Lines Matching refs:shard

53  * Compute per-LLC shard layout. Each shard holds at most @shard_size cids, and
284 				 * Advance to a new shard when either core or
286 				 * shard sizes under uneven SMT. Never start an
287 				 * empty shard.
332 	 * SCX_CID_SHARD_MAX_CPUS) cids so that every cid has a valid shard
374 	 * shard, nr_cids is the count.
451  * scx_bpf_cid_override - Install an explicit cpu->cid mapping with shard info
454  * @shard_start__arena: array of first-cid-of-each-shard, one entry per shard
459  * topology-probed cid mapping and shard layout with caller-provided ones. Each
460  * possible cpu must map to a unique cid in [0, num_possible_cpus()). The shard
462  * num_possible_cpus(). The last shard extends to num_possible_cpus() and no
463  * shard may span more than SCX_CID_SHARD_MAX_CPUS cids. Topo info
464  * (core/LLC/node) is cleared and the shard layout is set from the input. On
546 			scx_error(sch, "scx_bpf_cid_override: shard[%d] span %d exceeds max %d",
553 		scx_error(sch, "scx_bpf_cid_override: shard[%d] span %d exceeds max %d",
579 	 * Derive shard_node[] by majority count: an overridden shard may
596 	 * Invalidate stale topo info and install shard layout from
1047  * scx_cmask_ref_shard - Read one shard from @ref into @out
1049  * @shard_idx: target shard, in [@ref->shard_first, @ref->shard_end)
1050  * @out: output cmask whose @out->alloc_words must hold the shard
1055  * shard.
1060 	const struct scx_cid_shard *shard =
1062 	u32 shard_base = shard->base_cid;
1063 	u32 shard_end = shard_base + shard->nr_cids;
1078 		scx_error(ref->sch, "scx_cmask_ref_shard: out alloc_words=%u < %u for shard %d",