libitm: Fix handling of reentrancy in the HTM fastpath.

PR libitm/57643
	* beginend.cc (gtm_thread::begin_transaction): Handle reentrancy in
	the HTM fastpath.

From-SVN: r200250
This commit is contained in:
Torvald Riegel 2013-06-20 16:40:38 +00:00 committed by Torvald Riegel
parent dcef09b5d1
commit b1db457bf9
2 changed files with 16 additions and 0 deletions

View File

@ -1,3 +1,9 @@
2013-06-20 Torvald Riegel <triegel@redhat.com>
PR libitm/57643
* beginend.cc (gtm_thread::begin_transaction): Handle reentrancy in
the HTM fastpath.
2013-03-31 Gerald Pfeifer <gerald@pfeifer.com>
PR bootstrap/56714

View File

@ -197,6 +197,8 @@ GTM::gtm_thread::begin_transaction (uint32_t prop, const gtm_jmpbuf *jb)
// We are executing a transaction now.
// Monitor the writer flag in the serial-mode lock, and abort
// if there is an active or waiting serial-mode transaction.
// Note that this can also happen due to an enclosing
// serial-mode transaction; we handle this case below.
if (unlikely(serial_lock.is_write_locked()))
htm_abort();
else
@ -219,6 +221,14 @@ GTM::gtm_thread::begin_transaction (uint32_t prop, const gtm_jmpbuf *jb)
tx = new gtm_thread();
set_gtm_thr(tx);
}
// Check whether there is an enclosing serial-mode transaction;
// if so, we just continue as a nested transaction and don't
// try to use the HTM fastpath. This case can happen when an
// outermost relaxed transaction calls unsafe code that starts
// a transaction.
if (tx->nesting > 0)
break;
// Another thread is running a serial-mode transaction. Wait.
serial_lock.read_lock(tx);
serial_lock.read_unlock(tx);
// TODO We should probably reset the retry count t here, unless