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

ctxgv100.c (7460B)


      1/*
      2 * Copyright 2018 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#include "ctxgf100.h"
     23
     24/*******************************************************************************
     25 * PGRAPH context implementation
     26 ******************************************************************************/
     27
     28const struct gf100_gr_init
     29gv100_grctx_init_sw_veid_bundle_init_0[] = {
     30	{ 0x00001000, 64, 0x00100000, 0x00000008 },
     31	{ 0x00000941, 64, 0x00100000, 0x00000000 },
     32	{ 0x0000097e, 64, 0x00100000, 0x00000000 },
     33	{ 0x0000097f, 64, 0x00100000, 0x00000100 },
     34	{ 0x0000035c, 64, 0x00100000, 0x00000000 },
     35	{ 0x0000035d, 64, 0x00100000, 0x00000000 },
     36	{ 0x00000a08, 64, 0x00100000, 0x00000000 },
     37	{ 0x00000a09, 64, 0x00100000, 0x00000000 },
     38	{ 0x00000a0a, 64, 0x00100000, 0x00000000 },
     39	{ 0x00000352, 64, 0x00100000, 0x00000000 },
     40	{ 0x00000353, 64, 0x00100000, 0x00000000 },
     41	{ 0x00000358, 64, 0x00100000, 0x00000000 },
     42	{ 0x00000359, 64, 0x00100000, 0x00000000 },
     43	{ 0x00000370, 64, 0x00100000, 0x00000000 },
     44	{ 0x00000371, 64, 0x00100000, 0x00000000 },
     45	{ 0x00000372, 64, 0x00100000, 0x000fffff },
     46	{ 0x00000366, 64, 0x00100000, 0x00000000 },
     47	{ 0x00000367, 64, 0x00100000, 0x00000000 },
     48	{ 0x00000368, 64, 0x00100000, 0x00000fff },
     49	{ 0x00000623, 64, 0x00100000, 0x00000000 },
     50	{ 0x00000624, 64, 0x00100000, 0x00000000 },
     51	{ 0x0001e100,  1, 0x00000001, 0x02000001 },
     52	{}
     53};
     54
     55static const struct gf100_gr_pack
     56gv100_grctx_pack_sw_veid_bundle_init[] = {
     57	{ gv100_grctx_init_sw_veid_bundle_init_0 },
     58	{}
     59};
     60
     61void
     62gv100_grctx_generate_attrib(struct gf100_grctx *info)
     63{
     64	struct gf100_gr *gr = info->gr;
     65	const struct gf100_grctx_func *grctx = gr->func->grctx;
     66	const u32  alpha = grctx->alpha_nr;
     67	const u32 attrib = grctx->attrib_nr;
     68	const u32   gfxp = grctx->gfxp_nr;
     69	const int s = 12;
     70	u32 size = grctx->alpha_nr_max * gr->tpc_total;
     71	u32 ao = 0;
     72	u32 bo = ao + size;
     73	int gpc, ppc, b, n = 0;
     74
     75	for (gpc = 0; gpc < gr->gpc_nr; gpc++)
     76		size += grctx->gfxp_nr * gr->ppc_nr[gpc] * gr->ppc_tpc_max;
     77	size = ((size * 0x20) + 127) & ~127;
     78	b = mmio_vram(info, size, (1 << s), false);
     79
     80	mmio_refn(info, 0x418810, 0x80000000, s, b);
     81	mmio_refn(info, 0x419848, 0x10000000, s, b);
     82	mmio_refn(info, 0x419c2c, 0x10000000, s, b);
     83	mmio_refn(info, 0x419e00, 0x00000000, s, b);
     84	mmio_wr32(info, 0x419e04, 0x80000000 | size >> 7);
     85	mmio_wr32(info, 0x405830, attrib);
     86	mmio_wr32(info, 0x40585c, alpha);
     87
     88	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
     89		for (ppc = 0; ppc < gr->ppc_nr[gpc]; ppc++, n++) {
     90			const u32 as =  alpha * gr->ppc_tpc_nr[gpc][ppc];
     91			const u32 bs = attrib * gr->ppc_tpc_max;
     92			const u32 gs =   gfxp * gr->ppc_tpc_max;
     93			const u32 u = 0x418ea0 + (n * 0x04);
     94			const u32 o = PPC_UNIT(gpc, ppc, 0);
     95			if (!(gr->ppc_mask[gpc] & (1 << ppc)))
     96				continue;
     97			mmio_wr32(info, o + 0xc0, gs);
     98			mmio_wr32(info, o + 0xf4, bo);
     99			mmio_wr32(info, o + 0xf0, bs);
    100			bo += gs;
    101			mmio_wr32(info, o + 0xe4, as);
    102			mmio_wr32(info, o + 0xf8, ao);
    103			ao += grctx->alpha_nr_max * gr->ppc_tpc_nr[gpc][ppc];
    104			mmio_wr32(info, u, bs);
    105		}
    106	}
    107
    108	mmio_wr32(info, 0x4181e4, 0x00000100);
    109	mmio_wr32(info, 0x41befc, 0x00000100);
    110}
    111
    112void
    113gv100_grctx_generate_rop_mapping(struct gf100_gr *gr)
    114{
