NVIDIA will officially start providing signed firmwares through linux-firmware. Change the GR firmware lookup function to look them up in addition to the extracted firmwares.

Signed-off-by: Alexandre Courbot acourbot@nvidia.com --- drm/nouveau/nvkm/engine/gr/gf100.c | 66 +++++++++++++++++++++++++++++--------- 1 file changed, 51 insertions(+), 15 deletions(-)

diff --git a/drm/nouveau/nvkm/engine/gr/gf100.c b/drm/nouveau/nvkm/engine/gr/gf100.c index ca11ddb..39d482f 100644 --- a/drm/nouveau/nvkm/engine/gr/gf100.c +++ b/drm/nouveau/nvkm/engine/gr/gf100.c @@ -1544,26 +1544,62 @@ gf100_gr_dtor_fw(struct gf100_gr_fuc *fuc) fuc->data = NULL; }

+/** + * gf100_gr_ctor_fw - helper for loading external GR firmwares + * + * A firmware can either be officially provided by NVIDIA (in which case it will + * use a "NVIDIA name", or be extracted from the binary blob (and use a + * "Nouveau name". The fwname and nvfwname are to be given the Nouveau and + * NVIDIA names of a given firmware, respectively. This function will then + * try to load the NVIDIA firmware, then the extracted one, in that order. + * + */ int gf100_gr_ctor_fw(struct gf100_gr_priv *priv, const char *fwname, - struct gf100_gr_fuc *fuc) + const char *nvfwname, struct gf100_gr_fuc *fuc) { struct nvkm_device *device = nv_device(priv); const struct firmware *fw; - char f[32]; - int ret; + char f[64]; + int ret = -EINVAL; + int i;

- snprintf(f, sizeof(f), "nouveau/nv%02x_%s", device->chipset, fwname); - ret = request_firmware(&fw, f, nv_device_base(device)); - if (ret) { - snprintf(f, sizeof(f), "nouveau/%s", fwname); - ret = request_firmware(&fw, f, nv_device_base(device)); - if (ret) { - nv_error(priv, "failed to load %s\n", fwname); - return ret; + /* + * NVIDIA firmware name provided - try to load it + * We try this first since most chips that require external firmware + * are supported by NVIDIA + */ + if (nvfwname) { + snprintf(f, sizeof(f), "nvidia/%s/%s.bin", device->cname, + nvfwname); + i = strlen(f); + while (i) { + --i; + f[i] = tolower(f[i]); } + ret = request_firmware_direct(&fw, f, nv_device_base(device)); + if (!ret) + goto found; + } + + /* Nouveau firmware name provided - try to load it */ + if (fwname) { + snprintf(f, sizeof(f), "nouveau/nv%02x_%s", device->chipset, + fwname); + ret = request_firmware_direct(&fw, f, nv_device_base(device)); + if (!ret) + goto found; + + snprintf(f, sizeof(f), "nouveau/%s", fwname); + ret = request_firmware_direct(&fw, f, nv_device_base(device)); + if (!ret) + goto found; }

+ nv_error(priv, "failed to load %s / %s\n", fwname, nvfwname); + return ret; + +found: fuc->size = fw->size; fuc->data = kmemdup(fw->data, fuc->size, GFP_KERNEL); release_firmware(fw); @@ -1615,10 +1651,10 @@ gf100_gr_ctor(struct nvkm_object *parent, struct nvkm_object *engine,

if (use_ext_fw) { nv_info(priv, "using external firmware\n"); - if (gf100_gr_ctor_fw(priv, "fuc409c", &priv->fuc409c) || - gf100_gr_ctor_fw(priv, "fuc409d", &priv->fuc409d) || - gf100_gr_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) || - gf100_gr_ctor_fw(priv, "fuc41ad", &priv->fuc41ad)) + if (gf100_gr_ctor_fw(priv, "fuc409c", "fecs_inst", &priv->fuc409c) || + gf100_gr_ctor_fw(priv, "fuc409d", "fecs_data", &priv->fuc409d) || + gf100_gr_ctor_fw(priv, "fuc41ac", "gpccs_inst", &priv->fuc41ac) || + gf100_gr_ctor_fw(priv, "fuc41ad", "gpccs_data", &priv->fuc41ad)) return -ENODEV; priv->firmware = true; }

GK20A's initialization was based on GK104, but differences exist in the way the initial context is built and the initialization process itself.

This patch follows the same initialization sequence as nvgpu performs to avoid bad surprises. Since the register bundles initialization also differ considerably from GK104, the register packs are now loaded from firmware files, again similarly to what is done with nvgpu.

