diff options
34 files changed, 362 insertions, 260 deletions
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index fcaf9c9..7de069af 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -60,7 +60,7 @@ | X86_CR4_PSE | X86_CR4_PAE | X86_CR4_MCE \ | X86_CR4_PGE | X86_CR4_PCE | X86_CR4_OSFXSR | X86_CR4_PCIDE \ | X86_CR4_OSXSAVE | X86_CR4_SMEP | X86_CR4_FSGSBASE \ - | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE)) + | X86_CR4_OSXMMEXCPT | X86_CR4_VMXE | X86_CR4_SMAP)) #define CR8_RESERVED_BITS (~(unsigned long)X86_CR8_TPR) diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c index bea6067..f47a104 100644 --- a/arch/x86/kvm/cpuid.c +++ b/arch/x86/kvm/cpuid.c @@ -308,7 +308,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, const u32 kvm_supported_word9_x86_features = F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) | F(BMI2) | F(ERMS) | f_invpcid | F(RTM) | f_mpx | F(RDSEED) | - F(ADX); + F(ADX) | F(SMAP); /* all calls to cpuid_count() should be made on the same cpu */ get_cpu(); diff --git a/arch/x86/kvm/cpuid.h b/arch/x86/kvm/cpuid.h index a2a1bb7..eeecbed 100644 --- a/arch/x86/kvm/cpuid.h +++ b/arch/x86/kvm/cpuid.h @@ -48,6 +48,14 @@ static inline bool guest_cpuid_has_smep(struct kvm_vcpu *vcpu) return best && (best->ebx & bit(X86_FEATURE_SMEP)); } +static inline bool guest_cpuid_has_smap(struct kvm_vcpu *vcpu) +{ + struct kvm_cpuid_entry2 *best; + + best = kvm_find_cpuid_entry(vcpu, 7, 0); + return best && (best->ebx & bit(X86_FEATURE_SMAP)); +} + static inline bool guest_cpuid_has_fsgsbase(struct kvm_vcpu *vcpu) { struct kvm_cpuid_entry2 *best; diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index f5704d9..813d310 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -3601,20 +3601,27 @@ static void reset_rsvds_bits_mask_ept(struct kvm_vcpu *vcpu, } } -static void update_permission_bitmask(struct kvm_vcpu *vcpu, +void update_permission_bitmask(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, bool ept) { unsigned bit, byte, pfec; u8 map; - bool fault, x, w, u, wf, uf, ff, smep; + bool fault, x, w, u, wf, uf, ff, smapf, cr4_smap, cr4_smep, smap = 0; - smep = kvm_read_cr4_bits(vcpu, X86_CR4_SMEP); + cr4_smep = kvm_read_cr4_bits(vcpu, X86_CR4_SMEP); + cr4_smap = kvm_read_cr4_bits(vcpu, X86_CR4_SMAP); for (byte = 0; byte < ARRAY_SIZE(mmu->permissions); ++byte) { pfec = byte << 1; map = 0; wf = pfec & PFERR_WRITE_MASK; uf = pfec & PFERR_USER_MASK; ff = pfec & PFERR_FETCH_MASK; + /* + * PFERR_RSVD_MASK bit is set in PFEC if the access is not + * subject to SMAP restrictions, and cleared otherwise. The + * bit is only meaningful if the SMAP bit is set in CR4. + */ + smapf = !(pfec & PFERR_RSVD_MASK); for (bit = 0; bit < 8; ++bit) { x = bit & ACC_EXEC_MASK; w = bit & ACC_WRITE_MASK; @@ -3626,12 +3633,33 @@ static void update_permission_bitmask(struct kvm_vcpu *vcpu, /* Allow supervisor writes if !cr0.wp */ w |= !is_write_protection(vcpu) && !uf; /* Disallow supervisor fetches of user code if cr4.smep */ - x &= !(smep && u && !uf); + x &= !(cr4_smep && u && !uf); + + /* + * SMAP:kernel-mode data accesses from user-mode + * mappings should fault. A fault is considered + * as a SMAP violation if all of the following + * conditions are ture: + * - X86_CR4_SMAP is set in CR4 + * - An user page is accessed + * - Page fault in kernel mode + * - if CPL = 3 or X86_EFLAGS_AC is clear + * + * Here, we cover the first three conditions. + * The fourth is computed dynamically in + * permission_fault() and is in smapf. + * + * Also, SMAP does not affect instruction + * fetches, add the !ff check here to make it + * clearer. + */ + smap = cr4_smap && u && !uf && !ff; } else /* Not really needed: no U/S accesses on ept */ u = 1; - fault = (ff && !x) || (uf && !u) || (wf && !w); + fault = (ff && !x) || (uf && !u) || (wf && !w) || + (smapf && smap); map |= fault << bit; } mmu->permissions[byte] = map; diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h index 2926152..3842e70 100644 --- a/arch/x86/kvm/mmu.h +++ b/arch/x86/kvm/mmu.h @@ -44,11 +44,17 @@ #define PT_DIRECTORY_LEVEL 2 #define PT_PAGE_TABLE_LEVEL 1 -#define PFERR_PRESENT_MASK (1U << 0) -#define PFERR_WRITE_MASK (1U << 1) -#define PFERR_USER_MASK (1U << 2) -#define PFERR_RSVD_MASK (1U << 3) -#define PFERR_FETCH_MASK (1U << 4) +#define PFERR_PRESENT_BIT 0 +#define PFERR_WRITE_BIT 1 +#define PFERR_USER_BIT 2 +#define PFERR_RSVD_BIT 3 +#define PFERR_FETCH_BIT 4 + +#define PFERR_PRESENT_MASK (1U << PFERR_PRESENT_BIT) +#define PFERR_WRITE_MASK (1U << PFERR_WRITE_BIT) +#define PFERR_USER_MASK (1U << PFERR_USER_BIT) +#define PFERR_RSVD_MASK (1U << PFERR_RSVD_BIT) +#define PFERR_FETCH_MASK (1U << PFERR_FETCH_BIT) int kvm_mmu_get_spte_hierarchy(struct kvm_vcpu *vcpu, u64 addr, u64 sptes[4]); void kvm_mmu_set_mmio_spte_mask(u64 mmio_mask); @@ -73,6 +79,8 @@ int handle_mmio_page_fault_common(struct kvm_vcpu *vcpu, u64 addr, bool direct); void kvm_init_shadow_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *context); void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, struct kvm_mmu *context, bool execonly); +void update_permission_bitmask(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, + bool ept); static inline unsigned int kvm_mmu_available_pages(struct kvm *kvm) { @@ -110,10 +118,30 @@ static inline bool is_write_protection(struct kvm_vcpu *vcpu) * Will a fault with a given page-fault error code (pfec) cause a permission * fault with the given access (in ACC_* format)? */ -static inline bool permission_fault(struct kvm_mmu *mmu, unsigned pte_access, - unsigned pfec) +static inline bool permission_fault(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, + unsigned pte_access, unsigned pfec) { - return (mmu->permissions[pfec >> 1] >> pte_access) & 1; + int cpl = kvm_x86_ops->get_cpl(vcpu); + unsigned long rflags = kvm_x86_ops->get_rflags(vcpu); + + /* + * If CPL < 3, SMAP prevention are disabled if EFLAGS.AC = 1. + * + * If CPL = 3, SMAP applies to all supervisor-mode data accesses + * (these are implicit supervisor accesses) regardless of the value + * of EFLAGS.AC. + * + * This computes (cpl < 3) && (rflags & X86_EFLAGS_AC), leaving + * the result in X86_EFLAGS_AC. We then insert it in place of + * the PFERR_RSVD_MASK bit; this bit will always be zero in pfec, + * but it will be one in index if SMAP