linux/drivers/gpu/drm/i915/intel_overlay.c

1598 lines
40 KiB
C
Raw Normal View History

/*
* Copyright © 2009
*
* 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:
* Daniel Vetter <daniel@ffwll.ch>
*
* Derived from Xorg ddx, xf86-video-intel, src/i830_video.c
*/
#include <drm/drmP.h>
#include <drm/i915_drm.h>
#include "i915_drv.h"
#include "i915_reg.h"
#include "intel_drv.h"
/* Limits for overlay size. According to intel doc, the real limits are:
* Y width: 4095, UV width (planar): 2047, Y height: 2047,
* UV width (planar): * 1023. But the xorg thinks 2048 for height and width. Use
* the mininum of both. */
#define IMAGE_MAX_WIDTH 2048
#define IMAGE_MAX_HEIGHT 2046 /* 2 * 1023 */
/* on 830 and 845 these large limits result in the card hanging */
#define IMAGE_MAX_WIDTH_LEGACY 1024
#define IMAGE_MAX_HEIGHT_LEGACY 1088
/* overlay register definitions */
/* OCMD register */
#define OCMD_TILED_SURFACE (0x1<<19)
#define OCMD_MIRROR_MASK (0x3<<17)
#define OCMD_MIRROR_MODE (0x3<<17)
#define OCMD_MIRROR_HORIZONTAL (0x1<<17)
#define OCMD_MIRROR_VERTICAL (0x2<<17)
#define OCMD_MIRROR_BOTH (0x3<<17)
#define OCMD_BYTEORDER_MASK (0x3<<14) /* zero for YUYV or FOURCC YUY2 */
#define OCMD_UV_SWAP (0x1<<14) /* YVYU */
#define OCMD_Y_SWAP (0x2<<14) /* UYVY or FOURCC UYVY */
#define OCMD_Y_AND_UV_SWAP (0x3<<14) /* VYUY */
#define OCMD_SOURCE_FORMAT_MASK (0xf<<10)
#define OCMD_RGB_888 (0x1<<10) /* not in i965 Intel docs */
#define OCMD_RGB_555 (0x2<<10) /* not in i965 Intel docs */
#define OCMD_RGB_565 (0x3<<10) /* not in i965 Intel docs */
#define OCMD_YUV_422_PACKED (0x8<<10)
#define OCMD_YUV_411_PACKED (0x9<<10) /* not in i965 Intel docs */
#define OCMD_YUV_420_PLANAR (0xc<<10)
#define OCMD_YUV_422_PLANAR (0xd<<10)
#define OCMD_YUV_410_PLANAR (0xe<<10) /* also 411 */
#define OCMD_TVSYNCFLIP_PARITY (0x1<<9)
#define OCMD_TVSYNCFLIP_ENABLE (0x1<<7)
#define OCMD_BUF_TYPE_MASK (0x1<<5)
#define OCMD_BUF_TYPE_FRAME (0x0<<5)
#define OCMD_BUF_TYPE_FIELD (0x1<<5)
#define OCMD_TEST_MODE (0x1<<4)
#define OCMD_BUFFER_SELECT (0x3<<2)
#define OCMD_BUFFER0 (0x0<<2)
#define OCMD_BUFFER1 (0x1<<2)
#define OCMD_FIELD_SELECT (0x1<<2)
#define OCMD_FIELD0 (0x0<<1)
#define OCMD_FIELD1 (0x1<<1)
#define OCMD_ENABLE (0x1<<0)
/* OCONFIG register */
#define OCONF_PIPE_MASK (0x1<<18)
#define OCONF_PIPE_A (0x0<<18)
#define OCONF_PIPE_B (0x1<<18)
#define OCONF_GAMMA2_ENABLE (0x1<<16)
#define OCONF_CSC_MODE_BT601 (0x0<<5)
#define OCONF_CSC_MODE_BT709 (0x1<<5)
#define OCONF_CSC_BYPASS (0x1<<4)
#define OCONF_CC_OUT_8BIT (0x1<<3)
#define OCONF_TEST_MODE (0x1<<2)
#define OCONF_THREE_LINE_BUFFER (0x1<<0)
#define OCONF_TWO_LINE_BUFFER (0x0<<0)
/* DCLRKM (dst-key) register */
#define DST_KEY_ENABLE (0x1<<31)
#define CLK_RGB24_MASK 0x0
#define CLK_RGB16_MASK 0x070307
#define CLK_RGB15_MASK 0x070707
#define CLK_RGB8I_MASK 0xffffff
#define RGB16_TO_COLORKEY(c) \
(((c & 0xF800) << 8) | ((c & 0x07E0) << 5) | ((c & 0x001F) << 3))
#define RGB15_TO_COLORKEY(c) \
(((c & 0x7c00) << 9) | ((c & 0x03E0) << 6) | ((c & 0x001F) << 3))
/* overlay flip addr flag */
#define OFC_UPDATE 0x1
/* polyphase filter coefficients */
#define N_HORIZ_Y_TAPS 5
#define N_VERT_Y_TAPS 3
#define N_HORIZ_UV_TAPS 3
#define N_VERT_UV_TAPS 3
#define N_PHASES 17
#define MAX_TAPS 5
/* memory bufferd overlay registers */
struct overlay_registers {
u32 OBUF_0Y;
u32 OBUF_1Y;
u32 OBUF_0U;
u32 OBUF_0V;
u32 OBUF_1U;
u32 OBUF_1V;
u32 OSTRIDE;
u32 YRGB_VPH;
u32 UV_VPH;
u32 HORZ_PH;
u32 INIT_PHS;
u32 DWINPOS;
u32 DWINSZ;
u32 SWIDTH;
u32 SWIDTHSW;
u32 SHEIGHT;
u32 YRGBSCALE;
u32 UVSCALE;
u32 OCLRC0;
u32 OCLRC1;
u32 DCLRKV;
u32 DCLRKM;
u32 SCLRKVH;
u32 SCLRKVL;
u32 SCLRKEN;
u32 OCONFIG;
u32 OCMD;
u32 RESERVED1; /* 0x6C */
u32 OSTART_0Y;
u32 OSTART_1Y;
u32 OSTART_0U;
u32 OSTART_0V;
u32 OSTART_1U;
u32 OSTART_1V;
u32 OTILEOFF_0Y;
u32 OTILEOFF_1Y;
u32 OTILEOFF_0U;
u32 OTILEOFF_0V;
u32 OTILEOFF_1U;
u32 OTILEOFF_1V;
u32 FASTHSCALE; /* 0xA0 */
u32 UVSCALEV; /* 0xA4 */
u32 RESERVEDC[(0x200 - 0xA8) / 4]; /* 0xA8 - 0x1FC */
u16 Y_VCOEFS[N_VERT_Y_TAPS * N_PHASES]; /* 0x200 */
u16 RESERVEDD[0x100 / 2 - N_VERT_Y_TAPS * N_PHASES];
u16 Y_HCOEFS[N_HORIZ_Y_TAPS * N_PHASES]; /* 0x300 */
u16 RESERVEDE[0x200 / 2 - N_HORIZ_Y_TAPS * N_PHASES];
u16 UV_VCOEFS[N_VERT_UV_TAPS * N_PHASES]; /* 0x500 */
u16 RESERVEDF[0x100 / 2 - N_VERT_UV_TAPS * N_PHASES];
u16 UV_HCOEFS[N_HORIZ_UV_TAPS * N_PHASES]; /* 0x600 */
u16 RESERVEDG[0x100 / 2 - N_HORIZ_UV_TAPS * N_PHASES];
};
struct intel_overlay {
struct drm_device *dev;
struct intel_crtc *crtc;
struct drm_i915_gem_object *vid_bo;
struct drm_i915_gem_object *old_vid_bo;
bool active;
bool pfit_active;
u32 pfit_vscale_ratio; /* shifted-point number, (1<<12) == 1.0 */
u32 color_key:24;
u32 color_key_enabled:1;
u32 brightness, contrast, saturation;
u32 old_xscale, old_yscale;
/* register access */
u32 flip_addr;
struct drm_i915_gem_object *reg_bo;
/* flip handling */
struct drm_i915_gem_request *last_flip_req;
void (*flip_tail)(struct intel_overlay *);
};
static struct overlay_registers __iomem *
intel_overlay_map_regs(struct intel_overlay *overlay)
{
struct drm_i915_private *dev_priv = to_i915(overlay->dev);
struct i915_ggtt *ggtt = &dev_priv->ggtt;
struct overlay_registers __iomem *regs;
if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
regs = (struct overlay_registers __iomem *)overlay->reg_bo->phys_handle->vaddr;
else
regs = io_mapping_map_wc(ggtt->mappable,
i915_gem_obj_ggtt_offset(overlay->reg_bo));
return regs;
}
static void intel_overlay_unmap_regs(struct intel_overlay *overlay,
struct overlay_registers __iomem *regs)
{
if (!OVERLAY_NEEDS_PHYSICAL(overlay->dev))
io_mapping_unmap(regs);
}
static int intel_overlay_do_wait_request(struct intel_overlay *overlay,
struct drm_i915_gem_request *req,
void (*tail)(struct intel_overlay *))
{
int ret;
WARN_ON(overlay->last_flip_req);
i915_gem_request_assign(&overlay->last_flip_req, req);
i915_add_request(req);
overlay->flip_tail = tail;
ret = i915_wait_request(overlay->last_flip_req);
if (ret)
return ret;
i915_gem_request_assign(&overlay->last_flip_req, NULL);
return 0;
}
/* overlay needs to be disable in OCMD reg */
static int intel_overlay_on(struct intel_overlay *overlay)
{
struct drm_device *dev = overlay->dev;
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_engine_cs *engine = &dev_priv->engine[RCS];
struct drm_i915_gem_request *req;
int ret;
WARN_ON(overlay->active);
WARN_ON(IS_I830(dev) && !(dev_priv->quirks & QUIRK_PIPEA_FORCE));
req = i915_gem_request_alloc(engine, NULL);
drm/i915: simplify allocation of driver-internal requests There are a number of places where the driver needs a request, but isn't working on behalf of any specific user or in a specific context. At present, we associate them with the per-engine default context. A future patch will abolish those per-engine context pointers; but we can already eliminate a lot of the references to them, just by making the allocator allow NULL as a shorthand for "an appropriate context for this ring", which will mean that the callers don't need to know anything about how the "appropriate context" is found (e.g. per-ring vs per-device, etc). So this patch renames the existing i915_gem_request_alloc(), and makes it local (static inline), and replaces it with a wrapper that provides a default if the context is NULL, and also has a nicer calling convention (doesn't require a pointer to an output parameter). Then we change all callers to use the new convention: OLD: err = i915_gem_request_alloc(ring, user_ctx, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, user_ctx); if (IS_ERR(req)) ... OLD: err = i915_gem_request_alloc(ring, ring->default_context, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, NULL); if (IS_ERR(req)) ... v4: Rebased Signed-off-by: Dave Gordon <david.s.gordon@intel.com> Reviewed-by: Nick Hoath <nicholas.hoath@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1453230175-19330-2-git-send-email-david.s.gordon@intel.com Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2016-01-20 03:02:53 +08:00
if (IS_ERR(req))
return PTR_ERR(req);
ret = intel_ring_begin(req, 4);
if (ret) {
drm/i915: Late request cancellations are harmful Conceptually, each request is a record of a hardware transaction - we build up a list of pending commands and then either commit them to hardware, or cancel them. However, whilst building up the list of pending commands, we may modify state outside of the request and make references to the pending request. If we do so and then cancel that request, external objects then point to the deleted request leading to both graphical and memory corruption. The easiest example is to consider object/VMA tracking. When we mark an object as active in a request, we store a pointer to this, the most recent request, in the object. Then we want to free that object, we wait for the most recent request to be idle before proceeding (otherwise the hardware will write to pages now owned by the system, or we will attempt to read from those pages before the hardware is finished writing). If the request was cancelled instead, that wait completes immediately. As a result, all requests must be committed and not cancelled if the external state is unknown. All that remains of i915_gem_request_cancel() users are just a couple of extremely unlikely allocation failures, so remove the API entirely. A consequence of committing all incomplete requests is that we generate excess breadcrumbs and fill the ring much more often with dummy work. We have completely undone the outstanding_last_seqno optimisation. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: stable@vger.kernel.org Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1460565315-7748-16-git-send-email-chris@chris-wilson.co.uk
