linux/drivers/gpu/drm/nouveau/nv84_fence.c

260 lines
7.0 KiB
C
Raw Normal View History

/*
* Copyright 2012 Red Hat Inc.
*
* 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 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 COPYRIGHT HOLDER(S) OR AUTHOR(S) 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: Ben Skeggs
*/
#include "nouveau_drv.h"
#include "nouveau_dma.h"
#include "nouveau_fence.h"
#include "nv50_display.h"
static int
nv84_fence_emit32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
{
int ret = RING_SPACE(chan, 8);
if (ret == 0) {
BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
OUT_RING (chan, chan->vram.handle);
BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 5);
OUT_RING (chan, upper_32_bits(virtual));
OUT_RING (chan, lower_32_bits(virtual));
OUT_RING (chan, sequence);
OUT_RING (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG);
OUT_RING (chan, 0x00000000);
FIRE_RING (chan);
}
return ret;
}
static int
nv84_fence_sync32(struct nouveau_channel *chan, u64 virtual, u32 sequence)
{
int ret = RING_SPACE(chan, 7);
if (ret == 0) {
BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1);
OUT_RING (chan, chan->vram.handle);
BEGIN_NV04(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4);
OUT_RING (chan, upper_32_bits(virtual));
OUT_RING (chan, lower_32_bits(virtual));
OUT_RING (chan, sequence);
OUT_RING (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_GEQUAL);
FIRE_RING (chan);
}
return ret;
}
static int
nv84_fence_emit(struct nouveau_fence *fence)
{
struct nouveau_channel *chan = fence->channel;
struct nv84_fence_chan *fctx = chan->fence;
u64 addr = chan->chid * 16;
if (fence->sysmem)
addr += fctx->vma_gart.offset;
else
addr += fctx->vma.offset;
return fctx->base.emit32(chan, addr, fence->base.seqno);
}
static int
nv84_fence_sync(struct nouveau_fence *fence,
struct nouveau_channel *prev, struct nouveau_channel *chan)
{
struct nv84_fence_chan *fctx = chan->fence;
u64 addr = prev->chid * 16;
if (fence->sysmem)
addr += fctx->vma_gart.offset;
else
addr += fctx->vma.offset;
return fctx->base.sync32(chan, addr, fence->base.seqno);
}
static u32
nv84_fence_read(struct nouveau_channel *chan)
{
struct nv84_fence_priv *priv = chan->drm->fence;
return nouveau_bo_rd32(priv->bo, chan->chid * 16/4);
}
static void
nv84_fence_context_del(struct nouveau_channel *chan)
{
struct nv84_fence_priv *priv = chan->drm->fence;
struct nv84_fence_chan *fctx = chan->fence;
nouveau_bo_wr32(priv->bo, chan->chid * 16 / 4, fctx->base.sequence);
mutex_lock(&priv->mutex);
nouveau_bo_vma_del(priv->bo, &fctx->vma_gart);
nouveau_bo_vma_del(priv->bo, &fctx->vma);
mutex_unlock(&priv->mutex);
nouveau_fence_context_del(&fctx->base);
chan->fence = NULL;
nouveau_fence_context_free(&fctx->base);
}
int
nv84_fence_context_new(struct nouveau_channel *chan)
{
struct nouveau_cli *cli = (void *)chan->user.client;
struct nv84_fence_priv *priv = chan->drm->fence;
struct nv84_fence_chan *fctx;
int ret;
fctx = chan->fence = kzalloc(sizeof(*fctx), GFP_KERNEL);
if (!fctx)
return -ENOMEM;
nouveau_fence_context_new(chan, &fctx->base);
fctx->base.emit = nv84_fence_emit;
fctx->base.sync = nv84_fence_sync;
fctx->base.read = nv84_fence_read;
fctx->base.emit32 = nv84_fence_emit32;
fctx->base.sync32 = nv84_fence_sync32;
fctx->base.sequence = nv84_fence_read(chan);
mutex_lock(&priv->mutex);
ret = nouveau_bo_vma_add(priv->bo, cli->vm, &fctx->vma);
if (ret == 0) {
ret = nouveau_bo_vma_add(priv->bo_gart, cli->vm,
&fctx->vma_gart);
}
mutex_unlock(&priv->mutex);
if (ret)
nv84_fence_context_del(chan);
return ret;
}
static bool
nv84_fence_suspend(struct nouveau_drm *drm)
{
struct nv84_fence_priv *priv = drm->fence;
int i;
priv->suspend = vmalloc(priv->base.contexts * sizeof(u32));
if (priv->suspend) {
for (i = 0; i < priv->base.contexts; i++)
priv->suspend[i] = nouveau_bo_rd32(priv->bo, i*4);
}
return priv->suspend != NULL;
}
static void
nv84_fence_resume(struct nouveau_drm *drm)
{
struct nv84_fence_priv *priv = drm->fence;
int i;
if (priv->suspend) {
for (i = 0; i < priv->base.contexts; i++)
nouveau_bo_wr32(priv->bo, i*4, priv->suspend[i]);
vfree(priv->suspend);
priv->suspend = NULL;
}
}
static void
