[PATCH openEuler-5.10 1/7] drm/i915: Flush TLBs before releasing backing store

From: Tvrtko Ursulin tvrtko.ursulin@intel.com

stable inclusion from stable-v5.10.95 commit 6a6acf927895c38bdd9f3cd76b8dbfc25ac03e88 bugzilla: https://gitee.com/openeuler/kernel/issues/I55EDV

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=...

--------------------------------

commit 7938d61591d33394a21bdd7797a245b65428f44c upstream.

We need to flush TLBs before releasing backing store otherwise userspace is able to encounter stale entries if a) it is not declaring access to certain buffers and b) it races with the backing store release from a such undeclared execution already executing on the GPU in parallel.

The approach taken is to mark any buffer objects which were ever bound to the GPU and to trigger a serialized TLB flush when their backing store is released.

Alternatively the flushing could be done on VMA unbind, at which point we would be able to ascertain whether there is potential a parallel GPU execution (which could race), but essentially it boils down to paying the cost of TLB flushes potentially needlessly at VMA unbind time (when the backing store is not known to be going away so not needed for safety), versus potentially needlessly at backing store relase time (since we at that point cannot tell whether there is anything executing on the GPU which uses that object).

Thereforce simplicity of implementation has been chosen for now with scope to benchmark and refine later as required.

Signed-off-by: Tvrtko Ursulin tvrtko.ursulin@intel.com Reported-by: Sushma Venkatesh Reddy sushma.venkatesh.reddy@intel.com Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch Acked-by: Dave Airlie airlied@redhat.com Cc: Daniel Vetter daniel.vetter@ffwll.ch Cc: Jon Bloomfield jon.bloomfield@intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Cc: Jani Nikula jani.nikula@intel.com Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds torvalds@linux-foundation.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org Signed-off-by: Yu Liao liaoyu15@huawei.com Reviewed-by: Wei Li liwei391@huawei.com Signed-off-by: Zheng Zengkai zhengzengkai@huawei.com --- .../gpu/drm/i915/gem/i915_gem_object_types.h | 1 + drivers/gpu/drm/i915/gem/i915_gem_pages.c | 10 ++ drivers/gpu/drm/i915/gt/intel_gt.c | 102 ++++++++++++++++++ drivers/gpu/drm/i915/gt/intel_gt.h | 2 + drivers/gpu/drm/i915/gt/intel_gt_types.h | 2 + drivers/gpu/drm/i915/i915_reg.h | 11 ++ drivers/gpu/drm/i915/i915_vma.c | 3 + drivers/gpu/drm/i915/intel_uncore.c | 26 ++++- drivers/gpu/drm/i915/intel_uncore.h | 2 + 9 files changed, 155 insertions(+), 4 deletions(-)

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 d6711caa7f39..dbc88fc7136b 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h @@ -159,6 +159,7 @@ struct drm_i915_gem_object { #define I915_BO_ALLOC_VOLATILE BIT(1) #define I915_BO_ALLOC_FLAGS (I915_BO_ALLOC_CONTIGUOUS | I915_BO_ALLOC_VOLATILE) #define I915_BO_READONLY BIT(2) +#define I915_BO_WAS_BOUND_BIT 3

/* * Is the object to be mapped as read-only to the GPU diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pages.c b/drivers/gpu/drm/i915/gem/i915_gem_pages.c index f60ca6dc911f..27d24cb38c0d 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_pages.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_pages.c @@ -10,6 +10,8 @@ #include "i915_gem_lmem.h" #include "i915_gem_mman.h"

+#include "gt/intel_gt.h" + void __i915_gem_object_set_pages(struct drm_i915_gem_object *obj, struct sg_table *pages, unsigned int sg_page_sizes) @@ -186,6 +188,14 @@ __i915_gem_object_unset_pages(struct drm_i915_gem_object *obj) __i915_gem_object_reset_page_iter(obj); obj->mm.page_sizes.phys = obj->mm.page_sizes.sg = 0;

+ if (test_and_clear_bit(I915_BO_WAS_BOUND_BIT, &obj->flags)) { + struct drm_i915_private *i915 = to_i915(obj->base.dev); + intel_wakeref_t wakeref; + + with_intel_runtime_pm_if_active(&i915->runtime_pm, wakeref) + intel_gt_invalidate_tlbs(&i915->gt); + } + return pages; }

diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c index 39b428c5049c..6615eb5147e2 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt.c +++ b/drivers/gpu/drm/i915/gt/intel_gt.c @@ -26,6 +26,8 @@ void intel_gt_init_early(struct intel_gt *gt, struct drm_i915_private *i915)

spin_lock_init(&gt->irq_lock);

