summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/nvgpu/vgpu/css_vgpu.c
blob: d18ed10091d9504bb68e93ec66e389c2a13413cb (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
/*
 * Copyright (c) 2016-2017, NVIDIA CORPORATION.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
 * more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */
#if defined(CONFIG_GK20A_CYCLE_STATS)

#include <linux/tegra-ivc.h>
#include <linux/tegra_vgpu.h>

#include "gk20a/gk20a.h"
#include "gk20a/channel_gk20a.h"
#include "gk20a/platform_gk20a.h"
#include "gk20a/css_gr_gk20a.h"
#include "vgpu.h"

static struct tegra_hv_ivm_cookie *css_cookie;

static int vgpu_css_init_snapshot_buffer(struct gr_gk20a *gr)
{
	struct gk20a *g = gr->g;
	struct device *dev = g->dev;
	struct gk20a_cs_snapshot *data = gr->cs_data;
	struct device_node *np = dev->of_node;
	struct of_phandle_args args;
	struct device_node *hv_np;
	void *buf = NULL;
	u32 mempool;
	int err;

	gk20a_dbg_fn("");

	if (data->hw_snapshot)
		return 0;

	err = of_parse_phandle_with_fixed_args(np,
			"mempool-css", 1, 0, &args);
	if (err) {
		nvgpu_info(g, "dt missing mempool-css");
		goto fail;
	}

	hv_np = args.np;
	mempool = args.args[0];
	css_cookie = tegra_hv_mempool_reserve(hv_np, mempool);
	if (IS_ERR(css_cookie)) {
		nvgpu_info(g,
			"mempool  %u reserve failed", mempool);
		err = -EINVAL;
		goto fail;
	}

	/* Make sure buffer size is large enough */
	if (css_cookie->size < CSS_MIN_HW_SNAPSHOT_SIZE) {
		nvgpu_info(g, "mempool size %lld too small",
			css_cookie->size);
		err = -ENOMEM;
		goto fail;
	}

#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 18, 0)
	buf = ioremap_cached(css_cookie->ipa, css_cookie->size);
#else
	buf = ioremap_cache(css_cookie->ipa, css_cookie->size);
#endif
	if (!buf) {
		nvgpu_info(g, "ioremap_cache failed");
		err = -EINVAL;
		goto fail;
	}

	data->hw_snapshot = buf;
	data->hw_end = data->hw_snapshot +
		css_cookie->size / sizeof(struct gk20a_cs_snapshot_fifo_entry);
	data->hw_get = data->hw_snapshot;
	memset(data->hw_snapshot, 0xff, css_cookie->size);
	return 0;
fail:
	if (!IS_ERR_OR_NULL(css_cookie))
		tegra_hv_mempool_unreserve(css_cookie);
	return err;
}

static void vgpu_css_release_snapshot_buffer(struct gr_gk20a *gr)
{
	struct gk20a_cs_snapshot *data = gr->cs_data;

	if (!data->hw_snapshot)
		return;

	iounmap(data->hw_snapshot);
	data->hw_snapshot = NULL;

	tegra_hv_mempool_unreserve(css_cookie);

	gk20a_dbg_info("cyclestats(vgpu): buffer for snapshots released\n");
}

static int vgpu_css_flush_snapshots(struct channel_gk20a *ch,
			u32 *pending, bool *hw_overflow)
{
	struct gk20a *g = ch->g;
	struct tegra_vgpu_cmd_msg msg = {};
	struct tegra_vgpu_channel_cyclestats_snapshot_params *p;
	struct gr_gk20a *gr = &g->gr;
	struct gk20a_cs_snapshot *data = gr->cs_data;
	int err;

	gk20a_dbg_fn("");

	msg.cmd = TEGRA_VGPU_CMD_CHANNEL_CYCLESTATS_SNAPSHOT;
	msg.handle = vgpu_get_handle(g);
	p = &msg.params.cyclestats_snapshot;
	p->handle = ch->virt_ctx;
	p->subcmd = NVGPU_IOCTL_CHANNEL_CYCLE_STATS_SNAPSHOT_CMD_FLUSH;
	p->buf_info = (uintptr_t)data->hw_get - (uintptr_t)data->hw_snapshot;

	err = vgpu_comm_sendrecv(&msg, sizeof(msg), sizeof(msg));

	err = (err || msg.ret) ? -1 : 0;

	*pending = p->buf_info;
	*hw_overflow = p->hw_overflow;

	return err;
}

static int vgpu_css_attach(struct channel_gk20a *ch,
		struct gk20a_cs_snapshot_client *cs_client)
{
	struct gk20a *g = ch->g;
	struct tegra_vgpu_cmd_msg msg = {};
	struct tegra_vgpu_channel_cyclestats_snapshot_params *p =
				&msg.params.cyclestats_snapshot;
	int err;

	gk20a_dbg_fn("");

	msg.cmd = TEGRA_VGPU_CMD_CHANNEL_CYCLESTATS_SNAPSHOT;
	msg.handle = vgpu_get_handle(g);
	p->handle = ch->virt_ctx;
	p->subcmd = NVGPU_IOCTL_CHANNEL_CYCLE_STATS_SNAPSHOT_CMD_ATTACH;
	p->perfmon_count = cs_client->perfmon_count;

	err = vgpu_comm_sendrecv(&msg, sizeof(msg), sizeof(msg));
	err = err ? err : msg.ret;
	if (err)
		nvgpu_err(g, "failed");
	else
		cs_client->perfmon_start = p->perfmon_start;

	return err;
}

static int vgpu_css_detach(struct channel_gk20a *ch,
		struct gk20a_cs_snapshot_client *cs_client)
{
	struct gk20a *g = ch->g;
	struct tegra_vgpu_cmd_msg msg = {};
	struct tegra_vgpu_channel_cyclestats_snapshot_params *p =
				&msg.params.cyclestats_snapshot;
	int err;

	gk20a_dbg_fn("");

	msg.cmd = TEGRA_VGPU_CMD_CHANNEL_CYCLESTATS_SNAPSHOT;
	msg.handle = vgpu_get_handle(g);
	p->handle = ch->virt_ctx;
	p->subcmd = NVGPU_IOCTL_CHANNEL_CYCLE_STATS_SNAPSHOT_CMD_DETACH;
	p->perfmon_start = cs_client->perfmon_start;
	p->perfmon_count = cs_client->perfmon_count;

	err = vgpu_comm_sendrecv(&msg, sizeof(msg), sizeof(msg));
	err = err ? err : msg.ret;
	if (err)
		nvgpu_err(g, "failed");

	return err;
}

static int vgpu_css_enable_snapshot_buffer(struct channel_gk20a *ch,
				struct gk20a_cs_snapshot_client *cs_client)
{
	int ret;

	ret = vgpu_css_attach(ch, cs_client);
	if (ret)
		return ret;

	ret = vgpu_css_init_snapshot_buffer(&ch->g->gr);
	return ret;
}

void vgpu_init_css_ops(struct gpu_ops *gops)
{
	gops->css.enable_snapshot = vgpu_css_enable_snapshot_buffer;
	gops->css.disable_snapshot = vgpu_css_release_snapshot_buffer;
	gops->css.check_data_available = vgpu_css_flush_snapshots;
	gops->css.detach_snapshot = vgpu_css_detach;

	/* Following entries are not used when virtual, NULL them */
	gops->css.set_handled_snapshots = NULL;
	gops->css.allocate_perfmon_ids = NULL;
	gops->css.release_perfmon_ids = NULL;
}
#endif /* CONFIG_GK20A_CYCLE_STATS */