The purpose of this change is not to write superblock to the cache
drive untill all other sections are initilized on disk in attach()
path. Combined with superblock clearing at the erarlier stage of
attach(), this assures there are no residual mappings in the collision
section in case of power failure during attach with pre-existing
metadata.
This is implemented by removing ocf_metadata_flush_all_set_status() step
at the beginning of ocf_metadata_flush_all().
ocf_metadata_flush_all() is called, except for the attach() case described
above, in two cases:
1. at the end of cache load - potentially after cache recovery
2. during detaching cache drive in cache stop.
To make sure there are no regressions in the first case, an explicit
_ocf_mngt_attach_shutdown_status() is added to load pipeline before
ocf_metadata_flush_all(). The second case is always ran after cache
drive is attached, so dirty status bit must have already be written to
the disk.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Because of metadata flapping it is much more complicated to capture those
sections in flight in standby mode, so we read them directly from the cache
volume during the activate.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
This feature provides double buffering of config sections to prevent
situation when power failure during metadata flush leads to partially
updated metadata. Flapping mechanism makes it always possible to perform
graceful rollback to previous config metadata content in such situation.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
Request submitted in fast path may be freed before the sequential cutoff stats
are updated. Increment request reference counter to prevent it.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Move error print to where it belongs, preventing this message to
pop up when same error code is reported elsewhere for other reason.
Signed-off-by: Rafal Stefanowski <rafal.stefanowski@intel.com>
This patch fixes the issue 988 (and 997) causing a kernel stack
overflow.
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
To allow the fastest switching from the passive-standby to active mode, the
runtime metadata must be kept 100% synced with the metadata on the drive and in
the RAM thus recovery is required after each collision section update.
To avoid long-lasting recovering of all the cachelines each time the collision
section is being updated, the passive update procedure recovers only those
which have its MD entries on the updated pages.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Starting cache in a standby mode requires access to a valid cleaning policy
type. If the policy is stored only in the superblock, it may be overridden by
one of the metadata passive updates.
To prevent losing the information it should be stored in cache's runtime
metadata.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Initializing cleaning policy is very time consuming. To reduce the time required
for activating cache instance the initialization sholud be done during passitve
start
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Since part of the recovery is done during `standby init`, the correct shutdown
status has to be set
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
The unsafe mode is useful if the metadata of added cores is incomplete.
Such scenario is possible when starting cache to standby mode from partially
vaild metadata.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Make sure all the invalid cachelines have reset status bits. This allows to
recognize invalid cachelines easily during populate.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Recovery during passive start is based on the assuption that metadata collision
section stored on disk might be partially valid. Reseting this data would make
rebuilding metadata impossible.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Decide whether to promote sequential cutoff stream
to global structures when threshold is reached
Signed-off-by: Rafal Stefanowski <rafal.stefanowski@intel.com>
GCC 11 static check finds an array size mismatch and prevents OCF from
compiling correctly. This fixes OpenCAS Linux issue #968
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
Error for an invalid cache operation while in passive mode added
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
Error name correction
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
API changes for passive cache mode
Moved the passive cache error return source to the api for flush and
set_param
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
Further API changes for passive cache mode
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
Passive api - review changes
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
Management pipelines tend to consist of multiple asynchronous steps.
Allowing synchronous queue kick results in massive call stacks (e.g.
almost 500 functions deep in case of cache stop). Since async kick
is required anyway, it seems reasonable to switch to async kick
in pipeline implementation.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Cache name is needed for logging in passive mode, when config metadata
is still not accessible.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
To prevent deinitializing cleaner context (i.e. during switching policy) during
processing requests, access to cleaner should be protected with reference
counter
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Making the operation asynchronous will allow to use refcnt utility as an
synchronization mechanism between processing cachelines and deinitializing
cleaning policy.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
The change should unify access to cleaning policy resources and facilitate
synchronization when switching cleaning policies
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
All remapped cachelines are write locked. If the operation fails cachelines has
to be unlocked during rollback
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
ocf_lru_get_list() now returs clean list for freelist partition to
provide common interface regardless of partition type.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Flushing metadata in WT is required only if at least of the request's cacheline
changed its state to clean.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
There's a possibility that WT write is performed to dirty cache line (i.e. after
switching WB->WT without flush) and status bits change from dirty to clean. If
power failure occurs it might happen that recovery would ignore recent data from
cache and would assume that data is clean while backend storage data is out of
date.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
In WB mode metadata should be updated only if the actuall data had been saved
on disk. Otherwise metadata might be flushed too early and consequently data
corruption might occur.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
src/eviction/lru.c -> src/ocf_lru.c
src/eviction/lru.h -> src/ocf_lru.h
src/eviction/lru_structs.h -> src/ocf_lru_structs.h
src/eviction/eviction.c -> src/ocf_space.c
src/eviction/eviction.h -> src/ocf_space.h
.. as well as corresponding UT files.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
... in UT as well
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
New structure ocf_part is added to contain all the data common for both
user partitions and freelist partition: part_runtime and part_id.
