At least some users report seeing that error on the llw0 device on macOS
Catalina if the Wi-Fi is off; don't let that cause pcap_findalldevs() to
fail.
See
https://ask.wireshark.org/question/14711/need-debugging-suggestions-no-longer-see-ethernet-interfaces-after-upgrading-mac-from-10151-to-10153/
and
https://ask.wireshark.org/question/13572/updated-catalina-no-interfaces-found/
strncpy(req.ifm_name, name, sizeof(req.ifm_name));
if (ioctl(sock, SIOCGIFMEDIA, &req) < 0) {
if (errno == EOPNOTSUPP || errno == EINVAL || errno == ENOTTY ||
strncpy(req.ifm_name, name, sizeof(req.ifm_name));
if (ioctl(sock, SIOCGIFMEDIA, &req) < 0) {
if (errno == EOPNOTSUPP || errno == EINVAL || errno == ENOTTY ||
- errno == ENODEV || errno == EPERM) {
+ errno == ENODEV || errno == EPERM
+#ifdef EPWROFF
+ || errno == EPWROFF
+#endif
+ ) {
/*
* Not supported, so we can't provide any
* additional information. Assume that
/*
* Not supported, so we can't provide any
* additional information. Assume that
* So, just as we do for some ethtool ioctls
* on Linux, which makes the same mistake, we
* also treat EPERM as meaning "not supported".
* So, just as we do for some ethtool ioctls
* on Linux, which makes the same mistake, we
* also treat EPERM as meaning "not supported".
+ *
+ * And it appears that Apple's llw0 device, which
+ * appears to be part of the Skywalk subsystem:
+ *
+ * http://newosxbook.com/bonus/vol1ch16.html
+ *
+ * can sometimes return EPWROFF ("Device power
+ * is off") for that ioctl, so we treat *that*
+ * as another indication that we can't get a
+ * connection status. (If it *isn't* "powered
+ * off", it's reported as a wireless device,
+ * complete with an active/inactive state.)
*/
*flags |= PCAP_IF_CONNECTION_STATUS_NOT_APPLICABLE;
close(sock);
*/
*flags |= PCAP_IF_CONNECTION_STATUS_NOT_APPLICABLE;
close(sock);