[Intel-gfx] [PATCH] drm/i915/tgl: Register state context definition for Gen12
Daniele Ceraolo Spurio
daniele.ceraolospurio at intel.com
Thu Sep 5 22:48:44 UTC 2019
On 9/5/19 9:13 AM, Daniele Ceraolo Spurio wrote:
>
>
> On 9/5/2019 4:34 AM, Mika Kuoppala wrote:
>> From: Michel Thierry <michel.thierry at intel.com>
>>
>> Gen12 has subtle changes in the reg state context offsets (some fields
>> are gone, some are in a different location), compared to previous Gens.
>>
>> The simplest approach seems to be keeping Gen12 (and future platform)
>> changes apart from the previous gens, while keeping the registers that
>> are contiguous in functions we can reuse.
>>
>> v2: alias, virtual engine, rpcs, prune unused regs (Mika)
>>
>> Bspec: 46255
>> Cc: Michal Wajdeczko <michal.wajdeczko at intel.com>
>> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio at intel.com>
>> Cc: Chris Wilson <chris at chris-wilson.co.uk>
>> Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
>> Signed-off-by: Michel Thierry <michel.thierry at intel.com>
>> Signed-off-by: Lucas De Marchi <lucas.demarchi at intel.com>
>> Signed-off-by: Mika Kuoppala <mika.kuoppala at linux.intel.com>
>> ---
>> drivers/gpu/drm/i915/gt/intel_lrc.c | 194 +++++++++++++++++-------
>> drivers/gpu/drm/i915/gt/intel_lrc.h | 2 +
>> drivers/gpu/drm/i915/gt/intel_lrc_reg.h | 7 +-
>> 3 files changed, 146 insertions(+), 57 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c
>> b/drivers/gpu/drm/i915/gt/intel_lrc.c
>> index 87b7473a6dfb..f44c78a82915 100644
>> --- a/drivers/gpu/drm/i915/gt/intel_lrc.c
>> +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
>> @@ -810,6 +810,7 @@ static void virtual_update_register_offsets(u32
I've been looking a bit more into the relative MMIO stuff and I've
realized that we've been using it all along without realizing. When the
context is saved by the HW, the Add CS MMIO Start Offset flag is set in
the context LRIs (according to the specs this is true Gen11+, but I've
only explicitly verified it on TGL), so updating the offsets like we do
below doesn't make much sense. It all works only because during the ctx
restore (and only then, see Bspec 20206 & 45728) the HW is smart enough
to implicitly convert absolute offsets to relative ones and then add
back the correct MMIO base.
I've tried setting the flag in the context LRIs we write (to cover the
first submission) and disable this function on a TGL and I was able to
run gem_exec_balancer at nop, with the same context being submitted to
different engines without issues. Maybe that's a quicker solution
compared to keeping this function in sync with the context changes for
gen12, considering that a couple more updates to the context image are
going to be required for new functionality?
Daniele
>> *regs,
>> /* Must match execlists_init_reg_state()! */
>> + /* Common part */
>> regs[CTX_CONTEXT_CONTROL] =
>> i915_mmio_reg_offset(RING_CONTEXT_CONTROL(base));
>> regs[CTX_RING_HEAD] = i915_mmio_reg_offset(RING_HEAD(base));
>> @@ -820,13 +821,16 @@ static void virtual_update_register_offsets(u32
>> *regs,
>> regs[CTX_BB_HEAD_U] = i915_mmio_reg_offset(RING_BBADDR_UDW(base));
>> regs[CTX_BB_HEAD_L] = i915_mmio_reg_offset(RING_BBADDR(base));
>> regs[CTX_BB_STATE] = i915_mmio_reg_offset(RING_BBSTATE(base));
>> +
>> regs[CTX_SECOND_BB_HEAD_U] =
>> i915_mmio_reg_offset(RING_SBBADDR_UDW(base));
>> regs[CTX_SECOND_BB_HEAD_L] =
>> i915_mmio_reg_offset(RING_SBBADDR(base));
>> regs[CTX_SECOND_BB_STATE] =
>> i915_mmio_reg_offset(RING_SBBSTATE(base));
>
> We should do those 3 only on pre-gen12, or at least add a comment saying
> that those are not applicable to gen12+ and we rely to follow-up writes
> to overwrite them.
