diff options
Diffstat (limited to 'drivers/gpu/drm/i915')
27 files changed, 331 insertions, 128 deletions
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 829b2a40a482..99e682563d47 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -10636,6 +10636,10 @@ skl_get_initial_plane_config(struct intel_crtc *crtc, val & PLANE_CTL_FLIP_HORIZONTAL) plane_config->rotation |= DRM_MODE_REFLECT_X; + /* 90/270 degree rotation would require extra work */ + if (drm_rotation_90_or_270(plane_config->rotation)) + goto error; + base = intel_de_read(dev_priv, PLANE_SURF(pipe, plane_id)) & 0xfffff000; plane_config->base = base; @@ -12874,10 +12878,11 @@ compute_sink_pipe_bpp(const struct drm_connector_state *conn_state, case 10 ... 11: bpp = 10 * 3; break; - case 12: + case 12 ... 16: bpp = 12 * 3; break; default: + MISSING_CASE(conn_state->max_bpc); return -EINVAL; } diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c index 8a9d0bdde1bf..40e9cb29233d 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.c +++ b/drivers/gpu/drm/i915/display/intel_psr.c @@ -1754,7 +1754,7 @@ void intel_psr_atomic_check(struct drm_connector *connector, return; intel_connector = to_intel_connector(connector); - dig_port = enc_to_dig_port(intel_attached_encoder(intel_connector)); + dig_port = enc_to_dig_port(to_intel_encoder(new_state->best_encoder)); if (dev_priv->psr.dp != &dig_port->dp) return; diff --git a/drivers/gpu/drm/i915/gem/i915_gem_domain.c b/drivers/gpu/drm/i915/gem/i915_gem_domain.c index 7c90a63c273d..fcce6909f201 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_domain.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_domain.c @@ -509,21 +509,6 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data, return -ENOENT; /* - * Already in the desired write domain? Nothing for us to do! - * - * We apply a little bit of cunning here to catch a broader set of - * no-ops. If obj->write_domain is set, we must be in the same - * obj->read_domains, and only that domain. Therefore, if that - * obj->write_domain matches the request read_domains, we are - * already in the same read/write domain and can skip the operation, - * without having to further check the requested write_domain. - */ - if (READ_ONCE(obj->write_domain) == read_domains) { - err = 0; - goto out; - } - - /* * Try to flush the object off the GPU without holding the lock. * We will repeat the flush holding the lock in the normal manner * to catch cases where we are gazumped. @@ -560,6 +545,19 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data, if (err) goto out; + /* + * Already in the desired write domain? Nothing for us to do! + * + * We apply a little bit of cunning here to catch a broader set of + * no-ops. If obj->write_domain is set, we must be in the same + * obj->read_domains, and only that domain. Therefore, if that + * obj->write_domain matches the request read_domains, we are + * already in the same read/write domain and can skip the operation, + * without having to further check the requested write_domain. + */ + if (READ_ONCE(obj->write_domain) == read_domains) + goto out_unpin; + err = i915_gem_object_lock_interruptible(obj, NULL); if (err) goto out_unpin; diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h index b5c15557cc87..d6711caa7f39 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h @@ -56,6 +56,8 @@ struct drm_i915_gem_object_ops { void (*truncate)(struct drm_i915_gem_object *obj); void (*writeback)(struct drm_i915_gem_object *obj); + int (*pread)(struct drm_i915_gem_object *obj, + const struct drm_i915_gem_pread *arg); int (*pwrite)(struct drm_i915_gem_object *obj, const struct drm_i915_gem_pwrite *arg); diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c b/drivers/gpu/drm/i915/gem/i915_gem_phys.c index 28147aab47b9..3a4dfe2ef1da 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_phys.