[Intel-gfx] [PATCH 2/2] drm/i915: Consolidate TLB invalidation flow
Andrzej Hajda
andrzej.hajda at intel.com
Wed Dec 14 09:02:54 UTC 2022
On 13.12.2022 16:22, Tvrtko Ursulin wrote:
>
> On 13/12/2022 14:52, Andrzej Hajda wrote:
>> On 13.12.2022 13:39, Tvrtko Ursulin wrote:
>>> From: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>>
>>> As the logic for selecting the register and corresponsing values
>>> grew, the
>>> code become a bit unsightly. Consolidate by storing the required
>>> values at
>>> engine init time in the engine itself, and by doing so minimise the
>>> amount
>>> of invariant platform and engine checks during each and every TLB
>>> invalidation.
>>>
>>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>> Cc: Andrzej Hajda <andrzej.hajda at intel.com>
>>> ---
>>> I think this looks nicer, but I don't really want to merge any
>>> flavour of
>>> consolidation until we get IGT coverage for the issue upstreamed.
>>
>>
>> Yep, the important is to have 1st patch merged.
>
> Agreed.
>
> 1)
> Can you send it standalone so it gets CI results and can get merged?
Done.
>
> 2)
> Could you check internally where did gem_exec_tlb get stuck? >
>>> ---
>>> drivers/gpu/drm/i915/gt/intel_engine_cs.c | 81 ++++++++++++
>>> drivers/gpu/drm/i915/gt/intel_engine_types.h | 16 +++
>>> drivers/gpu/drm/i915/gt/intel_gt.c | 126 ++++---------------
>>> 3 files changed, 120 insertions(+), 103 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
>>> b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
>>> index 99c4b866addd..97cdd9853e38 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
>>> +++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
>>> @@ -1143,6 +1143,84 @@ static int init_status_page(struct
>>> intel_engine_cs *engine)
>>> return ret;
>>> }
>>> +static bool intel_engine_init_tlb_invalidation(struct
>>> intel_engine_cs *engine)
>>> +{
>>> + static const union intel_engine_tlb_inv_reg gen8_regs[] = {
>>> + [RENDER_CLASS].reg = GEN8_RTCR,
>>> + [VIDEO_DECODE_CLASS].reg = GEN8_M1TCR, /* , GEN8_M2TCR */
>>> + [VIDEO_ENHANCEMENT_CLASS].reg = GEN8_VTCR,
>>> + [COPY_ENGINE_CLASS].reg = GEN8_BTCR,
>>> + };
>>> + static const union intel_engine_tlb_inv_reg gen12_regs[] = {
>>> + [RENDER_CLASS].reg = GEN12_GFX_TLB_INV_CR,
>>> + [VIDEO_DECODE_CLASS].reg = GEN12_VD_TLB_INV_CR,
>>> + [VIDEO_ENHANCEMENT_CLASS].reg = GEN12_VE_TLB_INV_CR,
>>> + [COPY_ENGINE_CLASS].reg = GEN12_BLT_TLB_INV_CR,
>>> + [COMPUTE_CLASS].reg = GEN12_COMPCTX_TLB_INV_CR,
>>> + };
>>> + static const union intel_engine_tlb_inv_reg xehp_regs[] = {
>>> + [RENDER_CLASS].mcr_reg = XEHP_GFX_TLB_INV_CR,
>>> + [VIDEO_DECODE_CLASS].mcr_reg = XEHP_VD_TLB_INV_CR,
>>> + [VIDEO_ENHANCEMENT_CLASS].mcr_reg = XEHP_VE_TLB_INV_CR,
>>> + [COPY_ENGINE_CLASS].mcr_reg = XEHP_BLT_TLB_INV_CR,
>>> + [COMPUTE_CLASS].mcr_reg = XEHP_COMPCTX_TLB_INV_CR,
>>> + };
>>> + struct drm_i915_private *i915 = engine->i915;
>>> + const union intel_engine_tlb_inv_reg *regs;
>>> + union intel_engine_tlb_inv_reg reg;
>>> + unsigned int class = engine->class;
>>> + unsigned int num = 0;
>>> + u32 val;
>>> +
>>> + if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50)) {
>>> + regs = xehp_regs;
>>> + num = ARRAY_SIZE(xehp_regs);
>>> + } else if (GRAPHICS_VER(i915) == 12) {
>>> + regs = gen12_regs;
>>> + num = ARRAY_SIZE(gen12_regs);
>>> + } else if (GRAPHICS_VER(i915) >= 8 && GRAPHICS_VER(i915) <= 11) {
>>
>> "GRAPHICS_VER(i915) <= 11" seems redundant.