Signed-off-by: Alexandre Courbot acourbot@nvidia.com --- drm/nouveau/nvkm/engine/gr/ctxgk20a.c | 65 +++++-- drm/nouveau/nvkm/engine/gr/gf100.c | 3 +- drm/nouveau/nvkm/engine/gr/gf100.h | 12 ++ drm/nouveau/nvkm/engine/gr/gk20a.c | 336 ++++++++++++++++++++++++++++++++-- drm/nouveau/nvkm/engine/gr/gk20a.h | 35 ++++ 5 files changed, 421 insertions(+), 30 deletions(-) create mode 100644 drm/nouveau/nvkm/engine/gr/gk20a.h

diff --git a/drm/nouveau/nvkm/engine/gr/ctxgk20a.c b/drm/nouveau/nvkm/engine/gr/ctxgk20a.c index 2f241f6..3fe080e 100644 --- a/drm/nouveau/nvkm/engine/gr/ctxgk20a.c +++ b/drm/nouveau/nvkm/engine/gr/ctxgk20a.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved. + * Copyright (c) 2014-2015, NVIDIA CORPORATION. All rights reserved. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -19,14 +19,56 @@ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER * DEALINGS IN THE SOFTWARE. */ + #include "ctxgf100.h" +#include "gk20a.h" + +#include <subdev/mc.h> + +static void +gk20a_grctx_generate_main(struct gf100_gr_priv *priv, struct gf100_grctx *info) +{ + struct gf100_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass; + int idle_timeout_save; + int i; + + gf100_gr_mmio(priv, priv->fuc_sw_ctx); + + gf100_gr_wait_idle(priv); + + idle_timeout_save = nv_rd32(priv, 0x404154); + nv_wr32(priv, 0x404154, 0x00000000); + + oclass->attrib(info); + + oclass->unkn(priv); + + gf100_grctx_generate_tpcid(priv); + gf100_grctx_generate_r406028(priv); + gk104_grctx_generate_r418bb8(priv); + gf100_grctx_generate_r406800(priv); + + for (i = 0; i < 8; i++) + nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000); + + nv_wr32(priv, 0x405b00, (priv->tpc_total << 8) | priv->gpc_nr); + + gk104_grctx_generate_rop_active_fbps(priv); + + nv_mask(priv, 0x5044b0, 0x8000000, 0x8000000); + + gf100_gr_wait_idle(priv); + + nv_wr32(priv, 0x404154, idle_timeout_save); + gf100_gr_wait_idle(priv); + + gf100_gr_mthd(priv, priv->fuc_method); + gf100_gr_wait_idle(priv);

-static const struct gf100_gr_pack -gk20a_grctx_pack_mthd[] = { - { gk104_grctx_init_a097_0, 0xa297 }, - { gf100_grctx_init_902d_0, 0x902d }, - {} -}; + gf100_gr_icmd(priv, priv->fuc_bundle); + oclass->pagepool(info); + oclass->bundle(info); +}

struct nvkm_oclass * gk20a_grctx_oclass = &(struct gf100_grctx_oclass) { @@ -39,15 +81,8 @@ gk20a_grctx_oclass = &(struct gf100_grctx_oclass) { .rd32 = _nvkm_gr_context_rd32, .wr32 = _nvkm_gr_context_wr32, }, - .main = gk104_grctx_generate_main, + .main = gk20a_grctx_generate_main, .unkn = gk104_grctx_generate_unkn, - .hub = gk104_grctx_pack_hub, - .gpc = gk104_grctx_pack_gpc, - .zcull = gf100_grctx_pack_zcull, - .tpc = gk104_grctx_pack_tpc, - .ppc = gk104_grctx_pack_ppc, - .icmd = gk104_grctx_pack_icmd, - .mthd = gk20a_grctx_pack_mthd, .bundle = gk104_grctx_generate_bundle, .bundle_size = 0x1800, .bundle_min_gpm_fifo_depth = 0x62, diff --git a/drm/nouveau/nvkm/engine/gr/gf100.c b/drm/nouveau/nvkm/engine/gr/gf100.c index 39d482f..d5c148f 100644 --- a/drm/nouveau/nvkm/engine/gr/gf100.c +++ b/drm/nouveau/nvkm/engine/gr/gf100.c @@ -1537,7 +1537,7 @@ gf100_gr_init(struct nvkm_object *object) return gf100_gr_init_ctxctl(priv); }

-static void +void gf100_gr_dtor_fw(struct gf100_gr_fuc *fuc) { kfree(fuc->data); @@ -1719,6 +1719,7 @@ gf100_gr_ctor(struct nvkm_object *parent, struct nvkm_object *engine, break; case 0xd7: case 0xd9: /* 1/0/0/0, 1 */ + case 0xea: /* gk20a */ priv->magic_not_rop_nr = 0x01; break; } diff --git a/drm/nouveau/nvkm/engine/gr/gf100.h b/drm/nouveau/nvkm/engine/gr/gf100.h index c9533fd..af4d70f 100644 --- a/drm/nouveau/nvkm/engine/gr/gf100.h +++ b/drm/nouveau/nvkm/engine/gr/gf100.h @@ -76,6 +76,15 @@ struct gf100_gr_priv { struct gf100_gr_fuc fuc41ad; bool firmware;

+ /* + * Used if the register packs are loaded from NVIDIA fw instead of + * using hardcoded arrays. + */ + struct gf100_gr_pack *fuc_sw_nonctx; + struct gf100_gr_pack *fuc_sw_ctx; + struct gf100_gr_pack *fuc_bundle; + struct gf100_gr_pack *fuc_method; + struct gf100_gr_zbc_color zbc_color[NVKM_LTC_MAX_ZBC_CNT]; struct gf100_gr_zbc_depth zbc_depth[NVKM_LTC_MAX_ZBC_CNT];