>
>> + /* PPGTT part */
>> regs[CTX_CTX_TIMESTAMP] =
>> i915_mmio_reg_offset(RING_CTX_TIMESTAMP(base));
>> +
>> regs[CTX_PDP3_UDW] =
>> i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, 3));
>> regs[CTX_PDP3_LDW] =
>> i915_mmio_reg_offset(GEN8_RING_PDP_LDW(base, 3));
>> regs[CTX_PDP2_UDW] =
>> i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, 2));
>> @@ -844,8 +848,12 @@ static void virtual_update_register_offsets(u32
>> *regs,
>> regs[CTX_BB_PER_CTX_PTR] =
>> i915_mmio_reg_offset(RING_BB_PER_CTX_PTR(base));
>
> CTX_BB_PER_CTX_PTR and the 2 indirect ctx entries are different on gen12
> as well
>
>> - regs[CTX_R_PWR_CLK_STATE] =
>> - i915_mmio_reg_offset(GEN8_R_PWR_CLK_STATE);
>> + if (INTEL_GEN(engine->i915) < 11)
>
> This should be < 12
>
>> + regs[CTX_R_PWR_CLK_STATE] =
>> + i915_mmio_reg_offset(GEN8_R_PWR_CLK_STATE);
>> + else
>> + regs[GEN12_CTX_R_PWR_CLK_STATE] =
>> + i915_mmio_reg_offset(GEN8_R_PWR_CLK_STATE);
>> }
>> }
>> @@ -1747,8 +1755,12 @@ __execlists_update_reg_state(struct
>> intel_context *ce,
>> /* RPCS */
>> if (engine->class == RENDER_CLASS) {
>> - regs[CTX_R_PWR_CLK_STATE + 1] =
>> - intel_sseu_make_rpcs(engine->i915, &ce->sseu);
>> + const u32 rpcs = intel_sseu_make_rpcs(engine->i915, &ce->sseu);
>> + const u32 offset = INTEL_GEN(engine->i915) < 11 ?
>
> < 12
>
>> + CTX_R_PWR_CLK_STATE + 1 :
>> + GEN12_CTX_R_PWR_CLK_STATE + 1;
>> +
>> + regs[offset] = rpcs;
>> i915_oa_init_reg_state(engine, ce, regs);
>> }
>> @@ -3116,35 +3128,13 @@ static u32 intel_lr_indirect_ctx_offset(struct
>> intel_engine_cs *engine)
>> return indirect_ctx_offset;
>> }
>> -static struct i915_ppgtt *vm_alias(struct i915_address_space *vm)
>> -{
>> - if (i915_is_ggtt(vm))
>> - return i915_vm_to_ggtt(vm)->alias;
>> - else
>> - return i915_vm_to_ppgtt(vm);
>> -}
>> -static void execlists_init_reg_state(u32 *regs,
>> - struct intel_context *ce,
>> - struct intel_engine_cs *engine,
>> - struct intel_ring *ring)
>> +static void init_common_reg_state(u32 * const regs,
>> + struct i915_ppgtt * const ppgtt,
>> + struct intel_engine_cs *engine,
>> + struct intel_ring *ring)
>> {
>> - struct i915_ppgtt *ppgtt = vm_alias(ce->vm);
>> - bool rcs = engine->class == RENDER_CLASS;
>> - u32 base = engine->mmio_base;
>> -
>> - /*
>> - * A context is actually a big batch buffer with several
>> - * MI_LOAD_REGISTER_IMM commands followed by (reg, value) pairs. The
>> - * values we are setting here are only for the first context
>> restore:
>> - * on a subsequent save, the GPU will recreate this batchbuffer
>> with new
>> - * values (including all the missing MI_LOAD_REGISTER_IMM
>> commands that
>> - * we are not initializing here).
