The purpose of this patch is to make spdk_bs_opts
for compatiblity issue.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I26d2a6bc644feede64d48890c7903f224b1fc306
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5681
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Search for md_page to be used as extent page started
from 0, which is completely valid md_page.
This page can be free when for example blob with id 0
was deleted and some other requested a new page for extent.
There are already existing blobs that have extents
pointing to 0, which means unallocated.
Unfortunetly it means 0 can never mean md page 0.
If that already occured for someone, this extent page was already
lost during blob/bs reload and nothing can be done.
With this in mind following assumptions are made for extent pages:
- 0 means unallocated extent page
- UINT32_MAX means we ran out of md pages, and should not be persisted
- [NEW] extent page can never occupy md page 0
That last one is new addition in this patch.
bs_allocate_cluster will now always try to find md page from 1 or higher.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ia17ce5bbca2fab4fb4487e4e263f3a0aa120bf17
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4314
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
When loading a blob and parsing its metadata,
the array of pages was not updated.
Serialization was unaffected, since the current pages
array is unused there.
Behavior was working correctly for first page,
but did not for any blob with more than one page.
Unfortunetly blob_persist_zero_pages() never zeroed out
the pages, neither blob_persist_zero_pages_cpl()
released the md pages.
Resulting in md pages being claimed even after blobs
deletion.
This patch now fills out the active pages array with
the appropriate page numbers from metadata.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I6ff1f4fe95684119d283c2471fdbbea464da8151
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4504
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
We still need to be able to explicitly set specific
bits in the cluster array during initialization and
loading (especially recovery), so we use a bit_array
during load, and then convert it to a bit_pool just
before calling the user's cmopletion callback.
This gives a roughly 300% improvement over baseline
on a benchmark which does continuous resize operations.
The benefit is primarily from saving the lowest free
bit rather than having to always start at bit 0. We
may be able to further improve this by saving extents
in the bit pool as well, although after this patch,
the benchmark shows other hot spots different from the
bit search.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Idb1d75d8348bc50560b1f42d49dbe4d79d024619
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3975
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This test repeatedly creates and deletes blobs to
confirm that the used_md_pages and used_clusters
bit arrays are managed properly. This test exposed
the bug reported in #1590 which has now been fixed.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-ID: If5d7b7f03cc2fb9daac928573c73430b74ea9a51
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/4038
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
We only create one spdk_blob object for a given blob, and just
increase the ref_count if it is opened multiple times. bs_open_blob
would do the lookup for existing opened blobs.
But if the blob is opened again, before the previous open operation
has completed, we would end up with two spdk_blob objects for the same
blob.
Solution is to do another lookup when the open operation completes.
If we find the blob, free the one we just finished opening and return
the existing one instead.
Also added unit test that failed on the existing code but passes now
with this patch.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reported-by: Mike Cui
Change-Id: I00c3a913b413deddf06f0b63f7a669efb2b5658f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3855
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Blobid and md_page is claimed as first step of blob creation.
If blob creation failed, both should returned to be used by
other blobs.
This caused multiple reports of:
"Metadata page 1 crc mismatch"
when loading blobstore due to md_pages not actually containing
the written out md pages.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I495452c578d879f749281cebf8975eb2c1c7f79a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/2057
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
super->clean value signifies if blobstore was unloaded
cleanly.
If it was not, then during bs_load the _spdk_bs_recover()
procedure if called.
Meanwhile bs->clean is always set to 1 after load, causing very first
blob_persist to also re-write super block with the super->clean
set to 0. To signify that md has changed and possibly trigger
the recovery if clean bs unload does not occur.
When the re-write of super block succeeds the bs->clean is set to 0,
because further re-writes of super block are not needed on next
blob persist.
This patch resolves issue when:
1) reading super block fails - execution should backoff, to prevent
writing an empty buffer as super block !
