hmm someone posted on the openbsd mailing list having issues booting openbsd on arp oh he's on old kvm node actually his cpu flags look diff he seems to have sse4 support and i don't. mercutio, I tried 5.6 the other day and was not able to get it booted mercutio: this is what my flags look like on kvr30 : flags : fpu de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx lm up unfair_spinlock pni cx16 popcnt hypervisor lahf_lm abm sse4a i'm actually using a snapshot OpenBSD 5.6-current (MEH) #0: Sun Nov 9 15:22:48 NZDT 2014 but it's a recent snapshot cpu0: QEMU Virtual CPU version 1.0, 2667.27 MHz cpu0: FPU,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,MMX,FXSR,SSE,SSE2,SSE3,VMX,CX16,NXE,LONG,LAHF That's what she said!! i wouldn't want to say that reading it is ok they're not in the same order, but it seems i have vmx and you don't? unless vmx and hypervisor is the same thing maybe it's sse4.2 that shows as SSE4 on openbsd. his showed as: cpu0: QEMU Virtual CPU version 0.12.3, 2600.56 MHz cpu0: FPU,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,MM X,FXSR,SSE,SSE2,SSE3,CX16,POPCNT,NXE,LONG,LAHF,ABM,SSE4A,PERF That's what she said!! err it has ABM but I don't. actually that means AMD I think. mkb: where was your boot stopping? mercutio, nvram: invalid checksum; same place as the mailing list kvr29 i get invalid checksum too scsibus3 at softraid0: 256 targets root on wd0a (f68d22bdb905441b.a) swap on wd0b dump on wd0b clock: unknown CMOS layout his seemed to get up to scsibus3 get on up and DANCE and not say the root or clock i'm on kvr15 i think kvr29 is amd are you 2.6 ghz too? cpu0: QEMU Virtual CPU version 0.12.3, 2600.66 MHz it's 2.66ghz on intel it seems hmm ok why does it say different virtual cpu version i assumed 1.0 was more recent than 0.12.3 but it looks like i was wrong, and that signifys amd not intel. the version? maybe when up_the_irons he can clue us in more yeah. I suppose someone needs to figure out how to get ddb or the virtualization software to dump the registers damn amazon is expensive for shipping to here :( $28.34 for a video card. yeah i dunno if dumping would help much did you try the -usb thing? and -mpbios for good measure I did try mpbios; didn't help. Then I gate up and installed 5.5 back heh i jumped from 5.5 i386 to 5.6 amd64 well it was in betwen 5.5 and 5.6 on i386 i did it a while ago, but i just recompiled my kernel shifting from GENERIC mercutio: I thought 1.0 was newer than 0.12.3 also phlux: omg love the ascii art from the other day :) up_the_irons: i htink it's ths same and diff cpu definition? up_the_irons: is kvr29 and kvr15 the same version of kvm? $ ssh kvr15 lsb_release -rd Description: Ubuntu 12.04.4 LTS Release: 12.04 $ ssh kvr29 lsb_release -rd Description: Ubuntu 10.04.4 LTS Release: 10.04 $ kvr15 is newer oh ok thanks :) np:) is kvr29 amd? yes ok kvr15 is intel that's where i went wrong amd doesn't necessarily mean new kvm yeah up_the_irons: at least two people are having problems with openbsd 5.6 on the old kvm metal detector: http://i.imgur.com/9he0xN8.jpg unless that was the same guy as on the mailing list :) mercutio: yeah i saw a ticket or two about that going to need to migrate those guys or they downgrade migrate seems most sensible to me then virtio is supported too mind oyu there's no knob in cp to enable that $60,000 AUD in CAD @wa $60,000 AUD in CAD convert convert $60000 (US dollars) to Australian dollars to Canadian dollars;C$68230 (Canadian dollars) (at current quoted rate);$60000 (US dollars) (at current quoted rate);, 1-year minimum->C$65660 (December 20, 2013->11 months ago), 1-year maximum->C$71690 (April 23, 2014->7 months ago), 1-year average->C$69150 (annualized volatility: 6.3%) @exch 60000 AUD CAD ^ also 60000 AUD -> 51769.748864948 USD -> 58843.722659101 CAD (as of Mon, 10 Nov 2014 17:00:48 -0800) mercutio: well yeah i'd have to do it manually on my end @exch 5000 MYR USD 5000 MYR -> 1496.961617006 USD (as of Mon, 10 Nov 2014 17:00:48 -0800) @exch 60 EUR USD 60 EUR -> 74.567601122988 USD (as of Mon, 10 Nov 2014 17:00:48 -0800) @exch 100 EUR USD 100 EUR -> 124.27933520498 USD (as of Mon, 10 Nov 2014 17:00:48 -0800) @exch 16.90 USD MYR 16.90 USD -> 56.4476731 MYR (as of Mon, 10 Nov 2014 17:00:48 -0800) anyone know if the CPU NX flag can be passed on to guests? I can see our CPUs support it, but guests don't see it. I've been doing some googling and not coming up with much... nothing bad should happen up_the_irons it just means that the OS can mark pages as non executable. so that data pages etc can be marked that way to reduce the chances of security issues mercutio: yeah i understand, but how do i enable that flag for a guest... can't figure it out i think you haev to chhange from defining generic cpus to specific flags qemu -cpu qemu64,+ssse3,+sse4.1,+sse4.2,+x2apic err try +nx? qemu -cpu host actually that may be better hmm.. the only problem with cpu host is if doing live migration which you don't do anyway. it just means it exposes the raw cpu stuff yeah up_the_irons: Here is my pull request: https://github.com/up_the_irons/arpnet-toolbox/pull/3 Made some changes to the asn.sh file as per your input I feel like a lot of this is redundant For instance the if two= "AS" ... else ASN=${1##AS} Those 5 lines alone could be pared down to 2 lines But then we get to the point where you extract digits from $1... Why did we strip off "AS" if we can just pull the number out anyways? to limit input like 123anything What do you mean? if you pull the number out alone, then even text after the number will be valid input say AS123textgoeshere If I feed AS1234ABCD I get 1234 back out up_the_irons didnt like that behaviour it should be invalid only valid input should be as1234, AS1234 and 1234 I would argue that we could/should leave the value validation to the backend (once it's been massaged to a suitable format) mnathani: This will will replace all of that logic: egrep -o '^(AS)?[0-9]+$' ASN=$(echo "$1" | egrep -o '^(AS)?[0-9]+$') if[ "x$ASN" = "x" ] ; then usage ; exit 1 ; fi ASN will either have a valid number, or it will be empty oh forgot the -i to egrep egrep -i -o '^(AS)?[0-9]+$' -i case insensitive, -o to print only the portion that matched the regex The regex itself will only match "as1234" and "1234", any trailing non-numbers will cause the pattern to fail and return a blank (technically the check alone could be done with: if ! echo "$1" | egrep -q -i '^(AS)?[0-9]+$' ; then ... ; fi) (or if you like to have fun, [ echo "$1" | egrep -o '^(AS)?[0-9]+$' ] || { usage ; exit 1 } yea, looks like that will work more elegant, less code And arguably more readable (because it's concise) and you didnt need to use like 5 extra variables either Huh, looks like all (most?) of my ipv4 states were cut v4 IRC connections, OpenVPN connections (v4 udp) all flopped damn that dead beef.