-
Notifications
You must be signed in to change notification settings - Fork 1.1k
Memory usage limited for outgoing eth responses #1057
Conversation
@mkalinin ready for review and merge |
a3535b0#diff-c66ab184b22490e4ae62cc3f62b9b234R44 is a bit non-effective with memory, but we could enhance it later. Usually I see not a big number of pending txs on Mainnet, so slicing is not really used. |
@@ -423,4 +425,23 @@ public boolean equals(Object o) { | |||
public int hashCode() { | |||
return Arrays.hashCode(getHash()); | |||
} | |||
|
|||
public static final MemSizeEstimator<BlockHeader> MemEstimator = bh -> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Seems unused. Let's remove it
|
||
Iterator<byte[]> bodiesIterator = blockchain.getIteratorOfBodiesByHashes(msg.getBlockHashes()); | ||
List<byte[]> bodies = new ArrayList<>(); | ||
int sizeSum = 80; // ArrayList skeleton |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
i) it's a magic number ii) does 80
bytes matter much with regard to 32Kb
limit?
msg.getBlockIdentifier(), | ||
msg.getSkipBlocks(), | ||
min(msg.getMaxHeaders(), MAX_HASHES_TO_SEND), | ||
msg.isReverse() | ||
); | ||
|
||
BlockHeadersMessage response = new BlockHeadersMessage(headers); | ||
BlockHeadersMessage response = new BlockHeadersMessage(headersIterator); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why do we need to change BlockHeadersMessage
?
|
||
/** | ||
* Sames as {@link #sendTransactions(List)} but input list is randomly sliced to | ||
* contain not more than {@link MAX_SAFE_TXS} if needed |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
{@link MAX_SAFE_TXS}
-> {@link #MAX_SAFE_TXS}
* contain not more than {@link MAX_SAFE_TXS} if needed | ||
* @param txs List of txs to send | ||
*/ | ||
public void sendTransactionsSafely(List<Transaction> txs) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Safely
sounds too abstract for me and confusing on the other hand (is regular sendTransactions
unsafe??)
|
||
return rlpEncoded; | ||
} | ||
|
||
public synchronized void purgeEncoded() { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it's not used anywhere except tests. I'd remove it therefore
@mkalinin done! |
for (byte[] blockHash : msg.getBlockHashes()) { | ||
Block block = blockchain.getBlockByHash(blockHash); | ||
if (block == null) continue; | ||
|
||
List<TransactionReceipt> blockReceipts = new ArrayList<>(); | ||
sizeSum += 0; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
😀
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
byte[] plainMsg = this.getEncoded(); | ||
return HashUtil.sha3(plainMsg); | ||
getEncoded(); | ||
return hash; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This change causes an empty array to be returned when (hash == null && parsed == true && rlpEncoded != null). Happens all the time with the StandaloneBlockchain.
…rlpEncoded != null). See pull request ethereum#1057
…true && rlpEncoded != null). See pull request ethereum#1057" This reverts commit 6bce3d0.
…rlpEncoded != null). See pull request ethereum#1057
…true && rlpEncoded != null). See pull request ethereum#1057" This reverts commit 6bce3d0.
@mkalinin please, take a look