@@ -116,6 +125,9 @@ void gf100_gr_context_dtor(struct nvkm_object *);

void gf100_gr_ctxctl_debug(struct gf100_gr_priv *);

+void gf100_gr_dtor_fw(struct gf100_gr_fuc *); +int gf100_gr_ctor_fw(struct gf100_gr_priv *, const char *, const char *, + struct gf100_gr_fuc *); u64 gf100_gr_units(struct nvkm_gr *); int gf100_gr_ctor(struct nvkm_object *, struct nvkm_object *, struct nvkm_oclass *, void *data, u32 size, diff --git a/drm/nouveau/nvkm/engine/gr/gk20a.c b/drm/nouveau/nvkm/engine/gr/gk20a.c index 40ff5eb..65f1089 100644 --- a/drm/nouveau/nvkm/engine/gr/gk20a.c +++ b/drm/nouveau/nvkm/engine/gr/gk20a.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2014, NVIDIA CORPORATION. All rights reserved. + * Copyright (c) 2014-2015, NVIDIA CORPORATION. All rights reserved. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -19,10 +19,11 @@ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER * DEALINGS IN THE SOFTWARE. */ -#include "gf100.h" +#include "gk20a.h" #include "ctxgf100.h"

#include <nvif/class.h> +#include <subdev/timer.h>

static struct nvkm_oclass gk20a_gr_sclass[] = { @@ -33,17 +34,324 @@ gk20a_gr_sclass[] = { {} };

