Date: Fri, 1 Mar 2019 20:16:44 -0700 From: Rebecca Cran <rebecca@bluestop.org> To: freebsd-stable@freebsd.org Subject: EHCI interrupt storm (~24000 per second) on 12.0 (Lynx Point chipset) Message-ID: <af8240ea-fed9-e030-0e3f-6c82c337fcf8@bluestop.org>
next in thread | raw e-mail | index | archive | help
I have an Intel Server board (it's a UEFI Development Kit system, based on the S1200V3RPS) that I recently upgraded to 12.0-RELEASE, and today noticed the intr task taking up too much CPU time. I can't say whether this is a regression from 11.x or not. Looking at the output of "vmstat -i" there appears to be an interrupt storm: bcran@muon:~ % vmstat -i interrupt total rate irq4: uart0 248 4 irq16: ehci0 147 2 irq23: ehci1 1689451 24000 cpu0:timer 6398 91 cpu1:timer 3696 53 cpu2:timer 2246 32 cpu3:timer 2386 34 cpu4:timer 3152 45 cpu5:timer 2341 33 cpu6:timer 75101 1067 cpu7:timer 1646 23 irq265: igb0:rxq0 118 2 irq266: igb0:rxq1 133 2 irq267: igb0:rxq2 17 0 irq268: igb0:rxq3 259 4 irq269: igb0:aq 2 0 irq270: igb1:rxq0 2 0 irq275: ahci0 17450 248 Total 1804793 25639 This is after I tried setting hw.usb.ehci.lostintrbug=1 in /boot/loader and rebooting. The USB devices connected are: bcran@muon:~ % sudo usbconfig ugen1.1: <Intel EHCI root HUB> at usbus1, cfg=0 md=HOST spd=HIGH (480Mbps) pwr=SAVE (0mA) ugen0.1: <Intel EHCI root HUB> at usbus0, cfg=0 md=HOST spd=HIGH (480Mbps) pwr=SAVE (0mA) ugen1.2: <vendor 0x8087 product 0x8000> at usbus1, cfg=0 md=HOST spd=HIGH (480Mbps) pwr=SAVE (0mA) ugen0.2: <vendor 0x8087 product 0x8008> at usbus0, cfg=0 md=HOST spd=HIGH (480Mbps) pwr=SAVE (0mA) ugen1.3: <American Power Conversion Back-UPS XS 1300G FW:864.L9 .D USB FW:L9> at usbus1, cfg=0 md=HOST spd=FULL (12Mbps) pwr=ON (2mA) dmesg shows it being a Lynx Point USB controller: ehci1: <Intel Lynx Point USB 2.0 controller USB-A> mem 0xc2b29000-0xc2b293ff at device 29.0 on pci0 usbus1: EHCI version 1.0 usbus1 on ehci1 usbus1: 480Mbps High Speed USB v2.0 Strangely, em0 is failing to attach, but I guess that's unrelated? em0: <Intel(R) PRO/1000 Network Connection> port 0x5040-0x505f mem 0xc2b00000-0xc2b1ffff,0xc2b2b000-0xc2b2bfff at device 25.0 on pci0 em0: attach_pre capping queues at 1 em0: Setup of Shared code failed, error -2 em0: IFDI_ATTACH_PRE failed 6 device_attach: em0 attach returned 6 -- Rebecca Cran
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?af8240ea-fed9-e030-0e3f-6c82c337fcf8>