Fixes the deadlock noted by the "TODO this function will hang if it's
called twice" markers in Tree.Disable and Tree.Rebuild. Both paths walk
every layer in the tree and unconditionally invoke stopGeneration on
each disk layer, so a layer reachable through both must tolerate a
second call.
The previous implementation sent on the unbuffered genAbort channel
whenever genMarker was non-nil. After the first abort handshake the
generator goroutine exits, but genMarker is only cleared on the
successful-completion path (generate.go), not on the aborted-mid-flight
path. A second stopGeneration therefore saw generating=true, sent on
genAbort with no receiver, and blocked forever.
Wrap the abort handshake in a sync.Once on the disk layer. The first
call drives the handshake exactly as before; subsequent calls are
no-ops. Remove both TODO comments now that the contract is honoured.
Adds TestStopGenerationIdempotent: stands in a mock generator goroutine
that consumes a single abort, calls stopGeneration twice, and fails
with an explicit message at a 5s deadline rather than hanging until
the test runner's outer timeout.
Fixes#33233.
This pull request removes the destruct flag from the state snapshot to
simplify the code.
Previously, this flag indicated that an account was removed during a
state transition, making all associated storage slots inaccessible.
Because storage deletion can involve a large number of slots, the actual
deletion is deferred until the end of the process, where it is handled
in batches.
With the deprecation of self-destruct in the Cancun fork, storage
deletions are no longer expected. Historically, the largest storage
deletion event in Ethereum was around 15 megabytes—manageable in memory.
In this pull request, the single destruct flag is replaced by a set of
deletion markers for individual storage slots. Each deleted storage slot
will now appear in the Storage set with a nil value.
This change will simplify a lot logics, such as storage accessing,
storage flushing, storage iteration and so on.
#29995 has been reverted due to an unexpected flaw in the state snapshot
process.
Specifically, it attempts to stop the state snapshot generation, which
could potentially
cause the system to halt if the generation is not currently running.
This pull request ports the changes made in #29995 and fixes the flaw.
