* [PATCH v2 0/2] Make terminal read() thread-safe. @ 2021-01-28 12:20 Takashi Yano 2021-01-28 12:20 ` [PATCH v2 1/2] Cygwin: console: Make " Takashi Yano 2021-01-28 12:20 ` [PATCH v2 2/2] Cygwin: pty: Make slave " Takashi Yano 0 siblings, 2 replies; 4+ messages in thread From: Takashi Yano @ 2021-01-28 12:20 UTC (permalink / raw) To: cygwin-patches Currently read() for console and pty slave are somehow not thread-safe. These patches fix the issue. Takashi Yano (2): Cygwin: console: Make read() thread-safe. Cygwin: pty: Make slave read() thread-safe. winsup/cygwin/fhandler_console.cc | 10 +++++----- winsup/cygwin/fhandler_tty.cc | 6 ++++++ 2 files changed, 11 insertions(+), 5 deletions(-) -- 2.30.0 ^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH v2 1/2] Cygwin: console: Make read() thread-safe. 2021-01-28 12:20 [PATCH v2 0/2] Make terminal read() thread-safe Takashi Yano @ 2021-01-28 12:20 ` Takashi Yano 2021-01-28 12:24 ` Takashi Yano 2021-01-28 12:20 ` [PATCH v2 2/2] Cygwin: pty: Make slave " Takashi Yano 1 sibling, 1 reply; 4+ messages in thread From: Takashi Yano @ 2021-01-28 12:20 UTC (permalink / raw) To: cygwin-patches - Currently read() is somehow not thread-safe. This patch fixes the issue. --- winsup/cygwin/fhandler_console.cc | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/winsup/cygwin/fhandler_console.cc b/winsup/cygwin/fhandler_console.cc index 02d0ac052..d0e5bb33a 100644 --- a/winsup/cygwin/fhandler_console.cc +++ b/winsup/cygwin/fhandler_console.cc @@ -530,12 +530,14 @@ fhandler_console::read (void *pv, size_t& buflen) if (wincap.has_con_24bit_colors ()) request_xterm_mode_input (true, &handle_set); + acquire_input_mutex (INFINITE); while (!input_ready && !get_cons_readahead_valid ()) { int bgres; if ((bgres = bg_check (SIGTTIN)) <= bg_eof) { buflen = bgres; + release_input_mutex (); return; } @@ -552,6 +554,7 @@ wait_retry: pthread::static_cancel_self (); /*NOTREACHED*/ case WAIT_TIMEOUT: + release_input_mutex (); set_sig_errno (EAGAIN); buflen = (size_t) -1; return; @@ -572,9 +575,7 @@ wait_retry: int ret; acquire_attach_mutex (INFINITE); - acquire_input_mutex (INFINITE); ret = process_input_message (); - release_input_mutex (); release_attach_mutex (); switch (ret) { @@ -595,7 +596,6 @@ wait_retry: } /* Check console read-ahead buffer filled from terminal requests */ - acquire_input_mutex (INFINITE); while (con.cons_rapoi && *con.cons_rapoi && buflen) { buf[copied_chars++] = *con.cons_rapoi++; @@ -615,11 +615,13 @@ wait_retry: return; err: + release_input_mutex (); __seterrno (); buflen = (size_t) -1; return; sig_exit: + release_input_mutex (); set_sig_errno (EINTR); buflen = (size_t) -1; } @@ -984,9 +986,7 @@ fhandler_console::process_input_message (void) if (toadd) { ssize_t ret; - release_input_mutex (); line_edit_status res = line_edit (toadd, nread, *ti, &ret); - acquire_input_mutex (INFINITE); if (res == line_edit_signalled) { stat = input_signalled; -- 2.30.0 ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v2 1/2] Cygwin: console: Make read() thread-safe. 2021-01-28 12:20 ` [PATCH v2 1/2] Cygwin: console: Make " Takashi Yano @ 2021-01-28 12:24 ` Takashi Yano 0 siblings, 0 replies; 4+ messages in thread From: Takashi Yano @ 2021-01-28 12:24 UTC (permalink / raw) To: cygwin-patches On Thu, 28 Jan 2021 21:20:09 +0900 Takashi Yano wrote: > - Currently read() is somehow not thread-safe. This patch fixes the > issue. > --- > winsup/cygwin/fhandler_console.cc | 10 +++++----- > 1 file changed, 5 insertions(+), 5 deletions(-) Sorry again. This also has a problem...... -- Takashi Yano <takashi.yano@nifty.ne.jp> ^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH v2 2/2] Cygwin: pty: Make slave read() thread-safe. 2021-01-28 12:20 [PATCH v2 0/2] Make terminal read() thread-safe Takashi Yano 2021-01-28 12:20 ` [PATCH v2 1/2] Cygwin: console: Make " Takashi Yano @ 2021-01-28 12:20 ` Takashi Yano 1 sibling, 0 replies; 4+ messages in thread From: Takashi Yano @ 2021-01-28 12:20 UTC (permalink / raw) To: cygwin-patches - Currently slave read() is somehow not thread-safe. This patch fixes the issue. --- winsup/cygwin/fhandler_tty.cc | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/winsup/cygwin/fhandler_tty.cc b/winsup/cygwin/fhandler_tty.cc index 06fc19ac2..48b89ae77 100644 --- a/winsup/cygwin/fhandler_tty.cc +++ b/winsup/cygwin/fhandler_tty.cc @@ -1241,6 +1241,7 @@ fhandler_pty_slave::read (void *ptr, size_t& len) time_to_wait = !vtime ? INFINITE : 100 * vtime; } +wait_retry: while (len) { switch (cygwait (input_available_event, time_to_wait)) @@ -1319,6 +1320,11 @@ fhandler_pty_slave::read (void *ptr, size_t& len) } goto out; } + if (!IsEventSignalled (input_available_event)) + { /* Maybe another thread has processed input. */ + ReleaseMutex (input_mutex); + goto wait_retry; + } if (!bytes_available (bytes_in_pipe)) { -- 2.30.0 ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2021-01-28 12:24 UTC | newest] Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2021-01-28 12:20 [PATCH v2 0/2] Make terminal read() thread-safe Takashi Yano 2021-01-28 12:20 ` [PATCH v2 1/2] Cygwin: console: Make " Takashi Yano 2021-01-28 12:24 ` Takashi Yano 2021-01-28 12:20 ` [PATCH v2 2/2] Cygwin: pty: Make slave " Takashi Yano
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).