This function used to contain feature gate activation checks that
required access to a bank. Those checks have been cleaned up, so we no
longer need access to a full Bank. Rather, we can momentarily get a Bank
from BankForks, calculate the necessary results and then drop the Bank
along with the BankForks read lock.
- BankForks is not an optional argument, so remove dated comment
- Given that BankForks is always present, no need for special values to
initialize variables before the loop
- Root slot can be retrieved from root bank, no need to call
BankForks::root() which will load the underlying atomic a second time
- Use BankForks::highest_slot() instead of .slot() on .working_bank() to
avoid the extra clone that .working_bank() performs
- Move several operations outside of BankForks read lock scope to
minimize lock time
* separates out turbine QUIC from TPU implementation
Turbine being tied to QUIC implementation for TPU hinders development
and makes it hard to optimize QUIC specifically for turbine.
The commit separates out turbine QUIC from TPU implementation.
* Update core/src/validator.rs
Co-authored-by: Jon Cinque <me@jonc.dev>
* Update turbine/src/retransmit_stage.rs
Co-authored-by: Jon Cinque <me@jonc.dev>
---------
Co-authored-by: Jon Cinque <me@jonc.dev>
`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.
Shreds arriving at tvu/tvu_forward/repair sockets are each processed in
a separate thread, and since each thread has its own deduper, the
duplicates across these sockets are not filtered out.
Using a common deduper across these threads will require an RwLock
wrapper and may introduce lock contention.
The commit instead moves the shred-deduper to shred-sigverify-stage
where all these shreds arrive through the same channel.
If turbine_disabled is true, the commit discards turbine packets
earlier in the pipeline so that they won't interfere with the deduper
and the packets can get through once turbine is enabled again.
This is a prerequisite of:
https://github.com/solana-labs/solana/pull/30786
so that local-cluster tests pass.
generalizes Deduper to work with any hashable type
Current Deduper is hard-coded only for Packet type. In order to use
Deduper in retransmit-stage, we need to dedup types other than Packet.
The commit generalizes Deduper to any hashable type.
removes the false_positive_rate field from the Deduper
Deduper.false_positive_rate field is misleading because it is not
enforced until maybe_reset is called. But then maybe_reset can be
invoked with an explicit argument.
Tenets:
1. Limit thread names to 15 characters
2. Prefix all Solana-controlled threads with "sol"
3. Use Camel case. It's more character dense than Snake or Kebab case
Given the 32:32 erasure recovery schema, current implementation requires
exactly 32 data shreds to generate coding shreds for the batch (except
for the final erasure batch in each slot).
As a result, when serializing ledger entries to data shreds, if the
number of data shreds is not a multiple of 32, the coding shreds for the
last batch cannot be generated until there are more data shreds to
complete the batch to 32 data shreds. This adds latency in generating
and broadcasting coding shreds.
In addition, with Merkle variants for shreds, data shreds cannot be
signed and broadcasted until coding shreds are also generated. As a
result *both* code and data shreds will be delayed before broadcast if
we still require exactly 32 data shreds for each batch.
This commit instead always generates and broadcast coding shreds as soon
as there any number of data shreds available. When serializing entries
to shreds:
* if the number of resulting data shreds is less than 32, then more
coding shreds will be generated so that the resulting erasure batch
has the same recovery probabilities as a 32:32 batch.
* if the number of data shreds is more than 32, then the data shreds are
split uniformly into erasure batches with _at least_ 32 data shreds in
each batch. Each erasure batch will have the same number of code and
data shreds.
For example:
* If there are 19 data shreds, 27 coding shreds are generated. The
resulting 19(data):27(code) erasure batch has the same recovery
probabilities as a 32:32 batch.
* If there are 107 data shreds, they are split into 3 batches of 36:36,
36:36 and 35:35 data:code shreds each.
A consequence of this change is that code and data shreds indices will
no longer align as there will be more coding shreds than data shreds
(not only in the last batch in each slot but also in the intermediate
ones);
Shreds have different workload and traffic pattern from TPU vote and
transaction packets. Some of recent changes to SigVerifyStage are not
suitable or at least optimal for shreds sig-verify; e.g. random discard,
dedup with false positives, discard excess by IP-address, ...
SigVerifier trait is meant to abstract out the distinctions between the
two pipelines, but in practice it has led to more verbose and convoluted
code.
This commit discards SigVerifier implementation for shreds sig-verify
and instead provides a standalone stage for verifying shreds signatures.
Shred slot and parent are not verified until window-service where
resources are already wasted to sig-verify and deserialize shreds.
This commit moves above verification to earlier in the pipeline in fetch
stage.
Shred versions are not verified until window-service where resources are
already wasted to sig-verify and deserialize shreds.
The commit verifies shred-version earlier in the pipeline in fetch stage.
Upcoming changes to PacketBatch to support variable sized packets will
modify the internals of PacketBatch. So, this change removes usage of
the internal packet struct and instead uses accessors (which are
currently just wrappers of Vector functions but will change down the
road).
The indices for erasure coding shreds are tied to data shreds:
https://github.com/solana-labs/solana/blob/90f41fd9b/ledger/src/shred.rs#L921
However with the upcoming changes to erasure schema, there will be more
erasure coding shreds than data shreds and we can no longer infer coding
shreds indices from data shreds.
The commit adds constructs to track coding shreds indices explicitly.