| 1 |
On Wed, Dec 26, 2012 at 9:03 AM, Bruce Hill |
| 2 |
<daddy@×××××××××××××××××××××.com> wrote: |
| 3 |
> On Wed, Dec 26, 2012 at 09:24:20AM -0500, Todd Goodman wrote: |
| 4 |
>> * Bruce Hill <daddy@×××××××××××××××××××××.com> [121225 18:30]: |
| 5 |
>> > > |
| 6 |
>> > > Try reading the kernel Documentation. (e.g., |
| 7 |
>> > > /usr/src/linux/Documentation/filesystems/ramfs-rootfs-initramfs.txt.) |
| 8 |
>> > > |
| 9 |
>> > > initramfs is an improvement over initrd. |
| 10 |
>> > > |
| 11 |
>> > > Todd |
| 12 |
>> > |
| 13 |
>> > Having read it years ago it still fails to give me a good reason for using it. |
| 14 |
>> |
| 15 |
>> It gives plenty of good reasons. |
| 16 |
>> |
| 17 |
>> If they aren't good for you then fine. |
| 18 |
>> |
| 19 |
>> But if you read it you wouldn't be asking why initrd went away and was |
| 20 |
>> replaced by initramfs. |
| 21 |
>> |
| 22 |
>> Todd |
| 23 |
> |
| 24 |
> Actually I had not read it in quite a number of years, did this morning, and |
| 25 |
> you are entirely correct. Perhaps all those years ago when an initrd was |
| 26 |
> required at times, I'd just held onto my mkinitrd script and didn't want to |
| 27 |
> change; and since there's no need for an initrd now, I didn't actually read |
| 28 |
> it, but clung to incorrect memories. |
| 29 |
|
| 30 |
|
| 31 |
One interesting small point I got out of the docs that Neil pointed me |
| 32 |
toward: That since linux-2.6 we're all using an initramfs |
| 33 |
|
| 34 |
"The 2.6 kernel build process always creates a gzipped cpio format initramfs |
| 35 |
archive and links it into the resulting kernel binary. By default, this |
| 36 |
archive is empty (consuming 134 bytes on x86)." |
| 37 |
|
| 38 |
So it's a nit but no one should be saying "I don't use an init thingy" |
| 39 |
but rather "My init thingy is empty and has no jobs to do on my |
| 40 |
system". (Or at least that's my understanding...) |
| 41 |
|
| 42 |
- Mark |
Archives