cachepc-linux

Fork of AMDESE/linux with modifications for CachePC side-channel attack
git clone https://git.sinitax.com/sinitax/cachepc-linux
Log | Files | Refs | README | LICENSE | sfeed.txt

ctxgf117.c (9265B)


      1/*
      2 * Copyright 2013 Red Hat Inc.
      3 *
      4 * Permission is hereby granted, free of charge, to any person obtaining a
      5 * copy of this software and associated documentation files (the "Software"),
      6 * to deal in the Software without restriction, including without limitation
      7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
      8 * and/or sell copies of the Software, and to permit persons to whom the
      9 * Software is furnished to do so, subject to the following conditions:
     10 *
     11 * The above copyright notice and this permission notice shall be included in
     12 * all copies or substantial portions of the Software.
     13 *
     14 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
     15 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
     16 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
     17 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
     18 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
     19 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
     20 * OTHER DEALINGS IN THE SOFTWARE.
     21 *
     22 * Authors: Ben Skeggs <bskeggs@redhat.com>
     23 */
     24#include "ctxgf100.h"
     25
     26#include <subdev/fb.h>
     27#include <subdev/mc.h>
     28
     29/*******************************************************************************
     30 * PGRAPH context register lists
     31 ******************************************************************************/
     32
     33static const struct gf100_gr_init
     34gf117_grctx_init_ds_0[] = {
     35	{ 0x405800,   1, 0x04, 0x0f8000bf },
     36	{ 0x405830,   1, 0x04, 0x02180324 },
     37	{ 0x405834,   1, 0x04, 0x08000000 },
     38	{ 0x405838,   1, 0x04, 0x00000000 },
     39	{ 0x405854,   1, 0x04, 0x00000000 },
     40	{ 0x405870,   4, 0x04, 0x00000001 },
     41	{ 0x405a00,   2, 0x04, 0x00000000 },
     42	{ 0x405a18,   1, 0x04, 0x00000000 },
     43	{}
     44};
     45
     46static const struct gf100_gr_init
     47gf117_grctx_init_pd_0[] = {
     48	{ 0x406020,   1, 0x04, 0x000103c1 },
     49	{ 0x406028,   4, 0x04, 0x00000001 },
     50	{ 0x4064a8,   1, 0x04, 0x00000000 },
     51	{ 0x4064ac,   1, 0x04, 0x00003fff },
     52	{ 0x4064b4,   3, 0x04, 0x00000000 },
     53	{ 0x4064c0,   1, 0x04, 0x801a0078 },
     54	{ 0x4064c4,   1, 0x04, 0x00c9ffff },
     55	{ 0x4064d0,   8, 0x04, 0x00000000 },
     56	{}
     57};
     58
     59static const struct gf100_gr_pack
     60gf117_grctx_pack_hub[] = {
     61	{ gf100_grctx_init_main_0 },
     62	{ gf119_grctx_init_fe_0 },
     63	{ gf100_grctx_init_pri_0 },
     64	{ gf100_grctx_init_memfmt_0 },
     65	{ gf117_grctx_init_ds_0 },
     66	{ gf117_grctx_init_pd_0 },
     67	{ gf100_grctx_init_rstr2d_0 },
     68	{ gf100_grctx_init_scc_0 },
     69	{ gf119_grctx_init_be_0 },
     70	{}
     71};
     72
     73static const struct gf100_gr_init
     74gf117_grctx_init_setup_0[] = {
     75	{ 0x418800,   1, 0x04, 0x7006860a },
     76	{ 0x418808,   3, 0x04, 0x00000000 },
     77	{ 0x418828,   1, 0x04, 0x00008442 },
     78	{ 0x418830,   1, 0x04, 0x10000001 },
     79	{ 0x4188d8,   1, 0x04, 0x00000008 },
     80	{ 0x4188e0,   1, 0x04, 0x01000000 },
     81	{ 0x4188e8,   5, 0x04, 0x00000000 },
     82	{ 0x4188fc,   1, 0x04, 0x20100018 },
     83	{}
     84};
     85
     86static const struct gf100_gr_pack
     87gf117_grctx_pack_gpc_0[] = {
     88	{ gf100_grctx_init_gpc_unk_0 },
     89	{ gf119_grctx_init_prop_0 },
     90	{ gf119_grctx_init_gpc_unk_1 },
     91	{ gf117_grctx_init_setup_0 },
     92	{ gf100_grctx_init_zcull_0 },
     93	{}
     94};
