DragonFly bugs List (threaded) for 2004-07
Re: Every IRQ is listed as stray in 1.1-CURRENT
I'm not sure why you sometimes get them and sometimes don't, but
all they are are placemarkers. Internally the system has a 'FAST'
and a 'SLOW' interrupt vector for each irq, which is why you are seeing
the irqs listed twice, but the strays don't actually do anything unless
a real stray interrupt occurs (and then you get a count).
So, basically, just ignore them. Stray interrupts generally occur
when the kernel has not properly routed interrupts during boot (which
can happen for all sorts of reasons on modern machines, especially
laptops). The stray functions are like safeties in the end-zone... they
prevent the machine from livelocking on an improperly routed interrupt.
IRQ7 will also sometimes generate stray interrupts.
Everything else looks fine. The clock interrupt works very differently
in DragonFly then in FreeBSD. FreeBSD uses a fixed periodic timer
interrupt while DragonFly uses a fine-grained variable-interval timer
interrupt. The number of interrupts/sec you get will depend on a
number of factors but will generally not exceed 250 interrupts/sec.
:Not sure if this is an actual problem or just a cosmetic visual
:holdover from the FBSD sources, but every IRQ on my laptops is listed
:twice: once as stray, once as associated with a device. This occurs
:with ACPI enabled and disabled.
:Here's a sample (ACPI disabled):
:[fcash@spark ~]$ vmstat -i
:interrupt total rate
:??? 0 0
:stray irq0 0 0
:stray irq1 0 0
:stray irq2 0 0
:stray irq3 0 0
:stray irq4 0 0
:stray irq5 0 0
:stray irq6 0 0
:stray irq7 0 0
:stray irq8 0 0
:stray irq9 0 0
:stray irq10 0 0
:stray irq11 0 0
:stray irq12 0 0
:stray irq13 0 0
:stray irq14 0 0
:stray irq15 0 0
:cbb0 irq10 0 0
:cbb1 irq5 0 0
:ata0 irq14 989 31
:ata1 irq15 5 0
:uhci0 irq11 0 0
:mux irq5 0 0
:mux irq11 106 3
:fdc0 irq6 0 0
:atkbd0 irq1 54 1
:psm0 irq12 0 0
:ppc0 irq7 0 0
:clk irq0 6568 211
:Total 7722 249
:Every now and then I can boot and have a list without stray IRQs, but
:I can find no rhyme or reason for it.
:I've also noticed that the rate for the clock is entirely random
:running from 180-ish to 225-ish. Under FreeBSD 5-CURRENT, it was
:This is on a 1.1-CURRENT system with sources from this morning (just
:finished a quickworld/kernel build).
:Freddie Cash, CCNT CCLP Helpdesk / Network Support Tech.
:School District 73 (250) 377-HELP [377-4357]