2016-04-14 00:35:15 +08:00
i915_add_request_no_flush(req);
return ret;
}
overlay->active = true;
intel_ring_emit(engine, MI_OVERLAY_FLIP | MI_OVERLAY_ON);
intel_ring_emit(engine, overlay->flip_addr | OFC_UPDATE);
intel_ring_emit(engine, MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
intel_ring_emit(engine, MI_NOOP);
intel_ring_advance(engine);
return intel_overlay_do_wait_request(overlay, req, NULL);
}
/* overlay needs to be enabled in OCMD reg */
static int intel_overlay_continue(struct intel_overlay *overlay,
bool load_polyphase_filter)
{
struct drm_device *dev = overlay->dev;
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_engine_cs *engine = &dev_priv->engine[RCS];
struct drm_i915_gem_request *req;
u32 flip_addr = overlay->flip_addr;
u32 tmp;
int ret;
WARN_ON(!overlay->active);
if (load_polyphase_filter)
flip_addr |= OFC_UPDATE;
/* check for underruns */
tmp = I915_READ(DOVSTA);
if (tmp & (1 << 17))
DRM_DEBUG("overlay underrun, DOVSTA: %x\n", tmp);
req = i915_gem_request_alloc(engine, NULL);
drm/i915: simplify allocation of driver-internal requests There are a number of places where the driver needs a request, but isn't working on behalf of any specific user or in a specific context. At present, we associate them with the per-engine default context. A future patch will abolish those per-engine context pointers; but we can already eliminate a lot of the references to them, just by making the allocator allow NULL as a shorthand for "an appropriate context for this ring", which will mean that the callers don't need to know anything about how the "appropriate context" is found (e.g. per-ring vs per-device, etc). So this patch renames the existing i915_gem_request_alloc(), and makes it local (static inline), and replaces it with a wrapper that provides a default if the context is NULL, and also has a nicer calling convention (doesn't require a pointer to an output parameter). Then we change all callers to use the new convention: OLD: err = i915_gem_request_alloc(ring, user_ctx, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, user_ctx); if (IS_ERR(req)) ... OLD: err = i915_gem_request_alloc(ring, ring->default_context, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, NULL); if (IS_ERR(req)) ... v4: Rebased Signed-off-by: Dave Gordon <david.s.gordon@intel.com> Reviewed-by: Nick Hoath <nicholas.hoath@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1453230175-19330-2-git-send-email-david.s.gordon@intel.com Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2016-01-20 03:02:53 +08:00
if (IS_ERR(req))
return PTR_ERR(req);
ret = intel_ring_begin(req, 2);
if (ret) {
drm/i915: Late request cancellations are harmful Conceptually, each request is a record of a hardware transaction - we build up a list of pending commands and then either commit them to hardware, or cancel them. However, whilst building up the list of pending commands, we may modify state outside of the request and make references to the pending request. If we do so and then cancel that request, external objects then point to the deleted request leading to both graphical and memory corruption. The easiest example is to consider object/VMA tracking. When we mark an object as active in a request, we store a pointer to this, the most recent request, in the object. Then we want to free that object, we wait for the most recent request to be idle before proceeding (otherwise the hardware will write to pages now owned by the system, or we will attempt to read from those pages before the hardware is finished writing). If the request was cancelled instead, that wait completes immediately. As a result, all requests must be committed and not cancelled if the external state is unknown. All that remains of i915_gem_request_cancel() users are just a couple of extremely unlikely allocation failures, so remove the API entirely. A consequence of committing all incomplete requests is that we generate excess breadcrumbs and fill the ring much more often with dummy work. We have completely undone the outstanding_last_seqno optimisation. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: stable@vger.kernel.org Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1460565315-7748-16-git-send-email-chris@chris-wilson.co.uk
2016-04-14 00:35:15 +08:00
i915_add_request_no_flush(req);
return ret;
}
intel_ring_emit(engine, MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE);
intel_ring_emit(engine, flip_addr);
intel_ring_advance(engine);
WARN_ON(overlay->last_flip_req);
i915_gem_request_assign(&overlay->last_flip_req, req);
i915_add_request(req);
drm/i915: i915_add_request must not fail The i915_add_request() function is called to keep track of work that has been written to the ring buffer. It adds epilogue commands to track progress (seqno updates and such), moves the request structure onto the right list and other such house keeping tasks. However, the work itself has already been written to the ring and will get executed whether or not the add request call succeeds. So no matter what goes wrong, there isn't a whole lot of point in failing the call. At the moment, this is fine(ish). If the add request does bail early on and not do the housekeeping, the request will still float around in the ring->outstanding_lazy_request field and be picked up next time. It means multiple pieces of work will be tagged as the same request and driver can't actually wait for the first piece of work until something else has been submitted. But it all sort of hangs together. This patch series is all about removing the OLR and guaranteeing that each piece of work gets its own personal request. That means that there is no more 'hoovering up of forgotten requests'. If the request does not get tracked then it will be leaked. Thus the add request call _must_ not fail. The previous patch should have already ensured that it _will_ not fail by removing the potential for running out of ring space. This patch enforces the rule by actually removing the early exit paths and the return code. Note that if something does manage to fail and the epilogue commands don't get written to the ring, the driver will still hang together. The request will be added to the tracking lists. And as in the old case, any subsequent work will generate a new seqno which will suffice for marking the old one as complete. v2: Improved WARNings (Tomas Elf review request). For: VIZ-5115 Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Reviewed-by: Tomas Elf <tomas.elf@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-05-30 00:43:24 +08:00
return 0;
}
static void intel_overlay_release_old_vid_tail(struct intel_overlay *overlay)
{
struct drm_i915_gem_object *obj = overlay->old_vid_bo;
i915_gem_object_ggtt_unpin(obj);
drm_gem_object_unreference(&obj->base);
overlay->old_vid_bo = NULL;
}
static void intel_overlay_off_tail(struct intel_overlay *overlay)
{
struct drm_i915_gem_object *obj = overlay->vid_bo;
/* never have the overlay hw on without showing a frame */
if (WARN_ON(!obj))
return;
i915_gem_object_ggtt_unpin(obj);
drm_gem_object_unreference(&obj->base);
overlay->vid_bo = NULL;
overlay->crtc->overlay = NULL;
overlay->crtc = NULL;
overlay->active = false;
}
/* overlay needs to be disabled in OCMD reg */
static int intel_overlay_off(struct intel_overlay *overlay)
{
struct drm_device *dev = overlay->dev;
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_engine_cs *engine = &dev_priv->engine[RCS];
struct drm_i915_gem_request *req;
u32 flip_addr = overlay->flip_addr;
int ret;
WARN_ON(!overlay->active);
/* According to intel docs the overlay hw may hang (when switching
* off) without loading the filter coeffs. It is however unclear whether
* this applies to the disabling of the overlay or to the switching off
* of the hw. Do it in both cases */
flip_addr |= OFC_UPDATE;
req = i915_gem_request_alloc(engine, NULL);
drm/i915: simplify allocation of driver-internal requests There are a number of places where the driver needs a request, but isn't working on behalf of any specific user or in a specific context. At present, we associate them with the per-engine default context. A future patch will abolish those per-engine context pointers; but we can already eliminate a lot of the references to them, just by making the allocator allow NULL as a shorthand for "an appropriate context for this ring", which will mean that the callers don't need to know anything about how the "appropriate context" is found (e.g. per-ring vs per-device, etc). So this patch renames the existing i915_gem_request_alloc(), and makes it local (static inline), and replaces it with a wrapper that provides a default if the context is NULL, and also has a nicer calling convention (doesn't require a pointer to an output parameter). Then we change all callers to use the new convention: OLD: err = i915_gem_request_alloc(ring, user_ctx, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, user_ctx); if (IS_ERR(req)) ... OLD: err = i915_gem_request_alloc(ring, ring->default_context, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, NULL); if (IS_ERR(req)) ... v4: Rebased Signed-off-by: Dave Gordon <david.s.gordon@intel.com> Reviewed-by: Nick Hoath <nicholas.hoath@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1453230175-19330-2-git-send-email-david.s.gordon@intel.com Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2016-01-20 03:02:53 +08:00
if (IS_ERR(req))
return PTR_ERR(req);
ret = intel_ring_begin(req, 6);
if (ret) {
