[HN Gopher] A kernel update broke my stylus
___________________________________________________________________
A kernel update broke my stylus
Author : yorwba
Score : 74 points
Date : 2023-11-01 17:48 UTC (5 hours ago)
(HTM) web link (www.davidrevoy.com)
(TXT) w3m dump (www.davidrevoy.com)
| thanatos519 wrote:
| Sounds like someone broke userland!
| Aleklart wrote:
| "No, we fixed the glitch."
| tux3 wrote:
| That doesn't normally fly in kernel land
|
| If a 'fix' broke something (and it's not for something like a
| major security issue), then it is customary to revert the fix
| and go back to the drawing board
|
| I can see that this is a complicated situation and OP was a
| special case that had not been affected by acccident, but
| those are excuses
|
| If you upgrade your kernel and you don't have any out of tree
| stuff, by policy nothing should break, orherwise it's the
| kernel's problem.
| smelendez wrote:
| > go back to the drawing board
|
| I see what you did there
| appleskeptic wrote:
| "Never break userland" is more of a strong guideline than a
| hard rule, even if Linus has sometimes made it sound
| otherwise during a heated gamer moment. The fact is, kernel
| devs break userland code all the time, sometimes because
| the alternative is worse, or sometimes because a bugfix
| broke something and no one noticed until much later and it
| was decided not to revert the fix because it would now
| break other things to do that. Try to run an entire large
| userland from 2001 on a modern kernel and tell me how that
| goes.
| tux3 wrote:
| Mistakes are human. Certainly a lot of bugs slip through,
| and that's understandable, software is hard.
|
| That being said, I find that people tend to take the no
| regression rule pretty seriously, if you bring a valid
| complaint. It makes good sense. We don't want to give
| people a reason to stay on old kernels even more than
| they do today (this post written on an ancient Android
| kernel, chock full of out of tree drivers). We can at
| least save people from pain when it comes to in-tree
| drivers.
|
| I think we really all benefit from that policy. Even if
| it sometimes keeps some less than perfect workarounds in
| the code
| tedunangst wrote:
| Every userland program is equal, but some are more equal than
| others.
| kaetemi wrote:
| Now Windows isn't alone in changing or resetting drawing tablet
| settings on random updates.
| DarkmSparks wrote:
| difference with Linux is this will be fixed most likely before
| new year.
|
| windows will always stay broken.
| pizzapim wrote:
| A smarter person than me is perhaps able to hack an eBPF HID
| program together to temporarily remediate the problem
| oakwhiz wrote:
| It kind of sounds like there needs to be a quirks mode for pens.
| abdullahkhalids wrote:
| I recently started a remote scientist job, and its kind of insane
| that I am the only person among dozens who use a tablet+stylus
| during meetings. People have such poor discussions, because
| unlike a irl meeting where multiple people will write stuff on a
| whiteboard, here people are just communicating verbally (read
| vaguely) or painfully scratching unintelligible
| diagrams/equations with their mouse.
|
| I think every remote job that sends employees a laptop should
| also be sending them a wacom tablet.
| jon-wood wrote:
| What are you using for note taking? I while back I bought a
| small Wacom tablet for this very purpose, particularly being
| able to quickly sketch out diagrams while in in meetings, but
| ended up giving up because I couldn't get the hang of using the
| tablet.
|
| I'm pretty sure it was a tooling issue rather than the tablet
| itself, and wouldn't mind taking another punt at it.
| orbital-decay wrote:
| What's your OS? On Windows, Nebo is pretty good. Rnote is the
| best of FOSS, but it lacks handwriting recognition if you
| need it, at least for now. For hybrid workflows (tablet +
| keyboard) Obsidian also has something to offer, Excalidraw
| plugin in particular.
|
| _> I'm pretty sure it was a tooling issue_
|
| Depends on whether you want shape/diagram recognition or not.
| If the only thing you need is free-form drawing, you can use
| just about anything.
| abdullahkhalids wrote:
| I just tried Rnote. Seems quite decent. But on MacOS, it
| seems to be bugged. Instead of a tiny pointer/eraser/other
| tool showing as the pointer, it continues to show the OS
| mouse pointer.
|
| I will try it on linux later. Hopefully it will work there.
|
| My initial gripe is that there are three different
| toolbars, all separated from each other. I want to change
| tools and their settings quickly, not move my hand across
| the entire screen. For example, select brush tool on bottom
| toolbar, select size on the left toolbar, then select color
| on the top toolbar.
| abdullahkhalids wrote:
| I have the medium Wacom. I think the small size would be
| uncomfortable.
