BBB (cpsw(4)) seems to be broken in the latest 11-current
Updated my BBB to the latest -current, immediately got this while trying to make world over ssh console: 06:02:17 CPSW watchdog cpswss0: watchdog timeout cpswss0: Unable to cleanly shutdown transmitter Interface seems to be locked after that, no traffic comes in or out. This is: FreeBSD 11.0-ALPHA3 #1 ba7edef(tps65217x)-dirty: Fri Jun 17 16:22:07 PDT 2016, svn revision 301898 The previous version that was rock-solid was: FreeBSD 11.0-CURRENT #0 9d390ee(tps65217x)-dirty: Mon Jul 6 19:31:30 PDT 2015, svn revision 284878 I've been running buildworlds for literally days on that board, because it's how long it takes to build on that hardware. :) I'll run it again and see if the issue re-appears. If anyone seen this or if it's known issue please let me know. Thanks! -Max ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
On Sat, Jun 18, 2016 at 8:50 AM, Maxim Sobolev wrote: > Updated my BBB to the latest -current, immediately got this while trying to > make world over ssh console: > > 06:02:17 CPSW watchdog cpswss0: watchdog timeout > cpswss0: Unable to cleanly shutdown transmitter My BBB stucks in cpsw0 during boot rarely, and even soft reset (reset button) does not help. Only hard reset (power-off) helps. I have never had time to discover where a problem is. I'm not even sure if this is related to your problem as I did not remember exact dmesg in my case. Svata > > Interface seems to be locked after that, no traffic comes in or out. > > This is: > > FreeBSD 11.0-ALPHA3 #1 ba7edef(tps65217x)-dirty: Fri Jun 17 16:22:07 PDT > 2016, svn revision 301898 > > The previous version that was rock-solid was: > > FreeBSD 11.0-CURRENT #0 9d390ee(tps65217x)-dirty: Mon Jul 6 19:31:30 PDT > 2015, svn revision 284878 > > I've been running buildworlds for literally days on that board, because > it's how long it takes to build on that hardware. :) > > I'll run it again and see if the issue re-appears. > > If anyone seen this or if it's known issue please let me know. > > Thanks! > > -Max > ___ > freebsd-...@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-arm > To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org" ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
Well, I am not sure either as I don't have any issue restarting it afterwards. Yes, it seems to be happening fairly reliably here. :( Happened for me again, I left it running overnight. I am 99% positive it was not the case before kernel upgrade.. 07:11:52 CPSW watchdog cpswss0: watchdog timeout cpswss0: Unable to cleanly shutdown transmitter On Sat, Jun 18, 2016 at 1:09 AM, Svatopluk Kraus wrote: > On Sat, Jun 18, 2016 at 8:50 AM, Maxim Sobolev > wrote: > > Updated my BBB to the latest -current, immediately got this while trying > to > > make world over ssh console: > > > > 06:02:17 CPSW watchdog cpswss0: watchdog timeout > > cpswss0: Unable to cleanly shutdown transmitter > > My BBB stucks in cpsw0 during boot rarely, and even soft reset (reset > button) does not help. Only hard reset (power-off) helps. I have never > had time to discover where a problem is. I'm not even sure if this is > related to your problem as I did not remember exact dmesg in my case. > > Svata > > > > > > Interface seems to be locked after that, no traffic comes in or out. > > > > This is: > > > > FreeBSD 11.0-ALPHA3 #1 ba7edef(tps65217x)-dirty: Fri Jun 17 16:22:07 PDT > > 2016, svn revision 301898 > > > > The previous version that was rock-solid was: > > > > FreeBSD 11.0-CURRENT #0 9d390ee(tps65217x)-dirty: Mon Jul 6 19:31:30 PDT > > 2015, svn revision 284878 > > > > I've been running buildworlds for literally days on that board, because > > it's how long it takes to build on that hardware. :) > > > > I'll run it again and see if the issue re-appears. > > > > If anyone seen this or if it's known issue please let me know. > > > > Thanks! > > > > -Max > > ___ > > freebsd-...