>
> Code movement... Fixes: "patch which added >= 12,50" to clean just that
> part up. Before it was deliberately made to hit the below warn on so
> that any new platform added requires a human checking if the registers
> are still same and compatible. (Instead of assuming future platforms
> remain compatible - which is IMO too big risk to take.)
>
>>
>>> + regs = gen8_regs;
>>> + num = ARRAY_SIZE(gen8_regs);
>>> + } else if (GRAPHICS_VER(i915) < 8) {
>>
>> ditto
>
> How? We want to do nothing before gen 8.
} else {
return false;
}
?
As I understand there are three 'rules' we want to follow:
1. start from the newest hw, end at the oldest.
2. warn for not yet supported platforms, lets assume it is ver >= 13.
3. cover all cases
Why not use >= consistently, till exhaustion:
if (ver >= 13) {
warn
return false;
} else if (ver >= 12.50) {
...
} else if (ver >= 12) {
...
} else if (ver >= 8) {
...
} else {
return false;
}
It looks cleaner, ordered, no redundant checks, no doubts, hopefuly
future-proof.
Regards
Andrzej
>
>>> + return false;
>>> + }
>>> +
>>> + if (drm_WARN_ONCE(&i915->drm, !num,
>>> + "Platform does not implement TLB invalidation!"))
>>> + return false;
>>
>> It never happens.
>
> As above a) it's just code movement and b) it would be safer if it could
> happen.
>
>>> +
>>> + if (drm_WARN_ON_ONCE(&i915->drm,
>>> + class >= num ||
>>> + (!regs[class].reg.reg &&
>>> + !regs[class].mcr_reg.reg)))
>>> + return false;
>>> +
>>> + reg = regs[class];
>>> +
>>> + if (GRAPHICS_VER(i915) == 8 && class == VIDEO_DECODE_CLASS) {
>>> + reg.reg.reg += 4 * engine->instance; /* GEN8_M2TCR */
>>> + val = 0;
>>> + } else {
>>> + val = engine->instance;
>>> + }
>>> +
>>> + val = BIT(val);
>>> +
>>> + engine->tlb_inv.mcr = regs == xehp_regs;
>>> + engine->tlb_inv.reg = reg;
>>> + engine->tlb_inv.done = val;
>>> +
>>> + if (GRAPHICS_VER(i915) >= 12 &&
>>> + (engine->class == VIDEO_DECODE_CLASS ||
>>> + engine->class == VIDEO_ENHANCEMENT_CLASS ||
>>> + engine->class == COMPUTE_CLASS))
>>> + engine->tlb_inv.request = _MASKED_BIT_ENABLE(val);
>>> + else
>>> + engine->tlb_inv.request = val;
>>> +
>>> + return true;
>>> +}
>>> +
>>> static int engine_setup_common(struct intel_engine_cs *engine)
>>> {
>>> int err;
>>> @@ -1182,6 +1260,9 @@ static int engine_setup_common(struct
>>> intel_engine_cs *engine)
>>> intel_engine_init_whitelist(engine);
>>> intel_engine_init_ctx_wa(engine);
>>> + if (intel_engine_init_tlb_invalidation(engine))
>>> + engine->flags |= I915_ENGINE_HAS_TLB_INVALIDATION;
>>> +
>>> if (GRAPHICS_VER(engine->i915) >= 12)
>>> engine->flags |= I915_ENGINE_HAS_RELATIVE_MMIO;
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>> b/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>> index 4fd54fb8810f..8df4a09a6022 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>> +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
>>> @@ -341,6 +341,19 @@ struct intel_engine_guc_stats {
>>> u64 start_gt_clk;
>>> };
>>> +union intel_engine_tlb_inv_reg {
>>> + i915_reg_t reg;
>>> + i915_mcr_reg_t mcr_reg;
>>> +};
>>> +
>>> +struct intel_engine_tlb_inv
>>> +{
>>> + bool mcr;
>>> + union intel_engine_tlb_inv_reg reg;
>>> + u32 request;
>>> + u32 done;
>>> +};
>>> +
>>> struct intel_engine_cs {
>>> struct drm_i915_private *i915;
>>> struct intel_gt *gt;
>>> @@ -372,6 +385,8 @@ struct intel_engine_cs {
>>> u32 context_size;
>>> u32 mmio_base;
>>> + struct intel_engine_tlb_inv tlb_inv;
>>> +
>>> /*
