[Intel-gfx] [PATCH] drm/i915: Engine relative MMIO
Tvrtko Ursulin
tvrtko.ursulin at linux.intel.com
Wed May 15 08:52:17 UTC 2019
On 13/05/2019 20:45, John.C.Harrison at Intel.com wrote:
> From: John Harrison <John.C.Harrison at Intel.com>
>
> With virtual engines, it is no longer possible to know which specific
> physical engine a given request will be executed on at the time that
> request is generated. This means that the request itself must be engine
> agnostic - any direct register writes must be relative to the engine
> and not absolute addresses.
>
> The LRI command has support for engine relative addressing. However,
> the mechanism is not transparent to the driver. The scheme for Gen11
> (MI_LRI_ADD_CS_MMIO_START) requires the LRI address to have no
> absolute engine base component. The hardware then adds on the correct
> engine offset at execution time.
>
> Due to the non-trivial and differing schemes on different hardware, it
> is not possible to simply update the code that creates the LRI
> commands to set a remap flag and let the hardware get on with it.
> Instead, this patch adds function wrappers for generating the LRI
> command itself and then for constructing the correct address to use
> with the LRI.
>
> v2: Fix build break in GVT. Remove flags parameter [review feedback
> from Chris W].
>
> v3: Fix build break in selftest. Rebase to newer base tree and fix
> merge conflict.
>
> v4: More rebasing. Rmoved relative addressing support from Gen7-9 only
> code paths [review feedback from Chris W].
>
> v5: More rebasing (new 'gt' directory). Fix white space issue. Use
> COPY class rather than BCS0 id for checking against BCS engine.
>
> Signed-off-by: John Harrison <John.C.Harrison at Intel.com>
> CC: Rodrigo Vivi <rodrigo.vivi at intel.com>
> CC: Tvrtko Ursulin <tvrtko.ursulin at linux.intel.com>
> CC: Wilson, Chris P <chris.p.wilson at intel.com>
> ---
> drivers/gpu/drm/i915/gt/intel_engine.h | 4 +
> drivers/gpu/drm/i915/gt/intel_engine_cs.c | 11 +++
> drivers/gpu/drm/i915/gt/intel_gpu_commands.h | 6 +-
> drivers/gpu/drm/i915/gt/intel_lrc.c | 79 ++++++++++---------
> drivers/gpu/drm/i915/gt/intel_lrc_reg.h | 4 +-
> drivers/gpu/drm/i915/gt/intel_mocs.c | 17 ++--
> drivers/gpu/drm/i915/gt/intel_ringbuffer.c | 45 +++++++++--
> drivers/gpu/drm/i915/gt/intel_workarounds.c | 4 +-
> .../gpu/drm/i915/gt/selftest_workarounds.c | 9 ++-
> drivers/gpu/drm/i915/gvt/mmio_context.c | 16 +++-
> drivers/gpu/drm/i915/i915_cmd_parser.c | 4 +-
> drivers/gpu/drm/i915/i915_gem_context.c | 12 +--
> drivers/gpu/drm/i915/i915_gem_execbuffer.c | 3 +-
> drivers/gpu/drm/i915/i915_perf.c | 19 +++--
> 14 files changed, 154 insertions(+), 79 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_engine.h b/drivers/gpu/drm/i915/gt/intel_engine.h
> index 9359b3a7ad9c..3506c992182c 100644
> --- a/drivers/gpu/drm/i915/gt/intel_engine.h
> +++ b/drivers/gpu/drm/i915/gt/intel_engine.h
> @@ -546,4 +546,8 @@ static inline bool inject_preempt_hang(struct intel_engine_execlists *execlists)
>
> #endif
>
> +bool i915_engine_has_relative_lri(const struct intel_engine_cs *engine);
> +u32 i915_get_lri_cmd(const struct intel_engine_cs *engine, u32 word_count);
> +u32 i915_get_lri_reg(const struct intel_engine_cs *engine, i915_reg_t reg);
> +
> #endif /* _INTEL_RINGBUFFER_H_ */
> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
> index 4c3753c1b573..233295d689d2 100644
> --- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
> +++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
> @@ -253,6 +253,17 @@ static u32 __engine_mmio_base(struct drm_i915_private *i915,
> return bases[i].base;
> }
>
> +bool i915_engine_has_relative_lri(const struct intel_engine_cs *engine)
> +{
> + if (INTEL_GEN(engine->i915) < 11)
> + return false;
> +
> + if (engine->class == COPY_ENGINE_CLASS)
> + return false;
> +
> + return true;
I think engine->flags would be better. At least it is one conditional
instead of two at runtime, even one too many for something that is
invariant.
