linux/drivers/gpu/drm/nouveau/nva3_copy.c
Ben Skeggs 6c320fef58 drm/nouveau: pass flag to engine fini() method on suspend
It may not be necessary to fail in certain cases (such as failing to idle)
on module unload, whereas on suspend it's important to ensure a consistent
state can be restored on resume.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2011-07-25 09:43:22 +10:00

227 lines
6.3 KiB
C

/*
* Copyright 2011 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 <linux/firmware.h>
#include "drmP.h"
#include "nouveau_drv.h"
#include "nouveau_util.h"
#include "nouveau_vm.h"
#include "nouveau_ramht.h"
#include "nva3_copy.fuc.h"
struct nva3_copy_engine {
struct nouveau_exec_engine base;
};
static int
nva3_copy_context_new(struct nouveau_channel *chan, int engine)
{
struct drm_device *dev = chan->dev;
struct drm_nouveau_private *dev_priv = dev->dev_private;
struct nouveau_gpuobj *ramin = chan->ramin;
struct nouveau_gpuobj *ctx = NULL;
int ret;
NV_DEBUG(dev, "ch%d\n", chan->id);
ret = nouveau_gpuobj_new(dev, chan, 256, 0, NVOBJ_FLAG_ZERO_ALLOC |
NVOBJ_FLAG_ZERO_FREE, &ctx);
if (ret)
return ret;
nv_wo32(ramin, 0xc0, 0x00190000);
nv_wo32(ramin, 0xc4, ctx->vinst + ctx->size - 1);
nv_wo32(ramin, 0xc8, ctx->vinst);
nv_wo32(ramin, 0xcc, 0x00000000);
nv_wo32(ramin, 0xd0, 0x00000000);
nv_wo32(ramin, 0xd4, 0x00000000);
dev_priv->engine.instmem.flush(dev);
atomic_inc(&chan->vm->engref[engine]);
chan->engctx[engine] = ctx;
return 0;
}
static int
nva3_copy_object_new(struct nouveau_channel *chan, int engine,
u32 handle, u16 class)
{
struct nouveau_gpuobj *ctx = chan->engctx[engine];
/* fuc engine doesn't need an object, our ramht code does.. */
ctx->engine = 3;
ctx->class = class;
return nouveau_ramht_insert(chan, handle, ctx);
}
static void
nva3_copy_context_del(struct nouveau_channel *chan, int engine)
{
struct nouveau_gpuobj *ctx = chan->engctx[engine];
struct drm_device *dev = chan->dev;
u32 inst;
inst = (chan->ramin->vinst >> 12);
inst |= 0x40000000;
/* disable fifo access */
nv_wr32(dev, 0x104048, 0x00000000);
/* mark channel as unloaded if it's currently active */
if (nv_rd32(dev, 0x104050) == inst)
nv_mask(dev, 0x104050, 0x40000000, 0x00000000);
/* mark next channel as invalid if it's about to be loaded */
if (nv_rd32(dev, 0x104054) == inst)
nv_mask(dev, 0x104054, 0x40000000, 0x00000000);
/* restore fifo access */
nv_wr32(dev, 0x104048, 0x00000003);
for (inst = 0xc0; inst <= 0xd4; inst += 4)
nv_wo32(chan->ramin, inst, 0x00000000);
nouveau_gpuobj_ref(NULL, &ctx);
atomic_dec(&chan->vm->engref[engine]);
chan->engctx[engine] = ctx;
}
static void
nva3_copy_tlb_flush(struct drm_device *dev, int engine)
{
nv50_vm_flush_engine(dev, 0x0d);
}
static int
nva3_copy_init(struct drm_device *dev, int engine)
{
int i;
nv_mask(dev, 0x000200, 0x00002000, 0x00000000);
