/*
 * Copyright © 2012 Intel Corporation
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice (including the next
 * paragraph) shall be included in all copies or substantial portions of the
 * Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 *
 * Authors:
 *    Eugeni Dodonov <eugeni.dodonov@intel.com>
 *
 */

#include <linux/cpufreq.h>
#include "i915_drv.h"
#include "intel_drv.h"
#include "../../../platform/x86/intel_ips.h"
#include <linux/module.h>
#include <linux/vgaarb.h>
#include <drm/i915_powerwell.h>
#include <linux/pm_runtime.h>

/**
 * RC6 is a special power stage which allows the GPU to enter an very
 * low-voltage mode when idle, using down to 0V while at this stage.  This
 * stage is entered automatically when the GPU is idle when RC6 support is
 * enabled, and as soon as new workload arises GPU wakes up automatically as well.
 *
 * There are different RC6 modes available in Intel GPU, which differentiate
 * among each other with the latency required to enter and leave RC6 and
 * voltage consumed by the GPU in different states.
 *
 * The combination of the following flags define which states GPU is allowed
 * to enter, while RC6 is the normal RC6 state, RC6p is the deep RC6, and
 * RC6pp is deepest RC6. Their support by hardware varies according to the
 * GPU, BIOS, chipset and platform. RC6 is usually the safest one and the one
 * which brings the most power savings; deeper states save more power, but
 * require higher latency to switch to and wake up.
 */
#define INTEL_RC6_ENABLE                        (1<<0)
#define INTEL_RC6p_ENABLE                       (1<<1)
#define INTEL_RC6pp_ENABLE                      (1<<2)

/* FBC, or Frame Buffer Compression, is a technique employed to compress the
 * framebuffer contents in-memory, aiming at reducing the required bandwidth
 * during in-memory transfers and, therefore, reduce the power packet.
 *
 * The benefits of FBC are mostly visible with solid backgrounds and
 * variation-less patterns.
 *
 * FBC-related functionality can be enabled by the means of the
 * i915.i915_enable_fbc parameter
 */

static void i8xx_disable_fbc(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u32 fbc_ctl;

        /* Disable compression */
        fbc_ctl = I915_READ(FBC_CONTROL);
        if ((fbc_ctl & FBC_CTL_EN) == 0)
                return;

        fbc_ctl &= ~FBC_CTL_EN;
        I915_WRITE(FBC_CONTROL, fbc_ctl);

        /* Wait for compressing bit to clear */
        if (wait_for((I915_READ(FBC_STATUS) & FBC_STAT_COMPRESSING) == 0, 10)) {
                DRM_DEBUG_KMS("FBC idle timed out\n");
                return;
        }

        DRM_DEBUG_KMS("disabled FBC\n");
}

static void i8xx_enable_fbc(struct drm_crtc *crtc)
{
        struct drm_device *dev = crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_framebuffer *fb = crtc->primary->fb;
        struct drm_i915_gem_object *obj = intel_fb_obj(fb);
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        int cfb_pitch;
        int i;
        u32 fbc_ctl;

        cfb_pitch = dev_priv->fbc.size / FBC_LL_SIZE;
        if (fb->pitches[0] < cfb_pitch)
                cfb_pitch = fb->pitches[0];

        /* FBC_CTL wants 32B or 64B units */
        if (IS_GEN2(dev))
                cfb_pitch = (cfb_pitch / 32) - 1;
        else
                cfb_pitch = (cfb_pitch / 64) - 1;

        /* Clear old tags */
        for (i = 0; i < (FBC_LL_SIZE / 32) + 1; i++)
                I915_WRITE(FBC_TAG + (i * 4), 0);

        if (IS_GEN4(dev)) {
                u32 fbc_ctl2;

                /* Set it up... */
                fbc_ctl2 = FBC_CTL_FENCE_DBL | FBC_CTL_IDLE_IMM | FBC_CTL_CPU_FENCE;
                fbc_ctl2 |= FBC_CTL_PLANE(intel_crtc->plane);
                I915_WRITE(FBC_CONTROL2, fbc_ctl2);
                I915_WRITE(FBC_FENCE_OFF, crtc->y);
        }

        /* enable it... */
        fbc_ctl = I915_READ(FBC_CONTROL);
        fbc_ctl &= 0x3fff << FBC_CTL_INTERVAL_SHIFT;
        fbc_ctl |= FBC_CTL_EN | FBC_CTL_PERIODIC;
        if (IS_I945GM(dev))
                fbc_ctl |= FBC_CTL_C3_IDLE; /* 945 needs special SR handling */
        fbc_ctl |= (cfb_pitch & 0xff) << FBC_CTL_STRIDE_SHIFT;
        fbc_ctl |= obj->fence_reg;
        I915_WRITE(FBC_CONTROL, fbc_ctl);

        DRM_DEBUG_KMS("enabled FBC, pitch %d, yoff %d, plane %c\n",
                      cfb_pitch, crtc->y, plane_name(intel_crtc->plane));
}

static bool i8xx_fbc_enabled(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        return I915_READ(FBC_CONTROL) & FBC_CTL_EN;
}

static void g4x_enable_fbc(struct drm_crtc *crtc)
{
        struct drm_device *dev = crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_framebuffer *fb = crtc->primary->fb;
        struct drm_i915_gem_object *obj = intel_fb_obj(fb);
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        u32 dpfc_ctl;

        dpfc_ctl = DPFC_CTL_PLANE(intel_crtc->plane) | DPFC_SR_EN;
        if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
                dpfc_ctl |= DPFC_CTL_LIMIT_2X;
        else
                dpfc_ctl |= DPFC_CTL_LIMIT_1X;
        dpfc_ctl |= DPFC_CTL_FENCE_EN | obj->fence_reg;

        I915_WRITE(DPFC_FENCE_YOFF, crtc->y);

        /* enable it... */
        I915_WRITE(DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN);

        DRM_DEBUG_KMS("enabled fbc on plane %c\n", plane_name(intel_crtc->plane));
}

static void g4x_disable_fbc(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u32 dpfc_ctl;

        /* Disable compression */
        dpfc_ctl = I915_READ(DPFC_CONTROL);
        if (dpfc_ctl & DPFC_CTL_EN) {
                dpfc_ctl &= ~DPFC_CTL_EN;
                I915_WRITE(DPFC_CONTROL, dpfc_ctl);

                DRM_DEBUG_KMS("disabled FBC\n");
        }
}

static bool g4x_fbc_enabled(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        return I915_READ(DPFC_CONTROL) & DPFC_CTL_EN;
}

static void sandybridge_blit_fbc_update(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u32 blt_ecoskpd;

        /* Make sure blitter notifies FBC of writes */

        /* Blitter is part of Media powerwell on VLV. No impact of
         * his param in other platforms for now */
        gen6_gt_force_wake_get(dev_priv, FORCEWAKE_MEDIA);

        blt_ecoskpd = I915_READ(GEN6_BLITTER_ECOSKPD);
        blt_ecoskpd |= GEN6_BLITTER_FBC_NOTIFY <<
                GEN6_BLITTER_LOCK_SHIFT;
        I915_WRITE(GEN6_BLITTER_ECOSKPD, blt_ecoskpd);
        blt_ecoskpd |= GEN6_BLITTER_FBC_NOTIFY;
        I915_WRITE(GEN6_BLITTER_ECOSKPD, blt_ecoskpd);
        blt_ecoskpd &= ~(GEN6_BLITTER_FBC_NOTIFY <<
                         GEN6_BLITTER_LOCK_SHIFT);
        I915_WRITE(GEN6_BLITTER_ECOSKPD, blt_ecoskpd);
        POSTING_READ(GEN6_BLITTER_ECOSKPD);

        gen6_gt_force_wake_put(dev_priv, FORCEWAKE_MEDIA);
}

static void ironlake_enable_fbc(struct drm_crtc *crtc)
{
        struct drm_device *dev = crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_framebuffer *fb = crtc->primary->fb;
        struct drm_i915_gem_object *obj = intel_fb_obj(fb);
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        u32 dpfc_ctl;

        dpfc_ctl = DPFC_CTL_PLANE(intel_crtc->plane);
        if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
                dev_priv->fbc.threshold++;

        switch (dev_priv->fbc.threshold) {
        case 4:
        case 3:
                dpfc_ctl |= DPFC_CTL_LIMIT_4X;
                break;
        case 2:
                dpfc_ctl |= DPFC_CTL_LIMIT_2X;
                break;
        case 1:
                dpfc_ctl |= DPFC_CTL_LIMIT_1X;
                break;
        }
        dpfc_ctl |= DPFC_CTL_FENCE_EN;
        if (IS_GEN5(dev))
                dpfc_ctl |= obj->fence_reg;

        I915_WRITE(ILK_DPFC_FENCE_YOFF, crtc->y);
        I915_WRITE(ILK_FBC_RT_BASE, i915_gem_obj_ggtt_offset(obj) | ILK_FBC_RT_VALID);
        /* enable it... */
        I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN);

        if (IS_GEN6(dev)) {
                I915_WRITE(SNB_DPFC_CTL_SA,
                           SNB_CPU_FENCE_ENABLE | obj->fence_reg);
                I915_WRITE(DPFC_CPU_FENCE_OFFSET, crtc->y);
                sandybridge_blit_fbc_update(dev);
        }

        DRM_DEBUG_KMS("enabled fbc on plane %c\n", plane_name(intel_crtc->plane));
}

static void ironlake_disable_fbc(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u32 dpfc_ctl;

        /* Disable compression */
        dpfc_ctl = I915_READ(ILK_DPFC_CONTROL);
        if (dpfc_ctl & DPFC_CTL_EN) {
                dpfc_ctl &= ~DPFC_CTL_EN;
                I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl);

                DRM_DEBUG_KMS("disabled FBC\n");
        }
}

static bool ironlake_fbc_enabled(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        return I915_READ(ILK_DPFC_CONTROL) & DPFC_CTL_EN;
}

static void gen7_enable_fbc(struct drm_crtc *crtc)
{
        struct drm_device *dev = crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_framebuffer *fb = crtc->primary->fb;
        struct drm_i915_gem_object *obj = intel_fb_obj(fb);
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        u32 dpfc_ctl;

        dpfc_ctl = IVB_DPFC_CTL_PLANE(intel_crtc->plane);
        if (drm_format_plane_cpp(fb->pixel_format, 0) == 2)
                dev_priv->fbc.threshold++;

        switch (dev_priv->fbc.threshold) {
        case 4:
        case 3:
                dpfc_ctl |= DPFC_CTL_LIMIT_4X;
                break;
        case 2:
                dpfc_ctl |= DPFC_CTL_LIMIT_2X;
                break;
        case 1:
                dpfc_ctl |= DPFC_CTL_LIMIT_1X;
                break;
        }

        dpfc_ctl |= IVB_DPFC_CTL_FENCE_EN;

        I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN);

        if (IS_IVYBRIDGE(dev)) {
                /* WaFbcAsynchFlipDisableFbcQueue:ivb */
                I915_WRITE(ILK_DISPLAY_CHICKEN1,
                           I915_READ(ILK_DISPLAY_CHICKEN1) |
                           ILK_FBCQ_DIS);
        } else {
                /* WaFbcAsynchFlipDisableFbcQueue:hsw,bdw */
                I915_WRITE(CHICKEN_PIPESL_1(intel_crtc->pipe),
                           I915_READ(CHICKEN_PIPESL_1(intel_crtc->pipe)) |
                           HSW_FBCQ_DIS);
        }

        I915_WRITE(SNB_DPFC_CTL_SA,
                   SNB_CPU_FENCE_ENABLE | obj->fence_reg);
        I915_WRITE(DPFC_CPU_FENCE_OFFSET, crtc->y);

        sandybridge_blit_fbc_update(dev);

        DRM_DEBUG_KMS("enabled fbc on plane %c\n", plane_name(intel_crtc->plane));
}

bool intel_fbc_enabled(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        if (!dev_priv->display.fbc_enabled)
                return false;

        return dev_priv->display.fbc_enabled(dev);
}

static void intel_fbc_work_fn(struct work_struct *__work)
{
        struct intel_fbc_work *work =
                container_of(to_delayed_work(__work),
                             struct intel_fbc_work, work);
        struct drm_device *dev = work->crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;

        mutex_lock(&dev->struct_mutex);
        if (work == dev_priv->fbc.fbc_work) {
                /* Double check that we haven't switched fb without cancelling
                 * the prior work.
                 */
                if (work->crtc->primary->fb == work->fb) {
                        dev_priv->display.enable_fbc(work->crtc);

                        dev_priv->fbc.plane = to_intel_crtc(work->crtc)->plane;
                        dev_priv->fbc.fb_id = work->crtc->primary->fb->base.id;
                        dev_priv->fbc.y = work->crtc->y;
                }

                dev_priv->fbc.fbc_work = NULL;
        }
        mutex_unlock(&dev->struct_mutex);

        kfree(work);
}

static void intel_cancel_fbc_work(struct drm_i915_private *dev_priv)
{
        if (dev_priv->fbc.fbc_work == NULL)
                return;

        DRM_DEBUG_KMS("cancelling pending FBC enable\n");

        /* Synchronisation is provided by struct_mutex and checking of
         * dev_priv->fbc.fbc_work, so we can perform the cancellation
         * entirely asynchronously.
         */
        if (cancel_delayed_work(&dev_priv->fbc.fbc_work->work))
                /* tasklet was killed before being run, clean up */
                kfree(dev_priv->fbc.fbc_work);

