|
Tankakern posted:hm this is kind of a big deal Linux catching up with windows 7 I see
|
#
?
Jun 26, 2022 01:20
|
|
|
|
| # ? Apr 19, 2024 00:05 |
|
|
Lets hope we never catch up to 8, 10 or 11
|
|
#
?
Jun 26, 2022 07:42
|
|
|
Mr. Crow posted:Lets hope we never catch up to 8, 10 or 11 really baiting the monkey's paw here
|
|
#
?
Jun 26, 2022 08:29
|
|
|
akadajet posted:Linux catching up with windows 7 I see old apps are still regularly blurry on windows 11 wtf are you talking about lmao
|
|
#
?
Jun 26, 2022 11:49
|
|
|
Best Bi Geek Squid posted:https://jayfax.neocities.org/mediocrity/gnome-has-no-thumbnails-in-the-file-picker.html i’m the serious linux user who has had two blocked toilets in his house for a year
|
|
#
?
Jun 26, 2022 18:01
|
|
|
Gentle Autist posted:i’m the serious linux user who has had two blocked toilets in his house for a year …and posts like a moron
|
|
#
?
Jun 26, 2022 19:26
|
|
|
you know the British aren’t known for good plumbing or computing
|
|
#
?
Jun 26, 2022 19:31
|
|
|
I haven't checked but all these Linux auto-scaling things for 2xDPI always seem to use linear filtering instead of nearest neighbor Do these people save all their screenshots as JPEGs too? It talks a lot about things being "blurry". Use nearest-neighbor for integer scaling scenarios, then you have 2x2 pixels that look identical to how they'd look on a standard-DPI screen why is this so loving difficult to understand
|
|
#
?
Jun 26, 2022 19:42
|
|
|
Sapozhnik posted:I haven't checked but all these Linux auto-scaling things for 2xDPI always seem to use linear filtering instead of nearest neighbor while preferable it absolutely does not look identical unless you have a real bad hidpi screen. a good hidpi screen ironically makes the large pixels extremely clear and obvious.
|
|
#
?
Jun 26, 2022 19:53
|
|
|
Can someone tell me why my linux desktop, exact same hardware, has a zippier mouse cursor when booted to windows? Ubuntu, what the ever loving gently caress. It's noticeably less input lag on the god damned mouse cursor. Is there a gnome extension I can install or something to fix this?
|
|
#
?
Jul 3, 2022 03:27
|
|
|
is it actually “input lag” you’re hitting or is it the different acceleration curve
|
|
#
?
Jul 3, 2022 03:32
|
|
|
Yeah of all the problems linux has, i've never seen that one. mouse accel (like echi said) or speed? or maybe wrong resolution + scaling lag?
|
|
#
?
Jul 3, 2022 04:12
|
|
|
there's also a weird thing that can happen to libinput where it enables something called "hysteresis" mode that can not be disabled again (without rebooting), and it makes the cursor feel slow/laggy because every movement is calculated as a function of its previous position or something like that it happens on one of my laptops sometimes and i just don't use the mouse enough to bother to fix it, but somewhere on the freedesktop gitlab there's a patch where somebody just removed the whole "feature" (i'm sure it's a legit feature for some use-case but i've no idea what that would be)
|
|
#
?
Jul 3, 2022 12:05
|
|
|
it's for touchpads that sends in small position changes all the time (e.g. noise) regardless if there's any movement or not, it's so you'll have to go over a certain threshold before it's counted as actual movement
|
|
#
?
Jul 3, 2022 15:16
|
|
|
oh that's what that was. I thought gnome shell had fragmented its heap to poo poo after the running gnome-shell process had reached the positively Methuselah-like age of four days. Older mutter was absolutely awful for lagging out like that although I think it's being rearchitected lately. Yeah I had something where the mouse was stuttering along at 10 fps or something like that. good to know ty.
|
|
#
?
Jul 3, 2022 15:31
|
|
|
Systemd Creator Lands At Microsoft i ... have no words .. oh well maybe this'll mean we'll get proper bitlocker integration and azure ad support in systemd :p would make things easier for business users at least
|
|
#
?
Jul 7, 2022 08:06
|
|
|
Tankakern posted:oh well maybe this'll mean we'll get proper bitlocker integration and azure ad support in systemd :p would make things easier for business users at least On Fedora I can open and use my old Bitlocker NTFS drives without problems, is there any issue that you are aware of?
|
|
#
?
Jul 7, 2022 08:53
|
|
|
we can't use bitlocker natively yet, making dual boot hackish
|
|
#
?
Jul 7, 2022 09:02
|
|
|
Tankakern posted:Systemd Creator Lands At Microsoft [img-billgatesborg] embrace extEND EXTINGUISH
|
|
#
?
Jul 7, 2022 09:13
|
|
|
Tankakern posted:Systemd Creator Lands At Microsoft "Embrace, extend, and extinguish"... watch your back guy 
|
|
#
?
Jul 7, 2022 09:17
|
|
|
Tankakern posted:we can't use bitlocker natively yet, making dual boot hackish What does "natively" mean in this case? Can't enable Bitlocker on a Linux boot partition (only mount Bitlocker drives after boot), something like that? e: I don't dual boot anymore but I would just use Bitlocker for the Windows partition and LUKS for the Linux one. Don't set them to auto-mount and you get an extra layer of security against accidentally messing up the other installation. Shared data can go in an auto-mount Bitlocker drive since both can open it (though if you really want LUKS for some reason, I guess you could mount it in WSL in order to see it via Windows Explorer, why though). NihilCredo fucked around with this message at 09:32 on Jul 7, 2022 |
|
#
?
