On Friday, March 05, 2021 11:56:24 AM David Wright wrote: > On Thu 04 Mar 2021 at 15:47:37 (-0500), rhkra...@gmail.com wrote:
<cutting the context ;-) > > If a device is sold on a separate card, it's not necessarily > enough to know the model number of the card. Many "identical" > models are sold with various different chips, which will > require different firmware. You might not know which chip you've > got until you look at the board, or even read its codes from > the dmesg output. > > Being non-free, the firmware usually originates/d from some > manufacturer or other. If the firmware fails to work with the > device, there's not much that Debian can do about it. It might > be something for some sub-sub-group of the linux kernel people, > if the problem lies in how the driver and firmware interact. > > So in your scheme, the "unofficial installers" that have to be > "vetted" by someone to confirm they "indeed work on those > hardware configurations" are actually hundreds of different > combinations, each one comprising one particular firmware blob, > plus the same old official installer image: > > iwlwifi-100-5.ucode + official installer ✓ Vetted ✓ Passed > iwlwifi-105-6.ucode + official installer ✓ Vetted ✓ Passed > iwlwifi-135-6.ucode + official installer ✓ Vetted ✓ Passed > iwlwifi-1000-5.ucode + official installer ✓ Vetted ✓ Passed > iwlwifi-2000-6.ucode + official installer ✓ Vetted ✓ Passed > iwlwifi-2030-6.ucode + official installer ✓ Vetted ✓ Passed > … … … … … > > ad infinitum … It would be nice (imho), but may be difficult. ;-)