nv84_fence_destroy(struct nouveau_drm *drm)
{
struct nv84_fence_priv *priv = drm->fence;
nouveau_bo_unmap(priv->bo_gart);
if (priv->bo_gart)
nouveau_bo_unpin(priv->bo_gart);
nouveau_bo_ref(NULL, &priv->bo_gart);
nouveau_bo_unmap(priv->bo);
if (priv->bo)
nouveau_bo_unpin(priv->bo);
nouveau_bo_ref(NULL, &priv->bo);
drm->fence = NULL;
kfree(priv);
}
int
nv84_fence_create(struct nouveau_drm *drm)
{
struct nvkm_fifo *fifo = nvxx_fifo(&drm->client.device);
struct nv84_fence_priv *priv;
u32 domain;
int ret;
priv = drm->fence = kzalloc(sizeof(*priv), GFP_KERNEL);
if (!priv)
return -ENOMEM;
priv->base.dtor = nv84_fence_destroy;
priv->base.suspend = nv84_fence_suspend;
priv->base.resume = nv84_fence_resume;
priv->base.context_new = nv84_fence_context_new;
priv->base.context_del = nv84_fence_context_del;
priv->base.contexts = fifo->nr;
dma-buf: Rename struct fence to dma_fence I plan to usurp the short name of struct fence for a core kernel struct, and so I need to rename the specialised fence/timeline for DMA operations to make room. A consensus was reached in https://lists.freedesktop.org/archives/dri-devel/2016-July/113083.html that making clear this fence applies to DMA operations was a good thing. Since then the patch has grown a bit as usage increases, so hopefully it remains a good thing! (v2...: rebase, rerun spatch) v3: Compile on msm, spotted a manual fixup that I broke. v4: Try again for msm, sorry Daniel coccinelle script: @@ @@ - struct fence + struct dma_fence @@ @@ - struct fence_ops + struct dma_fence_ops @@ @@ - struct fence_cb + struct dma_fence_cb @@ @@ - struct fence_array + struct dma_fence_array @@ @@ - enum fence_flag_bits + enum dma_fence_flag_bits @@ @@ ( - fence_init + dma_fence_init | - fence_release + dma_fence_release | - fence_free + dma_fence_free | - fence_get + dma_fence_get | - fence_get_rcu + dma_fence_get_rcu | - fence_put + dma_fence_put | - fence_signal + dma_fence_signal | - fence_signal_locked + dma_fence_signal_locked | - fence_default_wait + dma_fence_default_wait | - fence_add_callback + dma_fence_add_callback | - fence_remove_callback + dma_fence_remove_callback | - fence_enable_sw_signaling + dma_fence_enable_sw_signaling | - fence_is_signaled_locked + dma_fence_is_signaled_locked | - fence_is_signaled + dma_fence_is_signaled | - fence_is_later + dma_fence_is_later | - fence_later + dma_fence_later | - fence_wait_timeout + dma_fence_wait_timeout | - fence_wait_any_timeout + dma_fence_wait_any_timeout | - fence_wait + dma_fence_wait | - fence_context_alloc + dma_fence_context_alloc | - fence_array_create + dma_fence_array_create | - to_fence_array + to_dma_fence_array | - fence_is_array + dma_fence_is_array | - trace_fence_emit + trace_dma_fence_emit | - FENCE_TRACE + DMA_FENCE_TRACE | - FENCE_WARN + DMA_FENCE_WARN | - FENCE_ERR + DMA_FENCE_ERR ) ( ... ) Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Acked-by: Sumit Semwal <sumit.semwal@linaro.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161025120045.28839-1-chris@chris-wilson.co.uk
2016-10-25 20:00:45 +08:00
priv->base.context_base = dma_fence_context_alloc(priv->base.contexts);
priv->base.uevent = true;
mutex_init(&priv->mutex);
/* Use VRAM if there is any ; otherwise fallback to system memory */
domain = drm->client.device.info.ram_size != 0 ? TTM_PL_FLAG_VRAM :
/*
* fences created in sysmem must be non-cached or we
* will lose CPU/GPU coherency!
*/
TTM_PL_FLAG_TT | TTM_PL_FLAG_UNCACHED;
ret = nouveau_bo_new(&drm->client, 16 * priv->base.contexts, 0,
domain, 0, 0, NULL, NULL, &priv->bo);
if (ret == 0) {
ret = nouveau_bo_pin(priv->bo, domain, false);
if (ret == 0) {
ret = nouveau_bo_map(priv->bo);
if (ret)
nouveau_bo_unpin(priv->bo);
}
if (ret)
nouveau_bo_ref(NULL, &priv->bo);
}
if (ret == 0)
ret = nouveau_bo_new(&drm->client, 16 * priv->base.contexts, 0,
TTM_PL_FLAG_TT | TTM_PL_FLAG_UNCACHED, 0,
0, NULL, NULL, &priv->bo_gart);
if (ret == 0) {
ret = nouveau_bo_pin(priv->bo_gart, TTM_PL_FLAG_TT, false);
if (ret == 0) {
ret = nouveau_bo_map(priv->bo_gart);
if (ret)
nouveau_bo_unpin(priv->bo_gart);
}
if (ret)
nouveau_bo_ref(NULL, &priv->bo_gart);
}
if (ret)
nv84_fence_destroy(drm);
return ret;
}