andreacavalli/rocksdb
0
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Assert valid linked list for write group (#7375)

Browse Source 
Summary:
We've seen some segfaults in db_write_test, with at least one
suggesting corruption of a write group linked list. Adding an assertion
to have this fail in a more specific way if that is the broken
invariant.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/7375

Test Plan: make check

Reviewed By: jay-zhuang

Differential Revision: D23638477

Pulled By: pdillinger

fbshipit-source-id: a76fd677cad60a3a516bd363947bfd9ce418edc1
This commit is contained in:
Peter Dillinger 2020-09-11 07:55:00 -07:00 committed by Facebook GitHub Bot
parent c4e2066dbd
commit be8445eea8
1 changed files with 4 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
db/write_thread.cc
View File

@ -438,6 +438,7 @@ size_t WriteThread::EnterAsBatchGroupLeader(Writer* leader,
// (newest_writer) is inclusive. Iteration goes from old to new.
Writer* w = leader;
while (w != newest_writer) {
assert(w->link_newer);
w = w->link_newer;
if (w->sync && !leader->sync) {
@ -512,6 +513,7 @@ void WriteThread::EnterAsMemTableWriter(Writer* leader,
Writer* w = leader;
while (w != newest_writer) {
assert(w->link_newer);
w = w->link_newer;
if (w->batch == nullptr) {
@ -568,6 +570,7 @@ void WriteThread::ExitAsMemTableWriter(Writer* /*self*/,
if (w == last_writer) {
break;
}
assert(next);
w = next;
}
// Note that leader has to exit last, since it owns the write group.
@ -721,6 +724,7 @@ void WriteThread::ExitAsBatchGroupLeader(WriteGroup& write_group,
// leader now
while (last_writer != leader) {
assert(last_writer);
last_writer->status = status;
// we need to read link_older before calling SetState, because as soon
// as it is marked committed the other thread's Await may return and