3 ms·
> The most interesting part is of course the Modem. In addition, they are said to have replaced the WiFi/Bluetooth chip with one developed in-house. > Apple I
by GeekyBear 2y ago
> The most interesting part is of course the Modem.
In addition, they are said to have replaced the WiFi/Bluetooth chip with one developed in-house.
> Apple Inc.’s ambitious plan to create in-house components for its devices will include switching to a homegrown chip for Bluetooth and Wi-Fi connections starting next year, a move that will replace some parts currently provided by Broadcom Inc.
https://www.bloomberg.com/news/articles/2024-12-12/apple-nears-switch-to-in-house-bluetooth-and-wi-fi-chip-for-iphone-home https://www.bloomberg.com/news/articles/2024-12-12/apple-nea...
The rumors called out improved battery life as an upside.
- walterbell 2y agoHopefully these new cellular, wifi and bluetooth basebands can avoid the zero-day fame of their predecessors.
- GeekyBear 2y agoI would assume that Apple's new firmware would have been written in a memory safe language. They did make some noise about enabling Swift for embedded development at last years WWDC. Although Qualcomm's big zero day last year was related to the DSP and not the baseband, I believe.
- FirmwareBurner 2y ago>I would assume that Apple's new firmware would have been written in a memory safe language. Memory safe languages don't protect from human programmer complacency and stupidity, or from incidental alphabet agency backdoors.
- GeekyBear 2y agoThe patch notes for Qualcomm's big zero day exploit last year certainly gives the impression that this was yet another example of a memory safety error. > According to the patch instructions, the fix works by adding direct memory access handle references. https://www.techtarget.com/searchsecurity/news/366612994/High-severity-Qualcomm-zero-day-vulnerability-under-attack https://www.techtarget.com/searchsecurity/news/366612994/Hig...
- DrBenCarson 2y agoSure, but they do protect from a massive swath of real attacks Don’t guarantee complete safety but do eliminate a massive attack surface
- iknowstuff 2y agoHere’s hoping. Bets on Swift v Rust? I swear if they wrote a modem from scratch in C that’s a major own goal
- DrBenCarson 2y agoDefinitely Swift https://swift.org/blog/embedded-swift-examples/ https://swift.org/blog/embedded-swift-examples/
- scrlk 2y agoThere's a number of firmware jobs on their website at the moment. All of them ask for C or C++ experience - I haven't seen anything that's asking for Rust or Embedded Swift experience. E.g.: https://jobs.apple.com/en-us/details/200579953/embedded-5g-4g-cellular-rf-sw-fw-engineer?team=HRDWR https://jobs.apple.com/en-us/details/200579953/embedded-5g-4...
- walterbell 2y agoWas the modem codebase inherited from Intel?
- scrlk 2y agohttps://www.bloomberg.com/news/articles/2023-11-16/apple-project-to-replace-modems-made-by-qualcomm-falls-further-behind-to-2026 https://www.bloomberg.com/news/articles/2023-11-16/apple-pro... (https://archive.ph/A2UUI https://archive.ph/A2UUI) > One major development snag involves the software used to power the modem, some of which was acquired from Intel. The Intel code wasn’t up to the task, and most of it had to be rewritten from scratch, people involved in the project said. As Apple engineers sought to add new features, existing capabilities would break, and the modem wouldn’t work properly.
- walterbell 2y agoLet the Apple C++ vs Qualcomm C++ zero day defense games begin!
- saagarjha 2y agoIt’s C(++) judging by the strings in the firmware
- internet2000 2y agoIs the firmware available somewhere already?
- saagarjha 2y agoYes it’s on the update channels
- conradev 2y agoIn this specific case, I imagine the system is much more secure. The Broadcom chip and Qualcomm chip each represented their own separate attack surface. Even without hacking the AP those chips were still a problem.