     95
     96const struct gf100_gr_pack
     97gf117_grctx_pack_gpc_1[] = {
     98	{ gf119_grctx_init_crstr_0 },
     99	{ gf108_grctx_init_gpm_0 },
    100	{ gf100_grctx_init_gcc_0 },
    101	{}
    102};
    103
    104const struct gf100_gr_init
    105gf117_grctx_init_pe_0[] = {
    106	{ 0x419848,   1, 0x04, 0x00000000 },
    107	{ 0x419864,   1, 0x04, 0x00000129 },
    108	{ 0x419888,   1, 0x04, 0x00000000 },
    109	{}
    110};
    111
    112static const struct gf100_gr_init
    113gf117_grctx_init_tex_0[] = {
    114	{ 0x419a00,   1, 0x04, 0x000001f0 },
    115	{ 0x419a04,   1, 0x04, 0x00000001 },
    116	{ 0x419a08,   1, 0x04, 0x00000023 },
    117	{ 0x419a0c,   1, 0x04, 0x00020000 },
    118	{ 0x419a10,   1, 0x04, 0x00000000 },
    119	{ 0x419a14,   1, 0x04, 0x00000200 },
    120	{ 0x419a1c,   1, 0x04, 0x00008000 },
    121	{ 0x419a20,   1, 0x04, 0x00000800 },
    122	{ 0x419ac4,   1, 0x04, 0x0017f440 },
    123	{}
    124};
    125
    126static const struct gf100_gr_init
    127gf117_grctx_init_mpc_0[] = {
    128	{ 0x419c00,   1, 0x04, 0x0000000a },
    129	{ 0x419c04,   1, 0x04, 0x00000006 },
    130	{ 0x419c08,   1, 0x04, 0x00000002 },
    131	{ 0x419c20,   1, 0x04, 0x00000000 },
    132	{ 0x419c24,   1, 0x04, 0x00084210 },
    133	{ 0x419c28,   1, 0x04, 0x3efbefbe },
    134	{}
    135};
    136
    137static const struct gf100_gr_pack
    138gf117_grctx_pack_tpc[] = {
    139	{ gf117_grctx_init_pe_0 },
    140	{ gf117_grctx_init_tex_0 },
    141	{ gf117_grctx_init_mpc_0 },
    142	{ gf104_grctx_init_l1c_0 },
    143	{ gf119_grctx_init_sm_0 },
    144	{}
    145};
    146
    147static const struct gf100_gr_init
    148gf117_grctx_init_pes_0[] = {
    149	{ 0x41be24,   1, 0x04, 0x00000002 },
    150	{}
    151};
    152
    153static const struct gf100_gr_init
    154gf117_grctx_init_cbm_0[] = {
    155	{ 0x41bec0,   1, 0x04, 0x12180000 },
    156	{ 0x41bec4,   1, 0x04, 0x00003fff },
    157	{ 0x41bee4,   1, 0x04, 0x03240218 },
    158	{}
    159};
    160
    161const struct gf100_gr_init
    162gf117_grctx_init_wwdx_0[] = {
    163	{ 0x41bf00,   1, 0x04, 0x0a418820 },
    164	{ 0x41bf04,   1, 0x04, 0x062080e6 },
    165	{ 0x41bf08,   1, 0x04, 0x020398a4 },
    166	{ 0x41bf0c,   1, 0x04, 0x0e629062 },
    167	{ 0x41bf10,   1, 0x04, 0x0a418820 },
    168	{ 0x41bf14,   1, 0x04, 0x000000e6 },
    169	{ 0x41bfd0,   1, 0x04, 0x00900103 },
    170	{ 0x41bfe0,   1, 0x04, 0x00400001 },
    171	{ 0x41bfe4,   1, 0x04, 0x00000000 },
    172	{}
    173};
    174
    175static const struct gf100_gr_pack
    176gf117_grctx_pack_ppc[] = {
    177	{ gf117_grctx_init_pes_0 },
    178	{ gf117_grctx_init_cbm_0 },
    179	{ gf117_grctx_init_wwdx_0 },
    180	{}
    181};
    182
    183/*******************************************************************************
    184 * PGRAPH context implementation
    185 ******************************************************************************/
    186
    187void
    188gf117_grctx_generate_dist_skip_table(struct gf100_gr *gr)
    189{
    190	struct nvkm_device *device = gr->base.engine.subdev.device;
    191	int i;
    192
    193	for (i = 0; i < 8; i++)
    194		nvkm_wr32(device, 0x4064d0 + (i * 0x04), 0x00000000);
    195}
    196
    197void
    198gf117_grctx_generate_rop_mapping(struct gf100_gr *gr)
    199{
    200	struct nvkm_device *device = gr->base.engine.subdev.device;
    201	u32 data[6] = {}, data2[2] = {};
    202	u8  shift, ntpcv;
    203	int i;
    204
    205	/* Pack tile map into register format. */
    206	for (i = 0; i < 32; i++)
    207		data[i / 6] |= (gr->tile[i] & 0x07) << ((i % 6) * 5);
    208
    209	/* Magic. */
    210	shift = 0;
    211	ntpcv = gr->tpc_total;
    212	while (!(ntpcv & (1 << 4))) {
    213		ntpcv <<= 1;
    214		shift++;
    215	}
    216
    217	data2[0]  = (ntpcv << 16);
