| Summary: | ixl: hangs when adding it to a bridge or enabling iovctl VFs (SR-IOV) | ||
|---|---|---|---|
| Product: | Base System | Reporter: | Finn <ixbug> |
| Component: | kern | Assignee: | freebsd-net (Nobody) <net> |
| Status: | Closed Not A Bug | ||
| Severity: | Affects Only Me | CC: | erj, sbruno |
| Priority: | --- | Keywords: | IntelNetworking |
| Version: | 11.2-RELEASE | ||
| Hardware: | amd64 | ||
| OS: | Any | ||
The SR-IOV part of this was an error on my side
/etc/iov/ixl1.conf
---
PF {
device : "ixl1";
num_vfs : 2;
}
DEFAULT {
passthrough : true;
}
--
new version with an VF for the host (passthrough: false)
---
PF {
device : "ixl1";
num_vfs : 3;
}
DEFAULT {
passthrough : true;
}
#for the host:
VF-0 {
passthrough : false;
}
--
this creates a ixlv0 which can be used by the host (ixl1 becomes unusable after 'service iovctl start')
|
this is with driver version v1.9.5 (for v1.9.13: see the next section): when adding ixl1 to a bridge the interface hangs: # vm switch create public # vm switch add public ixl1 /var/log/messages: hh:39:49 kernel: bridge0: Ethernet address: ... hh:39:49 kernel: bridge0: changing name to 'vm-public' hh:40:12 kernel: ixl1: RX queue 3 still enabled! hh:40:12 kernel: vm-public: link state changed to UP hh:40:12 kernel: ixl1: promiscuous mode enabled hh:40:12 kernel: ixl1: Malicious Driver Detection event 2 on TX queue 773, pf number 1 hh:40:12 kernel: ixl1: MDD TX event is for this function! hh:40:12 kernel: ixl1: Malicious Driver Detection event 2 on TX queue 774, pf number 1 hh:40:12 kernel: ixl1: MDD TX event is for this function! hh:40:12 kernel: ixl1: Malicious Driver Detection event 2 on TX queue 772, pf number 1 hh:40:12 kernel: ixl1: MDD TX event is for this function! hh:40:23 kernel: ixl1: WARNING: queue 7 appears to be hung! hh:40:24 kernel: ixl1: WARNING: queue 6 appears to be hung! hh:40:30 kernel: ixl1: WARNING: queue 4 appears to be hung! hh:40:34 kernel: ixl1: WARNING: queue 5 appears to be hung! -> rebooted to recover --------- Since the bridge interface does not work I tried SR-IOV (this is with driver version v1.9.13 after the reboot) log when booting up: kernel: ixl1: <Intel(R) Ethernet Connection 700 Series PF Driver, Version - 1.9.13> mem 0xf8800000-0xf8ffffff,0xf9800000-0xf9807fff irq 32 at device 0.1 numa-domain 0 on pci3 kernel: ixl1: using 1024 tx descriptors and 1024 rx descriptors kernel: ixl1: fw 6.0.48442 api 1.7 nvm 6.01 etid 800035cf oem 1.262.0 kernel: ixl1: PF-ID[1]: VFs 64, MSIX 129, VF MSIX 5, QPs 768, I2C kernel: ixl1: Using MSIX interrupts with 9 vectors kernel: ixl1: Allocating 8 queues for PF LAN VSI; 8 queues active kernel: ixl1: Ethernet address: ... kernel: ixl1: PCI Express Bus: Speed 8.0GT/s Width x8 kernel: ixl1: SR-IOV ready kernel: ixl1: netmap queues/slots: TX 8/1024, RX 8/1024 kernel: ixl1: The device is not iWARP enabled /etc/iov/ixl1.conf --- PF { device : "ixl1"; num_vfs : 2; } DEFAULT { passthrough : true; } -- starting iovctl caused it to hang again: # service iovctl start logs: /var/log/messages: hh:mm:19 kernel: NMI ISA 38, EISA 0 hh:mm:19 kernel: NMI/cpu10 ... going to debugger hh:mm:21 kernel: ixl1: RX queue 0 still enabled! hh:mm:21 kernel: ixl1: RX queue 1 still enabled! hh:mm:21 kernel: ixl1: RX queue 2 still enabled! hh:mm:21 kernel: pci3: <network, ethernet> at device 0.80 (no driver attached) hh:mm:21 kernel: pci3: <network, ethernet> at device 0.81 (no driver attached) hh:mm:30 kernel: ixl1: WARNING: queue 0 appears to be hung! hh:mm:30 kernel: ixl1: WARNING: queue 1 appears to be hung! hh:mm:30 kernel: ixl1: WARNING: queue 2 appears to be hung! hh:mm:30 kernel: ixl1: WARNING: queue 3 appears to be hung! -> to recover I disabled iovctl in rc.conf and rebooted there is about 1Gbps of network traffic when we invoked these commands. I found similar errors here: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=221919 but I'm running a newer driver version than 1.9.9-k I did not try to disable TSO. Is there anything else I can provide you with to help solving this?