avsm changed the topic of #mirage to: mirage 2 released! party on!
brson has quit [Ping timeout: 265 seconds]
brson has joined #mirage
seangrove has joined #mirage
seangrove has quit [Ping timeout: 240 seconds]
brson has quit [Ping timeout: 265 seconds]
brson has joined #mirage
brson has quit [Ping timeout: 252 seconds]
brson has joined #mirage
brson has quit [Ping timeout: 265 seconds]
rgrinberg has quit [Ping timeout: 250 seconds]
rgrinberg has joined #mirage
brson has joined #mirage
copy` has quit [Quit: Connection closed for inactivity]
jermar has joined #mirage
rgrinberg has quit [Quit: WeeChat 1.5]
brson has quit [Quit: leaving]
jermar has quit [Ping timeout: 265 seconds]
tet has quit [Ping timeout: 244 seconds]
jermar has joined #mirage
tet has joined #mirage
mort___ has joined #mirage
GemmaG has joined #mirage
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted async-udp at b07730c: https://git.io/vilkm
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted bitstring at 854c81d: https://git.io/vilks
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted camlp4.opt at 408c62f: https://git.io/vilkn
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted dead-code-elim at 527bc83: https://git.io/vilkc
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted fs-probe at cb49ca4: https://git.io/vilk4
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted iteratee-io at 0b1c315: https://git.io/vilkE
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted master-20131208 at 047cf32: https://git.io/vilkg
miragebot has joined #mirage
<miragebot> [mirage] avsm deleted newlib_removal at 5c2c09b: https://git.io/vilka
miragebot has left #mirage [#mirage]
miragebot has joined #mirage
<miragebot> [mirage] avsm deleted ocamlbuild at 336feb7: https://git.io/vilkr
miragebot has left #mirage [#mirage]
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted ocamlclean at 76e797f: https://git.io/vilkK
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted ofelia at 460c560: https://git.io/vilk6
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted old-master at 3d886b5: https://git.io/vilkP
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted one-build at 3035408: https://git.io/vilk1
miragebot has joined #mirage
miragebot has joined #mirage
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted os-bigarray at 5e744af: https://git.io/vilkD
miragebot has joined #mirage
<miragebot> [mirage] avsm deleted perf-experiments at 316b735: https://git.io/vilky
miragebot has left #mirage [#mirage]
miragebot has joined #mirage
<miragebot> [mirage] avsm deleted test-framework at 95ea0c4: https://git.io/vilkQ
miragebot has left #mirage [#mirage]
miragebot has joined #mirage
<miragebot> [mirage] avsm deleted xen-blkif at 359c525: https://git.io/vilk5
miragebot has left #mirage [#mirage]
<miragebot> [mirage] avsm deleted ocp_build at 4f2628e: https://git.io/vilIR
miragebot has left #mirage [#mirage]
miragebot has joined #mirage
apache2_ has joined #mirage
apache2 has quit [Remote host closed the connection]
<mattg> https://github.com/mirage/mirage-clock/pull/14 is really fixed now, mirage-clock-unix and tests should work on both osx and linux
<mato> hannes: the binary you linked to in #80 triple faults for me as well
<hannes> mattg: travis has OSX support (look at e.g. opam-repository how to add OSX into the matrix), maybe mirage-clock-* repo should run a MacOSX travis build slave as well
<hannes> mato: good news!?
<mato> hannes: also, it doesn't look like it's built against the latest master, since the bss is too small (#80 statically allocates tons of pkt buffers, so i see a BSS of ~25mb)
<mato> hannes: (which i need to fix, but that's a different issue)
<mato> hannes: having said that, test_ping_serve on master doesn't acutally work, but at least it doesn't triple fault
<hannes> (and now lunch)
<mattg> oooh, interesting hannes, i will investigate. but first, lunch
<mato> hannes: interesting, your binary has xmit_bufs and recv_bufs in the wrong place (both at the same address and in COMMON rather than .bss)
<mato> hannes: so that's definitely wrong
<mato> ugh
<mato> your linker is putting a bunch of stuff in the wrong place, it seems
<mato> which linker are you using again, the ancient freebsd "ld"?
mort___ has quit [Quit: Leaving.]
