> -----Original Message----- > From: Don Cohen [mailto:don-linux77@isis.cs3-inc.com] > Sent: Saturday, May 31, 2003 7:20 AM > To: linux-net@vger.kernel.org > Subject: from where can you call dev_set_promiscuity ? > > > My debugging attempts lead me to believe that dev_set_promiscuity > causes crashes when called from softirq in a 2.4.18 kernel compiled > with SMP. Not with UP. > Does this expected? A known bug? > Or is dev_set_promiscuity not supposed to be called from softirq code? > Or is it supposed to be called with some lock? > (Does it sleep? Does that depend on the driver?) By "crashes", do you mean a panic or a hang ? We've seen a system hang happen when using tcpdump. The tool would put the device in promiscuous mode and then take it out upon terminating. At that point we sometimes get a system hang if traffic is running, that when analyzing with kdb seems to be a deadlock between the br_lock and the device's xmit_lock being held in opposing order. It seems that it must be called only after disabling bottom halves in order to prevent other softirqs from running. -- | Shmulik Hen | | Israel Design Center (Jerusalem) | | LAN Access Division | | Intel Communications Group, Intel corp. | - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