+static void +gk20a_gr_init_dtor(struct gf100_gr_pack *pack) +{ + vfree(pack); +} + +struct gk20a_fw_av +{ + u32 addr; + u32 data; +}; + +static struct gf100_gr_pack * +gk20a_gr_av_to_init(struct gf100_gr_fuc *fuc) +{ + struct gf100_gr_init *init; + struct gf100_gr_pack *pack; + const int nent = (fuc->size / sizeof(struct gk20a_fw_av)); + int i; + + pack = vzalloc((sizeof(*pack) * 2) + (sizeof(*init) * (nent + 1))); + if (!pack) + return ERR_PTR(-ENOMEM); + + init = (void *)(pack + 2); + + pack[0].init = init; + + for (i = 0; i < nent; i++) { + struct gf100_gr_init *ent = &init[i]; + struct gk20a_fw_av *av = &((struct gk20a_fw_av *)fuc->data)[i]; + + ent->addr = av->addr; + ent->data = av->data; + ent->count = 1; + ent->pitch = 1; + } + + return pack; +} + +struct gk20a_fw_aiv +{ + u32 addr; + u32 index; + u32 data; +}; + +static struct gf100_gr_pack * +gk20a_gr_aiv_to_init(struct gf100_gr_fuc *fuc) +{ + struct gf100_gr_init *init; + struct gf100_gr_pack *pack; + const int nent = (fuc->size / sizeof(struct gk20a_fw_aiv)); + int i; + + pack = vzalloc((sizeof(*pack) * 2) + (sizeof(*init) * (nent + 1))); + if (!pack) + return ERR_PTR(-ENOMEM); + + init = (void *)(pack + 2); + + pack[0].init = init; + + for (i = 0; i < nent; i++) { + struct gf100_gr_init *ent = &init[i]; + struct gk20a_fw_aiv *av = &((struct gk20a_fw_aiv *)fuc->data)[i]; + + ent->addr = av->addr; + ent->data = av->data; + ent->count = 1; + ent->pitch = 1; + } + + return pack; +} + +static struct gf100_gr_pack * +gk20a_gr_av_to_method(struct gf100_gr_fuc *fuc) +{ + struct gf100_gr_init *init; + struct gf100_gr_pack *pack; + /* We don't suppose we will initialize more than 16 classes here... */ + static const unsigned int max_classes = 16; + const int nent = (fuc->size / sizeof(struct gk20a_fw_av)); + int i, classidx = 0; + u32 prevclass = 0; + + pack = vzalloc((sizeof(*pack) * max_classes) + + (sizeof(*init) * (nent + 1))); + if (!pack) + return ERR_PTR(-ENOMEM); + + init = (void *)(pack + max_classes); + + for (i = 0; i < nent; i++) { + struct gf100_gr_init *ent = &init[i]; + struct gk20a_fw_av *av = &((struct gk20a_fw_av *)fuc->data)[i]; + u32 class = av->addr & 0xffff; + u32 addr = (av->addr & 0xffff0000) >> 14; + + if (prevclass != class) { + pack[classidx].init = ent; + pack[classidx].type = class; + prevclass = class; + if (++classidx >= max_classes) { + vfree(pack); + return ERR_PTR(-ENOSPC); + } + } + + ent->addr = addr; + ent->data = av->data; + ent->count = 1; + ent->pitch = 1; + } + + return pack; +} + +static int +gk20a_gr_ctor(struct nvkm_object *parent, struct nvkm_object *engine, + struct nvkm_oclass *oclass, void *data, u32 size, + struct nvkm_object **pobject) +{ + int err; + struct gf100_gr_priv *priv; + struct gf100_gr_fuc fuc; + + err = gf100_gr_ctor(parent, engine, oclass, data, size, pobject); + if (err) + return err; + + priv = (void *)*pobject; + + err = gf100_gr_ctor_fw(priv, NULL, "sw_nonctx", &fuc); + if (err) + return err; + priv->fuc_sw_nonctx = gk20a_gr_av_to_init(&fuc); + gf100_gr_dtor_fw(&fuc); + if (IS_ERR(priv->fuc_sw_nonctx)) + return PTR_ERR(priv->fuc_sw_nonctx); + + err = gf100_gr_ctor_fw(priv, NULL, "sw_ctx", &fuc); + if (err) + return err; + priv->fuc_sw_ctx = gk20a_gr_aiv_to_init(&fuc); + gf100_gr_dtor_fw(&fuc); + if (IS_ERR(priv->fuc_sw_ctx)) + return PTR_ERR(priv->fuc_sw_ctx); + + err = gf100_gr_ctor_fw(priv, NULL, "sw_bundle_init", &fuc); + if (err) + return err; + priv->fuc_bundle = gk20a_gr_av_to_init(&fuc); + gf100_gr_dtor_fw(&fuc); + if (IS_ERR(priv->fuc_bundle)) + return PTR_ERR(priv->fuc_bundle); + + err = gf100_gr_ctor_fw(priv, NULL, "sw_method_init", &fuc); + if (err) + return err; + priv->fuc_method = gk20a_gr_av_to_method(&fuc); + gf100_gr_dtor_fw(&fuc); + if (IS_ERR(priv->fuc_method)) + return PTR_ERR(priv->fuc_method); + + return 0; +} + +static void +gk20a_gr_dtor(struct nvkm_object *object) +{ + struct gf100_gr_priv *priv = (void *)object; + + gk20a_gr_init_dtor(priv->fuc_method); + gk20a_gr_init_dtor(priv->fuc_bundle); + gk20a_gr_init_dtor(priv->fuc_sw_ctx); + gk20a_gr_init_dtor(priv->fuc_sw_nonctx); + + gf100_gr_dtor(object); +} + +static int +gk20a_gr_wait_mem_scrubbing(struct gf100_gr_priv *priv) +{ + if (!nv_wait(priv, 0x40910c, 0x6, 0x0)) { + nv_error(priv, "FECS mem scrubbing timeout\n"); + return -ETIMEDOUT; + } + + if (!nv_wait(priv, 0x41a10c, 0x6, 0x0)) { + nv_error(priv, "GPCCS mem scrubbing timeout\n"); + return -ETIMEDOUT; + } + + return 0; +} + +static void +gk20a_gr_set_hww_esr_report_mask(struct gf100_gr_priv *priv) +{ + nv_wr32(priv, 0x419e44, 0x1ffffe); + nv_wr32(priv, 0x419e4c, 0x7f); +} + +static int +gk20a_gr_init(struct nvkm_object *object) +{ + struct gk20a_gr_oclass *oclass = (void *)object->oclass; + struct gf100_gr_priv *priv = (void *)object; + const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total); + u32 data[TPC_MAX / 8] = {}; + u8 tpcnr[GPC_MAX]; + int gpc, tpc; + int ret, i; + + ret = nvkm_gr_init(&priv->base); + if (ret) + return ret; + + /* Clear SCC RAM */ + nv_wr32(priv, 0x40802c, 0x1); + + gf100_gr_mmio(priv, priv->fuc_sw_nonctx); + + ret = gk20a_gr_wait_mem_scrubbing(priv); + if (ret) + return ret; + + ret = gf100_gr_wait_idle(priv); + if (ret) + return ret; + + /* MMU debug buffer */ + nv_wr32(priv, 0x100cc8, priv->unk4188b4->addr >> 8); + nv_wr32(priv, 0x100ccc, priv->unk4188b8->addr >> 8); + + if (oclass->init_gpc_mmu) + oclass->init_gpc_mmu(priv); + + /* Set the PE as stream master */ + nv_mask(priv, 0x503018, 0x1, 0x1); + + /* Zcull init */ + memset(data, 0x00, sizeof(data)); + memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr)); + for (i = 0, gpc = -1; i < priv->tpc_total; i++) { + do { + gpc = (gpc + 1) % priv->gpc_nr; + } while (!tpcnr[gpc]); + tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--; + + data[i / 8] |= tpc << ((i % 8) * 4); + } + + nv_wr32(priv, GPC_BCAST(0x0980), data[0]); + nv_wr32(priv, GPC_BCAST(0x0984), data[1]); + nv_wr32(priv, GPC_BCAST(0x0988), data[2]); + nv_wr32(priv, GPC_BCAST(0x098c), data[3]); + + for (gpc = 0; gpc < priv->gpc_nr; gpc++) { + nv_wr32(priv, GPC_UNIT(gpc, 0x0914), + priv->magic_not_rop_nr << 8 | priv->tpc_nr[gpc]); + nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 | + priv->tpc_total); + nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918); + } + + nv_wr32(priv, GPC_BCAST(0x3fd4), magicgpc918); + + /* Enable FIFO access */ + nv_wr32(priv, 0x400500, 0x00010001); + + /* Enable interrupts */ + nv_wr32(priv, 0x400100, 0xffffffff); + nv_wr32(priv, 0x40013c, 0xffffffff); + + /* Enable FECS error interrupts */ + nv_wr32(priv, 0x409c24, 0x000f0000); + + /* Enable hardware warning exceptions */ + nv_wr32(priv, 0x404000, 0xc0000000); + nv_wr32(priv, 0x404600, 0xc0000000); + + if (oclass->set_hww_esr_report_mask) + oclass->set_hww_esr_report_mask(priv); + + /* Enable TPC exceptions per GPC */ + nv_wr32(priv, 0x419d0c, 0x2); + nv_wr32(priv, 0x41ac94, (((1 << priv->tpc_total) - 1) & 0xff) << 16); + + /* Reset and enable all exceptions */ + nv_wr32(priv, 0x400108, 0xffffffff); + nv_wr32(priv, 0x400138, 0xffffffff); + nv_wr32(priv, 0x400118, 0xffffffff); + nv_wr32(priv, 0x400130, 0xffffffff); + nv_wr32(priv, 0x40011c, 0xffffffff); + nv_wr32(priv, 0x400134, 0xffffffff); + + gf100_gr_zbc_init(priv); + + return gf100_gr_init_ctxctl(priv); +} + struct nvkm_oclass * -gk20a_gr_oclass = &(struct gf100_gr_oclass) { - .base.handle = NV_ENGINE(GR, 0xea), - .base.ofuncs = &(struct nvkm_ofuncs) { - .ctor = gf100_gr_ctor, - .dtor = gf100_gr_dtor, - .init = gk104_gr_init, - .fini = _nvkm_gr_fini, +gk20a_gr_oclass = &(struct gk20a_gr_oclass) { + .gf100 = { + .base.handle = NV_ENGINE(GR, 0xea), + .base.ofuncs = &(struct nvkm_ofuncs) { + .ctor = gk20a_gr_ctor, + .dtor = gk20a_gr_dtor, + .init = gk20a_gr_init, + .fini = _nvkm_gr_fini, + }, + .cclass = &gk20a_grctx_oclass, + .sclass = gk20a_gr_sclass, + .ppc_nr = 1, }, - .cclass = &gk20a_grctx_oclass, - .sclass = gk20a_gr_sclass, - .mmio = gk104_gr_pack_mmio, - .ppc_nr = 1, -}.base; + .set_hww_esr_report_mask = gk20a_gr_set_hww_esr_report_mask, +}.gf100.base; diff --git a/drm/nouveau/nvkm/engine/gr/gk20a.h b/drm/nouveau/nvkm/engine/gr/gk20a.h new file mode 100644 index 0000000..b369585 --- /dev/null +++ b/drm/nouveau/nvkm/engine/gr/gk20a.h @@ -0,0 +1,35 @@ +/* + * Copyright (c) 2015, NVIDIA CORPORATION. All rights reserved. + * + * 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 AUTHORS OR COPYRIGHT HOLDERS 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. + */ + +#ifndef __GK20A_GR_H__ +#define __GK20A_GR_H__ + +#include "gf100.h" + +struct gk20a_gr_oclass { + struct gf100_gr_oclass gf100; + + void (*init_gpc_mmu)(struct gf100_gr_priv *); + void (*set_hww_esr_report_mask)(struct gf100_gr_priv *); +}; + +#endif