>>> * Some w/a require forcewake to be held (which prevents RC6)
>>> while
>>> * a particular engine is active. If so, we set fw_domain to which
>>> @@ -556,6 +571,7 @@ struct intel_engine_cs {
>>> #define I915_ENGINE_HAS_EU_PRIORITY BIT(10)
>>> #define I915_ENGINE_FIRST_RENDER_COMPUTE BIT(11)
>>> #define I915_ENGINE_USES_WA_HOLD_CCS_SWITCHOUT BIT(12)
>>> +#define I915_ENGINE_HAS_TLB_INVALIDATION BIT(13)
>>> unsigned int flags;
>>> /*
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c
>>> b/drivers/gpu/drm/i915/gt/intel_gt.c
>>> index 7eeee5a7cb33..df7afff16fd6 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_gt.c
>>> +++ b/drivers/gpu/drm/i915/gt/intel_gt.c
>>> @@ -983,36 +983,6 @@ void intel_gt_info_print(const struct
>>> intel_gt_info *info,
>>> intel_sseu_dump(&info->sseu, p);
>>> }
>>> -struct reg_and_bit {
>>> - union {
>>> - i915_reg_t reg;
>>> - i915_mcr_reg_t mcr_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;
>>> -}
>>> -
>>> /*
>>> * HW architecture suggest typical invalidation time at 40us,
>>> * with pessimistic cases up to 100us and a recommendation to
>>> @@ -1026,14 +996,20 @@ get_reg_and_bit(const struct intel_engine_cs
>>> *engine, const bool gen8,
>>> * but are now considered MCR registers. Since they exist within a
>>> GAM range,
>>> * the primary instance of the register rolls up the status from
>>> each unit.
>>> */
>>> -static int wait_for_invalidate(struct intel_gt *gt, struct
>>> reg_and_bit rb)
>>> +static int wait_for_invalidate(struct intel_engine_cs *engine)
>>> {
>>> - if (GRAPHICS_VER_FULL(gt->i915) >= IP_VER(12, 50))
>>> - return intel_gt_mcr_wait_for_reg(gt, rb.mcr_reg, rb.bit, 0,
>>> + if (engine->tlb_inv.mcr)
>>> + return intel_gt_mcr_wait_for_reg(engine->gt,
>>> + engine->tlb_inv.reg.mcr_reg,
>>> + engine->tlb_inv.done,
>>> + 0,
>>> TLB_INVAL_TIMEOUT_US,
>>> TLB_INVAL_TIMEOUT_MS);
>>> else
>>> - return __intel_wait_for_register_fw(gt->uncore, rb.reg,
>>> rb.bit, 0,
>>> + return __intel_wait_for_register_fw(engine->gt->uncore,
>>> + engine->tlb_inv.reg.reg,
>>> + engine->tlb_inv.done,
>>> + 0,
>>> TLB_INVAL_TIMEOUT_US,
>>> TLB_INVAL_TIMEOUT_MS,
>>> NULL);
>>> @@ -1041,50 +1017,14 @@ static int wait_for_invalidate(struct
>>> intel_gt *gt, struct reg_and_bit rb)
>>> static void mmio_invalidate_full(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,
>>> - [COMPUTE_CLASS] = GEN12_COMPCTX_TLB_INV_CR,
>>> - };
>>> - static const i915_mcr_reg_t xehp_regs[] = {
>>> - [RENDER_CLASS] = XEHP_GFX_TLB_INV_CR,
>>> - [VIDEO_DECODE_CLASS] = XEHP_VD_TLB_INV_CR,
>>> - [VIDEO_ENHANCEMENT_CLASS] = XEHP_VE_TLB_INV_CR,
>>> - [COPY_ENGINE_CLASS] = XEHP_BLT_TLB_INV_CR,
>>> - [COMPUTE_CLASS] = XEHP_COMPCTX_TLB_INV_CR,
>>> - };
>>> struct drm_i915_private *i915 = gt->i915;
>>> struct intel_uncore *uncore = gt->uncore;
>>> struct intel_engine_cs *engine;
>>> intel_engine_mask_t awake, tmp;
>>> enum intel_engine_id id;
>>> - const i915_reg_t *regs;
>>> - unsigned int num = 0;
>>> unsigned long flags;
>>> - if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50)) {
>>> - regs = NULL;
>>> - num = ARRAY_SIZE(xehp_regs);
>>> - } else if (GRAPHICS_VER(i915) == 12) {
>>> - regs = gen12_regs;
>>> - num = ARRAY_SIZE(gen12_regs);
>>> - } else if (GRAPHICS_VER(i915) >= 8 && GRAPHICS_VER(i915) <= 11) {
>>> - regs = gen8_regs;
>>> - num = ARRAY_SIZE(gen8_regs);
>>> - } else if (GRAPHICS_VER(i915) < 8) {
>>> - return;
>>> - }
>>> -
>>> - if (drm_WARN_ONCE(&i915->drm, !num,
>>> - "Platform does not implement TLB invalidation!"))