|
| When I was a prof, I used the Wacom on linux with OpenBoard
| [1] to teach online for two years of the pandemic. My work
| gave me a Macbooks, but OpenBoard has some weird bugs on
| MacOS. So I am using GoodNotes, which is very annoying, but
| the best of what I tried. In another life, I would dedicate a
| few years of my life writing a decent note taking app.
|
| [1] https://github.com/OpenBoard-org/OpenBoard
|
| Edit: Besides the software, I think it takes a few weeks of
| writing on it daily where it becomes natural.
| quectophoton wrote:
| > I have the medium Wacom. I think the small size would be
| uncomfortable.
|
| I have a Wacom Intuos S.
|
| Sample size 1, but I haven't been inconvenienced by its
| size, even while handwriting or drawing at Krita.
|
| On one hand, I don't think a thought like "this would be
| easier with a bigger tablet" has ever crossed my mind; but
| on the other hand, it's also true that I haven't tried
| anything bigger, so my only point of comparison is tablet
| vs mouse or tablet vs whiteboard/notebook.
| abdullahkhalids wrote:
| I haven't also tried the S one. But my intuition is the
| following.
|
| The surface of the tablet is mapped to your screen in
| terms of input. The Intuos M is just a bit smaller than
| 13 inch laptop screens. So what you see on the screen is
| roughly the same size as what you draw with your hand.
| Additionally, while writing at normal pen/paper size, my
| hand moves comfortably from the left of the tablet to the
| right as I write across the whole screen.
|
| If I had to use a Intuos S, then I imagine, I would have
| to write a lot smaller with my hand, but then have it
| appear as bigger on the screen. But my hand would not
| move a lot, and making tiny characters would tire my hand
| out faster. I think my brain would get used to it, but it
| will still be a jump.
| dataflow wrote:
| > I think every remote job that sends employees a laptop should
| also be sending them a wacom tablet.
|
| It's so annoying dealing with a separate tablet just for the
| sake of writing. Why not a tablet PC or such?
| dharmab wrote:
| Wacom style devices are extremely affordable. I have a $30
| drawing tablet which I use for digital art and works very
| well.
| dvdkon wrote:
| I have one, and flipping it from "laptop mode" to "tablet
| mode" adds noticeable friction, especially if you have e.g. a
| dock connected. Then again, so does having to carry around a
| tablet.
|
| For an office job I'd take the external tablet, since my
| laptop would likely be plugged in all the time.
|
| You could also keep it permanently in tablet mode and use an
| external monitor, mouse, and keyboard. Just don't
| underestimate the mode change.
| quectophoton wrote:
| People look at me like I'm crazy because I use a wacom tablet
| as mouse.
|
| I'm using a "cheap" one (I think it was like 80 EUR), and it's
| mind blowing how convenient it is despite its small size. Now
| trying to use a physical whiteboard for brainstorming, feels
| the same way as trying to use pen and paper for programming:
| technically doable, but I curse internally because the
| inconvenience.
|
| Just being able to select stuff and move it around, or undo
| with a quick Ctrl+Z, or rotate things, makes all the
| difference.
|
| Sure, I love using a paper notebook for my notes, and the feel
| of writing with a pencil on paper is really nice; and I would
| use it for drawing quick stuff if I don't have anything else at
| hand. But if I think I'm going to be erasing a lot, or that I
| will be doing the kind of stuff that is usually done with a
| whiteboard, a drawing tablet is (to me) a vastly superior
| option.
| abdullahkhalids wrote:
| Yup. Writing on paper with a nice fountain pen or pencil is
| primal love. So satisfying. But digital tablets are so much
| superior for that messy kind of work, you just have to use
| them.
| Terr_ wrote:
| Not really a solution as much as an alternative form of
| attack: Tools like graphviz where I can type certain limited
| types of diagrams into a text editor and xdot well instantly
| display the render.
|
| There are also some online ones, although I'm not sure how
| well they work in a multi-editor situation.
|
| Might not be flexible enough for a really open brainstorming,
| but enough for stuff like figuring out domain concepts.
| c0nfused wrote:
| I have been really tempted to sit down and write a ot version
| of edotor.net so more than one person could edit a graph at the
| same time in a meeting/ discussion.
|
| We have been trialing using it to draw graphviz graphs for
| discussion and it has been working very well both as an in
| meeting tool and as a reference later on.
|
| Would be interested to go the Wacom route but my my handwriting
| is very bad and only getting worse
| pengaru wrote:
| Maybe this can be worked around via the 'quirks=' usbhid module
| parameter:
|
| https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...
| $ modinfo -p usbhid | grep quirks quirks:Add/modify USB HID
| quirks by specifying quirks=vendorID:productID:quirks where
| vendorID, productID, and quirks are all in 0x-prefixed hex (array
| of charp)
| lantastic wrote:
| David reached out to the maintainers on the relevant ML [0].