drm/i915: Late request cancellations are harmful Conceptually, each request is a record of a hardware transaction - we build up a list of pending commands and then either commit them to hardware, or cancel them. However, whilst building up the list of pending commands, we may modify state outside of the request and make references to the pending request. If we do so and then cancel that request, external objects then point to the deleted request leading to both graphical and memory corruption. The easiest example is to consider object/VMA tracking. When we mark an object as active in a request, we store a pointer to this, the most recent request, in the object. Then we want to free that object, we wait for the most recent request to be idle before proceeding (otherwise the hardware will write to pages now owned by the system, or we will attempt to read from those pages before the hardware is finished writing). If the request was cancelled instead, that wait completes immediately. As a result, all requests must be committed and not cancelled if the external state is unknown. All that remains of i915_gem_request_cancel() users are just a couple of extremely unlikely allocation failures, so remove the API entirely. A consequence of committing all incomplete requests is that we generate excess breadcrumbs and fill the ring much more often with dummy work. We have completely undone the outstanding_last_seqno optimisation. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: stable@vger.kernel.org Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1460565315-7748-16-git-send-email-chris@chris-wilson.co.uk
2016-04-14 00:35:15 +08:00
i915_add_request_no_flush(req);
return ret;
}
/* wait for overlay to go idle */
intel_ring_emit(engine, MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE);
intel_ring_emit(engine, flip_addr);
intel_ring_emit(engine, MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
/* turn overlay off */
drm/i915: fix overlay on i830M The overlay on the i830M has a peculiar failure mode: It works the first time around after boot-up, but consistenly hangs the second time it's used. Chris Wilson has dug out a nice errata: "1.5.12 Clock Gating Disable for Display Register Address Offset: 06200h–06203h "Bit 3 Ovrunit Clock Gating Disable. 0 = Clock gating controlled by unit enabling logic 1 = Disable clock gating function DevALM Errata ALM049: Overlay Clock Gating Must be Disabled: Overlay & L2 Cache clock gating must be disabled in order to prevent device hangs when turning off overlay.SW must turn off Ovrunit clock gating (6200h) and L2 Cache clock gating (C8h)." Now I've nowhere found that 0xc8 register and hence couldn't apply the l2 cache workaround. But I've remembered that part of the magic that the OVERLAY_ON/OFF commands are supposed to do is to rearrange cache allocations so that the overlay scaler has some scratch space. And while pondering how that could explain the hang the 2nd time we enable the overlay, I've remembered that the old ums overlay code did _not_ issue the OVERLAY_OFF cmd. And indeed, disabling the OFF cmd results in the overlay working flawlessly, so I guess we can workaround the lack of the above workaround by simply never disabling the overlay engine once it's enabled. Note that we have the first part of the above w/a already implemented in i830_init_clock_gating - leave that as-is to avoid surprises. v2: Add a comment in the code. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47827 Cc: stable@vger.kernel.org Tested-by: Rhys <rhyspuk@gmail.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-10-22 18:55:55 +08:00
if (IS_I830(dev)) {
/* Workaround: Don't disable the overlay fully, since otherwise
* it dies on the next OVERLAY_ON cmd. */
intel_ring_emit(engine, MI_NOOP);
intel_ring_emit(engine, MI_NOOP);
intel_ring_emit(engine, MI_NOOP);
drm/i915: fix overlay on i830M The overlay on the i830M has a peculiar failure mode: It works the first time around after boot-up, but consistenly hangs the second time it's used. Chris Wilson has dug out a nice errata: "1.5.12 Clock Gating Disable for Display Register Address Offset: 06200h–06203h "Bit 3 Ovrunit Clock Gating Disable. 0 = Clock gating controlled by unit enabling logic 1 = Disable clock gating function DevALM Errata ALM049: Overlay Clock Gating Must be Disabled: Overlay & L2 Cache clock gating must be disabled in order to prevent device hangs when turning off overlay.SW must turn off Ovrunit clock gating (6200h) and L2 Cache clock gating (C8h)." Now I've nowhere found that 0xc8 register and hence couldn't apply the l2 cache workaround. But I've remembered that part of the magic that the OVERLAY_ON/OFF commands are supposed to do is to rearrange cache allocations so that the overlay scaler has some scratch space. And while pondering how that could explain the hang the 2nd time we enable the overlay, I've remembered that the old ums overlay code did _not_ issue the OVERLAY_OFF cmd. And indeed, disabling the OFF cmd results in the overlay working flawlessly, so I guess we can workaround the lack of the above workaround by simply never disabling the overlay engine once it's enabled. Note that we have the first part of the above w/a already implemented in i830_init_clock_gating - leave that as-is to avoid surprises. v2: Add a comment in the code. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47827 Cc: stable@vger.kernel.org Tested-by: Rhys <rhyspuk@gmail.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-10-22 18:55:55 +08:00
} else {
intel_ring_emit(engine, MI_OVERLAY_FLIP | MI_OVERLAY_OFF);
intel_ring_emit(engine, flip_addr);
intel_ring_emit(engine,
MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
drm/i915: fix overlay on i830M The overlay on the i830M has a peculiar failure mode: It works the first time around after boot-up, but consistenly hangs the second time it's used. Chris Wilson has dug out a nice errata: "1.5.12 Clock Gating Disable for Display Register Address Offset: 06200h–06203h "Bit 3 Ovrunit Clock Gating Disable. 0 = Clock gating controlled by unit enabling logic 1 = Disable clock gating function DevALM Errata ALM049: Overlay Clock Gating Must be Disabled: Overlay & L2 Cache clock gating must be disabled in order to prevent device hangs when turning off overlay.SW must turn off Ovrunit clock gating (6200h) and L2 Cache clock gating (C8h)." Now I've nowhere found that 0xc8 register and hence couldn't apply the l2 cache workaround. But I've remembered that part of the magic that the OVERLAY_ON/OFF commands are supposed to do is to rearrange cache allocations so that the overlay scaler has some scratch space. And while pondering how that could explain the hang the 2nd time we enable the overlay, I've remembered that the old ums overlay code did _not_ issue the OVERLAY_OFF cmd. And indeed, disabling the OFF cmd results in the overlay working flawlessly, so I guess we can workaround the lack of the above workaround by simply never disabling the overlay engine once it's enabled. Note that we have the first part of the above w/a already implemented in i830_init_clock_gating - leave that as-is to avoid surprises. v2: Add a comment in the code. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47827 Cc: stable@vger.kernel.org Tested-by: Rhys <rhyspuk@gmail.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-10-22 18:55:55 +08:00
}
intel_ring_advance(engine);
return intel_overlay_do_wait_request(overlay, req, intel_overlay_off_tail);
}
/* recover from an interruption due to a signal
* We have to be careful not to repeat work forever an make forward progess. */
static int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay)
{
int ret;
if (overlay->last_flip_req == NULL)
return 0;
ret = i915_wait_request(overlay->last_flip_req);
if (ret)
return ret;
if (overlay->flip_tail)
overlay->flip_tail(overlay);
i915_gem_request_assign(&overlay->last_flip_req, NULL);
return 0;
}
/* Wait for pending overlay flip and release old frame.
* Needs to be called before the overlay register are changed
* via intel_overlay_(un)map_regs
*/
static int intel_overlay_release_old_vid(struct intel_overlay *overlay)
{
struct drm_device *dev = overlay->dev;
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_engine_cs *engine = &dev_priv->engine[RCS];
int ret;
WARN_ON(!mutex_is_locked(&dev->struct_mutex));
/* Only wait if there is actually an old frame to release to
* guarantee forward progress.
*/
if (!overlay->old_vid_bo)
return 0;
if (I915_READ(ISR) & I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT) {
/* synchronous slowpath */
struct drm_i915_gem_request *req;
req = i915_gem_request_alloc(engine, NULL);
drm/i915: simplify allocation of driver-internal requests There are a number of places where the driver needs a request, but isn't working on behalf of any specific user or in a specific context. At present, we associate them with the per-engine default context. A future patch will abolish those per-engine context pointers; but we can already eliminate a lot of the references to them, just by making the allocator allow NULL as a shorthand for "an appropriate context for this ring", which will mean that the callers don't need to know anything about how the "appropriate context" is found (e.g. per-ring vs per-device, etc). So this patch renames the existing i915_gem_request_alloc(), and makes it local (static inline), and replaces it with a wrapper that provides a default if the context is NULL, and also has a nicer calling convention (doesn't require a pointer to an output parameter). Then we change all callers to use the new convention: OLD: err = i915_gem_request_alloc(ring, user_ctx, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, user_ctx); if (IS_ERR(req)) ... OLD: err = i915_gem_request_alloc(ring, ring->default_context, &req); if (err) ... NEW: req = i915_gem_request_alloc(ring, NULL); if (IS_ERR(req)) ... v4: Rebased Signed-off-by: Dave Gordon <david.s.gordon@intel.com> Reviewed-by: Nick Hoath <nicholas.hoath@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1453230175-19330-2-git-send-email-david.s.gordon@intel.com Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2016-01-20 03:02:53 +08:00
