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

ctxgp100.c (4945B)


      1/*
      2 * Copyright 2016 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
     28/*******************************************************************************
     29 * PGRAPH context implementation
     30 ******************************************************************************/
     31
     32void
     33gp100_grctx_generate_pagepool(struct gf100_grctx *info)
     34{
     35	const struct gf100_grctx_func *grctx = info->gr->func->grctx;
     36	const int s = 8;
     37	const int b = mmio_vram(info, grctx->pagepool_size, (1 << s), true);
     38	mmio_refn(info, 0x40800c, 0x00000000, s, b);
     39	mmio_wr32(info, 0x408010, 0x8007d800);
     40	mmio_refn(info, 0x419004, 0x00000000, s, b);
     41	mmio_wr32(info, 0x419008, 0x00000000);
     42}
     43
     44static void
     45gp100_grctx_generate_attrib(struct gf100_grctx *info)
     46{
     47	struct gf100_gr *gr = info->gr;
     48	const struct gf100_grctx_func *grctx = gr->func->grctx;
     49	const u32  alpha = grctx->alpha_nr;
     50	const u32 attrib = grctx->attrib_nr;
     51	const int s = 12;
     52	const int max_batches = 0xffff;
     53	u32 size = grctx->alpha_nr_max * gr->tpc_total;
     54	u32 ao = 0;
     55	u32 bo = ao + size;
     56	int gpc, ppc, b, n = 0;
     57
     58	for (gpc = 0; gpc < gr->gpc_nr; gpc++)
     59		size += grctx->attrib_nr_max * gr->ppc_nr[gpc] * gr->ppc_tpc_max;
     60	size = ((size * 0x20) + 128) & ~127;
     61	b = mmio_vram(info, size, (1 << s), false);
     62
     63	mmio_refn(info, 0x418810, 0x80000000, s, b);
     64	mmio_refn(info, 0x419848, 0x10000000, s, b);
     65	mmio_refn(info, 0x419c2c, 0x10000000, s, b);
     66	mmio_refn(info, 0x419b00, 0x00000000, s, b);
     67	mmio_wr32(info, 0x419b04, 0x80000000 | size >> 7);
     68	mmio_wr32(info, 0x405830, attrib);
     69	mmio_wr32(info, 0x40585c, alpha);
     70	mmio_wr32(info, 0x4064c4, ((alpha / 4) << 16) | max_batches);
     71
     72	for (gpc = 0; gpc < gr->gpc_nr; gpc++) {
     73		for (ppc = 0; ppc < gr->ppc_nr[gpc]; ppc++, n++) {
     74			const u32 as =  alpha * gr->ppc_tpc_nr[gpc][ppc];
     75			const u32 bs = attrib * gr->ppc_tpc_max;
     76			const u32 u = 0x418ea0 + (n * 0x04);
     77			const u32 o = PPC_UNIT(gpc, ppc, 0);
     78			if (!(gr->ppc_mask[gpc] & (1 << ppc)))
     79				continue;
     80			mmio_wr32(info, o + 0xc0, bs);
     81			mmio_wr32(info, o + 0xf4, bo);
     82			mmio_wr32(info, o + 0xf0, bs);
     83			bo += grctx->attrib_nr_max * gr->ppc_tpc_max;
     84			mmio_wr32(info, o + 0xe4, as);
     85			mmio_wr32(info, o + 0xf8, ao);
     86			ao += grctx->alpha_nr_max * gr->ppc_tpc_nr[gpc][ppc];
     87			mmio_wr32(info, u, bs);
     88		}
     89	}
     90
     91	mmio_wr32(info, 0x418eec, 0x00000000);
     92	mmio_wr32(info, 0x41befc, 0x00000000);
     93}
     94
     95void
     96gp100_grctx_generate_smid_config(struct gf100_gr *gr)
     97{
     98	struct nvkm_device *device = gr->base.engine.subdev.device;
     99	const u32 dist_nr = DIV_ROUND_UP(gr->tpc_total, 4);
    100	u32 dist[TPC_MAX / 4] = {}, gpcs[16] = {};
    101	u8  sm, i;
    102
    103	for (sm = 0; sm < gr->sm_nr; sm++) {
    104		const u8 gpc = gr->sm[sm].gpc;
    105		const u8 tpc = gr->sm[sm].tpc;
    106		dist[sm / 4] |= ((gpc << 4) | tpc) << ((sm % 4) * 8);
    107		gpcs[gpc + (gr->func->gpc_nr * (tpc / 4))] |= sm << ((tpc % 4) * 8);
    108	}
    109
    110	for (i = 0; i < dist_nr; i++)
    111		nvkm_wr32(device, 0x405b60 + (i * 4), dist[i]);
    112	for (i = 0; i < ARRAY_SIZE(gpcs); i++)
    113		nvkm_wr32(device, 0x405ba0 + (i * 4), gpcs[i]);
    114}
    115
    116const struct gf100_grctx_func
    117gp100_grctx = {
    118	.main  = gf100_grctx_generate_main,
    119	.unkn  = gk104_grctx_generate_unkn,
    120	.bundle = gm107_grctx_generate_bundle,
    121	.bundle_size = 0x3000,
    122	.bundle_min_gpm_fifo_depth = 0x180,
    123	.bundle_token_limit = 0x1080,
    124	.pagepool = gp100_grctx_generate_pagepool,
    125	.pagepool_size = 0x20000,
    126	.attrib = gp100_grctx_generate_attrib,
    127	.attrib_nr_max = 0x660,
    128	.attrib_nr = 0x440,
    129	.alpha_nr_max = 0xc00,
    130	.alpha_nr = 0x800,
    131	.sm_id = gm107_grctx_generate_sm_id,
    132	.rop_mapping = gf117_grctx_generate_rop_mapping,
    133	.dist_skip_table = gm200_grctx_generate_dist_skip_table,
    134	.r406500 = gm200_grctx_generate_r406500,
    135	.gpc_tpc_nr = gk104_grctx_generate_gpc_tpc_nr,
    136	.tpc_mask = gm200_grctx_generate_tpc_mask,
    137	.smid_config = gp100_grctx_generate_smid_config,
    138	.r419a3c = gm200_grctx_generate_r419a3c,
    139};