@freebsd.org mailing list > > https://lists.freebsd.org/mailman/listinfo/freebsd-arm > > To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org" > > ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
Jim, Yes, I've seen those. There were just handful of revision into the driver between my old good kernel and now, most of them are from you guys: r299477 | gonzo | 2016-05-11 11:20:02 -0700 (ср, 11 май 2016) | 16 lines r298352 | pfg | 2016-04-20 08:45:55 -0700 (ср, 20 апр 2016) | 6 lines r297132 | loos | 2016-03-20 20:16:56 -0700 (вс, 20 мар 2016) | 5 lines r297043 | loos | 2016-03-18 13:24:31 -0700 (пт, 18 мар 2016) | 4 lines r297042 | loos | 2016-03-18 13:09:54 -0700 (пт, 18 мар 2016) | 4 lines r297041 | loos | 2016-03-18 13:04:34 -0700 (пт, 18 мар 2016) | 4 lines r296993 | loos | 2016-03-17 12:35:08 -0700 (чт, 17 мар 2016) | 24 lines r296980 | loos | 2016-03-16 23:23:48 -0700 (ср, 16 мар 2016) | 6 lines r283287 | andrew | 2015-05-22 07:25:23 -0700 (пт, 22 май 2015) | 4 lines (last known good one) I've reverted the driver to the state all way down to r283287, while keeping the rest of the kernel intact and soon will see if it works better. If that works fine, I'll try to bi-sect it to a single troublesome revision. -Max On Sat, Jun 18, 2016 at 12:26 PM, Jim Thompson wrote: > There are recent changes to enable the switch and two port MAC mode. > > These were lightly tested on BBB prior to being committed. > > -- Jim > > > On Jun 18, 2016, at 11:49 AM, Maxim Sobolev wrote: > > > > Well, I am not sure either as I don't have any issue restarting it > > afterwards. > > > > Yes, it seems to be happening fairly reliably here. :( Happened for me > > again, I left it running overnight. I am 99% positive it was not the case > > before kernel upgrade.. > > > > 07:11:52 CPSW watchdog cpswss0: watchdog timeout > > cpswss0: Unable to cleanly shutdown transmitter > > > > > >> On Sat, Jun 18, 2016 at 1:09 AM, Svatopluk Kraus > wrote: > >> > >> On Sat, Jun 18, 2016 at 8:50 AM, Maxim Sobolev > >> wrote: > >>> Updated my BBB to the latest -current, immediately got this while > trying > >> to > >>> make world over ssh console: > >>> > >>> 06:02:17 CPSW watchdog cpswss0: watchdog timeout > >>> cpswss0: Unable to cleanly shutdown transmitter > >> > >> My BBB stucks in cpsw0 during boot rarely, and even soft reset (reset > >> button) does not help. Only hard reset (power-off) helps. I have never > >> had time to discover where a problem is. I'm not even sure if this is > >> related to your problem as I did not remember exact dmesg in my case. > >> > >> Svata > >> > >> > >>> > >>> Interface seems to be locked after that, no traffic comes in or out. > >>> > >>> This is: > >>> > >>> FreeBSD 11.0-ALPHA3 #1 ba7edef(tps65217x)-dirty: Fri Jun 17 16:22:07 > PDT > >>> 2016, svn revision 301898 > >>> > >>> The previous version that was rock-solid was: > >>> > >>> FreeBSD 11.0-CURRENT #0 9d390ee(tps65217x)-dirty: Mon Jul 6 19:31:30 > PDT > >>> 2015, svn revision 284878 > >>> > >>> I've been running buildworlds for literally days on that board, because > >>> it's how long it takes to build on that hardware. :) > >>> > >>> I'll run it again and see if the issue re-appears. > >>> > >>> If anyone seen this or if it's known issue please let me know. > >>> > >>> Thanks! > >>> > >>> -Max > >>> ___ > >>> freebsd-...