2) writing super->clean = 0 to the super block fails - execution
again should fail, and bs->clean should not be set to 0. It will
cause next persist to attempt re-write again.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ia07cc5c6c107310059b50886edb7283c176b9169
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1164
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This change verifies that not only it is possible to open
the snapshot, but also the spdk_bs_create_snapshot() reported
success.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ifb7fcc91802c838710f1f9be41090057268a5900
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1262
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
blob_create_snapshot_power_failure() test always attempted
snapshot creation until the snapshot did open.
Which meant that on bs reload after dirty shutdown,
the original blob might be in unrecoverable state.
UT did not verify create call beyond that point.
Next patch will stop after snapshot is possible to be opened,
AND the spdk_bs_create_snapshot() returned success.
This patch makes it easier to recreate same starting
conditions after snapshot did not open.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I1901b29a8319b2203a855e7879821ed3850f4917
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1261
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
In general it is not possible to delete snapshot when
there are clones on top of it.
There is special case when there is just a single clone
on top that snapshot.
In such case the clone is 'merged' with snapshot.
Unallocated clusters in clone, are filled with the ones
in snapshot (if allocated there).
Similar behavior should have occurred for extent pages.
This patch adds the implementation for moving EP from
snapshot to clone along with UT.
The UT exposes the issue by allowing delete_blob
to proceed beyond just unrecoverable snapshot blob.
Fixes#1291
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ib2824c5737021f8e8d9b533a4cd245c12e6fe9fa
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1163
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
blob_delete_snapshot_power_failure() test always attempted
snapshot deletion until the snapshot did not open.
Which meant that on bs reload after dirty shutdown,
the snapshot was in unrecoverable state.
UT did not verify delete call beyond that point.
Next patch will stop after snapshot is unrecoverable,
AND the spdk_bs_delete_blob() returned success.
This patch makes it easier to recreate same starting
conditions after snapshot did not load.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I7e14e8dadaaef4aa3287ebd376465253466b2362
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1162
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Added new setup/cleanup functions for a suite of
unit tests that use a particular blob through out the
tests.
This simplifies the contents of particular UTs,
and brings forward focus of each UT.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ib7d69b90fcbe8e6a0ba303475fcbaeb90fa110a7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/945
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Some of the unit tests were only closing a blob,
without deleting it at the end of unit tests.
This is valid behaviour, but to later simplify
the setup/cleanup - a delete blob was added.
This change was made to unit tests where,
it was not focus of particular UT.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I7e338ba6fcb63af26658e8968971a6d56f70807b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1203
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Operation of closing and deleting a blob is common
throughout the tests.
This patch creates helper function ut_blob_close_and_delete(),
to do exactly that with the right asserts.
Throughout the UT this function replaces all instances where
blob is closed and deleted, when it is not the subject of
particular unit test (focus is on other functionality).
It will allow to later use it in setup/cleanup functions
for UT suite - similar to bs init.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ic56594da8cd4c442b62890d62961b44771059679
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1202
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Operation of creating a blob (with or w/o opts),
then opening it through out the tests is quite common.
This patch creates helper function ut_blob_create_and_open(),
to do exactly that with the right asserts.
Throughout the UT this function replaces all instances where
blob is created and opened, when it is not the subject of
particular unit test (focus is on other functionality).
It will allow to later use it in setup/cleanup functions
for UT suite - similar to bs init.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ia541b855d154503f824bfa3909bcfeac649c5853
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1201
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Always call API on blobid that was received from blob
create, instead of relying on global one.
Where necessary added declaration and assignment.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I3529b4d5e86c1301bb64268029a57873f9f600cb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/944
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Always call API on pointer that was received from blob
open, instead of relying on global one.
Where necessary added declaration and assignment.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I15531111201d071c471f5be79ce18ff508631677
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/943
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
A lot of unit tests require only creation of a blobstore
then unload it at the end.
This code can be made common and invoked automatically by
CUnit with CU_add_suite_with_setup_and_teardown().
Those functions are called before and after each test in suite.
This should simplify the unit tests by removing the common
code.