if (IS_ERR(req))
return PTR_ERR(req);
ret = intel_ring_begin(req, 2);
if (ret) {
drm/i915: Late request cancellations are harmful Conceptually, each request is a record of a hardware transaction - we build up a list of pending commands and then either commit them to hardware, or cancel them. However, whilst building up the list of pending commands, we may modify state outside of the request and make references to the pending request. If we do so and then cancel that request, external objects then point to the deleted request leading to both graphical and memory corruption. The easiest example is to consider object/VMA tracking. When we mark an object as active in a request, we store a pointer to this, the most recent request, in the object. Then we want to free that object, we wait for the most recent request to be idle before proceeding (otherwise the hardware will write to pages now owned by the system, or we will attempt to read from those pages before the hardware is finished writing). If the request was cancelled instead, that wait completes immediately. As a result, all requests must be committed and not cancelled if the external state is unknown. All that remains of i915_gem_request_cancel() users are just a couple of extremely unlikely allocation failures, so remove the API entirely. A consequence of committing all incomplete requests is that we generate excess breadcrumbs and fill the ring much more often with dummy work. We have completely undone the outstanding_last_seqno optimisation. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93907 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: stable@vger.kernel.org Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1460565315-7748-16-git-send-email-chris@chris-wilson.co.uk
2016-04-14 00:35:15 +08:00
i915_add_request_no_flush(req);
return ret;
}
intel_ring_emit(engine,
MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
intel_ring_emit(engine, MI_NOOP);
intel_ring_advance(engine);
ret = intel_overlay_do_wait_request(overlay, req,
intel_overlay_release_old_vid_tail);
if (ret)
return ret;
}
intel_overlay_release_old_vid_tail(overlay);
drm/i915: Introduce accurate frontbuffer tracking So from just a quick look we seem to have enough information to accurately figure out whether a given gem bo is used as a frontbuffer and where exactly: We have obj->pin_count as a first check with no false negatives and only negligible false positives. And then we can just walk the modeset objects and figure out where exactly a buffer is used as scanout. Except that we can't due to locking order: If we already hold dev->struct_mutex we can't acquire any modeset locks, so could potential chase freed pointers and other evil stuff. So we need something else. For that introduce a new set of bits obj->frontbuffer_bits to track where a buffer object is used. That we can then chase without grabbing any modeset locks. Of course the consumers of this (DRRS, PSR, FBC, ...) still need to be able to do their magic both when called from modeset and from gem code. But that can be easily achieved by adding locks for these specific subsystems which always nest within either kms or gem locking. This patch just adds the relevant update code to all places. Note that if we ever support multi-planar scanout targets then we need one frontbuffer tracking bit per attachment point that we expose to userspace. v2: - Fix more oopsen. Oops. - WARN if we leak obj->frontbuffer_bits when freeing a gem buffer. Fix the bugs this brought to light. - s/update_frontbuffer_bits/update_fb_bits/. More consistent with the fb tracking functions (fb for gem object, frontbuffer for raw bits). And the function name was way too long. v3: Size obj->frontbuffer_bits correctly so that all pipes fit in. v4: Don't update fb bits in set_base on failure. Noticed by Chris. v5: s/i915_gem_update_fb_bits/i915_gem_track_fb/ Also remove a few local enum pipe variables which are now no longer needed to make the function arguments no drop over the 80 char limit. Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-06-19 05:28:09 +08:00
i915_gem_track_fb(overlay->old_vid_bo, NULL,
INTEL_FRONTBUFFER_OVERLAY(overlay->crtc->pipe));
return 0;
}
void intel_overlay_reset(struct drm_i915_private *dev_priv)
{
struct intel_overlay *overlay = dev_priv->overlay;
if (!overlay)
return;
intel_overlay_release_old_vid(overlay);
overlay->last_flip_req = NULL;
overlay->old_xscale = 0;
overlay->old_yscale = 0;
overlay->crtc = NULL;
overlay->active = false;
}
struct put_image_params {
int format;
short dst_x;
short dst_y;
short dst_w;
short dst_h;
short src_w;
short src_scan_h;
short src_scan_w;
short src_h;
short stride_Y;
short stride_UV;
int offset_Y;
int offset_U;
int offset_V;
};
static int packed_depth_bytes(u32 format)
{
switch (format & I915_OVERLAY_DEPTH_MASK) {
case I915_OVERLAY_YUV422:
return 4;
case I915_OVERLAY_YUV411:
/* return 6; not implemented */
default:
return -EINVAL;
}
}
static int packed_width_bytes(u32 format, short width)
{
switch (format & I915_OVERLAY_DEPTH_MASK) {
case I915_OVERLAY_YUV422:
return width << 1;
default:
return -EINVAL;
}
}
static int uv_hsubsampling(u32 format)
{
switch (format & I915_OVERLAY_DEPTH_MASK) {
case I915_OVERLAY_YUV422:
case I915_OVERLAY_YUV420:
return 2;
case I915_OVERLAY_YUV411:
case I915_OVERLAY_YUV410:
return 4;
default:
return -EINVAL;
}
}
static int uv_vsubsampling(u32 format)
{
switch (format & I915_OVERLAY_DEPTH_MASK) {
case I915_OVERLAY_YUV420:
case I915_OVERLAY_YUV410:
return 2;
case I915_OVERLAY_YUV422:
case I915_OVERLAY_YUV411:
return 1;
default:
return -EINVAL;
}
}
static u32 calc_swidthsw(struct drm_device *dev, u32 offset, u32 width)
{
u32 mask, shift, ret;
if (IS_GEN2(dev)) {
mask = 0x1f;
shift = 5;
} else {
mask = 0x3f;
shift = 6;
}
ret = ((offset + width + mask) >> shift) - (offset >> shift);
if (!IS_GEN2(dev))
ret <<= 1;
ret -= 1;
return ret << 2;
}
static const u16 y_static_hcoeffs[N_HORIZ_Y_TAPS * N_PHASES] = {
0x3000, 0xb4a0, 0x1930, 0x1920, 0xb4a0,
0x3000, 0xb500, 0x19d0, 0x1880, 0xb440,
0x3000, 0xb540, 0x1a88, 0x2f80, 0xb3e0,
0x3000, 0xb580, 0x1b30, 0x2e20, 0xb380,
0x3000, 0xb5c0, 0x1bd8, 0x2cc0, 0xb320,
0x3020, 0xb5e0, 0x1c60, 0x2b80, 0xb2c0,
0x3020, 0xb5e0, 0x1cf8, 0x2a20, 0xb260,
0x3020, 0xb5e0, 0x1d80, 0x28e0, 0xb200,
0x3020, 0xb5c0, 0x1e08, 0x3f40, 0xb1c0,
0x3020, 0xb580, 0x1e78, 0x3ce0, 0xb160,
0x3040, 0xb520, 0x1ed8, 0x3aa0, 0xb120,
0x3040, 0xb4a0, 0x1f30, 0x3880, 0xb0e0,
0x3040, 0xb400, 0x1f78, 0x3680, 0xb0a0,
0x3020, 0xb340, 0x1fb8, 0x34a0, 0xb060,
0x3020, 0xb240, 0x1fe0, 0x32e0, 0xb040,
0x3020, 0xb140, 0x1ff8, 0x3160, 0xb020,
0xb000, 0x3000, 0x0800, 0x3000, 0xb000
};
static const u16 uv_static_hcoeffs[N_HORIZ_UV_TAPS * N_PHASES] = {
0x3000, 0x1800, 0x1800, 0xb000, 0x18d0, 0x2e60,
0xb000, 0x1990, 0x2ce0, 0xb020, 0x1a68, 0x2b40,
0xb040, 0x1b20, 0x29e0, 0xb060, 0x1bd8, 0x2880,
0xb080, 0x1c88, 0x3e60, 0xb0a0, 0x1d28, 0x3c00,
0xb0c0, 0x1db8, 0x39e0, 0xb0e0, 0x1e40, 0x37e0,
0xb100, 0x1eb8, 0x3620, 0xb100, 0x1f18, 0x34a0,
0xb100, 0x1f68, 0x3360, 0xb0e0, 0x1fa8, 0x3240,
0xb0c0, 0x1fe0, 0x3140, 0xb060, 0x1ff0, 0x30a0,
0x3000, 0x0800, 0x3000
};
static void update_polyphase_filter(struct overlay_registers __iomem *regs)
{
memcpy_toio(regs->Y_HCOEFS, y_static_hcoeffs, sizeof(y_static_hcoeffs));
memcpy_toio(regs->UV_HCOEFS, uv_static_hcoeffs,
sizeof(uv_static_hcoeffs));
}
static bool update_scaling_factors(struct intel_overlay *overlay,
struct overlay_registers __iomem *regs,
struct put_image_params *params)
{
/* fixed point with a 12 bit shift */
u32 xscale, yscale, xscale_UV, yscale_UV;
#define FP_SHIFT 12
#define FRACT_MASK 0xfff
bool scale_changed = false;
int uv_hscale = uv_hsubsampling(params->format);
int uv_vscale = uv_vsubsampling(params->format);
if (params->dst_w > 1)
xscale = ((params->src_scan_w - 1) << FP_SHIFT)
/(params->dst_w);
else
xscale = 1 << FP_SHIFT;
if (params->dst_h > 1)
yscale = ((params->src_scan_h - 1) << FP_SHIFT)
/(params->dst_h);
else
yscale = 1 << FP_SHIFT;
/*if (params->format & I915_OVERLAY_YUV_PLANAR) {*/
xscale_UV = xscale/uv_hscale;
yscale_UV = yscale/uv_vscale;
/* make the Y scale to UV scale ratio an exact multiply */
xscale = xscale_UV * uv_hscale;
yscale = yscale_UV * uv_vscale;
/*} else {
xscale_UV = 0;
yscale_UV = 0;
}*/
if (xscale != overlay->old_xscale || yscale != overlay->old_yscale)
scale_changed = true;
overlay->old_xscale = xscale;
overlay->old_yscale = yscale;
iowrite32(((yscale & FRACT_MASK) << 20) |
((xscale >> FP_SHIFT) << 16) |
((xscale & FRACT_MASK) << 3),
&regs->YRGBSCALE);
iowrite32(((yscale_UV & FRACT_MASK) << 20) |
((xscale_UV >> FP_SHIFT) << 16) |
((xscale_UV & FRACT_MASK) << 3),
&regs->UVSCALE);
iowrite32((((yscale >> FP_SHIFT) << 16) |
((yscale_UV >> FP_SHIFT) << 0)),
&regs->UVSCALEV);
if (scale_changed)
update_polyphase_filter(regs);
return scale_changed;
}
static void update_colorkey(struct intel_overlay *overlay,
struct overlay_registers __iomem *regs)
{
u32 key = overlay->color_key;
u32 flags;
flags = 0;
if (overlay->color_key_enabled)
flags |= DST_KEY_ENABLE;
switch (overlay->crtc->base.primary->fb->bits_per_pixel) {
case 8:
key = 0;
flags |= CLK_RGB8I_MASK;
break;
case 16:
if (overlay->crtc->base.primary->fb->depth == 15) {
key = RGB15_TO_COLORKEY(key);
flags |= CLK_RGB15_MASK;
} else {
key = RGB16_TO_COLORKEY(key);
flags |= CLK_RGB16_MASK;
}
break;
case 24:
case 32:
flags |= CLK_RGB24_MASK;
break;
}
iowrite32(key, &regs->DCLRKV);
iowrite32(flags, &regs->DCLRKM);
}
