Hi On Tue, Nov 4, 2014 at 12:17 AM, Peter Hutterer <[email protected]> wrote: > On Mon, Nov 03, 2014 at 11:49:12AM +0100, David Herrmann wrote: >> Hi >> >> On Wed, Oct 29, 2014 at 3:56 PM, Derek Foreman <[email protected]> >> wrote: >> > Log a message when the kernel event queue overflows and events are dropped. >> > After 10 messages logging stops to avoid flooding the logs if the condition >> > is persistent. >> > --- >> > src/evdev.c | 11 +++++++++++ >> > src/evdev.h | 4 ++++ >> > 2 files changed, 15 insertions(+) >> > >> > diff --git a/src/evdev.c b/src/evdev.c >> > index 1b4ce10..9026f5c 100644 >> > --- a/src/evdev.c >> > +++ b/src/evdev.c >> > @@ -924,6 +924,17 @@ evdev_device_dispatch(void *data) >> > rc = libevdev_next_event(device->evdev, >> > LIBEVDEV_READ_FLAG_NORMAL, &ev); >> > if (rc == LIBEVDEV_READ_STATUS_SYNC) { >> > + if (device->syn_drops_received < 10) { >> > + device->syn_drops_received++; >> > + log_info(libinput, "SYN_DROPPED event from >> > " >> > + "\"%s\" - some input events have " >> > + "been lost.\n", device->devname); >> > + if (device->syn_drops_received == 10) >> > + log_info(libinput, "No longer >> > logging " >> > + "SYN_DROPPED events for " >> > + "\"%s\"\n", >> > device->devname); >> > + } >> > + >> >> I really appreciate those log-messages, but can we use rate-limiting >> here, rather than a per-device counter? I mean, my compositor usually >> runs for multiple days, or even weeks. I really don't want it to stop >> reporting SYN_DROPPED events just because it got 100 of those during >> startup (or something similar). >> >> Something like this: >> http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/lib/ratelimit.c >> >> I can prep a patch if you want. > > good idea, let's do that.
I sent 2 patches to wayland-devel. Thanks David _______________________________________________ wayland-devel mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/wayland-devel
