channel logs for 2004 - 2010 are archived at http://tunes.org/~nef/logs/old/ ·· can't be searched
#osdev2 = #osdev @ Libera from 23may2021 to present
#osdev @ OPN/FreeNode from 3apr2001 to 23may2021
all other channels are on OPN/FreeNode from 2004 to present
http://bespin.org/~qz/search/?view=1&c=osdev&y=18&m=11&d=13
12:10:56 * ybyourmom has to get ARM-hyp working on an ARM board
12:10:59 <ybyourmom> TX2
12:11:07 <ybyourmom> Out of the blue
12:11:09 <ybyourmom> kek
12:11:30 <klange> I have a TX2 here somewhere...
12:12:01 <klange> Was on my desk for months until we finally got a Xavier.
04:51:59 <ybyourmom> ARMv8 is cleaner if you're willing to actually read all the stuff they've put in the manual
04:52:08 <ybyourmom> Bit of a learning curve from v7
04:52:46 <geist> word to that
04:53:02 <geist> i enjoy it, but it's definitely a bit of a learning curve
04:54:42 <ybyourmom> The thing is that it's more complex feature wise, but it's easier to understand because there's less [ Implementation defined ]
04:54:56 <klange> unfortunately, I'm not going to get much lower level than spinning up Debian in Docker on these ARM boards...
04:55:26 <ybyourmom> So it ends up being easier to model in your mind because less of the manual is sort of "schroedinger's feature"
04:55:58 <ybyourmom> Where you see that it's a feature, but you can't do anything with that in your mind because you still have to go and read 2 other manuals before you know whether or not you can actually use it
06:52:48 <geist> ybyourmom: right. ARMv8 is pretty clean all in all. the main complexities are in caching and memory model and whatnot
06:52:56 <geist> which are somewhat more complex and ambiguous
08:32:35 <ArsenArsen> what boot protocol should I start with to implement a simple linux bootloader?
08:33:09 <klange> Why do you want to implement a Linux bootloader? And for what platform? Linux isn't even interesting to boot since it basically ignores you and reinitializes everything itself.
08:34:34 <lkurusa> ArsenArsen: linux has its own boot protocol
08:34:39 <lkurusa> you can find the documentation here: https://elixir.bootlin.com/linux/latest/source/Documentation/x86/boot.txt
08:35:04 <ArsenArsen> yes but which one should I start with in order to not burden myself too much?
08:35:16 <lkurusa> what do you mean *which one*
08:35:26 <lkurusa> do you mean the architecture?
08:35:29 <ArsenArsen> which version?
08:35:41 <ArsenArsen> nah I'm targeting x86(_64)
08:35:44 <lkurusa> i'm confused
08:35:46 <lkurusa> what version?
08:35:59 <klange> I suggest not starting at all.
08:36:35 <ArsenArsen> of the boot protocol
08:37:08 <klange> I don't believe the things you'd learn building a Linux bootloader would be useful. I already tell people not start osdev by writing a bootloader - there's definitely no reason to want to write a Linux bootloader.
08:37:08 <lkurusa> i guess... start with the most basic one?
08:37:12 <lkurusa> and expand from there
08:37:18 <ArsenArsen> probably will
08:37:28 <lkurusa> but really what klange said is spot on
08:37:33 <ArsenArsen> I have a feeling this will be a pain
08:37:34 <lkurusa> something to seriously consider
08:37:54 <ArsenArsen> I'd love to not do it but what I am doing is truly too weird not to
08:38:20 <klange> Linux boots directly on EFI, so any bootloader you write for Linux wouldn't be reflective of how the real world works.
08:39:34 <ArsenArsen> oh actually maybe I could do that instead, chain loading EFI might work better and be more useful altogether
09:00:26 <geist> huh, that's a new one
09:03:42 <lkurusa> http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2310.pdf
09:03:47 <lkurusa> looks like a new C standard?
10:12:13 <geist> well, i guess it's about time
01:51:45 <Vercas> I really appreciate the more clear wording about the volatile qualifier.
06:41:46 * geist yawns
07:46:11 * bauen1 yawns
07:48:01 * nchambers yodels
07:48:32 * _mjg burps
11:15:11 <ArsenArsen> can bochs leave a20 disabled so I can test a20 enabling routines
11:17:44 <bcos_> Yes - I think it's a configure/compile time option
11:36:00 <ArsenArsen> that's a slight bit annoying but okay; I'll check it out bcos_
11:40:49 <geist> frankly i wouldn't bother writing a test case for it. either it enables and code works, or it doesn't and the system explodes
11:41:00 <geist> and disabling it is not useful in pretty much any case
11:41:09 <geist> (unless you're writing DOS)
11:41:39 <_mjg> (don't write DOS)
11:47:58 <ArsenArsen> nah I'd not be doing that, I just want to make it more compatible
11:53:00 <ArsenArsen> there's --enable-a20-pin but I cannot tell whether that is to enable emulation or the line itself