        /* Mark the work as no longer wanted so that if it does
         * wake-up (because the work was already running and waiting
         * for our mutex), it will discover that is no longer
         * necessary to run.
         */
        dev_priv->fbc.fbc_work = NULL;
}

static void intel_enable_fbc(struct drm_crtc *crtc)
{
        struct intel_fbc_work *work;
        struct drm_device *dev = crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;

        if (!dev_priv->display.enable_fbc)
                return;

        intel_cancel_fbc_work(dev_priv);

        work = kzalloc(sizeof(*work), GFP_KERNEL);
        if (work == NULL) {
                DRM_ERROR("Failed to allocate FBC work structure\n");
                dev_priv->display.enable_fbc(crtc);
                return;
        }

        work->crtc = crtc;
        work->fb = crtc->primary->fb;
        INIT_DELAYED_WORK(&work->work, intel_fbc_work_fn);

        dev_priv->fbc.fbc_work = work;

        /* Delay the actual enabling to let pageflipping cease and the
         * display to settle before starting the compression. Note that
         * this delay also serves a second purpose: it allows for a
         * vblank to pass after disabling the FBC before we attempt
         * to modify the control registers.
         *
         * A more complicated solution would involve tracking vblanks
         * following the termination of the page-flipping sequence
         * and indeed performing the enable as a co-routine and not
         * waiting synchronously upon the vblank.
         *
         * WaFbcWaitForVBlankBeforeEnable:ilk,snb
         */
        schedule_delayed_work(&work->work, msecs_to_jiffies(50));
}

void intel_disable_fbc(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        intel_cancel_fbc_work(dev_priv);

        if (!dev_priv->display.disable_fbc)
                return;

        dev_priv->display.disable_fbc(dev);
        dev_priv->fbc.plane = -1;
}

static bool set_no_fbc_reason(struct drm_i915_private *dev_priv,
                              enum no_fbc_reason reason)
{
        if (dev_priv->fbc.no_fbc_reason == reason)
                return false;

        dev_priv->fbc.no_fbc_reason = reason;
        return true;
}

/**
 * intel_update_fbc - enable/disable FBC as needed
 * @dev: the drm_device
 *
 * Set up the framebuffer compression hardware at mode set time.  We
 * enable it if possible:
 *   - plane A only (on pre-965)
 *   - no pixel mulitply/line duplication
 *   - no alpha buffer discard
 *   - no dual wide
 *   - framebuffer <= max_hdisplay in width, max_vdisplay in height
 *
 * We can't assume that any compression will take place (worst case),
 * so the compressed buffer has to be the same size as the uncompressed
 * one.  It also must reside (along with the line length buffer) in
 * stolen memory.
 *
 * We need to enable/disable FBC on a global basis.
 */
void intel_update_fbc(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_crtc *crtc = NULL, *tmp_crtc;
        struct intel_crtc *intel_crtc;
        struct drm_framebuffer *fb;
        struct drm_i915_gem_object *obj;
        const struct drm_display_mode *adjusted_mode;
        unsigned int max_width, max_height;

        if (!HAS_FBC(dev)) {
                set_no_fbc_reason(dev_priv, FBC_UNSUPPORTED);
                return;
        }

        if (!i915.powersave) {
                if (set_no_fbc_reason(dev_priv, FBC_MODULE_PARAM))
                        DRM_DEBUG_KMS("fbc disabled per module param\n");
                return;
        }

        /*
         * If FBC is already on, we just have to verify that we can
         * keep it that way...
         * Need to disable if:
         *   - more than one pipe is active
         *   - changing FBC params (stride, fence, mode)
         *   - new fb is too large to fit in compressed buffer
         *   - going to an unsupported config (interlace, pixel multiply, etc.)
         */
        for_each_crtc(dev, tmp_crtc) {
                if (intel_crtc_active(tmp_crtc) &&
                    to_intel_crtc(tmp_crtc)->primary_enabled) {
                        if (crtc) {
                                if (set_no_fbc_reason(dev_priv, FBC_MULTIPLE_PIPES))
                                        DRM_DEBUG_KMS("more than one pipe active, disabling compression\n");
                                goto out_disable;
                        }
                        crtc = tmp_crtc;
                }
        }

        if (!crtc || crtc->primary->fb == NULL) {
                if (set_no_fbc_reason(dev_priv, FBC_NO_OUTPUT))
                        DRM_DEBUG_KMS("no output, disabling\n");
                goto out_disable;
        }

        intel_crtc = to_intel_crtc(crtc);
        fb = crtc->primary->fb;
        obj = intel_fb_obj(fb);
        adjusted_mode = &intel_crtc->config.adjusted_mode;

        if (i915.enable_fbc < 0) {
                if (set_no_fbc_reason(dev_priv, FBC_CHIP_DEFAULT))
                        DRM_DEBUG_KMS("disabled per chip default\n");
                goto out_disable;
        }
        if (!i915.enable_fbc) {
                if (set_no_fbc_reason(dev_priv, FBC_MODULE_PARAM))
                        DRM_DEBUG_KMS("fbc disabled per module param\n");
                goto out_disable;
        }
        if ((adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) ||
            (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)) {
                if (set_no_fbc_reason(dev_priv, FBC_UNSUPPORTED_MODE))
                        DRM_DEBUG_KMS("mode incompatible with compression, "
                                      "disabling\n");
                goto out_disable;
        }

        if (INTEL_INFO(dev)->gen >= 8 || IS_HASWELL(dev)) {
                max_width = 4096;
                max_height = 4096;
        } else if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
                max_width = 4096;
                max_height = 2048;
        } else {
                max_width = 2048;
                max_height = 1536;
        }
        if (intel_crtc->config.pipe_src_w > max_width ||
            intel_crtc->config.pipe_src_h > max_height) {
                if (set_no_fbc_reason(dev_priv, FBC_MODE_TOO_LARGE))
                        DRM_DEBUG_KMS("mode too large for compression, disabling\n");
                goto out_disable;
        }
        if ((INTEL_INFO(dev)->gen < 4 || HAS_DDI(dev)) &&
            intel_crtc->plane != PLANE_A) {
                if (set_no_fbc_reason(dev_priv, FBC_BAD_PLANE))
                        DRM_DEBUG_KMS("plane not A, disabling compression\n");
                goto out_disable;
        }

        /* The use of a CPU fence is mandatory in order to detect writes
         * by the CPU to the scanout and trigger updates to the FBC.
         */
        if (obj->tiling_mode != I915_TILING_X ||
            obj->fence_reg == I915_FENCE_REG_NONE) {
                if (set_no_fbc_reason(dev_priv, FBC_NOT_TILED))
                        DRM_DEBUG_KMS("framebuffer not tiled or fenced, disabling compression\n");
                goto out_disable;
        }

        /* If the kernel debugger is active, always disable compression */
        if (in_dbg_master())
                goto out_disable;

        if (i915_gem_stolen_setup_compression(dev, obj->base.size,
                                              drm_format_plane_cpp(fb->pixel_format, 0))) {
                if (set_no_fbc_reason(dev_priv, FBC_STOLEN_TOO_SMALL))
                        DRM_DEBUG_KMS("framebuffer too large, disabling compression\n");
                goto out_disable;
        }

        /* If the scanout has not changed, don't modify the FBC settings.
         * Note that we make the fundamental assumption that the fb->obj
         * cannot be unpinned (and have its GTT offset and fence revoked)
         * without first being decoupled from the scanout and FBC disabled.
         */
        if (dev_priv->fbc.plane == intel_crtc->plane &&
            dev_priv->fbc.fb_id == fb->base.id &&
            dev_priv->fbc.y == crtc->y)
                return;

        if (intel_fbc_enabled(dev)) {
                /* We update FBC along two paths, after changing fb/crtc
                 * configuration (modeswitching) and after page-flipping
                 * finishes. For the latter, we know that not only did
                 * we disable the FBC at the start of the page-flip
                 * sequence, but also more than one vblank has passed.
                 *
                 * For the former case of modeswitching, it is possible
                 * to switch between two FBC valid configurations
                 * instantaneously so we do need to disable the FBC
                 * before we can modify its control registers. We also
                 * have to wait for the next vblank for that to take
                 * effect. However, since we delay enabling FBC we can
                 * assume that a vblank has passed since disabling and
                 * that we can safely alter the registers in the deferred
                 * callback.
                 *
                 * In the scenario that we go from a valid to invalid
                 * and then back to valid FBC configuration we have
                 * no strict enforcement that a vblank occurred since
                 * disabling the FBC. However, along all current pipe
                 * disabling paths we do need to wait for a vblank at
                 * some point. And we wait before enabling FBC anyway.
                 */
                DRM_DEBUG_KMS("disabling active FBC for update\n");
                intel_disable_fbc(dev);
        }

        intel_enable_fbc(crtc);
        dev_priv->fbc.no_fbc_reason = FBC_OK;
        return;

out_disable:
        /* Multiple disables should be harmless */
        if (intel_fbc_enabled(dev)) {
                DRM_DEBUG_KMS("unsupported config, disabling FBC\n");
                intel_disable_fbc(dev);
        }
        i915_gem_stolen_cleanup_compression(dev);
}

static void i915_pineview_get_mem_freq(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u32 tmp;

        tmp = I915_READ(CLKCFG);

        switch (tmp & CLKCFG_FSB_MASK) {
        case CLKCFG_FSB_533:
                dev_priv->fsb_freq = 533; /* 133*4 */
                break;
        case CLKCFG_FSB_800:
                dev_priv->fsb_freq = 800; /* 200*4 */
                break;
        case CLKCFG_FSB_667:
                dev_priv->fsb_freq =  667; /* 167*4 */
                break;
        case CLKCFG_FSB_400:
                dev_priv->fsb_freq = 400; /* 100*4 */
                break;
        }

        switch (tmp & CLKCFG_MEM_MASK) {
        case CLKCFG_MEM_533:
                dev_priv->mem_freq = 533;
                break;
        case CLKCFG_MEM_667:
                dev_priv->mem_freq = 667;
                break;
        case CLKCFG_MEM_800:
                dev_priv->mem_freq = 800;
                break;
        }

        /* detect pineview DDR3 setting */
        tmp = I915_READ(CSHRDDR3CTL);
        dev_priv->is_ddr3 = (tmp & CSHRDDR3CTL_DDR3) ? 1 : 0;
}

static void i915_ironlake_get_mem_freq(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u16 ddrpll, csipll;

        ddrpll = I915_READ16(DDRMPLL1);
        csipll = I915_READ16(CSIPLL0);

        switch (ddrpll & 0xff) {
        case 0xc:
                dev_priv->mem_freq = 800;
                break;
        case 0x10:
                dev_priv->mem_freq = 1066;
                break;
        case 0x14:
                dev_priv->mem_freq = 1333;
                break;
        case 0x18:
                dev_priv->mem_freq = 1600;
                break;
        default:
                DRM_DEBUG_DRIVER("unknown memory frequency 0x%02x\n",
                                 ddrpll & 0xff);
                dev_priv->mem_freq = 0;
                break;
        }

        dev_priv->ips.r_t = dev_priv->mem_freq;

        switch (csipll & 0x3ff) {
        case 0x00c:
                dev_priv->fsb_freq = 3200;
                break;
        case 0x00e:
                dev_priv->fsb_freq = 3733;
                break;
        case 0x010:
                dev_priv->fsb_freq = 4266;
                break;
        case 0x012:
                dev_priv->fsb_freq = 4800;
                break;
        case 0x014:
                dev_priv->fsb_freq = 5333;
                break;
        case 0x016:
                dev_priv->fsb_freq = 5866;
                break;
        case 0x018:
                dev_priv->fsb_freq = 6400;
                break;
        default:
                DRM_DEBUG_DRIVER("unknown fsb frequency 0x%04x\n",
                                 csipll & 0x3ff);
                dev_priv->fsb_freq = 0;
                break;
        }

        if (dev_priv->fsb_freq == 3200) {
                dev_priv->ips.c_m = 0;
        } else if (dev_priv->fsb_freq > 3200 && dev_priv->fsb_freq <= 4800) {
                dev_priv->ips.c_m = 1;
        } else {
                dev_priv->ips.c_m = 2;
        }
}

static const struct cxsr_latency cxsr_latency_table[] = {
        {1, 0, 800, 400, 3382, 33382, 3983, 33983},    /* DDR2-400 SC */
        {1, 0, 800, 667, 3354, 33354, 3807, 33807},    /* DDR2-667 SC */
        {1, 0, 800, 800, 3347, 33347, 3763, 33763},    /* DDR2-800 SC */
        {1, 1, 800, 667, 6420, 36420, 6873, 36873},    /* DDR3-667 SC */
        {1, 1, 800, 800, 5902, 35902, 6318, 36318},    /* DDR3-800 SC */

        {1, 0, 667, 400, 3400, 33400, 4021, 34021},    /* DDR2-400 SC */
        {1, 0, 667, 667, 3372, 33372, 3845, 33845},    /* DDR2-667 SC */
        {1, 0, 667, 800, 3386, 33386, 3822, 33822},    /* DDR2-800 SC */
        {1, 1, 667, 667, 6438, 36438, 6911, 36911},    /* DDR3-667 SC */
        {1, 1, 667, 800, 5941, 35941, 6377, 36377},    /* DDR3-800 SC */

