<nekobot>
• Begasus (b22ffde8): minuet, bump to 25.04.0 (#12307)
bjorkintosh has quit [Ping timeout: 480 seconds]
bjorkintosh has joined #haiku
bjorkint0sh has joined #haiku
bjorkintosh has quit [Ping timeout: 480 seconds]
bjorkint0sh has quit [Remote host closed the connection]
bjorkint0sh has joined #haiku
HaikuUser has joined #haiku
HaikuUser has quit []
Guest14842 has left #haiku [#haiku]
Guest14842 has joined #haiku
bjorkint0sh has quit [Ping timeout: 480 seconds]
bjorkint0sh has joined #haiku
bjorkint0sh has quit [Ping timeout: 480 seconds]
bjorkint0sh has joined #haiku
Guest14842 has left #haiku [#haiku]
Guest14842 has joined #haiku
dalme has joined #haiku
Aedil has quit [Quit: leaving]
Aedil has joined #haiku
Guest14842 has left #haiku [#haiku]
Guest14842 has joined #haiku
<Habbie>
waddlesplash, re 'mov ax,3; int 10' reboots my machine but is required before the BIOS on another machine reports all modes; i was wondering if we could do "get modes, see if we are happy, if not, -then- do ax=3 and get modes again'
<Habbie>
waddlesplash, this argument works fine if the problem was 'that other machine reports -no- modes' but the comment says 'not all'
duncsauce_ has quit [Remote host closed the connection]
duncsauce has joined #haiku
<Habbie>
i should make a ticket to track this
<PulkoMandy>
Habbie: That looks like the command to set a 80x25 mode with 8x8 fonts and should be supported since the original ibm pc
<Ellenor>
why would it cause a reboot?
<PulkoMandy>
So if your machine reboots, I'd say fix your bios, or don't use bios as a bootloader?
<Habbie>
PulkoMandy, it is that call, yes
<Habbie>
Ellenor, i have no clue
<Habbie>
but it does
<Habbie>
PulkoMandy, all valid options, but this BIOS is quite popular on chromebooks
<Habbie>
PulkoMandy, so -if- we can work with it from our end, that would be cool
<PulkoMandy>
Yes. An option is to check for some bios signature, or use a special build of the bootloader
<Habbie>
oh bios signature is a nice one, i like that
<Habbie>
a very narrowly targeted quirk
<Habbie>
special build is what i'm doing now, yes :)
<Habbie>
i've also been wondering if i can pass arguments from GRUB
B2IA has quit [Quit: Vision[]: i've been blurred!]
B2IA has joined #haiku
B2IA has quit [Quit: Vision[]: i've been blurred!]
<Habbie>
i can't figure out -why- haiku's boot loader doesn't like my usb-nvme
<Habbie>
it just doesn't list it in Select boot volume/state
<gordonjcp>
Habbie: what do other OSes see it as?
<Habbie>
'cmp' tells me my two USB sticks are identical for at least the first few megabytes
<gordonjcp>
Habbie: presumably a generic USB block device?
<Habbie>
yes, but this one is usb3 while the working one is usb2
<Habbie>
which should not matter at this stage, Haiku relies on the BIOS to this point
<Habbie>
boot loader debug log also says 'drives: 2'
<Habbie>
(internal eMMC + one or the other USB)
<x512[m]>
Habbie: USB 3 devices may not work if not connected to root hub (some internal or external USB hubs are present).
<Habbie>
only two root hubs in this laptop - one usb2, one usb3
<Habbie>
but the haiku usb driver has not been loaded yet at this point
dalme_ has joined #haiku
<x512[m]>
So BIOS bug?
dalme has quit [Ping timeout: 480 seconds]
<Habbie>
it's possible. wouldn't be the first or second BIOS bug i ran into these two weeks :)
<Habbie>
what i -can- do is load stage1 and the boot loader from the usb3 stick, and then mount root from the usb2 stick ;)
<Habbie>
PulkoMandy, opt-alt-t works! this saves me so much trouble!
<Habbie>
ok, and haiku booted from the usb2 stick can mount and see the usb3 stick just fine
<Habbie>
so clearly some fuckery is afoot
<Habbie>
and the code that decides what is bootable and what is not has zero logging - this i could fix :)
<Habbie>
huh. there's -another- stage1 loader in src/bin/writembr
dalme_ has quit [Read error: Connection reset by peer]
dalme has joined #haiku
<PulkoMandy>
That's not a stage1 loader, that's a master boot record. It will load stage1 frcm the active partition (or whatever else is in the first sector of the active partition)
<PulkoMandy>
A stage0 if you want
<Habbie>
oh right
<Habbie>
something that does -not- need the partition offset written into it at a magic offset (506 iirc)
<Habbie>
or perhaps, what gets targeted still does
<Habbie>
but indeed, it tries to understand MBR and GPT - stage1 does nothing like that
<PulkoMandy>
yes, stage1 always needs that. At some point there was a convention that the mbr would pass the partition offset in a register, and stage1 could use that, but the latest version of microsoft mbr code doesn't do that anymore
<Habbie>
i have been wondering about that
<Habbie>
'that guy who called us knew this, he could have just said so'
<PulkoMandy>
So we try to be compatible with whatever you may use as the mbr code :(
<Habbie>
-if- we still relied on this, it would explain that GRUB can boot haiku from my USB but the seabios just before it can't
<Habbie>
but that's not the reason
<Habbie>
however, debugging that is low on my list, as grub works
<PulkoMandy>
All the things we have to do for bios booting to work... and then people on the forums complain that efi is complicated and doesn't solve a problem :>
<Habbie>
hehe
<Habbie>
i read that haiku via efi is much easier on chromebooks than what i'm trying
<PulkoMandy>
I have no experience of chromebooks, but what I saw of seabios on real hardware (in coreboot or libreboot systems), it didn't seem to go well with haiku
<PulkoMandy>
Not sure if it works better in qemu, I assume so?
<Habbie>
whatever qemu does works better than this, yes
<PulkoMandy>
Yes, put I don't remember what they use, is it not seabios?
<Habbie>
i vaguely suspect it is, but i did not check
<Habbie>
trying UEFI means screwing or soldering, and i'm in a two week stretch of the unreasonable position "linux works, so haiku should also work"
<Habbie>
which with some luck will improve some things for everybody
<x512[m]>
I experienced that Haiku works, but Linux doesn't on several machines.
<Habbie>
oh fun
<Habbie>
i also realised tonight that the haiku and linux i2c_elan touchpad drivers are both broken, just in different ways
<x512[m]>
But it was ~10 years ago.
<Habbie>
in response to the patch you posted just above, i have a different patch that works -for me- so that's not a focus right now
<x512[m]>
I think that Coreboot project makes more harm than good. It spreads bad practice of hardcoding OS with hardware and not adhering specifications.
neoncortex has quit []
Babaj has joined #haiku
<Habbie>
+#define TRACE_DEVICES breaks compile of src/system/boot/platform/bios_ia32/devices.cpp