ocf_user_part now contains ocf_part structure as well as pointer to
cleaning partition runtime metadata (moved out from part_runtime) and
user partition config (no change here).
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Lookup is repeated after request is identified as miss and hash bucket
lock is upgraded (in order to map missing cachelines). At this point
cachelines status might change and the request might turn out to be
a hit after all. Adding check for this condition removes unnecessary
calls to remap logic.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
This allows access to it in ctx_metadata_updater_init, which is
done in the same call stack during initalization.
Signed-off-by: Kozlowski Mateusz <mateusz.kozlowski@intel.com>
After detaching a core if user wanted to remove inactive cores the
cleaning policy data would not be initialized and would bug-out on next
core add.
This check was incorrect, as cleaning policy core metadata lifetime is
not bound to core volume being open or not.
Signed-off-by: Jan Musial <jan.musial@intel.com>
Eviction should decrement occupancy statistics for the
core from which a cacheline is being evicted rather than
from the I/O target core.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
update_req_info() should include REMAPPED cachelines
in repart stats (number of cachelines within request
belonging to other than the target partition).
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
This check is incorrect as cacheline status may change
from dirty to clean at any point during cleaning, except for
when the hash bucket is locked.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
This isn't strictly required in current implementation as
nodes are always re-initialized before inserting to LRU list.
However it seems to make sense to zero the flag anyway to
make the code easier to reason about.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Removing conditional early return from engine_map() function
in case of insufficient free cachelines. The reasons are:
1. current implementation does not treat unssufficient free
cachelines condition as an error,
2. the check is based on stale request info, so it is inaccurate,
3. it is easier to hit more paths with functional tests,
4. partially mapping request from the freelist becomes more common
rather than being a corner case dependent on racy timings between
threads
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Move ref counts to their own cacheline - otherwise they pollute and cause
false sharing to fields nearby and cause a lot of cache bouncing between
physical CPUs.
Signed-off-by: Kozlowski Mateusz <mateusz.kozlowski@intel.com>
Grouping static fields together, while often changing ones get their own
cacheline, or some not used often/important fields.
Signed-off-by: Kozlowski Mateusz <mateusz.kozlowski@intel.com>
set_hot() depends on cacheline metadata status to determine
on which list the element is located (dirty cs clean list).
Thus at least hash bucket lock is required when calling
set_hot().
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
When issuing discard request over 512KiB OCF would trim this request and
overwrite req->core_line_count which would then cause this request to be
freed from wrong mpool.
This is fixed now by saving core_line_count that was set when allocating
this request that is never overwritten. This alloc_core_line_count is
then used to free the request from correct mpool.
Signed-off-by: Jan Musial <jan.musial@intel.com>
If user thread is preempted during tree/list update and another IO
is issued on the same CPU, the structure will be in undefined state.
This may result in hung tasks, if the tree stops being a tree and a loop exists -
tree search functions won't be able to end properly; or panics if a NULL value appears
suddenly in the preempted thread, after a null-check was already done.
Signed-off-by: Kozlowski Mateusz <mateusz.kozlowski@intel.com>
When allocation of request with map fails, we fallback to allocating
request with no map, and then allocate map separately. During request
put we need to distinguish between those two cases in order to deallocate
request properly.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
Early return from engine_map() in case of insufficient free
cachelines on the freelist is opportunistic, as both request
map info and freelist count are not accurate. Map info is stale
as it is to be refreshed in engine_map() after hash bucket
lock had been upgraded. Freelist count on other hand is subject
to change asynchronously.
The implementation assumption however is that after engine_map()
request is fully traversed (engine_map() is equivalent to
engine_lookup() followed by an attempt to map missing cachelines).
So in case of early return we must take care of repeating the
lookup.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
At this point cacheline status in request map is stale,
as lookup was performed before upgrading hash bucket lock.
If indeed all cachelines are mapped, this will be determined
in the main loop of engine_map().
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
This assures that cacheline with LOOKUP_INSERTED status
is always present on the LRU list.
