Hi,
if BIOCGRTIMEOUT is set, we should wait for the timeout and not just
set EWOULDBLOCK. From my point of view the whole 'if' is unnecessary.
If the User doesn't set a timeout we have to wait anyway with tsleep()
and loop again. Well, and if the timeout is set, we should tsleep()
for the set timeout.
Index: sys/net//bpf.c
===================================================================
RCS file: /home/cvs/src/sys/net/bpf.c,v
retrieving revision 1.103
diff -u -p -u -r1.103 bpf.c
--- sys/net//bpf.c 12 Jul 2014 18:44:22 -0000 1.103
+++ sys/net//bpf.c 17 Dec 2014 08:10:18 -0000
@@ -434,11 +434,9 @@ bpfread(dev_t dev, struct uio *uio, int
/* User requested non-blocking I/O */
error = EWOULDBLOCK;
} else {
- if ((d->bd_rdStart + d->bd_rtout) < ticks) {
- error = tsleep((caddr_t)d, PRINET|PCATCH, "bpf",
- d->bd_rtout);
- } else
- error = EWOULDBLOCK;
+ /* User requested timeout */
+ error = tsleep((caddr_t)d, PRINET|PCATCH, "bpf",
+ d->bd_rtout);
}
if (error == EINTR || error == ERESTART) {
D_PUT(d);
