cygthread: suspend thread before terminating.

This addresses an extremely difficult to debug deadlock when running
under emulation on ARM64.

A relatively easy way to trigger this bug is to call `fork()`, then within the
child process immediately call another `fork()` and then `exit()` the
intermediate process.

It would seem that there is a "code emulation" lock on the wait thread at
this stage, and if the thread is terminated too early, that lock still exists
albeit without a thread, and nothing moves forward.

It seems that a `SuspendThread()` combined with a `GetThreadContext()`
(to force the thread to _actually_ be suspended, for more details see
https://devblogs.microsoft.com/oldnewthing/20150205-00/?p=44743)
makes sure the thread is "booted" from emulation before it is suspended.

Hopefully this means it won't be holding any locks or otherwise leave
emulation in a bad state when the thread is terminated.

Also, attempt to use `CancelSynchonousIo()` (as seen in `flock.cc`) to avoid
the need for `TerminateThread()` altogether.  This doesn't always work,
however, so was not a complete fix for the deadlock issue.

Addresses: https://cygwin.com/pipermail/cygwin-developers/2024-May/012694.html
Signed-off-by: Jeremy Drake <cygwin@jdrake.com>
This commit is contained in:
Jeremy Drake 2024-11-19 11:06:42 -08:00 committed by Corinna Vinschen
parent b2672642c1
commit b091b47b9e
3 changed files with 31 additions and 5 deletions

View File

@ -302,6 +302,20 @@ cygthread::terminate_thread ()
if (!inuse)
goto force_notterminated;
if (_my_tls._ctinfo != this)
{
CONTEXT context;
context.ContextFlags = CONTEXT_CONTROL;
/* SuspendThread makes sure a thread is "booted" from emulation before
it is suspended. As such, the emulator hopefully won't be in a bad
state (aka, holding any locks) when the thread is terminated. */
SuspendThread (h);
/* We need to call GetThreadContext, even though we don't care about the
context, because SuspendThread is asynchronous and GetThreadContext
will make sure the thread is *really* suspended before returning */
GetThreadContext (h, &context);
}
TerminateThread (h, 0);
WaitForSingleObject (h, INFINITE);
CloseHandle (h);

View File

@ -1252,13 +1252,17 @@ proc_waiter (void *arg)
for (;;)
{
DWORD nb;
DWORD nb, err;
char buf = '\0';
if (!ReadFile (vchild.rd_proc_pipe, &buf, 1, &nb, NULL)
&& GetLastError () != ERROR_BROKEN_PIPE)
&& (err = GetLastError ()) != ERROR_BROKEN_PIPE)
{
system_printf ("error on read of child wait pipe %p, %E", vchild.rd_proc_pipe);
/* ERROR_OPERATION_ABORTED is expected due to the possibility that
CancelSynchronousIo interruped the ReadFile call, so don't output
that error */
if (err != ERROR_OPERATION_ABORTED)
system_printf ("error on read of child wait pipe %p, %E", vchild.rd_proc_pipe);
break;
}

View File

@ -409,7 +409,11 @@ proc_terminate ()
to 1 iff it is a Cygwin process. */
if (!have_execed || !have_execed_cygwin)
chld_procs[i]->ppid = 1;
if (chld_procs[i].wait_thread)
/* Attempt to exit the wait_thread cleanly via CancelSynchronousIo
before falling back to the (explicitly dangerous) cross-thread
termination */
if (chld_procs[i].wait_thread
&& !CancelSynchronousIo (chld_procs[i].wait_thread->thread_handle ()))
chld_procs[i].wait_thread->terminate_thread ();
/* Release memory associated with this process unless it is 'myself'.
'myself' is only in the chld_procs table when we've execed. We
@ -1174,7 +1178,11 @@ remove_proc (int ci)
{
if (have_execed)
{
if (_my_tls._ctinfo != chld_procs[ci].wait_thread)
/* Attempt to exit the wait_thread cleanly via CancelSynchronousIo
before falling back to the (explicitly dangerous) cross-thread
termination */
if (_my_tls._ctinfo != chld_procs[ci].wait_thread
&& !CancelSynchronousIo (chld_procs[ci].wait_thread->thread_handle ()))
chld_procs[ci].wait_thread->terminate_thread ();
}
else if (chld_procs[ci] && chld_procs[ci]->exists ())