view tests/test-gendoc-ja.t @ 43709:039fbd14d4e2

lock: fix race in lock-breaking code

With low frequency, I see hg pulls fail with output like:
abort: no such file or directory: .hg/store/lock

I think what happens is, in lock.py, in:

    def _testlock(self, locker):
        if not self._lockshouldbebroken(locker):
            return locker

        # if locker dead, break lock.  must do this with another lock
        # held, or can race and break valid lock.
        try:
            with lock(self.vfs, self.f + b'.break', timeout=0):
                self.vfs.unlink(self.f)
        except error.LockError:
            return locker

if a lock is breakable on disk, and two hg processes concurrently get
to the "if locker dead" comment, a possible interleaving is: process1
finishes executing the function and then process2 finishes executing
the function. If that happens, process2 will either get ENOENT in
self.vfs.unlink (resulting in the spurious failure above), or break a
valid lock and potentially cause repository corruption.

The fix is simple enough: make sure the lock is breakable _inside_ the
critical section, because only then can we know that no other process
can invalidate our knowledge on the lock on disk.

I don't think there are tests for this. I've tested this manually
with:

diff --git a/mercurial/lock.py b/mercurial/lock.py
--- a/mercurial/lock.py
+++ b/mercurial/lock.py
@@ -351,6 +351,8 @@ class lock(object):
         if not self._lockshouldbebroken(locker):
             return locker

+        import random
+        time.sleep(1. + random.random())
         # if locker dead, break lock.  must do this with another lock
         # held, or can race and break valid lock.
         try:
@@ -358,6 +360,7 @@ class lock(object):
                 self.vfs.unlink(self.f)
         except error.LockError:
             return locker
+        time.sleep(1)

     def testlock(self):
         """return id of locker if lock is valid, else None.

and I see this change of behavior before/after this commit:

 $ $hg init repo
 $ cd repo
 $ ln -s $HOSTNAME/effffffc:987654321 .hg/wlock
 $ touch a
 $ $hg commit -Am_ & $hg commit -Am _; wait
-abort: No such file or directory: '/tmp/repo/.hg/wlock'
 adding a
+warning: ignoring unknown working parent 679a8959a8ca!
+nothing changed

Differential Revision: https://phab.mercurial-scm.org/D7199

#require docutils gettext

  $ $TESTDIR/check-gendoc ja
  checking for parse errors
  \xe8\xa4\x87\xe8\xa3\xbd\xef\xbc\x8f\xe6\x94\xb9\xe5\x90\x8d\xe3\x81\xab\xe9\x96\xa2\xe3\x81\x99\xe3\x82\x8b\xe5\xb1\xa5\xe6\xad\xb4\xe4\xbf\x9d\xe5\xad\x98\xe3\x81\xae\xe3\x81\x9f\xe3\x82\x81\xe3\x81\xab git \xe5\xb7\xae\xe5\x88\x86\xe5\xbd\xa2\xe5\xbc\x8f\xe3\x82\x92\xe4\xbd\xbf\xe7\x94\xa8(-g/--git \xe6\x8c\x87\xe5\xae\x9a\xe3\x82\x84 \xe8\xa8\xad\xe5\xae\x9a\xe3\x83\x95\xe3\x82\xa1\xe3\x82\xa4\xe3\x83\xab\xe3\x81\xa7\xe3\x81\xae [diff] git=1 \xe8\xa8\x98\xe8\xbf\xb0)\xe3\x81\x99\xe3\x82\x8b\xe3\x81\xae\xe3\x81\xa7\xe3\x81\x82\xe3\x82\x8c\xe3\x81\xb0\xe3\x80\x81 add/remove/copy/rename \xe3\x81\xa8\xe3\x81\x84\xe3\x81\xa3\xe3\x81\x9f hg \xe3\x81\xae\xe3\x82\xb3\xe3\x83\x9e\xe3\x83\xb3\xe3\x83\x89\xe3\x81\xab\xe3\x82\x88\xe3\x82\x8b\xe5\xb1\xa5\xe6\xad\xb4\xe8\xa8\x98\xe9\x8c\xb2\xe3\x82\x82\xe3\x80\x81 \xe9\x80\x9a\xe5\xb8\xb8\xe3\x81\xa8\xe5\xa4\x89\xe3\x82\x8f\xe3\x82\x8a\xe3\x81\xaa\xe3\x81\x8f\xe6\xa9\x9f\xe8\x83\xbd\xe3\x81\x97\xe3\x81\xbe\xe3\x81\x99\xe3\x80\x82 git \xe5\xb7\xae\xe5\x88\x86\xe5\xbd\xa2\xe5\xbc\x8f\xe3\x81\xae\xe8\xa9\xb3\xe7\xb4\xb0\xe3\x81\xab\xe9\x96\xa2\xe3\x81\x97\xe3\x81\xa6\xe3\x81\xaf\xe3\x80\x81 'help diffs' \xe3\x82\x92\xe5\x8f\x82\xe7\x85\xa7\xe3\x81\x97\xe3\x81\xa6\xe3\x81\x8f\xe3\x81\xa0\xe3\x81\x95\xe3\x81\x84\xe3\x80\x82 (esc)
  warning: please use " instead of ' for hg ... "..."