[PATCH 2/2] kernel/cobalt/thread: Confine condition for sending SIGSHADOW on xnthread_suspend

From: Jan Kiszka <[email protected]>

In fact, XNDELAY is only self-applied. Drop it from the condition set
and also fix the comment explaining this as well as some other bits.

Signed-off-by: Jan Kiszka <[email protected]>
---
 kernel/cobalt/thread.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/kernel/cobalt/thread.c b/kernel/cobalt/thread.c
index b605781c64..5626e7655f 100644
--- a/kernel/cobalt/thread.c
+++ b/kernel/cobalt/thread.c
@@ -996,7 +996,7 @@ void xnthread_suspend(struct xnthread *thread, int mask,
 	 * assumptions regarding suspended threads.
 	 *
 	 * We only care for threads that are not current, and for
-	 * XNSUSP, XNDELAY, XNDORMANT and XNHELD conditions, because:
+	 * XNSUSP and XNHELD conditions, because:
 	 *
 	 * - There is no point in dealing with a relaxed thread which
 	 * is current, since personalities have to ask for primary
@@ -1005,15 +1005,15 @@ void xnthread_suspend(struct xnthread *thread, int mask,
 	 *
 	 * - among all blocking bits (XNTHREAD_BLOCK_BITS), only
 	 * XNSUSP, XNDELAY, XNHELD and XNDBGSTOP may be applied by the
-	 * current thread to a non-current thread. XNPEND is always
-	 * added by the caller to its own state, XNMIGRATE, XNRELAX
-	 * and XNDBGSTOP have special semantics escaping this issue.
+	 * current thread to a non-current thread. XNPEND, XNDELAY and
+	 * XNDORMANT are always added by the caller to its own state,
+	 * XNMIGRATE, XNRELAX and XNDBGSTOP have special semantics escaping this issue.
 	 *
 	 * We don't signal threads which are already in a dormant
 	 * state, since they are suspended by definition.
 	 */
 	if (((oldstate & (XNTHREAD_BLOCK_BITS|XNUSER)) == (XNRELAX|XNUSER)) &&
-	    (mask & (XNDELAY | XNSUSP | XNHELD)) != 0) {
+	    (mask & (XNSUSP | XNHELD)) != 0) {
 		if (mask & XNSUSP)
 			xnthread_set_info(thread, XNPNDSUSP);
 		if (mask & XNHELD)
-- 
2.47.3