IRC channel logs

<ArneBab>Is the publishing of the wiki broken? I do not see this change on the website: https://git.savannah.gnu.org/cgit/hurd/web.git/commit/?id=09ca4da2317d428d027c1f464534a470dbc06236 — see https://www.gnu.org/software/hurd/community/meetings/fosdem_2016.html (missing: '(including video)')

<youpi>the publishing of the wiki on gnu.org is done by hand by tschwinge

<ArneBab>it looks like something got lost — i.e. https://www.gnu.org/software/hurd/community/meetings/fosdem_2019.html is not there

<ArneBab>thank you!

<damo22>youpi: i dont think its a good idea for me to mail patches in for librump, it has a huge new tarball, but its all here http://git.zammit.org/rumpkernel-debian.git/log/?h=develop

<damo22>as mentioned here https://mail.gnu.org/archive/html/bug-hurd/2021-12/msg00022.html

<ArneBab>I now blogged about the checkperms translator: https://www.draketo.de/software/hurd-authorization-translator — it’s cool to see what’s actually practicable to build!

<Pellescours>youpi: in gnumach, ares the files util/putcharsh{h,c} and util/puts.c used ?

<youpi>Pellescours: possibly, depending whether gcc optimizes printfs into puts

<Pellescours>youpi: it seems to be used, the pagefault I had when removing linux disapeared when I restored the puts & putchar files

<Guest51>hello

<Guest51>I would like to to contribute to GNU hurd porject but I don't know where to start; therefore I joined this irc to ask what steps should I take to be able to develop HURD which books should I read or content

<damo22>Pellescours: maybe check all the warnings in gnumach compilation? if the symbol is missing wont it complain?

<damo22>how would it even link

<Pellescours>I don’t know but when I restored the files, my gnumach stops to page fault

<Pellescours>and otherwise I’m running in a gnumach without a piece of linux inside :D

<damo22>nice

<damo22>git grep for the symbols

<Pellescours>I just have 1 issue when I poweroff but I don’t know why yet

<Pellescours>wait actually I just removed it again and re-tested, and it worked… So the panic was not related to this and a recent gnumach commit seems to have fix the real issue

<damo22>ok sounds better

<Pellescours>I checked in the gnumach binary (build with the files), puts and putchars are not in, so we can remove them safely

<damo22>maybe there is a way to use the linker that drops unused symbols

<damo22>then you can run nm and see which ones are not needed

<Pellescours>oh yeah probably

<damo22>by comparing to previous one

<Pellescours>damo22: do you have an idea of why the read timeout when I poweroff? It’s a bit anoying because the poweroff is long and at the end ext2fs crashes and instead of shuting down the vm reboots

<damo22>read timeout?

<damo22>mine unmounts i think

<Pellescours>I posted a message yesterday about it

<Pellescours>wd0d: device timeout reading fsbn 2491836 of 2491836-2491839 (wd0 bn 2491836; cn 2472 tn 0 sn 60) retrying

<damo22>hmm not sure

<Pellescours>it appears just after DHCPRELEASE … and after theses logs it says that ext2fs crashed

<damo22>sounds like an interrupt issue?

<Pellescours>I though to this to but I don’t know how I can investigate since it’s during poweroff and I don’t have a shell at this step

<damo22>you could try simulating releasing the net irq

<damo22>see if the disk irq is affected

<Pellescours>how do I do this?

<damo22>uhh, somehow stop using the net translator?

<Pellescours>I will try to mount another disk and unmount it to see if it happens

<damo22>kill netdde?

<damo22>but you want it to shutdown the netdde clean i think

<damo22>or maybe just kill it and wait for the timeout on the irq

<Pellescours>killing netdde makes appear a message about an undelivered irq

<damo22>that is the last net irq

<damo22>yes

<Pellescours>and the netdde actualy still running

<damo22>ah it probably restarted

<Pellescours>settrans -s /dev/netdde <<< resource busy

<damo22>im not sure

<damo22>was the undelivered irq the net one?

<damo22>or a disk one

<Pellescours>I don’t know and I scratch my term

<damo22>it should print the number

<Pellescours>I rebooted and I didn’t had the read timeout issue

<damo22>maybe it was related to your gnumach?

<Pellescours>no, it appears with the packaged gnumach too

<damo22>ok

<damo22>have we confirmed that interrupts are working correctly in librump 9.99?

<damo22>i mean, all i did was compile and mount it

<damo22>maybe attach gdb to rumpdisk and see if you can find the interrupt thread etc

<damo22>but you cant boot off the same disk

<Pellescours>I will mount another disk using rumpdisk and use ide for /

<damo22>you could compile pci-userspace with the debug flag

<damo22>that would print something on the console before and after each disk interrupt

IRC channel logs

2021-12-30.log