        {1, 0, 400, 400, 3472, 33472, 4173, 34173},    /* DDR2-400 SC */
        {1, 0, 400, 667, 3443, 33443, 3996, 33996},    /* DDR2-667 SC */
        {1, 0, 400, 800, 3430, 33430, 3946, 33946},    /* DDR2-800 SC */
        {1, 1, 400, 667, 6509, 36509, 7062, 37062},    /* DDR3-667 SC */
        {1, 1, 400, 800, 5985, 35985, 6501, 36501},    /* DDR3-800 SC */

        {0, 0, 800, 400, 3438, 33438, 4065, 34065},    /* DDR2-400 SC */
        {0, 0, 800, 667, 3410, 33410, 3889, 33889},    /* DDR2-667 SC */
        {0, 0, 800, 800, 3403, 33403, 3845, 33845},    /* DDR2-800 SC */
        {0, 1, 800, 667, 6476, 36476, 6955, 36955},    /* DDR3-667 SC */
        {0, 1, 800, 800, 5958, 35958, 6400, 36400},    /* DDR3-800 SC */

        {0, 0, 667, 400, 3456, 33456, 4103, 34106},    /* DDR2-400 SC */
        {0, 0, 667, 667, 3428, 33428, 3927, 33927},    /* DDR2-667 SC */
        {0, 0, 667, 800, 3443, 33443, 3905, 33905},    /* DDR2-800 SC */
        {0, 1, 667, 667, 6494, 36494, 6993, 36993},    /* DDR3-667 SC */
        {0, 1, 667, 800, 5998, 35998, 6460, 36460},    /* DDR3-800 SC */

        {0, 0, 400, 400, 3528, 33528, 4255, 34255},    /* DDR2-400 SC */
        {0, 0, 400, 667, 3500, 33500, 4079, 34079},    /* DDR2-667 SC */
        {0, 0, 400, 800, 3487, 33487, 4029, 34029},    /* DDR2-800 SC */
        {0, 1, 400, 667, 6566, 36566, 7145, 37145},    /* DDR3-667 SC */
        {0, 1, 400, 800, 6042, 36042, 6584, 36584},    /* DDR3-800 SC */
};

static const struct cxsr_latency *intel_get_cxsr_latency(int is_desktop,
                                                         int is_ddr3,
                                                         int fsb,
                                                         int mem)
{
        const struct cxsr_latency *latency;
        int i;

        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 &&
                    is_ddr3 == latency->is_ddr3 &&
                    fsb == latency->fsb_freq && mem == latency->mem_freq)
                        return latency;
        }

        DRM_DEBUG_KMS("Unknown FSB/MEM found, disable CxSR\n");

        return NULL;
}

void intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable)
{
        struct drm_device *dev = dev_priv->dev;
        u32 val;

        if (IS_VALLEYVIEW(dev)) {
                I915_WRITE(FW_BLC_SELF_VLV, enable ? FW_CSPWRDWNEN : 0);
        } else if (IS_G4X(dev) || IS_CRESTLINE(dev)) {
                I915_WRITE(FW_BLC_SELF, enable ? FW_BLC_SELF_EN : 0);
        } else if (IS_PINEVIEW(dev)) {
                val = I915_READ(DSPFW3) & ~PINEVIEW_SELF_REFRESH_EN;
                val |= enable ? PINEVIEW_SELF_REFRESH_EN : 0;
                I915_WRITE(DSPFW3, val);
        } else if (IS_I945G(dev) || IS_I945GM(dev)) {
                val = enable ? _MASKED_BIT_ENABLE(FW_BLC_SELF_EN) :
                               _MASKED_BIT_DISABLE(FW_BLC_SELF_EN);
                I915_WRITE(FW_BLC_SELF, val);
        } else if (IS_I915GM(dev)) {
                val = enable ? _MASKED_BIT_ENABLE(INSTPM_SELF_EN) :
                               _MASKED_BIT_DISABLE(INSTPM_SELF_EN);
                I915_WRITE(INSTPM, val);
        } else {
                return;
        }

        DRM_DEBUG_KMS("memory self-refresh is %s\n",
                      enable ? "enabled" : "disabled");
}

/*
 * 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.
 */
static const int latency_ns = 5000;

static int i9xx_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;

        size = dsparb & 0x7f;
        if (plane)
                size = ((dsparb >> DSPARB_CSTART_SHIFT) & 0x7f) - size;

        DRM_DEBUG_KMS("FIFO size - (0x%08x) %s: %d\n", dsparb,
                      plane ? "B" : "A", size);

        return size;
}

static int i830_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;

        size = dsparb & 0x1ff;
        if (plane)
                size = ((dsparb >> DSPARB_BEND_SHIFT) & 0x1ff) - size;
        size >>= 1; /* Convert to cachelines */

        DRM_DEBUG_KMS("FIFO size - (0x%08x) %s: %d\n", dsparb,
                      plane ? "B" : "A", size);

        return size;
}

static int i845_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;

        size = dsparb & 0x7f;
        size >>= 2; /* Convert to cachelines */

        DRM_DEBUG_KMS("FIFO size - (0x%08x) %s: %d\n", dsparb,
                      plane ? "B" : "A",
                      size);

        return size;
}

/* Pineview has different values for various configs */
static const struct intel_watermark_params pineview_display_wm = {
        .fifo_size = PINEVIEW_DISPLAY_FIFO,
        .max_wm = PINEVIEW_MAX_WM,
        .default_wm = PINEVIEW_DFT_WM,
        .guard_size = PINEVIEW_GUARD_WM,
        .cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params pineview_display_hplloff_wm = {
        .fifo_size = PINEVIEW_DISPLAY_FIFO,
        .max_wm = PINEVIEW_MAX_WM,
        .default_wm = PINEVIEW_DFT_HPLLOFF_WM,
        .guard_size = PINEVIEW_GUARD_WM,
        .cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params pineview_cursor_wm = {
        .fifo_size = PINEVIEW_CURSOR_FIFO,
        .max_wm = PINEVIEW_CURSOR_MAX_WM,
        .default_wm = PINEVIEW_CURSOR_DFT_WM,
        .guard_size = PINEVIEW_CURSOR_GUARD_WM,
        .cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params pineview_cursor_hplloff_wm = {
        .fifo_size = PINEVIEW_CURSOR_FIFO,
        .max_wm = PINEVIEW_CURSOR_MAX_WM,
        .default_wm = PINEVIEW_CURSOR_DFT_WM,
        .guard_size = PINEVIEW_CURSOR_GUARD_WM,
        .cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params g4x_wm_info = {
        .fifo_size = G4X_FIFO_SIZE,
        .max_wm = G4X_MAX_WM,
        .default_wm = G4X_MAX_WM,
        .guard_size = 2,
        .cacheline_size = G4X_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params g4x_cursor_wm_info = {
        .fifo_size = I965_CURSOR_FIFO,
        .max_wm = I965_CURSOR_MAX_WM,
        .default_wm = I965_CURSOR_DFT_WM,
        .guard_size = 2,
        .cacheline_size = G4X_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params valleyview_wm_info = {
        .fifo_size = VALLEYVIEW_FIFO_SIZE,
        .max_wm = VALLEYVIEW_MAX_WM,
        .default_wm = VALLEYVIEW_MAX_WM,
        .guard_size = 2,
        .cacheline_size = G4X_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params valleyview_cursor_wm_info = {
        .fifo_size = I965_CURSOR_FIFO,
        .max_wm = VALLEYVIEW_CURSOR_MAX_WM,
        .default_wm = I965_CURSOR_DFT_WM,
        .guard_size = 2,
        .cacheline_size = G4X_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params i965_cursor_wm_info = {
        .fifo_size = I965_CURSOR_FIFO,
        .max_wm = I965_CURSOR_MAX_WM,
        .default_wm = I965_CURSOR_DFT_WM,
        .guard_size = 2,
        .cacheline_size = I915_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params i945_wm_info = {
        .fifo_size = I945_FIFO_SIZE,
        .max_wm = I915_MAX_WM,
        .default_wm = 1,
        .guard_size = 2,
        .cacheline_size = I915_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params i915_wm_info = {
        .fifo_size = I915_FIFO_SIZE,
        .max_wm = I915_MAX_WM,
        .default_wm = 1,
        .guard_size = 2,
        .cacheline_size = I915_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params i830_wm_info = {
        .fifo_size = I855GM_FIFO_SIZE,
        .max_wm = I915_MAX_WM,
        .default_wm = 1,
        .guard_size = 2,
        .cacheline_size = I830_FIFO_LINE_SIZE,
};
static const struct intel_watermark_params i845_wm_info = {
        .fifo_size = I830_FIFO_SIZE,
        .max_wm = I915_MAX_WM,
        .default_wm = 1,
        .guard_size = 2,
        .cacheline_size = 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,
                                        const struct intel_watermark_params *wm,
                                        int fifo_size,
                                        int pixel_size,
                                        unsigned long latency_ns)
{
        long entries_required, wm_size;

        /*
         * Note: we need to make sure we don't overflow for various clock &
         * latency values.
         * clocks go from a few thousand to several hundred thousand.
         * latency is usually a few thousand
         */
        entries_required = ((clock_in_khz / 1000) * pixel_size * latency_ns) /
                1000;
        entries_required = DIV_ROUND_UP(entries_required, wm->cacheline_size);

        DRM_DEBUG_KMS("FIFO entries required for mode: %ld\n", entries_required);

        wm_size = fifo_size - (entries_required + wm->guard_size);

        DRM_DEBUG_KMS("FIFO watermark level: %ld\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;
}

static struct drm_crtc *single_enabled_crtc(struct drm_device *dev)
{
        struct drm_crtc *crtc, *enabled = NULL;

        for_each_crtc(dev, crtc) {
                if (intel_crtc_active(crtc)) {
                        if (enabled)
                                return NULL;
                        enabled = crtc;
                }
        }

        return enabled;
}

static void pineview_update_wm(struct drm_crtc *unused_crtc)
{
        struct drm_device *dev = unused_crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_crtc *crtc;
        const struct cxsr_latency *latency;
        u32 reg;
        unsigned long wm;

        latency = intel_get_cxsr_latency(IS_PINEVIEW_G(dev), dev_priv->is_ddr3,
                                         dev_priv->fsb_freq, dev_priv->mem_freq);
        if (!latency) {
                DRM_DEBUG_KMS("Unknown FSB/MEM found, disable CxSR\n");
                intel_set_memory_cxsr(dev_priv, false);
                return;
        }

        crtc = single_enabled_crtc(dev);
        if (crtc) {
                const struct drm_display_mode *adjusted_mode;
                int pixel_size = crtc->primary->fb->bits_per_pixel / 8;
                int clock;

                adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
                clock = adjusted_mode->crtc_clock;

                /* Display SR */
                wm = intel_calculate_wm(clock, &pineview_display_wm,
                                        pineview_display_wm.fifo_size,
                                        pixel_size, latency->display_sr);
                reg = I915_READ(DSPFW1);
                reg &= ~DSPFW_SR_MASK;
                reg |= wm << DSPFW_SR_SHIFT;
                I915_WRITE(DSPFW1, reg);
                DRM_DEBUG_KMS("DSPFW1 register is %x\n", reg);

                /* cursor SR */
                wm = intel_calculate_wm(clock, &pineview_cursor_wm,
                                        pineview_display_wm.fifo_size,
                                        pixel_size, latency->cursor_sr);
                reg = I915_READ(DSPFW3);
                reg &= ~DSPFW_CURSOR_SR_MASK;
                reg |= (wm & 0x3f) << DSPFW_CURSOR_SR_SHIFT;
                I915_WRITE(DSPFW3, reg);

                /* Display HPLL off SR */
                wm = intel_calculate_wm(clock, &pineview_display_hplloff_wm,
                                        pineview_display_hplloff_wm.fifo_size,
                                        pixel_size, latency->display_hpll_disable);
                reg = I915_READ(DSPFW3);
                reg &= ~DSPFW_HPLL_SR_MASK;
                reg |= wm & DSPFW_HPLL_SR_MASK;
                I915_WRITE(DSPFW3, reg);

                /* cursor HPLL off SR */
                wm = intel_calculate_wm(clock, &pineview_cursor_hplloff_wm,
                                        pineview_display_hplloff_wm.fifo_size,
                                        pixel_size, latency->cursor_hpll_disable);
                reg = I915_READ(DSPFW3);
                reg &= ~DSPFW_HPLL_CURSOR_MASK;
                reg |= (wm & 0x3f) << DSPFW_HPLL_CURSOR_SHIFT;
                I915_WRITE(DSPFW3, reg);
                DRM_DEBUG_KMS("DSPFW3 register is %x\n", reg);

                intel_set_memory_cxsr(dev_priv, true);
        } else {
                intel_set_memory_cxsr(dev_priv, false);
        }
}

static bool g4x_compute_wm0(struct drm_device *dev,
                            int plane,
                            const struct intel_watermark_params *display,
                            int display_latency_ns,
                            const struct intel_watermark_params *cursor,
                            int cursor_latency_ns,
                            int *plane_wm,
                            int *cursor_wm)
{
        struct drm_crtc *crtc;
        const struct drm_display_mode *adjusted_mode;
        int htotal, hdisplay, clock, pixel_size;
        int line_time_us, line_count;
        int entries, tlb_miss;

        crtc = intel_get_crtc_for_plane(dev, plane);
        if (!intel_crtc_active(crtc)) {
                *cursor_wm = cursor->guard_size;
                *plane_wm = display->guard_size;
                return false;
        }

        adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
        clock = adjusted_mode->crtc_clock;
        htotal = adjusted_mode->crtc_htotal;
        hdisplay = to_intel_crtc(crtc)->config.pipe_src_w;
        pixel_size = crtc->primary->fb->bits_per_pixel / 8;