>> - *
>> - * Must keep consistent with virtual_update_register_offsets().
>> - */
>> - regs[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(rcs ? 14 : 11) |
>> - MI_LRI_FORCE_POSTED;
>> + const u32 base = engine->mmio_base;
>> CTX_REG(regs, CTX_CONTEXT_CONTROL, RING_CONTEXT_CONTROL(base),
>> _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT) |
>> @@ -3162,38 +3152,44 @@ static void execlists_init_reg_state(u32 *regs,
>> CTX_REG(regs, CTX_BB_HEAD_U, RING_BBADDR_UDW(base), 0);
>> CTX_REG(regs, CTX_BB_HEAD_L, RING_BBADDR(base), 0);
>> CTX_REG(regs, CTX_BB_STATE, RING_BBSTATE(base), RING_BB_PPGTT);
>> - CTX_REG(regs, CTX_SECOND_BB_HEAD_U, RING_SBBADDR_UDW(base), 0);
>> - CTX_REG(regs, CTX_SECOND_BB_HEAD_L, RING_SBBADDR(base), 0);
>> - CTX_REG(regs, CTX_SECOND_BB_STATE, RING_SBBSTATE(base), 0);
>> - if (rcs) {
>> - struct i915_ctx_workarounds *wa_ctx = &engine->wa_ctx;
>> -
>> - CTX_REG(regs, CTX_RCS_INDIRECT_CTX, RING_INDIRECT_CTX(base), 0);
>> - CTX_REG(regs, CTX_RCS_INDIRECT_CTX_OFFSET,
>> - RING_INDIRECT_CTX_OFFSET(base), 0);
>> - if (wa_ctx->indirect_ctx.size) {
>> - u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
>> +}
>> - regs[CTX_RCS_INDIRECT_CTX + 1] =
>> - (ggtt_offset + wa_ctx->indirect_ctx.offset) |
>> - (wa_ctx->indirect_ctx.size / CACHELINE_BYTES);
>> +static void init_wa_bb_reg_state(u32 * const regs,
>> + struct intel_engine_cs *engine,
>> + u32 pos_bb_per_ctx)
>> +{
>> + struct i915_ctx_workarounds * const wa_ctx = &engine->wa_ctx;
>> + const u32 base = engine->mmio_base;
>> + const u32 pos_indirect_ctx = pos_bb_per_ctx + 2;
>> + const u32 pos_indirect_ctx_offset = pos_indirect_ctx + 2;
>> - regs[CTX_RCS_INDIRECT_CTX_OFFSET + 1] =
>> - intel_lr_indirect_ctx_offset(engine) << 6;
>> - }
>> + GEM_BUG_ON(engine->class != RENDER_CLASS);
>> + CTX_REG(regs, pos_indirect_ctx, RING_INDIRECT_CTX(base), 0);
>> + CTX_REG(regs, pos_indirect_ctx_offset,
>> + RING_INDIRECT_CTX_OFFSET(base), 0);
>> + if (wa_ctx->indirect_ctx.size) {
>> + const u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
>> - CTX_REG(regs, CTX_BB_PER_CTX_PTR, RING_BB_PER_CTX_PTR(base), 0);
>> - if (wa_ctx->per_ctx.size) {
>> - u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
>> + regs[pos_indirect_ctx + 1] =
>> + (ggtt_offset + wa_ctx->indirect_ctx.offset) |
>> + (wa_ctx->indirect_ctx.size / CACHELINE_BYTES);
>> - regs[CTX_BB_PER_CTX_PTR + 1] =
>> - (ggtt_offset + wa_ctx->per_ctx.offset) | 0x01;
>> - }
>> + regs[pos_indirect_ctx_offset + 1] =
>> + intel_lr_indirect_ctx_offset(engine) << 6;
>> }
>> - regs[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9) |
>> MI_LRI_FORCE_POSTED;
>> + CTX_REG(regs, pos_bb_per_ctx, RING_BB_PER_CTX_PTR(base), 0);
>> + if (wa_ctx->per_ctx.size) {
>> + const u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
>> - CTX_REG(regs, CTX_CTX_TIMESTAMP, RING_CTX_TIMESTAMP(base), 0);
>> + regs[pos_bb_per_ctx + 1] =
>> + (ggtt_offset + wa_ctx->per_ctx.offset) | 0x01;