static u32 overlay_cmd_reg(struct put_image_params *params)
{
u32 cmd = OCMD_ENABLE | OCMD_BUF_TYPE_FRAME | OCMD_BUFFER0;
if (params->format & I915_OVERLAY_YUV_PLANAR) {
switch (params->format & I915_OVERLAY_DEPTH_MASK) {
case I915_OVERLAY_YUV422:
cmd |= OCMD_YUV_422_PLANAR;
break;
case I915_OVERLAY_YUV420:
cmd |= OCMD_YUV_420_PLANAR;
break;
case I915_OVERLAY_YUV411:
case I915_OVERLAY_YUV410:
cmd |= OCMD_YUV_410_PLANAR;
break;
}
} else { /* YUV packed */
switch (params->format & I915_OVERLAY_DEPTH_MASK) {
case I915_OVERLAY_YUV422:
cmd |= OCMD_YUV_422_PACKED;
break;
case I915_OVERLAY_YUV411:
cmd |= OCMD_YUV_411_PACKED;
break;
}
switch (params->format & I915_OVERLAY_SWAP_MASK) {
case I915_OVERLAY_NO_SWAP:
break;
case I915_OVERLAY_UV_SWAP:
cmd |= OCMD_UV_SWAP;
break;
case I915_OVERLAY_Y_SWAP:
cmd |= OCMD_Y_SWAP;
break;
case I915_OVERLAY_Y_AND_UV_SWAP:
cmd |= OCMD_Y_AND_UV_SWAP;
break;
}
}
return cmd;
}
static int intel_overlay_do_put_image(struct intel_overlay *overlay,
struct drm_i915_gem_object *new_bo,
struct put_image_params *params)
{
int ret, tmp_width;
struct overlay_registers __iomem *regs;
bool scale_changed = false;
struct drm_device *dev = overlay->dev;
u32 swidth, swidthsw, sheight, ostride;
drm/i915: Introduce accurate frontbuffer tracking So from just a quick look we seem to have enough information to accurately figure out whether a given gem bo is used as a frontbuffer and where exactly: We have obj->pin_count as a first check with no false negatives and only negligible false positives. And then we can just walk the modeset objects and figure out where exactly a buffer is used as scanout. Except that we can't due to locking order: If we already hold dev->struct_mutex we can't acquire any modeset locks, so could potential chase freed pointers and other evil stuff. So we need something else. For that introduce a new set of bits obj->frontbuffer_bits to track where a buffer object is used. That we can then chase without grabbing any modeset locks. Of course the consumers of this (DRRS, PSR, FBC, ...) still need to be able to do their magic both when called from modeset and from gem code. But that can be easily achieved by adding locks for these specific subsystems which always nest within either kms or gem locking. This patch just adds the relevant update code to all places. Note that if we ever support multi-planar scanout targets then we need one frontbuffer tracking bit per attachment point that we expose to userspace. v2: - Fix more oopsen. Oops. - WARN if we leak obj->frontbuffer_bits when freeing a gem buffer. Fix the bugs this brought to light. - s/update_frontbuffer_bits/update_fb_bits/. More consistent with the fb tracking functions (fb for gem object, frontbuffer for raw bits). And the function name was way too long. v3: Size obj->frontbuffer_bits correctly so that all pipes fit in. v4: Don't update fb bits in set_base on failure. Noticed by Chris. v5: s/i915_gem_update_fb_bits/i915_gem_track_fb/ Also remove a few local enum pipe variables which are now no longer needed to make the function arguments no drop over the 80 char limit. Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-06-19 05:28:09 +08:00
enum pipe pipe = overlay->crtc->pipe;
WARN_ON(!mutex_is_locked(&dev->struct_mutex));
WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
ret = intel_overlay_release_old_vid(overlay);
if (ret != 0)
return ret;
ret = i915_gem_object_pin_to_display_plane(new_bo, 0,
&i915_ggtt_view_normal);
if (ret != 0)
return ret;
ret = i915_gem_object_put_fence(new_bo);
if (ret)
goto out_unpin;
if (!overlay->active) {
u32 oconfig;
regs = intel_overlay_map_regs(overlay);
if (!regs) {
ret = -ENOMEM;
goto out_unpin;
}
oconfig = OCONF_CC_OUT_8BIT;
if (IS_GEN4(overlay->dev))
oconfig |= OCONF_CSC_MODE_BT709;
drm/i915: Introduce accurate frontbuffer tracking So from just a quick look we seem to have enough information to accurately figure out whether a given gem bo is used as a frontbuffer and where exactly: We have obj->pin_count as a first check with no false negatives and only negligible false positives. And then we can just walk the modeset objects and figure out where exactly a buffer is used as scanout. Except that we can't due to locking order: If we already hold dev->struct_mutex we can't acquire any modeset locks, so could potential chase freed pointers and other evil stuff. So we need something else. For that introduce a new set of bits obj->frontbuffer_bits to track where a buffer object is used. That we can then chase without grabbing any modeset locks. Of course the consumers of this (DRRS, PSR, FBC, ...) still need to be able to do their magic both when called from modeset and from gem code. But that can be easily achieved by adding locks for these specific subsystems which always nest within either kms or gem locking. This patch just adds the relevant update code to all places. Note that if we ever support multi-planar scanout targets then we need one frontbuffer tracking bit per attachment point that we expose to userspace. v2: - Fix more oopsen. Oops. - WARN if we leak obj->frontbuffer_bits when freeing a gem buffer. Fix the bugs this brought to light. - s/update_frontbuffer_bits/update_fb_bits/. More consistent with the fb tracking functions (fb for gem object, frontbuffer for raw bits). And the function name was way too long. v3: Size obj->frontbuffer_bits correctly so that all pipes fit in. v4: Don't update fb bits in set_base on failure. Noticed by Chris. v5: s/i915_gem_update_fb_bits/i915_gem_track_fb/ Also remove a few local enum pipe variables which are now no longer needed to make the function arguments no drop over the 80 char limit. Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-06-19 05:28:09 +08:00
oconfig |= pipe == 0 ?
OCONF_PIPE_A : OCONF_PIPE_B;
iowrite32(oconfig, &regs->OCONFIG);
intel_overlay_unmap_regs(overlay, regs);
ret = intel_overlay_on(overlay);
if (ret != 0)
goto out_unpin;
}
regs = intel_overlay_map_regs(overlay);
if (!regs) {
ret = -ENOMEM;
goto out_unpin;
}
iowrite32((params->dst_y << 16) | params->dst_x, &regs->DWINPOS);
iowrite32((params->dst_h << 16) | params->dst_w, &regs->DWINSZ);
if (params->format & I915_OVERLAY_YUV_PACKED)
tmp_width = packed_width_bytes(params->format, params->src_w);
else
tmp_width = params->src_w;
swidth = params->src_w;
swidthsw = calc_swidthsw(overlay->dev, params->offset_Y, tmp_width);
sheight = params->src_h;
iowrite32(i915_gem_obj_ggtt_offset(new_bo) + params->offset_Y, &regs->OBUF_0Y);
ostride = params->stride_Y;
if (params->format & I915_OVERLAY_YUV_PLANAR) {
int uv_hscale = uv_hsubsampling(params->format);
int uv_vscale = uv_vsubsampling(params->format);
u32 tmp_U, tmp_V;
swidth |= (params->src_w/uv_hscale) << 16;
tmp_U = calc_swidthsw(overlay->dev, params->offset_U,
params->src_w/uv_hscale);
tmp_V = calc_swidthsw(overlay->dev, params->offset_V,
params->src_w/uv_hscale);
swidthsw |= max_t(u32, tmp_U, tmp_V) << 16;
sheight |= (params->src_h/uv_vscale) << 16;
iowrite32(i915_gem_obj_ggtt_offset(new_bo) + params->offset_U, &regs->OBUF_0U);
iowrite32(i915_gem_obj_ggtt_offset(new_bo) + params->offset_V, &regs->OBUF_0V);
ostride |= params->stride_UV << 16;
}
iowrite32(swidth, &regs->SWIDTH);
iowrite32(swidthsw, &regs->SWIDTHSW);
iowrite32(sheight, &regs->SHEIGHT);
iowrite32(ostride, &regs->OSTRIDE);
scale_changed = update_scaling_factors(overlay, regs, params);
update_colorkey(overlay, regs);
iowrite32(overlay_cmd_reg(params), &regs->OCMD);
intel_overlay_unmap_regs(overlay, regs);
ret = intel_overlay_continue(overlay, scale_changed);
if (ret)
goto out_unpin;
drm/i915: Introduce accurate frontbuffer tracking So from just a quick look we seem to have enough information to accurately figure out whether a given gem bo is used as a frontbuffer and where exactly: We have obj->pin_count as a first check with no false negatives and only negligible false positives. And then we can just walk the modeset objects and figure out where exactly a buffer is used as scanout. Except that we can't due to locking order: If we already hold dev->struct_mutex we can't acquire any modeset locks, so could potential chase freed pointers and other evil stuff. So we need something else. For that introduce a new set of bits obj->frontbuffer_bits to track where a buffer object is used. That we can then chase without grabbing any modeset locks. Of course the consumers of this (DRRS, PSR, FBC, ...) still need to be able to do their magic both when called from modeset and from gem code. But that can be easily achieved by adding locks for these specific subsystems which always nest within either kms or gem locking. This patch just adds the relevant update code to all places. Note that if we ever support multi-planar scanout targets then we need one frontbuffer tracking bit per attachment point that we expose to userspace. v2: - Fix more oopsen. Oops. - WARN if we leak obj->frontbuffer_bits when freeing a gem buffer. Fix the bugs this brought to light. - s/update_frontbuffer_bits/update_fb_bits/. More consistent with the fb tracking functions (fb for gem object, frontbuffer for raw bits). And the function name was way too long. v3: Size obj->frontbuffer_bits correctly so that all pipes fit in. v4: Don't update fb bits in set_base on failure. Noticed by Chris. v5: s/i915_gem_update_fb_bits/i915_gem_track_fb/ Also remove a few local enum pipe variables which are now no longer needed to make the function arguments no drop over the 80 char limit. Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-06-19 05:28:09 +08:00
i915_gem_track_fb(overlay->vid_bo, new_bo,
INTEL_FRONTBUFFER_OVERLAY(pipe));
overlay->old_vid_bo = overlay->vid_bo;
overlay->vid_bo = new_bo;
drm/i915: Track frontbuffer invalidation/flushing So these are the guts of the new beast. This tracks when a frontbuffer gets invalidated (due to frontbuffer rendering) and hence should be constantly scaned out, and when it's flushed again and can be compressed/one-shot-upload. Rules for flushing are simple: The frontbuffer needs one more full upload starting from the next vblank. Which means that the flushing can _only_ be called once the frontbuffer update has been latched. But this poses a problem for pageflips: We can't just delay the flushing until the pageflip is latched, since that would pose the risk that we override frontbuffer rendering that has been scheduled in-between the pageflip ioctl and the actual latching. To handle this track asynchronous invalidations (and also pageflip) state per-ring and delay any in-between flushing until the rendering has completed. And also cancel any delayed flushing if we get a new invalidation request (whether delayed or not). Also call intel_mark_fb_busy in both cases in all cases to make sure that we keep the screen at the highest refresh rate both on flips, synchronous plane updates and for frontbuffer rendering. v2: Lots of improvements Suggestions from Chris: - Move invalidate/flush in flush_*_domain and set_to_*_domain. - Drop the flush in busy_ioctl since it's redundant. Was a leftover from an earlier concept to track flips/delayed flushes. - Don't forget about the initial modeset enable/final disable. Suggested by Chris. Track flips accurately, too. Since flips complete independently of rendering we need to track pending flips in a separate mask. Again if an invalidate happens we need to cancel the evenutal flush to avoid races. v3: Provide correct header declarations for flip functions. Currently not needed outside of intel_display.c, but part of the proper interface. v4: Add proper domain management to fbcon so that the fbcon buffer is also tracked correctly. v5: Fixup locking around the fbcon set_to_gtt_domain call. v6: More comments from Chris: - Split out fbcon changes. - Drop superflous checks for potential scanout before calling intel_fb functions - we can micro-optimize this later. - s/intel_fb_/intel_fb_obj_/ to make it clear that this deals in gem object. We already have precedence for fb_obj in the pin_and_fence functions. v7: Clarify the semantics of the flip flush handling by renaming things a bit: - Don't go through a gem object but take the relevant frontbuffer bits directly. These functions center on the plane, the actual object is irrelevant - even a flip to the same object as already active should cause a flush. - Add a new intel_frontbuffer_flip for synchronous plane updates. It currently just calls intel_frontbuffer_flush since the implemenation differs. This way we achieve a clear split between one-shot update events on one side and frontbuffer rendering with potentially a very long delay between the invalidate and flush. Chris and I also had some discussions about mark_busy and whether it is appropriate to call from flush. But mark busy is a state which should be derived from the 3 events (invalidate, flush, flip) we now have by the users, like psr does by tracking relevant information in psr.busy_frontbuffer_bits. DRRS (the only real use of mark_busy for frontbuffer) needs to have similar logic. With that the overall mark_busy in the core could be removed. v8: Only when retiring gpu buffers only flush frontbuffer bits we actually invalidated in a batch. Just for safety since before any additional usage/invalidate we should always retire current rendering. Suggested by Chris Wilson. v9: Actually use intel_frontbuffer_flip in all appropriate places. Spotted by Chris. v10: Address more comments from Chris: - Don't call _flip in set_base when the crtc is inactive, avoids redunancy in the modeset case with the initial enabling of all planes. - Add comments explaining that the initial/final plane enable/disable still has work left to do before it's fully generic. v11: Only invalidate for gtt/cpu access when writing. Spotted by Chris. v12: s/_flush/_flip/ in intel_overlay.c per Chris' comment. Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-06-19 22:01:59 +08:00
intel_frontbuffer_flip(dev,
INTEL_FRONTBUFFER_OVERLAY(pipe));
return 0;
out_unpin:
i915_gem_object_ggtt_unpin(new_bo);
return ret;
}
int intel_overlay_switch_off(struct intel_overlay *overlay)
{
struct overlay_registers __iomem *regs;
struct drm_device *dev = overlay->dev;
int ret;
WARN_ON(!mutex_is_locked(&dev->struct_mutex));
WARN_ON(!drm_modeset_is_locked(&dev->mode_config.connection_mutex));
ret = intel_overlay_recover_from_interrupt(overlay);
if (ret != 0)
return ret;
if (!overlay->active)
return 0;
ret = intel_overlay_release_old_vid(overlay);
if (ret != 0)
return ret;
regs = intel_overlay_map_regs(overlay);
iowrite32(0, &regs->OCMD);
intel_overlay_unmap_regs(overlay, regs);
ret = intel_overlay_off(overlay);
if (ret != 0)
return ret;
intel_overlay_off_tail(overlay);
return 0;
}
static int check_overlay_possible_on_crtc(struct intel_overlay *overlay,
struct intel_crtc *crtc)
{
if (!crtc->active)
return -EINVAL;
/* can't use the overlay with double wide pipe */
drm/i915: Make intel_crtc->config a pointer To match the semantics of drm_crtc->state, which this will eventually become. The allocation of the memory for config will be fixed in a followup patch. By adding the extra _config field to intel_crtc it was possible to generate this entire patch with the cocci script below. @@ @@ struct intel_crtc { ... -struct intel_crtc_state config; +struct intel_crtc_state _config; +struct intel_crtc_state *config; ... } @@ struct intel_crtc *crtc; @@ -memset(&crtc->config, 0, sizeof(crtc->config)); +memset(crtc->config, 0, sizeof(*crtc->config)); @@ @@ __intel_set_mode(...) { <... -to_intel_crtc(crtc)->config = *pipe_config; +(*(to_intel_crtc(crtc)->config)) = *pipe_config; ...> } @@ @@ intel_crtc_init(...) { ... WARN_ON(drm_crtc_index(&intel_crtc->base) != intel_crtc->pipe); +intel_crtc->config = &intel_crtc->_config; return; ... } @@ struct intel_crtc *crtc; @@ -&crtc->config +crtc->config @@ struct intel_crtc *crtc; identifier member; @@ -crtc->config.member +crtc->config->member @@ expression E; @@ -&(to_intel_crtc(E)->config) +to_intel_crtc(E)->config @@ expression E; identifier member; @@ -to_intel_crtc(E)->config.member +to_intel_crtc(E)->config->member v2: Clarify manual changes by splitting them into another patch. (Matt) Improve cocci script to generate even more of the changes. (Ander) Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> Reviewed-by: Matt Roper <matthew.d.roper@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-01-15 20:55:25 +08:00
if (crtc->config->double_wide)
return -EINVAL;
return 0;
}
static void update_pfit_vscale_ratio(struct intel_overlay *overlay)
{
struct drm_device *dev = overlay->dev;
struct drm_i915_private *dev_priv = dev->dev_private;
u32 pfit_control = I915_READ(PFIT_CONTROL);
u32 ratio;
/* XXX: This is not the same logic as in the xorg driver, but more in
* line with the intel documentation for the i965
*/
if (INTEL_INFO(dev)->gen >= 4) {
/* on i965 use the PGM reg to read out the autoscaler values */
ratio = I915_READ(PFIT_PGM_RATIOS) >> PFIT_VERT_SCALE_SHIFT_965;
} else {
if (pfit_control & VERT_AUTO_SCALE)
ratio = I915_READ(PFIT_AUTO_RATIOS);
else
ratio = I915_READ(PFIT_PGM_RATIOS);
ratio >>= PFIT_VERT_SCALE_SHIFT;
}
overlay->pfit_vscale_ratio = ratio;
}
static int check_overlay_dst(struct intel_overlay *overlay,
struct drm_intel_overlay_put_image *rec)
{
struct drm_display_mode *mode = &overlay->crtc->base.mode;
if (rec->dst_x < mode->hdisplay &&
rec->dst_x + rec->dst_width <= mode->hdisplay &&
rec->dst_y < mode->vdisplay &&
rec->dst_y + rec->dst_height <= mode->vdisplay)
return 0;
else
return -EINVAL;
}
static int check_overlay_scaling(struct put_image_params *rec)
{
u32 tmp;
/* downscaling limit is 8.0 */
tmp = ((rec->src_scan_h << 16) / rec->dst_h) >> 16;
if (tmp > 7)
return -EINVAL;
tmp = ((rec->src_scan_w << 16) / rec->dst_w) >> 16;
if (tmp > 7)
return -EINVAL;
return 0;
}
static int check_overlay_src(struct drm_device *dev,
struct drm_intel_overlay_put_image *rec,
struct drm_i915_gem_object *new_bo)
{
int uv_hscale = uv_hsubsampling(rec->flags);
int uv_vscale = uv_vsubsampling(rec->flags);
u32 stride_mask;
int depth;
u32 tmp;
/* check src dimensions */
if (IS_845G(dev) || IS_I830(dev)) {
if (rec->src_height > IMAGE_MAX_HEIGHT_LEGACY ||
rec->src_width > IMAGE_MAX_WIDTH_LEGACY)
return -EINVAL;
} else {
if (rec->src_height > IMAGE_MAX_HEIGHT ||
rec->src_width > IMAGE_MAX_WIDTH)
return -EINVAL;
}
/* better safe than sorry, use 4 as the maximal subsampling ratio */
if (rec->src_height < N_VERT_Y_TAPS*4 ||
rec->src_width < N_HORIZ_Y_TAPS*4)
return -EINVAL;
/* check alignment constraints */
switch (rec->flags & I915_OVERLAY_TYPE_MASK) {
case I915_OVERLAY_RGB:
/* not implemented */
return -EINVAL;
case I915_OVERLAY_YUV_PACKED:
if (uv_vscale != 1)
return -EINVAL;
depth = packed_depth_bytes(rec->flags);
if (depth < 0)
return depth;
/* ignore UV planes */
rec->stride_UV = 0;
rec->offset_U = 0;
rec->offset_V = 0;
/* check pixel alignment */
if (rec->offset_Y % depth)
return -EINVAL;
break;
case I915_OVERLAY_YUV_PLANAR:
if (uv_vscale < 0 || uv_hscale < 0)
return -EINVAL;
/* no offset restrictions for planar formats */
break;
default:
return -EINVAL;
}
if (rec->src_width % uv_hscale)
return -EINVAL;
/* stride checking */
if (IS_I830(dev) || IS_845G(dev))
stride_mask = 255;
else
stride_mask = 63;
if (rec->stride_Y & stride_mask || rec->stride_UV & stride_mask)
return -EINVAL;
if (IS_GEN4(dev) && rec->stride_Y < 512)
return -EINVAL;
tmp = (rec->flags & I915_OVERLAY_TYPE_MASK) == I915_OVERLAY_YUV_PLANAR ?