| Could also CC the regressions@lists.linux.dev [1].
|
| [0] https://lore.kernel.org/linux-
| input/nycvar.YFH.7.76.23110120... [1]
| https://www.kernel.org/doc/html/v6.6/admin-guide/reporting-r...
| molticrystal wrote:
| It didn't hurt that the comment section[0] for the blog is
| powered by the fediverse [1], so Greg KH responded to the post
| [2] pointing him towards writing that email and tips on doing
| so along with many other tech savvy individuals. This issue is
| being seen all over the place.
|
| [0] https://www.davidrevoy.com/article995/how-a-kernel-update-
| br...
|
| [1] https://framapiaf.org/@davidrevoy/111336038253784524
|
| [2] https://social.kernel.org/notice/AbN55QfONFCPweYq7E
| chabad360 wrote:
| While very annoying, it is possible to create an evdev based
| wrapper around the input, that can remap any of the signals as
| you wish (I know it's possible cause I've done it).
| gertlex wrote:
| We had a fun one a couple years ago where a kernel bump turned
| our USB touchscreen monitors into trackpad monitors... (i.e.
| touch input behaved like your laptop trackpad) Thankfully I was
| able to figure out a udev rule via trial and error and get that
| fixed easily enough. ChatGPT probably would have saved me some
| time there had I had it at the time. Deploying that to hardware
| in the field was the messy bit (but still automated, so that was
| a win).
| 1970-01-01 wrote:
| Great example of how Linux evenings randomly come about. Someone,
| somewhere, did a thing, and now user needs to find out who did
| what, and why it happend, and how to fix it by themselves.
| Galanwe wrote:
| If it's just a driver problem, why not checkout the old one,
| build it as a new module with a different ID, and insmod it on
| the new kernel?
| dharmab wrote:
| Because an artist using, say, Krita on Linux to paint shouldn't
| be expected to rebuild kernel modules.
| doublerabbit wrote:
| > Krita on Linux to paint shouldn't be expected to rebuild
| kernel modules.
|
| The other-side of the coin is that if your using Linux, you
| should expect that such things happen without your desire.
| And that if you wish for the previous feature you'll have to
| nose-dive in to if you want to fix or just revert back to the
| previous version. You shouldn't, but here we are.
|
| At the risk of downvotes, I would also say you shouldn't
| really need to update the kernel either. Unless there is
| really a feature required, or extreme-security vulnerability
| of your current version, exclude it.
|
| That was the past-beauty of Linux, you didn't need to be on
| $latest; that was Microsoft's realm, always requiring
| updates. If it worked you left it be for this exact reason.
| Every day I look at my iPhone and everyday there's a batch of
| updates for my apps sometimes twice on the same day for the
| same app and I don't use many.
|
| The current fearing update culture we live in is terrible.
|
| /vent
| jdhendrickson wrote:
| I love his comic, I hope this is addressed soon.
| qwery wrote:
| Software all the way up the input stack seems to be a bit
| confused about a few things, or there is a lack of understanding
| that:
|
| - graphics tablet digitisers generally are not and don't include
| touch digitisers
|
| - tablets are not touch screens
|
| - a stylus (tip) isn't a button
|
| - a stylus can have buttons
|
| By convention, stylus tip interactions are "tools", which seems
| to be what Linux' BTN_TOOL_* codes are meant to mean. A button on
| a stylus should never be (directly) translated to a tool -- like
| how the ~5th button on a mouse is `BTN_EXTRA`, not `BTN_MOUSE`.
| In this particular case it looks like the thinking is: the device
| has 2 event codes (which we know as the stylus tip and the barrel
| button) and it's a stylus^wdigitiser, so map device[0] to
| `BTN_DIGI` and device[1] to `BTN_DIGI + 1`.
|
| editendum: the thing about the BTN_EXTRA doesn't really map...
| The "digitiser" event codes _are_ being treated in the same way
| as mouse event codes (device[i] = > event[CLASS + i]). The issue
| is really that the barrel buttons have no (dedicated) event
| codes.
| javier_e06 wrote:
| Oh don't get me started. I got a Corsair Claw Gaming mouse and
| Rocky Linux 9 does not have a simple way to map the buttons to
| events. Corsair does not provide Linux software for it. Yes I
| tried all kind of mapping programs to no avail. My chromebook old
| chromebook tablet (32 bit) has a stylus and works okay. I won't
| upgrade, I know how those upgrades break non-run-of-the-mill
| external devices APIs.
| paxys wrote:
| Ah, Linux.
| pipeline_peak wrote:
| Definitely wouldn't recommend using Linux for artwork...
___________________________________________________________________
(page generated 2023-11-01 23:00 UTC)