> +}
> +
> static void __sprint_engine_name(char *name, const struct engine_info *info)
> {
> WARN_ON(snprintf(name, INTEL_ENGINE_CS_MAX_NAME, "%s%u",
> diff --git a/drivers/gpu/drm/i915/gt/intel_gpu_commands.h b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> index a34ece53a771..e7784b3fb759 100644
> --- a/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> +++ b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
> @@ -123,9 +123,13 @@
> * simply ignores the register load under certain conditions.
> * - One can actually load arbitrary many arbitrary registers: Simply issue x
> * address/value pairs. Don't overdue it, though, x <= 2^4 must hold!
> + * - Newer hardware supports engine relative addresses but older hardware does
> + * not. So never call MI_LRI directly, always use the i915_get_lri_cmd()
> + * and i915_get_lri_reg() helper functions.
> */
> -#define MI_LOAD_REGISTER_IMM(x) MI_INSTR(0x22, 2*(x)-1)
> +#define __MI_LOAD_REGISTER_IMM(x) MI_INSTR(0x22, 2*(x)-1)
So we end up with code using __MI_LOAD_REGISTER_IMM for absolute, and
i915_get_lri_cmd for relative addressing. One is a macro, another is a
function call, and naming/case is different.
How about we static inline it like:
static inline u32
MI_LOAD_REGISTER_IMM(struct intel_engine_cs *engine,
u32 count,
unsigned int flags)
#define MI_LRI_ABSOLUTE (0)
#define MI_LRI_RELATIVE (1)
{
u32 cmd = __MI_LOAD_REGISTER_IMM(count);
if ((flags & MI_LRI_RELATIVE) &&
(engine->flags & ...HAS_RELATIVE_LRI))
cmd |= ...;
return cmd;
}
Then all call sites can use just this single helper and the naming
remains familiar.
Register offset can use the same approach - again it is also small
enough to be either static inline or a macro. Function calls for both
seem excessive, but maybe it is just me.
> #define MI_LRI_FORCE_POSTED (1<<12)
> +#define MI_LRI_ADD_CS_MMIO_START_GEN11 (1<<19)
> #define MI_STORE_REGISTER_MEM MI_INSTR(0x24, 1)
> #define MI_STORE_REGISTER_MEM_GEN8 MI_INSTR(0x24, 2)
> #define MI_SRM_LRM_GLOBAL_GTT (1<<22)
> diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
> index e18623def282..49a9a6648b9c 100644
> --- a/drivers/gpu/drm/i915/gt/intel_lrc.c
> +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
> @@ -1388,14 +1388,15 @@ static int emit_pdps(struct i915_request *rq)
> return PTR_ERR(cs);
>
> /* Ensure the LRI have landed before we invalidate & continue */
> - *cs++ = MI_LOAD_REGISTER_IMM(2 * GEN8_3LVL_PDPES) | MI_LRI_FORCE_POSTED;
> + *cs++ = i915_get_lri_cmd(engine, 2 * GEN8_3LVL_PDPES) |
> + MI_LRI_FORCE_POSTED;
> for (i = GEN8_3LVL_PDPES; i--; ) {
> const dma_addr_t pd_daddr = i915_page_dir_dma_addr(ppgtt, i);
> u32 base = engine->mmio_base;
>
> - *cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, i));
> + *cs++ = i915_get_lri_reg(engine, GEN8_RING_PDP_UDW(base, i));
> *cs++ = upper_32_bits(pd_daddr);
> - *cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(base, i));
> + *cs++ = i915_get_lri_reg(engine, GEN8_RING_PDP_LDW(base, i));
> *cs++ = lower_32_bits(pd_daddr);
> }
> *cs++ = MI_NOOP;
> @@ -1469,7 +1470,8 @@ gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine, u32 *batch)
> *batch++ = i915_scratch_offset(engine->i915) + 256;
> *batch++ = 0;
>
> - *batch++ = MI_LOAD_REGISTER_IMM(1);