Following patches will add further suites with different
setup/cleanup to reduce common code even more.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I984cde2381f4461e188e06050355024a88ebfca9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/932
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Next patch in series adds setup/cleanup functions
that use g_bs for its operation.
Particular unit tests should not leave the g_bs
unassigned, if cleanup is supposed to unload it.
It applies to helper functions to reload/dirty_load.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I6e022b6c762981c2c70ebf31f4981a17f54f3591
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1182
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Some of the unit tests initalized bs opts,
event if they were left unmodified.
There is no need to perform this operation,
as spdk_bs_init() can accept NULL as bs_opts parameter
to just initialize default values.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Iacea1fa9d53796b8c359b611092cc63c6bce4094
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1181
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This patch adds better check if the xattr was removed.
It will identify potential bugs when first md sync
takes precedence over the second md sync.
As per suggestion in
https://review.spdk.io/gerrit/c/spdk/spdk/+/774/6/test/unit/lib/blob/blob.c/blob_ut.c#7655
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I463f2455614ef11cc1512dc2fb1972ba1f024337
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1109
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Operation of dirty shutdown of blobstore and reloading
it occurs often enough in UT to provide a common function
to perform it.
Added ut_bs_dirty_load() to facilitate this.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I0079dfabd64eaec6495db02fe200be1d6116f0c7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1092
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Operation of unloading and loading blobstore
occurs often enough in UT to provide a common function
to perform it.
Added ut_bs_reload() to facilitate this.
Couple occurences in this patch actually fix
test cases where mistakenly opts were not passed.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ib46ed395cd134feaa540e00a334ae861872b3ef4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1091
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Couple unit tests were still refering to g_bs pointer,
rather than the one local to particular UT.
This patch fixes this.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: If343eb561a48c4547d499139d76e1db6b2d7f3bf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1090
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Always call API on pointer that was received from blobstore
initialization or load, instead of relying on global one.
Where necessary added declaration, assignment and asserts.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ic4198ac984c13a8b77428deaa32ce9864e6750c1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/942
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
CUnit provides a helper macro CU_ADD_TEST() that
simplifies using CU_add_test() function.
Test name no longer needs to be provided,
instead test is named after test function name.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I9f5d9e0d8e72ab36b2a5c04e33c336e6c0a1fa6a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/931
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Failure to prepare CUnit for tests is always a fatal error,
and application exits.
Most test managment functions in CUnit set some CU_ErrorCode,
that can be retrived by return code from a function or CU_get_error().
Thus usual structure of UT in SPDK checked every output from every
test managment function.
There is a helper function CU_set_error_action() that when set to
CUEA_ABORT, will abort the application run with proper error return code.
Along with nice error message.
For example if one were to add same UT twice:
"Aborting due to error #32: Test having this name already in suite."
Application return code set to 32.
Using CU_set_error_action(CUEA_ABORT), removes the need for error checking
on test managment functions in CUnit.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ida7de5e040b714509e79957d47a006ee518a42b8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/930
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Expands blob_simultaneous_operations to cover case
when second md sync is perfomed without waiting for the first
one to complete.
The blob state is forced in UT, but in normal usage could be
changed by modifying xattr, allocating thin clusters or other.
Related to #1170
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I880d289cdf957d4513c080efdcb7a5d8c6b3b447
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/775
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This new UT is checking validity of md sync being run
along another md sync with changed metadata.
At this time added case where first md sync is being
interrupted by removal of xattr with second md sync.
This interruption is perfomed at increasing number of
poller executions, until the number is enough to
complete first md sync.
There are two expected states of used_md_pages array,
either with xattr or without. The state is verified
after md syncs.
This UT will be expaned in similar manner by other
operations (than xattr) that casue changes in persisted metadata.
Related to #960
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I251dca92ffc3080d8dc503a7f1ff342aa59adef9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/774
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Making use of recently added utility to poll exact number of times.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I23a9dc702c899285626522ace8ec77a9ba0f7757
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/773
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
spdk_put_io_channel() was designed to be called on the same thread
that called spdk_get_io_channel(). spdk_put_io_channel() sends a
message to its own thread, to allow the context to unwind before
releasing the resources. This had the side effect to allow an
incorrect thread to call spdk_put_io_channel(). This patch will fix
that.
