On Thu, Aug 01, 2013 at 12:58:58PM +0200, John Crispin wrote: > On 01/08/13 13:02, Yousong Zhou wrote: > >Hi, this series tries to fix unintended failsafe triggering by initializing > >buttons' last_status with its actual value instead of always 0 (released). > >Otherwise devices with like TL-WR720N-v3 would always go into failsafe as > >some of > >its GPIO buttons are always in pressed state. It's weird that this issue > >never > >happened before. > >
Now I guess it's because the order of execution. Previously, we first start hotplug2 in preinit_essential, then insmod gpio-button-hotplug (20_device_fs_daemons). Shortly after that the first state event happened which would sync last_status and the actual state, but at which moment no signal handler (USR1) was yet installed until later by 30_failsafe_wait. Anyway, glad that this helps. ;) yousong > > Thank you, > > the bug has always been there. before procd it seems to have never > been triggered/noticed. > > John > _______________________________________________ > openwrt-devel mailing list > openwrt-devel@lists.openwrt.org > https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel