drm/nouveau/core: pull in most of the new core infrastructure

This commit provides most of the infrastructure to support a major overhaul
of Nouveau's internals coming in the following commits.  This work aims to
take all the things we've learned over the last several years, and turn that
into a cleaner architecture that's more maintainable going forward.

RAMHT and MM bits of the new core have been left out for the moment, and
will be pulled in as I go through the process of porting the code to
become either subdev or engine modules.

There are several main goals I wanted to achieve through this work:

-- Reduce complexity

The goal here was to make each component of the driver as independent as
possible, which will ease maintainability and readability, and provide a
good base for resetting locked up GPU units in the future.

-- Better tracking of GPU units that are required at any given time

This is for future PM work, we'll be able to tell exactly what parts of the
GPU we need powered at any given point (etc).

-- Expose all available NVIDIA GPUs to the client

In order to support things such as multi-GPU channels, we want to be able
to expose all the NVIDIA GPUs to the client over a single file descriptor
so it can send a single push buffer to multiple GPUs.

-- Untangle the core hardware support code from the DRM implementation

This happened initially as an unexpected side-effect of developing the
initial core infrastructure in userspace, but it turned into a goal of
the whole project.  Initial benefits will be the availablility of a
number of userspace tools and tests using the same code as the driver
itself, but will also be important as I look into some virtualisation
ideas.

v2: Ben Skeggs <bskeggs@redhat.com>
- fix duplicate assignments noticed by clang
- implement some forgotten yelling in error path
- ensure 64-bit engine mask is used everywhere

v3: Marcin Slusarz <marcin.slusarz@gmail.com>
- sparse fixes
- inline nv_printk into nv_assert to prevent recursive inlining issues

v4: Ben Skeggs <bskeggs@redhat.com>
- fixed minor memory leak on gpuobj destruction

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

drivers/gpu/drm/nouveau/Makefile

@@ -2,13 +2,31 @@
 # Makefile for the drm device driver.  This driver provides support for the
 # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
 
-ccflags-y := -Iinclude/drm
+ccflags-y := -Iinclude/drm -DCONFIG_NOUVEAU_DEBUG=7 -DCONFIG_NOUVEAU_DEBUG_DEFAULT=3
 ccflags-y += -I$(src)/core/include
 ccflags-y += -I$(src)
 
-nouveau-y := core/core/mm.o
+nouveau-y := core/core/client.o
+nouveau-y += core/core/engine.o
+nouveau-y += core/core/handle.o
+nouveau-y += core/core/mm.o
+nouveau-y += core/core/namedb.o
+nouveau-y += core/core/object.o
+nouveau-y += core/core/option.o
+nouveau-y += core/core/parent.o
+nouveau-y += core/core/printk.o
 nouveau-y += core/core/ramht.o
+nouveau-y += core/core/subdev.o
 
+nouveau-y += core/subdev/device/base.o
+nouveau-y += core/subdev/device/nv04.o
+nouveau-y += core/subdev/device/nv10.o
+nouveau-y += core/subdev/device/nv20.o
+nouveau-y += core/subdev/device/nv30.o
+nouveau-y += core/subdev/device/nv40.o
+nouveau-y += core/subdev/device/nv50.o
+nouveau-y += core/subdev/device/nvc0.o
+nouveau-y += core/subdev/device/nve0.o
 nouveau-y += core/subdev/fb/nv04.o
 nouveau-y += core/subdev/fb/nv10.o
 nouveau-y += core/subdev/fb/nv20.o

