Commit f2bd5188 authored by Yalong Wang's avatar Yalong Wang Committed by Tomasz Zawadzki
Browse files

lib/blob: do not release eP when it is all zero 



Releasing an extpage with all-zero entries is unnecessary:
1. It increases code complexity and may race with cluster allocation.
2. The space usage of extpages is negligible: at most 1/512k,
   while blobstore already reserves 1/1k for mdpages.
3. extpage will be released when bdev_lvol_delete.

This change simplifies metadata management without affecting space usage.

Change-Id: I8f6a8fcf8bf35a9d2f04346f7ee4932814c1f666
Signed-off-by: default avatarYalong Wang <yalong9@staff.sina.com.cn>
Reviewed-on: https://review.spdk.io/c/spdk/spdk/+/26682


Community-CI: Mellanox Build Bot
Reviewed-by: default avatarGangCao <gang.cao@intel.com>
Reviewed-by: default avatarTomasz Zawadzki <tomasz@tzawadzki.com>
Reviewed-by: default avatarJim Harris <jim.harris@nvidia.com>
Tested-by: default avatarSPDK Automated Test System <spdkbot@gmail.com>
parent 7c0720d1

lib/blob/blobstore.c

+2 −37
Original line number Diff line number Diff line
@@ -8876,19 +8876,6 @@ blob_free_cluster_update_ep_cb(void *arg, int bserrno) 
	blob_sync_md(ctx->blob, blob_free_cluster_msg_cb, ctx);

}



static void

blob_free_cluster_free_ep_cb(void *arg, int bserrno)

{

	struct spdk_blob_cluster_op_ctx *ctx = arg;



	spdk_spin_lock(&ctx->blob->bs->used_lock);

	assert(spdk_bit_array_get(ctx->blob->bs->used_md_pages, ctx->extent_page) == true);

	bs_release_md_page(ctx->blob->bs, ctx->extent_page);

	spdk_spin_unlock(&ctx->blob->bs->used_lock);

	ctx->blob->state = SPDK_BLOB_STATE_DIRTY;

	blob_sync_md(ctx->blob, blob_free_cluster_msg_cb, ctx);

}



static void

blob_persist_extent_page_cpl(spdk_bs_sequence_t *seq, void *cb_arg, int bserrno)

{
@@ -9029,9 +9016,6 @@ blob_free_cluster_msg(void *arg) 
{

	struct spdk_blob_cluster_op_ctx *ctx = arg;

	uint32_t *extent_page;

	uint32_t start_cluster_idx;

	bool free_extent_page = true;

	size_t i;



	ctx->cluster = bs_lba_to_cluster(ctx->blob->bs, ctx->blob->active.clusters[ctx->cluster_num]);
@@ -9061,28 +9045,9 @@ blob_free_cluster_msg(void *arg) 
	/* There shouldn't be parallel release operations on same cluster */

	assert(*extent_page == ctx->extent_page);



	start_cluster_idx = (ctx->cluster_num / SPDK_EXTENTS_PER_EP) * SPDK_EXTENTS_PER_EP;

	for (i = 0; i < SPDK_EXTENTS_PER_EP; ++i) {

		if (spdk_unlikely(start_cluster_idx + i >= ctx->blob->active.num_clusters)) {

			break;

		}

		if (ctx->blob->active.clusters[start_cluster_idx + i] != 0) {

			free_extent_page = false;

			break;

		}

	}



	if (free_extent_page) {

		assert(ctx->extent_page != 0);

		assert(spdk_bit_array_get(ctx->blob->bs->used_md_pages, ctx->extent_page) == true);

		ctx->blob->active.extent_pages[bs_cluster_to_extent_table_id(ctx->cluster_num)] = 0;

		blob_write_extent_page(ctx->blob, ctx->extent_page, ctx->cluster_num, ctx->page,

				       blob_free_cluster_free_ep_cb, ctx);

	} else {

	blob_write_extent_page(ctx->blob, *extent_page, ctx->cluster_num, ctx->page,

			       blob_free_cluster_update_ep_cb, ctx);

}

}





static void