| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
ee58362 Used legacy category to match insert_key_images behavior (vtnerd)
|
|
|
|
|
|
|
|
- New flag in NOTIFY_NEW_TRANSACTION to indicate stem mode
- Stem loops detected in tx_pool.cpp
- Embargo timeout for a blackhole attack during stem phase
|
|
5de2295 Correct key image check in tx_pool (vtnerd)
|
|
760ecf2 console_handler: do not let exception past the dor (moneromooo-monero)
09c8111 threadpool: lock mutex in create (moneromooo-monero)
e377977 tx_pool: catch theoretical error in get_block_reward (moneromooo-monero)
|
|
054b4c7 protocol: request txpool contents when synced (moneromooo-monero)
|
|
A newly synced Alice sends a (typically quite small) list of
txids in the local tpxool to a random peer Bob, who then uses
the existing tx relay system to send Alice any tx in his txpool
which is not in the list Alice sent
|
|
|
|
for example, in the RPC server
|
|
Coverity 196626
|
|
c96b7ee tx_pool: fix error message assuming incorrectly (moneromooo-monero)
|
|
|
|
|
|
Coverity 205394
|
|
88c9d90 protocol: initialize block_weight in block_complete_entry ctor (moneromooo-monero)
fe443bb cryptonote: don't leave block_weight uninitialized (moneromooo-monero)
1ba9baf tx_pool: do not divide by 0 (moneromooo-monero)
|
|
01f660f blockchain: fill in cumulative block weight for alt blocks (moneromooo-monero)
|
|
ec14abd tx_pool: make spent return-by-reference as intended (stoffu)
|
|
In case of a 0 tx weight, we use a placeholder value to insert in the
fee-per-byte set. This is used for pruning and mining, and those txes
are pruned, so will not be too large, nor added to the block template
if mining, so this is safe.
CID 204465
|
|
|
|
8330e77 monerod can now sync from pruned blocks (moneromooo-monero)
|
|
|
|
If the peer (whether pruned or not itself) supports sending pruned blocks
to syncing nodes, the pruned version will be sent along with the hash
of the pruned data and the block weight. The original tx hashes can be
reconstructed from the pruned txes and theur prunable data hash. Those
hashes and the block weights are hashes and checked against the set of
precompiled hashes, ensuring the data we received is the original data.
It is currently not possible to use this system when not using the set
of precompiled hashes, since block weights can not otherwise be checked
for validity.
This is off by default for now, and is enabled by --sync-pruned-blocks
|
|
The 98th percentile position in the agebytes map was incorrectly
calculated: it assumed the transactions in the mempool all have unique
timestamps at second-granularity. This commit fixes this by correctly
finding the right cumulative number of transactions in the map suffix.
This bug could lead to an out-of-bounds write in the rare case that
all transactions in the mempool were received (and added to the mempool)
at a rate of at least 50 transactions per second. (More specifically,
the number of *unique* receive_time values, which have second-
granularity, must be at most 2% of the number of transactions in the
mempool for this crash to trigger.) If this condition is satisfied, 'it'
points to *before* the agebytes map, 'delta' gets a nonsense value, and
the value of 'i' in the first stats.histo-filling loop will be out of
bounds of stats.histo.
|
|
|
|
|
|
|
|
6644b9b blockchain_db: remove a couple unused functions (moneromooo-monero)
ce594f5 blockchain_db: allocate known size vector only once (moneromooo-monero)
8332698 db_lmdb: inline check_open, it's trivial and called everywhere (moneromooo-monero)
5511563 db_lmdb: avoid pointless division (moneromooo-monero)
d1efe3d cryptonote: set tx hash on newly parsed txes when known (moneromooo-monero)
9cc68a2 tx_pool: add a few std::move where it can make a difference (moneromooo-monero)
|
|
|
|
|
|
|
|
cf552c75 tx_pool: allow take_tx to work without m_txs_by_fee_and_receive_time (moneromooo-monero)
|
|
5808530f blockchain: remove unused output_scan_worker parameter (moneromooo-monero)
1426209a blockchain: don't run threads if we have just one function to run (moneromooo-monero)
6f7a5fd4 db_lmdb: slight speedup getting array data from the blockchain (moneromooo-monero)
99fbe100 db_lmdb: save some string copies for readonly db keys/values (moneromooo-monero)
bf31447e tx_pool: speed up take_tx for transactions from blocks (moneromooo-monero)
4f005a77 tx_pool: remove unnecessary get_transaction_hash (moneromooo-monero)
593ef598 perf_timer: call reserve on new timer array (moneromooo-monero)
6ecc99ad core: avoid unnecessary tx/blob conversions (moneromooo-monero)
00cc1a16 unit_tests: notify test special case for the usual weirdo (moneromooo-monero)
|
|
5a2e54a1 Fix prune using receive time as tx_weight (doy-lee)
|
|
This happens for every historical tx when syncing, and the
unnecessary parsing is actually showing up on profile.