checks are being overridden. + * It is important to keep this branchless. + */ + unsigned long smap = (cpl - 3) & (rflags & X86_EFLAGS_AC); + int index = (pfec >> 1) + + (smap >> (X86_EFLAGS_AC_BIT - PFERR_RSVD_BIT + 1)); + + return (mmu->permissions[index] >> pte_access) & 1; } void kvm_mmu_invalidate_zap_all_pages(struct kvm *kvm); diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h index b1e6c1b..123efd3 100644 --- a/arch/x86/kvm/paging_tmpl.h +++ b/arch/x86/kvm/paging_tmpl.h @@ -353,7 +353,7 @@ retry_walk: walker->ptes[walker->level - 1] = pte; } while (!is_last_gpte(mmu, walker->level, pte)); - if (unlikely(permission_fault(mmu, pte_access, access))) { + if (unlikely(permission_fault(vcpu, mmu, pte_access, access))) { errcode |= PFERR_PRESENT_MASK; goto error; } diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 1320e0f..1f68c58 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -3484,13 +3484,14 @@ static int vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4) hw_cr4 &= ~X86_CR4_PAE; hw_cr4 |= X86_CR4_PSE; /* - * SMEP is disabled if CPU is in non-paging mode in - * hardware. However KVM always uses paging mode to + * SMEP/SMAP is disabled if CPU is in non-paging mode + * in hardware. However KVM always uses paging mode to * emulate guest non-paging mode with TDP. - * To emulate this behavior, SMEP needs to be manually - * disabled when guest switches to non-paging mode. + * To emulate this behavior, SMEP/SMAP needs to be + * manually disabled when guest switches to non-paging + * mode. */ - hw_cr4 &= ~X86_CR4_SMEP; + hw_cr4 &= ~(X86_CR4_SMEP | X86_CR4_SMAP); } else if (!(cr4 & X86_CR4_PAE)) { hw_cr4 &= ~X86_CR4_PAE; } diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 9d1b5cd..8b8fc0b 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -652,6 +652,9 @@ int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4) if (!guest_cpuid_has_smep(vcpu) && (cr4 & X86_CR4_SMEP)) return 1; + if (!guest_cpuid_has_smap(vcpu) && (cr4 & X86_CR4_SMAP)) + return 1; + if (!guest_cpuid_has_fsgsbase(vcpu) && (cr4 & X86_CR4_FSGSBASE)) return 1; @@ -680,6 +683,9 @@ int kvm_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4) (!(cr4 & X86_CR4_PCIDE) && (old_cr4 & X86_CR4_PCIDE))) kvm_mmu_reset_context(vcpu); + if ((cr4 ^ old_cr4) & X86_CR4_SMAP) + update_permission_bitmask(vcpu, vcpu->arch.walk_mmu, false); + if ((cr4 ^ old_cr4) & X86_CR4_OSXSAVE) kvm_update_cpuid(vcpu); @@ -1117,7 +1123,6 @@ static inline u64 get_kernel_ns(void) { struct timespec ts; - WARN_ON(preemptible()); ktime_get_ts(&ts); monotonic_to_bootbased(&ts); return timespec_to_ns(&ts); @@ -4164,7 +4169,8 @@ static int vcpu_mmio_gva_to_gpa(struct kvm_vcpu *vcpu, unsigned long gva, | (write ? PFERR_WRITE_MASK : 0); if (vcpu_match_mmio_gva(vcpu, gva) - && !permission_fault(vcpu->arch.walk_mmu, vcpu->arch.access, access)) { + && !permission_fault(vcpu, vcpu->arch.walk_mmu, + vcpu->arch.access, access)) { *gpa = vcpu->arch.mmio_gfn << PAGE_SHIFT | (gva & (PAGE_SIZE - 1)); trace_vcpu_match_mmio(gva, *gpa, write, false); diff --git a/drivers/char/hw_random/bcm2835-rng.c b/drivers/char/hw_random/bcm2835-rng.c index 8c3b255..e900961 100644 --- a/drivers/char/hw_random/bcm2835-rng.c +++ b/drivers/char/hw_random/bcm2835-rng.c @@ -61,18 +61,18 @@ static int bcm2835_rng_probe(struct platform_device *pdev) } bcm2835_rng_ops.priv = (unsigned long)rng_base; + /* set warm-up count & enable */ + __raw_writel(RNG_WARMUP_COUNT, rng_base + RNG_STATUS); + __raw_writel(RNG_RBGEN, rng_base + RNG_CTRL); + /* register driver */ err = hwrng_register(&bcm2835_rng_ops); if (err) { dev_err(dev, "hwrng registration failed\n"); iounmap(rng_base); - } else { + } else dev_info(dev, "hwrng registered\n"); - /* set warm-up count & enable */ - __raw_writel(RNG_WARMUP_COUNT, rng_base + RNG_STATUS); - __raw_writel(RNG_RBGEN, rng_base + RNG_CTRL); - } return err; } diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c index 71e2d3f..04a209e 100644 --- a/drivers/gpu/drm/drm_mm.c +++ b/drivers/gpu/drm/drm_mm.c @@ -207,8 +207,6 @@ int drm_mm_reserve_node(struct drm_mm *mm, struct drm_mm_node *node) return 0; } - WARN(1, "no hole found for node 0x%lx + 0x%lx\n", - node->start, node->size); return -ENOSPC; } EXPORT_SYMBOL(drm_mm_reserve_node); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 0905cd9..ec82f6b 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1308,6 +1308,7 @@ struct intel_vbt_data { struct { u16 pwm_freq_hz; + bool present; bool active_low_pwm; } backlight; @@ -2431,20 +2432,18 @@ int i915_gem_context_open(struct drm_device *dev, struct drm_file *file); int i915_gem_context_enable(struct drm_i915_private *dev_priv); void i915_gem_context_close(struct drm_device *dev, struct drm_file *file); int i915_switch_context(struct intel_ring_buffer *ring, - struct drm_file *file, struct i915_hw_context *to); + struct i915_hw_context *to); struct i915_hw_context * i915_gem_context_get(struct drm_i915_file_private *file_priv, u32 id); void i915_gem_context_free(struct kref *ctx_ref); static inline void i915_gem_context_reference(struct i915_hw_context *ctx) { - if (ctx->obj && HAS_HW_CONTEXTS(ctx->obj->base.dev)) - kref_get(&ctx->ref); + kref_get(&ctx->ref); } static inline void i915_gem_context_unreference(struct i915_hw_context *ctx) { - if (ctx->obj && HAS_HW_CONTEXTS(ctx->obj->base.dev)) - kref_put(&ctx->ref, i915_gem_context_free); + kref_put(&ctx->ref, i915_gem_context_free); } static inline bool i915_gem_context_is_default(const struct i915_hw_context *c) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 6370a76..2871ce7 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2790,7 +2790,7 @@ int i915_gpu_idle(struct drm_device *dev) /* Flush everything onto the inactive list. */ for_each_ring(ring, dev_priv, i) { - ret = i915_switch_context(ring, NULL, ring->default_context); + ret = i915_switch_context(ring, ring->default_context); if (ret) return ret; diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 6043062..d72db15 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -96,9 +96,6 @@ #define GEN6_CONTEXT_ALIGN (64<<10) #define GEN7_CONTEXT_ALIGN 4096 -static int do_switch(struct intel_ring_buffer *ring, - struct i915_hw_context *to); - static void do_ppgtt_cleanup(struct i915_hw_ppgtt *ppgtt) { struct drm_device *dev = ppgtt->base.dev; @@ -185,13 +182,15 @@ void i915_gem_context_free(struct kref *ctx_ref) typeof(*ctx), ref); struct i915_hw_ppgtt *ppgtt = NULL; - /* We refcount even the aliasing PPGTT to keep the code symmetric */ - if (USES_PPGTT(ctx->obj->base.dev)) - ppgtt = ctx_to_ppgtt(ctx); + if (ctx->obj) { + /* We refcount even the aliasing PPGTT to keep the code symmetric */ + if (USES_PPGTT(ctx->obj->base.dev)) + ppgtt = ctx_to_ppgtt(ctx); - /* XXX: Free up the object before tearing down the address space, in - * case we're bound in the PPGTT */ - drm_gem_object_unreference(&ctx->obj->base); + /* XXX: Free up the object before tearing down the address space, in + * case we're bound in the PPGTT */ + drm_gem_object_unreference(&ctx->obj->base); + } if (ppgtt) kref_put(&ppgtt->ref, ppgtt_release); @@ -232,32 +231,32 @@ __create_hw_context(struct drm_device *dev, return ERR_PTR(-ENOMEM); kref_init(&ctx->ref); - ctx->obj = i915_gem_alloc_object(dev, dev_priv->hw_context_size); - INIT_LIST_HEAD(&ctx->link); - if (ctx->obj == NULL) { - kfree(ctx); - DRM_DEBUG_DRIVER("Context object allocated failed\n"); - return ERR_PTR(-ENOMEM); - } + list_add_tail(&ctx->link, &dev_priv->context_list); - if (INTEL_INFO(dev)->gen >= 7) { - ret = i915_gem_object_set_cache_level(ctx->obj, - I915_CACHE_L3_LLC); - /* Failure shouldn't ever happen this early */ - if (WARN_ON(ret)) + if (dev_priv->hw_context_size) { + ctx->obj = i915_gem_alloc_object(dev, dev_priv->hw_context_size); + if (ctx->obj == NULL) { + ret = -ENOMEM; goto err_out; - } + } - list_add_tail(&ctx->link, &dev_priv->context_list); + if (INTEL_INFO(dev)->gen >= 7) { + ret = i915_gem_object_set_cache_level(ctx->obj, + I915_CACHE_L3_LLC); + /* Failure shouldn't ever happen this early */ + if (WARN_ON(ret)) + goto err_out; + } + } /* Default context will never have a file_priv */ - if (file_priv == NULL) - return ctx; - - ret = idr_alloc(&file_priv->context_idr, ctx, DEFAULT_CONTEXT_ID, 0, - GFP_KERNEL); - if (ret < 0) - goto err_out; + if (file_priv != NULL) { + ret = idr_alloc(&file_priv->context_idr, ctx, + DEFAULT_CONTEXT_ID, 0, GFP_KERNEL); + if (ret < 0) + goto err_out; + } else + ret = DEFAULT_CONTEXT_ID; ctx->file_priv = file_priv; ctx->id = ret; @@ -294,7 +293,7 @@ i915_gem_create_context(struct drm_device *dev, if (IS_ERR(ctx)) return ctx; - if (is_global_default_ctx) { + if (is_global_default_ctx && ctx->obj) { /* We may need to do things with the shrinker which * require us to immediately switch back to the default * context. This can cause a problem as pinning the @@ -342,7 +341,7 @@ i915_gem_create_context(struct drm_device *dev, return ctx; err_unpin: - if (is_global_default_ctx) + if (is_global_default_ctx && ctx->obj) i915_gem_object_ggtt_unpin(ctx->obj); err_destroy: i915_gem_context_unreference(ctx); @@ -352,32 +351,22 @@ err_destroy: void i915_gem_context_reset(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - struct intel_ring_buffer *ring; int i; - if (!HAS_HW_CONTEXTS(dev)) - return; - /* Prevent the hardware from restoring the last context (which hung) on * the next switch */ for (i = 0; i < I915_NUM_RINGS; i++) { - struct i915_hw_context *dctx; - if (!(INTEL_INFO(dev)->ring_mask & (1<<i))) - continue; + struct intel_ring_buffer *ring = &dev_priv->ring[i]; + struct i915_hw_context *dctx = ring->default_context; /* Do a fake switch to the default context */ - ring = &dev_priv->ring[i]; - dctx = ring->default_context; - if (WARN_ON(!dctx)) + if (ring->last_context == dctx) continue; if (!ring->last_context) continue; - if (ring->last_context == dctx) - continue; - - if (i == RCS) { + if (dctx->obj && i == RCS) { WARN_ON(i915_gem_obj_ggtt_pin(dctx->obj, get_context_alignment(dev), 0)); /* Fake a finish/inactive */ @@ -394,44 +383,35 @@ void i915_gem_context_reset(struct drm_device *dev) int i915_gem_context_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - struct intel_ring_buffer *ring; + struct i915_hw_context *ctx; int i; - if (!HAS_HW_CONTEXTS(dev)) - return 0; - /* Init should only be called once per module load. Eventually the * restriction on the context_disabled check can be loosened. */ if (WARN_ON(dev_priv->ring[RCS].default_context)) return 0; - dev_priv->hw_context_size = round_up(get_context_size(dev), 4096); - - if (dev_priv->hw_context_size > (1<<20)) { - DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size\n"); - return -E2BIG; + if (HAS_HW_CONTEXTS(dev)) { + dev_priv->hw_context_size = round_up(get_context_size(dev), 4096); + if (dev_priv->hw_context_size > (1<<20)) { + DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size %d\n", + dev_priv->hw_context_size); + dev_priv->hw_context_size = 0; + } } - dev_priv->ring[RCS].default_context = - i915_gem_create_context(dev, NULL, USES_PPGTT(dev)); - - if (IS_ERR_OR_NULL(dev_priv->ring[RCS].default_context)) { - DRM_DEBUG_DRIVER("Disabling HW Contexts; create failed %ld\n", - PTR_ERR(dev_priv->ring[RCS].default_context)); - return PTR_ERR(dev_priv->ring[RCS].default_context); + ctx = i915_gem_create_context(dev, NULL, USES_PPGTT(dev)); + if (IS_ERR(ctx)) { + DRM_ERROR("Failed to create default global context (error %ld)\n", + PTR_ERR(ctx)); + return PTR_ERR(ctx); } - for (i = RCS + 1; i < I915_NUM_RINGS; i++) { - if (!(INTEL_INFO(dev)->ring_mask & (1<<i))) - continue; - - ring = &dev_priv->ring[i]; + /* NB: RCS will hold a ref for all rings */ + for (i = 0; i < I915_NUM_RINGS; i++) + dev_priv->ring[i].default_context = ctx; - /* NB: RCS will hold a ref for all rings */ - ring->default_context = dev_priv->ring[RCS].default_context; - } - - DRM_DEBUG_DRIVER("HW context support initialized\n"); + DRM_DEBUG_DRIVER("%s context support initialized\n", dev_priv->hw_context_size ? "HW" : "fake"); return 0; } @@ -441,33 +421,30 @@ void i915_gem_context_fini(struct drm_device *dev) struct i915_hw_context *dctx = dev_priv->ring[RCS].default_context; int i; - if (!HAS_HW_CONTEXTS(dev)) - return; - - /* The only known way to stop the gpu from accessing the hw context is - * to reset it. Do this as the very last operation to avoid confusing - * other code, leading to spurious errors. */ - intel_gpu_reset(dev); - - /* When default context is created and switched to, base object refcount - * will be 2 (+1 from object creation and +1 from do_switch()). - * i915_gem_context_fini() will be called