On 24.01.21 17:08, John Scott wrote: > Changing the firmware on an EEPROM is far less practical for the user or > manufacturer (they're on similar footing), and if it's not electronically > erasable, it's merely an object that can't be practically changed of which > you'd need to make a new one anyway.
LVFS is a thing now (kudos to Richard Hughes) and firmware updates can nowadays be pretty seamless, even on Linux. So I don't think I agree that EEPROM updates are far less practical. And I think I'd still prefer if the kernel pushes the (driver-)appropriate firmware to the device as it sees fit rather than having explicit EEPROM update cycles independent from driver updates. > 1. Unlike with SSD firmware, there are wireless cards that use libre firmware > and some are still manufactured and quite easy to attain. The goalpost for > free software moves with what has been achieved. I guess to make your point stronger you could also have linked to those products that work with libre firmware. A brief research then finds two abgn cards from Atheros that is not available through normal retail channels anymore, because they are 8 to 10 years old (at least) and do not support contemporary wifi standards. And the same research turns up that it took many years from the point were it existed (2013) until it got uploaded to Debian (2017) and released (2019). I think its existence is super interesting from a research point of view. But I don't think it makes a strong case for availability of libre firmware for wifi cards. Especially if you care about spectral efficiency, i.e. using a shared medium efficiently. Kind regards Philipp Kern [1] https://libreplanet.org/wiki/LinuxLibre:Devices_that_require_non-free_firmware