+ mutex_init(&gt->tlb_invalidate_lock); + INIT_LIST_HEAD(&gt->closed_vma); spin_lock_init(&gt->closed_lock);

@@ -661,3 +663,103 @@ void intel_gt_info_print(const struct intel_gt_info *info,

intel_sseu_dump(&info->sseu, p); } + +struct reg_and_bit { + i915_reg_t reg; + u32 bit; +}; + +static struct reg_and_bit +get_reg_and_bit(const struct intel_engine_cs *engine, const bool gen8, + const i915_reg_t *regs, const unsigned int num) +{ + const unsigned int class = engine->class; + struct reg_and_bit rb = { }; + + if (drm_WARN_ON_ONCE(&engine->i915->drm, + class >= num || !regs[class].reg)) + return rb; + + rb.reg = regs[class]; + if (gen8 && class == VIDEO_DECODE_CLASS) + rb.reg.reg += 4 * engine->instance; /* GEN8_M2TCR */ + else + rb.bit = engine->instance; + + rb.bit = BIT(rb.bit); + + return rb; +} + +void intel_gt_invalidate_tlbs(struct intel_gt *gt) +{ + static const i915_reg_t gen8_regs[] = { + [RENDER_CLASS] = GEN8_RTCR, + [VIDEO_DECODE_CLASS] = GEN8_M1TCR, /* , GEN8_M2TCR */ + [VIDEO_ENHANCEMENT_CLASS] = GEN8_VTCR, + [COPY_ENGINE_CLASS] = GEN8_BTCR, + }; + static const i915_reg_t gen12_regs[] = { + [RENDER_CLASS] = GEN12_GFX_TLB_INV_CR, + [VIDEO_DECODE_CLASS] = GEN12_VD_TLB_INV_CR, + [VIDEO_ENHANCEMENT_CLASS] = GEN12_VE_TLB_INV_CR, + [COPY_ENGINE_CLASS] = GEN12_BLT_TLB_INV_CR, + }; + struct drm_i915_private *i915 = gt->i915; + struct intel_uncore *uncore = gt->uncore; + struct intel_engine_cs *engine; + enum intel_engine_id id; + const i915_reg_t *regs; + unsigned int num = 0; + + if (I915_SELFTEST_ONLY(gt->awake == -ENODEV)) + return; + + if (INTEL_GEN(i915) == 12) { + regs = gen12_regs; + num = ARRAY_SIZE(gen12_regs); + } else if (INTEL_GEN(i915) >= 8 && INTEL_GEN(i915) <= 11) { + regs = gen8_regs; + num = ARRAY_SIZE(gen8_regs); + } else if (INTEL_GEN(i915) < 8) { + return; + } + + if (drm_WARN_ONCE(&i915->drm, !num, + "Platform does not implement TLB invalidation!")) + return; + + GEM_TRACE("\n"); + + assert_rpm_wakelock_held(&i915->runtime_pm); + + mutex_lock(&gt->tlb_invalidate_lock); + intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL); + + for_each_engine(engine, gt, id) { + /* + * HW architecture suggest typical invalidation time at 40us, + * with pessimistic cases up to 100us and a recommendation to + * cap at 1ms. We go a bit higher just in case. + */ + const unsigned int timeout_us = 100; + const unsigned int timeout_ms = 4; + struct reg_and_bit rb; + + rb = get_reg_and_bit(engine, regs == gen8_regs, regs, num); + if (!i915_mmio_reg_offset(rb.reg)) + continue; + + intel_uncore_write_fw(uncore, rb.reg, rb.bit); + if (__intel_wait_for_register_fw(uncore, + rb.reg, rb.bit, 0, + timeout_us, timeout_ms, + NULL)) + drm_err_ratelimited(&gt->i915->drm, + "%s TLB invalidation did not complete in %ums!\n", + engine->name, timeout_ms); + } + + intel_uncore_forcewake_put_delayed(uncore, FORCEWAKE_ALL); + mutex_unlock(&gt->tlb_invalidate_lock); +} diff --git a/drivers/gpu/drm/i915/gt/intel_gt.h b/drivers/gpu/drm/i915/gt/intel_gt.h index 9157c7411f60..d9a1168172ae 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt.h +++ b/drivers/gpu/drm/i915/gt/intel_gt.h @@ -77,4 +77,6 @@ static inline bool intel_gt_is_wedged(const struct intel_gt *gt) void intel_gt_info_print(const struct intel_gt_info *info, struct drm_printer *p);

+void intel_gt_invalidate_tlbs(struct intel_gt *gt); + #endif /* __INTEL_GT_H__ */ diff --git a/drivers/gpu/drm/i915/gt/intel_gt_types.h b/drivers/gpu/drm/i915/gt/intel_gt_types.h index 6d39a4a11bf3..78c061614d8b 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_types.h +++ b/drivers/gpu/drm/i915/gt/intel_gt_types.h @@ -36,6 +36,8 @@ struct intel_gt {

struct intel_uc uc;

