Commit b5f96b0e authored by Tomasz Zawadzki's avatar Tomasz Zawadzki Committed by Jim Harris
Browse files

ut/blobstore: write and allocate clusters from multiple threads 



This patch verifies that functionality of only releasing cluster
that was originaly claimed in race thread condition.

Change-Id: I4d60ffdb4fd6612d61a761835aae2fd1bf81b7fd
Signed-off-by: default avatarTomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/452626


Tested-by: default avatarSPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: default avatarBen Walker <benjamin.walker@intel.com>
Reviewed-by: default avatarJim Harris <james.r.harris@intel.com>
parent fde382d1

test/unit/lib/blob/blob.c/blob_ut.c

+17 −5
Original line number Diff line number Diff line
@@ -4413,7 +4413,7 @@ blob_thin_prov_rw(void) 
	struct spdk_blob_store *bs;

	struct spdk_bs_dev *dev;

	struct spdk_blob *blob;

	struct spdk_io_channel *channel;

	struct spdk_io_channel *channel, *channel_thread1;

	struct spdk_blob_opts opts;

	spdk_blob_id blobid;

	uint64_t free_clusters;
@@ -4478,14 +4478,25 @@ blob_thin_prov_rw(void) 
	write_bytes = g_dev_write_bytes;

	read_bytes = g_dev_read_bytes;



	/* Perform write on thread 1. That will allocate cluster on thread 0 via send_msg */

	set_thread(1);

	channel_thread1 = spdk_bs_alloc_io_channel(bs);

	CU_ASSERT(channel_thread1 != NULL);

	memset(payload_write, 0xE5, sizeof(payload_write));

	spdk_blob_io_write(blob, channel_thread1, payload_write, 4, 10, blob_op_complete, NULL);

	CU_ASSERT(free_clusters - 1 == spdk_bs_free_cluster_count(bs));

	/* Perform write on thread 0. That will try to allocate cluster,

	 * but fail due to another thread issuing the cluster allocation first. */

	set_thread(0);

	memset(payload_write, 0xE5, sizeof(payload_write));

	spdk_blob_io_write(blob, channel, payload_write, 4, 10, blob_op_complete, NULL);

	CU_ASSERT(free_clusters - 2 == spdk_bs_free_cluster_count(bs));

	poll_threads();

	CU_ASSERT(g_bserrno == 0);

	CU_ASSERT(free_clusters != spdk_bs_free_cluster_count(bs));

	/* For thin-provisioned blob we need to write 10 pages plus one page metadata and

	CU_ASSERT(free_clusters - 1 == spdk_bs_free_cluster_count(bs));

	/* For thin-provisioned blob we need to write 20 pages plus one page metadata and

	 * read 0 bytes */

	CU_ASSERT(g_dev_write_bytes - write_bytes == page_size * 11);

	CU_ASSERT(g_dev_write_bytes - write_bytes == page_size * 21);

	CU_ASSERT(g_dev_read_bytes - read_bytes == 0);



	spdk_blob_io_read(blob, channel, payload_read, 4, 10, blob_op_complete, NULL);
@@ -4502,6 +4513,7 @@ blob_thin_prov_rw(void) 
	CU_ASSERT(g_bserrno == 0);

	CU_ASSERT(free_clusters == spdk_bs_free_cluster_count(bs));



	spdk_bs_free_io_channel(channel_thread1);

	spdk_bs_free_io_channel(channel);

	poll_threads();
@@ -6661,7 +6673,7 @@ int main(int argc, char **argv) 
		return CU_get_error();

	}



	allocate_threads(1);

	allocate_threads(2);

	set_thread(0);



	g_dev_buffer = calloc(1, DEV_BUFFER_SIZE);