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c @@ -134,6 +134,58 @@ i915_gem_object_put_pages_phys(struct drm_i915_gem_object *obj, vaddr, dma); } +static int +phys_pwrite(struct drm_i915_gem_object *obj, + const struct drm_i915_gem_pwrite *args) +{ + void *vaddr = sg_page(obj->mm.pages->sgl) + args->offset; + char __user *user_data = u64_to_user_ptr(args->data_ptr); + int err; + + err = i915_gem_object_wait(obj, + I915_WAIT_INTERRUPTIBLE | + I915_WAIT_ALL, + MAX_SCHEDULE_TIMEOUT); + if (err) + return err; + + /* + * We manually control the domain here and pretend that it + * remains coherent i.e. in the GTT domain, like shmem_pwrite. + */ + i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU); + + if (copy_from_user(vaddr, user_data, args->size)) + return -EFAULT; + + drm_clflush_virt_range(vaddr, args->size); + intel_gt_chipset_flush(&to_i915(obj->base.dev)->gt); + + i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU); + return 0; +} + +static int +phys_pread(struct drm_i915_gem_object *obj, + const struct drm_i915_gem_pread *args) +{ + void *vaddr = sg_page(obj->mm.pages->sgl) + args->offset; + char __user *user_data = u64_to_user_ptr(args->data_ptr); + int err; + + err = i915_gem_object_wait(obj, + I915_WAIT_INTERRUPTIBLE, + MAX_SCHEDULE_TIMEOUT); + if (err) + return err; + + drm_clflush_virt_range(vaddr, args->size); + if (copy_to_user(user_data, vaddr, args->size)) + return -EFAULT; + + return 0; +} + static void phys_release(struct drm_i915_gem_object *obj) { fput(obj->base.filp); @@ -144,6 +196,9 @@ static const struct drm_i915_gem_object_ops i915_gem_phys_ops = { .get_pages = i915_gem_object_get_pages_phys, .put_pages = i915_gem_object_put_pages_phys, + .pread = phys_pread, + .pwrite = phys_pwrite, + .release = phys_release, }; diff --git a/drivers/gpu/drm/i915/gt/intel_engine.h b/drivers/gpu/drm/i915/gt/intel_engine.h index 7c3a1012e702..760fefdfe392 100644 --- a/drivers/gpu/drm/i915/gt/intel_engine.h +++ b/drivers/gpu/drm/i915/gt/intel_engine.h @@ -245,22 +245,14 @@ static inline u32 *gen12_emit_pipe_control(u32 *batch, u32 flags0, u32 flags1, u } static inline u32 * -__gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags0, u32 flags1) +__gen8_emit_write_rcs(u32 *cs, u32 value, u32 offset, u32 flags0, u32 flags1) { - /* We're using qword write, offset should be aligned to 8 bytes. */ - GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8)); - - /* w/a for post sync ops following a GPGPU operation we - * need a prior CS_STALL, which is emitted by the flush - * following the batch. - */ *cs++ = GFX_OP_PIPE_CONTROL(6) | flags0; - *cs++ = flags1 | PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_GLOBAL_GTT_IVB; - *cs++ = gtt_offset; + *cs++ = flags1 | PIPE_CONTROL_QW_WRITE; + *cs++ = offset; *cs++ = 0; *cs++ = value; - /* We're thrashing one dword of HWS. */ - *cs++ = 0; + *cs++ = 0; /* We're thrashing one extra dword. */ return cs; } @@ -268,13 +260,38 @@ __gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags0, u32 f static inline u32* gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags) { - return __gen8_emit_ggtt_write_rcs(cs, value, gtt_offset, 0, flags); + /* We're using qword write, offset should be aligned to 8 bytes. */ + GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8)); + + return __gen8_emit_write_rcs(cs, + value, + gtt_offset, + 0, + flags | PIPE_CONTROL_GLOBAL_GTT_IVB); } static inline u32* gen12_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags0, u32 flags1) { - return __gen8_emit_ggtt_write_rcs(cs, value, gtt_offset, flags0, flags1); + /* We're using qword write, offset should be aligned to 8 bytes. */ + GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8)); + + return __gen8_emit_write_rcs(cs, + value, + gtt_offset, + flags0, + flags1 | PIPE_CONTROL_GLOBAL_GTT_IVB); +} + +static inline u32 * +__gen8_emit_flush_dw(u32 *cs, u32 value, u32 gtt_offset, u32 flags) +{ + *cs++ = (MI_FLUSH_DW + 1) | flags; + *cs++ = gtt_offset; + *cs++ = 0; + *cs++ = value; + + return cs; } static inline u32 * @@ -285,12 +302,10 @@ gen8_emit_ggtt_write(u32 *cs, u32 value, u32 gtt_offset, u32 flags) /* Offset