I occasionally use rsync to transfer folders/files; however, it seems to cause the wlan interface to go down. I performed the same action, forcing IPv4, and it did not replicate the issue. Mar 18 17:01:55 localhost wpa_supplicant[37136]: wlan0: CTRL-EVENT-DISCONNECTED bssid=82:2a:a8:ab:cd:ef reason=0 Mar 18 17:01:55 localhost kernel: wlan0: link state changed to DOWN Mar 18 17:01:55 localhost kernel: rtwn0: Mar 18 17:01:55 localhost kernel: ieee80211_crypto_encap returns NULL. Mar 18 17:01:55 localhost kernel: . Mar 18 17:01:55 localhost syslogd: last message repeated 1 times Mar 18 17:01:56 localhost syslogd: exiting on signal 15 Mar 18 17:41:28 localhost syslogd: kernel boot file is /boot/kernel/kernel The work around is to perform an ACPI powerbutton off then powerbutton on. % kldstat Id Refs Address Size Name 1 25 0xffffffff80200000 243cd00 kernel 2 1 0xffffffff8263d000 82a0 if_tap.ko 3 1 0xffffffff82646000 1c930 if_rtwn_usb.ko 4 2 0xffffffff82663000 23710 rtwn.ko 5 1 0xffffffff82687000 3a9a10 zfs.ko 6 2 0xffffffff82a31000 a4f0 opensolaris.ko 7 1 0xffffffff82a3d000 566b8 pf.ko 8 1 0xffffffff82c21000 6fb3 if_bridge.ko 9 1 0xffffffff82c28000 4410 bridgestp.ko % sudo usbconfig -d ugen4.2 dump_device_desc ugen4.2: <Realtek Edimax AC1200 USB> at usbus4, cfg=0 md=HOST spd=HIGH (480Mbps) pwr=ON (500mA) bLength = 0x0012 bDescriptorType = 0x0001 bcdUSB = 0x0200 bDeviceClass = 0x0000 <Probed by interface class> bDeviceSubClass = 0x0000 bDeviceProtocol = 0x0000 bMaxPacketSize0 = 0x0040 idVendor = 0x7392 idProduct = 0xa822 bcdDevice = 0x0000 iManufacturer = 0x0001 <Realtek> iProduct = 0x0002 <Edimax AC1200 USB> iSerialNumber = 0x0003 <123456> bNumConfigurations = 0x0001 % uname -a FreeBSD localhost 12.0-RELEASE-p3 FreeBSD 12.0-RELEASE-p3 GENERIC amd64
Is there some changes after 1) ifconfig wlan0 -rxcsum -rxcsum6 2) kenv dev.rtwn.0.hwcrypto=0 && kldunload if_rtwn_usb && kldload if_rtwn_usb ?The work around is to perform an ACPI powerbutton off then powerbutton on. > The work around is to perform an ACPI powerbutton off then powerbutton on. kldunload if_rtwn_usb && kldload if_rtwn_usb ? (the last part is only needed if you don't have devd(8) running).
devd(8) is running. Mar 28 14:46:08 localhost syslogd: kernel boot file is /boot/kernel/kernel Mar 28 14:46:08 localhost kernel: rtwn0 on uhub3 Mar 28 14:46:08 localhost kernel: rtwn0: <Realtek Edimax AC1200 USB, class 0/0, rev 2.00/0.00, addr 2> on usbus4 Mar 28 14:46:08 localhost kernel: rtwn0: MAC/BB RTL8812AU, RF 6052 2T2R Mar 28 14:49:38 localhost reboot[51260]: rebooted by root Mar 28 14:52:13 localhost syslogd: kernel boot file is /boot/kernel/kernel Mar 28 14:52:13 localhost kernel: rtwn0 on uhub1 Mar 28 14:52:13 localhost kernel: rtwn0: <Realtek Edimax AC1200 USB, class 0/0, rev 2.00/0.00, addr 2> on usbus4 Mar 28 14:52:13 localhost kernel: rtwn0: MAC/BB RTL8812AU, RF 6052 2T2R Mar 29 10:58:33 localhost kernel: rtwn0: Mar 29 10:58:34 localhost kernel: rtwn0: ieee80211_crypto_encap returns NULL. Mar 29 10:58:34 localhost kernel: rtwn0: Mar 29 10:58:34 localhost kernel: rtwn0: ieee80211_crypto_encap returns NULL. Mar 29 11:01:42 localhost syslogd: kernel boot file is /boot/kernel/kernel Mar 29 11:01:42 localhost kernel: rtwn0 on uhub0 Mar 29 11:01:42 localhost kernel: rtwn0: <Realtek Edimax AC1200 USB, class 0/0, rev 2.00/0.00, addr 2> on usbus4 Mar 29 11:01:42 localhost kernel: rtwn0: MAC/BB RTL8812AU, RF 6052 2T2R I've disabled rxcsum for both IP and IPv6, to no avail. Step #2 doesn't seem to do anything; I'd need to add 'dev.rtwn.0.hwcrypto=0' to /boot/loader.conf because the OID is only tunable at boot, otherwise it's RO.
I don't think that it's rtwn_usb(4) issue. The card - probably - at least, I don't know about the opposite - isn't aware about IPv6 and / or IPv4. However, it can validate checksums (and -rxcsum[6] may reveal the issue) and cannot inspect frame contents when software encryption is used (dev.rtwn.0.hwcrypto=0). If you cannot reproduce with the latter (actually, it implicitly implies -rxcsum -rxcsum6) then this issue should be reassigned to the net@ team.
Oops, wrong statement: if you *still* can reproduce the issue - then ifnet(9) or net80211 should be inspected instead.
I could not repeat the issue after disabling hwcrypto. % % grep rtwn /boot/loader.conf if_rtwn_usb_load="YES" dev.rtwn.0.hwcrypto=0 % % sysctl dev.rtwn.0 dev.rtwn.0.rx_buf_size: 24 dev.rtwn.0.ratectl_selected: 1 dev.rtwn.0.ratectl: 1 dev.rtwn.0.hwcrypto: 0 dev.rtwn.0.debug: 0 dev.rtwn.0.ht40: 0 dev.rtwn.0.%parent: uhub2 dev.rtwn.0.%pnpinfo: vendor=0x7392 product=0xa822 devclass=0x00 devsubclass=0x00 devproto=0x00 sernum="123456" release=0x0000 mode=host intclass=0xff intsubclass=0xff intprotocol=0xff dev.rtwn.0.%location: bus=4 hubaddr=1 port=2 devaddr=2 interface=0 ugen=ugen4.2 dev.rtwn.0.%driver: rtwn dev.rtwn.0.%desc: Realtek Edimax AC1200 USB, class 0/0, rev 2.00/0.00, addr 2 % % rsync -6avz --progress --delete /usr/home/streams matt@remotehost:/media/usb/localhost/