public inbox for cygwin-cvs@sourceware.org
help / color / mirror / Atom feed
* [newlib-cygwin] Cygwin: FIFO: hit_eof: add a call to fifo_client_lock
@ 2019-04-16 11:15 Corinna Vinschen
0 siblings, 0 replies; only message in thread
From: Corinna Vinschen @ 2019-04-16 11:15 UTC (permalink / raw)
To: cygwin-cvs
https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin.git;h=528169992668a7e292b8c1c5a39e2c1e24ff6683
commit 528169992668a7e292b8c1c5a39e2c1e24ff6683
Author: Ken Brown <kbrown@cornell.edu>
Date: Sun Apr 14 19:15:56 2019 +0000
Cygwin: FIFO: hit_eof: add a call to fifo_client_lock
The second check of nconnected needs to be protected by a lock as well
as the first.
Diff:
---
winsup/cygwin/fhandler_fifo.cc | 21 +++++++++++++--------
1 file changed, 13 insertions(+), 8 deletions(-)
diff --git a/winsup/cygwin/fhandler_fifo.cc b/winsup/cygwin/fhandler_fifo.cc
index 9016964..70551eb 100644
--- a/winsup/cygwin/fhandler_fifo.cc
+++ b/winsup/cygwin/fhandler_fifo.cc
@@ -705,15 +705,20 @@ fhandler_fifo::raw_write (const void *ptr, size_t len)
bool
fhandler_fifo::hit_eof ()
{
- fifo_client_lock ();
- bool eof = (nconnected == 0);
- fifo_client_unlock ();
- if (eof)
- {
- /* Give the listen_client thread time to catch up, then recheck. */
- Sleep (1);
+ bool eof;
+ bool retry = true;
+
+retry:
+ fifo_client_lock ();
eof = (nconnected == 0);
- }
+ fifo_client_unlock ();
+ if (eof && retry)
+ {
+ retry = false;
+ /* Give the listen_client thread time to catch up. */
+ Sleep (1);
+ goto retry;
+ }
return eof;
}
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2019-04-16 11:15 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-04-16 11:15 [newlib-cygwin] Cygwin: FIFO: hit_eof: add a call to fifo_client_lock Corinna Vinschen
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).