Cygwin: pipe: Fix a regression that raw_write() slows down

After the commit 7f3c225325, writing to pipe extremely slows down.
This is because cygwait(select_sem, 10, cw_cancel) is called even
when write operation is already completed. With this patch, the
cygwait() is called only if the write operation is not completed.

Addresses: https://cygwin.com/pipermail/cygwin/2024-August/256398.html
Fixes: 7f3c225325 ("Cygwin: pipe: handle signals explicitely in raw_write")
Reported-by: Jim Reisert AD1C <jjreisert@alum.mit.edu>
Reviewed-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>

winsup/cygwin/fhandler/pipe.cc

@@ -518,8 +518,9 @@ fhandler_pipe_fifo::raw_write (const void *ptr, size_t len)
 		      raise (SIGPIPE);
 		      goto out;
 		    }
-		  else
-		    cygwait (select_sem, 10, cw_cancel);
+		  /* Break out on completion */
+		  if (waitret == WAIT_OBJECT_0)
+		    break;
 		  /* If we got a timeout in the blocking case, and we already
 		     did a short write, we got a signal in the previous loop. */
 		  if (waitret == WAIT_TIMEOUT && short_write_once)
@@ -527,6 +528,7 @@ fhandler_pipe_fifo::raw_write (const void *ptr, size_t len)
 		      waitret = WAIT_SIGNALED;
 		      break;
 		    }
+		  cygwait (select_sem, 10, cw_cancel);
 		}
 	      /* Loop in case of blocking write or SA_RESTART */
 	      while (waitret == WAIT_TIMEOUT || waitret == WAIT_SIGNALED);

winsup/cygwin/release/3.5.5

@@ -4,3 +4,6 @@ Fixes:
 - Fix undesired behaviour of console master thread in win32-input-mode
   which is supported by Windows Termainal.
   Addresses: https://cygwin.com/pipermail/cygwin/2024-August/256380.html
+
+- Fix a regression in 3.5.4 that writing to pipe extremely slows down.
+  Addresses: https://cygwin.com/pipermail/cygwin/2024-August/256398.html