        /* Use the small buffer method to calculate plane watermark */
        entries = ((clock * pixel_size / 1000) * display_latency_ns) / 1000;
        tlb_miss = display->fifo_size*display->cacheline_size - hdisplay * 8;
        if (tlb_miss > 0)
                entries += tlb_miss;
        entries = DIV_ROUND_UP(entries, display->cacheline_size);
        *plane_wm = entries + display->guard_size;
        if (*plane_wm > (int)display->max_wm)
                *plane_wm = display->max_wm;

        /* Use the large buffer method to calculate cursor watermark */
        line_time_us = max(htotal * 1000 / clock, 1);
        line_count = (cursor_latency_ns / line_time_us + 1000) / 1000;
        entries = line_count * to_intel_crtc(crtc)->cursor_width * pixel_size;
        tlb_miss = cursor->fifo_size*cursor->cacheline_size - hdisplay * 8;
        if (tlb_miss > 0)
                entries += tlb_miss;
        entries = DIV_ROUND_UP(entries, cursor->cacheline_size);
        *cursor_wm = entries + cursor->guard_size;
        if (*cursor_wm > (int)cursor->max_wm)
                *cursor_wm = (int)cursor->max_wm;

        return true;
}

/*
 * Check the wm result.
 *
 * If any calculated watermark values is larger than the maximum value that
 * can be programmed into the associated watermark register, that watermark
 * must be disabled.
 */
static bool g4x_check_srwm(struct drm_device *dev,
                           int display_wm, int cursor_wm,
                           const struct intel_watermark_params *display,
                           const struct intel_watermark_params *cursor)
{
        DRM_DEBUG_KMS("SR watermark: display plane %d, cursor %d\n",
                      display_wm, cursor_wm);

        if (display_wm > display->max_wm) {
                DRM_DEBUG_KMS("display watermark is too large(%d/%ld), disabling\n",
                              display_wm, display->max_wm);
                return false;
        }

        if (cursor_wm > cursor->max_wm) {
                DRM_DEBUG_KMS("cursor watermark is too large(%d/%ld), disabling\n",
                              cursor_wm, cursor->max_wm);
                return false;
        }

        if (!(display_wm || cursor_wm)) {
                DRM_DEBUG_KMS("SR latency is 0, disabling\n");
                return false;
        }

        return true;
}

static bool g4x_compute_srwm(struct drm_device *dev,
                             int plane,
                             int latency_ns,
                             const struct intel_watermark_params *display,
                             const struct intel_watermark_params *cursor,
                             int *display_wm, int *cursor_wm)
{
        struct drm_crtc *crtc;
        const struct drm_display_mode *adjusted_mode;
        int hdisplay, htotal, pixel_size, clock;
        unsigned long line_time_us;
        int line_count, line_size;
        int small, large;
        int entries;

        if (!latency_ns) {
                *display_wm = *cursor_wm = 0;
                return false;
        }

        crtc = intel_get_crtc_for_plane(dev, plane);
        adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
        clock = adjusted_mode->crtc_clock;
        htotal = adjusted_mode->crtc_htotal;
        hdisplay = to_intel_crtc(crtc)->config.pipe_src_w;
        pixel_size = crtc->primary->fb->bits_per_pixel / 8;

        line_time_us = max(htotal * 1000 / clock, 1);
        line_count = (latency_ns / line_time_us + 1000) / 1000;
        line_size = hdisplay * pixel_size;

        /* Use the minimum of the small and large buffer method for primary */
        small = ((clock * pixel_size / 1000) * latency_ns) / 1000;
        large = line_count * line_size;

        entries = DIV_ROUND_UP(min(small, large), display->cacheline_size);
        *display_wm = entries + display->guard_size;

        /* calculate the self-refresh watermark for display cursor */
        entries = line_count * pixel_size * to_intel_crtc(crtc)->cursor_width;
        entries = DIV_ROUND_UP(entries, cursor->cacheline_size);
        *cursor_wm = entries + cursor->guard_size;

        return g4x_check_srwm(dev,
                              *display_wm, *cursor_wm,
                              display, cursor);
}

static bool vlv_compute_drain_latency(struct drm_device *dev,
                                     int plane,
                                     int *plane_prec_mult,
                                     int *plane_dl,
                                     int *cursor_prec_mult,
                                     int *cursor_dl)
{
        struct drm_crtc *crtc;
        int clock, pixel_size;
        int entries;

        crtc = intel_get_crtc_for_plane(dev, plane);
        if (!intel_crtc_active(crtc))
                return false;

        clock = to_intel_crtc(crtc)->config.adjusted_mode.crtc_clock;
        pixel_size = crtc->primary->fb->bits_per_pixel / 8;     /* BPP */

        entries = (clock / 1000) * pixel_size;
        *plane_prec_mult = (entries > 128) ?
                DRAIN_LATENCY_PRECISION_64 : DRAIN_LATENCY_PRECISION_32;
        *plane_dl = (64 * (*plane_prec_mult) * 4) / entries;

        entries = (clock / 1000) * 4;   /* BPP is always 4 for cursor */
        *cursor_prec_mult = (entries > 128) ?
                DRAIN_LATENCY_PRECISION_64 : DRAIN_LATENCY_PRECISION_32;
        *cursor_dl = (64 * (*cursor_prec_mult) * 4) / entries;

        return true;
}

/*
 * Update drain latency registers of memory arbiter
 *
 * Valleyview SoC has a new memory arbiter and needs drain latency registers
 * to be programmed. Each plane has a drain latency multiplier and a drain
 * latency value.
 */

static void vlv_update_drain_latency(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        int planea_prec, planea_dl, planeb_prec, planeb_dl;
        int cursora_prec, cursora_dl, cursorb_prec, cursorb_dl;
        int plane_prec_mult, cursor_prec_mult; /* Precision multiplier is
                                                        either 16 or 32 */

        /* For plane A, Cursor A */
        if (vlv_compute_drain_latency(dev, 0, &plane_prec_mult, &planea_dl,
                                      &cursor_prec_mult, &cursora_dl)) {
                cursora_prec = (cursor_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
                        DDL_CURSORA_PRECISION_32 : DDL_CURSORA_PRECISION_64;
                planea_prec = (plane_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
                        DDL_PLANEA_PRECISION_32 : DDL_PLANEA_PRECISION_64;

                I915_WRITE(VLV_DDL1, cursora_prec |
                                (cursora_dl << DDL_CURSORA_SHIFT) |
                                planea_prec | planea_dl);
        }

        /* For plane B, Cursor B */
        if (vlv_compute_drain_latency(dev, 1, &plane_prec_mult, &planeb_dl,
                                      &cursor_prec_mult, &cursorb_dl)) {
                cursorb_prec = (cursor_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
                        DDL_CURSORB_PRECISION_32 : DDL_CURSORB_PRECISION_64;
                planeb_prec = (plane_prec_mult == DRAIN_LATENCY_PRECISION_32) ?
                        DDL_PLANEB_PRECISION_32 : DDL_PLANEB_PRECISION_64;

                I915_WRITE(VLV_DDL2, cursorb_prec |
                                (cursorb_dl << DDL_CURSORB_SHIFT) |
                                planeb_prec | planeb_dl);
        }
}

#define single_plane_enabled(mask) is_power_of_2(mask)

static void valleyview_update_wm(struct drm_crtc *crtc)
{
        struct drm_device *dev = crtc->dev;
        static const int sr_latency_ns = 12000;
        struct drm_i915_private *dev_priv = dev->dev_private;
        int planea_wm, planeb_wm, cursora_wm, cursorb_wm;
        int plane_sr, cursor_sr;
        int ignore_plane_sr, ignore_cursor_sr;
        unsigned int enabled = 0;
        bool cxsr_enabled;

        vlv_update_drain_latency(dev);

        if (g4x_compute_wm0(dev, PIPE_A,
                            &valleyview_wm_info, latency_ns,
                            &valleyview_cursor_wm_info, latency_ns,
                            &planea_wm, &cursora_wm))
                enabled |= 1 << PIPE_A;

        if (g4x_compute_wm0(dev, PIPE_B,
                            &valleyview_wm_info, latency_ns,
                            &valleyview_cursor_wm_info, latency_ns,
                            &planeb_wm, &cursorb_wm))
                enabled |= 1 << PIPE_B;

        if (single_plane_enabled(enabled) &&
            g4x_compute_srwm(dev, ffs(enabled) - 1,
                             sr_latency_ns,
                             &valleyview_wm_info,
                             &valleyview_cursor_wm_info,
                             &plane_sr, &ignore_cursor_sr) &&
            g4x_compute_srwm(dev, ffs(enabled) - 1,
                             2*sr_latency_ns,
                             &valleyview_wm_info,
                             &valleyview_cursor_wm_info,
                             &ignore_plane_sr, &cursor_sr)) {
                cxsr_enabled = true;
        } else {
                cxsr_enabled = false;
                intel_set_memory_cxsr(dev_priv, false);
                plane_sr = cursor_sr = 0;
        }

        DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
                      planea_wm, cursora_wm,
                      planeb_wm, cursorb_wm,
                      plane_sr, cursor_sr);

        I915_WRITE(DSPFW1,
                   (plane_sr << DSPFW_SR_SHIFT) |
                   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
                   (planeb_wm << DSPFW_PLANEB_SHIFT) |
                   planea_wm);
        I915_WRITE(DSPFW2,
                   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
                   (cursora_wm << DSPFW_CURSORA_SHIFT));
        I915_WRITE(DSPFW3,
                   (I915_READ(DSPFW3) & ~DSPFW_CURSOR_SR_MASK) |
                   (cursor_sr << DSPFW_CURSOR_SR_SHIFT));

        if (cxsr_enabled)
                intel_set_memory_cxsr(dev_priv, true);
}

static void g4x_update_wm(struct drm_crtc *crtc)
{
        struct drm_device *dev = crtc->dev;
        static const int sr_latency_ns = 12000;
        struct drm_i915_private *dev_priv = dev->dev_private;
        int planea_wm, planeb_wm, cursora_wm, cursorb_wm;
        int plane_sr, cursor_sr;
        unsigned int enabled = 0;
        bool cxsr_enabled;

        if (g4x_compute_wm0(dev, PIPE_A,
                            &g4x_wm_info, latency_ns,
                            &g4x_cursor_wm_info, latency_ns,
                            &planea_wm, &cursora_wm))
                enabled |= 1 << PIPE_A;

        if (g4x_compute_wm0(dev, PIPE_B,
                            &g4x_wm_info, latency_ns,
                            &g4x_cursor_wm_info, latency_ns,
                            &planeb_wm, &cursorb_wm))
                enabled |= 1 << PIPE_B;

        if (single_plane_enabled(enabled) &&
            g4x_compute_srwm(dev, ffs(enabled) - 1,
                             sr_latency_ns,
                             &g4x_wm_info,
                             &g4x_cursor_wm_info,
                             &plane_sr, &cursor_sr)) {
                cxsr_enabled = true;
        } else {
                cxsr_enabled = false;
                intel_set_memory_cxsr(dev_priv, false);
                plane_sr = cursor_sr = 0;
        }

        DRM_DEBUG_KMS("Setting FIFO watermarks - A: plane=%d, cursor=%d, B: plane=%d, cursor=%d, SR: plane=%d, cursor=%d\n",
                      planea_wm, cursora_wm,
                      planeb_wm, cursorb_wm,
                      plane_sr, cursor_sr);

        I915_WRITE(DSPFW1,
                   (plane_sr << DSPFW_SR_SHIFT) |
                   (cursorb_wm << DSPFW_CURSORB_SHIFT) |
                   (planeb_wm << DSPFW_PLANEB_SHIFT) |
                   planea_wm);
        I915_WRITE(DSPFW2,
                   (I915_READ(DSPFW2) & ~DSPFW_CURSORA_MASK) |
                   (cursora_wm << DSPFW_CURSORA_SHIFT));
        /* HPLL off in SR has some issues on G4x... disable it */
        I915_WRITE(DSPFW3,
                   (I915_READ(DSPFW3) & ~(DSPFW_HPLL_SR_EN | DSPFW_CURSOR_SR_MASK)) |
                   (cursor_sr << DSPFW_CURSOR_SR_SHIFT));

        if (cxsr_enabled)
                intel_set_memory_cxsr(dev_priv, true);
}

static void i965_update_wm(struct drm_crtc *unused_crtc)
{
        struct drm_device *dev = unused_crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_crtc *crtc;
        int srwm = 1;
        int cursor_sr = 16;
        bool cxsr_enabled;

        /* Calc sr entries for one plane configs */
        crtc = single_enabled_crtc(dev);
        if (crtc) {
                /* self-refresh has much higher latency */
                static const int sr_latency_ns = 12000;
                const struct drm_display_mode *adjusted_mode =
                        &to_intel_crtc(crtc)->config.adjusted_mode;
                int clock = adjusted_mode->crtc_clock;
                int htotal = adjusted_mode->crtc_htotal;
                int hdisplay = to_intel_crtc(crtc)->config.pipe_src_w;
                int pixel_size = crtc->primary->fb->bits_per_pixel / 8;
                unsigned long line_time_us;
                int entries;

                line_time_us = max(htotal * 1000 / clock, 1);