> + /* Gen8/9 only so no need to support relative offsets */
> + *batch++ = __MI_LOAD_REGISTER_IMM(1);
> *batch++ = i915_mmio_reg_offset(GEN8_L3SQCREG4);
> *batch++ = 0x40400000 | GEN8_LQSC_FLUSH_COHERENT_LINES;
>
> @@ -1540,13 +1542,14 @@ struct lri {
> u32 value;
> };
>
> -static u32 *emit_lri(u32 *batch, const struct lri *lri, unsigned int count)
> +static u32 *emit_lri(struct intel_engine_cs *engine, u32 *batch,
> + const struct lri *lri, unsigned int count)
> {
> GEM_BUG_ON(!count || count > 63);
>
> - *batch++ = MI_LOAD_REGISTER_IMM(count);
> + *batch++ = i915_get_lri_cmd(engine, count);
> do {
> - *batch++ = i915_mmio_reg_offset(lri->reg);
> + *batch++ = i915_get_lri_reg(engine, lri->reg);
> *batch++ = lri->value;
> } while (lri++, --count);
> *batch++ = MI_NOOP;
> @@ -1584,7 +1587,7 @@ static u32 *gen9_init_indirectctx_bb(struct intel_engine_cs *engine, u32 *batch)
> /* WaFlushCoherentL3CacheLinesAtContextSwitch:skl,bxt,glk */
> batch = gen8_emit_flush_coherentl3_wa(engine, batch);
>
> - batch = emit_lri(batch, lri, ARRAY_SIZE(lri));
> + batch = emit_lri(engine, batch, lri, ARRAY_SIZE(lri));
>
> /* WaMediaPoolStateCmdInWABB:bxt,glk */
> if (HAS_POOLED_EU(engine->i915)) {
> @@ -2537,10 +2540,10 @@ static void execlists_init_reg_state(u32 *regs,
> * values (including all the missing MI_LOAD_REGISTER_IMM commands that
> * we are not initializing here).
> */
> - regs[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(rcs ? 14 : 11) |
> - MI_LRI_FORCE_POSTED;
> + regs[CTX_LRI_HEADER_0] = i915_get_lri_cmd(engine, rcs ? 14 : 11) |
> + MI_LRI_FORCE_POSTED;
>
> - CTX_REG(regs, CTX_CONTEXT_CONTROL, RING_CONTEXT_CONTROL(base),
> + CTX_REG(engine, regs, CTX_CONTEXT_CONTROL, RING_CONTEXT_CONTROL(base),
> _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT) |
> _MASKED_BIT_ENABLE(CTX_CTRL_INHIBIT_SYN_CTX_SWITCH));
> if (INTEL_GEN(engine->i915) < 11) {
> @@ -2548,22 +2551,23 @@ static void execlists_init_reg_state(u32 *regs,
> _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT |
> CTX_CTRL_RS_CTX_ENABLE);
> }
> - CTX_REG(regs, CTX_RING_HEAD, RING_HEAD(base), 0);
> - CTX_REG(regs, CTX_RING_TAIL, RING_TAIL(base), 0);
> - CTX_REG(regs, CTX_RING_BUFFER_START, RING_START(base), 0);
> - CTX_REG(regs, CTX_RING_BUFFER_CONTROL, RING_CTL(base),
> + CTX_REG(engine, regs, CTX_RING_HEAD, RING_HEAD(base), 0);
> + CTX_REG(engine, regs, CTX_RING_TAIL, RING_TAIL(base), 0);
> + CTX_REG(engine, regs, CTX_RING_BUFFER_START, RING_START(base), 0);
> + CTX_REG(engine, regs, CTX_RING_BUFFER_CONTROL, RING_CTL(base),
> RING_CTL_SIZE(ring->size) | RING_VALID);
> - 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);
> + CTX_REG(engine, regs, CTX_BB_HEAD_U, RING_BBADDR_UDW(base), 0);
> + CTX_REG(engine, regs, CTX_BB_HEAD_L, RING_BBADDR(base), 0);
> + CTX_REG(engine, regs, CTX_BB_STATE, RING_BBSTATE(base), RING_BB_PPGTT);
> + CTX_REG(engine, regs, CTX_SECOND_BB_HEAD_U, RING_SBBADDR_UDW(base), 0);
> + CTX_REG(engine, regs, CTX_SECOND_BB_HEAD_L, RING_SBBADDR(base), 0);
> + CTX_REG(engine, 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,
> + CTX_REG(engine, regs, CTX_RCS_INDIRECT_CTX,
> + RING_INDIRECT_CTX(base), 0);
> + CTX_REG(engine, 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);
> @@ -2576,7 +2580,8 @@ static void execlists_init_reg_state(u32 *regs,
