eth/catalyst: reset to current header if chain is rewound (in dev mode) #27992
Conversation
There was a problem hiding this comment.
An edge case is if SetHead is invoked after the check you have added, and before the end of sealBlock. One change that needs to be made is adding a check that returns from sealBlock if the ForkChoiceStatus returned from ForkChoiceUpdatedV2 (on line 147) is engine.STATUS_SYNCING.
I'm looking to see if anything else needs to be done.
|
We will also want to check if |
Signed-off-by: jsvisa <delweng@gmail.com>
Signed-off-by: jsvisa <delweng@gmail.com>
jsvisa
force-pushed
the
eth-catalyst-reset
branch
from
b9864b8
to
dfd8a79
Compare
|
@jwasinger thanks for the advice, PTAL :) |
holiman
changed the title
holiman
changed the title
Co-authored-by: Martin Holst Swende <martin@swende.se>
Signed-off-by: jsvisa <delweng@gmail.com>
Signed-off-by: jsvisa <delweng@gmail.com>
…in rewind. move calculation of finalized block hash into its own function
Signed-off-by: jsvisa <delweng@gmail.com>
| finalizedHash := c.finalizedBlockHash(header.Number.Uint64()) | ||
| c.setCurrentState(header.Hash(), *finalizedHash) |
There was a problem hiding this comment.
finalizedHash can be nil, should have nil check here.
There was a problem hiding this comment.
Yes, it could happen in theory but should be extremely unlikely: a setHead which rolls back to before the latest finalized block would have to occur between the execution of line 147 and 148.
…e) (ethereum#27992) Signed-off-by: jsvisa <delweng@gmail.com> Co-authored-by: Jared Wasinger <j-wasinger@hotmail.com>
… dev mode) (ethereum#27992)" This reverts commit f796c8f.
… dev mode) (ethereum#27992)" This reverts commit f796c8f.
…e) (ethereum#27992) Signed-off-by: jsvisa <delweng@gmail.com> Co-authored-by: Jared Wasinger <j-wasinger@hotmail.com> (cherry picked from commit cde462c)
fix #27990