>>> + if (GRAPHICS_VER(i915) < 8)
>>> return;
>>> intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
>>> @@ -1094,33 +1034,22 @@ static void mmio_invalidate_full(struct
>>> intel_gt *gt)
>>> awake = 0;
>>> for_each_engine(engine, gt, id) {
>>> - struct reg_and_bit rb;
>>> -
>>> if (!intel_engine_pm_is_awake(engine))
>>> continue;
>>> - if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50)) {
>>> - u32 val = BIT(engine->instance);
>>> + if (drm_WARN_ON_ONCE(&i915->drm,
>>> + !(engine->flags &
>>> I915_ENGINE_HAS_TLB_INVALIDATION)))
>>> + continue;
>>
>> Hmm, can this flag change dynamically? If not why not put it in
>> initialization phase.
>
> Could do. In that case I couldn't have any asserts that engine->tlb_inv
> is valid but perhaps that is okay. Could just fail engine probe from
> there if register table is incomplete. Sounds better indeed, I'll change
> it.
>
>>
>>> - if (engine->class == VIDEO_DECODE_CLASS ||
>>> - engine->class == VIDEO_ENHANCEMENT_CLASS ||
>>> - engine->class == COMPUTE_CLASS)
>>> - val = _MASKED_BIT_ENABLE(val);
>>> + if (engine->tlb_inv.mcr)
>>> intel_gt_mcr_multicast_write_fw(gt,
>>> - xehp_regs[engine->class],
>>> - val);
>>> - } else {
>>> - rb = get_reg_and_bit(engine, regs == gen8_regs, regs, num);
>>> - if (!i915_mmio_reg_offset(rb.reg))
>>> - continue;
>>> -
>>> - if (GRAPHICS_VER(i915) == 12 && (engine->class ==
>>> VIDEO_DECODE_CLASS ||
>>> - engine->class == VIDEO_ENHANCEMENT_CLASS ||
>>> - engine->class == COMPUTE_CLASS))
>>> - rb.bit = _MASKED_BIT_ENABLE(rb.bit);
>>> -
>>> - intel_uncore_write_fw(uncore, rb.reg, rb.bit);
>>> - }
>>> + engine->tlb_inv.reg.mcr_reg,
>>> + engine->tlb_inv.request);
>>> + else
>>> + intel_uncore_write_fw(uncore,
>>> + engine->tlb_inv.reg.reg,
>>> + engine->tlb_inv.request);
>>> +
>>> awake |= engine->mask;
>>> }
>>> @@ -1139,16 +1068,7 @@ static void mmio_invalidate_full(struct
>>> intel_gt *gt)
>>> intel_gt_mcr_unlock(gt, flags);
>>> for_each_engine_masked(engine, gt, awake, tmp) {
>>> - struct reg_and_bit rb;
>>> -
>>> - if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50)) {
>>> - rb.mcr_reg = xehp_regs[engine->class];
>>> - rb.bit = BIT(engine->instance);
>>> - } else {
>>> - rb = get_reg_and_bit(engine, regs == gen8_regs, regs, num);
>>> - }
>>> -
>>> - if (wait_for_invalidate(gt, rb))
>>> + if (wait_for_invalidate(engine))
>>> drm_err_ratelimited(>->i915->drm,
>>> "%s TLB invalidation did not complete in
>>> %ums!\n",
>>> engine->name, TLB_INVAL_TIMEOUT_MS);
>>
>> With minor comments addressed:
>>
>> Reviewed-by: Andrzej Hajda <andrzej.hajda at intel.com>
>
> I'll send v2 at some point, please stand by.
>
> Regards,
>
> Tvrtko
More information about the Intel-gfx
mailing list