Silly question: if Linux kernel interfaces e.g. syscalls are stable, why is there a need to create an LTS release? How can non-LTS updates affect differently than updates within the LTS branch?
539 karma · joined July 9, 2014
Writing emulators and hypervisors.
Alexandro Sanchez <alexandro@phi.nz> https://phi.nz/
Do you trust videocall participants because you recognize their faces and voices? ...Or because a server certified by a root CA has authenticated the other participants?
The age of deepfakes has started, nobody can stop it. Improving our mental security models will become as essential as literacy.
Throwing a dead battery into a dump -> Battery acid gets dragged by the rain -> Polluting a river -> Flows polluting a sea -> Currents polluting all oceans.
Of course effects become smaller as scope grows, but still in many instances it's measurable, see Fukushima.
- PS3 (Nvidia RSX): https://github.com/RPCS3/rpcs3/blob/master/rpcs3/Emu/RSX/rsx...
- PS4 (AMD GCN): https://github.com/AlexAltea/orbital-qemu/blob/master/hw/ps4...
The hardware is known ahead of time, so console platforms take advantage of that: GPU drivers allocate a context and hand it over to userspace which can do nearly anything it wants with it (abstraction layers are quite thin), and graphics APIs are essentially a 1:1 mapping of the hardware.
Because of that there's little difference between emulating the API or emulating the GPU (the subset of the GPU accessible through the API).
The only major difference I'd say is emulating the entire GPU (including VM, GART/IOMMU, display controller, multi-contexts, etc.) vs just emulating a single GPU context (which feels quite similar to implementing a graphics API).
Which leaves me no choice but to trust all my future projects to services run by people as you put it, "without a moral compass".
If I'm into (say) 3D printing and create software for it, I simply will not risk getting de-platformed just because some dude is upset people are 3D printing guns and guns are bad. The day 3D printing is illegal/restricted, I'll stop or accommodate to regulations.
I chose "3D printing" as something innocuous, but feel free to replace it with anything divisive and controversial... People just do not realize who *exhausting* it is dealing with founders with strong moral compasses. (And yes, I get it's their/your prerogative and if I don't like it I can GTFO. Just venting off.)
I want the predictability that as long as I am not breaking the law, my data can be safely hosted. Am I supposed to now pay attention to law and the countless individual moral compasses of people/vendors I interact with?
Moral compasses are tricky, some people let them run amok, and end up infecting computers with revenge-ransomware just for having their system language set to Russian, as it happened in March 2022. Was that also justified for the "greater good" of humanity?
Of course, it's your project, your service... anybody can fork it or self-host it, but this only screams at me: "be wary of this guy and his services, you could be next".
PS: I do not maintain cryptocurrency projects or trade any.
People that engineer the future of countries they will never live at, happy to displace cultures for a fucking basis point, make me deeply sick.
Who in the bloody hell are you, or anyone at Davos for that matter, to dictate what my nation, my town, or my family needs? Have you even been asked? Or is this just the usual arrogance ingrained in all globalists of just "knowing better"?
[Citation needed]
Over the past years I've been able to book flights, shop hardware, get hotels and rental cars, renew my domain, pay for my server with cryptocurrencies.
Also, you can express disagreement without throwing Reddit-style sarcastic snarks.
Depends on the device, I've had low-quality USB-C devices whose transfer speed inexplicably decreased when connected "the wrong way". Not knowing which orientation is the right one until you initiate the transfer is frustrating, and worse IMHO than physically preventing the wrong orientation.
Of course, this issue affects only a subset of mediocre hardware... But the fact that [1] this is possible is quite a downside to USB-C.
[1] https://hackaday.com/2021/03/22/cursed-usb-c-when-plug-orien...
What do you mean by "growing up"? Hopefully not embracing the pervasive mediocrity of institutions and defending the current status quo.
You are replying to marcan from fail0verflow; he likely does. Many people from the early PS3 hacking days did.
Crippling use-cases is quite difficult: how could you distinguish at hardware/firmware-level object detection for fighter jets vs object detection for cars. Under the hood everything is just a bunch of compute units with extremely wide ALUs. I would even say, it's next to impossible to cripple "AI" without crippling graphics engines and most GPGPU kernels.
EDIT: Ah, you meant drivers. Yeah, that's perhaps more doable (since the OS can provide context on the calling application), also more detectable by the end-users: many people diff drivers to find patched vulnerabilities, security researchers would eventually notice it.
Mostly in Europe and North America, and territories culturally-influenced by them, largely as a side-effect of historical events 80 years ago.
Other parts of the world have different opinions on the matter.
I wonder what we could make out of many high-end cameras if only we had enough patience and motivated people to reverse engineer everything that has been hidden away by comfort abstractions.
Anecdotal, but plenty of contacts who would have never accepted an on-site position in UAE, Hong Kong, Germany (for reasons ranging from personal, political, or just sheer affordability), and are now contracting as individuals at companies based in such regions, for substantially above the Bay Area average for their experience.
Companies will pick from a larger pool of people. People will pick from a larger pool of companies.
Why can't they switch to P2P audio/video transmission instead of funneling everything through their servers?
What an odd way of representing data transfer bandwidth: Was it so hard to just say 1.2 Tbps (or better 150 GB/s)?
Also, maybe an ignorant question from my side: Is that bandwidth such a big deal at nation-scales? It feels like barely 1000 homes maxing out the fiber usage. All submarine cables being deployed today are already in the 100+ Tbps range, aren't they?
Such nonsense just ruins the Internet, maybe we should find better approaches than "yet another banner": accepting cookies, accepting privacy policy, verifying age, registering, logging-in, etc. all to watch some dumb 10 second video... Yet, porn and violence are to this day still a Google search away. So, what is YouTube exactly trying to stop here?
Under that logic, shouldn't we replace <img> and <video> elements with WASM modules for ffmpeg/libjpeg/libpng rendering into a <canvas> too?
Or hey, maybe we could drop HTML altogether and just render all content from WebAssembly straight into the <body>?
Granted, it was an automated email, but an automated "No" would have been more appropriate. Is it so hard to offer different standard responses? Why does the company consistently fail at human interaction, be it candidates or customers?
From my circles, most share the feeling that FAANG/MSFT are insulting for candidates (I can only speak for Google/Microsoft).
That's absolute nonsense. There's so many platforms, Android/iOS, where applications are compartmentalized: local privilege escalation (LPE) vulnerabilities is what enables a rogue application (or remote code execution into one, such as a browser/chat), into full control of the device and its data. Sometimes LPEs are needed to achieve some level of persistence. In VMs, a guest LPE is frequently needed to interface (and exploit) directly the emulated/paravirtualized devices and VM-escape into the host OS.
Yes, it is a big deal.
> The new CPU family marks one of the largest architectural jumps we’ve had in years, as the company is now baselining all three new CPU IPs on Armv9.0.
I believe the deprecation of AArch32 is quite an important change and by itself already warrants issuing a new major version (then there's all mandatory extensions from v8.2+ and SVE2).
They have not claimed to bump the major version on a regular basis from now as you seem to suggest.
I don't think https://libvirt.org/ is a solution to the fragmentation problem. Whenever a new virtualization feature appears (and this happens quite often!), having to update N different hypervisors, and expecting their interfaces to consistent enough and consumed by libvirt is wishful thinking: In practice it just does not happen. This hurts the development of cross-platform software that require cutting-edge hardware virtualization features.