Bdevperf tool had a design flaw that needed the side effect, but
it was fixed recently. We do not know if we have any other case.
Hence add assert to spdk_put_io_channel() to find other case.
We found that unit test for blobstore had called
spdk_put_io_channel() and fix it together in this patch.
Besides, correct the comment for spdk_put_io_channel() in
include/spdk/thread.h not to create any other case in future.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6ec7bf074818abef43b23ca40bc9385adac70a75
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479390
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Extent table and extent page descriptors are now
set to be default way clusters are serialized on disk.
With this patch UT are ran with and without
extent table.
Changed two asserts in test, since amount is dependent on
which type of serialization is used.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ica58fce6a4effd014d7dd40ee26edd0fa3196d0f
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/481901
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
This will be used to add another run of whole UT suite
with extent pages on/off.
Next patch in series will be enabling both types of
extent serialization for all UT.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ia8b4b8822edefb90ffc13cf777885f9af95e4545
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482170
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
All the non-ext version of the call is doing, is calling
ext with NULL as opts. Then default opts are used in
its place.
This change was facilitated by next on in series,
where all blob opts will be initalized in UT
with parameter use_extent_table either set to
true or false.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I62b642c1808b38a5f7c94a5900f25f4978a4ec39
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482859
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
After creation of blobs in both tests, only clusters indexed from
0 to 10 are supposed to be used. Index 0 for md and 1-10 for data
of single blob since it was create thick provisoned.
Cluster allocations are done in order so if there was a bug for
overflow amount of clusters claimed, first in order would be
one with index 11.
This patch adds asserts after each bs load for first data cluster
that is supposed to be used and for first data cluster that is
not.
During the tests those should remain constant.
When creating/deleting snapshots, the blobs are affected
by changing their type to/from thin_provisioned.
Added asserts to verify their state at every blob open.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I38418da55850d5b8468e578b3c42c5b817ae8045
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482661
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
g_bserrno from blob deletion or snapshot creation,
should not be checked. It is implementation
dependent whether the error (or success) from those
calls actually means that enough data was persisted
on disk.
This test case should work even if we set the threshold
high enough that no failed opperations occur.
On the other hand some parts of those calls do cleanup
in them, meanwhile there is enough metadata data on disk already.
Such as cleaning up unused clusters or pages issue
writes, but at that point the blobs already are in expected
state.
Thus removed assert for g_bserrno, as failure is not indicative
of impossibility to recover.
While here, removed the spdk_bs_unload(). This UT are for
testing power fail safety. Never should it be the case that
enough writes occured in create/delete, but blobs are not
in the expected state.
When such bug would be introduced, it could be covered up
by spdk_bs_unload() cleanly closing up the blobstore.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ic69c3061f2cc1fe04bf895632cdb11efb2fe6912
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/482660
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Extent Pages claim and insertion can be asynchronous
when cluster allocation happens due to writing to a new cluster.
In such case lowest free cluster and lowest free md page
is claimed, and message is passed to md_thread.
Where inserting both into the arrays and md_sycn happens.
This patch adds parameters to pass the Extent Page offset
in such case.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I46d8ace9cd5abc0bfe48174c2f2ec218145b9c75
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/479849
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Accept a clear method option on blob create by adding clear_method
to the opts structure passed in to _spdk_bs_create_blob(). Store
these 2 bits in md_ro_flags so that earlier versions without an
understanding of these bits can not alter metadata.
The new metadata values will be used later in the series.
Signed-off-by: paul luse <paul.e.luse@intel.com>
Change-Id: I5440645ca20b426778d13b2e544b65dc2b3b83c7
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/472204
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Previously this UT was only checking the snapshot counts.
This was correct, but since the blobs in this tests are
clones of each other - it would make sense to check it as well.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Id9dec7f0a292cbb52da8196bfc77679fc0f49d8b
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478976
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
For deletion to actually deliver results, poll_threads()
has to be called.
g_bserrno came from previous API call, not delete itself.