                /* Use ns/us then divide to preserve precision */
                entries = (((sr_latency_ns / line_time_us) + 1000) / 1000) *
                        pixel_size * hdisplay;
                entries = DIV_ROUND_UP(entries, I915_FIFO_LINE_SIZE);
                srwm = I965_FIFO_SIZE - entries;
                if (srwm < 0)
                        srwm = 1;
                srwm &= 0x1ff;
                DRM_DEBUG_KMS("self-refresh entries: %d, wm: %d\n",
                              entries, srwm);

                entries = (((sr_latency_ns / line_time_us) + 1000) / 1000) *
                        pixel_size * to_intel_crtc(crtc)->cursor_width;
                entries = DIV_ROUND_UP(entries,
                                          i965_cursor_wm_info.cacheline_size);
                cursor_sr = i965_cursor_wm_info.fifo_size -
                        (entries + i965_cursor_wm_info.guard_size);

                if (cursor_sr > i965_cursor_wm_info.max_wm)
                        cursor_sr = i965_cursor_wm_info.max_wm;

                DRM_DEBUG_KMS("self-refresh watermark: display plane %d "
                              "cursor %d\n", srwm, cursor_sr);

                cxsr_enabled = true;
        } else {
                cxsr_enabled = false;
                /* Turn off self refresh if both pipes are enabled */
                intel_set_memory_cxsr(dev_priv, false);
        }

        DRM_DEBUG_KMS("Setting FIFO watermarks - A: 8, B: 8, C: 8, SR %d\n",
                      srwm);

        /* 965 has limitations... */
        I915_WRITE(DSPFW1, (srwm << DSPFW_SR_SHIFT) |
                   (8 << 16) | (8 << 8) | (8 << 0));
        I915_WRITE(DSPFW2, (8 << 8) | (8 << 0));
        /* update cursor SR watermark */
        I915_WRITE(DSPFW3, (cursor_sr << DSPFW_CURSOR_SR_SHIFT));

        if (cxsr_enabled)
                intel_set_memory_cxsr(dev_priv, true);
}

static void i9xx_update_wm(struct drm_crtc *unused_crtc)
{
        struct drm_device *dev = unused_crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        const struct intel_watermark_params *wm_info;
        uint32_t fwater_lo;
        uint32_t fwater_hi;
        int cwm, srwm = 1;
        int fifo_size;
        int planea_wm, planeb_wm;
        struct drm_crtc *crtc, *enabled = NULL;

        if (IS_I945GM(dev))
                wm_info = &i945_wm_info;
        else if (!IS_GEN2(dev))
                wm_info = &i915_wm_info;
        else
                wm_info = &i830_wm_info;

        fifo_size = dev_priv->display.get_fifo_size(dev, 0);
        crtc = intel_get_crtc_for_plane(dev, 0);
        if (intel_crtc_active(crtc)) {
                const struct drm_display_mode *adjusted_mode;
                int cpp = crtc->primary->fb->bits_per_pixel / 8;
                if (IS_GEN2(dev))
                        cpp = 4;

                adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
                planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
                                               wm_info, fifo_size, cpp,
                                               latency_ns);
                enabled = crtc;
        } else
                planea_wm = fifo_size - wm_info->guard_size;

        fifo_size = dev_priv->display.get_fifo_size(dev, 1);
        crtc = intel_get_crtc_for_plane(dev, 1);
        if (intel_crtc_active(crtc)) {
                const struct drm_display_mode *adjusted_mode;
                int cpp = crtc->primary->fb->bits_per_pixel / 8;
                if (IS_GEN2(dev))
                        cpp = 4;

                adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
                planeb_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
                                               wm_info, fifo_size, cpp,
                                               latency_ns);
                if (enabled == NULL)
                        enabled = crtc;
                else
                        enabled = NULL;
        } else
                planeb_wm = fifo_size - wm_info->guard_size;

        DRM_DEBUG_KMS("FIFO watermarks - A: %d, B: %d\n", planea_wm, planeb_wm);

        if (IS_I915GM(dev) && enabled) {
                struct drm_i915_gem_object *obj;

                obj = intel_fb_obj(enabled->primary->fb);

                /* self-refresh seems busted with untiled */
                if (obj->tiling_mode == I915_TILING_NONE)
                        enabled = NULL;
        }

        /*
         * Overlay gets an aggressive default since video jitter is bad.
         */
        cwm = 2;

        /* Play safe and disable self-refresh before adjusting watermarks. */
        intel_set_memory_cxsr(dev_priv, false);

        /* Calc sr entries for one plane configs */
        if (HAS_FW_BLC(dev) && enabled) {
                /* self-refresh has much higher latency */
                static const int sr_latency_ns = 6000;
                const struct drm_display_mode *adjusted_mode =
                        &to_intel_crtc(enabled)->config.adjusted_mode;
                int clock = adjusted_mode->crtc_clock;
                int htotal = adjusted_mode->crtc_htotal;
                int hdisplay = to_intel_crtc(enabled)->config.pipe_src_w;
                int pixel_size = enabled->primary->fb->bits_per_pixel / 8;
                unsigned long line_time_us;
                int entries;

                line_time_us = max(htotal * 1000 / clock, 1);

                /* Use ns/us then divide to preserve precision */
                entries = (((sr_latency_ns / line_time_us) + 1000) / 1000) *
                        pixel_size * hdisplay;
                entries = DIV_ROUND_UP(entries, wm_info->cacheline_size);
                DRM_DEBUG_KMS("self-refresh entries: %d\n", entries);
                srwm = wm_info->fifo_size - entries;
                if (srwm < 0)
                        srwm = 1;

                if (IS_I945G(dev) || IS_I945GM(dev))
                        I915_WRITE(FW_BLC_SELF,
                                   FW_BLC_SELF_FIFO_MASK | (srwm & 0xff));
                else if (IS_I915GM(dev))
                        I915_WRITE(FW_BLC_SELF, srwm & 0x3f);
        }

        DRM_DEBUG_KMS("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);

        if (enabled)
                intel_set_memory_cxsr(dev_priv, true);
}

static void i845_update_wm(struct drm_crtc *unused_crtc)
{
        struct drm_device *dev = unused_crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct drm_crtc *crtc;
        const struct drm_display_mode *adjusted_mode;
        uint32_t fwater_lo;
        int planea_wm;

        crtc = single_enabled_crtc(dev);
        if (crtc == NULL)
                return;

        adjusted_mode = &to_intel_crtc(crtc)->config.adjusted_mode;
        planea_wm = intel_calculate_wm(adjusted_mode->crtc_clock,
                                       &i845_wm_info,
                                       dev_priv->display.get_fifo_size(dev, 0),
                                       4, latency_ns);
        fwater_lo = I915_READ(FW_BLC) & ~0xfff;
        fwater_lo |= (3<<8) | planea_wm;

        DRM_DEBUG_KMS("Setting FIFO watermarks - A: %d\n", planea_wm);

        I915_WRITE(FW_BLC, fwater_lo);
}

static uint32_t ilk_pipe_pixel_rate(struct drm_device *dev,
                                    struct drm_crtc *crtc)
{
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        uint32_t pixel_rate;

        pixel_rate = intel_crtc->config.adjusted_mode.crtc_clock;

        /* We only use IF-ID interlacing. If we ever use PF-ID we'll need to
         * adjust the pixel_rate here. */

        if (intel_crtc->config.pch_pfit.enabled) {
                uint64_t pipe_w, pipe_h, pfit_w, pfit_h;
                uint32_t pfit_size = intel_crtc->config.pch_pfit.size;

                pipe_w = intel_crtc->config.pipe_src_w;
                pipe_h = intel_crtc->config.pipe_src_h;
                pfit_w = (pfit_size >> 16) & 0xFFFF;
                pfit_h = pfit_size & 0xFFFF;
                if (pipe_w < pfit_w)
                        pipe_w = pfit_w;
                if (pipe_h < pfit_h)
                        pipe_h = pfit_h;

                pixel_rate = div_u64((uint64_t) pixel_rate * pipe_w * pipe_h,
                                     pfit_w * pfit_h);
        }

        return pixel_rate;
}

/* latency must be in 0.1us units. */
static uint32_t ilk_wm_method1(uint32_t pixel_rate, uint8_t bytes_per_pixel,
                               uint32_t latency)
{
        uint64_t ret;

        if (WARN(latency == 0, "Latency value missing\n"))
                return UINT_MAX;

        ret = (uint64_t) pixel_rate * bytes_per_pixel * latency;
        ret = DIV_ROUND_UP_ULL(ret, 64 * 10000) + 2;

        return ret;
}

/* latency must be in 0.1us units. */
static uint32_t ilk_wm_method2(uint32_t pixel_rate, uint32_t pipe_htotal,
                               uint32_t horiz_pixels, uint8_t bytes_per_pixel,
                               uint32_t latency)
{
        uint32_t ret;

        if (WARN(latency == 0, "Latency value missing\n"))
                return UINT_MAX;

        ret = (latency * pixel_rate) / (pipe_htotal * 10000);
        ret = (ret + 1) * horiz_pixels * bytes_per_pixel;
        ret = DIV_ROUND_UP(ret, 64) + 2;
        return ret;
}

static uint32_t ilk_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels,
                           uint8_t bytes_per_pixel)
{
        return DIV_ROUND_UP(pri_val * 64, horiz_pixels * bytes_per_pixel) + 2;
}

struct ilk_pipe_wm_parameters {
        bool active;
        uint32_t pipe_htotal;
        uint32_t pixel_rate;
        struct intel_plane_wm_parameters pri;
        struct intel_plane_wm_parameters spr;
        struct intel_plane_wm_parameters cur;
};

struct ilk_wm_maximums {
        uint16_t pri;
        uint16_t spr;
        uint16_t cur;
        uint16_t fbc;
};

/* used in computing the new watermarks state */
struct intel_wm_config {
        unsigned int num_pipes_active;
        bool sprites_enabled;
        bool sprites_scaled;
};

/*
 * For both WM_PIPE and WM_LP.
 * mem_value must be in 0.1us units.
 */
static uint32_t ilk_compute_pri_wm(const struct ilk_pipe_wm_parameters *params,
                                   uint32_t mem_value,
                                   bool is_lp)
{
        uint32_t method1, method2;

        if (!params->active || !params->pri.enabled)
                return 0;

        method1 = ilk_wm_method1(params->pixel_rate,
                                 params->pri.bytes_per_pixel,
                                 mem_value);

        if (!is_lp)
                return method1;

        method2 = ilk_wm_method2(params->pixel_rate,
                                 params->pipe_htotal,
                                 params->pri.horiz_pixels,
                                 params->pri.bytes_per_pixel,
                                 mem_value);

        return min(method1, method2);
}

/*
 * For both WM_PIPE and WM_LP.
 * mem_value must be in 0.1us units.
 */
static uint32_t ilk_compute_spr_wm(const struct ilk_pipe_wm_parameters *params,
                                   uint32_t mem_value)
{
        uint32_t method1, method2;

        if (!params->active || !params->spr.enabled)
                return 0;

        method1 = ilk_wm_method1(params->pixel_rate,
                                 params->spr.bytes_per_pixel,
                                 mem_value);
        method2 = ilk_wm_method2(params->pixel_rate,
                                 params->pipe_htotal,
                                 params->spr.horiz_pixels,
                                 params->spr.bytes_per_pixel,
                                 mem_value);
        return min(method1, method2);
}

/*
 * For both WM_PIPE and WM_LP.
 * mem_value must be in 0.1us units.
 */
static uint32_t ilk_compute_cur_wm(const struct ilk_pipe_wm_parameters *params,
                                   uint32_t mem_value)
{
        if (!params->active || !params->cur.enabled)
                return 0;

        return ilk_wm_method2(params->pixel_rate,
                              params->pipe_htotal,
                              params->cur.horiz_pixels,
                              params->cur.bytes_per_pixel,
                              mem_value);
}

/* Only for WM_LP. */
static uint32_t ilk_compute_fbc_wm(const struct ilk_pipe_wm_parameters *params,
                                   uint32_t pri_val)
{
        if (!params->active || !params->pri.enabled)
                return 0;

        return ilk_wm_fbc(pri_val,
                          params->pri.horiz_pixels,
                          params->pri.bytes_per_pixel);
}

static unsigned int ilk_display_fifo_size(const struct drm_device *dev)
{
        if (INTEL_INFO(dev)->gen >= 8)
                return 3072;
        else if (INTEL_INFO(dev)->gen >= 7)
                return 768;
        else
                return 512;
}

static unsigned int ilk_plane_wm_reg_max(const struct drm_device *dev,
                                         int level, bool is_sprite)
{
        if (INTEL_INFO(dev)->gen >= 8)
                /* BDW primary/sprite plane watermarks */
                return level == 0 ? 255 : 2047;
        else if (INTEL_INFO(dev)->gen >= 7)
                /* IVB/HSW primary/sprite plane watermarks */
                return level == 0 ? 127 : 1023;
        else if (!is_sprite)
                /* ILK/SNB primary plane watermarks */
                return level == 0 ? 127 : 511;
        else
                /* ILK/SNB sprite plane watermarks */
                return level == 0 ? 63 : 255;
}

static unsigned int ilk_cursor_wm_reg_max(const struct drm_device *dev,
                                          int level)
{
        if (INTEL_INFO(dev)->gen >= 7)
                return level == 0 ? 63 : 255;
        else
                return level == 0 ? 31 : 63;
}

static unsigned int ilk_fbc_wm_reg_max(const struct drm_device *dev)
{
        if (INTEL_INFO(dev)->gen >= 8)
                return 31;
        else
                return 15;
}

