IRC channel logs


back to list of logs

***Server sets mode: +nt
<damo22>hmm pthread_mutex_lock(lock) locks up the terminal
<damo22>i mean it locks up gdb
<damo22>i found something, but cant fix it... the second time the same region gets mapped it maps to 0x0:
<damo22>the first mapping is the probe of the device, which works, and the second time it is mapped it is to try to open the device
<damo22>could it be some logic in libpciaccess that stops you mapping the same region twice?
<damo22>when i step through it, it seems to call pci_device_map_range in the middle of the routine
<damo22>and the gdb listing makes no sense
<damo22> cant figure this out
***rekado_ is now known as rekado
<damo22>heh fixed!!!
<damo22>libpciaccess was not allowing reuse of mappings
<damo22>i mounted a real disk with rump debian packages, theres still a small bug but i think i know why
<damo22>requires patches to libpciaccess, hurd and rumpkernel
<youpi>yay :)
<damo22>i guess i should fix the remaining bug tomorrow and then work on upstreaming the patches that dont involve rump first
<damo22>because rumpkernel is a huge upgrade, basically a new upstream tree
<damo22>problem is, hurd would require rumpkernel as a build-dep to build rumpdisk
<damo22><-- bed
<youpi>on the long run we want to keep rumpkernel updated anyway
<Pellescours>hi, is someone have knoledge about x86_64 bit architecture ? I'm trying to work on gnumach x86_64 but I have linking issues and my knowledges are limited on this part
***dp is now known as bq