This patch fixes it so that g_bserrno is from the delete itself.
Next patch in series adds checks for number of clones.
Having such checks in first place,
would allow to notice lack of poll_threads() here before.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I3477f5c5b99106b70ee3d5706906a4dc94283180
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/478975
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This patch fixes minor unit test deallocations around io_channels.
It has not exposed any issues in blobstore code, but should always be right
to prevent covering up any issues later.
1) Two spdk_bs_free_io_channel() were missing
2) Dirty shutdown should still free bs resources with _spdk_bs_free()
Sample log at the end of UT, before this patch:
thread.c: 200:_free_thread: *ERROR*: thread 0x617000000080 still has channel for io_device blobstore
thread.c: 200:_free_thread: *ERROR*: thread 0x617000000080 still has channel for io_device blobstore
thread.c: 200:_free_thread: *ERROR*: thread 0x617000000080 still has channel for io_device blobstore
thread.c: 180:spdk_thread_lib_fini: *ERROR*: io_device blobstore not unregistered
thread.c: 180:spdk_thread_lib_fini: *ERROR*: io_device blobstore not unregistered
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I187bc61bb6e094c9c740a987e7d14760551a0503
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/475872
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Previously blob_insert_cluster_msg UT specified the cluster
which will be used to store LBA. So for blob the cluster [1],
was always stored as cluster 0xF in blobstore.
This patch changes _spdk_bs_claim_cluster() with preselected
cluster in blobstore to _spdk_bs_allocate_cluster() that
will choose first free one in the map.
It will help in further patches that add more logic
on extent table and extent pages in _spdk_bs_allocate_cluster().
Otherwise this patch 'hard coded' values might get more complex
than needed.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I238329b2191e7ebd3f73c1fe85d7e1da0249979e
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/475493
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When serializing extents, run-length encoding is supposed to
1) RLE all sequential LBAs
2) RLE zero LBAs (unallocated)
There is one special case, with sequential LBAs that start
with 0 LBA. This is RLE as 1) case, but results in descriptor
matching case 2). Which causes loss of allocated clusters.
This requires following conditions to be met:
- blobstore has just a single cluster reserved for MD
- blob is thin provisioned
- first allocation occurs on cluster_num=1
For last part to be true, very first write for blob has to be
issued to LBA between cluster_size and 2*cluster_size.
Causing allocation of second cluster in blobstore and assiging
it LBA equal to number of LBAs per cluster.
To fix this, case 1) disallows to RLE zeroes.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I136282407966310c882ca97c960e9a71c442c469
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/475494
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
This was inconsistent on all CU_ASSERT macros,
so now it is fixed.
Change-Id: I6d20f98528c4041baf02cb419ab9644a45b807c4
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/459760
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Blob id only is matched to the very first page of md for
that particular blob.
During loading blobstore, we shouldn't verify
further pages in chain against the blobid.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Ifc7863ddcb403aedc264c14e6b4c3915bd30dc41
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/460607
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Length of xattr descriptor is equal to length of xattr struct,
xattr name and the len of stored value.
There is no limit to how much can be stored in memory for xattr.
On disk xattr size is limited to single page and within that to
max descriptors that can fit in it.
This size is known at compile time.
Before this patch it was possible to add xattr exceeding
what was possible to be written to disk. This caused issues
when serializing the metadata during spdk_blob_sync_md()
or spdk_blob_close(). Making those fail without specific info
to the user and not actually writting such descriptor.
Since maximum length of xattr descriptor is known at compile time,
this patch compares against this value when setting the xattr.
It will immediately report back to user with error, and will
not store xattr in memory (thus not serialize it).
This patch should not affect any backward compatibility for blobs.
Too large xattrs weren't written to disk before,
API for blobstore stays the same - only reporting ENOMEM
when it should.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I6f4af4d079e47f084e20d7a4969d9a78ec1f8610
Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/460450
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
