Off-topic but curious if you experience this issue my Librem5 exhibits - I use my L5 mostly just to play mp3s. Regardless of being powered over usb-c or running on battery, it frequently has audio blips reminiscent of some old laptops briefly going in and out of deeper sleep states long enough for an audio blip.
Have you noticed something similar? do you listen to music on your L5?
Edit: I should also mention it does this regardless of audio out the internal DAC+3.5mm TRS, or over a USB MBox attached to a dock. It seems to be something quite low-level disabling interrupts for too long, something like that.
So you may want to consider GNU/Linux phones (Librem 5 and Pinephone). By the way, they have a real desktop mode with desktop apps, not mobile apps on a large screen.
The plaintext TSS/ECID and the plaintext OCSP issues have been fixed, which IMO were the only meaningful security gripes of the article.
The iMessage/ADP/Metadata stuff I think is more of an implementation decision than a meaningful attempt at data collection. Using clear text file names and hashes for dealing with collisions and deduplicating is a reasonable first pass at something like this. Sure, they could probably roll some end-to-end obfuscation for this, but with how big their stack and cloud integrations are, I’m sure that’s non-trivial.
While this phone still does not have an Armv9-A CPU, but only a quadruple Cortex-A78, that is nonetheless like 4 to 10 times faster than the phones mentioned by you and faster than a light notebook of a decade ago.
The other Linux phones are barely competitive with a laptop of 25 years ago and they are not really usable as a personal computer today, unless you like waiting for your computer.
While I partially agree with what it says at your link, about specs depending on the context, I am also well aware about the actual performance of quadruple Cortex-A53 CPUs and how it compares with alternatives, because about a decade ago I have used many single-board computers with various kinds of such CPUs (including Raspberry Pi, but also other very different SoCs). I eventually abandoned them for better alternatives.
If you have modest requirements, you can be content with such a CPU, e.g. for reading and editing simple documents or browsing the Internet with scripting disabled.
The point is that there is no need to restrain yourself to cope with its limitations, because for more than a decade there have been much better alternatives.
Even when your target is a sub-$100 computer, it makes no sense to use any CPU weaker than a quadruple Cortex-A76, like in the many computer models using Rockchip CPUs or in the current Raspberry Pi. While there are much more sources of Cortex-A76 based computers, the number of those offering much faster Cortex-A78 based CPUs using Qualcomm or Mediatek SoCs, including this smartphone, are steadily increasing.
> The point is that there is no need to restrain yourself to cope with its limitations, because for more than a decade there have been much better alternatives.
Where are those alternatives? Which phones with a better CPU can run GNU/Linux?
Good call. It seems this is just one way to run it, https://wiki.postmarketos.org/wiki/Hybris. Another way is to run mainline kernel, which is currently supported by few devices, including Librem 5, Pinephone, Fairphone 2, OnePlus. It seems only the first two run with all FLOSS drivers and full support: https://wiki.postmarketos.org/wiki/Devices
So, for OnePlus, you may stop getting updates for proprietary drivers whenever the vendor decides so. Also, not all features are supported due to the difficulty of reverse-engineering the drivers.
Never heard anything about "driver stabilization" before, and couldn't find anything confirming that it exists. In my understanding, the longer some piece of software is used, the higher the probability that a vulnerability can be found.
I can imagine that after a really long time, most vulnerabilities might be found, but I don't believe I will live long enough to see that for relevant drivers.
> Never heard anything about "driver stabilization" before, and couldn't find anything confirming that it exists.
I don't know that it's a known phrase, I find it odd you searched for it as I would have thought the meaning would have been obvious. Bugs are found generally within the developer mandated lifespan of the device, more bugs are found closer to release with less being found over time.
There might still be vulnerabilities, but that would be just as true for open source ones, because in both cases no one is really looking for them.
Can you find any unpatched vulnerabilities for any of the OnePlus devices that PMOS runs on? I doubt it. They may exist, but they may exist i your Librem also.
Stastistically, vulnerabilities should exist in both OnePlus and Librem. However, as the latter has free drivers, anyone will be able to fix them, whereas for OnePlus you have to hope that the vendor helps you.
Right, so you're just going off vague generalizations and possibilities. There's an entire army of people looking for vulnerabilities in this area, and they haven't found any. Until you can show a vulnerability for the devices I mentioned, you don't have much of a point.
Why are you so sure? Not everybody shares vulnerabilities with the world. Some people keep them for themselves or sell on the black market. I doubt OnePlus offer a reasonable payment for that.
Because it's not a high target device. Valuable targets are not using outdated OnePlus devices. There might not be money to be made, but there is reputation to be earned by discovering a new vulnerability. I doubt a vulnerability in some old 6T hardware would be worth much, while disclosing it and padding a resume would be.
People who use non-mainstream systems and rely very little on proprietary software are automatically more interesting. They may have something to hide.
It is hard to associate in the same sentence the word "powerful" with a quadruple Cortex-A53 CPU.
Cortex-A78 is in a totally different class of "powerful", when compared with Cortex-A53. Cortex-A53 has not been designed as a top performer among Arm CPUs, but only as a low-power core. Even at its launch, in 2012, 14 years ago, Cortex-A53 was much slower than the older big Arm cores, like Cortex-A15 from 2010. I am aware of this from direct experience, because a decade ago I have developed software on several SBCs with Cortex-A53, including Raspberry Pi, and also on an ODROID model with Cortex-A15, which ran circles around them, and unlike them it delivered a passable desktop experience.
On the other hand, at its launch in 2020, Cortex-A78 was faster than any older Arm CPU cores. It was then surpassed by the Cortex-X1 launched simultaneously with it, and then by the Apple M1, launched later that year.
Even if some enthusiasts have used old Raspberry Pi and similar SBCs with Cortex-A53 as PCs, that was really not wise as they were too slow for a comfortable use and there were faster alternatives with a similar price (after you added the cost of all required peripherals).
It's more expensive and not sufficiently more powerful: I would also have to pay customs duties to import it into Serbia. The above are actual devices I own or have owned, it's natural I did not get all of them (including like PinePhone Pro).
They do not dominate the development Like MS for Windows. Independent people from all over the world review their contribution. This is a small problem related to other things.
reply