> intel_lr_indirect_ctx_offset(engine) << 6;
> }
>
> - CTX_REG(regs, CTX_BB_PER_CTX_PTR, RING_BB_PER_CTX_PTR(base), 0);
> + CTX_REG(engine, 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);
>
> @@ -2585,18 +2590,19 @@ static void execlists_init_reg_state(u32 *regs,
> }
> }
>
> - regs[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9) | MI_LRI_FORCE_POSTED;
> + regs[CTX_LRI_HEADER_1] = i915_get_lri_cmd(engine, 9) |
> + MI_LRI_FORCE_POSTED;
>
> - CTX_REG(regs, CTX_CTX_TIMESTAMP, RING_CTX_TIMESTAMP(base), 0);
> + CTX_REG(engine, regs, CTX_CTX_TIMESTAMP, RING_CTX_TIMESTAMP(base), 0);
> /* 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);
> - CTX_REG(regs, CTX_PDP2_UDW, GEN8_RING_PDP_UDW(base, 2), 0);
> - CTX_REG(regs, CTX_PDP2_LDW, GEN8_RING_PDP_LDW(base, 2), 0);
> - CTX_REG(regs, CTX_PDP1_UDW, GEN8_RING_PDP_UDW(base, 1), 0);
> - CTX_REG(regs, CTX_PDP1_LDW, GEN8_RING_PDP_LDW(base, 1), 0);
> - CTX_REG(regs, CTX_PDP0_UDW, GEN8_RING_PDP_UDW(base, 0), 0);
> - CTX_REG(regs, CTX_PDP0_LDW, GEN8_RING_PDP_LDW(base, 0), 0);
> + CTX_REG(engine, regs, CTX_PDP3_UDW, GEN8_RING_PDP_UDW(base, 3), 0);
> + CTX_REG(engine, regs, CTX_PDP3_LDW, GEN8_RING_PDP_LDW(base, 3), 0);
> + CTX_REG(engine, regs, CTX_PDP2_UDW, GEN8_RING_PDP_UDW(base, 2), 0);
> + CTX_REG(engine, regs, CTX_PDP2_LDW, GEN8_RING_PDP_LDW(base, 2), 0);
> + CTX_REG(engine, regs, CTX_PDP1_UDW, GEN8_RING_PDP_UDW(base, 1), 0);
> + CTX_REG(engine, regs, CTX_PDP1_LDW, GEN8_RING_PDP_LDW(base, 1), 0);
> + CTX_REG(engine, regs, CTX_PDP0_UDW, GEN8_RING_PDP_UDW(base, 0), 0);
> + CTX_REG(engine, regs, CTX_PDP0_LDW, GEN8_RING_PDP_LDW(base, 0), 0);
>
> if (i915_vm_is_4lvl(&ppgtt->vm)) {
> /* 64b PPGTT (48bit canonical)
> @@ -2612,8 +2618,9 @@ static void execlists_init_reg_state(u32 *regs,
> }
>
> if (rcs) {
> - regs[CTX_LRI_HEADER_2] = MI_LOAD_REGISTER_IMM(1);
> - CTX_REG(regs, CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE, 0);
> + regs[CTX_LRI_HEADER_2] = i915_get_lri_cmd(engine, 1);
> + CTX_REG(engine, regs, CTX_R_PWR_CLK_STATE,
> + GEN8_R_PWR_CLK_STATE, 0);
>
> i915_oa_init_reg_state(engine, ce, regs);
> }
> diff --git a/drivers/gpu/drm/i915/gt/intel_lrc_reg.h b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
> index 5ef932d810a7..40b1142d0d74 100644
> --- a/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
> +++ b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
> @@ -39,10 +39,10 @@
> #define CTX_R_PWR_CLK_STATE 0x42
> #define CTX_END 0x44
>
> -#define CTX_REG(reg_state, pos, reg, val) do { \
> +#define CTX_REG(engine, reg_state, pos, reg, val) do { \
> u32 *reg_state__ = (reg_state); \
> const u32 pos__ = (pos); \
> - (reg_state__)[(pos__) + 0] = i915_mmio_reg_offset(reg); \
> + (reg_state__)[(pos__) + 0] = i915_get_lri_reg((engine), (reg)); \
> (reg_state__)[(pos__) + 1] = (val); \
> } while (0)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_mocs.c b/drivers/gpu/drm/i915/gt/intel_mocs.c
> index 79df66022d3a..5dae6333481d 100644
> --- a/drivers/gpu/drm/i915/gt/intel_mocs.c
> +++ b/drivers/gpu/drm/i915/gt/intel_mocs.c
> @@ -324,9 +324,6 @@ static u32 get_entry_control(const struct drm_i915_mocs_table *table,
> /**
> * intel_mocs_init_engine() - emit the mocs control table
> * @engine: The engine for whom to emit the registers.