4096 : 8192;
if (rec->stride_Y > tmp || rec->stride_UV > 2*1024)
return -EINVAL;
/* check buffer dimensions */
switch (rec->flags & I915_OVERLAY_TYPE_MASK) {
case I915_OVERLAY_RGB:
case I915_OVERLAY_YUV_PACKED:
/* always 4 Y values per depth pixels */
if (packed_width_bytes(rec->flags, rec->src_width) > rec->stride_Y)
return -EINVAL;
tmp = rec->stride_Y*rec->src_height;
if (rec->offset_Y + tmp > new_bo->base.size)
return -EINVAL;
break;
case I915_OVERLAY_YUV_PLANAR:
if (rec->src_width > rec->stride_Y)
return -EINVAL;
if (rec->src_width/uv_hscale > rec->stride_UV)
return -EINVAL;
tmp = rec->stride_Y * rec->src_height;
if (rec->offset_Y + tmp > new_bo->base.size)
return -EINVAL;
tmp = rec->stride_UV * (rec->src_height / uv_vscale);
if (rec->offset_U + tmp > new_bo->base.size ||
rec->offset_V + tmp > new_bo->base.size)
return -EINVAL;
break;
}
return 0;
}
/**
* Return the pipe currently connected to the panel fitter,
* or -1 if the panel fitter is not present or not in use
*/
static int intel_panel_fitter_pipe(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 pfit_control;
/* i830 doesn't have a panel fitter */
if (INTEL_INFO(dev)->gen <= 3 && (IS_I830(dev) || !IS_MOBILE(dev)))
return -1;
pfit_control = I915_READ(PFIT_CONTROL);
/* See if the panel fitter is in use */
if ((pfit_control & PFIT_ENABLE) == 0)
return -1;
/* 965 can place panel fitter on either pipe */
if (IS_GEN4(dev))
return (pfit_control >> 29) & 0x3;
/* older chips can only use pipe 1 */
return 1;
}
int intel_overlay_put_image(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_intel_overlay_put_image *put_image_rec = data;
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_overlay *overlay;
struct drm_crtc *drmmode_crtc;
struct intel_crtc *crtc;
struct drm_i915_gem_object *new_bo;
struct put_image_params *params;
int ret;
overlay = dev_priv->overlay;
if (!overlay) {
DRM_DEBUG("userspace bug: no overlay\n");
return -ENODEV;
}
if (!(put_image_rec->flags & I915_OVERLAY_ENABLE)) {
drm_modeset_lock_all(dev);
mutex_lock(&dev->struct_mutex);
ret = intel_overlay_switch_off(overlay);
mutex_unlock(&dev->struct_mutex);
drm_modeset_unlock_all(dev);
return ret;
}
params = kmalloc(sizeof(*params), GFP_KERNEL);
if (!params)
return -ENOMEM;
drmmode_crtc = drm_crtc_find(dev, put_image_rec->crtc_id);
if (!drmmode_crtc) {
ret = -ENOENT;
goto out_free;
}
crtc = to_intel_crtc(drmmode_crtc);
new_bo = to_intel_bo(drm_gem_object_lookup(dev, file_priv,
put_image_rec->bo_handle));
if (&new_bo->base == NULL) {
ret = -ENOENT;
goto out_free;
}
drm_modeset_lock_all(dev);
mutex_lock(&dev->struct_mutex);
if (new_bo->tiling_mode) {
DRM_DEBUG_KMS("buffer used for overlay image can not be tiled\n");
ret = -EINVAL;
goto out_unlock;
}
ret = intel_overlay_recover_from_interrupt(overlay);
if (ret != 0)
goto out_unlock;
if (overlay->crtc != crtc) {
struct drm_display_mode *mode = &crtc->base.mode;
ret = intel_overlay_switch_off(overlay);
if (ret != 0)
goto out_unlock;
ret = check_overlay_possible_on_crtc(overlay, crtc);
if (ret != 0)
goto out_unlock;
overlay->crtc = crtc;
crtc->overlay = overlay;
/* line too wide, i.e. one-line-mode */
if (mode->hdisplay > 1024 &&
intel_panel_fitter_pipe(dev) == crtc->pipe) {
overlay->pfit_active = true;
update_pfit_vscale_ratio(overlay);
} else
overlay->pfit_active = false;
}
ret = check_overlay_dst(overlay, put_image_rec);
if (ret != 0)
goto out_unlock;
if (overlay->pfit_active) {
params->dst_y = ((((u32)put_image_rec->dst_y) << 12) /
overlay->pfit_vscale_ratio);
/* shifting right rounds downwards, so add 1 */
params->dst_h = ((((u32)put_image_rec->dst_height) << 12) /
overlay->pfit_vscale_ratio) + 1;
} else {
params->dst_y = put_image_rec->dst_y;
params->dst_h = put_image_rec->dst_height;
}
params->dst_x = put_image_rec->dst_x;
params->dst_w = put_image_rec->dst_width;
params->src_w = put_image_rec->src_width;
params->src_h = put_image_rec->src_height;
params->src_scan_w = put_image_rec->src_scan_width;
params->src_scan_h = put_image_rec->src_scan_height;
if (params->src_scan_h > params->src_h ||
params->src_scan_w > params->src_w) {
ret = -EINVAL;
goto out_unlock;
}
ret = check_overlay_src(dev, put_image_rec, new_bo);
if (ret != 0)
goto out_unlock;
params->format = put_image_rec->flags & ~I915_OVERLAY_FLAGS_MASK;
params->stride_Y = put_image_rec->stride_Y;
params->stride_UV = put_image_rec->stride_UV;
params->offset_Y = put_image_rec->offset_Y;
params->offset_U = put_image_rec->offset_U;
params->offset_V = put_image_rec->offset_V;
/* Check scaling after src size to prevent a divide-by-zero. */
ret = check_overlay_scaling(params);
if (ret != 0)
goto out_unlock;
ret = intel_overlay_do_put_image(overlay, new_bo, params);
if (ret != 0)
goto out_unlock;
mutex_unlock(&dev->struct_mutex);
drm_modeset_unlock_all(dev);
kfree(params);
return 0;
out_unlock:
mutex_unlock(&dev->struct_mutex);
drm_modeset_unlock_all(dev);
drm_gem_object_unreference_unlocked(&new_bo->base);
out_free:
kfree(params);
return ret;
}
static void update_reg_attrs(struct intel_overlay *overlay,
struct overlay_registers __iomem *regs)
{
iowrite32((overlay->contrast << 18) | (overlay->brightness & 0xff),
&regs->OCLRC0);
iowrite32(overlay->saturation, &regs->OCLRC1);
}
static bool check_gamma_bounds(u32 gamma1, u32 gamma2)
{
int i;
if (gamma1 & 0xff000000 || gamma2 & 0xff000000)
return false;
for (i = 0; i < 3; i++) {
if (((gamma1 >> i*8) & 0xff) >= ((gamma2 >> i*8) & 0xff))
return false;
}
return true;
}
static bool check_gamma5_errata(u32 gamma5)
{
int i;
for (i = 0; i < 3; i++) {
if (((gamma5 >> i*8) & 0xff) == 0x80)
return false;
}
return true;
}
static int check_gamma(struct drm_intel_overlay_attrs *attrs)
{
if (!check_gamma_bounds(0, attrs->gamma0) ||
!check_gamma_bounds(attrs->gamma0, attrs->gamma1) ||
!check_gamma_bounds(attrs->gamma1, attrs->gamma2) ||
!check_gamma_bounds(attrs->gamma2, attrs->gamma3) ||
!check_gamma_bounds(attrs->gamma3, attrs->gamma4) ||
!check_gamma_bounds(attrs->gamma4, attrs->gamma5) ||
!check_gamma_bounds(attrs->gamma5, 0x00ffffff))
return -EINVAL;
if (!check_gamma5_errata(attrs->gamma5))
return -EINVAL;
return 0;
}
int intel_overlay_attrs(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_intel_overlay_attrs *attrs = data;
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_overlay *overlay;
struct overlay_registers __iomem *regs;
int ret;
overlay = dev_priv->overlay;
if (!overlay) {
DRM_DEBUG("userspace bug: no overlay\n");
return -ENODEV;
}
drm_modeset_lock_all(dev);
mutex_lock(&dev->struct_mutex);
ret = -EINVAL;
if (!(attrs->flags & I915_OVERLAY_UPDATE_ATTRS)) {
attrs->color_key = overlay->color_key;
attrs->brightness = overlay->brightness;
attrs->contrast = overlay->contrast;
attrs->saturation = overlay->saturation;
if (!IS_GEN2(dev)) {
attrs->gamma0 = I915_READ(OGAMC0);
attrs->gamma1 = I915_READ(OGAMC1);
attrs->gamma2 = I915_READ(OGAMC2);
attrs->gamma3 = I915_READ(OGAMC3);
attrs->gamma4 = I915_READ(OGAMC4);
attrs->gamma5 = I915_READ(OGAMC5);
}
} else {
if (attrs->brightness < -128 || attrs->brightness > 127)
goto out_unlock;
if (attrs->contrast > 255)
goto out_unlock;
if (attrs->saturation > 1023)
goto out_unlock;
overlay->color_key = attrs->color_key;
overlay->brightness = attrs->brightness;
overlay->contrast = attrs->contrast;
overlay->saturation = attrs->saturation;
regs = intel_overlay_map_regs(overlay);
if (!regs) {
ret = -ENOMEM;
goto out_unlock;
}
update_reg_attrs(overlay, regs);
intel_overlay_unmap_regs(overlay, regs);
if (attrs->flags & I915_OVERLAY_UPDATE_GAMMA) {
if (IS_GEN2(dev))
goto out_unlock;
if (overlay->active) {
ret = -EBUSY;
goto out_unlock;
}
ret = check_gamma(attrs);
if (ret)
goto out_unlock;
I915_WRITE(OGAMC0, attrs->gamma0);
I915_WRITE(OGAMC1, attrs->gamma1);
I915_WRITE(OGAMC2, attrs->gamma2);
I915_WRITE(OGAMC3, attrs->gamma3);
I915_WRITE(OGAMC4, attrs->gamma4);
I915_WRITE(OGAMC5, attrs->gamma5);
}
}
overlay->color_key_enabled = (attrs->flags & I915_OVERLAY_DISABLE_DEST_COLORKEY) == 0;
ret = 0;
out_unlock:
mutex_unlock(&dev->struct_mutex);
drm_modeset_unlock_all(dev);
return ret;
}
void intel_setup_overlay(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_overlay *overlay;
struct drm_i915_gem_object *reg_bo;
struct overlay_registers __iomem *regs;
int ret;
if (!HAS_OVERLAY(dev))
return;
overlay = kzalloc(sizeof(*overlay), GFP_KERNEL);
if (!overlay)
return;
mutex_lock(&dev->struct_mutex);
if (WARN_ON(dev_priv->overlay))
goto out_free;
overlay->dev = dev;
reg_bo = NULL;
if (!OVERLAY_NEEDS_PHYSICAL(dev))
reg_bo = i915_gem_object_create_stolen(dev, PAGE_SIZE);
if (reg_bo == NULL)
reg_bo = i915_gem_object_create(dev, PAGE_SIZE);
if (IS_ERR(reg_bo))
goto out_free;
overlay->reg_bo = reg_bo;
if (OVERLAY_NEEDS_PHYSICAL(dev)) {
ret = i915_gem_object_attach_phys(reg_bo, PAGE_SIZE);
if (ret) {
DRM_ERROR("failed to attach phys overlay regs\n");
goto out_free_bo;
}
overlay->flip_addr = reg_bo->phys_handle->busaddr;
} else {
ret = i915_gem_obj_ggtt_pin(reg_bo, PAGE_SIZE, PIN_MAPPABLE);
if (ret) {
DRM_ERROR("failed to pin overlay register bo\n");
goto out_free_bo;
}
overlay->flip_addr = i915_gem_obj_ggtt_offset(reg_bo);
ret = i915_gem_object_set_to_gtt_domain(reg_bo, true);
if (ret) {
DRM_ERROR("failed to move overlay register bo into the GTT\n");
goto out_unpin_bo;
}
}
/* init all values */
overlay->color_key = 0x0101fe;
overlay->color_key_enabled = true;
overlay->brightness = -19;
overlay->contrast = 75;
overlay->saturation = 146;
regs = intel_overlay_map_regs(overlay);
if (!regs)
goto out_unpin_bo;
memset_io(regs, 0, sizeof(struct overlay_registers));
update_polyphase_filter(regs);
update_reg_attrs(overlay, regs);
intel_overlay_unmap_regs(overlay, regs);
dev_priv->overlay = overlay;
mutex_unlock(&dev->struct_mutex);
DRM_INFO("initialized overlay support\n");
return;
out_unpin_bo:
if (!OVERLAY_NEEDS_PHYSICAL(dev))
i915_gem_object_ggtt_unpin(reg_bo);
out_free_bo:
drm_gem_object_unreference(&reg_bo->base);
out_free:
mutex_unlock(&dev->struct_mutex);
kfree(overlay);
return;
}
void intel_cleanup_overlay(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (!dev_priv->overlay)
return;
/* The bo's should be free'd by the generic code already.
