Fix deadlock with CancellationChecker

[antonio2368]

Changelog category (leave one):

• Bug Fix (user-visible misbehavior in an official stable release)

Changelog entry (a user-readable short description of the changes that goes into CHANGELOG.md):

Fix deadlock caused by background cancellation checker thread.

Deadlock can happen in following situation:
T1: ProcessListEntry in its destructor takes ProcessList::mutex and wants to mark task as done in CancellationChecker but CancellationChecker::mutex is taken by T2.
T2: CancellationChecker::workerFunction took CancellationChecker::mutex and started cancelling tasks. Cancellation waits on ExecutingGraph::mutex to cancel the task but it's taken by T3.
T3: ExecutingGraph::cancel was called and it took ExecutingGraph::mutex. Inside cancel, OverCommitTracker is triggered which tries to take ProcessList::mutex but it's already taken by T1.

Documentation entry for user-facing changes

• Documentation is written (mandatory for new features)

[antonio2368]

here is the fix, we cancel tasks without lock

[clickhouse-gh]

Workflow [PR], commit [c90cc4e]

Summary: ❌

job_name	test_name	status	info	comment
Unit tests (tsan)		error
Stress test (amd_debug)		failure
	Server died	FAIL
	Hung check failed, possible deadlock found (see hung_check.log)	FAIL
	Killed by signal (in clickhouse-server.log)	FAIL
	Fatal message in clickhouse-server.log (see fatal_messages.txt)	FAIL
	Killed by signal (output files)	FAIL
	Found signal in gdb.log	FAIL
Finish Workflow		failure
	python3 ./ci/jobs/scripts/workflow_hooks/new_tests_check.py	failure