ivampiresp/Spdk
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

blob: remove short-circuiting path for blob_freeze

Browse Source 
If blob_freeze_io() is called twice in a row,
and the second time occurs before the for_each_channel
for the first completes, the second caller will
receive its callback too soon.

Instead just simplify the whole process, always do
the for_each_channel and don't try to optimize it
at all.  These are infrequent operations - correctness
and simplicity are in order.

A few additional changes:

1) Make same changes for unfreeze path.
2) Add blob_verify_md_op() calls, just to be sure
   these are only called from md_thread.  This was
   already checked in calling functions, but as these
   functions get called from new code paths (i.e.
   esnap clones) it can't hurt to add additional
   checks.
3) Add unit test that failed with original code, but
   passes with this patch.

Fixes issue #2935.

Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ibefba554547ddf3e26aaabfa4288c8073d3c04ff
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17148
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Mike Gerdts <mgerdts@nvidia.com>
Community-CI: Mellanox Build Bot
This commit is contained in:
Jim Harris 2023-03-10 16:49:18 +00:00 committed by Konrad Sztyber
parent d4d57898cb
commit 037c8b01a1
2 changed files with 98 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
lib/blob/blobstore.c
View File

@ -388,6 +388,8 @@ blob_freeze_io(struct spdk_blob *blob, spdk_blob_op_complete cb_fn, void *cb_arg
{
struct freeze_io_ctx *ctx;
blob_verify_md_op(blob);
ctx = calloc(1, sizeof(*ctx));
if (!ctx) {
cb_fn(cb_arg, -ENOMEM);
@ -402,12 +404,7 @@ blob_freeze_io(struct spdk_blob *blob, spdk_blob_op_complete cb_fn, void *cb_arg
/* Freeze I/O on blob */
blob->frozen_refcnt++;
if (blob->frozen_refcnt == 1) {
spdk_for_each_channel(blob->bs, blob_io_sync, ctx, blob_io_cpl);
} else {
cb_fn(cb_arg, 0);
free(ctx);
}
spdk_for_each_channel(blob->bs, blob_io_sync, ctx, blob_io_cpl);
}
static void
@ -415,6 +412,8 @@ blob_unfreeze_io(struct spdk_blob *blob, spdk_blob_op_complete cb_fn, void *cb_a
{
struct freeze_io_ctx *ctx;
blob_verify_md_op(blob);
ctx = calloc(1, sizeof(*ctx));
if (!ctx) {
cb_fn(cb_arg, -ENOMEM);
@ -430,12 +429,7 @@ blob_unfreeze_io(struct spdk_blob *blob, spdk_blob_op_complete cb_fn, void *cb_a
blob->frozen_refcnt--;
if (blob->frozen_refcnt == 0) {
spdk_for_each_channel(blob->bs, blob_execute_queued_io, ctx, blob_io_cpl);
} else {
cb_fn(cb_arg, 0);
free(ctx);
}
spdk_for_each_channel(blob->bs, blob_execute_queued_io, ctx, blob_io_cpl);
}
static int

92
test/unit/lib/blob/blob.c/blob_ut.c
View File

@ -7524,6 +7524,97 @@ blob_esnap_create(void)
g_blob = NULL;
}
static void
freeze_done(void *cb_arg, int bserrno)
{
uint32_t *freeze_cnt = cb_arg;
CU_ASSERT(bserrno == 0);
(*freeze_cnt)++;
}
static void
unfreeze_done(void *cb_arg, int bserrno)
{
uint32_t *unfreeze_cnt = cb_arg;
CU_ASSERT(bserrno == 0);
(*unfreeze_cnt)++;
}
static void
blob_nested_freezes(void)
{
struct spdk_blob_store *bs = g_bs;
struct spdk_blob *blob;
struct spdk_io_channel *channel[2];
struct spdk_blob_opts opts;
uint32_t freeze_cnt, unfreeze_cnt;
int i;
for (i = 0; i < 2; i++) {
set_thread(i);
channel[i] = spdk_bs_alloc_io_channel(bs);
SPDK_CU_ASSERT_FATAL(channel[i] != NULL);
}
set_thread(0);
ut_spdk_blob_opts_init(&opts);
blob = ut_blob_create_and_open(bs, &opts);
/* First just test a single freeze/unfreeze. */
freeze_cnt = 0;
unfreeze_cnt = 0;
CU_ASSERT(blob->frozen_refcnt == 0);
blob_freeze_io(blob, freeze_done, &freeze_cnt);
CU_ASSERT(blob->frozen_refcnt == 1);
CU_ASSERT(freeze_cnt == 0);
poll_threads();
CU_ASSERT(freeze_cnt == 1);
blob_unfreeze_io(blob, unfreeze_done, &unfreeze_cnt);
CU_ASSERT(blob->frozen_refcnt == 0);
CU_ASSERT(unfreeze_cnt == 0);
poll_threads();
CU_ASSERT(unfreeze_cnt == 1);
/* Now nest multiple freeze/unfreeze operations. We should
* expect a callback for each operation, but only after
* the threads have been polled to ensure a for_each_channel()
* was executed.
*/
freeze_cnt = 0;
unfreeze_cnt = 0;
CU_ASSERT(blob->frozen_refcnt == 0);
blob_freeze_io(blob, freeze_done, &freeze_cnt);
CU_ASSERT(blob->frozen_refcnt == 1);
CU_ASSERT(freeze_cnt == 0);
blob_freeze_io(blob, freeze_done, &freeze_cnt);
CU_ASSERT(blob->frozen_refcnt == 2);
CU_ASSERT(freeze_cnt == 0);
poll_threads();
CU_ASSERT(freeze_cnt == 2);
blob_unfreeze_io(blob, unfreeze_done, &unfreeze_cnt);
CU_ASSERT(blob->frozen_refcnt == 1);
CU_ASSERT(unfreeze_cnt == 0);
blob_unfreeze_io(blob, unfreeze_done, &unfreeze_cnt);
CU_ASSERT(blob->frozen_refcnt == 0);
CU_ASSERT(unfreeze_cnt == 0);
poll_threads();
CU_ASSERT(unfreeze_cnt == 2);
for (i = 0; i < 2; i++) {
set_thread(i);
spdk_bs_free_io_channel(channel[i]);
}
set_thread(0);
ut_blob_close_and_delete(bs, blob);
poll_threads();
g_blob = NULL;
g_blobid = 0;
}
static void
suite_bs_setup(void)
{
@ -7721,6 +7812,7 @@ main(int argc, char **argv)
CU_ADD_TEST(suite_bs, blob_decouple_snapshot);
CU_ADD_TEST(suite_bs, blob_seek_io_unit);
CU_ADD_TEST(suite_esnap_bs, blob_esnap_create);
CU_ADD_TEST(suite_bs, blob_nested_freezes);
allocate_threads(2);
set_thread(0);