/* Calculate the maximum primary/sprite plane watermark */
static unsigned int ilk_plane_wm_max(const struct drm_device *dev,
                                     int level,
                                     const struct intel_wm_config *config,
                                     enum intel_ddb_partitioning ddb_partitioning,
                                     bool is_sprite)
{
        unsigned int fifo_size = ilk_display_fifo_size(dev);

        /* if sprites aren't enabled, sprites get nothing */
        if (is_sprite && !config->sprites_enabled)
                return 0;

        /* HSW allows LP1+ watermarks even with multiple pipes */
        if (level == 0 || config->num_pipes_active > 1) {
                fifo_size /= INTEL_INFO(dev)->num_pipes;

                /*
                 * For some reason the non self refresh
                 * FIFO size is only half of the self
                 * refresh FIFO size on ILK/SNB.
                 */
                if (INTEL_INFO(dev)->gen <= 6)
                        fifo_size /= 2;
        }

        if (config->sprites_enabled) {
                /* level 0 is always calculated with 1:1 split */
                if (level > 0 && ddb_partitioning == INTEL_DDB_PART_5_6) {
                        if (is_sprite)
                                fifo_size *= 5;
                        fifo_size /= 6;
                } else {
                        fifo_size /= 2;
                }
        }

        /* clamp to max that the registers can hold */
        return min(fifo_size, ilk_plane_wm_reg_max(dev, level, is_sprite));
}

/* Calculate the maximum cursor plane watermark */
static unsigned int ilk_cursor_wm_max(const struct drm_device *dev,
                                      int level,
                                      const struct intel_wm_config *config)
{
        /* HSW LP1+ watermarks w/ multiple pipes */
        if (level > 0 && config->num_pipes_active > 1)
                return 64;

        /* otherwise just report max that registers can hold */
        return ilk_cursor_wm_reg_max(dev, level);
}

static void ilk_compute_wm_maximums(const struct drm_device *dev,
                                    int level,
                                    const struct intel_wm_config *config,
                                    enum intel_ddb_partitioning ddb_partitioning,
                                    struct ilk_wm_maximums *max)
{
        max->pri = ilk_plane_wm_max(dev, level, config, ddb_partitioning, false);
        max->spr = ilk_plane_wm_max(dev, level, config, ddb_partitioning, true);
        max->cur = ilk_cursor_wm_max(dev, level, config);
        max->fbc = ilk_fbc_wm_reg_max(dev);
}

static void ilk_compute_wm_reg_maximums(struct drm_device *dev,
                                        int level,
                                        struct ilk_wm_maximums *max)
{
        max->pri = ilk_plane_wm_reg_max(dev, level, false);
        max->spr = ilk_plane_wm_reg_max(dev, level, true);
        max->cur = ilk_cursor_wm_reg_max(dev, level);
        max->fbc = ilk_fbc_wm_reg_max(dev);
}

static bool ilk_validate_wm_level(int level,
                                  const struct ilk_wm_maximums *max,
                                  struct intel_wm_level *result)
{
        bool ret;

        /* already determined to be invalid? */
        if (!result->enable)
                return false;

        result->enable = result->pri_val <= max->pri &&
                         result->spr_val <= max->spr &&
                         result->cur_val <= max->cur;

        ret = result->enable;

        /*
         * HACK until we can pre-compute everything,
         * and thus fail gracefully if LP0 watermarks
         * are exceeded...
         */
        if (level == 0 && !result->enable) {
                if (result->pri_val > max->pri)

                        DRM_DEBUG_KMS("Primary WM%d too large %u (max %u)\n",
                                      level, result->pri_val, max->pri);
                if (result->spr_val > max->spr)
                        DRM_DEBUG_KMS("Sprite WM%d too large %u (max %u)\n",
                                      level, result->spr_val, max->spr);
                if (result->cur_val > max->cur)
                        DRM_DEBUG_KMS("Cursor WM%d too large %u (max %u)\n",
                                      level, result->cur_val, max->cur);

                result->pri_val = min_t(uint32_t, result->pri_val, max->pri);
                result->spr_val = min_t(uint32_t, result->spr_val, max->spr);
                result->cur_val = min_t(uint32_t, result->cur_val, max->cur);
                result->enable = true;
        }

        return ret;
}

static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv,
                                 int level,
                                 const struct ilk_pipe_wm_parameters *p,
                                 struct intel_wm_level *result)
{
        uint16_t pri_latency = dev_priv->wm.pri_latency[level];
        uint16_t spr_latency = dev_priv->wm.spr_latency[level];
        uint16_t cur_latency = dev_priv->wm.cur_latency[level];

        /* WM1+ latency values stored in 0.5us units */
        if (level > 0) {
                pri_latency *= 5;
                spr_latency *= 5;
                cur_latency *= 5;
        }

        result->pri_val = ilk_compute_pri_wm(p, pri_latency, level);
        result->spr_val = ilk_compute_spr_wm(p, spr_latency);
        result->cur_val = ilk_compute_cur_wm(p, cur_latency);
        result->fbc_val = ilk_compute_fbc_wm(p, result->pri_val);
        result->enable = true;
}

static uint32_t
hsw_compute_linetime_wm(struct drm_device *dev, struct drm_crtc *crtc)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        struct drm_display_mode *mode = &intel_crtc->config.adjusted_mode;
        u32 linetime, ips_linetime;

        if (!intel_crtc_active(crtc))
                return 0;

        /* The WM are computed with base on how long it takes to fill a single
         * row at the given clock rate, multiplied by 8.
         * */
        linetime = DIV_ROUND_CLOSEST(mode->crtc_htotal * 1000 * 8,
                                     mode->crtc_clock);
        ips_linetime = DIV_ROUND_CLOSEST(mode->crtc_htotal * 1000 * 8,
                                         intel_ddi_get_cdclk_freq(dev_priv));

        return PIPE_WM_LINETIME_IPS_LINETIME(ips_linetime) |
               PIPE_WM_LINETIME_TIME(linetime);
}

static void intel_read_wm_latency(struct drm_device *dev, uint16_t wm[5])
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
                uint64_t sskpd = I915_READ64(MCH_SSKPD);

                wm[0] = (sskpd >> 56) & 0xFF;
                if (wm[0] == 0)
                        wm[0] = sskpd & 0xF;
                wm[1] = (sskpd >> 4) & 0xFF;
                wm[2] = (sskpd >> 12) & 0xFF;
                wm[3] = (sskpd >> 20) & 0x1FF;
                wm[4] = (sskpd >> 32) & 0x1FF;
        } else if (INTEL_INFO(dev)->gen >= 6) {
                uint32_t sskpd = I915_READ(MCH_SSKPD);

                wm[0] = (sskpd >> SSKPD_WM0_SHIFT) & SSKPD_WM_MASK;
                wm[1] = (sskpd >> SSKPD_WM1_SHIFT) & SSKPD_WM_MASK;
                wm[2] = (sskpd >> SSKPD_WM2_SHIFT) & SSKPD_WM_MASK;
                wm[3] = (sskpd >> SSKPD_WM3_SHIFT) & SSKPD_WM_MASK;
        } else if (INTEL_INFO(dev)->gen >= 5) {
                uint32_t mltr = I915_READ(MLTR_ILK);

                /* ILK primary LP0 latency is 700 ns */
                wm[0] = 7;
                wm[1] = (mltr >> MLTR_WM1_SHIFT) & ILK_SRLT_MASK;
                wm[2] = (mltr >> MLTR_WM2_SHIFT) & ILK_SRLT_MASK;
        }
}

static void intel_fixup_spr_wm_latency(struct drm_device *dev, uint16_t wm[5])
{
        /* ILK sprite LP0 latency is 1300 ns */
        if (INTEL_INFO(dev)->gen == 5)
                wm[0] = 13;
}

static void intel_fixup_cur_wm_latency(struct drm_device *dev, uint16_t wm[5])
{
        /* ILK cursor LP0 latency is 1300 ns */
        if (INTEL_INFO(dev)->gen == 5)
                wm[0] = 13;

        /* WaDoubleCursorLP3Latency:ivb */
        if (IS_IVYBRIDGE(dev))
                wm[3] *= 2;
}

int ilk_wm_max_level(const struct drm_device *dev)
{
        /* how many WM levels are we expecting */
        if (IS_HASWELL(dev) || IS_BROADWELL(dev))
                return 4;
        else if (INTEL_INFO(dev)->gen >= 6)
                return 3;
        else
                return 2;
}

static void intel_print_wm_latency(struct drm_device *dev,
                                   const char *name,
                                   const uint16_t wm[5])
{
        int level, max_level = ilk_wm_max_level(dev);

        for (level = 0; level <= max_level; level++) {
                unsigned int latency = wm[level];

                if (latency == 0) {
                        DRM_ERROR("%s WM%d latency not provided\n",
                                  name, level);
                        continue;
                }

                /* WM1+ latency values in 0.5us units */
                if (level > 0)
                        latency *= 5;

                DRM_DEBUG_KMS("%s WM%d latency %u (%u.%u usec)\n",
                              name, level, wm[level],
                              latency / 10, latency % 10);
        }
}

static bool ilk_increase_wm_latency(struct drm_i915_private *dev_priv,
                                    uint16_t wm[5], uint16_t min)
{
        int level, max_level = ilk_wm_max_level(dev_priv->dev);

        if (wm[0] >= min)
                return false;

        wm[0] = max(wm[0], min);
        for (level = 1; level <= max_level; level++)
                wm[level] = max_t(uint16_t, wm[level], DIV_ROUND_UP(min, 5));

        return true;
}

static void snb_wm_latency_quirk(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        bool changed;

        /*
         * The BIOS provided WM memory latency values are often
         * inadequate for high resolution displays. Adjust them.
         */
        changed = ilk_increase_wm_latency(dev_priv, dev_priv->wm.pri_latency, 12) |
                ilk_increase_wm_latency(dev_priv, dev_priv->wm.spr_latency, 12) |
                ilk_increase_wm_latency(dev_priv, dev_priv->wm.cur_latency, 12);

        if (!changed)
                return;

        DRM_DEBUG_KMS("WM latency values increased to avoid potential underruns\n");
        intel_print_wm_latency(dev, "Primary", dev_priv->wm.pri_latency);
        intel_print_wm_latency(dev, "Sprite", dev_priv->wm.spr_latency);
        intel_print_wm_latency(dev, "Cursor", dev_priv->wm.cur_latency);
}

static void ilk_setup_wm_latency(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        intel_read_wm_latency(dev, dev_priv->wm.pri_latency);

        memcpy(dev_priv->wm.spr_latency, dev_priv->wm.pri_latency,
               sizeof(dev_priv->wm.pri_latency));
        memcpy(dev_priv->wm.cur_latency, dev_priv->wm.pri_latency,
               sizeof(dev_priv->wm.pri_latency));

        intel_fixup_spr_wm_latency(dev, dev_priv->wm.spr_latency);
        intel_fixup_cur_wm_latency(dev, dev_priv->wm.cur_latency);

        intel_print_wm_latency(dev, "Primary", dev_priv->wm.pri_latency);
        intel_print_wm_latency(dev, "Sprite", dev_priv->wm.spr_latency);
        intel_print_wm_latency(dev, "Cursor", dev_priv->wm.cur_latency);

        if (IS_GEN6(dev))
                snb_wm_latency_quirk(dev);
}

static void ilk_compute_wm_parameters(struct drm_crtc *crtc,
                                      struct ilk_pipe_wm_parameters *p)
{
        struct drm_device *dev = crtc->dev;
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        enum pipe pipe = intel_crtc->pipe;
        struct drm_plane *plane;

        if (!intel_crtc_active(crtc))
                return;

        p->active = true;
        p->pipe_htotal = intel_crtc->config.adjusted_mode.crtc_htotal;
        p->pixel_rate = ilk_pipe_pixel_rate(dev, crtc);
        p->pri.bytes_per_pixel = crtc->primary->fb->bits_per_pixel / 8;
        p->cur.bytes_per_pixel = 4;
        p->pri.horiz_pixels = intel_crtc->config.pipe_src_w;
        p->cur.horiz_pixels = intel_crtc->cursor_width;
        /* TODO: for now, assume primary and cursor planes are always enabled. */
        p->pri.enabled = true;
        p->cur.enabled = true;

        drm_for_each_legacy_plane(plane, &dev->mode_config.plane_list) {
                struct intel_plane *intel_plane = to_intel_plane(plane);

                if (intel_plane->pipe == pipe) {
                        p->spr = intel_plane->wm;
                        break;
                }
        }
}

static void ilk_compute_wm_config(struct drm_device *dev,
                                  struct intel_wm_config *config)
{
        struct intel_crtc *intel_crtc;

        /* Compute the currently _active_ config */
        for_each_intel_crtc(dev, intel_crtc) {
                const struct intel_pipe_wm *wm = &intel_crtc->wm.active;

                if (!wm->pipe_enabled)
                        continue;

                config->sprites_enabled |= wm->sprites_enabled;
                config->sprites_scaled |= wm->sprites_scaled;
                config->num_pipes_active++;
        }
}

/* Compute new watermarks for the pipe */
static bool intel_compute_pipe_wm(struct drm_crtc *crtc,
                                  const struct ilk_pipe_wm_parameters *params,
                                  struct intel_pipe_wm *pipe_wm)
{
        struct drm_device *dev = crtc->dev;
        const struct drm_i915_private *dev_priv = dev->dev_private;
        int level, max_level = ilk_wm_max_level(dev);
        /* LP0 watermark maximums depend on this pipe alone */
        struct intel_wm_config config = {
                .num_pipes_active = 1,
                .sprites_enabled = params->spr.enabled,
                .sprites_scaled = params->spr.scaled,
        };
        struct ilk_wm_maximums max;

        pipe_wm->pipe_enabled = params->active;
        pipe_wm->sprites_enabled = params->spr.enabled;
        pipe_wm->sprites_scaled = params->spr.scaled;

