On [31/01/18] [01:53P], Kevin Lo wrote:
; On Wed, Jan 31, 2018 at 01:25:10PM +0100, Stefan Sperling wrote:
; >
; > If iwn(4) is in associated state (status: active), and I run:
; >
; > ifconfig iwn0 scan
; >
; > it loses link (status: no network) and never comes back while the
; > LED keeps blinking. I need to run 'ifconfig iwn0 down' to recover.
;
; I'm getting the same problem.
;
; > What happens is the driver goes to SCAN state and sends a scan
; > command to the firmware and waits for a STOP_SCAN response.
; > This happens while the firmware is in associated state, with a
; > scan command which does *not* run a background scan. It seems the
; > firmware doesn't like this and never signals STOP_SCAN, so the
; > driver remains in SCAN state until the interface gets reset.
; >
; > I can fix this problem by making the driver move the firmware out
; > of associated state during RUN->SCAN transitions. For completeness
; > we can do the same for RUN->{ASSOC,AUTH} and AUTH->ASSOC. i.e. always
; > keep the firmware's association state in sync with the driver's state
; > machine. The firmware should only be associated in RUN state.
; >
; > The diff below does that, and also adds support for SCAN->SCAN
; > transitions to the state machine (which are a no-op in this driver).
; >
; > This seems to be working well here with:
; > iwn0 at pci3 dev 0 function 0 "Intel Centrino Advanced-N 6200" rev 0x35:
msi, MIMO 2T2R, MoW, address xx:xx:xx:xx:xx:xx
; >
; > OK to commit? Does anybody else want to test this beforehand?
;
; Your diff works for me as well:
; iwn0 at pci2 dev 0 function 0 "Intel Wireless WiFi Link 4965" rev 0x61: msi,
MIMO 2T3R, MoW1, address 00:1d:e0:b3:e6:c1
;
; Thanks. ok kevlo@
;
; > Index: if_iwn.c
; > ===================================================================
; > RCS file: /cvs/src/sys/dev/pci/if_iwn.c,v
; > retrieving revision 1.198
; > diff -u -p -r1.198 if_iwn.c
; > --- if_iwn.c 9 Jan 2018 10:00:12 -0000 1.198
; > +++ if_iwn.c 31 Jan 2018 12:13:41 -0000
; > @@ -1772,6 +1772,25 @@ iwn_newstate(struct ieee80211com *ic, en
; > iwn_scan_abort(sc);
; > }
; >
; > + if (ic->ic_state == IEEE80211_S_SCAN) {
; > + if (nstate == IEEE80211_S_SCAN)
; > + return 0;
; > + /* Turn LED off when leaving scan state. */
; > + iwn_set_led(sc, IWN_LED_LINK, 1, 0);
; > + }
; > +
; > + if (ic->ic_state >= IEEE80211_S_ASSOC &&
; > + nstate <= IEEE80211_S_ASSOC) {
; > + /* Reset state to handle re- and disassociations. */
; > + sc->rxon.associd = 0;
; > + sc->rxon.filter &= ~htole32(IWN_FILTER_BSS);
; > + sc->calib.state = IWN_CALIB_STATE_INIT;
; > + error = iwn_cmd(sc, IWN_CMD_RXON, &sc->rxon, sc->rxonsz, 1);
; > + if (error != 0)
; > + printf("%s: RXON command failed\n",
; > + sc->sc_dev.dv_xname);
; > + }
; > +
; > switch (nstate) {
; > case IEEE80211_S_SCAN:
; > /* Make the link LED blink while we're scanning. */
; > @@ -1790,11 +1809,6 @@ iwn_newstate(struct ieee80211com *ic, en
; > break;
; > /* FALLTHROUGH */
; > case IEEE80211_S_AUTH:
; > - /* Reset state to handle reassociations correctly. */
; > - sc->rxon.associd = 0;
; > - sc->rxon.filter &= ~htole32(IWN_FILTER_BSS);
; > - sc->calib.state = IWN_CALIB_STATE_INIT;
; > -
; > if ((error = iwn_auth(sc, arg)) != 0) {
; > printf("%s: could not move to auth state\n",
; > sc->sc_dev.dv_xname);
; >
; >
;
Same here, with: "Intel Dual Band Wireless AC 7265"
Thanks!
--
Sending from my toaster.