> - *
> - * This function simply emits a MI_LOAD_REGISTER_IMM command for the
> - * given table starting at the given address.
> */
> void intel_mocs_init_engine(struct intel_engine_cs *engine)
> {
> @@ -380,18 +377,20 @@ static int emit_mocs_control_table(struct i915_request *rq,
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(table->n_entries);
> + *cs++ = i915_get_lri_cmd(rq->engine, table->n_entries);
>
> for (index = 0; index < table->size; index++) {
> u32 value = get_entry_control(table, index);
>
> - *cs++ = i915_mmio_reg_offset(mocs_register(engine, index));
> + *cs++ = i915_get_lri_reg(rq->engine,
> + mocs_register(engine, index));
> *cs++ = value;
> }
>
> /* All remaining entries are also unused */
> for (; index < table->n_entries; index++) {
> - *cs++ = i915_mmio_reg_offset(mocs_register(engine, index));
> + *cs++ = i915_get_lri_reg(rq->engine,
> + mocs_register(engine, index));
> *cs++ = unused_value;
> }
>
> @@ -449,7 +448,11 @@ static int emit_mocs_l3cc_table(struct i915_request *rq,
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(table->n_entries / 2);
> + /*
> + * GEN9_LNCFCMOCS is not engine relative, therefore there is no
> + * need for relative addressing?
> + */
> + *cs++ = __MI_LOAD_REGISTER_IMM(table->n_entries / 2);
>
> for (i = 0; i < table->size / 2; i++) {
> u16 low = get_entry_l3cc(table, 2 * i);
> diff --git a/drivers/gpu/drm/i915/gt/intel_ringbuffer.c b/drivers/gpu/drm/i915/gt/intel_ringbuffer.c
> index f0d60affdba3..e98c2fe727a5 100644
> --- a/drivers/gpu/drm/i915/gt/intel_ringbuffer.c
> +++ b/drivers/gpu/drm/i915/gt/intel_ringbuffer.c
> @@ -1516,12 +1516,13 @@ static int load_pd_dir(struct i915_request *rq,
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(1);
> - *cs++ = i915_mmio_reg_offset(RING_PP_DIR_DCLV(engine->mmio_base));
> + /* Can these not be merged into a single LRI??? */
> + *cs++ = i915_get_lri_cmd(engine, 1);
> + *cs++ = i915_get_lri_reg(engine, RING_PP_DIR_DCLV(engine->mmio_base));
> *cs++ = PP_DIR_DCLV_2G;
>
> - *cs++ = MI_LOAD_REGISTER_IMM(1);
> - *cs++ = i915_mmio_reg_offset(RING_PP_DIR_BASE(engine->mmio_base));
> + *cs++ = i915_get_lri_cmd(engine, 1);
> + *cs++ = i915_get_lri_reg(engine, RING_PP_DIR_BASE(engine->mmio_base));
> *cs++ = ppgtt->pd.base.ggtt_offset << 10;
>
> intel_ring_advance(rq, cs);
> @@ -1589,7 +1590,11 @@ static inline int mi_set_context(struct i915_request *rq, u32 flags)
> if (num_engines) {
> struct intel_engine_cs *signaller;
>
> - *cs++ = MI_LOAD_REGISTER_IMM(num_engines);
> + /*
> + * Must use absolute engine address as the register
> + * write is targeting a different engine.
> + */
> + *cs++ = __MI_LOAD_REGISTER_IMM(num_engines);
> for_each_engine(signaller, i915, id) {
> if (signaller == engine)
> continue;
> @@ -1643,7 +1648,11 @@ static inline int mi_set_context(struct i915_request *rq, u32 flags)
> struct intel_engine_cs *signaller;
> i915_reg_t last_reg = {}; /* keep gcc quiet */
>
> - *cs++ = MI_LOAD_REGISTER_IMM(num_engines);
> + /*
> + * Must use absolute engine address as the register
> + * write is targeting a different engine.