        /* ILK/SNB: LP2+ watermarks only w/o sprites */
        if (INTEL_INFO(dev)->gen <= 6 && params->spr.enabled)
                max_level = 1;

        /* ILK/SNB/IVB: LP1+ watermarks only w/o scaling */
        if (params->spr.scaled)
                max_level = 0;

        ilk_compute_wm_level(dev_priv, 0, params, &pipe_wm->wm[0]);

        if (IS_HASWELL(dev) || IS_BROADWELL(dev))
                pipe_wm->linetime = hsw_compute_linetime_wm(dev, crtc);

        /* LP0 watermarks always use 1/2 DDB partitioning */
        ilk_compute_wm_maximums(dev, 0, &config, INTEL_DDB_PART_1_2, &max);

        /* At least LP0 must be valid */
        if (!ilk_validate_wm_level(0, &max, &pipe_wm->wm[0]))
                return false;

        ilk_compute_wm_reg_maximums(dev, 1, &max);

        for (level = 1; level <= max_level; level++) {
                struct intel_wm_level wm = {};

                ilk_compute_wm_level(dev_priv, level, params, &wm);

                /*
                 * Disable any watermark level that exceeds the
                 * register maximums since such watermarks are
                 * always invalid.
                 */
                if (!ilk_validate_wm_level(level, &max, &wm))
                        break;

                pipe_wm->wm[level] = wm;
        }

        return true;
}

/*
 * Merge the watermarks from all active pipes for a specific level.
 */
static void ilk_merge_wm_level(struct drm_device *dev,
                               int level,
                               struct intel_wm_level *ret_wm)
{
        const struct intel_crtc *intel_crtc;

        ret_wm->enable = true;

        for_each_intel_crtc(dev, intel_crtc) {
                const struct intel_pipe_wm *active = &intel_crtc->wm.active;
                const struct intel_wm_level *wm = &active->wm[level];

                if (!active->pipe_enabled)
                        continue;

                /*
                 * The watermark values may have been used in the past,
                 * so we must maintain them in the registers for some
                 * time even if the level is now disabled.
                 */
                if (!wm->enable)
                        ret_wm->enable = false;

                ret_wm->pri_val = max(ret_wm->pri_val, wm->pri_val);
                ret_wm->spr_val = max(ret_wm->spr_val, wm->spr_val);
                ret_wm->cur_val = max(ret_wm->cur_val, wm->cur_val);
                ret_wm->fbc_val = max(ret_wm->fbc_val, wm->fbc_val);
        }
}

/*
 * Merge all low power watermarks for all active pipes.
 */
static void ilk_wm_merge(struct drm_device *dev,
                         const struct intel_wm_config *config,
                         const struct ilk_wm_maximums *max,
                         struct intel_pipe_wm *merged)
{
        int level, max_level = ilk_wm_max_level(dev);
        int last_enabled_level = max_level;

        /* ILK/SNB/IVB: LP1+ watermarks only w/ single pipe */
        if ((INTEL_INFO(dev)->gen <= 6 || IS_IVYBRIDGE(dev)) &&
            config->num_pipes_active > 1)
                return;

        /* ILK: FBC WM must be disabled always */
        merged->fbc_wm_enabled = INTEL_INFO(dev)->gen >= 6;

        /* merge each WM1+ level */
        for (level = 1; level <= max_level; level++) {
                struct intel_wm_level *wm = &merged->wm[level];

                ilk_merge_wm_level(dev, level, wm);

                if (level > last_enabled_level)
                        wm->enable = false;
                else if (!ilk_validate_wm_level(level, max, wm))
                        /* make sure all following levels get disabled */
                        last_enabled_level = level - 1;

                /*
                 * The spec says it is preferred to disable
                 * FBC WMs instead of disabling a WM level.
                 */
                if (wm->fbc_val > max->fbc) {
                        if (wm->enable)
                                merged->fbc_wm_enabled = false;
                        wm->fbc_val = 0;
                }
        }

        /* ILK: LP2+ must be disabled when FBC WM is disabled but FBC enabled */
        /*
         * FIXME this is racy. FBC might get enabled later.
         * What we should check here is whether FBC can be
         * enabled sometime later.
         */
        if (IS_GEN5(dev) && !merged->fbc_wm_enabled && intel_fbc_enabled(dev)) {
                for (level = 2; level <= max_level; level++) {
                        struct intel_wm_level *wm = &merged->wm[level];

                        wm->enable = false;
                }
        }
}

static int ilk_wm_lp_to_level(int wm_lp, const struct intel_pipe_wm *pipe_wm)
{
        /* LP1,LP2,LP3 levels are either 1,2,3 or 1,3,4 */
        return wm_lp + (wm_lp >= 2 && pipe_wm->wm[4].enable);
}

/* The value we need to program into the WM_LPx latency field */
static unsigned int ilk_wm_lp_latency(struct drm_device *dev, int level)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        if (IS_HASWELL(dev) || IS_BROADWELL(dev))
                return 2 * level;
        else
                return dev_priv->wm.pri_latency[level];
}

static void ilk_compute_wm_results(struct drm_device *dev,
                                   const struct intel_pipe_wm *merged,
                                   enum intel_ddb_partitioning partitioning,
                                   struct ilk_wm_values *results)
{
        struct intel_crtc *intel_crtc;
        int level, wm_lp;

        results->enable_fbc_wm = merged->fbc_wm_enabled;
        results->partitioning = partitioning;

        /* LP1+ register values */
        for (wm_lp = 1; wm_lp <= 3; wm_lp++) {
                const struct intel_wm_level *r;

                level = ilk_wm_lp_to_level(wm_lp, merged);

                r = &merged->wm[level];

                /*
                 * Maintain the watermark values even if the level is
                 * disabled. Doing otherwise could cause underruns.
                 */
                results->wm_lp[wm_lp - 1] =
                        (ilk_wm_lp_latency(dev, level) << WM1_LP_LATENCY_SHIFT) |
                        (r->pri_val << WM1_LP_SR_SHIFT) |
                        r->cur_val;

                if (r->enable)
                        results->wm_lp[wm_lp - 1] |= WM1_LP_SR_EN;

                if (INTEL_INFO(dev)->gen >= 8)
                        results->wm_lp[wm_lp - 1] |=
                                r->fbc_val << WM1_LP_FBC_SHIFT_BDW;
                else
                        results->wm_lp[wm_lp - 1] |=
                                r->fbc_val << WM1_LP_FBC_SHIFT;

                /*
                 * Always set WM1S_LP_EN when spr_val != 0, even if the
                 * level is disabled. Doing otherwise could cause underruns.
                 */
                if (INTEL_INFO(dev)->gen <= 6 && r->spr_val) {
                        WARN_ON(wm_lp != 1);
                        results->wm_lp_spr[wm_lp - 1] = WM1S_LP_EN | r->spr_val;
                } else
                        results->wm_lp_spr[wm_lp - 1] = r->spr_val;
        }

        /* LP0 register values */
        for_each_intel_crtc(dev, intel_crtc) {
                enum pipe pipe = intel_crtc->pipe;
                const struct intel_wm_level *r =
                        &intel_crtc->wm.active.wm[0];

                if (WARN_ON(!r->enable))
                        continue;

                results->wm_linetime[pipe] = intel_crtc->wm.active.linetime;

                results->wm_pipe[pipe] =
                        (r->pri_val << WM0_PIPE_PLANE_SHIFT) |
                        (r->spr_val << WM0_PIPE_SPRITE_SHIFT) |
                        r->cur_val;
        }
}

/* Find the result with the highest level enabled. Check for enable_fbc_wm in
 * case both are at the same level. Prefer r1 in case they're the same. */
static struct intel_pipe_wm *ilk_find_best_result(struct drm_device *dev,
                                                  struct intel_pipe_wm *r1,
                                                  struct intel_pipe_wm *r2)
{
        int level, max_level = ilk_wm_max_level(dev);
        int level1 = 0, level2 = 0;

        for (level = 1; level <= max_level; level++) {
                if (r1->wm[level].enable)
                        level1 = level;
                if (r2->wm[level].enable)
                        level2 = level;
        }

        if (level1 == level2) {
                if (r2->fbc_wm_enabled && !r1->fbc_wm_enabled)
                        return r2;
                else
                        return r1;
        } else if (level1 > level2) {
                return r1;
        } else {
                return r2;
        }
}

/* dirty bits used to track which watermarks need changes */
#define WM_DIRTY_PIPE(pipe) (1 << (pipe))
#define WM_DIRTY_LINETIME(pipe) (1 << (8 + (pipe)))
#define WM_DIRTY_LP(wm_lp) (1 << (15 + (wm_lp)))
#define WM_DIRTY_LP_ALL (WM_DIRTY_LP(1) | WM_DIRTY_LP(2) | WM_DIRTY_LP(3))
#define WM_DIRTY_FBC (1 << 24)
#define WM_DIRTY_DDB (1 << 25)

static unsigned int ilk_compute_wm_dirty(struct drm_device *dev,
                                         const struct ilk_wm_values *old,
                                         const struct ilk_wm_values *new)
{
        unsigned int dirty = 0;
        enum pipe pipe;
        int wm_lp;

        for_each_pipe(pipe) {
                if (old->wm_linetime[pipe] != new->wm_linetime[pipe]) {
                        dirty |= WM_DIRTY_LINETIME(pipe);
                        /* Must disable LP1+ watermarks too */
                        dirty |= WM_DIRTY_LP_ALL;
                }

                if (old->wm_pipe[pipe] != new->wm_pipe[pipe]) {
                        dirty |= WM_DIRTY_PIPE(pipe);
                        /* Must disable LP1+ watermarks too */
                        dirty |= WM_DIRTY_LP_ALL;
                }
        }

        if (old->enable_fbc_wm != new->enable_fbc_wm) {
                dirty |= WM_DIRTY_FBC;
                /* Must disable LP1+ watermarks too */
                dirty |= WM_DIRTY_LP_ALL;
        }

        if (old->partitioning != new->partitioning) {
                dirty |= WM_DIRTY_DDB;
                /* Must disable LP1+ watermarks too */
                dirty |= WM_DIRTY_LP_ALL;
        }

        /* LP1+ watermarks already deemed dirty, no need to continue */
        if (dirty & WM_DIRTY_LP_ALL)
                return dirty;

        /* Find the lowest numbered LP1+ watermark in need of an update... */
        for (wm_lp = 1; wm_lp <= 3; wm_lp++) {
                if (old->wm_lp[wm_lp - 1] != new->wm_lp[wm_lp - 1] ||
                    old->wm_lp_spr[wm_lp - 1] != new->wm_lp_spr[wm_lp - 1])
                        break;
        }

        /* ...and mark it and all higher numbered LP1+ watermarks as dirty */
        for (; wm_lp <= 3; wm_lp++)
                dirty |= WM_DIRTY_LP(wm_lp);

        return dirty;
}

static bool _ilk_disable_lp_wm(struct drm_i915_private *dev_priv,
                               unsigned int dirty)
{
        struct ilk_wm_values *previous = &dev_priv->wm.hw;
        bool changed = false;

        if (dirty & WM_DIRTY_LP(3) && previous->wm_lp[2] & WM1_LP_SR_EN) {
                previous->wm_lp[2] &= ~WM1_LP_SR_EN;
                I915_WRITE(WM3_LP_ILK, previous->wm_lp[2]);
                changed = true;
        }
        if (dirty & WM_DIRTY_LP(2) && previous->wm_lp[1] & WM1_LP_SR_EN) {
                previous->wm_lp[1] &= ~WM1_LP_SR_EN;
                I915_WRITE(WM2_LP_ILK, previous->wm_lp[1]);
                changed = true;
        }
        if (dirty & WM_DIRTY_LP(1) && previous->wm_lp[0] & WM1_LP_SR_EN) {
                previous->wm_lp[0] &= ~WM1_LP_SR_EN;
                I915_WRITE(WM1_LP_ILK, previous->wm_lp[0]);
                changed = true;
        }

        /*
         * Don't touch WM1S_LP_EN here.
         * Doing so could cause underruns.
         */

        return changed;
}

/*
 * The spec says we shouldn't write when we don't need, because every write
 * causes WMs to be re-evaluated, expending some power.
 */
static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
                                struct ilk_wm_values *results)
{
        struct drm_device *dev = dev_priv->dev;
        struct ilk_wm_values *previous = &dev_priv->wm.hw;
        unsigned int dirty;
        uint32_t val;

        dirty = ilk_compute_wm_dirty(dev, previous, results);
        if (!dirty)
                return;