Since these are kept cached for just one block, this does
not increase memory usage after syncing.
|
|
|
|
|
|
This should make it possible to have two daemons running on the
same database again.
|
|
|
|
|
|
6844ae1b tx_pool: avoid parsing a whole tx if only the prefix is needed (moneromooo-monero)
|
|
get_transaction_pool RPC
Inspired by https://github.com/masari-project/masari/issues/93
|
|
|
|
|
|
|
|
|
|
48a7bc4 tx_pool: fix infinite loop when failing to find a meta record (moneromooo-monero)
|
|
|
|
This avoids constant rechecking of the same things each time
a miner asks for the block template. The tx pool maintains
a cookie to allow users to detect when the pool state changed,
which means the block template needs rebuilding.
|
|
41b4bf9 tx_pool: cache check_tx_inputs results (moneromooo-monero)
|
|
a2b557f 6795bd0 209ec96 ed2c81e a830db2 57ea902 31a895e ba8331c f7f1917 41be339 f025ae9 ef2cb63 dcfd299 5d3e702 2704624 2771a18 0e4c7d0 (moneromooo-monero)
|
|
also use reserve where appropriate
|
|
d81e042 tx_pool: initialize bitflags padding since it gets written to storage (moneromooo-monero)
|
|
This is called a lot when creating a block template, and
does not change until the blockchain changes.
This also avoids tx parsing when cached.
|
|
Avoids valgrind reporting uninitialized data usage
|
|
|
|
|
|
A key image may be present more than once if all but one of the
txes spending that key image are coming from blocks. When loading
a txpool from storage, we must load the one that's not from a
block first to avoid rejection
|
|
|
|
Previously, when blob_size == tx_size_limit, the "m_too_big" property was set
and the transaction was rejected. This should not have been the case.
|
|
39992134 txpool: Properly bail out when outputs_amount == inputs_amount (Leon Klingele)
|
|
|
|
Previously, when outputs_amount == inputs_amount, the "m_overspend" property
was set, whereas "m_fee_too_low" would have been the correct property to set.
This is unlikely to ever occur and just something I've noticed while reading
through the code.
|
|
|
|
|
|
It's nasty, and actually breaks on Solaris, where if.h fails to
build due to:
struct map *if_memmap;
|
|
10013e94 Protect node privacy by proper filtering in restricted-mode RPC answers (binaryFate)
|
|
This patch allows to filter out sensitive information for queries that rely on the pool state, when running in restricted mode.
This filtering is only applied to data sent back to RPC queries. Results of inline commands typed locally in the daemon are not affected.
In practice, when running with `--restricted-rpc`:
* get_transaction_pool will list relayed transactions with the fields "last relayed time" and "received time" set to zero.
* get_transaction_pool will not list transaction that have do_not_relay set to true, and will not list key images that are used only for such transactions
* get_transaction_pool_hashes.bin will not list such transaction
* get_transaction_pool_stats will not count such transactions in any of the aggregated values that are computed
The implementation does not make filtering the default, so developers should be mindful of this if they add new RPC functionality.
Fixes #2590.
|
|
Transactions in the txpool are marked when another transaction
is seen double spending one or more of its inputs.
This is then exposed wherever appropriate.
Note that being marked with this "double spend seen" flag does
NOT mean this transaction IS a double spend and will never be
mined: it just means that the network has seen at least another
transaction spending at least one of the same inputs, so care
should be taken to wait for a few confirmations before acting
upon that transaction (ie, mostly of use for merchants wanting
to accept unconfirmed transactions).
|
|
0aaaca29 tx_pool: set the "invalid input" bit when check_tx_inputs fails (moneromooo-monero)
9236823b simplewallet: print tx rejection reason where it was missing (moneromooo-monero)
3dee3301 core_rpc_server: print tx rejection reason at L0 too (moneromooo-monero)
|
|
CID 175316
|
|
Should be impossible in practice, but easy change
CID 175282
|
|
instead of just failing
This is a workaround for bad tx blobs being inserted in the
pool for unknown reasons
|
|
f761dbae Remove 1.25x multiplier from tx_pool (Nano Akron)
|
|
0299cb77 Fix various oversights/bugs in ZMQ RPC server code (Thomas Winget)
77986023 json serialization for rpc-relevant monero types (Thomas Winget)
5c1e08fe Refactor some things into more composable (smaller) functions (Thomas Winget)
9ac2ad07 DRY refactoring (Thomas Winget)
|
|
|
|
|
|
Structured {de-,}serialization methods for (many new) types
which are used for requests or responses in the RPC.