>> + }
>> +}
>> +
>> +static void init_ppgtt_reg_state(u32 *regs, u32 base,
>> + struct i915_ppgtt *ppgtt)
>> +{
>> /* PDP values well be assigned later if needed */
>> CTX_REG(regs, CTX_PDP3_UDW, GEN8_RING_PDP_UDW(base, 3), 0);
>> CTX_REG(regs, CTX_PDP3_LDW, GEN8_RING_PDP_LDW(base, 3), 0);
>> @@ -3216,6 +3212,40 @@ static void execlists_init_reg_state(u32 *regs,
>> ASSIGN_CTX_PDP(ppgtt, regs, 1);
>> ASSIGN_CTX_PDP(ppgtt, regs, 0);
>> }
>> +}
>> +
>> +static struct i915_ppgtt *vm_alias(struct i915_address_space *vm)
>> +{
>> + if (i915_is_ggtt(vm))
>> + return i915_vm_to_ggtt(vm)->alias;
>> + else
>> + return i915_vm_to_ppgtt(vm);
>> +}
>> +
>> +static void gen8_init_reg_state(u32 * const regs,
>> + struct intel_context *ce,
>> + struct intel_engine_cs *engine,
>> + struct intel_ring *ring)
>> +{
>> + struct i915_ppgtt * const ppgtt = vm_alias(ce->vm);
>> + const bool rcs = engine->class == RENDER_CLASS;
>> + const u32 base = engine->mmio_base;
>> +
>> + regs[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(rcs ? 14 : 11) |
>> + MI_LRI_FORCE_POSTED;
>> +
>> + init_common_reg_state(regs, ppgtt, engine, ring);
>> + CTX_REG(regs, CTX_SECOND_BB_HEAD_U, RING_SBBADDR_UDW(base), 0);
>> + CTX_REG(regs, CTX_SECOND_BB_HEAD_L, RING_SBBADDR(base), 0);
>> + CTX_REG(regs, CTX_SECOND_BB_STATE, RING_SBBSTATE(base), 0);
>> + if (rcs)
>> + init_wa_bb_reg_state(regs, engine, CTX_BB_PER_CTX_PTR);
>> +
>> + regs[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9) |
>> MI_LRI_FORCE_POSTED;
>> +
>> + CTX_REG(regs, CTX_CTX_TIMESTAMP, RING_CTX_TIMESTAMP(base), 0);
>> +
>> + init_ppgtt_reg_state(regs, base, ppgtt);
>> if (rcs) {
>> regs[CTX_LRI_HEADER_2] = MI_LOAD_REGISTER_IMM(1);
>> @@ -3227,6 +3257,58 @@ static void execlists_init_reg_state(u32 *regs,
>> regs[CTX_END] |= BIT(0);
>> }
>> +static void gen12_init_reg_state(u32 * const regs,
>> + struct intel_context *ce,
>> + struct intel_engine_cs *engine,
>> + struct intel_ring *ring)
>> +{
>> + struct i915_ppgtt * const ppgtt = vm_alias(ce->vm);
>
> Do we expect appgtt gen12+? If not we could use i915_vm_to_ppgtt() directly
>
>> + const bool rcs = engine->class == RENDER_CLASS;
>> + const u32 base = engine->mmio_base;
>> +
>> + GEM_DEBUG_EXEC(DRM_INFO_ONCE("Using GEN12 Register State
>> Context\n"));
>> +
>> + regs[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(13) |
>> + MI_LRI_FORCE_POSTED;
>> +
>> + init_common_reg_state(regs, ppgtt, engine, ring);
>> + if (rcs)
>> + init_wa_bb_reg_state(regs, engine, GEN12_CTX_BB_PER_CTX_PTR);
>> +
>> + regs[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9) |
>> + MI_LRI_FORCE_POSTED;
>> +
>> + CTX_REG(regs, CTX_CTX_TIMESTAMP, RING_CTX_TIMESTAMP(base), 0);
>> +
>> + init_ppgtt_reg_state(regs, base, ppgtt);
>> +
>> + if (rcs) {
>> + regs[GEN12_CTX_LRI_HEADER_3] = MI_LOAD_REGISTER_IMM(1);
>> + CTX_REG(regs, GEN12_CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE,
>> + 0);
>> +
>> + /* TODO: oa_init_reg_state ? */
