[Intel-gfx] [PATCH v10] drm/i915: Squash repeated awaits on the same fence

[Tvrtko Ursulin]

On 28/04/2017 08:41, Chris Wilson wrote:

[snip]

> +static int igt_sync(void *arg)
> +{
> +	const struct {
> +		const char *name;
> +		u32 seqno;
> +		bool expected;
> +		bool set;
> +	} pass[] = {
> +		{ "unset", 0, false, false },
> +		{ "new", 0, false, true },
> +		{ "0a", 0, true, true },
> +		{ "1a", 1, false, true },
> +		{ "1b", 1, true, true },
> +		{ "0b", 0, true, false },
> +		{ "2a", 2, false, true },
> +		{ "4", 4, false, true },
> +		{ "INT_MAX", INT_MAX, false, true },
> +		{ "INT_MAX-1", INT_MAX-1, true, false },
> +		{ "INT_MAX+1", (u32)INT_MAX+1, false, true },
> +		{ "INT_MAX", INT_MAX, true, false },
> +		{ "UINT_MAX", UINT_MAX, false, true },
> +		{ "wrap", 0, false, true },
> +		{ "unwrap", UINT_MAX, true, false },
> +		{},
> +	}, *p;
> +	struct intel_timeline *tl;
> +	int order, offset;
> +	int ret;
> +
> +	tl = mock_timeline(0);
> +	if (!tl)
> +		return -ENOMEM;
> +
> +	for (p = pass; p->name; p++) {
> +		for (order = 1; order < 64; order++) {
> +			for (offset = -1; offset <= (order > 1); offset++) {
> +				u64 ctx = BIT_ULL(order) + offset;
> +
> +				if (__intel_timeline_sync_is_later
> +				    (tl, ctx, p->seqno) != p->expected) {
> +					pr_err("1: %s(ctx=%llu, seqno=%u) expected passed %s but failed\n",
> +					       p->name, ctx, p->seqno, yesno(p->expected));
> +					ret = -EINVAL;
> +					goto out;
> +				}
> +
> +				if (p->set) {
> +					ret = __intel_timeline_sync_set(tl, ctx, p->seqno);
> +					if (ret)
> +						goto out;
> +				}
> +			}
> +		}
> +	}

I think verification that the tree height matches the expectation, and 
also total number of nodes, is required.

If too complicated in the current structure you could add a simpler 
iteration over another table of steps, which would include the ctx id 
and seqno in it, and expected height/total nodes after each step.

Regards,

Tvrtko