after gpu_idle() has switched - * to default context. So we need to unreference the base object once - * to offset the do_switch part, so that i915_gem_context_unreference() - * can then free the base object correctly. */ - WARN_ON(!dev_priv->ring[RCS].last_context); - if (dev_priv->ring[RCS].last_context == dctx) { - /* Fake switch to NULL context */ - WARN_ON(dctx->obj->active); - i915_gem_object_ggtt_unpin(dctx->obj); - i915_gem_context_unreference(dctx); - dev_priv->ring[RCS].last_context = NULL; + if (dctx->obj) { + /* The only known way to stop the gpu from accessing the hw context is + * to reset it. Do this as the very last operation to avoid confusing + * other code, leading to spurious errors. */ + intel_gpu_reset(dev); + + /* When default context is created and switched to, base object refcount + * will be 2 (+1 from object creation and +1 from do_switch()). + * i915_gem_context_fini() will be called after gpu_idle() has switched + * to default context. So we need to unreference the base object once + * to offset the do_switch part, so that i915_gem_context_unreference() + * can then free the base object correctly. */ + WARN_ON(!dev_priv->ring[RCS].last_context); + if (dev_priv->ring[RCS].last_context == dctx) { + /* Fake switch to NULL context */ + WARN_ON(dctx->obj->active); + i915_gem_object_ggtt_unpin(dctx->obj); + i915_gem_context_unreference(dctx); + dev_priv->ring[RCS].last_context = NULL; + } } for (i = 0; i < I915_NUM_RINGS; i++) { struct intel_ring_buffer *ring = &dev_priv->ring[i]; - if (!(INTEL_INFO(dev)->ring_mask & (1<<i))) - continue; if (ring->last_context) i915_gem_context_unreference(ring->last_context); @@ -478,7 +455,6 @@ void i915_gem_context_fini(struct drm_device *dev) i915_gem_object_ggtt_unpin(dctx->obj); i915_gem_context_unreference(dctx); - dev_priv->mm.aliasing_ppgtt = NULL; } int i915_gem_context_enable(struct drm_i915_private *dev_priv) @@ -486,9 +462,6 @@ int i915_gem_context_enable(struct drm_i915_private *dev_priv) struct intel_ring_buffer *ring; int ret, i; - if (!HAS_HW_CONTEXTS(dev_priv->dev)) - return 0; - /* This is the only place the aliasing PPGTT gets enabled, which means * it has to happen before we bail on reset */ if (dev_priv->mm.aliasing_ppgtt) { @@ -503,7 +476,7 @@ int i915_gem_context_enable(struct drm_i915_private *dev_priv) BUG_ON(!dev_priv->ring[RCS].default_context); for_each_ring(ring, dev_priv, i) { - ret = do_switch(ring, ring->default_context); + ret = i915_switch_context(ring, ring->default_context); if (ret) return ret; } @@ -526,19 +499,6 @@ static int context_idr_cleanup(int id, void *p, void *data) int i915_gem_context_open(struct drm_device *dev, struct drm_file *file) { struct drm_i915_file_private *file_priv = file->driver_priv; - struct drm_i915_private *dev_priv = dev->dev_private; - - if (!HAS_HW_CONTEXTS(dev)) { - /* Cheat for hang stats */ - file_priv->private_default_ctx = - kzalloc(sizeof(struct i915_hw_context), GFP_KERNEL); - - if (file_priv->private_default_ctx == NULL) - return -ENOMEM; - - file_priv->private_default_ctx->vm = &dev_priv->gtt.base; - return 0; - } idr_init(&file_priv->context_idr); @@ -559,14 +519,10 @@ void i915_gem_context_close(struct drm_device *dev, struct drm_file *file) { struct drm_i915_file_private *file_priv = file->driver_priv; - if (!HAS_HW_CONTEXTS(dev)) { - kfree(file_priv->private_default_ctx); - return; - } - idr_for_each(&file_priv->context_idr, context_idr_cleanup, NULL); - i915_gem_context_unreference(file_priv->private_default_ctx); idr_destroy(&file_priv->context_idr); + + i915_gem_context_unreference(file_priv->private_default_ctx); } struct i915_hw_context * @@ -574,9 +530,6 @@ i915_gem_context_get(struct drm_i915_file_private *file_priv, u32 id) { struct i915_hw_context *ctx; - if (!HAS_HW_CONTEXTS(file_priv->dev_priv->dev)) - return file_priv->private_default_ctx; - ctx = (struct i915_hw_context *)idr_find(&file_priv->context_idr, id); if (!ctx) return ERR_PTR(-ENOENT); @@ -758,7 +711,6 @@ unpin_out: /** * i915_switch_context() - perform a GPU context switch. * @ring: ring for which we'll execute the context switch - * @file_priv: file_priv associated with the context, may be NULL * @to: the context to switch to * * The context life cycle is simple. The context refcount is incremented and @@ -767,24 +719,30 @@ unpin_out: * object while letting the normal object tracking destroy the backing BO. */ int i915_switch_context(struct intel_ring_buffer *ring, - struct drm_file *file, struct i915_hw_context *to) { struct drm_i915_private *dev_priv = ring->dev->dev_private; WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex)); - BUG_ON(file && to == NULL); - - /* We have the fake context */ - if (!HAS_HW_CONTEXTS(ring->dev)) { - ring->last_context = to; + if (to->obj == NULL) { /* We have the fake context */ + if (to != ring->last_context) { + i915_gem_context_reference(to); + if (ring->last_context) + i915_gem_context_unreference(ring->last_context); + ring->last_context = to; + } return 0; } return do_switch(ring, to); } +static bool hw_context_enabled(struct drm_device *dev) +{ + return to_i915(dev)->hw_context_size; +} + int i915_gem_context_create_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { @@ -793,7 +751,7 @@ int i915_gem_context_create_ioctl(struct drm_device *dev, void *data, struct i915_hw_context *ctx; int ret; - if (!HAS_HW_CONTEXTS(dev)) + if (!hw_context_enabled(dev)) return -ENODEV; ret = i915_mutex_lock_interruptible(dev); diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 7447160..2c9d9cb 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -1221,7 +1221,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, if (ret) goto err; - ret = i915_switch_context(ring, file, ctx); + ret = i915_switch_context(ring, ctx); if (ret) goto err; diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index 4867f4c..fa486c5 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -287,6 +287,9 @@ parse_lfp_backlight(struct drm_i915_private *dev_priv, struct bdb_header *bdb) const struct bdb_lfp_backlight_data *backlight_data; const struct bdb_lfp_backlight_data_entry *entry; + /* Err to enabling backlight if no backlight block. */ + dev_priv->vbt.backlight.present = true; + backlight_data = find_section(bdb, BDB_LVDS_BACKLIGHT); if (!backlight_data) return; @@ -299,6 +302,13 @@ parse_lfp_backlight(struct drm_i915_private *dev_priv, struct bdb_header *bdb) entry = &backlight_data->data[panel_type]; + dev_priv->vbt.backlight.present = entry->type == BDB_BACKLIGHT_TYPE_PWM; + if (!dev_priv->vbt.backlight.present) { + DRM_DEBUG_KMS("PWM backlight not present in VBT (type %u)\n", + entry->type); + return; + } + dev_priv->vbt.backlight.pwm_freq_hz = entry->pwm_freq_hz; dev_priv->vbt.backlight.active_low_pwm = entry->active_low_pwm; DRM_DEBUG_KMS("VBT backlight PWM modulation frequency %u Hz, " diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h index 83b7629..f27f7b2 100644 --- a/drivers/gpu/drm/i915/intel_bios.h +++ b/drivers/gpu/drm/i915/intel_bios.h @@ -374,6 +374,9 @@ struct bdb_lvds_lfp_data { struct bdb_lvds_lfp_data_entry data[16]; } __packed; +#define BDB_BACKLIGHT_TYPE_NONE 0 +#define BDB_BACKLIGHT_TYPE_PWM 2 + struct bdb_lfp_backlight_data_entry { u8 type:2; u8 active_low_pwm:1; diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c index cb05840..0eead16 100644 --- a/drivers/gpu/drm/i915/intel_panel.c +++ b/drivers/gpu/drm/i915/intel_panel.c @@ -1065,6 +1065,11 @@ int intel_panel_setup_backlight(struct drm_connector *connector) unsigned long flags; int ret; + if (!dev_priv->vbt.backlight.present) { + DRM_DEBUG_KMS("native backlight control not available per VBT\n"); + return 0; + } + /* set level and max in panel struct */ spin_lock_irqsave(&dev_priv->backlight_lock, flags); ret = dev_priv->display.setup_backlight(intel_connector); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 5874716..19e94c3 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -1545,6 +1545,16 @@ static void i9xx_update_wm(struct drm_crtc *unused_crtc) DRM_DEBUG_KMS("FIFO watermarks - A: %d, B: %d\n", planea_wm, planeb_wm); + if (IS_I915GM(dev) && enabled) { + struct intel_framebuffer *fb; + + fb = to_intel_framebuffer(enabled->primary->fb); + + /* self-refresh seems busted with untiled */ + if (fb->obj->tiling_mode == I915_TILING_NONE) + enabled = NULL; + } + /* * Overlay gets an aggressive default since video jitter is bad. */ diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index cad89a9..10dae41 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -21,8 +21,10 @@ * */ +#include <linux/firmware.h> #include "drmP.h" #include "radeon.h" +#include "radeon_ucode.h" #include "cikd.h" #include "r600_dpm.h" #include "ci_dpm.h" @@ -202,24 +204,29 @@ static void ci_initialize_powertune_defaults(struct radeon_device *rdev) struct ci_power_info *pi = ci_get_pi(rdev); switch (rdev->pdev->device) { + case 0x6649: case 0x6650: + case 0x6651: case 0x6658: case 0x665C: + case 0x665D: default: pi->powertune_defaults = &defaults_bonaire_xt; break; - case 0x6651: - case 0x665D: - pi->powertune_defaults = &defaults_bonaire_pro; - break; case 0x6640: - pi->powertune_defaults = &defaults_saturn_xt; - break; case 0x6641: - pi->powertune_defaults = &defaults_saturn_pro; + case 0x6646: + case 0x6647: + pi->powertune_defaults = &defaults_saturn_xt; break; case 0x67B8: case 0x67B0: + pi->powertune_defaults = &defaults_hawaii_xt; + break; + case 0x67BA: + case 0x67B1: + pi->powertune_defaults = &defaults_hawaii_pro; + break; case 0x67A0: case 0x67A1: case 0x67A2: @@ -228,11 +235,7 @@ static void ci_initialize_powertune_defaults(struct radeon_device *rdev) case 0x67AA: case 0x67B9: case 0x67BE: - pi->powertune_defaults = &defaults_hawaii_xt; - break; - case 0x67BA: - case 0x67B1: - pi->powertune_defaults = &defaults_hawaii_pro; + pi->powertune_defaults = &defaults_bonaire_xt; break; } @@ -5146,6 +5149,12 @@ int ci_dpm_init(struct radeon_device *rdev) pi->mclk_dpm_key_disabled = 0; pi->pcie_dpm_key_disabled = 0; + /* mclk dpm is unstable on some R7 260X cards with the old mc ucode */ + if ((rdev->pdev->device == 0x6658) && + (rdev->mc_fw->size == (BONAIRE_MC_UCODE_SIZE * 4))) { + pi->mclk_dpm_key_disabled = 1; + } + pi->caps_sclk_ds = true; pi->mclk_strobe_mode_threshold = 40000; diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index 745143c..199eb19 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -38,6 +38,7 @@ MODULE_FIRMWARE("radeon/BONAIRE_me.bin"); MODULE_FIRMWARE("radeon/BONAIRE_ce.bin"); MODULE_FIRMWARE("radeon/BONAIRE_mec.bin"); MODULE_FIRMWARE("radeon/BONAIRE_mc.bin"); +MODULE_FIRMWARE("radeon/BONAIRE_mc2.bin"); MODULE_FIRMWARE("radeon/BONAIRE_rlc.bin"); MODULE_FIRMWARE("radeon/BONAIRE_sdma.bin"); MODULE_FIRMWARE("radeon/BONAIRE_smc.bin"); @@ -46,6 +47,7 @@ MODULE_FIRMWARE("radeon/HAWAII_me.bin"); MODULE_FIRMWARE("radeon/HAWAII_ce.bin"); MODULE_FIRMWARE("radeon/HAWAII_mec.bin"); MODULE_FIRMWARE("radeon/HAWAII_mc.bin"); +MODULE_FIRMWARE("radeon/HAWAII_mc2.bin"); MODULE_FIRMWARE("radeon/HAWAII_rlc.bin"); MODULE_FIRMWARE("radeon/HAWAII_sdma.bin"); MODULE_FIRMWARE("radeon/HAWAII_smc.bin"); @@ -1703,20 +1705,20 @@ int ci_mc_load_microcode(struct radeon_device *rdev) const __be32 *fw_data; u32 running, blackout = 0; u32 *io_mc_regs; - int i, ucode_size, regs_size; + int i, regs_size, ucode_size; if (!rdev->mc_fw) return -EINVAL; + ucode_size = rdev->mc_fw->size / 4; + switch (rdev->family) { case CHIP_BONAIRE: io_mc_regs = (u32 *)&bonaire_io_mc_regs; - ucode_size = CIK_MC_UCODE_SIZE; regs_size = BONAIRE_IO_MC_REGS_SIZE; break; case CHIP_HAWAII: io_mc_regs = (u32 *)&hawaii_io_mc_regs; - ucode_size = HAWAII_MC_UCODE_SIZE; regs_size = HAWAII_IO_MC_REGS_SIZE; break; default: @@ -1783,7 +1785,7 @@ static int cik_init_microcode(struct radeon_device *rdev) const char *chip_name; size_t pfp_req_size, me_req_size, ce_req_size, mec_req_size, rlc_req_size, mc_req_size = 0, - sdma_req_size, smc_req_size = 0; + sdma_req_size, smc_req_size = 0, mc2_req_size = 0; char fw_name[30]; int err; @@ -1797,7 +1799,8 @@ static int cik_init_microcode(struct radeon_device *rdev) ce_req_size = CIK_CE_UCODE_SIZE * 4; mec_req_size = CIK_MEC_UCODE_SIZE * 4; rlc_req_size = BONAIRE_RLC_UCODE_SIZE * 4; - mc_req_size = CIK_MC_UCODE_SIZE * 4; + mc_req_size = BONAIRE_MC_UCODE_SIZE * 4; + mc2_req_size = BONAIRE_MC2_UCODE_SIZE * 4; sdma_req_size = CIK_SDMA_UCODE_SIZE * 4; smc_req_size = ALIGN(BONAIRE_SMC_UCODE_SIZE, 4); break; @@ -1809,6 +1812,7 @@ static int cik_init_microcode(struct radeon_device *rdev) mec_req_size = CIK_MEC_UCODE_SIZE * 4; rlc_req_size = BONAIRE_RLC_UCODE_SIZE * 4; mc_req_size = HAWAII_MC_UCODE_SIZE * 4; + mc2_req_size = HAWAII_MC2_UCODE_SIZE * 4; sdma_req_size = CIK_SDMA_UCODE_SIZE * 4; smc_req_size = ALIGN(HAWAII_SMC_UCODE_SIZE, 4); break; @@ -1904,16 +1908,22 @@ static int cik_init_microcode(struct radeon_device *rdev) /* No SMC, MC ucode on APUs */ if (!