>> + }
>> +}
>> +
>> +static void execlists_init_reg_state(u32 *regs,
>> + struct intel_context *ce,
>> + struct intel_engine_cs *engine,
>> + struct intel_ring *ring)
>> +{
>> + /* A context is actually a big batch buffer with several
>> + * MI_LOAD_REGISTER_IMM commands followed by (reg, value) pairs. The
>> + * values we are setting here are only for the first context
>> restore:
>> + * on a subsequent save, the GPU will recreate this batchbuffer
>> with new
>> + * values (including all the missing MI_LOAD_REGISTER_IMM
>> commands that
>> + * we are not initializing here).
>> + */
>> + if (INTEL_GEN(engine->i915) >= 12)
>> + gen12_init_reg_state(regs, ce, engine, ring);
>> + else
>> + gen8_init_reg_state(regs, ce, engine, ring);
>> +}
>> +
>> static int
>> populate_lr_context(struct intel_context *ce,
>> struct drm_i915_gem_object *ctx_obj,
>> diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.h
>> b/drivers/gpu/drm/i915/gt/intel_lrc.h
>> index dc0252e0589e..a4bde38e35d8 100644
>> --- a/drivers/gpu/drm/i915/gt/intel_lrc.h
>> +++ b/drivers/gpu/drm/i915/gt/intel_lrc.h
>> @@ -49,6 +49,8 @@ struct intel_engine_cs;
>> #define EL_CTRL_LOAD (1 << 0)
>> +#define GEN12_ENGINE_SEMAPHORE_TOKEN(engine)
>> _MMIO((engine)->mmio_base + 0x2b4)
>> +
>
> As I commented on the previous revision, defining this in this patch
> without using it or setting it in the context doesn't make much sense.
>
>> /* The docs specify that the write pointer wraps around after 5h,
>> "After status
>> * is written out to the last available status QW at offset 5h, this
>> pointer
>> * wraps to 0."
>> diff --git a/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
>> b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
>> index b8f20ad71169..8cb6655744cd 100644
>> --- a/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
>> +++ b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
>> @@ -9,7 +9,7 @@
>> #include <linux/types.h>
>> -/* GEN8+ Reg State Context */
>> +/* GEN8 to GEN11 Reg State Context */
>> #define CTX_LRI_HEADER_0 0x01
>> #define CTX_CONTEXT_CONTROL 0x02
>> #define CTX_RING_HEAD 0x04
>> @@ -39,6 +39,11 @@
>> #define CTX_R_PWR_CLK_STATE 0x42
>> #define CTX_END 0x44
>> +/* GEN12+ Reg State Context */
>> +#define GEN12_CTX_BB_PER_CTX_PTR 0x12
>> +#define GEN12_CTX_LRI_HEADER_3 0x41
>
> Gen11 has the same number and positioning of LRIs as gen12, so marking
> this as a gen12 change doesn't feel too reflective of teh HW changes
> IMO, but not a blocker.
>
>
>> +#define GEN12_CTX_R_PWR_CLK_STATE 0x42
>
> This hasn't moved compared to gen11:
> #define CTX_R_PWR_CLK_STATE 0x42
>
> so no need for a different defs and all those ifs higher up in the patch.
>
> Daniele
>
>> +
>> #define CTX_REG(reg_state, pos, reg, val) do { \
>> u32 *reg_state__ = (reg_state); \
>> const u32 pos__ = (pos); \
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
More information about the Intel-gfx
mailing list