+ struct mutex tlb_invalidate_lock; + struct intel_gt_timelines { spinlock_t lock; /* protects active_list */ struct list_head active_list; diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index ce8c91c5fdd3..12488996a7f4 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -2639,6 +2639,12 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING (1 << 28) #define GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT (1 << 24)

+#define GEN8_RTCR _MMIO(0x4260) +#define GEN8_M1TCR _MMIO(0x4264) +#define GEN8_M2TCR _MMIO(0x4268) +#define GEN8_BTCR _MMIO(0x426c) +#define GEN8_VTCR _MMIO(0x4270) + #if 0 #define PRB0_TAIL _MMIO(0x2030) #define PRB0_HEAD _MMIO(0x2034) @@ -2728,6 +2734,11 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define FAULT_VA_HIGH_BITS (0xf << 0) #define FAULT_GTT_SEL (1 << 4)

+#define GEN12_GFX_TLB_INV_CR _MMIO(0xced8) +#define GEN12_VD_TLB_INV_CR _MMIO(0xcedc) +#define GEN12_VE_TLB_INV_CR _MMIO(0xcee0) +#define GEN12_BLT_TLB_INV_CR _MMIO(0xcee4) + #define GEN12_AUX_ERR_DBG _MMIO(0x43f4)

#define FPGA_DBG _MMIO(0x42300) diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index caa9b041616b..50a86fd89d00 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -439,6 +439,9 @@ int i915_vma_bind(struct i915_vma *vma, vma->ops->bind_vma(vma->vm, NULL, vma, cache_level, bind_flags); }

+ if (vma->obj) + set_bit(I915_BO_WAS_BOUND_BIT, &vma->obj->flags); + atomic_or(bind_flags, &vma->flags); return 0; } diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index 97ded2a59cf4..01849840ac56 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -694,7 +694,8 @@ void intel_uncore_forcewake_get__locked(struct intel_uncore *uncore, }

static void __intel_uncore_forcewake_put(struct intel_uncore *uncore, - enum forcewake_domains fw_domains) + enum forcewake_domains fw_domains, + bool delayed) { struct intel_uncore_forcewake_domain *domain; unsigned int tmp; @@ -709,7 +710,11 @@ static void __intel_uncore_forcewake_put(struct intel_uncore *uncore, continue; }

- uncore->funcs.force_wake_put(uncore, domain->mask); + if (delayed && + !(domain->uncore->fw_domains_timer & domain->mask)) + fw_domain_arm_timer(domain); + else + uncore->funcs.force_wake_put(uncore, domain->mask); } }

@@ -730,7 +735,20 @@ void intel_uncore_forcewake_put(struct intel_uncore *uncore, return;

spin_lock_irqsave(&uncore->lock, irqflags); - __intel_uncore_forcewake_put(uncore, fw_domains); + __intel_uncore_forcewake_put(uncore, fw_domains, false); + spin_unlock_irqrestore(&uncore->lock, irqflags); +} + +void intel_uncore_forcewake_put_delayed(struct intel_uncore *uncore, + enum forcewake_domains fw_domains) +{ + unsigned long irqflags; + + if (!uncore->funcs.force_wake_put) + return; + + spin_lock_irqsave(&uncore->lock, irqflags); + __intel_uncore_forcewake_put(uncore, fw_domains, true); spin_unlock_irqrestore(&uncore->lock, irqflags); }

@@ -772,7 +790,7 @@ void intel_uncore_forcewake_put__locked(struct intel_uncore *uncore, if (!uncore->funcs.force_wake_put) return;

- __intel_uncore_forcewake_put(uncore, fw_domains); + __intel_uncore_forcewake_put(uncore, fw_domains, false); }

void assert_forcewakes_inactive(struct intel_uncore *uncore) diff --git a/drivers/gpu/drm/i915/intel_uncore.h b/drivers/gpu/drm/i915/intel_uncore.h index c4b22d9d0b45..034f04e0de8b 100644 --- a/drivers/gpu/drm/i915/intel_uncore.h +++ b/drivers/gpu/drm/i915/intel_uncore.h @@ -211,6 +211,8 @@ void intel_uncore_forcewake_get(struct intel_uncore *uncore, enum forcewake_domains domains); void intel_uncore_forcewake_put(struct intel_uncore *uncore, enum forcewake_domains domains); +void intel_uncore_forcewake_put_delayed(struct intel_uncore *uncore, + enum forcewake_domains domains); void intel_uncore_forcewake_flush(struct intel_uncore *uncore, enum forcewake_domains fw_domains);