<mato> ok, i think i know how to fix this, and the bitrot in test_ping_serve along the way
<hannes> mato: oh, you might be right that the test there is still using ancient ld..
<hannes> sorry about that, but I got the same issue for mirage-skeleton/ping
<mato> yes, you will get the same
<hannes> also using /usr/local/bin/ld I get an issue at the same location..
<mato> right
<mato> i don't particularly want to rabbithole into fixing the COMMON issue, so i'll just eliminate all truly global variables instead
<mato> looks like i'll have to kill test_ping_serve entirely, not really worth trying to "fix" it
<hannes> fair enough
<hannes> mato: I uploaded the one with new ld to http://berlin.ccc.de/~hannes/test_ping_server.virtio.newld
<hannes> (in case this is of any interest)
<mato> same problem
<mato> hannes: are you building with clang or with gcc?
<mato> hannes: ok, can you try with this branch? https://github.com/mato/solo5/tree/kill-all-globals
<mato> (you'll need to use mirage to provide the tcp stack)
<hannes> clang
<hannes> mato: I get lots of output in the form of got irq 5 at 0x7275b777a15
<mato> ok, that's better
<hannes> and no ping gets replied, but also no crash
<mato> what does the start of the console log say?
<mato> (about interrupt lines)
<hannes> [13:21 hannes@aternefas:ping] sudo bhyve -A -H -P -s 0:0,hostbridge -s 1:0,lpc -s 2:0,virtio-net,tap0 -l com1,stdio -m 256M test
<hannes> | ___|
<hannes> \__ \ ( | | ( | ) |
<hannes> __| _ \ | _ \ __ \
<hannes> ____/\___/ _|\___/____/
<hannes> multiboot: Using memory: 0x100000 - 0x10000000
<hannes> TSC frequency estimate is 2594162380 Hz
<hannes> virtio_config: device_id=1, interrupt_line=5
<hannes> Found virtio network device with MAC: 00 a0 98 cc 6a 96
<hannes> Solo5: new bindings
<hannes> STUB: getenv() called
<hannes> Netif: plugging into tap0 with mac 00:a0:98:cc:6a:96
<hannes> Netif: connect tap0
<hannes> starting...
<hannes> 2016-09-08 15:34.50: INF [ethif] Connected Ethernet interface 00:a0:98:cc:6a:96
<hannes> 2016-09-08 15:34.50: INF [arpv4] Connected arpv4 device on 00:a0:98:cc:6a:96
<hannes> got irq 5 at 0x746b28bcea6
<hannes> got irq 5 at 0x746b28f11fa
<mato> ok, can you add "case 0x05:" to the switch in kernel/virtio/low_level_interrupts.c:low_level_handle_irq() and rebuild?
<mato> (horrible hack, but should fix the problem)
<hannes> rebuilding
<hannes> no data via tcp..
<hannes> 2016-09-08 15:48.06: INF [arpv4] ARP: timeout 10.0.0.1
<hannes> is what the ping unikernel tells me
<mato> that's normal
<mato> (i think)
<mato> can you ping the unikernel at all?
<hannes> on the tap0 interface I get via tcpdump:
<hannes> 13:29:10.625080 ARP, Ethernet (len 6), IPv4 (len 4), Request who-has 10.0.0.2 tell 10.0.0.1, length 28
<hannes> 0x0020: 0000 0000 0000 0a00 0002
<hannes> 0x0010: 0800 0604 0001 00bd c781 f900 0a00 0001
<hannes> 0x0000: ffff ffff ffff 00bd c781 f900 0806 0001
<hannes> 13:29:10.625217 00:00:00:00:00:00 > 00:00:00:00:00:00 Null Information, send seq 0, rcv seq 0, Flags [Command], length 46
<hannes> 0x0000: 0000 0000 0000 0000 0000 0000 0000 0000
<hannes> 0x0010: 0000 0000 0000 0000 0000 0000 0000 0000
<hannes> 0x0020: 0000 0000 0000 0000 0000 0000 0000 0000
<hannes> 0x0030: 0000 0000 0000 0000 0000 0000
<hannes> I configured host to be 10.0.0.1 (I should then see arp requests and replies from the unikernel to the host, shouldn't I)?