@freebsd.org mailing list > >>> https://lists.freebsd.org/mailman/listinfo/freebsd-arm > >>> To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org" > > ___ > > freebsd-...@freebsd.org mailing list > > https://lists.freebsd.org/mailman/listinfo/freebsd-arm > > To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org" > > ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
There are recent changes to enable the switch and two port MAC mode. These were lightly tested on BBB prior to being committed. -- Jim > On Jun 18, 2016, at 11:49 AM, Maxim Sobolev wrote: > > Well, I am not sure either as I don't have any issue restarting it > afterwards. > > Yes, it seems to be happening fairly reliably here. :( Happened for me > again, I left it running overnight. I am 99% positive it was not the case > before kernel upgrade.. > > 07:11:52 CPSW watchdog cpswss0: watchdog timeout > cpswss0: Unable to cleanly shutdown transmitter > > >> On Sat, Jun 18, 2016 at 1:09 AM, Svatopluk Kraus wrote: >> >> On Sat, Jun 18, 2016 at 8:50 AM, Maxim Sobolev >> wrote: >>> Updated my BBB to the latest -current, immediately got this while trying >> to >>> make world over ssh console: >>> >>> 06:02:17 CPSW watchdog cpswss0: watchdog timeout >>> cpswss0: Unable to cleanly shutdown transmitter >> >> My BBB stucks in cpsw0 during boot rarely, and even soft reset (reset >> button) does not help. Only hard reset (power-off) helps. I have never >> had time to discover where a problem is. I'm not even sure if this is >> related to your problem as I did not remember exact dmesg in my case. >> >> Svata >> >> >>> >>> Interface seems to be locked after that, no traffic comes in or out. >>> >>> This is: >>> >>> FreeBSD 11.0-ALPHA3 #1 ba7edef(tps65217x)-dirty: Fri Jun 17 16:22:07 PDT >>> 2016, svn revision 301898 >>> >>> The previous version that was rock-solid was: >>> >>> FreeBSD 11.0-CURRENT #0 9d390ee(tps65217x)-dirty: Mon Jul 6 19:31:30 PDT >>> 2015, svn revision 284878 >>> >>> I've been running buildworlds for literally days on that board, because >>> it's how long it takes to build on that hardware. :) >>> >>> I'll run it again and see if the issue re-appears. >>> >>> If anyone seen this or if it's known issue please let me know. >>> >>> Thanks! >>> >>> -Max >>> ___ >>> freebsd-...@freebsd.org mailing list >>> https://lists.freebsd.org/mailman/listinfo/freebsd-arm >>> To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org" > ___ > freebsd-...@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-arm > To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org" ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
Jim, some update from here. Running r283287 of the driver, I still see the same "watchdog timeout" messages, but they do not lead to the interface lockout. The traffic resumes momentarily. Which is probably why I never paid much attention to those warnings before. Therefore, I suspect that the new MAC code does not deal with watchdog-triggered interface reset as good as the old code. Does it give you any ideas about what could be wrong there by any chance? 