The Blockstore currently maintains a RwLock<Slot> of the maximum root
it has seen inserted. The value is initialized during
Blockstore::open() and updated during calls to Blockstore::set_roots().
The max root is queried fairly often for several use cases, and caching
the value is cheaper than constructing an iterator to look it up every
time.
However, the access patterns of these RwLock match that of an atomic.
That is, there is no critical section of code that is run while the
lock is head. Rather, read/write locks are acquired in order to read/
update, respectively. So, change the RwLock<u64> to an AtomicU64.
* allow test feature to skip rewrites
* hook up cli arg for test skip rewrites, update tests
* fix sanity checker
* add account hash to abi to fix a test
* reviews
* use hashmap to collect skip_rewrites. exclude skip_rewrites from dirty
pubkey set
* accumulate skipped_rewrite in reduce
* mutex
* fmt
* skip hash verify for this test flag
* add skipped rewrites num stat
* skip bank hash verify not account hash verify
* reviews
---------
Co-authored-by: HaoranYi <haoran.yi@solana.com>
* Do not unnecessarily re-verify unloaded program
* clippy fixes
* new unit test for extract
* fixes after rebase
* fixes after rebase
* address review comments
* fix clippy
* allow pedantic invalid cast lint
* allow lint with false-positive triggered by `test-case` crate
* nightly `fmt` correction
* adapt to rust layout changes
* remove dubious test
* Use transmute instead of pointer cast and de/ref when check_aligned is false.
* Renames clippy::integer_arithmetic to clippy::arithmetic_side_effects.
* bump rust nightly to 2023-08-25
* Upgrades Rust to 1.72.0
---------
Co-authored-by: Trent Nelson <trent@solana.com>
In most cases, either a &Bank or an Arc<Bank> is more proper.
- &Bank is used if the function only needs a momentary reference
- Arc<Bank> is used if the function needs its' own copy
This PR leaves several instances of &Arc<Bank> around; these instances
are situations where a clone may only happen conditionally.
* Integrate program loader-v4 with bank
* fix tests
* new struct for ProgramRuntimeEnvironments
* remove environment from program_runtime_environment_v
* move find_program_in_cache() to invoke_context
* cleanup
When ledger-tool runs, it may create secondary directories for things
like accounts, accounts-index, etc as not to potentially interfere with
solana-validator's directories. These would show as multiple directories
with ".ledger-tool" appended to the typical directory name.
To more clearly group these items, make the default directories for
snapshots, accounts and accounts-index all within a common "ledger-tool"
directory.
When a consensus divergance occurs, the current workflow involves a
handful of manual steps to hone in on the offending slot and
transaction. This process isn't overly difficult to execute; however, it
is tedious and currently involves creating and parsing logs.
This change introduces functionality to output a debug file that
contains the components go into the bank hash. The file can be generated
in two ways:
- Via solana-validator when the node realizes it has diverged
- Via solana-ledger-tool verify by passing a flag
When a divergance occurs now, the steps to debug would be:
- Grab the file from the node that diverged
- Generate a file for the same slot with ledger-tool with a known good
version
- Diff the files, they are pretty-printed json
Several of the program command subcommands do not require a ledger as
they act solely on the program object. So, defer checking the ledger
path until we know we need to load. Additionally, remove genesis arg
from these commands that do not load a genesis.bin
The current desired open file descriptor limit is 1,000,000. This is
quite a large number, and not needed for every command. Namely, commands
that do not unpack a snapshot and create an AccountsDB will likely not
use this many files.
There is already an option in BlockstoreOptions to ignore errors if the
desired value cannot be set; this PR just bubbles that option up to a
CLI flag in ledger-tool.
The --encoding flag only makes sense for the genesis command if the
genesis accounts will be printed. Hence, the encoding flag used the
Arg::requires() function on --accounts argument.
However, in what appears to be a clap bug, this made the --accounts flag
required even though it should not be. So, remove the .requires()
dependency between these two arguments.