drivers/gpu/drm/nouveau/core/core/client.c

@@ -0,0 +1,103 @@
+/*
+ * 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 <core/object.h>
+#include <core/client.h>
+#include <core/handle.h>
+#include <core/option.h>
+
+#include <subdev/device.h>
+
+static void
+nouveau_client_dtor(struct nouveau_object *object)
+{
+	struct nouveau_client *client = (void *)object;
+	nouveau_object_ref(NULL, &client->device);
+	nouveau_handle_destroy(client->root);
+	nouveau_namedb_destroy(&client->base);
+}
+
+static struct nouveau_oclass
+nouveau_client_oclass = {
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.dtor = nouveau_client_dtor,
+	},
+};
+
+int
+nouveau_client_create_(u32 name, u64 devname, const char *cfg, const char *dbg,
+		       int length, void **pobject)
+{
+	struct nouveau_object *device;
+	struct nouveau_client *client;
+	int ret;
+
+	device = (void *)nouveau_device_find(devname);
+	if (!device)
+		return -ENODEV;
+
+	ret = nouveau_namedb_create_(NULL, NULL, &nouveau_client_oclass,
+				     NV_CLIENT_CLASS, nouveau_device_sclass,
+				     0, length, pobject);
+	client = *pobject;
+	if (ret)
+		return ret;
+
+	ret = nouveau_handle_create(nv_object(client), ~0, ~0,
+				    nv_object(client), &client->root);
+	if (ret) {
+		nouveau_namedb_destroy(&client->base);
+		return ret;
+	}
+
+	/* prevent init/fini being called, os in in charge of this */
+	atomic_set(&nv_object(client)->usecount, 2);
+
+	nouveau_object_ref(device, &client->device);
+	client->handle = name;
+	client->debug = nouveau_dbgopt(dbg, "CLIENT");
+	return 0;
+}
+
+int
+nouveau_client_init(struct nouveau_client *client)
+{
+	int ret;
+	nv_debug(client, "init running\n");
+	ret = nouveau_handle_init(client->root);
+	nv_debug(client, "init completed with %d\n", ret);
+	return ret;
+}
+
+int
+nouveau_client_fini(struct nouveau_client *client, bool suspend)
+{
+	const char *name[2] = { "fini", "suspend" };
+	int ret;
+
+	nv_debug(client, "%s running\n", name[suspend]);
+	ret = nouveau_handle_fini(client->root, suspend);
+	nv_debug(client, "%s completed with %d\n", name[suspend], ret);
+	return ret;
+}