should be aligned to 8 bytes for both (QW/DW) write types */ GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8)); - *cs++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW | flags; - *cs++ = gtt_offset | MI_FLUSH_DW_USE_GTT; - *cs++ = 0; - *cs++ = value; - - return cs; + return __gen8_emit_flush_dw(cs, + value, + gtt_offset | MI_FLUSH_DW_USE_GTT, + flags | MI_FLUSH_DW_OP_STOREDW); } static inline void __intel_engine_reset(struct intel_engine_cs *engine, diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c index 5bfb5f7ed02c..efdeb7b7b2a0 100644 --- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c @@ -371,7 +371,8 @@ static void __setup_engine_capabilities(struct intel_engine_cs *engine) * instances. */ if ((INTEL_GEN(i915) >= 11 && - engine->gt->info.vdbox_sfc_access & engine->mask) || + (engine->gt->info.vdbox_sfc_access & + BIT(engine->instance))) || (INTEL_GEN(i915) >= 9 && engine->instance == 0)) engine->uabi_capabilities |= I915_VIDEO_AND_ENHANCE_CLASS_CAPABILITY_SFC; diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index a32aabce7901..9bb16bdf93cf 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -3547,6 +3547,19 @@ static const struct intel_context_ops execlists_context_ops = { .destroy = execlists_context_destroy, }; +static u32 hwsp_offset(const struct i915_request *rq) +{ + const struct intel_timeline_cacheline *cl; + + /* Before the request is executed, the timeline/cachline is fixed */ + + cl = rcu_dereference_protected(rq->hwsp_cacheline, 1); + if (cl) + return cl->ggtt_offset; + + return rcu_dereference_protected(rq->timeline, 1)->hwsp_offset; +} + static int gen8_emit_init_breadcrumb(struct i915_request *rq) { u32 *cs; @@ -3569,7 +3582,7 @@ static int gen8_emit_init_breadcrumb(struct i915_request *rq) *cs++ = MI_NOOP; *cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT; - *cs++ = i915_request_timeline(rq)->hwsp_offset; + *cs++ = hwsp_offset(rq); *cs++ = 0; *cs++ = rq->fence.seqno - 1; @@ -4886,11 +4899,9 @@ gen8_emit_fini_breadcrumb_tail(struct i915_request *request, u32 *cs) return gen8_emit_wa_tail(request, cs); } -static u32 *emit_xcs_breadcrumb(struct i915_request *request, u32 *cs) +static u32 *emit_xcs_breadcrumb(struct i915_request *rq, u32 *cs) { - u32 addr = i915_request_active_timeline(request)->hwsp_offset; - - return gen8_emit_ggtt_write(cs, request->fence.seqno, addr, 0); + return gen8_emit_ggtt_write(cs, rq->fence.seqno, hwsp_offset(rq), 0); } static u32 *gen8_emit_fini_breadcrumb(struct i915_request *rq, u32 *cs) @@ -4909,7 +4920,7 @@ static u32 *gen8_emit_fini_breadcrumb_rcs(struct i915_request *request, u32 *cs) /* XXX flush+write+CS_STALL all in one upsets gem_concurrent_blt:kbl */ cs = gen8_emit_ggtt_write_rcs(cs, request->fence.seqno, - i915_request_active_timeline(request)->hwsp_offset, + hwsp_offset(request), PIPE_CONTROL_FLUSH_ENABLE | PIPE_CONTROL_CS_STALL); @@ -4921,7 +4932,7 @@ gen11_emit_fini_breadcrumb_rcs(struct i915_request *request, u32 *cs) { cs = gen8_emit_ggtt_write_rcs(cs, request->fence.seqno, - i915_request_active_timeline(request)->hwsp_offset, + hwsp_offset(request), PIPE_CONTROL_CS_STALL | PIPE_CONTROL_TILE_CACHE_FLUSH | PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH | @@ -4983,7 +4994,9 @@ gen12_emit_fini_breadcrumb_tail(struct i915_request *request, u32 *cs) static u32 *gen12_emit_fini_breadcrumb(struct i915_request *rq, u32 *cs) { - return gen12_emit_fini_breadcrumb_tail(rq, emit_xcs_breadcrumb(rq, cs)); + /* XXX Stalling flush before seqno write; post-sync not */ + cs = emit_xcs_breadcrumb(rq, __gen8_emit_flush_dw(cs, 0, 0, 0)); + return gen12_emit_fini_breadcrumb_tail(rq, cs); } static u32 * @@ -4991,7 +5004,7 @@ gen12_emit_fini_breadcrumb_rcs(struct i915_request *request, u32 *cs) { cs = gen12_emit_ggtt_write_rcs(cs, request->fence.seqno, - i915_request_active_timeline(request)->hwsp_offset, + hwsp_offset(request), PIPE_CONTROL0_HDC_PIPELINE_FLUSH, PIPE_CONTROL_CS_STALL | PIPE_CONTROL_TILE_CACHE_FLUSH | @@ -5444,6 +5457,7 @@ static void virtual_context_destroy(struct kref *kref) __execlists_context_fini(&ve->context); intel_context_fini(&ve->context); + intel_breadcrumbs_free(ve->base.breadcrumbs); intel_engine_free_request_pool(&ve->base); kfree(ve->bonds); diff --git a/drivers/gpu/drm/i915/gt/intel_mocs.c b/drivers/gpu/drm/i915/gt/intel_mocs.c index b8f56e62158e..313e51e7d4f7 100644 --- a/drivers/gpu/drm/i915/gt/intel_mocs.c +++ b/drivers/gpu/drm/i915/gt/intel_mocs.c @@ -243,8 +243,9 @@ static const struct drm_i915_mocs_entry tgl_mocs_table[] = { * only, __init_mocs_table() take care to program unused index with * this entry. */ - MOCS_ENTRY(1, LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), - L3_3_WB), + MOCS_ENTRY(I915_MOCS_PTE, + LE_0_PAGETABLE | LE_TC_0_PAGETABLE, + L3_1_UC), GEN11_MOCS_ENTRIES, /* Implicitly enable L1 - HDC:L1 + L3 + LLC */ diff --git a/drivers/gpu/drm/i915/gt/intel_rc6.c b/drivers/gpu/drm/i915/gt/intel_rc6.c index ab675d35030d..d7b8e4457fc2 100644 --- a/drivers/gpu/drm/i915/gt/intel_rc6.c +++ b/drivers/gpu/drm/i915/gt/intel_rc6.c @@ -56,9 +56,12 @@ static inline void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val) static void gen11_rc6_enable(struct intel_rc6 *rc6) { - struct intel_uncore *uncore = rc6_to_uncore(rc6); + struct intel_gt *gt = rc6_to_gt(rc6); + struct intel_uncore *uncore = gt->uncore; struct intel_engine_cs *engine; enum intel_engine_id id; + u32 pg_enable; + int i; /* 2b: Program RC6 thresholds.*/ set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 54 << 16 | 85); @@ -102,10 +105,19 @@ static void gen11_rc6_enable(struct intel_rc6 *rc6) GEN6_RC_CTL_RC6_ENABLE | GEN6_RC_CTL_EI_MODE(1); - set(uncore, GEN9_PG_ENABLE, - GEN9_RENDER_PG_ENABLE | - GEN9_MEDIA_PG_ENABLE | - GEN11_MEDIA_SAMPLER_PG_ENABLE); + pg_enable = + GEN9_RENDER_PG_ENABLE | + GEN9_MEDIA_PG_ENABLE | + GEN11_MEDIA_SAMPLER_PG_ENABLE; + + if (INTEL_GEN(gt->i915) >= 12) { + for (i = 0; i < I915_MAX_VCS; i++) + if (HAS_ENGINE(gt, _VCS(i))) + pg_enable |= (VDN_HCP_POWERGATE_ENABLE(i) | + VDN_MFX_POWERGATE_ENABLE(i)); + } + + set(uncore, GEN9_PG_ENABLE, pg_enable); } static void gen9_rc6_enable(struct intel_rc6 *rc6) diff --git a/drivers/gpu/drm/i915/gt/intel_timeline.c b/drivers/gpu/drm/i915/gt/intel_timeline.c index a2f74cefe4c3..7ea94d201fe6 100644 --- a/drivers/gpu/drm/i915/gt/intel_timeline.c +++ b/drivers/gpu/drm/i915/gt/intel_timeline.c @@ -188,10 +188,14 @@ cacheline_alloc(struct intel_timeline_hwsp *hwsp, unsigned int cacheline) return cl; } -static void cacheline_acquire(struct intel_timeline_cacheline *cl) +static void cacheline_acquire(struct intel_timeline_cacheline *cl, + u32 ggtt_offset) { - if (cl) - i915_active_acquire(&cl->active); + if (!cl) + return; + + cl->ggtt_offset = ggtt_offset; + i915_active_acquire(&cl->active); } static void cacheline_release(struct intel_timeline_cacheline *cl) @@ -340,7 +344,7 @@ int intel_timeline_pin(struct intel_timeline *tl, struct i915_gem_ww_ctx *ww) GT_TRACE(tl->gt, "timeline:%llx using HWSP offset:%x\n", tl->fence_context, tl->hwsp_offset); - cacheline_acquire(tl->hwsp_cacheline); + cacheline_acquire(tl->hwsp_cacheline, tl->hwsp_offset); if (atomic_fetch_inc(&tl->pin_count)) { cacheline_release(tl->hwsp_cacheline); __i915_vma_unpin(tl->hwsp_ggtt); @@ -515,7 +519,7 @@ __intel_timeline_get_seqno(struct intel_timeline *tl, GT_TRACE(tl->gt, "timeline:%llx using HWSP offset:%x\n", tl->fence_context, tl->hwsp_offset); - cacheline_acquire(cl); + cacheline_acquire(cl, tl->hwsp_offset); tl->hwsp_cacheline = cl; *seqno = timeline_advance(tl); @@ -573,9 +577,7 @@ int intel_timeline_read_hwsp(struct i915_request *from, if (err) goto out; - *hwsp = i915_ggtt_offset(cl->hwsp->vma) + - ptr_unmask_bits(cl->vaddr, CACHELINE_BITS) * CACHELINE_BYTES; - + *hwsp = cl->ggtt_offset; out: i915_active_release(&cl->active); return err; diff --git a/drivers/gpu/drm/i915/gt/intel_timeline_types.h b/drivers/gpu/drm/i915/gt/intel_timeline_types.h index 02181c5020db..4474f487f589 100644 --- a/drivers/gpu/drm/i915/gt/intel_timeline_types.h +++ b/drivers/gpu/drm/i915/gt/intel_timeline_types.h @@ -94,6 +94,8 @@ struct intel_timeline_cacheline { struct intel_timeline_hwsp *hwsp; void *vaddr; + u32 ggtt_offset; + struct rcu_head rcu; }; diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 6c580d0d9ea8..4a3bde7c9f21 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -131,8 +131,10 @@ static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) return; } - if (wal->list) + if (wal->list) { memcpy(list, wal->list, sizeof(*wa) * wal->count); + kfree(wal->list); + } wal->list = list; } diff --git a/drivers/gpu/drm/i915/gvt/display.c b/drivers/gpu/drm/i915/gvt/display.c index 7ba16ddfe75f..d7898e87791f 100644 --- a/drivers/gpu/drm/i915/gvt/display.c +++ b/drivers/gpu/drm/i915/gvt/display.c @@ -164,7 +164,7 @@ static unsigned char virtual_dp_monitor_edid[GVT_EDID_NUM][EDID_SIZE] = { /* let the virtual display supports DP1.2 */ static u8 dpcd_fix_data[DPCD_HEADER_SIZE] = { - 0x12, 0x014, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 + 0x12, 0x014, 0x84, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }; static void emulate_monitor_status_change(struct intel_vgpu *vgpu) diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 3be37e6fe33d..eb342a759943 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -1489,7 +1489,8 @@ static int hws_pga_write(struct intel_vgpu *vgpu, unsigned int offset, const struct intel_engine_cs *engine = intel_gvt_render_mmio_to_engine(vgpu->gvt, offset); - if (!intel_gvt_ggtt_validate_range(vgpu, value, I915_GTT_PAGE_SIZE)) { + if (value != 0 && + !intel_gvt_ggtt_validate_range(vgpu, value, I915_GTT_PAGE_SIZE)) { gvt_vgpu_err("write invalid HWSP address, reg:0x%x, value:0x%x\n", offset, value); return -EINVAL; @@ -1650,6 +1651,34 @@ static int edp_psr_imr_iir_write(struct intel_vgpu *vgpu, return 0; } +/** + * FixMe: + * If guest fills non-priv batch buffer on ApolloLake/Broxton as Mesa i965 did: + * 717e7539124d (i965: Use a WC map and memcpy for the batch instead of pwrite.) + * Due to the missing flush of bb filled by VM vCPU, host GPU hangs on executing + * these MI_BATCH_BUFFER. + * Temporarily workaround this by setting SNOOP bit for PAT3 used by PPGTT + * PML4 PTE: PAT(0) PCD(1) PWT(1). + * The performance is still expected to be low, will need further improvement. + */ +static int bxt_ppat_low_write(struct intel_vgpu *vgpu, unsigned int offset, + void *p_data, unsigned int bytes) +{ + u64 pat = + GEN8_PPAT(0, CHV_PPAT_SNOOP) | + GEN8_PPAT(1, 0) | + GEN8_PPAT(2, 0) | + GEN8_PPAT(3, CHV_PPAT_SNOOP) | + GEN8_PPAT(4, CHV_PPAT_SNOOP) | + GEN8_PPAT(5, CHV_PPAT_SNOOP) | + GEN8_PPAT(6, CHV_PPAT_SNOOP) | + GEN8_PPAT(7, CHV_PPAT_SNOOP); + + vgpu_vreg(vgpu, offset) = lower_32_bits(pat); + + return 0; +} + static int guc_status_read(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes) @@ -2812,7 +2841,7 @@ static int init_bdw_mmio_info(struct intel_gvt *gvt) MMIO_DH(GEN6_PCODE_MAILBOX, D_BDW_PLUS, NULL, mailbox_write); - MMIO_D(GEN8_PRIVATE_PAT_LO, D_BDW_PLUS); + MMIO_D(GEN8_PRIVATE_PAT_LO, D_BDW_PLUS & ~D_BXT); MMIO_D(GEN8_PRIVATE_PAT_HI, D_BDW_PLUS); MMIO_D(GAMTARBMODE, D_BDW_PLUS); @@ -3139,7 +3168,7 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) NULL, NULL); MMIO_DFH(GAMT_CHKN_BIT_REG, D_KBL | D_CFL, F_CMD_ACCESS, NULL, NULL); - MMIO_D(GEN9_CTX_PREEMPT_REG, D_SKL_PLUS); + MMIO_D(GEN9_CTX_PREEMPT_REG, D_SKL_PLUS & ~D_BXT); return 0; } @@ -3313,9 +3342,21 @@ static int init_bxt_mmio_info(struct intel_gvt *gvt) MMIO_D(GEN8_PUSHBUS_SHIFT, D_BXT); MMIO_D(GEN6_GFXPAUSE, D_BXT); MMIO_DFH(GEN8_L3SQCREG1, D_BXT, F_CMD_ACCESS, NULL, NULL); + MMIO_DFH(GEN8_L3CNTLREG, D_BXT, F_CMD_ACCESS, NULL, NULL); + MMIO_DFH(_MMIO(0x20D8), D_BXT, F_CMD_ACCESS, NULL, NULL); + MMIO_F(GEN8_RING_CS_GPR(RENDER_RING_BASE, 0), 0x40, F_CMD_ACCESS, + 0, 0, D_BXT, NULL, NULL); + MMIO_F(GEN8_RING_CS_GPR(GEN6_BSD_RING_BASE, 0), 0x40, F_CMD_ACCESS, + 0, 0, D_BXT, NULL, NULL); + MMIO_F(GEN8_RING_CS_GPR(BLT_RING_BASE, 