By: Doug S (foo.delete@this.bar.bar), June 22, 2020 2:50 pm
Room: Moderated Discussions
Linus Torvalds (torvalds.delete@this.linux-foundation.org) on June 22, 2020 12:55 pm wrote:
> Linus Torvalds (torvalds.delete@this.linux-foundation.org) on June 22, 2020 12:46 pm wrote:
> >
> > Will be interesting to see the reactions from people leaking out the development
> > hardware (despite possibly NDA's in place for early hardware access).
>
> Side note: regardless, it's a big step for ARM. I've been complaining
> about lack of real developer hardware for a long long time.
>
> I don't personally like doing kernel development in virtual machines (which sounds odd: a lot of
> people love it because it makes some things easier), but if that VM is well done and not overly paravirtualized,
> it might be the best way to do at least some ARM kernel development so far (*).
>
> That perhaps says more about the state of non-apple ARM infrastructure than
> it says about the new Apple hardware. Let's just say that I've been (somewhat
> loudly) disappointed in the ARM ecosystem for a long long time...
>
> Linus
>
> (*) Running Linux native on the hardware would be better, but Apple has traditionally tried to lock down
> their hardware to annoying degrees, and with their own silicon that lockdown tends to be only stronger.
I think your problem might end up being drivers moreso than lockdown, since Apple won't have any reason or incentive to open source the drivers for their own designed hardware stuff. Obviously the GPU would be one such, though you can work around it since no doubt they'll still offer machines with discrete GPUs like Radeon or whatever. But it may be a show stopper elsewhere...
> Linus Torvalds (torvalds.delete@this.linux-foundation.org) on June 22, 2020 12:46 pm wrote:
> >
> > Will be interesting to see the reactions from people leaking out the development
> > hardware (despite possibly NDA's in place for early hardware access).
>
> Side note: regardless, it's a big step for ARM. I've been complaining
> about lack of real developer hardware for a long long time.
>
> I don't personally like doing kernel development in virtual machines (which sounds odd: a lot of
> people love it because it makes some things easier), but if that VM is well done and not overly paravirtualized,
> it might be the best way to do at least some ARM kernel development so far (*).
>
> That perhaps says more about the state of non-apple ARM infrastructure than
> it says about the new Apple hardware. Let's just say that I've been (somewhat
> loudly) disappointed in the ARM ecosystem for a long long time...
>
> Linus
>
> (*) Running Linux native on the hardware would be better, but Apple has traditionally tried to lock down
> their hardware to annoying degrees, and with their own silicon that lockdown tends to be only stronger.
I think your problem might end up being drivers moreso than lockdown, since Apple won't have any reason or incentive to open source the drivers for their own designed hardware stuff. Obviously the GPU would be one such, though you can work around it since no doubt they'll still offer machines with discrete GPUs like Radeon or whatever. But it may be a show stopper elsewhere...