03:54:50 CPSW watchdog cpsw0: watchdog timeout cpsw0: Unable to cleanly shutdown transmitter 03:58:03 CPSW watchdog cpsw0: watchdog timeout cpsw0: Unable to cleanly shutdown transmitter On Sat, Jun 18, 2016 at 2:09 PM, Maxim Sobolev wrote: > Jim, > > Yes, I've seen those. There were just handful of revision into the driver > between my old good kernel and now, most of them are from you guys: > > r299477 | gonzo | 2016-05-11 11:20:02 -0700 (ср, 11 май 2016) | 16 lines > r298352 | pfg | 2016-04-20 08:45:55 -0700 (ср, 20 апр 2016) | 6 lines > r297132 | loos | 2016-03-20 20:16:56 -0700 (вс, 20 мар 2016) | 5 lines > r297043 | loos | 2016-03-18 13:24:31 -0700 (пт, 18 мар 2016) | 4 lines > r297042 | loos | 2016-03-18 13:09:54 -0700 (пт, 18 мар 2016) | 4 lines > r297041 | loos | 2016-03-18 13:04:34 -0700 (пт, 18 мар 2016) | 4 lines > r296993 | loos | 2016-03-17 12:35:08 -0700 (чт, 17 мар 2016) | 24 lines > r296980 | loos | 2016-03-16 23:23:48 -0700 (ср, 16 мар 2016) | 6 lines > r283287 | andrew | 2015-05-22 07:25:23 -0700 (пт, 22 май 2015) | 4 lines > (last known good one) > > I've reverted the driver to the state all way down to r283287, while > keeping the rest of the kernel intact and soon will see if it works better. > If that works fine, I'll try to bi-sect it to a single troublesome revision. > > -Max > > On Sat, Jun 18, 2016 at 12:26 PM, Jim Thompson wrote: > >> There are recent changes to enable the switch and two port MAC mode. >> >> These were lightly tested on BBB prior to being committed. >> >> -- Jim >> >> > On Jun 18, 2016, at 11:49 AM, Maxim Sobolev >> wrote: >> > >> > Well, I am not sure either as I don't have any issue restarting it >> > afterwards. >> > >> > Yes, it seems to be happening fairly reliably here. :( Happened for me >> > again, I left it running overnight. I am 99% positive it was not the >> case >> > before kernel upgrade.. >> > >> > 07:11:52 CPSW watchdog cpswss0: watchdog timeout >> > cpswss0: Unable to cleanly shutdown transmitter >> > >> > >> >> On Sat, Jun 18, 2016 at 1:09 AM, Svatopluk Kraus >> wrote: >> >> >> >> On Sat, Jun 18, 2016 at 8:50 AM, Maxim Sobolev >> >> wrote: >> >>> Updated my BBB to the latest -current, immediately got this while >> trying >> >> to >> >>> make world over ssh console: >> >>> >> >>> 06:02:17 CPSW watchdog cpswss0: watchdog timeout >> >>> cpswss0: Unable to cleanly shutdown transmitter >> >> >> >> My BBB stucks in cpsw0 during boot rarely, and even soft reset (reset >> >> button) does not help. Only hard reset (power-off) helps. I have never >> >> had time to discover where a problem is. I'm not even sure if this is >> >> related to your problem as I did not remember exact dmesg in my case. >> >> >> >> Svata >> >> >> >> >> >>> >> >>> Interface seems to be locked after that, no traffic comes in or out. >> >>> >> >>> This is: >> >>> >> >>> FreeBSD 11.0-ALPHA3 #1 ba7edef(tps65217x)-dirty: Fri Jun 17 16:22:07 >> PDT >> >>> 2016, svn revision 301898 >> >>> >> >>> The previous version that was rock-solid was: >> >>> >> >>> FreeBSD 11.0-CURRENT #0 9d390ee(tps65217x)-dirty: Mon Jul 6 19:31:30 >> PDT >> >>> 2015, svn revision 284878 >> >>> >> >>> I've been running buildworlds for literally days on that board, >> because >> >>> it's how long it takes to build on that hardware. :) >> >>> >> >>> I'll run it again and see if the issue re-appears. >> >>> >> >>> If anyone seen this or if it's known issue please let me know. >> >>> >> >>> Thanks! >> >>> >> >>> -Max >> >>> ___ >> >>> freebsd-...@freebsd.org mailing list >> >>> https://lists.freebsd.org/mailman/listinfo/freebsd-arm >> >>> To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org >> " >> > ___ >> > freebsd-...@freebsd.org mailing list >> > https://lists.freebsd.org/mailman/listinfo/freebsd-arm >> > To unsubscribe, send any mail to "freebsd-arm-unsubscr...@freebsd.org" >> >> > -- Maksym Sobolyev Sippy Software, Inc. Internet Telephony (VoIP) Experts Tel (Canada): +1-778-783-0474 Tel (Toll-Free): +1-855-747-7779 Fax: +1-866-857-6942 Web: http://www.sippysoft.com MSN: sa...@sippysoft.com Skype: SippySoft ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
