Re: iflib/bridge kernel panic
Quoting Kristof Provost (from Mon, 28 Sep 2020 13:53:16 +0200): On 28 Sep 2020, at 12:45, Alexander Leidinger wrote: Quoting Kristof Provost (from Sun, 27 Sep 2020 17:51:32 +0200): Here’s an early version of a task queue based approach: http://people.freebsd.org/~kp/0001-bridge-Cope-with-if_ioctl-s-that-sleep.patch That still needs to be cleaned up, but this should resolve the sleep issue and the LOR. There are some issues... seems like inside a jail I can't ping systems outside of the hardware. Bridge setup: - member jail A - member jail B - member external_if of host If I ping the router from the host, it works. If I ping from one jail to another, it works. If I ping from the jail to the IP of the external_if, it works. If I ping from a jail to the router, I do not get a response. Can you check for 'failed ifpromisc' error messages in dmesg? And verify that all bridge member interfaces are in promiscuous mode? I have a panic for you...: - startup still in progress = 22 jails in startup, somewhere after a few jails started the panic happened - tcpdump was running on the external interface - a ping to a jail IP from another system was running, the first ping went through, then it paniced First regarding your questions about promisc mode: no error, but the promisc mode is directly disabled again on all interfaces. Data (external_if = igb0, jail epairs are j_X_Yif with X the ID of the jail and Y either h like host-side or j like jail-side): ---snip--- Host: # ifconfig -a igb0: flags=8863 metric 0 mtu 1500 options=4a520b9 ether [...]:a4 inet 192.168.1.x netmask 0xff00 broadcast 192.168.1.255 inet6 fe80::[...]a4%igb0 prefixlen 64 scopeid 0x1 inet6 fd73:[...] prefixlen 64 inet6 2003:[...] prefixlen 64 autoconf inet6 fd73:[...] prefixlen 64 autoconf media: Ethernet autoselect (1000baseT ) status: active nd6 options=23 igb1: flags=8822 metric 0 mtu 1500 options=4e527bb ether [...]:a5 media: Ethernet autoselect status: no carrier nd6 options=29 lo0: flags=8049 metric 0 mtu 16384 options=680003 inet6 ::1 prefixlen 128 inet6 fe80::1%lo0 prefixlen 64 scopeid 0x3 inet 127.0.0.1 netmask 0xff00 groups: lo nd6 options=21 vswitch0: flags=8843 metric 0 mtu 1500 ether [...]:a3 id 00:00:00:00:00:00 priority 32768 hellotime 2 fwddelay 15 maxage 20 holdcnt 6 proto stp-rstp maxaddr 2000 timeout 1200 root id 00:00:00:00:00:00 priority 32768 ifcost 0 port 0 member: j_weather_hif flags=143 ifmaxaddr 0 port 9 priority 128 path cost 2000 member: j_web_hif flags=143 ifmaxaddr 0 port 8 priority 128 path cost 2000 member: j_commit_hif flags=143 ifmaxaddr 0 port 7 priority 128 path cost 2000 member: j_video_hif flags=143 ifmaxaddr 0 port 6 priority 128 path cost 2000 member: j_dns_hif flags=143 ifmaxaddr 0 port 5 priority 128 path cost 2000 member: igb0 flags=143 ifmaxaddr 0 port 1 priority 128 path cost 2 groups: bridge nd6 options=9 j_dns_hif: flags=8843 metric 0 mtu 1500 options=8 ether [...]:0a hwaddr [...]:0a inet6 fe80::[...]0a%j_dns_hif prefixlen 64 scopeid 0x5 groups: epair media: Ethernet 10Gbase-T (10Gbase-T ) status: active nd6 options=21 [... some more jail interfaces ...] # dmesg | grep promis igb0: promiscuous mode enabled igb0: promiscuous mode disabled j_dns_hif: promiscuous mode enabled j_dns_hif: promiscuous mode disabled [... some more like this ...] # jexec 2 ifconfig -a lo0: flags=8049 metric 0 mtu 16384 options=680003 inet6 ::1 prefixlen 128 inet6 fe80::1%lo0 prefixlen 64 scopeid 0x1 inet 127.0.0.1 netmask 0xff00 groups: lo nd6 options=21 j_dns_jif: flags=8843 metric 0 mtu 1500 options=8 ether [...]:0b hwaddr [...]:0b inet 192.168.1.y netmask 0xff00 broadcast 192.168.1.255 inet6 fe80::[...]0b%j_dns_jif prefixlen 64 scopeid 0x2 inet6 fd73:[...]:y prefixlen 64 