Date: Mon, 1 Jun 2009 09:41:41 -0500 From: Dan Nelson <dnelson@allantgroup.com> To: Morgan =?utf-8?B?V2Vzc3Ryw7Zt?= <freebsd-questions@pp.dyndns.biz> Cc: FreeBSD Mailing List <freebsd-questions@freebsd.org> Subject: Re: vmstat questions Message-ID: <20090601144141.GD90077@dan.emsphone.com> In-Reply-To: <4A238FD1.1010708@pp.dyndns.biz> References: <4A238FD1.1010708@pp.dyndns.biz>
next in thread | previous in thread | raw e-mail | index | archive | help
In the last episode (Jun 01), Morgan Wesström said: > Hi. This is my "vmstat -i" from my newly installed 7.2-RELEASE-amd64: > > # vmstat -i > interrupt total rate > irq4: sio0 1105 0 > irq17: em1 uhci1+ 10921807 182 > irq19: uhci3++ 8196905 136 > cpu0: timer 117164643 1956 > irq256: em0 7346687 122 > irq257: re0 166625 2 > cpu1: timer 117164471 1956 > Total 260962243 4358 > > For a long time I've tried to find out what the + and ++ means. Can > anyone shed some light on that? > > Also, where did my atapcis on irq19 go? I'm pretty sure I had them > listed there in 7.0-RELEASE. There is a very small buffer in the kernel for recording which devices are assigned to which IRQs in a human-readable format (MAXCOMLEN: 19 characters per IRQ, including the "irqNNN:" text). Code in /sys/kern/kern_intr.c:intr_event_update() truncates the list and adds a + for each omitted device name. http://fxr.watson.org/fxr/source/kern/kern_intr.c#L194 I'm not even sure why MAXCOMLEN is being used at this point; it's not storing a command name. http://fxr.watson.org/fxr/source/sys/interrupt.h#L104 I don't know what the repercussions would be if you increased MAXCOMLEN to 40, but you can try bumping it and see what breaks :) -- Dan Nelson dnelson@allantgroup.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090601144141.GD90077>