IRC channel logs

2026-01-05.log

back to list of logs

<gnucode>nexussfan: the hurd.gnu.org site is very out of date.
<gnucode>:(
<gnucode>there's some technical reason for why that is.
<nexussfan>Hurdos.com is actually behind
<nexussfan>It doesn't have the latest qoth
<nexussfan>I don't use darnassus.sceen.net as it's unreliable
<gnucode>nexussfan: I can update it...I just need to craft the rsync command.
<gnucode>hurdos.com uses openBSD and I haven't figured out how to install ikiwiki on it.
<gnucode>I might have my guix-system server start using hurdos.com
<gnucode>it would be a little easier maybe to update hurdos.com/wiki
<nexussfan> https://www.gnu.org/software/hurd/community/weblogs/ArneBab/niches_for_the_hurd.html is a great article
<gnucode>nexussfan: I also feel like I should add my hurd-related blogs to the community/webblogs
<nexussfan>Yeah that would be nice
<damo22>(gdb) bt
<damo22>#0 thread_dispatch (thread=0x0) at ../kern/sched_prim.c:939
<damo22>#1 thread_invoke (old_thread=0x0, old_thread@entry=0xffffffffdc0edb40,
<damo22> continuation=continuation@entry=0xffffffff8105e010 <idle_thread_continue>, new_thread=0xffffffffdc0ed900)
<damo22> at ../kern/sched_prim.c:821
<damo22>#2 0xffffffff8105dec3 in thread_block (continuation=continuation@entry=0xffffffff8105e010 <idle_thread_continue>)
<damo22> at ../kern/sched_prim.c:891
<damo22>#3 0xffffffff8105e4a3 in idle_thread () at ../kern/sched_prim.c:1813
<damo22>#4 0xffffffff8105cfe9 in thread_continue (old_thread=<optimized out>) at ../kern/sched_prim.c:847
<damo22>#5 0xffffffff81073304 in Thread_continue () at ../x86_64/cswitch.S:103
<damo22>...
<damo22>#10 0xffffffff81071561 in switch_context (old=0x0, continuation=0xffffffff8105cfb0 <thread_continue>, new=0x0)
<damo22> at ../i386/i386/pcb.c:393
<damo22>Backtrace stopped: previous frame inner to this frame (corrupt stack?)
<damo22>(gdb)
<damo22>this almost boots but tries to resume a thread = 0x0
<damo22>ok so you cant ljmp from a register
<damo22>how annoying, that means you have to patch all the relocations from bsp once and then use one memory location shared among all APs
<rrq>20:07 < damo22> ok so you cant ljmp from a register
<rrq>(oops. the cat got busy)
<damo22>hmm but 0: ff 28 ljmp *(%eax)
<damo22>that seems to assemble
<damo22>ah but that is still a memory location
<damo22>ljmp *%eax does not assemble
<damo22>how do i jump to a 64 bit code segment with a 32 bit absolute address?
<damo22>ljmpq does not work in 32 bit mode
<damo22>i suppose the gdt specifies the segment with SZ_64
<youpi>you can ljmpq *foo
<youpi>where foo has a .quad ptr and a .word segment
<youpi>at least boothdr.S does some ljmp to 64b
<damo22>i think the ljmp has to be to a 32 bit absolute address but the gdt can specify a 64 bit code segment, so the result is jumping to 0x0000000xxxxxxxx in 64 bit segment?
<damo22>+0