\n",
+ intel_crtc->pipe, crtc->mode.clock);
+ planeb_clock = crtc->mode.clock;
+ }
+ sr_hdisplay = crtc->mode.hdisplay;
+ sr_clock = crtc->mode.clock;
+ if (crtc->fb)
+ pixel_size = crtc->fb->bits_per_pixel / 8;
+ else
+ pixel_size = 4; /* by default */
+ }
+ }
+
+ if (enabled <= 0)
+ return;
+
+ /* Single plane configs can enable self refresh */
+ if (enabled == 1 && IS_IGD(dev))
+ igd_enable_cxsr(dev, sr_clock, pixel_size);
+ else if (IS_IGD(dev))
+ igd_disable_cxsr(dev);
+
+ if (IS_G4X(dev))
+ g4x_update_wm(dev);
+ else if (IS_I965G(dev))
+ i965_update_wm(dev);
+ else if (IS_I9XX(dev) || IS_MOBILE(dev))
+ i9xx_update_wm(dev, planea_clock, planeb_clock, sr_hdisplay,
+ pixel_size);
+ else
+ i830_update_wm(dev, planea_clock, pixel_size);
+}
+
+static int intel_crtc_mode_set(struct drm_crtc *crtc,
+ struct drm_display_mode *mode,
+ struct drm_display_mode *adjusted_mode,
+ int x, int y,
+ struct drm_framebuffer *old_fb)