Add support for GM20B's graphics engine, based on GK20A. Note that this code alone will not allow the engine to initialize on released devices which require PMU-assisted secure boot.

Signed-off-by: Alexandre Courbot acourbot@nvidia.com --- drm/nouveau/include/nvkm/engine/gr.h | 1 + drm/nouveau/nvkm/engine/gr/Kbuild | 2 + drm/nouveau/nvkm/engine/gr/ctxgf100.h | 7 +++ drm/nouveau/nvkm/engine/gr/ctxgm107.c | 2 +- drm/nouveau/nvkm/engine/gr/ctxgm204.c | 4 +- drm/nouveau/nvkm/engine/gr/ctxgm20b.c | 110 ++++++++++++++++++++++++++++++++++ drm/nouveau/nvkm/engine/gr/gf100.c | 1 + drm/nouveau/nvkm/engine/gr/gf100.h | 6 ++ drm/nouveau/nvkm/engine/gr/gk20a.c | 6 +- drm/nouveau/nvkm/engine/gr/gm20b.c | 84 ++++++++++++++++++++++++++ 10 files changed, 217 insertions(+), 6 deletions(-) create mode 100644 drm/nouveau/nvkm/engine/gr/ctxgm20b.c create mode 100644 drm/nouveau/nvkm/engine/gr/gm20b.c

diff --git a/drm/nouveau/include/nvkm/engine/gr.h b/drm/nouveau/include/nvkm/engine/gr.h index 7cbe202..c772497 100644 --- a/drm/nouveau/include/nvkm/engine/gr.h +++ b/drm/nouveau/include/nvkm/engine/gr.h @@ -74,6 +74,7 @@ extern struct nvkm_oclass *gk208_gr_oclass; extern struct nvkm_oclass *gm107_gr_oclass; extern struct nvkm_oclass *gm204_gr_oclass; extern struct nvkm_oclass *gm206_gr_oclass; +extern struct nvkm_oclass *gm20b_gr_oclass;