Jul 7, 2022 09:26
|
|
|
tpm unlock of bitlocker booting linux, dual boot with windows with bitlocker without reboot uefi hacks
|
|
#
?
Jul 7, 2022 09:32
|
|
|
Tankakern posted:tpm unlock of bitlocker booting linux, dual boot with windows with bitlocker without reboot uefi hacks that's probably never going to work, simply due to the way tpm sealing works. massively oversimplified: a measured boot is part of the process of unsealing the tpm, and by design a tpm sealed with one set of measurements cannot be unsealed by a boot with a different set of measurements.
|
|
#
?
Jul 7, 2022 09:38
|
|
|
when microsoft hired guido van rossum he was handed a couple of engineers and told to work on whatever (which is how we're finally getting some minimal effort on performance optimizing python), so this might not be a bad thing.
|
|
#
?
Jul 7, 2022 09:49
|
|
|
Tankakern posted:Systemd Creator Lands At Microsoft Proper bitlocker integration won't happen before they stop supporting sending sensitive information over I²C et al.
|
|
#
?
Jul 7, 2022 11:10
|
|
|
nudgenudgetilt posted:that's probably never going to work, simply due to the way tpm sealing works. linux should add support for adding its own measurements, replacing windows' bootloader but still working with bitlocker, and that way have proper support for dual boot with tpm
|
|
#
?
Jul 7, 2022 13:13
|
|
|
BlankSystemDaemon posted:It makes complete sense, considering System500 and its friends are very similar to the registry hive, service handling and logging in Windows. idiot
|
|
#
?
Jul 7, 2022 13:13
|
|
|
i know you can use ntldr or whatever its called now to load a vhdx file, maybe it doesn't have to be windows? e: obviously i have not looked this up
|
|
#
?
Jul 7, 2022 13:33
|
|
|
Tankakern posted:linux should add support for adding its own measurements, replacing windows' bootloader but still working with bitlocker, and that way have proper support for dual boot with tpm linux *has* support for measured boot. the issue is that the way measured boot and tpm is designed, without a vulnerability, linux will never be able to directly unseal a tpm sealed by windows. this isn't even just a windows/linux thing. every change to your bootloader or kernel means a different value is written to the write-one pcr registers on the tpm at boot time. if those pcr registers have the wrong value, the tpm will refuse to unseal. bitlocker takes full advantage of the tpm, and the tpm, by design, ensures specific known-good (as in known to the tpm itself as informed by when the tpm was sealed) software stack is required. you literally cannot make linux look like windows to the tpm without executing all the windows boot process code prior to starting the linux boot process.
|
|
#
?
Jul 7, 2022 16:42
|
|
|
Tankakern posted:idiot pot/kettle, yo
|
|
#
?
Jul 7, 2022 16:44
|
|
|
*measures a boot*
|
|
#
?
Jul 7, 2022 16:45
|
|
|
The big stumbling block for measured boot on Linux is the initrd. This has to be built locally the way things are set up right now, so it can't be signed by the OS vendor. When a secret is sealed you can choose what TPM registers need to match to unseal it. PCR7 I think tracks the secure boot keystore state and I think that's enough to securely seal a LUKS key with. Firmware checks the signature on Shim, Shim checks a user-generated key's signature on GRUB and I guess GRUB can validate an OS vendor signature on the kernel and a universal initrd, and then you have a validated boot path that you know hasn't been backdoored.
|
|
#
?
Jul 7, 2022 16:51
|
|
|
yep my point was that you could get it to work properly if linux took it all the way to unlocking bitlocker for windows to boot from, doing the measuring and tpm unlocking and stuff. if you in linux could set up the bitlocker part with setting up the tpm and bitlocker metadata, that would be a solution that also could support proper tpm boot on linux at the same time but you're right, the initrd stuff is the most important thing to fix first
|
|
#
?
Jul 7, 2022 19:54
|
|
|
you'd have to create some windows extension to make it work, because if windows is left on its own devices to measure the boot it'll of course screw things up
|
|
#
?
Jul 7, 2022 19:57
|
|
|
It shouldn't matter anyway because I assume you can get the bitlocker recovery key from windows and then use that to set up an automount on linux, TPM be damned TPM LUKS would be nice under Linux although I don't know how much I'd trust it. You know where you stand with a good old fashioned passphrase. Better to do what macOS does and transparently add admin user accounts' passwords to the FDE keyring.
|
|
#
?
Jul 7, 2022 20:00
|
|
|
i want dual boot and proper tpm support drat it
|
|
#
?
Jul 7, 2022 20:01
|
|
|
oh well i hope poettering does well at ms, if anyone could make it happen it would be him (or mjg59 i guess)
|
|
#
?
Jul 7, 2022 20:03
|
|
|
poettering has some weird ideas about ... basically everything. his designs are always just a bit screwy and other people have to keep some of his weirder design decisions in check. i guess it's like george lucas syndrome or something. really wish we could use systemd-boot instead of grub though. or idk a kexec based bootstrap kernel+initrd file instead of grub.
|
|
#
?
Jul 7, 2022 20:13
|
|
|
you have to use systemd-boot if you dual boot with bitlocker, it's the only one that has built-in support for booting bitlocker-windows through bootnext-variable setting
|
|
#
?
Jul 7, 2022 20:32
|
|
|
|
| # ? Apr 19, 2024 00:05 |
|
|
phoronix comments is like golden-era slashdot for subtlety, lot of gnashing of teeth about poettering going to the great devil. i genuinely suspect that microsoft of 2022 is as good, if not better, a place to do good linux software than ibm is. might also change already 2023 for all i know, but we'll see what happens.
|
|
#
?
Jul 7, 2022 20:44
|
|