Commit 0ed4b3c2 authored by Lucas De Marchi's avatar Lucas De Marchi
Browse files

drm/xe/lrc: Prepare WA BB setup for more users



The post context restore (WA BB) is a mechanism in HW that may be used
for things other than the utilization setup. Create a new function
called setup_wa_bb() that wraps any function writing useful commands in
the buffer.

Reviewed-by: default avatarUmesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: default avatarTvrtko Ursulin <tvrtko.ursulin@igalia.com>
Link: https://lore.kernel.org/r/20250604-wa-bb-fix-v1-2-0dfc5dafcef0@intel.com


Signed-off-by: default avatarLucas De Marchi <lucas.demarchi@intel.com>
parent ef48715b
Loading
Loading
Loading
Loading
+55 −14
Original line number Diff line number Diff line
@@ -914,9 +914,8 @@ static void xe_lrc_finish(struct xe_lrc *lrc)
}

/*
 * xe_lrc_setup_utilization() - Setup wa bb to assist in calculating active
 * context run ticks.
 * @lrc: Pointer to the lrc.
 * wa_bb_setup_utilization() - Write commands to wa bb to assist
 * in calculating active context run ticks.
 *
 * Context Timestamp (CTX_TIMESTAMP) in the LRC accumulates the run ticks of the
 * context, but only gets updated when the context switches out. In order to
@@ -941,18 +940,13 @@ static void xe_lrc_finish(struct xe_lrc *lrc)
 * store it in the PPHSWP.
 */
#define CONTEXT_ACTIVE 1ULL
static int xe_lrc_setup_utilization(struct xe_lrc *lrc)
static ssize_t wa_bb_setup_utilization(struct xe_lrc *lrc, struct xe_hw_engine *hwe,
				       u32 *batch, size_t max_len)
{
	u32 *cmd, *buf = NULL;
	u32 *cmd = batch;

	if (lrc->bb_per_ctx_bo->vmap.is_iomem) {
		buf = kmalloc(lrc->bb_per_ctx_bo->size, GFP_KERNEL);
		if (!buf)
			return -ENOMEM;
		cmd = buf;
	} else {
		cmd = lrc->bb_per_ctx_bo->vmap.vaddr;
	}
	if (xe_gt_WARN_ON(lrc->gt, max_len < 12))
		return -ENOSPC;

	*cmd++ = MI_STORE_REGISTER_MEM | MI_SRM_USE_GGTT | MI_SRM_ADD_CS_OFFSET;
	*cmd++ = ENGINE_ID(0).addr;
@@ -971,6 +965,49 @@ static int xe_lrc_setup_utilization(struct xe_lrc *lrc)
		*cmd++ = upper_32_bits(CONTEXT_ACTIVE);
	}

	return cmd - batch;
}

struct wa_bb_setup {
	ssize_t (*setup)(struct xe_lrc *lrc, struct xe_hw_engine *hwe,
			 u32 *batch, size_t max_size);
};

static int setup_wa_bb(struct xe_lrc *lrc, struct xe_hw_engine *hwe)
{
	const size_t max_size = lrc->bb_per_ctx_bo->size;
	static const struct wa_bb_setup funcs[] = {
		{ .setup = wa_bb_setup_utilization },
	};
	ssize_t remain;
	u32 *cmd, *buf = NULL;

	if (lrc->bb_per_ctx_bo->vmap.is_iomem) {
		buf = kmalloc(max_size, GFP_KERNEL);
		if (!buf)
			return -ENOMEM;
		cmd = buf;
	} else {
		cmd = lrc->bb_per_ctx_bo->vmap.vaddr;
	}

	remain = max_size / sizeof(*cmd);

	for (size_t i = 0; i < ARRAY_SIZE(funcs); i++) {
		ssize_t len = funcs[i].setup(lrc, hwe, cmd, remain);

		remain -= len;

		/*
		 * There should always be at least 1 additional dword for
		 * the end marker
		 */
		if (len < 0 || xe_gt_WARN_ON(lrc->gt, remain < 1))
			goto fail;

		cmd += len;
	}

	*cmd++ = MI_BATCH_BUFFER_END;

	if (buf) {
@@ -983,6 +1020,10 @@ static int xe_lrc_setup_utilization(struct xe_lrc *lrc)
			     xe_bo_ggtt_addr(lrc->bb_per_ctx_bo) | 1);

	return 0;

fail:
	kfree(buf);
	return -ENOSPC;
}

#define PVC_CTX_ASID		(0x2e + 1)
@@ -1139,7 +1180,7 @@ static int xe_lrc_init(struct xe_lrc *lrc, struct xe_hw_engine *hwe,
	map = __xe_lrc_start_seqno_map(lrc);
	xe_map_write32(lrc_to_xe(lrc), &map, lrc->fence_ctx.next_seqno - 1);

	err = xe_lrc_setup_utilization(lrc);
	err = setup_wa_bb(lrc, hwe);
	if (err)
		goto err_lrc_finish;