| 1 |
On Thu, 29 Mar 2012 13:13:40 -0400, Michael Mol wrote: |
| 2 |
|
| 3 |
> On Thu, Mar 29, 2012 at 12:35 PM, Neil Bothwick <neil@××××××××××.uk> |
| 4 |
> wrote: |
| 5 |
|
| 6 |
> >> I'll articulate a few. (i) The initramfs involves having two copies |
| 7 |
> >> of lots of software around. |
| 8 |
> > |
| 9 |
> > Lots? For most people busybox is enough! If you want encrypted |
| 10 |
> > filesystems on LVM over RAID that rises to a total of four |
| 11 |
> > executables. |
| 12 |
> |
| 13 |
> And anything they might conceivably link to. Not everything supports |
| 14 |
> static linking. |
| 15 |
|
| 16 |
Those four all have static version, there are no libraries in my |
| 17 |
initramfs. |
| 18 |
|
| 19 |
> Don't forget boot-time X-based animation, too. That's an |
| 20 |
> extraordinarily common feature of mainstream desktop distributions. |
| 21 |
> And there will be other things, I'm sure. |
| 22 |
|
| 23 |
I don't get involved with those, but I'd hope something intended to be |
| 24 |
run so early would have minimal dependencies, if any. |
| 25 |
|
| 26 |
> >> (ii) What's more, these two copies are often |
| 27 |
> >> different, one being built with static libraries, the other with |
| 28 |
> >> dynamic ones. (iii) This situation is not (as far as I know) yet |
| 29 |
> >> handled by Portage, which means in building such software |
| 30 |
> >> statically, you've got to save the dynamic version somewhere else |
| 31 |
> >> whilst you're doing it. |
| 32 |
> > |
| 33 |
> > That's wrong. For example, LVM builds dynamic executable plus the |
| 34 |
> > lvm.static file for use in the initramfs. |
| 35 |
> |
| 36 |
> That's exactly what Alan just noted in (ii), but perhaps portage |
| 37 |
> handles (iii) in the case of LVM. |
| 38 |
|
| 39 |
Exactly, there are static and dynamic files, all handled by portage. |
| 40 |
|
| 41 |
> >> (iv) |
| 42 |
> >> The initramfs requires a potentially long script to make it work. |
| 43 |
> > |
| 44 |
> > Mount /proc, /sys and /dev. |
| 45 |
> > Mount root |
| 46 |
> > Unmount /proc, /sys and /dev. |
| 47 |
> > switch_root |
| 48 |
> |
| 49 |
> Things look much simpler when you abstract away the details. You still |
| 50 |
> have to manage lvm, mdraid and whatever else is necessary for mounting |
| 51 |
> things. That's where 'potentially long' came from, I expect. |
| 52 |
> |
| 53 |
> >> I think that qualifies the initramfs solution as ugly. |
| 54 |
> > |
| 55 |
> > Only if you build the initramfs with USE="fud". |
| 56 |
> |
| 57 |
> FUD: "Fear, uncertainty and doubt" |
| 58 |
> |
| 59 |
> In short, three things which are important to rationally examine and |
| 60 |
> deal with on a case-by-case basis. |
| 61 |
|
| 62 |
Yes, ideally before you start spreading them instead of vague handwaving |
| 63 |
about initramfs being ugly and using "lots of files" (four only counts at |
| 64 |
lots when applied to wives). |
| 65 |
|
| 66 |
|
| 67 |
|
| 68 |
-- |
| 69 |
Neil Bothwick |
| 70 |
|
| 71 |
Loose bits sink chips. |
Archives