    115	struct nvkm_device *device = gr->base.engine.subdev.device;
    116	u32 data;
    117	int i, j;
    118
    119	/* Pack tile map into register format. */
    120	nvkm_wr32(device, 0x418bb8, (gr->tpc_total << 8) |
    121				     gr->screen_tile_row_offset);
    122	for (i = 0; i < 11; i++) {
    123		for (data = 0, j = 0; j < 6; j++)
    124			data |= (gr->tile[i * 6 + j] & 0x1f) << (j * 5);
    125		nvkm_wr32(device, 0x418b08 + (i * 4), data);
    126		nvkm_wr32(device, 0x41bf00 + (i * 4), data);
    127		nvkm_wr32(device, 0x40780c + (i * 4), data);
    128	}
    129
    130	/* GPC_BROADCAST.TP_BROADCAST */
    131	nvkm_wr32(device, 0x41bfd0, (gr->tpc_total << 8) |
    132				     gr->screen_tile_row_offset);
    133	for (i = 0, j = 1; i < 5; i++, j += 4) {
    134		u8 v19 = (1 << (j + 0)) % gr->tpc_total;
    135		u8 v20 = (1 << (j + 1)) % gr->tpc_total;
    136		u8 v21 = (1 << (j + 2)) % gr->tpc_total;
    137		u8 v22 = (1 << (j + 3)) % gr->tpc_total;
    138		nvkm_wr32(device, 0x41bfb0 + (i * 4), (v22 << 24) |
    139						      (v21 << 16) |
    140						      (v20 <<  8) |
    141						       v19);
    142	}
    143
    144	/* UNK78xx */
    145	nvkm_wr32(device, 0x4078bc, (gr->tpc_total << 8) |
    146				     gr->screen_tile_row_offset);
    147}
    148
    149void
    150gv100_grctx_generate_r400088(struct gf100_gr *gr, bool on)
    151{
    152	struct nvkm_device *device = gr->base.engine.subdev.device;
    153	nvkm_mask(device, 0x400088, 0x00060000, on ? 0x00060000 : 0x00000000);
    154}
    155
    156static void
    157gv100_grctx_generate_sm_id(struct gf100_gr *gr, int gpc, int tpc, int sm)
    158{
    159	struct nvkm_device *device = gr->base.engine.subdev.device;
    160	nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x608), sm);
    161	nvkm_wr32(device, GPC_UNIT(gpc, 0x0c10 + tpc * 4), sm);
    162	nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x088), sm);
    163}
    164
    165void
    166gv100_grctx_generate_unkn(struct gf100_gr *gr)
    167{
    168	struct nvkm_device *device = gr->base.engine.subdev.device;
    169	nvkm_mask(device, 0x41980c, 0x00000010, 0x00000010);
    170	nvkm_mask(device, 0x41be08, 0x00000004, 0x00000004);
    171	nvkm_mask(device, 0x4064c0, 0x80000000, 0x80000000);
    172	nvkm_mask(device, 0x405800, 0x08000000, 0x08000000);
    173	nvkm_mask(device, 0x419c00, 0x00000008, 0x00000008);
    174}
    175
    176void
    177gv100_grctx_unkn88c(struct gf100_gr *gr, bool on)
    178{
    179	struct nvkm_device *device = gr->base.engine.subdev.device;
    180	const u32 mask = 0x00000010, data = on ? mask : 0x00000000;
    181	nvkm_mask(device, 0x40988c, mask, data);
    182	nvkm_rd32(device, 0x40988c);
    183	nvkm_mask(device, 0x41a88c, mask, data);
    184	nvkm_rd32(device, 0x41a88c);
    185	nvkm_mask(device, 0x408a14, mask, data);
    186	nvkm_rd32(device, 0x408a14);
    187}
    188
    189const struct gf100_grctx_func
    190gv100_grctx = {
    191	.unkn88c = gv100_grctx_unkn88c,
    192	.main = gf100_grctx_generate_main,
    193	.unkn = gv100_grctx_generate_unkn,
    194	.sw_veid_bundle_init = gv100_grctx_pack_sw_veid_bundle_init,
    195	.bundle = gm107_grctx_generate_bundle,
    196	.bundle_size = 0x3000,
    197	.bundle_min_gpm_fifo_depth = 0x180,
    198	.bundle_token_limit = 0x1680,
    199	.pagepool = gp100_grctx_generate_pagepool,
    200	.pagepool_size = 0x20000,
    201	.attrib = gv100_grctx_generate_attrib,
    202	.attrib_nr_max = 0x6c0,
    203	.attrib_nr = 0x480,
    204	.alpha_nr_max = 0xc00,
    205	.alpha_nr = 0x800,
    206	.gfxp_nr = 0xd10,
    207	.sm_id = gv100_grctx_generate_sm_id,
    208	.rop_mapping = gv100_grctx_generate_rop_mapping,
    209	.dist_skip_table = gm200_grctx_generate_dist_skip_table,
    210	.r406500 = gm200_grctx_generate_r406500,
    211	.gpc_tpc_nr = gk104_grctx_generate_gpc_tpc_nr,
    212	.smid_config = gp100_grctx_generate_smid_config,
    213	.r400088 = gv100_grctx_generate_r400088,
    214};