IRC channel logs

2021-03-31.log

back to list of logs

<youpi>it seems your user intr patch does work :D
<youpi>I can boot the debian installer with it, and netdde does work fine with it
<youpi>now trying to give it an ahci disk for rumpdisk to find it
<youpi>bleh, the installed doesn't see /dev/wd0
<youpi>but otherwise it does seem to be working completely fine!!
<youpi>(with machine q35, that is)
<Pellescours>yay
<youpi>539K 31 mars 00:31 gnumach*
<youpi>now that's a smaller kernel :)
<youpi>and free of gpl2-only source
<damo22>woot!
<youpi>bleh I made a mistake in partman-base
***snd is now known as snd1
<damo22>so that is with pic?
***Server sets mode: +nt
***Server sets mode: +nt
<Gooberpatrol66>very cool!
<youpi>damo22: with pic yes
<youpi>better test things one at a time
<youpi>always
<damo22>yes
<damo22>cool that it works with pic
<youpi>installation going on :D
<youpi>with rumpdisk
<damo22>:)
<damo22>rumpdisk supports cdroms too
<youpi>yes, though here it installs from the net
<youpi>not sure it actually detects the iso
<youpi>cd0(ahcisata0:2:0): Sense Error Code 0x50
<youpi>cd0(ahcisata0:2:0): Sense Error Code 0x50
<youpi>Mar 31 00:29:35 kernel: cd0(ahcisata0:2:0): Sense Error Code 0x50
<youpi>Mar 31 00:29:35 kernel: cd0(ahcisata0:2:0): Sense Error Code 0x50
<damo22>hmm weird
<damo22>is that in qemu?
<damo22>i fixed a bug in qemu for that i think
<youpi>ah, I didn't reapply the patch
<youpi>it seems quite slow, though
<youpi>like perhaps 2x-4x slower
<damo22>the OS?
<youpi>which is not really surprising for something that basically just writes on the disk :)
<youpi>the installation
<damo22>ok
<damo22>the write is single threaded and synchronous?
<youpi>it was as well in gnumach
<youpi>but yes, making it asynchronous will probably help
<youpi>and we really need to cluster page writes
<youpi>that is what kills our performance
<youpi>both read & write
<youpi>yes the cd0 warning is away with the patch
<damo22>:D
<damo22>what was the patch again? i forgot, was it related to IDE cache?
<youpi>there is also some validate_bcl
<youpi>also some MCH_HOST_BRIDGE_SMRAM_C_SIZE
<damo22>oh i think the io ports for the ahci overlapped with pci conf2
<damo22>so i did a hack to move them
<youpi>linux: Add option to disable Linux groups
<youpi>at last :D
<damo22>youpi: im getting somewhere with ACPICA in hurd userspace http://paste.debian.net/plain/1191783
<damo22>haha it compiles libacpica with zero warnings!
<damo22>and i only touched the glue code
<damo22>but it still segfaults at #1 0x08056180 in acpi_ns_root_initialize () at ../../libacpica/nsaccess.c:182
<damo22>youpi: can i give you a link to my hurd branch for libacpica, i dont really want to spam the mailing list for such a large patchset that may not be ready
<damo22> http://git.zammit.org/hurd-sv.git/log/?h=acpica-integrate
<damo22>it compiles with no warnings, but something is wrong
<damo22>maybe you can find the problem in the glue code i wrote, its much smaller than the acpica code itself
<youpi>you can post the glue to the list
<damo22>ok
<damo22>gnite