On Sun, Jun 19, 2016 at 1:11 AM, Maxim Sobolev wrote: > Jim, some update from here. Running r283287 of the driver, I still see the > same "watchdog timeout" messages, but they do not lead to the interface > lockout. The traffic resumes momentarily. Which is probably why I never paid > much attention to those warnings before. Therefore, I suspect that the new > MAC code does not deal with watchdog-triggered interface reset as good as > the old code. Does it give you any ideas about what could be wrong there by > any chance? Hi Maxim, My recent changes contributed somehow to expose the bug more frequently. There was a condition in tx packet reclamation where we aren't restarting the tx queue in one of the possible stall conditions. Please try the attached patch and let me know if it works for you. Luiz Index: sys/arm/ti/cpsw/if_cpsw.c === --- sys/arm/ti/cpsw/if_cpsw.c (revision 301975) +++ sys/arm/ti/cpsw/if_cpsw.c (working copy) @@ -1874,6 +1874,7 @@ return; } else if (last_old_slot == NULL) { /* Start a fresh queue. */ + sc->swsc->last_hdp = cpsw_cpdma_bd_paddr(sc->swsc, first_new_slot); cpsw_write_hdp_slot(sc->swsc, &sc->swsc->tx, first_new_slot); } else { /* Add buffers to end of current queue. */ @@ -1882,6 +1883,7 @@ /* If underrun, restart queue. */ if (cpsw_cpdma_read_bd_flags(sc->swsc, last_old_slot) & CPDMA_BD_EOQ) { + sc->swsc->last_hdp = cpsw_cpdma_bd_paddr(sc->swsc, first_new_slot); cpsw_write_hdp_slot(sc->swsc, &sc->swsc->tx, first_new_slot); } @@ -1897,6 +1899,7 @@ cpsw_tx_dequeue(struct cpsw_softc *sc) { struct cpsw_slot *slot, *last_removed_slot = NULL; + struct cpsw_cpdma_bd bd; uint32_t flags, removed = 0; slot = STAILQ_FIRST(&sc->tx.active); @@ -1931,7 +1934,8 @@ } /* TearDown complete is only marked on the SOP for the packet. */ - if (flags & CPDMA_BD_TDOWNCMPLT) { + if ((flags & (CPDMA_BD_SOP | CPDMA_BD_TDOWNCMPLT)) == + (CPDMA_BD_EOP | CPDMA_BD_TDOWNCMPLT)) { CPSW_DEBUGF(sc, ("TX teardown in progress")); cpsw_write_cp(sc, &sc->tx, 0xfffc); // TODO: Increment a count of dropped TX packets @@ -1938,6 +1942,16 @@ sc->tx.running = 0; break; } + + if ((flags & CPDMA_BD_EOP) == 0) + flags = cpsw_cpdma_read_bd_flags(sc, last_removed_slot); + if ((flags & (CPDMA_BD_EOP | CPDMA_BD_EOQ)) == + (CPDMA_BD_EOP | CPDMA_BD_EOQ)) { + cpsw_cpdma_read_bd(sc, last_removed_slot, &bd); + if (bd.next != 0 && bd.next != sc->last_hdp) + /* Restart the queue. */ + cpsw_write_4(sc, sc->tx.hdp_offset, bd.next); + } } if (removed != 0) { Index: sys/arm/ti/cpsw/if_cpswvar.h === --- sys/arm/ti/cpsw/if_cpswvar.h(revision 301975) +++ sys/arm/ti/cpsw/if_cpswvar.h(working copy) @@ -83,6 +83,7 @@ /* RX and TX buffer tracking */ struct cpsw_queue rx, tx; + uint32_tlast_hdp; /* We expect 1 memory resource and 4 interrupts from the device tree. */ int mem_rid; ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
