ChanServ changed the topic of #mirage to: https://mirage.io - bug cleaning day every first friday in month (14:00 UTC - late, next: July 6th) - next call June 27th 16:00 BST https://github.com/mirage/mirage-www/wiki/Call-Agenda - retreat 3rd-10th October http://retreat.mirage.io - this channel is logged at http://irclog.whitequark.org/mirage/ - MirageOS 3.1 is released - happy hacking!
yomimono has quit [Ping timeout: 268 seconds]
demonimin has joined #mirage
demonimin has joined #mirage
demonimin has quit [Changing host]
yomimono has joined #mirage
pagurus` has joined #mirage
pagurus has quit [Ping timeout: 240 seconds]
Haudegen has joined #mirage
Haudegen has quit [Read error: Connection reset by peer]
Haudegen has joined #mirage
mort___ has joined #mirage
Haudegen has quit [Remote host closed the connection]
mort___ has quit [Quit: Leaving.]
mort___ has joined #mirage
Haudegen has joined #mirage
mort___ has left #mirage [#mirage]
Haudegen has quit [Read error: Connection reset by peer]
yomimono has quit [Ping timeout: 265 seconds]
Haudegen has joined #mirage
Haudegen has quit [Read error: Connection reset by peer]
yomimono has joined #mirage
<yomimono> reynir: I totally failed to repro that uninstallability bug you mentioned on masto :( can you throw up output in a pastebin or something? just want to see what the solver says
<reynir> yomimono: Yes, what do you need the output from?
<reynir> And thanks for trying!!
jnavila has joined #mirage
<yomimono> reynir: probably `make depend` for the mirage-skeleton example you were trying (probably static_website_tls I guess?) is the best place to start
<yomimono> also `opam list`
<reynir> Ok! :)
<reynir> Oh, I can't remember if I mentioned I was doing this on arm64
<reynir> Output from opam list: http://p.reyn.ir/p/5ODhMkedBoM -- doing make depend now, but it takes some time on the raspberry pi
<reynir> Output from make depend (not very helpful): http://p.reyn.ir/p/F3PqmJvRIIc
<yomimono> reynir: oh! yes, that's likely to be relevant :)
<yomimono> in that case can I punt and ask you to report on solo5/solo5 ?
* yomimono looks at output
<yomimono> mato had this happen a lot too
<reynir> Oh
<yomimono> I don't think it's necessarily an arm64 problem but I don't remember what his solution was (it may have been "just move to opam2")
<reynir> I'm compiling a new switch just to try again without the packages I attempted to install manually
<reynir> bbl, groceries :)
<yomimono> there is an issue for this at https://github.com/ocaml/opam/issues/2717 and a bunch of debug output discussion in https://github.com/ocaml/opam/issues/3083 , but unless it happens with opam2 as well it's likely to be WONTFIX
<reynir> Ok
<reynir> I will try with opam2
jnavila has quit [Ping timeout: 240 seconds]
Abandoner58 has joined #mirage
<Abandoner58> Hi everyone, I've run into a problem trying to run the Mirage network example (mirage-skeleton/device-usage/network) on Xen and I wonder if anyone here could help. I've set up Xen networking with a bridge interface (which works with other unikernels, so it's not likely to be a Xen misconfiguration), configured with 'mirage configure xen --dhcp true' and specified the 'bridge' parameter in the 'vif' option of network.xl.
<Abandoner58> However, when I start the Xen guest, I see no DHCP traffic on the network bridge. The console of the Mirage unkernel shows very little, and configuring with parameter '--logs *:debug' doesn't give any more information than this:
<Abandoner58> MirageOS booting...
<Abandoner58> Initialising timer interface
<Abandoner58> Initialising console ... done.
<Abandoner58> Attempt to open(/dev/urandom)!
<Abandoner58> Unsupported function getpid called in Mini-OS kernel
<Abandoner58> Unsupported function getppid called in Mini-OS kernel
<Abandoner58> 2018-07-06 17:12:11 -00:00: INF [net-xen:frontend] connect 0
<Abandoner58> 2018-07-06 17:12:11 -00:00: INF [net-xen:frontend] create: id=0 domid=0
<Abandoner58> 2018-07-06 17:12:11 -00:00: INF [net-xen:frontend] sg:true gso_tcpv4:true rx_copy:true rx_flip:false smart_poll:false
<Abandoner58> Attempt to open(/dev/urandom)!
<Abandoner58> Unsupported function getpid called in Mini-OS kernel
<Abandoner58> Unsupported function getppid called in Mini-OS kernel
yomimono has quit [Ping timeout: 240 seconds]
<Abandoner58> ...and that's all there is. Since I have other non-mirage unkernels running on Xen with the network working just fine, it looks like this is an issue with Mirage, but I don't have any way to debug any further.
Haudegen has joined #mirage
<reynir> oh what the hell
<reynir> opam2 installed mirage 2.9.1
<reynir> Oh, older version of OCaml... -.-
yomimono has joined #mirage
<yomimono> meaning to clean bugs with an iterative deepening approach but accidentally doing depth-first search, the yomimono story
<reynir> :D
<yomimono> abandoner58: can you boot a non-networked unikernel like the hello world example?
