| 1 |
On Tue, Jan 29, 2019 at 20:58:37 +0000, Wol's lists wrote: |
| 2 |
> On 29/01/2019 19:41, Grant Taylor wrote: |
| 3 |
> > The kernel /must/ have (at least) the minimum drivers (and dependencies) |
| 4 |
> > to be able to boot strap. It doesn't matter if it's boot strapping an |
| 5 |
> > initramfs or otherwise. |
| 6 |
|
| 7 |
> > All of these issues about lack of a driver are avoided by having the |
| 8 |
> > driver statically compiled into the kernel. |
| 9 |
|
| 10 |
> I'm not sure to what extent it's true of 64-bit hardware, but one of the |
| 11 |
> big problems with non-module kernels is actually being able to load them |
| 12 |
> into the available ram ... something to do with BG's "640K should be |
| 13 |
> enough for anyone". |
| 14 |
|
| 15 |
Uh? I've never had problems with my hand-configured kernels fitting |
| 16 |
into RAM, regardless of whether it's a 32-bit or 64-bit processor. |
| 17 |
|
| 18 |
A modular kernel is a workaround for binary kernels needing to support |
| 19 |
any and all hardware devices. If you load drivers for all these devices |
| 20 |
at the same time, you may well run out of RAM (or have done so in the |
| 21 |
relatively recent past). |
| 22 |
|
| 23 |
If you're configuring the kernel for a specific machine, I can't see how |
| 24 |
you could run out of RAM, unless there's too little of it to run |
| 25 |
GNU/Linux anyway. |
| 26 |
|
| 27 |
> Cheers, |
| 28 |
> Wol |
| 29 |
|
| 30 |
-- |
| 31 |
Alan Mackenzie (Nuremberg, Germany). |
Archives