On Jun 18, 2016, at 2:50 AM, Maxim Sobolev wrote: > Updated my BBB to the latest -current, immediately got this while trying to > make world over ssh console: > > 06:02:17 CPSW watchdog cpswss0: watchdog timeout > cpswss0: Unable to cleanly shutdown transmitter > > Interface seems to be locked after that, no traffic comes in or out. > > This is: > > FreeBSD 11.0-ALPHA3 #1 ba7edef(tps65217x)-dirty: Fri Jun 17 16:22:07 PDT > 2016, svn revision 301898 > > The previous version that was rock-solid was: > > FreeBSD 11.0-CURRENT #0 9d390ee(tps65217x)-dirty: Mon Jul 6 19:31:30 PDT > 2015, svn revision 284878 > > I've been running buildworlds for literally days on that board, because > it's how long it takes to build on that hardware. :) > > I'll run it again and see if the issue re-appears. > > If anyone seen this or if it's known issue please let me know. I have experienced this problem recently, too, after updating to 11.0-ALPHA3. I get the watchdog timeout messages you give above when trying to buildworld with /usr/src and /usr/obj mounted via NFS. My last successful build that did not have this problem is this one: FreeBSD beaglebone 11.0-ALPHA2 FreeBSD 11.0-ALPHA2 #0 r301779: Mon Jun 13 01:30:05 EDT 2016 pmather@beaglebone:/usr/obj/usr/src/sys/BEAGLEBONE-NO_WITNESS arm The build where this started happening for me is this one: FreeBSD 11.0-ALPHA3 #0 r301876: Wed Jun 15 14:23:28 EDT 2016 That's just two days apart. Maybe that might help track down the potential revision that caused the problem. I've not been able to buildworld via NFS since the problem began, so I've reverted back to the r301779 kernel for now. Cheers, Paul. ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
On Mon, 20 Jun 2016, Luiz Otavio O Souza wrote: On Sun, Jun 19, 2016 at 1:11 AM, Maxim Sobolev wrote: Jim, some update from here. Running r283287 of the driver, I still see the same "watchdog timeout" messages, but they do not lead to the interface lockout. The traffic resumes momentarily. Which is probably why I never paid much attention to those warnings before. Therefore, I suspect that the new MAC code does not deal with watchdog-triggered interface reset as good as the old code. Does it give you any ideas about what could be wrong there by any chance? Hi Maxim, My recent changes contributed somehow to expose the bug more frequently. There was a condition in tx packet reclamation where we aren't restarting the tx queue in one of the possible stall conditions. Please try the attached patch and let me know if it works for you. Luiz Your patch fixes the problem for me. Thanks! FreeBSD beaglebone 11.0-ALPHA4 FreeBSD 11.0-ALPHA4 #0 r302028M: Mon Jun 20 18:19:55 EDT 2016 kwhite@freebsd11:/usr/obj/arm.armv6/usr/src/sys/BEAGLEBONE-LOCAL arm armv6 ...keith ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
Nice to hear. I am building it right now. Gotta get you some results in the next few hours. Thanks! -Max On Mon, Jun 20, 2016 at 3:33 PM, Keith White wrote: > On Mon, 20 Jun 2016, Luiz Otavio O Souza wrote: > > On Sun, Jun 19, 2016 at 1:11 AM, Maxim Sobolev wrote: >> >>> Jim, some update from here. Running r283287 of the driver, I still see >>> the >>> same "watchdog timeout" messages, but they do not lead to the interface >>> lockout. The traffic resumes momentarily. Which is probably why I never >>> paid >>> much attention to those warnings before. Therefore, I suspect that the >>> new >>> MAC code does not deal with watchdog-triggered interface reset as good as >>> the old code. Does it give you any ideas about what could be wrong there >>> by >>> any chance? >>> >> >> >> Hi Maxim, >> >> My recent changes contributed somehow to expose the bug more frequently. >> >> There was a condition in tx packet reclamation where we aren't >> restarting the tx queue in one of the possible stall conditions. >> >> Please try the attached patch and let me know if it works for you. >> >> Luiz >> > > Your patch fixes the problem for me. Thanks! > > FreeBSD beaglebone 11.0-ALPHA4 FreeBSD 11.0-ALPHA4 #0 r302028M: Mon Jun 20 > 18:19:55 EDT 2016 > kwhite@freebsd11:/usr/obj/arm.armv6/usr/src/sys/BEAGLEBONE-LOCAL > arm armv6 > > ...keith > ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