0), 0x40, F_CMD_ACCESS, + 0, 0, D_BXT, NULL, NULL); + MMIO_F(GEN8_RING_CS_GPR(VEBOX_RING_BASE, 0), 0x40, F_CMD_ACCESS, + 0, 0, D_BXT, NULL, NULL); MMIO_DFH(GEN9_CTX_PREEMPT_REG, D_BXT, F_CMD_ACCESS, NULL, NULL); + MMIO_DH(GEN8_PRIVATE_PAT_LO, D_BXT, NULL, bxt_ppat_low_write); + return 0; } diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c index ad8a9df49f29..778eb8cab610 100644 --- a/drivers/gpu/drm/i915/gvt/kvmgt.c +++ b/drivers/gpu/drm/i915/gvt/kvmgt.c @@ -829,8 +829,10 @@ static int intel_vgpu_open(struct mdev_device *mdev) /* Take a module reference as mdev core doesn't take * a reference for vendor driver. */ - if (!try_module_get(THIS_MODULE)) + if (!try_module_get(THIS_MODULE)) { + ret = -ENODEV; goto undo_group; + } ret = kvmgt_guest_init(mdev); if (ret) diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c index 1570eb8aa978..aed2ef6466a2 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.c +++ b/drivers/gpu/drm/i915/gvt/scheduler.c @@ -1277,7 +1277,7 @@ void intel_vgpu_clean_submission(struct intel_vgpu *vgpu) i915_context_ppgtt_root_restore(s, i915_vm_to_ppgtt(s->shadow[0]->vm)); for_each_engine(engine, vgpu->gvt->gt, id) - intel_context_unpin(s->shadow[id]); + intel_context_put(s->shadow[id]); kmem_cache_destroy(s->workloads); } @@ -1369,11 +1369,6 @@ int intel_vgpu_setup_submission(struct intel_vgpu *vgpu) ce->ring = __intel_context_ring_size(ring_size); } - ret = intel_context_pin(ce); - intel_context_put(ce); - if (ret) - goto out_shadow_ctx; - s->shadow[i] = ce; } @@ -1405,7 +1400,6 @@ out_shadow_ctx: if (IS_ERR(s->shadow[i])) break; - intel_context_unpin(s->shadow[i]); intel_context_put(s->shadow[i]); } i915_vm_put(&ppgtt->vm); @@ -1479,6 +1473,7 @@ void intel_vgpu_destroy_workload(struct intel_vgpu_workload *workload) { struct intel_vgpu_submission *s = &workload->vgpu->submission; + intel_context_unpin(s->shadow[workload->engine->id]); release_shadow_batch_buffer(workload); release_shadow_wa_ctx(&workload->wa_ctx); @@ -1724,6 +1719,12 @@ intel_vgpu_create_workload(struct intel_vgpu *vgpu, return ERR_PTR(ret); } + ret = intel_context_pin(s->shadow[engine->id]); + if (ret) { + intel_vgpu_destroy_workload(workload); + return ERR_PTR(ret); + } + return workload; } diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index f6d7e33c7099..399582aeeefb 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -439,7 +439,8 @@ static struct intel_vgpu *__intel_gvt_create_vgpu(struct intel_gvt *gvt, if (IS_BROADWELL(dev_priv)) ret = intel_gvt_hypervisor_set_edid(vgpu, PORT_B); - else + /* FixMe: Re-enable APL/BXT once vfio_edid enabled */ + else if (!IS_BROXTON(dev_priv)) ret = intel_gvt_hypervisor_set_edid(vgpu, PORT_D); if (ret) goto out_clean_sched_policy; diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index bb0c12975f38..58276694c848 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -180,30 +180,6 @@ try_again: } static int -i915_gem_phys_pwrite(struct drm_i915_gem_object *obj, - struct drm_i915_gem_pwrite *args, - struct drm_file *file) -{ - void *vaddr = sg_page(obj->mm.pages->sgl) + args->offset; - char __user *user_data = u64_to_user_ptr(args->data_ptr); - - /* - * We manually control the domain here and pretend that it - * remains coherent i.e. in the GTT domain, like shmem_pwrite. - */ - i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU); - - if (copy_from_user(vaddr, user_data, args->size)) - return -EFAULT; - - drm_clflush_virt_range(vaddr, args->size); - intel_gt_chipset_flush(&to_i915(obj->base.dev)->gt); - - i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU); - return 0; -} - -static int i915_gem_create(struct drm_file *file, struct intel_memory_region *mr, u64 *size_p, @@ -527,6 +503,12 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data, trace_i915_gem_object_pread(obj, args->offset, args->size); + ret = -ENODEV; + if (obj->ops->pread) + ret = obj->ops->pread(obj, args); + if (ret != -ENODEV) + goto out; + ret = i915_gem_object_wait(obj, I915_WAIT_INTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT); @@ -866,8 +848,6 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data, if (ret == -EFAULT || ret == -ENOSPC) { if (i915_gem_object_has_struct_page(obj)) ret = i915_gem_shmem_pwrite(obj, args); - else - ret = i915_gem_phys_pwrite(obj, args, file); } i915_gem_object_unpin_pages(obj); diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c index 366ddfc8df6b..fb5e30de78c2 100644 --- a/drivers/gpu/drm/i915/i915_pci.c +++ b/drivers/gpu/drm/i915/i915_pci.c @@ -389,6 +389,7 @@ static const struct intel_device_info ilk_m_info = { GEN5_FEATURES, PLATFORM(INTEL_IRONLAKE), .is_mobile = 1, + .has_rps = true, .display.has_fbc = 1, }; diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index d805d4da6181..664f3bf9af03 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -8971,10 +8971,6 @@ enum { #define GEN9_PWRGT_MEDIA_STATUS_MASK (1 << 0) #define GEN9_PWRGT_RENDER_STATUS_MASK (1 << 1) -#define POWERGATE_ENABLE _MMIO(0xa210) -#define VDN_HCP_POWERGATE_ENABLE(n) BIT(((n) * 2) + 3) -#define VDN_MFX_POWERGATE_ENABLE(n) BIT(((n) * 2) + 4) - #define GTFIFODBG _MMIO(0x120000) #define GT_FIFO_SBDEDICATE_FREE_ENTRY_CHV (0x1f << 20) #define GT_FIFO_FREE_ENTRIES_CHV (0x7f << 13) @@ -9114,9 +9110,11 @@ enum { #define GEN9_MEDIA_PG_IDLE_HYSTERESIS _MMIO(0xA0C4) #define GEN9_RENDER_PG_IDLE_HYSTERESIS _MMIO(0xA0C8) #define GEN9_PG_ENABLE _MMIO(0xA210) -#define GEN9_RENDER_PG_ENABLE REG_BIT(0) -#define GEN9_MEDIA_PG_ENABLE REG_BIT(1) -#define GEN11_MEDIA_SAMPLER_PG_ENABLE REG_BIT(2) +#define GEN9_RENDER_PG_ENABLE REG_BIT(0) +#define GEN9_MEDIA_PG_ENABLE REG_BIT(1) +#define GEN11_MEDIA_SAMPLER_PG_ENABLE REG_BIT(2) +#define VDN_HCP_POWERGATE_ENABLE(n) REG_BIT(3 + 2 * (n)) +#define VDN_MFX_POWERGATE_ENABLE(n) REG_BIT(4 + 2 * (n)) #define GEN8_PUSHBUS_CONTROL _MMIO(0xA248) #define GEN8_PUSHBUS_ENABLE _MMIO(0xA250) #define GEN8_PUSHBUS_SHIFT _MMIO(0xA25C) diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index ffb5287e055a..caa9b041616b 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -314,8 +314,10 @@ static void __vma_release(struct dma_fence_work *work) { struct i915_vma_work *vw = container_of(work, typeof(*vw), base); - if (vw->pinned) + if (vw->pinned) { __i915_gem_object_unpin_pages(vw->pinned); + i915_gem_object_put(vw->pinned); + } i915_vm_free_pt_stash(vw->vm, &vw->stash); i915_vm_put(vw->vm); @@ -431,7 +433,7 @@ int i915_vma_bind(struct i915_vma *vma, if (vma->obj) { __i915_gem_object_pin_pages(vma->obj); - work->pinned = vma->obj; + work->pinned = i915_gem_object_get(vma->obj); } } else { vma->ops->bind_vma(vma->vm, NULL, vma, cache_level, bind_flags); diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c index 6b5e9d88646d..180e1078ef7c 100644 --- a/drivers/gpu/drm/i915/intel_memory_region.c +++ b/drivers/gpu/drm/i915/intel_memory_region.c @@ -87,7 +87,7 @@ __intel_memory_region_get_pages_buddy(struct intel_memory_region *mem, min_order = ilog2(size) - ilog2(mem->mm.chunk_size); } - if (size > BIT(mem->mm.max_order) * mem->mm.chunk_size) + if (size > mem->mm.size) return -E2BIG; n_pages = size >> ilog2(mem->mm.chunk_size); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 34e0d22d456b..cfb806767fc5 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -7118,23 +7118,10 @@ static void icl_init_clock_gating(struct drm_i915_private *dev_priv) static void tgl_init_clock_gating(struct drm_i915_private *dev_priv) { - u32 vd_pg_enable = 0; - unsigned int i; - /* Wa_1409120013:tgl */ I915_WRITE(ILK_DPFC_CHICKEN, ILK_DPFC_CHICKEN_COMP_DUMMY_PIXEL); - /* This is not a WA. Enable VD HCP & MFX_ENC powergate */ - for (i = 0; i < I915_MAX_VCS; i++) { - if (HAS_ENGINE(&dev_priv->gt, _VCS(i))) - vd_pg_enable |= VDN_HCP_POWERGATE_ENABLE(i) | - VDN_MFX_POWERGATE_ENABLE(i); - } - - I915_WRITE(POWERGATE_ENABLE, - I915_READ(POWERGATE_ENABLE) | vd_pg_enable); - /* Wa_1409825376:tgl (pre-prod)*/ if (IS_TGL_DISP_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_B1)) I915_WRITE(GEN9_CLKGATE_DIS_3, I915_READ(GEN9_CLKGATE_DIS_3) | diff --git a/drivers/gpu/drm/i915/selftests/i915_request.c b/drivers/gpu/drm/i915/selftests/i915_request.c index 64bbb8288249..e424a6d1a68c 100644 --- a/drivers/gpu/drm/i915/selftests/i915_request.c +++ b/drivers/gpu/drm/i915/selftests/i915_request.c @@ -2293,8 +2293,10 @@ static int perf_request_latency(void *arg) struct intel_context *ce; ce = intel_context_create(engine); - if (IS_ERR(ce)) + if (IS_ERR(ce)) { + err = PTR_ERR(ce); goto out; + } err = intel_context_pin(ce); if (err) { @@ -2467,8 +2469,10 @@ static int perf_series_engines(void *arg) struct intel_context *ce; ce = intel_context_create(engine); - if (IS_ERR(ce)) + if (IS_ERR(ce)) { + err = PTR_ERR(ce); goto out; + } err = intel_context_pin(ce); if (err) { diff --git a/drivers/gpu/drm/i915/selftests/intel_memory_region.c b/drivers/gpu/drm/i915/selftests/intel_memory_region.c index 334b0648e253..0aeba8e3af28 100644 --- a/drivers/gpu/drm/i915/selftests/intel_memory_region.c +++ b/drivers/gpu/drm/i915/selftests/intel_memory_region.c @@ -261,6 +261,82 @@ err_close_objects: return err; } +static int igt_mock_splintered_region(void *arg) +{ + struct intel_memory_region *mem = arg; + struct drm_i915_private *i915 = mem->i915; + struct drm_i915_gem_object *obj; + unsigned int expected_order; + LIST_HEAD(objects); + u64 size; + int err = 0; + + /* + * Sanity check we can still allocate everything even if the + * mm.max_order != mm.size. i.e our starting address space size is not a + * power-of-two. + */ + + size = (SZ_4G - 1) & PAGE_MASK; + mem = mock_region_create(i915, 0, size, PAGE_SIZE, 0); + if (IS_ERR(mem)) + return PTR_ERR(mem); + + if (mem->mm.size != size) { + pr_err("%s size mismatch(%llu != %llu)\n", + __func__, mem->mm.size, size); + err = -EINVAL; + goto out_put; + } + + expected_order = get_order(rounddown_pow_of_two(size)); + if (mem->mm.max_order != expected_order) { + pr_err("%s order mismatch(%u != %u)\n", + __func__, mem->mm.max_order, expected_order); + err = -EINVAL; + goto out_put; + } + + obj = igt_object_create(mem, &objects, size, 0); + if (IS_ERR(obj)) { + err = PTR_ERR(obj); + goto out_close; + } + + close_objects(mem, &objects); + + /* + * While we should be able allocate everything without any flag + * restrictions, if we consider I915_BO_ALLOC_CONTIGUOUS then we are + * actually limited to the largest power-of-two for the region size i.e + * max_order, due to the inner workings of the buddy allocator. So make + * sure that does indeed hold true. + */ + + obj = igt_object_create(mem, &objects, size, I915_BO_ALLOC_CONTIGUOUS); + if (!IS_ERR(obj)) { + pr_err("%s too large contiguous allocation was not rejected\n", + __func__); + err = -EINVAL; + goto out_close; + } + + obj = igt_object_create(mem, &objects, rounddown_pow_of_two(size), + I915_BO_ALLOC_CONTIGUOUS); + if (IS_ERR(obj)) { + pr_err("%s largest possible contiguous allocation failed\n", + __func__); + err = PTR_ERR(obj); + goto out_close; + } + +out_close: + close_objects(mem, &objects); +out_put: + intel_memory_region_put(mem); + return err; +} + static int igt_gpu_write_dw(struct intel_context *ce, struct i915_vma *vma, u32 dword, @@ -771,6 +847,7 @@ int intel_memory_region_mock_selftests(void) static const struct i915_subtest tests[] = { SUBTEST(igt_mock_fill), SUBTEST(igt_mock_contiguous), + SUBTEST(igt_mock_splintered_region), }; struct intel_memory_region *mem; struct drm_i915_private *i915; diff --git a/drivers/gpu/drm/i915/selftests/mock_region.c b/drivers/gpu/drm/i915/selftests/mock_region.c index 09660f5a0a4c..979d96f27c43 100644 --- a/drivers/gpu/drm/i915/selftests/mock_region.c +++ b/drivers/gpu/drm/i915/selftests/mock_region.c @@ -24,7 +24,7 @@ mock_object_create(struct intel_memory_region *mem, struct drm_i915_private *i915 = mem->i915; struct drm_i915_gem_object *obj; - if (size > BIT(mem->mm.max_order) * mem->mm.chunk_size) + if (size > mem->mm.size) return ERR_PTR(-E2BIG); obj = i915_gem_object_alloc(); |