* Furthermore modesetting teardown happens beforehand so the
* hardware should be off already */
WARN_ON(dev_priv->overlay->active);
drm_gem_object_unreference_unlocked(&dev_priv->overlay->reg_bo->base);
kfree(dev_priv->overlay);
}
struct intel_overlay_error_state {
struct overlay_registers regs;
unsigned long base;
u32 dovsta;
u32 isr;
};
static struct overlay_registers __iomem *
Merge branch 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6 * 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (476 commits) vmwgfx: Implement a proper GMR eviction mechanism drm/radeon/kms: fix r6xx/7xx 1D tiling CS checker v2 drm/radeon/kms: properly compute group_size on 6xx/7xx drm/radeon/kms: fix 2D tile height alignment in the r600 CS checker drm/radeon/kms/evergreen: set the clear state to the blit state drm/radeon/kms: don't poll dac load detect. gpu: Add Intel GMA500(Poulsbo) Stub Driver drm/radeon/kms: MC vram map needs to be >= pci aperture size drm/radeon/kms: implement display watermark support for evergreen drm/radeon/kms/evergreen: add some additional safe regs v2 drm/radeon/r600: fix tiling issues in CS checker. drm/i915: Move gpu_write_list to per-ring drm/i915: Invalidate the to-ring, flush the old-ring when updating domains drm/i915/ringbuffer: Write the value passed in to the tail register agp/intel: Restore valid PTE bit for Sandybridge after bdd3072 drm/i915: Fix flushing regression from 9af90d19f drm/i915/sdvo: Remove unused encoding member i915: enable AVI infoframe for intel_hdmi.c [v4] drm/i915: Fix current fb blocking for page flip drm/i915: IS_IRONLAKE is synonymous with gen == 5 ... Fix up conflicts in - drivers/gpu/drm/i915/{i915_gem.c, i915/intel_overlay.c}: due to the new simplified stack-based kmap_atomic() interface - drivers/gpu/drm/vmwgfx/vmwgfx_drv.c: added .llseek entry due to BKL removal cleanups.
2010-10-27 09:57:59 +08:00
intel_overlay_map_regs_atomic(struct intel_overlay *overlay)
{
struct drm_i915_private *dev_priv = to_i915(overlay->dev);
struct i915_ggtt *ggtt = &dev_priv->ggtt;
struct overlay_registers __iomem *regs;
if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
/* Cast to make sparse happy, but it's wc memory anyway, so
* equivalent to the wc io mapping on X86. */
regs = (struct overlay_registers __iomem *)
overlay->reg_bo->phys_handle->vaddr;
else
regs = io_mapping_map_atomic_wc(ggtt->mappable,
i915_gem_obj_ggtt_offset(overlay->reg_bo));
return regs;
}
static void intel_overlay_unmap_regs_atomic(struct intel_overlay *overlay,
struct overlay_registers __iomem *regs)
{
if (!OVERLAY_NEEDS_PHYSICAL(overlay->dev))
Merge branch 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6 * 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (476 commits) vmwgfx: Implement a proper GMR eviction mechanism drm/radeon/kms: fix r6xx/7xx 1D tiling CS checker v2 drm/radeon/kms: properly compute group_size on 6xx/7xx drm/radeon/kms: fix 2D tile height alignment in the r600 CS checker drm/radeon/kms/evergreen: set the clear state to the blit state drm/radeon/kms: don't poll dac load detect. gpu: Add Intel GMA500(Poulsbo) Stub Driver drm/radeon/kms: MC vram map needs to be >= pci aperture size drm/radeon/kms: implement display watermark support for evergreen drm/radeon/kms/evergreen: add some additional safe regs v2 drm/radeon/r600: fix tiling issues in CS checker. drm/i915: Move gpu_write_list to per-ring drm/i915: Invalidate the to-ring, flush the old-ring when updating domains drm/i915/ringbuffer: Write the value passed in to the tail register agp/intel: Restore valid PTE bit for Sandybridge after bdd3072 drm/i915: Fix flushing regression from 9af90d19f drm/i915/sdvo: Remove unused encoding member i915: enable AVI infoframe for intel_hdmi.c [v4] drm/i915: Fix current fb blocking for page flip drm/i915: IS_IRONLAKE is synonymous with gen == 5 ... Fix up conflicts in - drivers/gpu/drm/i915/{i915_gem.c, i915/intel_overlay.c}: due to the new simplified stack-based kmap_atomic() interface - drivers/gpu/drm/vmwgfx/vmwgfx_drv.c: added .llseek entry due to BKL removal cleanups.
2010-10-27 09:57:59 +08:00
io_mapping_unmap_atomic(regs);
}
struct intel_overlay_error_state *
intel_overlay_capture_error_state(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_overlay *overlay = dev_priv->overlay;
struct intel_overlay_error_state *error;
struct overlay_registers __iomem *regs;
if (!overlay || !overlay->active)
return NULL;
error = kmalloc(sizeof(*error), GFP_ATOMIC);
if (error == NULL)
return NULL;
error->dovsta = I915_READ(DOVSTA);
error->isr = I915_READ(ISR);
if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
error->base = (__force long)overlay->reg_bo->phys_handle->vaddr;
else
error->base = i915_gem_obj_ggtt_offset(overlay->reg_bo);
regs = intel_overlay_map_regs_atomic(overlay);
if (!regs)
goto err;
memcpy_fromio(&error->regs, regs, sizeof(struct overlay_registers));
Merge branch 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6 * 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (476 commits) vmwgfx: Implement a proper GMR eviction mechanism drm/radeon/kms: fix r6xx/7xx 1D tiling CS checker v2 drm/radeon/kms: properly compute group_size on 6xx/7xx drm/radeon/kms: fix 2D tile height alignment in the r600 CS checker drm/radeon/kms/evergreen: set the clear state to the blit state drm/radeon/kms: don't poll dac load detect. gpu: Add Intel GMA500(Poulsbo) Stub Driver drm/radeon/kms: MC vram map needs to be >= pci aperture size drm/radeon/kms: implement display watermark support for evergreen drm/radeon/kms/evergreen: add some additional safe regs v2 drm/radeon/r600: fix tiling issues in CS checker. drm/i915: Move gpu_write_list to per-ring drm/i915: Invalidate the to-ring, flush the old-ring when updating domains drm/i915/ringbuffer: Write the value passed in to the tail register agp/intel: Restore valid PTE bit for Sandybridge after bdd3072 drm/i915: Fix flushing regression from 9af90d19f drm/i915/sdvo: Remove unused encoding member i915: enable AVI infoframe for intel_hdmi.c [v4] drm/i915: Fix current fb blocking for page flip drm/i915: IS_IRONLAKE is synonymous with gen == 5 ... Fix up conflicts in - drivers/gpu/drm/i915/{i915_gem.c, i915/intel_overlay.c}: due to the new simplified stack-based kmap_atomic() interface - drivers/gpu/drm/vmwgfx/vmwgfx_drv.c: added .llseek entry due to BKL removal cleanups.
2010-10-27 09:57:59 +08:00
intel_overlay_unmap_regs_atomic(overlay, regs);
return error;
err:
kfree(error);
return NULL;
}
void
intel_overlay_print_error_state(struct drm_i915_error_state_buf *m,
struct intel_overlay_error_state *error)
{
i915_error_printf(m, "Overlay, status: 0x%08x, interrupt: 0x%08x\n",
error->dovsta, error->isr);
i915_error_printf(m, " Register file at 0x%08lx:\n",
error->base);
#define P(x) i915_error_printf(m, " " #x ": 0x%08x\n", error->regs.x)
P(OBUF_0Y);
P(OBUF_1Y);
P(OBUF_0U);
P(OBUF_0V);
P(OBUF_1U);
P(OBUF_1V);
P(OSTRIDE);
P(YRGB_VPH);
P(UV_VPH);
P(HORZ_PH);
P(INIT_PHS);
P(DWINPOS);
P(DWINSZ);
P(SWIDTH);
P(SWIDTHSW);
P(SHEIGHT);
P(YRGBSCALE);
P(UVSCALE);
P(OCLRC0);
P(OCLRC1);
P(DCLRKV);
P(DCLRKM);
P(SCLRKVH);
P(SCLRKVL);
P(SCLRKEN);
P(OCONFIG);
P(OCMD);
P(OSTART_0Y);
P(OSTART_1Y);
P(OSTART_0U);
P(OSTART_0V);
P(OSTART_1U);
P(OSTART_1V);
P(OTILEOFF_0Y);
P(OTILEOFF_1Y);
P(OTILEOFF_0U);
P(OTILEOFF_0V);
P(OTILEOFF_1U);
P(OTILEOFF_1V);
P(FASTHSCALE);
P(UVSCALEV);
#undef P
}