On Jun 20, 2016, at 6:33 PM, Keith White wrote: > On Mon, 20 Jun 2016, Luiz Otavio O Souza wrote: > >> On Sun, Jun 19, 2016 at 1:11 AM, Maxim Sobolev wrote: >>> Jim, some update from here. Running r283287 of the driver, I still see the >>> same "watchdog timeout" messages, but they do not lead to the interface >>> lockout. The traffic resumes momentarily. Which is probably why I never paid >>> much attention to those warnings before. Therefore, I suspect that the new >>> MAC code does not deal with watchdog-triggered interface reset as good as >>> the old code. Does it give you any ideas about what could be wrong there by >>> any chance? >> >> >> Hi Maxim, >> >> My recent changes contributed somehow to expose the bug more frequently. >> >> There was a condition in tx packet reclamation where we aren't >> restarting the tx queue in one of the possible stall conditions. >> >> Please try the attached patch and let me know if it works for you. >> >> Luiz > > Your patch fixes the problem for me. Thanks! > > FreeBSD beaglebone 11.0-ALPHA4 FreeBSD 11.0-ALPHA4 #0 r302028M: Mon Jun 20 > 18:19:55 EDT 2016 > kwhite@freebsd11:/usr/obj/arm.armv6/usr/src/sys/BEAGLEBONE-LOCAL arm armv6 > > ...keith The patch also fixes the problem for me. FreeBSD beaglebone 11.0-ALPHA4 FreeBSD 11.0-ALPHA4 #2 r302030M: Tue Jun 21 10:20:59 EDT 2016 pmather@beaglebone:/usr/obj/usr/src/sys/BEAGLEBONE-NO_WITNESS arm Cheers, Paul. ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: BBB (cpsw(4)) seems to be broken in the latest 11-current
Luiz, the patch works for me beautifully. Has not been able to reproduce the issue since applying it. Can you please merge it before 11.0 gets branched? Thanks! -Max On Tue, Jun 21, 2016 at 10:33 AM, Paul Mather wrote: > On Jun 20, 2016, at 6:33 PM, Keith White wrote: > > > On Mon, 20 Jun 2016, Luiz Otavio O Souza wrote: > > > >> On Sun, Jun 19, 2016 at 1:11 AM, Maxim Sobolev wrote: > >>> Jim, some update from here. Running r283287 of the driver, I still see > the > >>> same "watchdog timeout" messages, but they do not lead to the interface > >>> lockout. The traffic resumes momentarily. Which is probably why I > never paid > >>> much attention to those warnings before. Therefore, I suspect that the > new > >>> MAC code does not deal with watchdog-triggered interface reset as good > as > >>> the old code. Does it give you any ideas about what could be wrong > there by > >>> any chance? > >> > >> > >> Hi Maxim, > >> > >> My recent changes contributed somehow to expose the bug more frequently. > >> > >> There was a condition in tx packet reclamation where we aren't > >> restarting the tx queue in one of the possible stall conditions. > >> > >> Please try the attached patch and let me know if it works for you. > >> > >> Luiz > > > > Your patch fixes the problem for me. Thanks! > > > > FreeBSD beaglebone 11.0-ALPHA4 FreeBSD 11.0-ALPHA4 #0 r302028M: Mon Jun > 20 18:19:55 EDT 2016 > kwhite@freebsd11:/usr/obj/arm.armv6/usr/src/sys/BEAGLEBONE-LOCAL > arm armv6 > > > > ...keith > > > The patch also fixes the problem for me. > > FreeBSD beaglebone 11.0-ALPHA4 FreeBSD 11.0-ALPHA4 #2 r302030M: Tue Jun 21 > 10:20:59 EDT 2016 > pmather@beaglebone:/usr/obj/usr/src/sys/BEAGLEBONE-NO_WITNESS > arm > > > Cheers, > > Paul. > > ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"