#include <core/enum.h>

diff --git a/drm/nouveau/nvkm/engine/gr/Kbuild b/drm/nouveau/nvkm/engine/gr/Kbuild index 2e1b92f..e91b4df 100644 --- a/drm/nouveau/nvkm/engine/gr/Kbuild +++ b/drm/nouveau/nvkm/engine/gr/Kbuild @@ -14,6 +14,7 @@ nvkm-y += nvkm/engine/gr/ctxgk208.o nvkm-y += nvkm/engine/gr/ctxgm107.o nvkm-y += nvkm/engine/gr/ctxgm204.o nvkm-y += nvkm/engine/gr/ctxgm206.o +nvkm-y += nvkm/engine/gr/ctxgm20b.o nvkm-y += nvkm/engine/gr/nv04.o nvkm-y += nvkm/engine/gr/nv10.o nvkm-y += nvkm/engine/gr/nv20.o @@ -38,3 +39,4 @@ nvkm-y += nvkm/engine/gr/gk208.o nvkm-y += nvkm/engine/gr/gm107.o nvkm-y += nvkm/engine/gr/gm204.o nvkm-y += nvkm/engine/gr/gm206.o +nvkm-y += nvkm/engine/gr/gm20b.o diff --git a/drm/nouveau/nvkm/engine/gr/ctxgf100.h b/drm/nouveau/nvkm/engine/gr/ctxgf100.h index 3676a33..f89ab37 100644 --- a/drm/nouveau/nvkm/engine/gr/ctxgf100.h +++ b/drm/nouveau/nvkm/engine/gr/ctxgf100.h @@ -91,6 +91,10 @@ void gk104_grctx_generate_r418bb8(struct gf100_gr_priv *); void gk104_grctx_generate_rop_active_fbps(struct gf100_gr_priv *);

+void gm107_grctx_generate_bundle(struct gf100_grctx *); +void gm107_grctx_generate_pagepool(struct gf100_grctx *); +void gm107_grctx_generate_attrib(struct gf100_grctx *); + extern struct nvkm_oclass *gk110_grctx_oclass; extern struct nvkm_oclass *gk110b_grctx_oclass; extern struct nvkm_oclass *gk208_grctx_oclass; @@ -102,8 +106,11 @@ void gm107_grctx_generate_attrib(struct gf100_grctx *);

extern struct nvkm_oclass *gm204_grctx_oclass; void gm204_grctx_generate_main(struct gf100_gr_priv *, struct gf100_grctx *); +void gm204_grctx_generate_tpcid(struct gf100_gr_priv *); +void gm204_grctx_generate_405b60(struct gf100_gr_priv *);

extern struct nvkm_oclass *gm206_grctx_oclass; +extern struct nvkm_oclass *gm20b_grctx_oclass;

/* context init value lists */

diff --git a/drm/nouveau/nvkm/engine/gr/ctxgm107.c b/drm/nouveau/nvkm/engine/gr/ctxgm107.c index fbeaae3..6bf2fd1 100644 --- a/drm/nouveau/nvkm/engine/gr/ctxgm107.c +++ b/drm/nouveau/nvkm/engine/gr/ctxgm107.c @@ -931,7 +931,7 @@ gm107_grctx_generate_attrib(struct gf100_grctx *info) } }

-static void +void gm107_grctx_generate_tpcid(struct gf100_gr_priv *priv) { int gpc, tpc, id; diff --git a/drm/nouveau/nvkm/engine/gr/ctxgm204.c b/drm/nouveau/nvkm/engine/gr/ctxgm204.c index ea8e661..efc76bf 100644 --- a/drm/nouveau/nvkm/engine/gr/ctxgm204.c +++ b/drm/nouveau/nvkm/engine/gr/ctxgm204.c @@ -918,7 +918,7 @@ gm204_grctx_pack_ppc[] = { * PGRAPH context implementation ******************************************************************************/

-static void +void gm204_grctx_generate_tpcid(struct gf100_gr_priv *priv) { int gpc, tpc, id; @@ -943,7 +943,7 @@ gm204_grctx_generate_rop_active_fbps(struct gf100_gr_priv *priv) nv_mask(priv, 0x408958, 0x0000000f, fbp_count); /* crop */ }