nv_mask(dev, 0x000200, 0x00002000, 0x00002000);
nv_wr32(dev, 0x104014, 0xffffffff); /* disable all interrupts */
/* upload ucode */
nv_wr32(dev, 0x1041c0, 0x01000000);
for (i = 0; i < sizeof(nva3_pcopy_data) / 4; i++)
nv_wr32(dev, 0x1041c4, nva3_pcopy_data[i]);
nv_wr32(dev, 0x104180, 0x01000000);
for (i = 0; i < sizeof(nva3_pcopy_code) / 4; i++) {
if ((i & 0x3f) == 0)
nv_wr32(dev, 0x104188, i >> 6);
nv_wr32(dev, 0x104184, nva3_pcopy_code[i]);
}
/* start it running */
nv_wr32(dev, 0x10410c, 0x00000000);
nv_wr32(dev, 0x104104, 0x00000000); /* ENTRY */
nv_wr32(dev, 0x104100, 0x00000002); /* TRIGGER */
return 0;
}
static int
nva3_copy_fini(struct drm_device *dev, int engine, bool suspend)
{
nv_mask(dev, 0x104048, 0x00000003, 0x00000000);
/* trigger fuc context unload */
nv_wait(dev, 0x104008, 0x0000000c, 0x00000000);
nv_mask(dev, 0x104054, 0x40000000, 0x00000000);
nv_wr32(dev, 0x104000, 0x00000008);
nv_wait(dev, 0x104008, 0x00000008, 0x00000000);
nv_wr32(dev, 0x104014, 0xffffffff);
return 0;
}
static struct nouveau_enum nva3_copy_isr_error_name[] = {
{ 0x0001, "ILLEGAL_MTHD" },
{ 0x0002, "INVALID_ENUM" },
{ 0x0003, "INVALID_BITFIELD" },
{}
};
static void
nva3_copy_isr(struct drm_device *dev)
{
u32 dispatch = nv_rd32(dev, 0x10401c);
u32 stat = nv_rd32(dev, 0x104008) & dispatch & ~(dispatch >> 16);
u32 inst = nv_rd32(dev, 0x104050) & 0x3fffffff;
u32 ssta = nv_rd32(dev, 0x104040) & 0x0000ffff;
u32 addr = nv_rd32(dev, 0x104040) >> 16;
u32 mthd = (addr & 0x07ff) << 2;
u32 subc = (addr & 0x3800) >> 11;
u32 data = nv_rd32(dev, 0x104044);
int chid = nv50_graph_isr_chid(dev, inst);
if (stat & 0x00000040) {
NV_INFO(dev, "PCOPY: DISPATCH_ERROR [");
nouveau_enum_print(nva3_copy_isr_error_name, ssta);
printk("] ch %d [0x%08x] subc %d mthd 0x%04x data 0x%08x\n",
chid, inst, subc, mthd, data);
nv_wr32(dev, 0x104004, 0x00000040);
stat &= ~0x00000040;
}
if (stat) {
NV_INFO(dev, "PCOPY: unhandled intr 0x%08x\n", stat);
nv_wr32(dev, 0x104004, stat);
}
nv50_fb_vm_trap(dev, 1);
}
static void
nva3_copy_destroy(struct drm_device *dev, int engine)
{
struct nva3_copy_engine *pcopy = nv_engine(dev, engine);
nouveau_irq_unregister(dev, 22);
NVOBJ_ENGINE_DEL(dev, COPY0);
kfree(pcopy);
}
int
nva3_copy_create(struct drm_device *dev)
{
struct nva3_copy_engine *pcopy;
pcopy = kzalloc(sizeof(*pcopy), GFP_KERNEL);
if (!pcopy)
return -ENOMEM;
pcopy->base.destroy = nva3_copy_destroy;
pcopy->base.init = nva3_copy_init;
pcopy->base.fini = nva3_copy_fini;
pcopy->base.context_new = nva3_copy_context_new;
pcopy->base.context_del = nva3_copy_context_del;
pcopy->base.object_new = nva3_copy_object_new;
pcopy->base.tlb_flush = nva3_copy_tlb_flush;
nouveau_irq_register(dev, 22, nva3_copy_isr);
NVOBJ_ENGINE_ADD(dev, COPY0, &pcopy->base);
NVOBJ_CLASS(dev, 0x85b5, COPY0);
return 0;
}