> + */
> + *cs++ = __MI_LOAD_REGISTER_IMM(num_engines);
> for_each_engine(signaller, i915, id) {
> if (signaller == engine)
> continue;
> @@ -1687,9 +1696,9 @@ static int remap_l3(struct i915_request *rq, int slice)
> * here because no other code should access these registers other than
> * at initialization time.
> */
> - *cs++ = MI_LOAD_REGISTER_IMM(GEN7_L3LOG_SIZE/4);
> + *cs++ = i915_get_lri_cmd(rq->engine, GEN7_L3LOG_SIZE/4);
> for (i = 0; i < GEN7_L3LOG_SIZE/4; i++) {
> - *cs++ = i915_mmio_reg_offset(GEN7_L3LOG(slice, i));
> + *cs++ = i915_get_lri_reg(rq->engine, GEN7_L3LOG(slice, i));
> *cs++ = remap_info[i];
> }
> *cs++ = MI_NOOP;
> @@ -2335,3 +2344,23 @@ int intel_ring_submission_init(struct intel_engine_cs *engine)
> intel_engine_cleanup_common(engine);
> return err;
> }
> +
> +u32 i915_get_lri_cmd(const struct intel_engine_cs *engine, u32 word_count)
> +{
> + u32 word;
> +
> + word = __MI_LOAD_REGISTER_IMM(word_count);
> +
> + if (i915_engine_has_relative_lri(engine))
> + word |= MI_LRI_ADD_CS_MMIO_START_GEN11;
> +
> + return word;
> +}
> +
> +u32 i915_get_lri_reg(const struct intel_engine_cs *engine, i915_reg_t reg)
> +{
> + if (!i915_engine_has_relative_lri(engine))
> + return i915_mmio_reg_offset(reg);
> +
> + return i915_mmio_reg_offset(reg) - engine->mmio_base;
> +}
> diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c
> index 43e290306551..d5edc10c860c 100644
> --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c
> +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c
> @@ -625,9 +625,9 @@ int intel_engine_emit_ctx_wa(struct i915_request *rq)
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(wal->count);
> + *cs++ = i915_get_lri_cmd(rq->engine, wal->count);
> for (i = 0, wa = wal->list; i < wal->count; i++, wa++) {
> - *cs++ = i915_mmio_reg_offset(wa->reg);
> + *cs++ = i915_get_lri_reg(rq->engine, wa->reg);
> *cs++ = wa->val;
> }
> *cs++ = MI_NOOP;
> diff --git a/drivers/gpu/drm/i915/gt/selftest_workarounds.c b/drivers/gpu/drm/i915/gt/selftest_workarounds.c
> index 9f7680b9984b..b0513c1de53c 100644
> --- a/drivers/gpu/drm/i915/gt/selftest_workarounds.c
> +++ b/drivers/gpu/drm/i915/gt/selftest_workarounds.c
> @@ -442,6 +442,7 @@ static int check_dirty_whitelist(struct i915_gem_context *ctx,
>
> for (i = 0; i < engine->whitelist.count; i++) {
> u32 reg = i915_mmio_reg_offset(engine->whitelist.list[i].reg);
> + u32 regLRI = i915_get_lri_reg(engine, engine->whitelist.list[i].reg);
> u64 addr = scratch->node.start;
> struct i915_request *rq;
> u32 srm, lrm, rsvd;
> @@ -474,8 +475,8 @@ static int check_dirty_whitelist(struct i915_gem_context *ctx,
> idx = 1;
> for (v = 0; v < ARRAY_SIZE(values); v++) {
> /* LRI garbage */
> - *cs++ = MI_LOAD_REGISTER_IMM(1);
> - *cs++ = reg;
> + *cs++ = i915_get_lri_cmd(engine, 1);
> + *cs++ = regLRI;
> *cs++ = values[v];
>
> /* SRM result */
> @@ -487,8 +488,8 @@ static int check_dirty_whitelist(struct i915_gem_context *ctx,
> }
> for (v = 0; v < ARRAY_SIZE(values); v++) {
> /* LRI garbage */
> - *cs++ = MI_LOAD_REGISTER_IMM(1);
> - *cs++ = reg;
> + *cs++ = i915_get_lri_cmd(engine, 1);
> + *cs++ = regLRI;
> *cs++ = ~values[v];
>
> /* SRM result */
> diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c
> index a27bdd3951f6..3807ce5fe564 100644
> --- a/drivers/gpu/drm/i915/gvt/mmio_context.c
> +++ b/drivers/gpu/drm/i915/gvt/mmio_context.c
> @@ -200,14 +200,14 @@ restore_context_mmio_for_inhibit(struct intel_vgpu *vgpu,
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(count);
> + *cs++ = i915_get_lri_cmd(req->engine, count);
> for (mmio = gvt->engine_mmio_list.mmio;
> i915_mmio_reg_valid(mmio->reg); mmio++) {
> if (mmio->ring_id != ring_id ||
> !mmio->in_context)
> continue;
>
> - *cs++ = i915_mmio_reg_offset(mmio->reg);
> + *cs++ = i915_get_lri_reg(req->engine, mmio->reg);
> *cs++ = vgpu_vreg_t(vgpu, mmio->reg) |
> (mmio->mask << 16);
> gvt_dbg_core("add lri reg pair 0x%x:0x%x in inhibit ctx, vgpu:%d, rind_id:%d\n",
> @@ -235,7 +235,11 @@ restore_render_mocs_control_for_inhibit(struct intel_vgpu *vgpu,
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(GEN9_MOCS_SIZE);
> + /*
> + * GEN9_GFX_MOCS is not engine relative, therefore there is no
> + * need for relative addressing.