-static void +void gm204_grctx_generate_405b60(struct gf100_gr_priv *priv) { const u32 dist_nr = DIV_ROUND_UP(priv->tpc_total, 4); diff --git a/drm/nouveau/nvkm/engine/gr/ctxgm20b.c b/drm/nouveau/nvkm/engine/gr/ctxgm20b.c new file mode 100644 index 0000000..c011bf3 --- /dev/null +++ b/drm/nouveau/nvkm/engine/gr/ctxgm20b.c @@ -0,0 +1,110 @@ +/* + * Copyright (c) 2015, NVIDIA CORPORATION. All rights reserved. + * + * 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 AUTHORS OR COPYRIGHT HOLDERS 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. + */ +#include "ctxgf100.h" + +static void +gm20b_grctx_generate_r406028(struct gf100_gr_priv *priv) +{ + u32 tpc_per_gpc = 0; + int i; + + for (i = 0; i < priv->gpc_nr; i++) + tpc_per_gpc |= priv->tpc_nr[i] << (4 * i); + + nv_wr32(priv, 0x406028, tpc_per_gpc); + nv_wr32(priv, 0x405870, tpc_per_gpc); +} + +static void +gm20b_grctx_generate_main(struct gf100_gr_priv *priv, struct gf100_grctx *info) +{ + struct gf100_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass; + int idle_timeout_save; + int i, tmp; + + gf100_gr_mmio(priv, priv->fuc_sw_ctx); + + gf100_gr_wait_idle(priv); + + idle_timeout_save = nv_rd32(priv, 0x404154); + nv_wr32(priv, 0x404154, 0x00000000); + + oclass->attrib(info); + + oclass->unkn(priv); + + gm204_grctx_generate_tpcid(priv); + gm20b_grctx_generate_r406028(priv); + gk104_grctx_generate_r418bb8(priv); + + for (i = 0; i < 8; i++) + nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000); + + nv_wr32(priv, 0x405b00, (priv->tpc_total << 8) | priv->gpc_nr); + + gk104_grctx_generate_rop_active_fbps(priv); + nv_wr32(priv, 0x408908, nv_rd32(priv, 0x410108) | 0x80000000); + + for (tmp = 0, i = 0; i < priv->gpc_nr; i++) + tmp |= ((1 << priv->tpc_nr[i]) - 1) << (i * 4); + nv_wr32(priv, 0x4041c4, tmp); + + gm204_grctx_generate_405b60(priv); + + gf100_gr_wait_idle(priv); + + nv_wr32(priv, 0x404154, idle_timeout_save); + gf100_gr_wait_idle(priv); + + gf100_gr_mthd(priv, priv->fuc_method); + gf100_gr_wait_idle(priv); + + gf100_gr_icmd(priv, priv->fuc_bundle); + oclass->pagepool(info); + oclass->bundle(info); +} + +struct nvkm_oclass * +gm20b_grctx_oclass = &(struct gf100_grctx_oclass) { + .base.handle = NV_ENGCTX(GR, 0x2b), + .base.ofuncs = &(struct nvkm_ofuncs) { + .ctor = gf100_gr_context_ctor, + .dtor = gf100_gr_context_dtor, + .init = _nvkm_gr_context_init, + .fini = _nvkm_gr_context_fini, + .rd32 = _nvkm_gr_context_rd32, + .wr32 = _nvkm_gr_context_wr32, + }, + .main = gm20b_grctx_generate_main, + .unkn = gk104_grctx_generate_unkn, + .bundle = gm107_grctx_generate_bundle, + .bundle_size = 0x1800, + .bundle_min_gpm_fifo_depth = 0x182, + .bundle_token_limit = 0x1c0, + .pagepool = gm107_grctx_generate_pagepool, + .pagepool_size = 0x8000, + .attrib = gm107_grctx_generate_attrib, + .attrib_nr_max = 0x600, + .attrib_nr = 0x400, + .alpha_nr_max = 0xc00, + .alpha_nr = 0x800, +}.base; \ No newline at end of file diff --git a/drm/nouveau/nvkm/engine/gr/gf100.c b/drm/nouveau/nvkm/engine/gr/gf100.c index d5c148f..359501f 100644 --- a/drm/nouveau/nvkm/engine/gr/gf100.c +++ b/drm/nouveau/nvkm/engine/gr/gf100.c @@ -1720,6 +1720,7 @@ gf100_gr_ctor(struct nvkm_object *parent, struct nvkm_object *engine, case 0xd7: case 0xd9: /* 1/0/0/0, 1 */ case 0xea: /* gk20a */ + case 0x12b: /* gm20b */ priv->magic_not_rop_nr = 0x01; break; } diff --git a/drm/nouveau/nvkm/engine/gr/gf100.h b/drm/nouveau/nvkm/engine/gr/gf100.h index af4d70f..719c7ab 100644 --- a/drm/nouveau/nvkm/engine/gr/gf100.h +++ b/drm/nouveau/nvkm/engine/gr/gf100.h @@ -141,6 +141,12 @@ int gk104_gr_ctor(struct nvkm_object *, struct nvkm_object *, struct nvkm_object **); int gk104_gr_init(struct nvkm_object *);