drivers/gpu/drm/nouveau/core/core/engctx.c

@@ -0,0 +1,230 @@
+/*
+ * 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 <core/object.h>
+#include <core/namedb.h>
+#include <core/handle.h>
+#include <core/client.h>
+#include <core/engctx.h>
+
+#include <subdev/vm.h>
+
+int
+nouveau_engctx_create_(struct nouveau_object *parent,
+		       struct nouveau_object *engobj,
+		       struct nouveau_oclass *oclass,
+		       struct nouveau_object *pargpu,
+		       u32 size, u32 align, u32 flags,
+		       int length, void **pobject)
+{
+	struct nouveau_client *client = nouveau_client(parent);
+	struct nouveau_engine *engine = nv_engine(engobj);
+	struct nouveau_subdev *subdev = nv_subdev(engine);
+	struct nouveau_engctx *engctx;
+	struct nouveau_object *ctxpar;
+	int ret;
+
+	/* use existing context for the engine if one is available */
+	mutex_lock(&subdev->mutex);
+	list_for_each_entry(engctx, &engine->contexts, head) {
+		ctxpar = nv_pclass(nv_object(engctx), NV_PARENT_CLASS);
+		if (ctxpar == parent) {
+			atomic_inc(&nv_object(engctx)->refcount);
+			*pobject = engctx;
+			mutex_unlock(&subdev->mutex);
+			return 1;
+		}
+	}
+	mutex_unlock(&subdev->mutex);
+
+	if (size) {
+		ret = nouveau_gpuobj_create_(parent, engobj, oclass,
+					     NV_ENGCTX_CLASS,
+					     pargpu, size, align, flags,
+					     length, pobject);
+	} else {
+		ret = nouveau_object_create_(parent, engobj, oclass,
+					     NV_ENGCTX_CLASS, length, pobject);
+	}
+
+	engctx = *pobject;
+	if (engctx && client->vm)
+		atomic_inc(&client->vm->engref[nv_engidx(engobj)]);
+	if (ret)
+		return ret;
+
+	list_add(&engctx->head, &engine->contexts);
+	return 0;
+}
+
+void
+nouveau_engctx_destroy(struct nouveau_engctx *engctx)
+{
+	struct nouveau_object *engine = nv_object(engctx)->engine;
+	struct nouveau_client *client = nouveau_client(engctx);
+
+	nouveau_gpuobj_unmap(&engctx->vma);
+	list_del(&engctx->head);
+	if (client->vm)
+		atomic_dec(&client->vm->engref[nv_engidx(engine)]);
+
+	if (engctx->base.size)
+		nouveau_gpuobj_destroy(&engctx->base);
+	else
+		nouveau_object_destroy(&engctx->base.base);
+}
+
+int
+nouveau_engctx_init(struct nouveau_engctx *engctx)
+{
+	struct nouveau_object *object = nv_object(engctx);
+	struct nouveau_subdev *subdev = nv_subdev(object->engine);
+	struct nouveau_object *parent;
+	struct nouveau_subdev *pardev;
+	int ret;
+
+	ret = nouveau_gpuobj_init(&engctx->base);
+	if (ret)
+		return ret;
+
+	parent = nv_pclass(object->parent, NV_PARENT_CLASS);
+	pardev = nv_subdev(parent->engine);
+	if (nv_parent(parent)->context_attach) {
+		mutex_lock(&pardev->mutex);
+		ret = nv_parent(parent)->context_attach(parent, object);
+		mutex_unlock(&pardev->mutex);
+	}
+
+	if (ret) {
+		nv_error(parent, "failed to attach %s context, %d\n",
+			 subdev->name, ret);
+		return ret;
+	}
+
+	nv_debug(parent, "attached %s context\n", subdev->name);
+	return 0;
+}
+
+int
+nouveau_engctx_fini(struct nouveau_engctx *engctx, bool suspend)
+{
+	struct nouveau_object *object = nv_object(engctx);
+	struct nouveau_subdev *subdev = nv_subdev(object->engine);
+	struct nouveau_object *parent;
+	struct nouveau_subdev *pardev;
+	int ret = 0;
+
+	parent = nv_pclass(object->parent, NV_PARENT_CLASS);
+	pardev = nv_subdev(parent->engine);
+	if (nv_parent(parent)->context_detach) {
+		mutex_lock(&pardev->mutex);
+		ret = nv_parent(parent)->context_detach(parent, suspend, object);
+		mutex_unlock(&pardev->mutex);
+	}
+
+	if (ret) {
+		nv_error(parent, "failed to detach %s context, %d\n",
+			 subdev->name, ret);
+		return ret;
+	}
+
+	nv_debug(parent, "detached %s context\n", subdev->name);
+	return nouveau_gpuobj_fini(&engctx->base, suspend);
+}
+
+void
+_nouveau_engctx_dtor(struct nouveau_object *object)
+{
+	nouveau_engctx_destroy(nv_engctx(object));
+}
+
+int
+_nouveau_engctx_init(struct nouveau_object *object)
+{
+	return nouveau_engctx_init(nv_engctx(object));
+}
+
+
+int
+_nouveau_engctx_fini(struct nouveau_object *object, bool suspend)
+{
+	return nouveau_engctx_fini(nv_engctx(object), suspend);
+}
+
+struct nouveau_object *
+nouveau_engctx_lookup(struct nouveau_engine *engine, u64 addr)
+{
+	struct nouveau_engctx *engctx;
+
+	list_for_each_entry(engctx, &engine->contexts, head) {
+		if (engctx->base.size &&
+		    nv_gpuobj(engctx)->addr == addr)
+			return nv_object(engctx);
+	}
+
+	return NULL;
+}
+
+struct nouveau_handle *
+nouveau_engctx_lookup_class(struct nouveau_engine *engine, u64 addr, u16 oclass)
+{
+	struct nouveau_object *engctx = nouveau_engctx_lookup(engine, addr);
+	struct nouveau_namedb *namedb;
+
+	if (engctx && (namedb = (void *)nv_pclass(engctx, NV_NAMEDB_CLASS)))
+		return nouveau_namedb_get_class(namedb, oclass);
+
+	return NULL;
+}
+
+struct nouveau_handle *
+nouveau_engctx_lookup_vinst(struct nouveau_engine *engine, u64 addr, u64 vinst)
+{
+	struct nouveau_object *engctx = nouveau_engctx_lookup(engine, addr);
+	struct nouveau_namedb *namedb;
+
+	if (engctx && (namedb = (void *)nv_pclass(engctx, NV_NAMEDB_CLASS)))
+		return nouveau_namedb_get_vinst(namedb, vinst);
+
+	return NULL;
+}
+
+struct nouveau_handle *
+nouveau_engctx_lookup_cinst(struct nouveau_engine *engine, u64 addr, u32 cinst)
+{
+	struct nouveau_object *engctx = nouveau_engctx_lookup(engine, addr);
+	struct nouveau_namedb *namedb;
+
+	if (engctx && (namedb = (void *)nv_pclass(engctx, NV_NAMEDB_CLASS)))
+		return nouveau_namedb_get_cinst(namedb, cinst);
+
+	return NULL;
+}
+
+void
+nouveau_engctx_handle_put(struct nouveau_handle *handle)
+{
+	if (handle)
+		nouveau_namedb_put(handle);
+}