<yomimono> should have said "non-networked MirageOS unikernel", e.g. mirage-skeleton/tutorial/hello
<Abandoner58> Yes, the hello world tutorial one works as expected
<yomimono> abandoner58: how about the same unikernel with a static IP configuration? does that give at least more boot messages?
<yomimono> I recently released a new version of the DHCP library which worked for me (tm) and passed tests but may have some stupid bug in it, is why I ask
<Abandoner58> I've tried it with a static IP before, and the logs say that the TDP and UDP interfaces were brought online with the address I assigned, but there is no actual traffic over the bridge interface and I can't telnet/netcat to any of the ports I expect to be open (as if the host just doesn't exist)
<yomimono> hang on, this sounds suspiciously like something from the last time I had to set up a Xen host, digging through my notes...
Abandoner58_ has joined #mirage
<Abandoner58_> oops, sorry, connection dropped. I'm back
Abandoner58 has quit [Ping timeout: 252 seconds]
<yomimono> abandoner58: are the other unikernels you have running on the same bridge as the mirageos ones?
<yomimono> (it should be 100% OK if they are, just want to make sure that the "other stuff works" data point is meaningful in the way I think it is)
<Abandoner58_> yes, they are on the same bridge
<Abandoner58_> the ones that work are all rumprun, if that matters
Abandoner58_ has quit [Client Quit]
Abandoner58 has joined #mirage
<Abandoner58> (dropped again...)
<Abandoner58> yes, they are on the same bridge
<reynir> yomimono: I get a conflict between mirage-no-solo5 and mirage-solo5 :D
<yomimono> abandoner58: cool, thanks. I might ask you to report at https://github.com/mirage/mirage since I'm failing to find whatever bit of arcana is tugging at my memory
<Abandoner58> OK, I'll put up an issue, thanks
<yomimono> thanks, sorry not to be of more help - away from my Xen machine (and all its notes) atm but can dig around some more later today
<yomimono> reynir: have you tried installing mirage-solo5 ? :)
<Abandoner58> No problem. Thanks for looking
<yomimono> opam prioritizes things you explicitly install ("roots") IIRC
<yomimono> abandoner58: lazy response is to check your dom0 firewall rules of course :)
<reynir> yomimono: Trying that now! Takes a while to do anything on an rpi :)
<yomimono> reynir: I hear that :D nocrypto is an especially painful install since it needs gmp, zarith, etc
<yomimono> when building for the cubieboards I'd set it to do that and then go do something else for half a day
<yomimono> lack of cross-compile hurts so much in that case :(
<reynir> haha
<reynir> I was wondering if you could do some qemu chroot thing to cross compile
<reynir> (I don't know what I'm talking about)
<yomimono> lortex is the expert :)
<Drup> yomimono: .... I just noticed the section you added
<yomimono> drup: :P
<Drup> Thanks, I'm extending it now
<yomimono> drup: it's OK to notice it by deleting it also :)
<reynir> Ah, found the cause: http://p.reyn.ir/p/KB/v+B0RiJ8
<Drup> No, the content is good, I'm just going to extend the explanations
<Drup> (also, result is not a GADT, but that's minor)
<yomimono> reynir: oh, indeed
<reynir> mirage-entropy says it can't run on arm64
<yomimono> https://github.com/mirage/mirage-entropy/issues/38 there's even an open issue!
<reynir> ooh, something to subscribe to!
<reynir> yomimono: thanks for the help, I would still be stuck & clueless :)
<yomimono> reynir: np, hope someone comes along who can help more materially soon :)
<kit_ty_kate> hannes: do you have something to retrieve the github handle from an email by any chance?
<yomimono> drup: I mean, that's not super minor, since it reveals that I totally don't understand GADTs ;)
<yomimono> although now that I think about it, that's a fairly minor concern for everyone else
<Drup> yomimono: well, from a purely syntactic point of view, if it's written "Foo of bar", it's not a GADT :D
<Drup> GADT must be "Foo : bar -> baz"
<Abandoner58> yomimono: I put up an issue on github just now https://github.com/mirage/mirage/issues/910
<yomimono> abandoner58: thanks :)
<yomimono> drup: also thanks :)
Abandoner58 has quit [Ping timeout: 252 seconds]
yomimono has quit [Ping timeout: 264 seconds]
<hannes> kit_ty_kate: I have some tooling for this, yes.
<hannes> kit_ty_kate: mostly hiding somewhere in the conex repository, if you let me know where and how to put it, I'm happy to do this.
<hannes> kit_ty_kate: my approach (as described in https://hannes.nqsb.io/Posts/Maintainers) is to (a) use opam authors etc., but more importantly the people who do the commits for the repo.
<hannes> (I think the github id should be stored in the opam file)
yomimono has joined #mirage
jnavila has joined #mirage
jnavila has quit [Ping timeout: 256 seconds]
Haudegen has quit [Remote host closed the connection]