This fixes an ENV_BUG() caused by an attempt to remove
a cacheline from LRU list which was not there. This
happened when cacheline was mapped from freelist
(LOOKUP_INSERTED) but the entire request mapping failed
and generic cleanup routines attempted to invalidate cacheline,
including removing it from the LRU list. As engine_set_hot()
is called after successfull mapping, the inserted cacheline was
not yet present on the LRU list.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Number of cachelines to evcit can't be greater than the number of unmapped
entries in request.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Split traversation into two distinct phases: lookup()
and lru set_hot(). prepare_cachelines() now only calls
set_hot() once after lookup and insert are finished.
lookup() is called explicitly only once in
prepare_cachelines() at the very beginning of the
procedure. If request is a miss then then map()
performs operations equivalent to lookup() supplemented
by an attempt to map cachelines. Both lookup() and
set_hot() are called via traverse() from the engines
which do not attempt mapping, and thus do not call
prepare_clines().
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Unmapping cachelines previously mapped from freelist before
eviction is a waste of resources. Also if map does not erarly
exit upon first mapping error, we can have request fully
traversed (and partially mapped) after mapping and thus
skip lookup in eviction.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Eviction changes allowing to evict (remap) cachelines while
holding hash bucket write lock instead of global metadata
write lock.
As eviction (replacement) is now tightly coupled with request,
each request uses eviction size equal to number of its
unmapped cachelines.
Evicting without global metadata write lock is possible
thanks to the fact that remaping is always performed
while exclusively holding cacheline (read or write) lock.
So for a cacheline on LRU list we acquire cacheline lock,
safely resolve hash and consequently write-lock hash bucket.
Since cacheline lock is acquired under hash bucket (everywhere
except for new eviction implementation), we are certain that
noone acquires cacheline lock behind our back. Concurrent
eviction threads are eliminated by holding eviction list
lock for the duration of critial locking operations.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
.. so that the main part, responsible strictly for mapping
given LBA to given collision index, is encapsulated in
a function ocf_map_cache_line with external linkage.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Changing sequential request detection so that a miss request is
recognized as sequential after needed cachelines are evicted
and mapped to the request in a sequential order.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
.. to make it clean that true means cleaner must lock
cachelines rather than the lock is already being held.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Allowing request cacheline lock to be called on partially
locked request. This is going to be usefull for upcomming
eviction improvements, where request will first have evicted
(LOOKUP_REMAPPED) cachelines assigned to it in a locked state,
followed by standard request cacheline lock call in order to
lock previously inserted (LOOKUP_HIT) or mapped from freelist
(LOOKUP_INSERTED) cachelines.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Function returns true if cacheline is locked (read
or write) by exactly one entity with no waiters.
This is usefull for eviction. Assuming caller holds
hash bucket write lock, having exlusive cacheline
lock (either read or write) allows holder to remap
cacheline safely. Typically during eviction hash
bucket is unknown until resolved under cacheline lock,
so locking cacheline exclusively (instead of locking
and checking for exclusive lock) is not possible.
More specifically this is the flow for synchronizing
cacheline remap using ocf_cache_line_is_locked_exclusively:
1. acquire a cacheline (read or write) lock
2. resolve hash bucket
3. write-lock hash bucket
4. verify cacheline lock is exclusive
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Reformat function that calculates how long cache/core is dirty
Update `dirty_for` types in functional tests
Values stored in info structs fields (both in cache and core structs)
are unsigned 64-bits ints but `dirty_for`s were unsigned 32-bits ints.
Use existing function to transform returned value to seconds.
Replace seconds stored in metadata with seconds.
Replacement was done if old value of replaced field was equal to zero.
Acquiring monotonic high precission timestamp is potentially
slow and it makes sense to compare the field's value
to zero before calling atomic function.
Signed-off-by: Slawomir Jankowski <slawomir.jankowski@intel.com>
Provide number of cachelines as the cacheline concurrency
construtor param instead of reading it from cache.
The purpose of this change is to improve testability.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Cacheline concurrency functions have their interface changed
so that the cacheline concurrency private context is
explicitly on the parameter list, rather than being taken
from cache->device->concurrency.cache_line.
Cache pointer is no longer provided as a parameter to these
functions. Cacheline concurrency context now has a pointer
to cache structure (for logging purposes only).
The purpose of this change is to facilitate unit testing.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
The main purpose of cacheline concurrency global lock
is to eliminate the possibility of deadlocks when
locking multiple cachelines.