> + */
> + *cs++ = __MI_LOAD_REGISTER_IMM(GEN9_MOCS_SIZE);
>
> for (index = 0; index < GEN9_MOCS_SIZE; index++) {
> *cs++ = i915_mmio_reg_offset(GEN9_GFX_MOCS(index));
> @@ -262,7 +266,11 @@ restore_render_mocs_l3cc_for_inhibit(struct intel_vgpu *vgpu,
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(GEN9_MOCS_SIZE / 2);
> + /*
> + * GEN9_LNCFCMOCS is not engine relative, therefore there is no
> + * need for relative addressing.
> + */
> + *cs++ = __MI_LOAD_REGISTER_IMM(GEN9_MOCS_SIZE / 2);
>
> for (index = 0; index < GEN9_MOCS_SIZE / 2; index++) {
> *cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(index));
> diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
> index e9fadcb4d592..fd183e72dace 100644
> --- a/drivers/gpu/drm/i915/i915_cmd_parser.c
> +++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
> @@ -221,7 +221,7 @@ static const struct drm_i915_cmd_descriptor common_cmds[] = {
> CMD( MI_SUSPEND_FLUSH, SMI, F, 1, S ),
> CMD( MI_SEMAPHORE_MBOX, SMI, !F, 0xFF, R ),
> CMD( MI_STORE_DWORD_INDEX, SMI, !F, 0xFF, R ),
> - CMD( MI_LOAD_REGISTER_IMM(1), SMI, !F, 0xFF, W,
> + CMD( __MI_LOAD_REGISTER_IMM(1), SMI, !F, 0xFF, W,
> .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 2 } ),
> CMD( MI_STORE_REGISTER_MEM, SMI, F, 3, W | B,
> .reg = { .offset = 1, .mask = 0x007FFFFC },
> @@ -1183,7 +1183,7 @@ static bool check_cmd(const struct intel_engine_cs *engine,
> return false;
> }
>
> - if (desc->cmd.value == MI_LOAD_REGISTER_IMM(1) &&
> + if (desc->cmd.value == __MI_LOAD_REGISTER_IMM(1) &&
> (offset + 2 > length ||
> (cmd[offset + 1] & reg->mask) != reg->value)) {
> DRM_DEBUG_DRIVER("CMD: Rejected LRI to masked register 0x%08X\n",
> diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
> index 65cefc520e79..98260d8a45be 100644
> --- a/drivers/gpu/drm/i915/i915_gem_context.c
> +++ b/drivers/gpu/drm/i915/i915_gem_context.c
> @@ -1026,11 +1026,11 @@ static int emit_ppgtt_update(struct i915_request *rq, void *data)
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(2);
> + *cs++ = i915_get_lri_cmd(engine, 2);
>
> - *cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, 0));
> + *cs++ = i915_get_lri_reg(engine, GEN8_RING_PDP_UDW(base, 0));
> *cs++ = upper_32_bits(pd_daddr);
> - *cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(base, 0));
> + *cs++ = i915_get_lri_reg(engine, GEN8_RING_PDP_LDW(base, 0));
> *cs++ = lower_32_bits(pd_daddr);
>
> *cs++ = MI_NOOP;
> @@ -1040,13 +1040,13 @@ static int emit_ppgtt_update(struct i915_request *rq, void *data)
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(2 * GEN8_3LVL_PDPES);
> + *cs++ = i915_get_lri_cmd(engine, 2 * GEN8_3LVL_PDPES);
> for (i = GEN8_3LVL_PDPES; i--; ) {
> const dma_addr_t pd_daddr = i915_page_dir_dma_addr(ppgtt, i);
>
> - *cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, i));
> + *cs++ = i915_get_lri_reg(engine, GEN8_RING_PDP_UDW(base, i));
> *cs++ = upper_32_bits(pd_daddr);