<mato> yes
<hannes> there is no icmp echo reply from the ping unikernel, but it seems a bunch of zeroes
<mato> interesting
<hannes> (there's not even arp replies from the unikernel)
<mato> right, so there's something wrong on the tx path
<mato> still, much better than triple-faulting
<hannes> +1
<hannes> well, not sure whether the rx path is good..
mort___ has joined #mirage
<mato> ok, one thing at a time, i'll finish fixing the linker issue first
<hannes> mato: fair enough, just mention my name here when you want me to test something :)
<mato> and see if i can make a simple fix for #32, though that really calls for a proper cleanup of the irq handler code
<mato> hannes: sure. thanks!
<hannes> (will procrastinate with some opam signing test cases again)
copy` has joined #mirage
GemmaG has quit [Quit: Leaving.]
yomimono has joined #mirage
yomimono has quit [Ping timeout: 276 seconds]
rgrinberg has joined #mirage
jermar has quit [Ping timeout: 265 seconds]
GemmaG has joined #mirage
GemmaG has quit [Client Quit]
GemmaG has joined #mirage
yomimono has joined #mirage
GemmaG has quit [Ping timeout: 276 seconds]
agarwal1975 has joined #mirage
mort___ has quit [Quit: Leaving.]
GemmaG has joined #mirage
GemmaG has quit [Ping timeout: 244 seconds]
yomimono has quit [Ping timeout: 244 seconds]
mort___ has joined #mirage
yomimono has joined #mirage
brson has joined #mirage
jermar has joined #mirage
brson has quit [Ping timeout: 260 seconds]
brson has joined #mirage
brson has quit [Ping timeout: 260 seconds]
djs55 has quit [Quit: Leaving.]
brson has joined #mirage
mort___ has quit [Quit: Leaving.]
rgrinberg has quit [Ping timeout: 240 seconds]
<mato> hannes: ok, current solo5 master now has the fixes for both globals and irq 5 (just the hack)
<mato> hannes: i've also rewritten test_ping_serve (since it *is* useful), so you should be able to test with that
<mato> hannes: however, it does not do ARP, so you want one of:
<mato> ping -b -I tap100 10.0.0.255
<mato> sudo arping -i tap100 52:54:00:12:34:56
<mato> (replacing the MAC with whichever MAC it reports on startup)
rgrinberg has joined #mirage
agarwal1975 has quit [Quit: agarwal1975]
agarwal1975 has joined #mirage
<hannes> mato: aha. so, the ping test actually sends back replies, but they're all 0
<hannes> (as reported 4.5 hours ago)
<hannes> and exactly 58 bytes
<mato> hannes: great..., so there's something wrong on the tx path. can you file an issue?
<mato> hannes: still much better than as of this morning :)
<hannes> mato: yes, thank you very much!
<hannes> I'll throw in some printfs and look whether I can come a bit further
djs55 has joined #mirage
yomimono has quit [Ping timeout: 240 seconds]
yomimono has joined #mirage
yomimono has quit [Ping timeout: 244 seconds]
GemmaG has joined #mirage
insitu has joined #mirage
GemmaG has quit [Ping timeout: 264 seconds]
insitu has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rgrinberg has quit [Ping timeout: 244 seconds]
rgrinberg has joined #mirage
agarwal1975 has quit [Quit: agarwal1975]
jermar has quit [Ping timeout: 244 seconds]
GemmaG has joined #mirage
GemmaG has quit [Ping timeout: 244 seconds]