| Message ID | 1352391981-15692-1-git-send-email-steve.glendinning@shawell.net |
|---|---|
| State | Accepted, archived |
| Delegated to: | David Miller |
| Headers | show |
From: Steve Glendinning <steve.glendinning@shawell.net> Date: Thu, 8 Nov 2012 16:26:21 +0000 > when something goes wrong, a flood of these messages can be > generated by usbnet (thousands per second). This doesn't > generally *help* the condition so this patch ratelimits the > rate of their generation. > > There's an underlying problem in usbnet's kevent deferral > mechanism which needs fixing, specifically that events *can* > get dropped and not handled. This patch doesn't address this, > but just mitigates fallout caused by the current implemention. > > Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net> Applied, thanks. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thursday 08 November 2012 16:26:21 Steve Glendinning wrote: > when something goes wrong, a flood of these messages can be > generated by usbnet (thousands per second). This doesn't > generally *help* the condition so this patch ratelimits the > rate of their generation. > > There's an underlying problem in usbnet's kevent deferral > mechanism which needs fixing, specifically that events *can* > get dropped and not handled. This patch doesn't address this, > but just mitigates fallout caused by the current implemention. > > Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net> Acked-by: Oliver Neukum <oneukum@suse.de> -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thursday 08 November 2012 16:26:21 Steve Glendinning wrote: > There's an underlying problem in usbnet's kevent deferral This is imported from schedule_work(). > mechanism which needs fixing, specifically that events *can* > get dropped and not handled. This patch doesn't address this, > but just mitigates fallout caused by the current implemention. All is not lost, as the flag is still set. How about starting a timer in the failure case? It feels kind of dirty, but a solution with locks would leave open a window to a race in any case. Regards Oliver -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c index 5bf7717..3565e6d 100644 --- a/drivers/net/usb/usbnet.c +++ b/drivers/net/usb/usbnet.c @@ -359,10 +359,12 @@ static enum skb_state defer_bh(struct usbnet *dev, struct sk_buff *skb, void usbnet_defer_kevent (struct usbnet *dev, int work) { set_bit (work, &dev->flags); - if (!schedule_work (&dev->kevent)) - netdev_err(dev->net, "kevent %d may have been dropped\n", work); - else + if (!schedule_work (&dev->kevent)) { + if (net_ratelimit()) + netdev_err(dev->net, "kevent %d may have been dropped\n", work); + } else { netdev_dbg(dev->net, "kevent %d scheduled\n", work); + } } EXPORT_SYMBOL_GPL(usbnet_defer_kevent);
when something goes wrong, a flood of these messages can be generated by usbnet (thousands per second). This doesn't generally *help* the condition so this patch ratelimits the rate of their generation. There's an underlying problem in usbnet's kevent deferral mechanism which needs fixing, specifically that events *can* get dropped and not handled. This patch doesn't address this, but just mitigates fallout caused by the current implemention. Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net> --- drivers/net/usb/usbnet.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)