Cacheline lock fast path does not need to acquire
this lock, as it is only opportunistically attempting
to lock all clines without wait. There is no risk
of deadlock, as:
* concurrent fast path will also only try_lock
cachelines, releasing all acquired locks if failed
to immediately acquire lock for any cacheline
* concurrent slow path is guaranteed to have
precedence in lock acquisition when conditions
for deadlock occure (both slowpath and fastpath
have acquired some locks required by the other
thread). This is because the fastpath thread will
back off (release acquired locks) if any one of the
cacheline locks is not acquired.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
If an ioclass is pinned but it exceeded its occupancy limit, it should be
evicted anyway.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
.. in order to move primitives intended to be accessed
concurrently in separate CPU cache line.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Divide single global lock instance into 4 to reduce contention
in multiple read-locks scenario.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
1. new abbreviated previx: ocf_hb (HB stands for hash bucket)
2. clear distinction between functions requiring caller to
hold metadata shared global lock ("naked") vs the ones
which acquire global lock on its own ("prot" for protected)
3. clear distinction between hash bucket locking functions
accepting hash bucket id ("id"), core line and lba ("cline")
and entire request ("req").
Resulting naming scheme:
ocf_hb_(id/cline/req)_(prot/naked)_(lock/unlock/trylock)_(rd/wr)
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Top 50% least recently used cachelines are not promoted
to list head upon access. Only after cacheline drops to
bottom 50% it is considered as a candidate to promote
to list head.
The purpose of this change is to reduce overhead of
LRU list maintanance for hot cachelines.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Overflown partitions now have precedence over others during
eviction, regardless of IO class priorities.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Removing the logic for oportunistic partition overflow
reduction by evicting more cachelines than actually
required by the request being serviced.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
If request is hit, simply try to acquire cachelines instead of verifying
whether target partition's size is not exceeded.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
If there is any dirty data on the cache associated with removed core,
we must flush collision metadata after removing core to make metadata
persistent in case of dirty shutdown.
This fixes the problem when recovery procedure erroneously interprets
cache lines that belonged to removed core as valid ones.
This also fixes the problem, when after removing core containing dirty
data another core is added, and then recovery procedure following dirty
shutdown assigns cache lines from removed core to the new one, effectively
leading to data corruption.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
Min and max values, keept as an explicit number of cachelines, are tightly
coupled with particular cache. This might lead to errors and mismatches after
reattaching cache of different size.
To prevent those errors, min and max should be calculated dynamically.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Since the request carries an explicit information about number of the
cacheliens to be reparted, no need of keeping the boolean information if some
of the request's cachelines are assigned to a wrong partition
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Instead of redunant calculating number of cachlines to be reparted, keep this
information in request's info
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
If partition's occupancy limit is reached, cachelines should be evicted from
request's target partition.
Information whether particular partition eviction should be triggered is
carried as a flag by request which triggered eviction.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Moving metadata implementation out of obsolete metadata_hash.c
to .c files corresponding to function declaration header files.
This requires adding shared header for metadata implementation
metadata_internal.h. Some metadata header files did not have
a corresponding .c file - in this case it is added in this
commit.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Metadata wrapper functions (calling iface->func) in header
files are changed to be declarations only. Hash interface
implementation functions in metadata_hash.c are given an
external linkage and are renamed to drop "hash" prefix.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Locks acquired in ocf_metadata_flush(/load)_all are
acquired only for the duration of queueing asynch
service for flush/load, no actual metadata accesses
are performed there.
Also flush/load all are always performed with metadata
marked as deinitialized (metadata reference counter freezed),
so no I/O is reading nor writing the metadata. The only source
of potential concurrent metadata accesses are other management
operations, which should be synchronized using management lock.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Initializing each stream with unique LBA ensures there are no initial
rbtree collisions, and thus helps to avoid clustering of all the streams
into one big linked list instead of forming performance friendly proper
tree structure.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
After writing metadata configuration to disk we must
send a flush request to make sure configuration sections
are commited to non-volatile storage.
Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
Load properties before checking memory needs and obtain cache line size
from context rather than from cache state.
Signed-off-by: Rafal Stefanowski <rafal.stefanowski@intel.com>
Rather then passing whole structs, supply
_ocf_mngt_calculate_ram_needed() with just the values it actually uses.
Signed-off-by: Rafal Stefanowski <rafal.stefanowski@intel.com>
During recovery procedure there is no guarantee that checksums
of runtime sections were flushed correctly before dirty shutdown.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