New types include RPC requests and responses, and structs which compose
types within those.
# Conflicts:
# src/cryptonote_core/blockchain.cpp
|
|
This might prevent some calls to terminate when the LockedTXN
dtor is called as part of stack unwinding caused by another
exception in the first place.
|
|
|
|
1914c999 txpool: update db tx metadata when it changes (moneromooo)
4dbf29bd txpool: add tx size median to the pool stats (moneromooo)
|
|
|
|
|
|
|
|
if tx receive_time == now. supersedes #2322
|
|
|
|
235df7f4 blockchain_db: add a txpool tx getter which returns existence (moneromooo-monero)
|
|
Avoids exception spam for the "nope, not found" case
|
|
|
|
31417d57 tx_pool: add missing blockchain lock in add_tx (moneromooo-monero)
|
|
Since we're just counting txs, there's no reason to deserialize all the blobs.
|
|
|
|
b52abd13 Move txpool to the database (moneromooo-monero)
|
|
Integration could go further (ie, return_tx_to_pool calls should
not be needed anymore, possibly other things).
poolstate.bin is now obsolete.
|
|
Only works from V5 fork onward - returns 0 before that block.
|
|
|
|
91d41090 tx_pool: ensure txes loaded from poolstate.bin have their txid cached (moneromooo-monero)
aaeb164c tx_pool: remove transactions if they're in the blockchain (moneromooo-monero)
558cfc31 core, wallet: faster tx pool scanning (moneromooo-monero)
f065234b core: cache tx and block hashes in the respective classes (moneromooo-monero)
|
|
The txid is not saved, and we want to make sure the transactions
have their txid cached while in the pool, since get_transactions
copies the transaction object, so any txid calculation on those
copies would not benefit any later caller, since the original tx
would be left without a cached txid.
|
|
When starting up, if the pool state was not saved, the pool
might contain transactions which made it into the blockchain,
so these need removing
|
|
Includes a new RPC to get tx pool hashes fast.
|
|
The result is not actually used when uninitialized
|
|
|
|
Minimum mixin 4 and enforced ringct is moved from v5 to v6.
v5 is now used for an increased minimum block size (from 60000
to 300000) to cater for larger typical/minimum transaction size.
The fee algorithm is also changed to decrease the base per kB
fee, and add a cheap tier for those transactions which we do
not care if they get delayed (or even included in a block).
|
|
|
|
|
|
They are not ready yet
|
|
|
|
|
|
58e82506 Blockfill - Sort tx pool correctly (Alexis Enston)
5f7a8741 Blockfill - Take TX fees into account properly (Alexis Enston)
4ecab0d8 Consider empty block when filling with TXs (Alexis Enston)
|
|
|
|
|
|
|
|
|
|
This replaces the epee and data_loggers logging systems with
a single one, and also adds filename:line and explicit severity
levels. Categories may be defined, and logging severity set
by category (or set of categories). epee style 0-4 log level
maps to a sensible severity configuration. Log files now also
rotate when reaching 100 MB.
To select which logs to output, use the MONERO_LOGS environment
variable, with a comma separated list of categories (globs are
supported), with their requested severity level after a colon.
If a log matches more than one such setting, the last one in
the configuration string applies. A few examples:
This one is (mostly) silent, only outputting fatal errors:
MONERO_LOGS=*:FATAL
This one is very verbose:
MONERO_LOGS=*:TRACE
This one is totally silent (logwise):
MONERO_LOGS=""
This one outputs all errors and warnings, except for the
"verify" category, which prints just fatal errors (the verify
category is used for logs about incoming transactions and
blocks, and it is expected that some/many will fail to verify,
hence we don't want the spam):
MONERO_LOGS=*:WARNING,verify:FATAL
Log levels are, in decreasing order of priority:
FATAL, ERROR, WARNING, INFO, DEBUG, TRACE
Subcategories may be added using prefixes and globs. This
example will output net.p2p logs at the TRACE level, but all
other net* logs only at INFO:
MONERO_LOGS=*:ERROR,net*:INFO,net.p2p:TRACE
Logs which are intended for the user (which Monero was using
a lot through epee, but really isn't a nice way to go things)
should use the "global" category. There are a few helper macros
for using this category, eg: MGINFO("this shows up by default")
or MGINFO_RED("this is red"), to try to keep a similar look
and feel for now.
Existing epee log macros still exist, and map to the new log
levels, but since they're used as a "user facing" UI element
as much as a logging system, they often don't map well to log
severities (ie, a log level 0 log may be an error, or may be
something we want the user to see, such as an important info).