> - *cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(base, i));
> + *cs++ = i915_get_lri_reg(engine, GEN8_RING_PDP_LDW(base, i));
> *cs++ = lower_32_bits(pd_daddr);
> }
> *cs++ = MI_NOOP;
> diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> index 679f7c1561ba..ac5b06d2ffdc 100644
> --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> @@ -1963,7 +1963,8 @@ static int i915_reset_gen7_sol_offsets(struct i915_request *rq)
> if (IS_ERR(cs))
> return PTR_ERR(cs);
>
> - *cs++ = MI_LOAD_REGISTER_IMM(4);
> + /* Gen7 only so no need to support relative offsets */
> + *cs++ = __MI_LOAD_REGISTER_IMM(4);
> for (i = 0; i < 4; i++) {
> *cs++ = i915_mmio_reg_offset(GEN7_SO_WRITE_OFFSET(i));
> *cs++ = 0;
> diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
> index c4995d5a16d2..86facbccdb02 100644
> --- a/drivers/gpu/drm/i915/i915_perf.c
> +++ b/drivers/gpu/drm/i915/i915_perf.c
> @@ -1636,7 +1636,8 @@ static void hsw_disable_metric_set(struct drm_i915_private *dev_priv)
> * in the case that the OA unit has been disabled.
> */
> static void
> -gen8_update_reg_state_unlocked(struct intel_context *ce,
> +gen8_update_reg_state_unlocked(struct intel_engine_cs *engine,
> + struct intel_context *ce,
ce->engine?
> u32 *reg_state,
> const struct i915_oa_config *oa_config)
> {
> @@ -1655,7 +1656,12 @@ gen8_update_reg_state_unlocked(struct intel_context *ce,
> };
> int i;
>
> - CTX_REG(reg_state, ctx_oactxctrl, GEN8_OACTXCONTROL,
> + /*
> + * NB: The LRI instruction is generated by the hardware.
> + * Should we read it in and assert that the offset flag is set?
> + */
> +
> + CTX_REG(engine, reg_state, ctx_oactxctrl, GEN8_OACTXCONTROL,
> (i915->perf.oa.period_exponent << GEN8_OA_TIMER_PERIOD_SHIFT) |
> (i915->perf.oa.periodic ? GEN8_OA_TIMER_ENABLE : 0) |
> GEN8_OA_COUNTER_RESUME);
> @@ -1682,10 +1688,10 @@ gen8_update_reg_state_unlocked(struct intel_context *ce,
> }
> }
>
> - CTX_REG(reg_state, state_offset, flex_regs[i], value);
> + CTX_REG(engine, reg_state, state_offset, flex_regs[i], value);
> }
>
> - CTX_REG(reg_state,
> + CTX_REG(engine, reg_state,
> CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE,
> intel_sseu_make_rpcs(i915, &ce->sseu));
> }
> @@ -1770,7 +1776,7 @@ static int gen8_configure_all_contexts(struct drm_i915_private *dev_priv,
> ce->state->obj->mm.dirty = true;
> regs += LRC_STATE_PN * PAGE_SIZE / sizeof(*regs);
>
> - gen8_update_reg_state_unlocked(ce, regs, oa_config);
> + gen8_update_reg_state_unlocked(dev_priv->engine[RCS0], ce, regs, oa_config);
>
> i915_gem_object_unpin_map(ce->state->obj);
> }
> @@ -2166,7 +2172,8 @@ void i915_oa_init_reg_state(struct intel_engine_cs *engine,
>
> stream = engine->i915->perf.oa.exclusive_stream;
> if (stream)
> - gen8_update_reg_state_unlocked(ce, regs, stream->oa_config);
> + gen8_update_reg_state_unlocked(engine, ce, regs,
> + stream->oa_config);
> }
>
> /**
>
Regards,
Tvrtko
More information about the Intel-gfx
mailing list