    218	data2[0] |= (shift << 21);
    219	data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
    220	for (i = 1; i < 7; i++)
    221		data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
    222
    223	/* GPC_BROADCAST */
    224	nvkm_wr32(device, 0x418bb8, (gr->tpc_total << 8) |
    225				     gr->screen_tile_row_offset);
    226	for (i = 0; i < 6; i++)
    227		nvkm_wr32(device, 0x418b08 + (i * 4), data[i]);
    228
    229	/* GPC_BROADCAST.TP_BROADCAST */
    230	nvkm_wr32(device, 0x41bfd0, (gr->tpc_total << 8) |
    231				     gr->screen_tile_row_offset | data2[0]);
    232	nvkm_wr32(device, 0x41bfe4, data2[1]);
    233	for (i = 0; i < 6; i++)
    234		nvkm_wr32(device, 0x41bf00 + (i * 4), data[i]);
    235
    236	/* UNK78xx */
    237	nvkm_wr32(device, 0x4078bc, (gr->tpc_total << 8) |
    238				     gr->screen_tile_row_offset);
    239	for (i = 0; i < 6; i++)
    240		nvkm_wr32(device, 0x40780c + (i * 4), data[i]);
    241}
    242
    243void
    244gf117_grctx_generate_attrib(struct gf100_grctx *info)
    245{
    246	struct gf100_gr *gr = info->gr;
    247	const struct gf100_grctx_func *grctx = gr->func->grctx;
    248	const u32  alpha = grctx->alpha_nr;
    249	const u32   beta = grctx->attrib_nr;
    250	const u32   size = 0x20 * (grctx->attrib_nr_max + grctx->alpha_nr_max);
    251	const int s = 12;
    252	const int b = mmio_vram(info, size * gr->tpc_total, (1 << s), false);
    253	const int timeslice_mode = 1;
    254	const int max_batches = 0xffff;
    255	u32 bo = 0;
    256	u32 ao = bo + grctx->attrib_nr_max * gr->tpc_total;
    257	int gpc, ppc;
    258
    259	mmio_refn(info, 0x418810, 0x80000000, s, b);
    260	mmio_refn(info, 0x419848, 0x10000000, s, b);
    261	mmio_wr32(info, 0x405830, (beta << 16) | alpha);
    262	mmio_wr32(info, 0x4064c4, ((alpha / 4) << 16) | max_batches);
    263
    264	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
    265		for (ppc = 0; ppc < gr->ppc_nr[gpc]; ppc++) {
    266			const u32 a = alpha * gr->ppc_tpc_nr[gpc][ppc];
    267			const u32 b =  beta * gr->ppc_tpc_nr[gpc][ppc];
    268			const u32 t = timeslice_mode;
    269			const u32 o = PPC_UNIT(gpc, ppc, 0);
    270			if (!(gr->ppc_mask[gpc] & (1 << ppc)))
    271				continue;
    272			mmio_skip(info, o + 0xc0, (t << 28) | (b << 16) | ++bo);
    273			mmio_wr32(info, o + 0xc0, (t << 28) | (b << 16) | --bo);
    274			bo += grctx->attrib_nr_max * gr->ppc_tpc_nr[gpc][ppc];
    275			mmio_wr32(info, o + 0xe4, (a << 16) | ao);
    276			ao += grctx->alpha_nr_max * gr->ppc_tpc_nr[gpc][ppc];
    277		}
    278	}
    279}
    280
    281const struct gf100_grctx_func
    282gf117_grctx = {
    283	.main  = gf100_grctx_generate_main,
    284	.unkn  = gk104_grctx_generate_unkn,
    285	.hub   = gf117_grctx_pack_hub,
    286	.gpc_0 = gf117_grctx_pack_gpc_0,
    287	.gpc_1 = gf117_grctx_pack_gpc_1,
    288	.zcull = gf100_grctx_pack_zcull,
    289	.tpc   = gf117_grctx_pack_tpc,
    290	.ppc   = gf117_grctx_pack_ppc,
    291	.icmd  = gf119_grctx_pack_icmd,
    292	.mthd  = gf119_grctx_pack_mthd,
    293	.bundle = gf100_grctx_generate_bundle,
    294	.bundle_size = 0x1800,
    295	.pagepool = gf100_grctx_generate_pagepool,
    296	.pagepool_size = 0x8000,
    297	.attrib = gf117_grctx_generate_attrib,
    298	.attrib_nr_max = 0x324,
    299	.attrib_nr = 0x218,
    300	.alpha_nr_max = 0x7ff,
    301	.alpha_nr = 0x324,
    302	.sm_id = gf100_grctx_generate_sm_id,
    303	.tpc_nr = gf100_grctx_generate_tpc_nr,
    304	.r4060a8 = gf100_grctx_generate_r4060a8,
    305	.rop_mapping = gf117_grctx_generate_rop_mapping,
    306	.alpha_beta_tables = gf100_grctx_generate_alpha_beta_tables,
    307	.max_ways_evict = gf100_grctx_generate_max_ways_evict,
    308	.dist_skip_table = gf117_grctx_generate_dist_skip_table,
    309	.r419cb8 = gf100_grctx_generate_r419cb8,
    310};