* Move CostModel and CostTracker to its own crate
* compile new crate and update imports
* update sbf Cargo.lock
* fix AbiExample
* fix cargo sort
* Fix AbiExample
If a slot is marked as optimistically confirmed, it is probable but not
guaranteed that its' ancestors will also be marked as optimistically
confirmed in the Blockstore. Given the importance of examining
optimistically confirmed slots around cluster restarts, manually walk
an AncestorIterator to avoid the chance of a slot improperly being
ignored in cluster restart scenarios.
The optional args allow reuse by ledger-tool repair roots command Also,
hold cleanup lock for duration of Blockstore::scan_and_fix_roots().
This prevents a scenario where scan_and_fix_roots() could identify a
slot as needing to be marked root, that slot getting cleaned by
LedgerCleanupService, and then scan_and_fix_roots() marking the slot as
root on the now purged slot.
The previous help message was pretty generic as it covered two potential
error scenarios. This change gives a more detailed error message to each
of those errors.
* Restrict access to Bank's HardForks
Callers could previously obtain a a lock to read/write HardForks from
any Bank. This would allow any caller to modify, and creates the
opportunity for inconsistent handling of what is considered a valid hard
fork (ie too old).
This PR adds a function to Bank so consistent sanity checks can be
applied; the caller will already have a Bank as that is where they would
have obtained the HardForks from in the first place. Additionally,
change the getter to return a copy of HardForks (simple Vec).
* Allow hard fork at bank slot if bank is not yet frozen
* Makes Bank::load_program() return correct tombstones.
* Removes early TX failure caused by closed and invalid programs.
* Adjusts the feature gate of simplify_writable_program_account_check.
`Arc` is already a reference internally, so it does not seem to be
beneficial to pass a reference to it. Just adds an extra layer of
indirection.
Functions that need to be able to increment `Arc` reference count need
to take `Arc<AtomicBool>`, but those that just want to read the
`AtomicBool` value can accept `&AtomicBool`, making them a bit more
generic.
This change focuses specifically on `Arc<AtomicBool>`. There are other
uses of `&Arc<T>` in the code base that could be converted in a similar
manner. But it would make the change even larger.
This function has morphed as it has been updated, and the existing
function name is a bit of misnomer. Update the function name to use
terminology that I think is more clear and more consistent with what we
use core/src/validator.rs. Additionally, adjust the error messages to be
more clear about what exactly is wrong for when we can't process.
The callstack updated in this PR passed an &Arc<...> down only to have
the bottom level clone the reference. Thus, we are giving shared
ownership so the reference is a bit redundant and arguably obscures the
intention to clone further down the callstack.
* Fix ledger-tool program command for non-runnable subcommands
These subcommands (cfg and disassemble) don't need to load a ledger
database for running, and neither do they need an input file
specifying instruction accounts etc. This change creates a simplified
invoke_context for processing the static actions of ledger-tool
program subcommand.
* Factor out program file loading
* Move entry_notifier_interface
* Add EntryNotifierService
* Use descriptive struct in sender/receiver
* Optionally initialize EntryNotifierService in validator
* Plumb EntryNotfierSender into Tvu, blockstore_processor
* Plumb EntryNotfierSender into Tpu
* Only return one option when constructing EntryNotifierService
* Add debug and trace capabilities to ledger-tool, previously rbpf-cli
* Dump nested logs from saved traces in invoke_context
* Change the ledger-tool run input format to represent keys as strings
* Enables JIT everywhere by default.
* Removes the CLI argument "no-bpf-jit" and its plumbering through out the validator code base.
* Removes with_jit bpf_loader variants.
* Removes the to_builtin!() macro.
Performing a manual ledger compaction can be very time consuming.
Although the compaction will optimize storage use, swapping to skip
compaction favors the command completing faster. Additionally, we no
longer do regular compaction in the validator (via LedgerCleanupService)
and instead make use of range deletion which is much more efficient.
Given the above, this changes alters the default behavior to skip ledger
compaction by default, and introduces a flag should a user really want
to run compaction.
PR #29975 previously combined some repeated logic around parsing
AccountsDb configuration options. That PR consolidated logic between
verify and create-snapshot, but left alone all of the other commands
that load up and process Bank(s).
Make all functions that will spin up Banks / AccountsDb share this
common function. Additionally, expose some of the AccountsDb
configuration flags to the other commands such as accounts, graph, etc.