(rdev->flags & RADEON_IS_IGP)) { - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name); err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); - if (err) - goto out; - if (rdev->mc_fw->size != mc_req_size) { + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) + goto out; + } + if ((rdev->mc_fw->size != mc_req_size) && + (rdev->mc_fw->size != mc2_req_size)){ printk(KERN_ERR "cik_mc: Bogus length %zu in firmware \"%s\"\n", rdev->mc_fw->size, fw_name); err = -EINVAL; } + DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size); snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index f21db7a..7014bdd 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -2321,6 +2321,7 @@ struct radeon_device { bool have_disp_power_ref; }; +bool radeon_is_px(struct drm_device *dev); int radeon_device_init(struct radeon_device *rdev, struct drm_device *ddev, struct pci_dev *pdev, diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c index fa9a9c0..dedea72 100644 --- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c +++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c @@ -59,7 +59,7 @@ struct atpx_mux { u16 mux; } __packed; -bool radeon_is_px(void) { +bool radeon_has_atpx(void) { return radeon_atpx_priv.atpx_detected; } diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index 835516d..511fe26 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -102,11 +102,14 @@ static const char radeon_family_name[][16] = { "LAST", }; -#if defined(CONFIG_VGA_SWITCHEROO) -bool radeon_is_px(void); -#else -static inline bool radeon_is_px(void) { return false; } -#endif +bool radeon_is_px(struct drm_device *dev) +{ + struct radeon_device *rdev = dev->dev_private; + + if (rdev->flags & RADEON_IS_PX) + return true; + return false; +} /** * radeon_program_register_sequence - program an array of registers. @@ -1082,7 +1085,7 @@ static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero { struct drm_device *dev = pci_get_drvdata(pdev); - if (radeon_is_px() && state == VGA_SWITCHEROO_OFF) + if (radeon_is_px(dev) && state == VGA_SWITCHEROO_OFF) return; if (state == VGA_SWITCHEROO_ON) { @@ -1301,9 +1304,7 @@ int radeon_device_init(struct radeon_device *rdev, * ignore it */ vga_client_register(rdev->pdev, rdev, NULL, radeon_vga_set_decode); - if (radeon_runtime_pm == 1) - runtime = true; - if ((radeon_runtime_pm == -1) && radeon_is_px()) + if (rdev->flags & RADEON_IS_PX) runtime = true; vga_switcheroo_register_client(rdev->pdev, &radeon_switcheroo_ops, runtime); if (runtime) diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 386cfa4..063d425 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -865,7 +865,7 @@ void radeon_compute_pll_avivo(struct radeon_pll *pll, unsigned post_div_min, post_div_max, post_div; unsigned ref_div_min, ref_div_max, ref_div; unsigned post_div_best, diff_best; - unsigned nom, den, tmp; + unsigned nom, den; /* determine allowed feedback divider range */ fb_div_min = pll->min_feedback_div; @@ -937,23 +937,27 @@ void radeon_compute_pll_avivo(struct radeon_pll *pll, } post_div = post_div_best; + /* limit reference * post divider to a maximum */ + ref_div_max = min(210 / post_div, ref_div_max); + /* get matching reference and feedback divider */ - ref_div = max(den / post_div, 1u); - fb_div = nom; + ref_div = max(DIV_ROUND_CLOSEST(den, post_div), 1u); + fb_div = DIV_ROUND_CLOSEST(nom * ref_div * post_div, den); /* we're almost done, but reference and feedback divider might be to large now */ - tmp = ref_div; + nom = fb_div; + den = ref_div; if (fb_div > fb_div_max) { - ref_div = ref_div * fb_div_max / fb_div; + ref_div = DIV_ROUND_CLOSEST(den * fb_div_max, nom); fb_div = fb_div_max; } if (ref_div > ref_div_max) { ref_div = ref_div_max; - fb_div = nom * ref_div_max / tmp; + fb_div = DIV_ROUND_CLOSEST(nom * ref_div_max, den); } /* reduce the numbers to a simpler ratio once more */ diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index d0eba48..25127ba 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -115,6 +115,7 @@ extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, unsigned int flags, int *vpos, int *hpos, ktime_t *stime, ktime_t *etime); +extern bool radeon_is_px(struct drm_device *dev); extern const struct drm_ioctl_desc radeon_ioctls_kms[]; extern int radeon_max_kms_ioctl; int radeon_mmap(struct file *filp, struct vm_area_struct *vma); @@ -144,11 +145,9 @@ void radeon_debugfs_cleanup(struct drm_minor *minor); #if defined(CONFIG_VGA_SWITCHEROO) void radeon_register_atpx_handler(void); void radeon_unregister_atpx_handler(void); -bool radeon_is_px(void); #else static inline void radeon_register_atpx_handler(void) {} static inline void radeon_unregister_atpx_handler(void) {} -static inline bool radeon_is_px(void) { return false; } #endif int radeon_no_wb; @@ -405,12 +404,7 @@ static int radeon_pmops_runtime_suspend(struct device *dev) struct drm_device *drm_dev = pci_get_drvdata(pdev); int ret; - if (radeon_runtime_pm == 0) { - pm_runtime_forbid(dev); - return -EBUSY; - } - - if (radeon_runtime_pm == -1 && !radeon_is_px()) { + if (!radeon_is_px(drm_dev)) { pm_runtime_forbid(dev); return -EBUSY; } @@ -434,10 +428,7 @@ static int radeon_pmops_runtime_resume(struct device *dev) struct drm_device *drm_dev = pci_get_drvdata(pdev); int ret; - if (radeon_runtime_pm == 0) - return -EINVAL; - - if (radeon_runtime_pm == -1 && !radeon_is_px()) + if (!radeon_is_px(drm_dev)) return -EINVAL; drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; @@ -462,14 +453,7 @@ static int radeon_pmops_runtime_idle(struct device *dev) struct drm_device *drm_dev = pci_get_drvdata(pdev); struct drm_crtc *crtc; - if (radeon_runtime_pm == 0) { - pm_runtime_forbid(dev); - return -EBUSY; - } - - /* are we PX enabled? */ - if (radeon_runtime_pm == -1 && !radeon_is_px()) { - DRM_DEBUG_DRIVER("failing to power off - not px\n"); + if (!radeon_is_px(drm_dev)) { pm_runtime_forbid(dev); return -EBUSY; } diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h index 614ad54..9da5da4 100644 --- a/drivers/gpu/drm/radeon/radeon_family.h +++ b/drivers/gpu/drm/radeon/radeon_family.h @@ -115,6 +115,7 @@ enum radeon_chip_flags { RADEON_NEW_MEMMAP = 0x00400000UL, RADEON_IS_PCI = 0x00800000UL, RADEON_IS_IGPGART = 0x01000000UL, + RADEON_IS_PX = 0x02000000UL, }; #endif diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 3e49342..fb3d13f 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -35,9 +35,9 @@ #include <linux/pm_runtime.h> #if defined(CONFIG_VGA_SWITCHEROO) -bool radeon_is_px(void); +bool radeon_has_atpx(void); #else -static inline bool radeon_is_px(void) { return false; } +static inline bool radeon_has_atpx(void) { return false; } #endif /** @@ -107,6 +107,13 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags) flags |= RADEON_IS_PCI; } + if (radeon_runtime_pm == 1) + flags |= RADEON_IS_PX; + else if ((radeon_runtime_pm == -1) && + radeon_has_atpx() && + ((flags & RADEON_IS_IGP) == 0)) + flags |= RADEON_IS_PX; + /* radeon_device_init should report only fatal error * like memory allocation failure or iomapping failure, * or memory manager initialization failure, it must @@ -137,8 +144,7 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags) "Error during ACPI methods call\n"); } - if ((radeon_runtime_pm == 1) || - ((radeon_runtime_pm == -1) && radeon_is_px())) { + if (radeon_is_px(dev)) { pm_runtime_use_autosuspend(dev->dev); pm_runtime_set_autosuspend_delay(dev->dev, 5000); pm_runtime_set_active(dev->dev); @@ -568,12 +574,17 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) } r = radeon_vm_init(rdev, &fpriv->vm); - if (r) + if (r) { + kfree(fpriv); return r; + } r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false); - if (r) + if (r) { + radeon_vm_fini(rdev, &fpriv->vm); + kfree(fpriv); return r; + } /* map the ib pool buffer read only into * virtual address space */ diff --git a/drivers/gpu/drm/radeon/radeon_ucode.h b/drivers/gpu/drm/radeon/radeon_ucode.h index a77cd27..58d1293 100644 --- a/drivers/gpu/drm/radeon/radeon_ucode.h +++ b/drivers/gpu/drm/radeon/radeon_ucode.h @@ -57,9 +57,14 @@ #define BTC_MC_UCODE_SIZE 6024 #define CAYMAN_MC_UCODE_SIZE 6037 #define SI_MC_UCODE_SIZE 7769 +#define TAHITI_MC_UCODE_SIZE 7808 +#define PITCAIRN_MC_UCODE_SIZE 7775 +#define VERDE_MC_UCODE_SIZE 7875 #define OLAND_MC_UCODE_SIZE 7863 -#define CIK_MC_UCODE_SIZE 7866 +#define BONAIRE_MC_UCODE_SIZE 7866 +#define BONAIRE_MC2_UCODE_SIZE 7948 #define HAWAII_MC_UCODE_SIZE 7933 +#define HAWAII_MC2_UCODE_SIZE 8091 /* SDMA */ #define CIK_SDMA_UCODE_SIZE 1050 diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c index 76e9904..ced53dd 100644 --- a/drivers/gpu/drm/radeon/radeon_vce.c +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -613,7 +613,7 @@ void radeon_vce_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence) { struct radeon_ring *ring = &rdev->ring[fence->ring]; - uint32_t addr = rdev->fence_drv[fence->ring].gpu_addr; + uint64_t addr = rdev->fence_drv[fence->ring].gpu_addr; radeon_ring_write(ring, VCE_CMD_FENCE); radeon_ring_write(ring, addr); diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index d589475..ac708e0 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -39,30 +39,35 @@ MODULE_FIRMWARE("radeon/TAHITI_pfp.bin"); MODULE_FIRMWARE("radeon/TAHITI_me.bin"); MODULE_FIRMWARE("radeon/TAHITI_ce.bin"); MODULE_FIRMWARE("radeon/TAHITI_mc.bin"); +MODULE_FIRMWARE("radeon/TAHITI_mc2.bin"); MODULE_FIRMWARE("radeon/TAHITI_rlc.bin"); MODULE_FIRMWARE("radeon/TAHITI_smc.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_me.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_mc2.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin"); MODULE_FIRMWARE("radeon/VERDE_pfp.bin"); MODULE_FIRMWARE("radeon/VERDE_me.bin"); MODULE_FIRMWARE("radeon/VERDE_ce.bin"); MODULE_FIRMWARE("radeon/VERDE_mc.bin"); +MODULE_FIRMWARE("radeon/VERDE_mc2.bin"); MODULE_FIRMWARE("radeon/VERDE_rlc.bin"); MODULE_FIRMWARE("radeon/VERDE_smc.bin"); MODULE_FIRMWARE("radeon/OLAND_pfp.bin"); MODULE_FIRMWARE("radeon/OLAND_me.bin"); MODULE_FIRMWARE("radeon/OLAND_ce.bin"); MODULE_FIRMWARE("radeon/OLAND_mc.bin"); +MODULE_FIRMWARE("radeon/OLAND_mc2.bin"); MODULE_FIRMWARE("radeon/OLAND_rlc.bin"); MODULE_FIRMWARE("radeon/OLAND_smc.bin"); MODULE_FIRMWARE("radeon/HAINAN_pfp.bin"); MODULE_FIRMWARE("radeon/HAINAN_me.bin"); MODULE_FIRMWARE("radeon/HAINAN_ce.bin"); MODULE_FIRMWARE("radeon/HAINAN_mc.bin"); +MODULE_FIRMWARE("radeon/HAINAN_mc2.bin"); MODULE_FIRMWARE("radeon/HAINAN_rlc.bin"); MODULE_FIRMWARE("radeon/HAINAN_smc.bin"); @@ -1467,36 +1472,33 @@ int si_mc_load_microcode(struct radeon_device *rdev) const __be32 *fw_data; u32 running, blackout = 0; u32 *io_mc_regs; - int i, ucode_size, regs_size; + int i, regs_size, ucode_size; if (!rdev->mc_fw) return -EINVAL; + ucode_size = rdev->mc_fw->size / 4; + switch (rdev->family) { case CHIP_TAHITI: io_mc_regs = (u32 *)&tahiti_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_PITCAIRN: io_mc_regs = (u32 *)&pitcairn_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_VERDE: default: io_mc_regs = (u32 *)&verde_io_mc_regs; - ucode_size = SI_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_OLAND: io_mc_regs = (u32 *)&oland_io_mc_regs; - ucode_size = OLAND_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; case CHIP_HAINAN: io_mc_regs = (u32 *)&hainan_io_mc_regs; - ucode_size = OLAND_MC_UCODE_SIZE; regs_size = TAHITI_IO_MC_REGS_SIZE; break; } @@ -1552,7 +1554,7 @@ static int si_init_microcode(struct radeon_device *rdev) const char *chip_name; const char *rlc_chip_name; size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size; - size_t smc_req_size; + size_t smc_req_size, mc2_req_size; char fw_name[30]; int err; @@ -1567,6 +1569,7 @@ static int si_init_microcode(struct radeon_device *rdev) ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = TAHITI_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(TAHITI_SMC_UCODE_SIZE, 4); break; case CHIP_PITCAIRN: @@ -1577,6 +1580,7 @@ static int si_init_microcode(struct radeon_device *rdev) ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = PITCAIRN_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(PITCAIRN_SMC_UCODE_SIZE, 4); break; case CHIP_VERDE: @@ -1587,6 +1591,7 @@ static int si_init_microcode(struct radeon_device *rdev) ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; mc_req_size = SI_MC_UCODE_SIZE * 4; + mc2_req_size = VERDE_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(VERDE_SMC_UCODE_SIZE, 4); break; case CHIP_OLAND: @@ -1596,7 +1601,7 @@ static int si_init_microcode(struct radeon_device *rdev) me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; - mc_req_size = OLAND_MC_UCODE_SIZE * 4; + mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(OLAND_SMC_UCODE_SIZE, 4); break; case CHIP_HAINAN: @@ -1606,7 +1611,7 @@ static int si_init_microcode(struct radeon_device *rdev) me_req_size = SI_PM4_UCODE_SIZE * 4; ce_req_size = SI_CE_UCODE_SIZE * 4; rlc_req_size = SI_RLC_UCODE_SIZE * 4; - mc_req_size = OLAND_MC_UCODE_SIZE * 4; + mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4; smc_req_size = ALIGN(HAINAN_SMC_UCODE_SIZE, 4); break; default: BUG(); @@ -1659,16 +1664,22 @@ static int si_init_microcode(struct radeon_device *rdev) err = -EINVAL; } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name); err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); - if (err) - goto out; - if (rdev->mc_fw->size != mc_req_size) { + if (err) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev); + if (err) + goto out; + } + if ((rdev->mc_fw->size != mc_req_size) && + (rdev->mc_fw->size != mc2_req_size)) { printk(KERN_ERR "si_mc: Bogus length %zu in firmware \"%s\"\n", rdev->mc_fw->size, fw_name); err = -EINVAL; } + DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size); snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); diff --git a/drivers/gpu/host1x/hw/intr_hw.c b/drivers/gpu/host1x/hw/intr_hw.c index db9017a..498b37e 100644 --- a/drivers/gpu/host1x/hw/intr_hw.c +++ b/drivers/gpu/host1x/hw/intr_hw.c @@ -47,7 +47,7 @@ static irqreturn_t syncpt_thresh_isr(int irq, void *dev_id) unsigned long reg; int i, id; - for (i = 0; i <= BIT_WORD(host->info->nb_pts); i++) { + for (i = 0; i < DIV_ROUND_UP(host->info->nb_pts, 32); i++) { reg = host1x_sync_readl(host, HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(i)); for_each_set_bit(id, ®, BITS_PER_LONG) { @@ -64,7 +64,7 @@ static void _host1x_intr_disable_all_syncpt_intrs(struct host1x *host) { u32 i; - for (i = 0; i <= BIT_WORD(host->info->nb_pts); ++i) { + for (i = 0; i < DIV_ROUND_UP(host->info->nb_pts, 32); ++i) { host1x_sync_writel(host, 0xffffffffu, HOST1X_SYNC_SYNCPT_THRESH_INT_DISABLE(i)); host1x_sync_writel(host, 0xffffffffu, diff --git a/include/uapi/drm/tegra_drm.h b/include/uapi/drm/tegra_drm.h index b042b48..b754821 100644 --- a/include/uapi/drm/tegra_drm.h +++ b/include/uapi/drm/tegra_drm.h @@ -120,7 +120,6 @@ struct drm_tegra_submit { __u32 num_waitchks; __u32 waitchk_mask; __u32 timeout; - __u32 pad; __u64 syncpts; __u64 cmdbufs; __u64 relocs; diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index 0d8f602..bf71b4b 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -152,7 +152,7 @@ static u32 map_id_range_down(struct uid_gid_map *map, u32 id, u32 count) /* Find the matching extent */ extents = map->nr_extents; - smp_read_barrier_depends(); + smp_rmb(); for (idx = 0; idx < extents; idx++) { first = map->extent[idx].first; last = first + map->extent[idx].count - 1; @@ -176,7 +176,7 @@ static u32 map_id_down(struct uid_gid_map *map, u32 id) /* Find the matching extent */ extents = map->nr_extents; - smp_read_barrier_depends(); + smp_rmb(); for (idx = 0; idx < extents; idx++) { first = map->extent[idx].first; last = first + map->extent[idx].count - 1; @@ -199,7 +199,7 @@ static u32 map_id_up(struct uid_gid_map *map, u32 id) /* Find the matching extent */ extents = map->nr_extents; - smp_read_barrier_depends(); + smp_rmb(); for (idx = 0; idx < extents; idx++) { first = map->extent[idx].lower_first; last = first + map->extent[idx].count - 1; @@ -615,9 +615,8 @@ static ssize_t map_write(struct file *file, const char __user *buf, * were written before the count of the extents. * * To achieve this smp_wmb() is used on guarantee the write - * order and smp_read_barrier_depends() is guaranteed that we - * don't have crazy architectures returning stale data. - * + * order and smp_rmb() is guaranteed that we don't have crazy + * architectures returning stale data. */ mutex_lock(&id_map_mutex); diff --git a/virt/kvm/ioapic.c b/virt/kvm/ioapic.c index d4b6015..2458a1d 100644 --- a/virt/kvm/ioapic.c +++ b/virt/kvm/ioapic.c @@ -97,6 +97,14 @@ static void rtc_irq_eoi_tracking_reset(struct kvm_ioapic *ioapic) bitmap_zero(ioapic->rtc_status.dest_map, KVM_MAX_VCPUS); } +static void kvm_rtc_eoi_tracking_restore_all(struct kvm_ioapic *ioapic); + +static void rtc_status_pending_eoi_check_valid(struct kvm_ioapic *ioapic) +{ + if (WARN_ON(ioapic->rtc_status.pending_eoi < 0)) + kvm_rtc_eoi_tracking_restore_all(ioapic); +} + static void __rtc_irq_eoi_tracking_restore_one(struct kvm_vcpu *vcpu) { bool new_val, old_val; @@ -120,9 +128,8 @@ static void __rtc_irq_eoi_tracking_restore_one(struct kvm_vcpu *vcpu) } else { __clear_bit(vcpu->vcpu_id, ioapic->rtc_status.dest_map); ioapic->rtc_status.pending_eoi--; + rtc_status_pending_eoi_check_valid(ioapic); } - - WARN_ON(ioapic->rtc_status.pending_eoi < 0); } void kvm_rtc_eoi_tracking_restore_one(struct kvm_vcpu *vcpu) @@ -149,10 +156,10 @@ static void kvm_rtc_eoi_tracking_restore_all(struct kvm_ioapic *ioapic) static void rtc_irq_eoi(struct kvm_ioapic *ioapic, struct kvm_vcpu *vcpu) { - if (test_and_clear_bit(vcpu->vcpu_id, ioapic->rtc_status.dest_map)) + if (test_and_clear_bit(vcpu->vcpu_id, ioapic->rtc_status.dest_map)) { --ioapic->rtc_status.pending_eoi; - - WARN_ON(ioapic->rtc_status.pending_eoi < 0); + rtc_status_pending_eoi_check_valid(ioapic); + } } static bool rtc_irq_check_coalesced(struct kvm_ioapic *ioapic) @@ -353,10 +360,16 @@ static int ioapic_service(struct kvm_ioapic *ioapic, int irq, bool line_status) ioapic->irr &= ~(1 << irq); if (irq == RTC_GSI && line_status) { + /* + * pending_eoi cannot ever become negative (see + * rtc_status_pending_eoi_check_valid) and the caller + * ensures that it is only called if it is >= zero, namely + * if rtc_irq_check_coalesced returns false). + */ BUG_ON(ioapic->rtc_status.pending_eoi != 0); ret = kvm_irq_delivery_to_apic(ioapic->kvm, NULL, &irqe, ioapic->rtc_status.dest_map); - ioapic->rtc_status.pending_eoi = ret; + ioapic->rtc_status.pending_eoi = (ret < 0 ? 0 : ret); } else ret = kvm_irq_delivery_to_apic(ioapic->kvm, NULL, &irqe, NULL); |