+int gk20a_gr_ctor(struct nvkm_object *, struct nvkm_object *, + struct nvkm_oclass *, void *data, u32 size, + struct nvkm_object **); +void gk20a_gr_dtor(struct nvkm_object *); +int gk20a_gr_init(struct nvkm_object *); + int gm204_gr_init(struct nvkm_object *);

extern struct nvkm_ofuncs gf100_fermi_ofuncs; diff --git a/drm/nouveau/nvkm/engine/gr/gk20a.c b/drm/nouveau/nvkm/engine/gr/gk20a.c index 65f1089..0c6bf83 100644 --- a/drm/nouveau/nvkm/engine/gr/gk20a.c +++ b/drm/nouveau/nvkm/engine/gr/gk20a.c @@ -154,7 +154,7 @@ gk20a_gr_av_to_method(struct gf100_gr_fuc *fuc) return pack; }

-static int +int gk20a_gr_ctor(struct nvkm_object *parent, struct nvkm_object *engine, struct nvkm_oclass *oclass, void *data, u32 size, struct nvkm_object **pobject) @@ -204,7 +204,7 @@ gk20a_gr_ctor(struct nvkm_object *parent, struct nvkm_object *engine, return 0; }

-static void +void gk20a_gr_dtor(struct nvkm_object *object) { struct gf100_gr_priv *priv = (void *)object; @@ -240,7 +240,7 @@ gk20a_gr_set_hww_esr_report_mask(struct gf100_gr_priv *priv) nv_wr32(priv, 0x419e4c, 0x7f); }

-static int +int gk20a_gr_init(struct nvkm_object *object) { struct gk20a_gr_oclass *oclass = (void *)object->oclass; diff --git a/drm/nouveau/nvkm/engine/gr/gm20b.c b/drm/nouveau/nvkm/engine/gr/gm20b.c new file mode 100644 index 0000000..8976280 --- /dev/null +++ b/drm/nouveau/nvkm/engine/gr/gm20b.c @@ -0,0 +1,84 @@ +/* + * Copyright (c) 2015, NVIDIA CORPORATION. All rights reserved. + * + * 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 AUTHORS OR COPYRIGHT HOLDERS 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. + */ +#include "gk20a.h" +#include "ctxgf100.h" + +#include <nvif/class.h> +#include <subdev/timer.h> + +static struct nvkm_oclass +gm20b_gr_sclass[] = { + { FERMI_TWOD_A, &nvkm_object_ofuncs }, + { KEPLER_INLINE_TO_MEMORY_B, &nvkm_object_ofuncs }, + { MAXWELL_B, &gf100_fermi_ofuncs, gf100_gr_9097_omthds }, + { MAXWELL_COMPUTE_B, &nvkm_object_ofuncs, gf100_gr_90c0_omthds }, + {} +}; + +static void +gm20b_gr_init_gpc_mmu(struct gf100_gr_priv *priv) +{ + u32 val; + + /* TODO this needs to be removed once secure boot works */ + if (1) { + nv_wr32(priv, 0x100ce4, 0xffffffff); + } + + /* TODO update once secure boot works */ + val = nv_rd32(priv, 0x100c80); + val &= 0xf000087f; + nv_wr32(priv, 0x418880, val); + nv_wr32(priv, 0x418890, 0); + nv_wr32(priv, 0x418894, 0); + + nv_wr32(priv, 0x4188b0, nv_rd32(priv, 0x100cc4)); + nv_wr32(priv, 0x4188b4, nv_rd32(priv, 0x100cc8)); + nv_wr32(priv, 0x4188b8, nv_rd32(priv, 0x100ccc)); + + nv_wr32(priv, 0x4188ac, nv_rd32(priv, 0x100800)); +} + +static void +gm20b_gr_set_hww_esr_report_mask(struct gf100_gr_priv *priv) +{ + nv_wr32(priv, 0x419e44, 0xdffffe); + nv_wr32(priv, 0x419e4c, 0x5); +} + +struct nvkm_oclass * +gm20b_gr_oclass = &(struct gk20a_gr_oclass) { + .gf100 = { + .base.handle = NV_ENGINE(GR, 0x2b), + .base.ofuncs = &(struct nvkm_ofuncs) { + .ctor = gk20a_gr_ctor, + .dtor = gf100_gr_dtor, + .init = gk20a_gr_init, + .fini = _nvkm_gr_fini, + }, + .cclass = &gm20b_grctx_oclass, + .sclass = gm20b_gr_sclass, + .ppc_nr = 1, + }, + .init_gpc_mmu = gm20b_gr_init_gpc_mmu, + .set_hww_esr_report_mask = gm20b_gr_set_hww_esr_report_mask, +}.gf100.base;

Allow the platform driver to recognize GM20B.

Signed-off-by: Alexandre Courbot acourbot@nvidia.com --- drm/nouveau/nouveau_platform.c | 1 + 1 file changed, 1 insertion(+)

diff --git a/drm/nouveau/nouveau_platform.c b/drm/nouveau/nouveau_platform.c index dcfbbfa..7a39d44 100644 --- a/drm/nouveau/nouveau_platform.c +++ b/drm/nouveau/nouveau_platform.c @@ -252,6 +252,7 @@ static int nouveau_platform_remove(struct platform_device *pdev) #if IS_ENABLED(CONFIG_OF) static const struct of_device_id nouveau_platform_match[] = { { .compatible = "nvidia,gk20a" }, + { .compatible = "nvidia,gm20b" }, { } };