In those cases, I tried to use the new macros. In other cases,
I left the existing macros in. When modifying logs, it is
probably best to switch to the new macros with explicit levels.
The --log-level options and set_log commands now also accept
category settings, in addition to the epee style log levels.
|
|
81c384e4 fix do_not_relay not preventing relaying on a timer (moneromooo-monero)
|
|
Also print its value when printing pool
|
|
The transactions are first prioritized by fee and in case the fees are
equal by receive_time.
|
|
1607cb7e tx_pool: better block template filling algorithm (moneromooo-monero)
9731b4e5 rpc: add block size to GET_BLOCK_HEADER RPC (moneromooo-monero)
9188b346 rpc: add current block size to the getinfo call (moneromooo-monero)
|
|
Continue filling until we reach the block size limit, or the
resulting coinbase decreases.
Also remove old sanity check on block size, which is now not
wanted anymore.
|
|
|
|
|
|
The fee will vary based on the base reward and the current
block size limit:
fee = (R/R0) * (M0/M) * F0
R: base reward
R0: reference base reward (10 monero)
M: block size limit
M0: minimum block size limit (60000)
F0: 0.002 monero
Starts applying at v4
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The whole rct data apart from the MLSAGs is now included in
the signed message, to avoid malleability issues.
Instead of passing the data that's not serialized as extra
parameters to the verification API, the transaction is modified
to fill all that information. This means the transaction can
not be const anymore, but it cleaner in other ways.
|
|
|
|
|
|
|
|
|
|
It is not yet constrained to a fork, so don't use on the real network
or you'll be orphaned or rejected.
|
|
|
|
|
|
24b3e90 Convey tx verification failure reasons to the RPC client (moneromooo-monero)
|
|
|
|
This allows appropriate action to be taken, like displaying
the reason to the user.
Do just that in simplewallet, which should help a lot in
determining why users fail to send.
Also make it so a tx which is accepted but not relayed is
seen as a success rather than a failure.
|
|
tx_pool.h doxygen documentation completed.
Many notes made on areas for improvement, be that functionality or
code clarity.
Commented code and unused code removed.
|
|
This is intended to avoid cases where a timed out tx will be
re-relayed by another peer for which it has not timed out yet,
which would cause the tx to stay in the network's pool for a
long time (until all peers time it out before another one
tries to relay it again).
|
|
|
|
Some of it uses hardcoded height, which will need some thinking
for next (voted upon) fork.
|
|
Fixes intermittent test failures when the pool contains
unexpected transactions that were brought in from the
live pool.
|
|
|
|
Remove trailing whitespace in same files.
|
|
|
|
The last relayed time of a transaction is maintained, and
transactions will be relayed again if they are still in the
pool after a certain amount of time, which increases with
the transaction's age. All such transactions are resent,
whether or not they originated on the local node.
|
|
The wallet decomposes fully as of now too.
|
|
Added option to cache tx-input verification results.
|
|
Fixed OSX compilation issues due to random lmdb resize points.
Fixed infinite loop bug when calling core::get_block_template(..).
|
|
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
|
|
to lock up.
|
|
|
|
|
|
|
|
|
|
When a stuck tx is removed from memory pool, first remove the associated
spent key images.
|
|
This is for the "print_pool" command and "get_transaction_pool" RPC
method.
Add mempool's spent key images to the results.
|
|
Usage:
default is lmdb for blockchain branch:
$ make release
same as:
$ DATABASE=lmdb make release
for original in-memory implementation:
$ DATABASE=memory make release
|
|
Probably needs more looking at -- lot of things were done...in a rushed
sort of way. That said, it all builds and *should* be at least
testable.
update for rebase (warptangent 2015-01-04)
fix conflicts with upstream CMakeLists.txt files
src/CMakeLists.txt (remove edits from original commit)
tests/CMakeLists.txt (remove edits from original commit)
src/cryptonote_core/CMakeLists.txt (edit)
- use blockchain db .cpp and .h files
- add LMDB_LIBRARIES
|
|
|
|
|
|
denominations based on DEFAULT_DUST_THRESHOLD, document fee arg to create_transactions as unused, se DEFAULT_DUST_THRESHOLD for wallet dust collection instead of calcualted tx fee
|
|
|
|
This reverts commit 4e2b2b942daa4206ec44c66e59863670dfe3fde4.
|
|
|
|
|
|
|
|
tx pool fixes, courtesy of Bytecoin
|
|
|
|
|
|
|
|
This is now done when adding transactions to the pool.
|
|
This is now done when transactions enter the pool.
|
|
|
|
|
|
|
|
|
|
Transactions larger than the limit will not enter the pool
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
