Benjamin Tissoires response is pretty awesome in style, tone, context, usefulness for both the layman and the expert. Kudos to him, and I really need to learn more about eBFF one of these days.
I love how he basically just explained what eBPF was, where I wasn't really sure beforehand. What a brilliant teacher; really inspires me to do some more kernel hacking, too.
eBPD seemed to be the only thing anyone was talking about at KubeCon and Cloud Native Con last week.
I hope people don’t get carried away and do too much kernel tomfoolery with their production kubernetes nodes. Not that it’s my problem, but they’re playing with fire.
"Agreed. I just bought one too and I'm sitting here practicing and I'm smelling a cat piss odor. Thought a cat had pissed on my clothes or something and then I smelled the pen."
"Yesterday, I was sure my cat pissed somewhere in my daughter's room. I literally spent hours trying to find the source of the stench, checked the whole feckin room and then I realised my hand stinks. The very last thing I checked was the bloody Wacom tablet pen that she got from us in the morning."
"Thank you! I thought I was going crazy! I was making everyone in my house smell the pen and they were like “yoooo that’s stinky.” I ended up returning it because the smell was making me gag. So weird."
If Linux was a commercial project we would have had to wait for this to go through a team of product managers, managers, organise a few meetings to prepare meetings to discuss the feature and then maybe few months a year down line it would be approved for work to start.
> But plug and play often means for a hardware maker: "let's do trial and error until Windows seems to behave in a correct way".
It would seem that the Microsoft way, at least in this case, would be to let the hardware manufacturer fall over themselves changing their device's behavior in order to fit however Microsoft implemented the generic drivers. No matter whether it is a correct implementation of the relevant standards.
This is quite efficient, in that Microsoft never even needs to claim there was a bug in the first place, let alone fix it ;)
I dream of Linux someday having a stable driver API. Would be a lot of work to accomplish (and an ongoing headache for the kernel maintainers), but ultimately worth it to no longer have these embarrassing and frequent breakages.