groups: epair media: Ethernet 10Gbase-T (10Gbase-T ) status: active nd6 options=21 ---snip--- And here the backtrace of the panic: ---snip--- panic: if_setflag: decrement non-positive refcount 0 for flag 256 cpuid = 4 time = 1601300532 KDB: stack backtrace: db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfe0378ea3920 vpanic() at vpanic+0x182/frame 0xfe0378ea3970 panic() at panic+0x43/frame 0xfe0378ea39d0 if_setflag() at if_setflag+0x137/frame 0xfe0378ea3a30 ifpromisc() at ifpromisc+0x2a/frame 0xfe0378ea3a60 bpf_detachd_locked() at bpf_detachd_locked+0x280/frame 0xfe0378ea3ab0
Re: iflib/bridge kernel panic
Quoting Kristof Provost (from Sun, 27 Sep 2020 17:51:32 +0200): On 21 Sep 2020, at 14:16, Shawn Webb wrote: On Mon, Sep 21, 2020 at 09:57:40AM +0200, Kristof Provost wrote: On 21 Sep 2020, at 2:52, Shawn Webb wrote: From latest HEAD on a Dell Precision 7550 laptop: https://gist.github.com/lattera/a0803f31f58bcf8ead51ac1ebbc447e2 The last working boot environment was 14 Aug 2020. If I get some time to bisect commits, I'll try to figure out the culprit. Try https://reviews.freebsd.org/D26418 That seems to fix the kernel panic. dmesg gets spammed with a freak ton of these LOR messages now: Here’s an early version of a task queue based approach: http://people.freebsd.org/~kp/0001-bridge-Cope-with-if_ioctl-s-that-sleep.patch That still needs to be cleaned up, but this should resolve the sleep issue and the LOR. There are some issues... seems like inside a jail I can't ping systems outside of the hardware. Bridge setup: - member jail A - member jail B - member external_if of host If I ping the router from the host, it works. If I ping from one jail to another, it works. If I ping from the jail to the IP of the external_if, it works. If I ping from a jail to the router, I do not get a response. Bye, Alexander. -- http://www.Leidinger.net alexan...@leidinger.net: PGP 0x8F31830F9F2772BF http://www.FreeBSD.orgnetch...@freebsd.org : PGP 0x8F31830F9F2772BF pgpZ4OpaNfO4d.pgp Description: Digitale PGP-Signatur
Re: iflib/bridge kernel panic
On 28 Sep 2020, at 12:45, Alexander Leidinger wrote: Quoting Kristof Provost (from Sun, 27 Sep 2020 17:51:32 +0200): Here’s an early version of a task queue based approach: http://people.freebsd.org/~kp/0001-bridge-Cope-with-if_ioctl-s-that-sleep.patch That still needs to be cleaned up, but this should resolve the sleep issue and the LOR. There are some issues... seems like inside a jail I can't ping systems outside of the hardware. Bridge setup: - member jail A - member jail B - member external_if of host If I ping the router from the host, it works. If I ping from one jail to another, it works. If I ping from the jail to the IP of the external_if, it works. If I ping from a jail to the router, I do not get a response. Can you check for 'failed ifpromisc' error messages in dmesg? And verify that all bridge member interfaces are in promiscuous mode? Kristof ___ 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: clang build buggy code with certain CPUTYPE setting
On 26/09/2020 22:55, Marek Zarychta wrote: > Thank you for the information and for the fix. Sadly I must admit it > doesn't work for me. I have tried two builds with fresh sources today to > be certain and it looks like the bug is still present on FreeBSD > 13-CURRENT r366186. Either the upstream fixed it only partially or it is > another bug. As a workaround, I will build worlds without > CPUTYPE?=amdfam10 for a while. I hope the problem will be resolved > before clang 11 is MFCed to 12-STABLE. Can you disassemble the faulting instruction in the core dump? Can you provide full CPU ID / features information from dmesg? -- Andriy Gapon ___ 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"