In cluster restart scenarios, it is desirable to know the latest
optimistic slot(s), which the subcommand latest-optimistic-slots will
return. However, it is also useful to know whether slots contain only
votes or if they contain votes and user transactions.
This PR adds an extra column of output to show whether an optimistically
confirmed slot is vote only (contains zero non-vote transactions).
Additionally, a flag has been added to enable filtering output to
exclude vote only slots.
* Add --{replaying,banking}-backend flags (noop atm)
* Greatly simplify enums with strum
* Update programs/sbf/Cargo.lock...
* Rely on Display, removing Debug
* constify cli_names()
* Don't allow omitting bankend value
* Rename to --block-{verification,production}-method
* Use more specific name
* Actually support missing value....
* Remove strictly-unnecessary flags
* Use lazy_static! instead of abusing DefaultArgs...
`poh_verify` actually disables transaction signature, tick count and
built in program argument verifications as well. It is somewhat
confusing to call it `poh_verify`.
* Clean up orphaned account snapshot hardlink dirs
* fix compilation issues
* debugged, now working. seeing the orphaned directories deleted
* change back to eprintln + exit for account_path error
* changed eprintln to panic for now
* add test_clean_orphaned_account_snapshot_dirs for codecov check
* address a few comments and nit isseus
* directly unzip, skipped the intermediate array of tuples
* let set_up_account_run_and_snapshot_paths return Result
* 'proper' typo, and comment on return
* use map_err
* use for loop in clean_orphaned_account_snapshot_dirs, removed panic
* add test_set_up_account_run_and_snapshot_paths
* minor, replace .for_each with .all
* rename set_up_account_run_and_snapshot_paths to create_all_accounts_run_and_snapshot_dirs
* remove unnecessary closure return type
* change to for loop
* change match to unwrap_or_else
* remove create_dir_all(&account_path) in create_all
* minor comment cleanup
Declare constants at top of function with comment, and stringify the
constants when needed to be in line with how other defaults are done in
ledger-tool main function.
The print command looks up program_ids from transactions by creating
SanitizedTransactions. But, print doesn't have access to a Bank so
transactions that include an address table lookup can't be sanitized.
Sanitizing the transaction isn't necessary, so this change just looks up
the program id from unsanitized transactions.
* introduce workspace.package
* introduce workspace.dependencies
* read version from root cargo.toml
* pass check when version = { workspace = true }
* don't bump version when version = { workspace = true }
* including workspace Cargo.toml when bump version
* programs/sbf use workspace inheritance
* fix increasing cargo version ignore program/sbf/Cargo.toml
The previous change used Primary access; however, it is better to use
PrimaryForMaintenance in this case to disallow any potential compaction
from happening.
* Add address lookup tables to minimized snapshot
* Add comment for future posterity
* Add reference to the issue
* Adjust comment a bit
Co-authored-by: Andrew Fitzgerald <apfitzge@gmail.com>
---------
Co-authored-by: Andrew Fitzgerald <apfitzge@gmail.com>
Previously, ledger-tool had a guardrail to disallow a custom accounts
path when access mode to the blockstore was Secondary. This was to avoid
potentially pulling the accounts out from underneath solana-validator.
When ledger-tool switched over to use Secondary blockstore access for
all commands that do not need write access, this removed the ability to
use custom accounts paths with ledger-tool at all for these commands.
Custom accounts paths are desirable, especially if that custom path is
in tmpfs to speed up processing.
With this change, when a custom accounts path is passed for a command
using Secondary access, ledger-tool now checks if Primary access is
being held by another process. If not, allow processing to proceed
with the custom accounts path.
The above check isn't fullproof, but it is about equal to the check that
previously existed when ledger-tool would run in Primary access mode
when it didn't need to.
Brooks
Alexander Meißner
steviez
Jeff Washington (jwash)
Justin Starry
Ryo Onodera
Andrew Fitzgerald
Pankaj Garg
Alessandro Decina
Jon Cinque
HaoranYi
behzad nouri
Illia Bobyr
Trent Nelson
Dmitri Makarov
Tyera
Xiang Zhu
cavemanloverboy
Yihau Chen
Michael Vines
Benno Fünfstück