        _ilk_disable_lp_wm(dev_priv, dirty);

        if (dirty & WM_DIRTY_PIPE(PIPE_A))
                I915_WRITE(WM0_PIPEA_ILK, results->wm_pipe[0]);
        if (dirty & WM_DIRTY_PIPE(PIPE_B))
                I915_WRITE(WM0_PIPEB_ILK, results->wm_pipe[1]);
        if (dirty & WM_DIRTY_PIPE(PIPE_C))
                I915_WRITE(WM0_PIPEC_IVB, results->wm_pipe[2]);

        if (dirty & WM_DIRTY_LINETIME(PIPE_A))
                I915_WRITE(PIPE_WM_LINETIME(PIPE_A), results->wm_linetime[0]);
        if (dirty & WM_DIRTY_LINETIME(PIPE_B))
                I915_WRITE(PIPE_WM_LINETIME(PIPE_B), results->wm_linetime[1]);
        if (dirty & WM_DIRTY_LINETIME(PIPE_C))
                I915_WRITE(PIPE_WM_LINETIME(PIPE_C), results->wm_linetime[2]);

        if (dirty & WM_DIRTY_DDB) {
                if (IS_HASWELL(dev) || IS_BROADWELL(dev)) {
                        val = I915_READ(WM_MISC);
                        if (results->partitioning == INTEL_DDB_PART_1_2)
                                val &= ~WM_MISC_DATA_PARTITION_5_6;
                        else
                                val |= WM_MISC_DATA_PARTITION_5_6;
                        I915_WRITE(WM_MISC, val);
                } else {
                        val = I915_READ(DISP_ARB_CTL2);
                        if (results->partitioning == INTEL_DDB_PART_1_2)
                                val &= ~DISP_DATA_PARTITION_5_6;
                        else
                                val |= DISP_DATA_PARTITION_5_6;
                        I915_WRITE(DISP_ARB_CTL2, val);
                }
        }

        if (dirty & WM_DIRTY_FBC) {
                val = I915_READ(DISP_ARB_CTL);
                if (results->enable_fbc_wm)
                        val &= ~DISP_FBC_WM_DIS;
                else
                        val |= DISP_FBC_WM_DIS;
                I915_WRITE(DISP_ARB_CTL, val);
        }

        if (dirty & WM_DIRTY_LP(1) &&
            previous->wm_lp_spr[0] != results->wm_lp_spr[0])
                I915_WRITE(WM1S_LP_ILK, results->wm_lp_spr[0]);

        if (INTEL_INFO(dev)->gen >= 7) {
                if (dirty & WM_DIRTY_LP(2) && previous->wm_lp_spr[1] != results->wm_lp_spr[1])
                        I915_WRITE(WM2S_LP_IVB, results->wm_lp_spr[1]);
                if (dirty & WM_DIRTY_LP(3) && previous->wm_lp_spr[2] != results->wm_lp_spr[2])
                        I915_WRITE(WM3S_LP_IVB, results->wm_lp_spr[2]);
        }

        if (dirty & WM_DIRTY_LP(1) && previous->wm_lp[0] != results->wm_lp[0])
                I915_WRITE(WM1_LP_ILK, results->wm_lp[0]);
        if (dirty & WM_DIRTY_LP(2) && previous->wm_lp[1] != results->wm_lp[1])
                I915_WRITE(WM2_LP_ILK, results->wm_lp[1]);
        if (dirty & WM_DIRTY_LP(3) && previous->wm_lp[2] != results->wm_lp[2])
                I915_WRITE(WM3_LP_ILK, results->wm_lp[2]);

        dev_priv->wm.hw = *results;
}

static bool ilk_disable_lp_wm(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;

        return _ilk_disable_lp_wm(dev_priv, WM_DIRTY_LP_ALL);
}

static void ilk_update_wm(struct drm_crtc *crtc)
{
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        struct drm_device *dev = crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct ilk_wm_maximums max;
        struct ilk_pipe_wm_parameters params = {};
        struct ilk_wm_values results = {};
        enum intel_ddb_partitioning partitioning;
        struct intel_pipe_wm pipe_wm = {};
        struct intel_pipe_wm lp_wm_1_2 = {}, lp_wm_5_6 = {}, *best_lp_wm;
        struct intel_wm_config config = {};

        ilk_compute_wm_parameters(crtc, &params);

        intel_compute_pipe_wm(crtc, &params, &pipe_wm);

        if (!memcmp(&intel_crtc->wm.active, &pipe_wm, sizeof(pipe_wm)))
                return;

        intel_crtc->wm.active = pipe_wm;

        ilk_compute_wm_config(dev, &config);

        ilk_compute_wm_maximums(dev, 1, &config, INTEL_DDB_PART_1_2, &max);
        ilk_wm_merge(dev, &config, &max, &lp_wm_1_2);

        /* 5/6 split only in single pipe config on IVB+ */
        if (INTEL_INFO(dev)->gen >= 7 &&
            config.num_pipes_active == 1 && config.sprites_enabled) {
                ilk_compute_wm_maximums(dev, 1, &config, INTEL_DDB_PART_5_6, &max);
                ilk_wm_merge(dev, &config, &max, &lp_wm_5_6);

                best_lp_wm = ilk_find_best_result(dev, &lp_wm_1_2, &lp_wm_5_6);
        } else {
                best_lp_wm = &lp_wm_1_2;
        }

        partitioning = (best_lp_wm == &lp_wm_1_2) ?
                       INTEL_DDB_PART_1_2 : INTEL_DDB_PART_5_6;

        ilk_compute_wm_results(dev, best_lp_wm, partitioning, &results);

        ilk_write_wm_values(dev_priv, &results);
}

static void
ilk_update_sprite_wm(struct drm_plane *plane,
                     struct drm_crtc *crtc,
                     uint32_t sprite_width, uint32_t sprite_height,
                     int pixel_size, bool enabled, bool scaled)
{
        struct drm_device *dev = plane->dev;
        struct intel_plane *intel_plane = to_intel_plane(plane);

        intel_plane->wm.enabled = enabled;
        intel_plane->wm.scaled = scaled;
        intel_plane->wm.horiz_pixels = sprite_width;
        intel_plane->wm.vert_pixels = sprite_width;
        intel_plane->wm.bytes_per_pixel = pixel_size;

        /*
         * IVB workaround: must disable low power watermarks for at least
         * one frame before enabling scaling.  LP watermarks can be re-enabled
         * when scaling is disabled.
         *
         * WaCxSRDisabledForSpriteScaling:ivb
         */
        if (IS_IVYBRIDGE(dev) && scaled && ilk_disable_lp_wm(dev))
                intel_wait_for_vblank(dev, intel_plane->pipe);

        ilk_update_wm(crtc);
}

static void ilk_pipe_wm_get_hw_state(struct drm_crtc *crtc)
{
        struct drm_device *dev = crtc->dev;
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct ilk_wm_values *hw = &dev_priv->wm.hw;
        struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
        struct intel_pipe_wm *active = &intel_crtc->wm.active;
        enum pipe pipe = intel_crtc->pipe;
        static const unsigned int wm0_pipe_reg[] = {
                [PIPE_A] = WM0_PIPEA_ILK,
                [PIPE_B] = WM0_PIPEB_ILK,
                [PIPE_C] = WM0_PIPEC_IVB,
        };

        hw->wm_pipe[pipe] = I915_READ(wm0_pipe_reg[pipe]);
        if (IS_HASWELL(dev) || IS_BROADWELL(dev))
                hw->wm_linetime[pipe] = I915_READ(PIPE_WM_LINETIME(pipe));

        active->pipe_enabled = intel_crtc_active(crtc);

        if (active->pipe_enabled) {
                u32 tmp = hw->wm_pipe[pipe];

                /*
                 * For active pipes LP0 watermark is marked as
                 * enabled, and LP1+ watermaks as disabled since
                 * we can't really reverse compute them in case
                 * multiple pipes are active.
                 */
                active->wm[0].enable = true;
                active->wm[0].pri_val = (tmp & WM0_PIPE_PLANE_MASK) >> WM0_PIPE_PLANE_SHIFT;
                active->wm[0].spr_val = (tmp & WM0_PIPE_SPRITE_MASK) >> WM0_PIPE_SPRITE_SHIFT;
                active->wm[0].cur_val = tmp & WM0_PIPE_CURSOR_MASK;
                active->linetime = hw->wm_linetime[pipe];
        } else {
                int level, max_level = ilk_wm_max_level(dev);

                /*
                 * For inactive pipes, all watermark levels
                 * should be marked as enabled but zeroed,
                 * which is what we'd compute them to.
                 */
                for (level = 0; level <= max_level; level++)
                        active->wm[level].enable = true;
        }
}

void ilk_wm_get_hw_state(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        struct ilk_wm_values *hw = &dev_priv->wm.hw;
        struct drm_crtc *crtc;

        for_each_crtc(dev, crtc)
                ilk_pipe_wm_get_hw_state(crtc);

        hw->wm_lp[0] = I915_READ(WM1_LP_ILK);
        hw->wm_lp[1] = I915_READ(WM2_LP_ILK);
        hw->wm_lp[2] = I915_READ(WM3_LP_ILK);

        hw->wm_lp_spr[0] = I915_READ(WM1S_LP_ILK);
        if (INTEL_INFO(dev)->gen >= 7) {
                hw->wm_lp_spr[1] = I915_READ(WM2S_LP_IVB);
                hw->wm_lp_spr[2] = I915_READ(WM3S_LP_IVB);
        }

        if (IS_HASWELL(dev) || IS_BROADWELL(dev))
                hw->partitioning = (I915_READ(WM_MISC) & WM_MISC_DATA_PARTITION_5_6) ?
                        INTEL_DDB_PART_5_6 : INTEL_DDB_PART_1_2;
        else if (IS_IVYBRIDGE(dev))
                hw->partitioning = (I915_READ(DISP_ARB_CTL2) & DISP_DATA_PARTITION_5_6) ?
                        INTEL_DDB_PART_5_6 : INTEL_DDB_PART_1_2;

        hw->enable_fbc_wm =
                !(I915_READ(DISP_ARB_CTL) & DISP_FBC_WM_DIS);
}

/**
 * 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
 *     surface width = hdisplay for normal plane and 64 for cursor
 *   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.
 */
void intel_update_watermarks(struct drm_crtc *crtc)
{
        struct drm_i915_private *dev_priv = crtc->dev->dev_private;

        if (dev_priv->display.update_wm)
                dev_priv->display.update_wm(crtc);
}

void intel_update_sprite_watermarks(struct drm_plane *plane,
                                    struct drm_crtc *crtc,
                                    uint32_t sprite_width,
                                    uint32_t sprite_height,
                                    int pixel_size,
                                    bool enabled, bool scaled)
{
        struct drm_i915_private *dev_priv = plane->dev->dev_private;

        if (dev_priv->display.update_sprite_wm)
                dev_priv->display.update_sprite_wm(plane, crtc,
                                                   sprite_width, sprite_height,
                                                   pixel_size, enabled, scaled);
}

static struct drm_i915_gem_object *
intel_alloc_context_page(struct drm_device *dev)
{
        struct drm_i915_gem_object *ctx;
        int ret;

        WARN_ON(!mutex_is_locked(&dev->struct_mutex));

        ctx = i915_gem_alloc_object(dev, 4096);
        if (!ctx) {
                DRM_DEBUG("failed to alloc power context, RC6 disabled\n");
                return NULL;
        }

        ret = i915_gem_obj_ggtt_pin(ctx, 4096, 0);
        if (ret) {
                DRM_ERROR("failed to pin power context: %d\n", ret);
                goto err_unref;
        }

        ret = i915_gem_object_set_to_gtt_domain(ctx, 1);
        if (ret) {
                DRM_ERROR("failed to set-domain on power context: %d\n", ret);
                goto err_unpin;
        }

        return ctx;

err_unpin:
        i915_gem_object_ggtt_unpin(ctx);
err_unref:
        drm_gem_object_unreference(&ctx->base);
        return NULL;
}

/**
 * Lock protecting IPS related data structures
 */
DEFINE_SPINLOCK(mchdev_lock);

/* Global for IPS driver to get at the current i915 device. Protected by
 * mchdev_lock. */
static struct drm_i915_private *i915_mch_dev;

bool ironlake_set_drps(struct drm_device *dev, u8 val)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u16 rgvswctl;

        assert_spin_locked(&mchdev_lock);

        rgvswctl = I915_READ16(MEMSWCTL);
        if (rgvswctl & MEMCTL_CMD_STS) {
                DRM_DEBUG("gpu busy, RCS change rejected\n");
                return false; /* still busy with another command */
        }

        rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) |
                (val << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM;
        I915_WRITE16(MEMSWCTL, rgvswctl);
        POSTING_READ16(MEMSWCTL);

        rgvswctl |= MEMCTL_CMD_STS;
        I915_WRITE16(MEMSWCTL, rgvswctl);

        return true;
}

static void ironlake_enable_drps(struct drm_device *dev)
{
        struct drm_i915_private *dev_priv = dev->dev_private;
        u32 rgvmodectl = I915_READ(MEMMODECTL);
        u8 fmax, fmin, fstart, vstart;

        spin_lock_irq(&mchdev_lock);

        /* Enable temp reporting */
        I915_WRITE16(PMMISC, I915_READ(PMMISC) | MCPPCE_EN);
        I915_WRITE16(TSC1, I915_READ(TSC1) | TSE);

        /* 100ms RC evaluation intervals */
        I915_WRITE(RCUPEI, 100000);
        I915_WRITE(RCDNEI, 100000);

        /* Set max/min thresholds to 90ms and 80ms respectively */
        I915_WRITE(RCBMAXAVG, 90000);
        I915_WRITE(RCBMINAVG, 80000);

        I915_WRITE(MEMIHYST, 1);

        /* Set up min, max, and cur for interrupt handling */
        fmax = (rgvmodectl & MEMMODE_FMAX_MASK) >> MEMMODE_FMAX_SHIFT;
        fmin = (rgvmodectl & MEMMODE_FMIN_MASK);
        fstart = (rgvmodectl & MEMMODE_FSTART_MASK) >>
                MEMMODE_FSTART_SHIFT;