H. Peter Anvin hpa at zytor.com
Mon Sep 5 12:55:23 PDT 2011

On 09/05/2011 08:19 AM, Shao Miller wrote:
> Do we leave the initramfs' in-memory representation intact, for the 
> potential purpose of actually "re-passing" the very same initramfs to a 
> Linux?  If so, then we either:
> - [potentially] walk the whole initramfs for every file access, or
> - do a one-time pass and allocate [a bunch of] memory to do file 
> accounting (such as tracking the position of file data for files 
> overwritten by concatenated archives)

I would say (b)... we can just create a hash table or something like that.

> Or:
> Can I [ab]use the initramfs' range of memory to [slightly] re-organize 
> into a more efficient "filesystem" with a one-time pass?
> Also: Do we make this a "cpio FS" driver that uses the block-caching 
> abstraction (and hook the disk I/O to actually do memory access, 
> instead) or do we make this an "initramfs" driver that will never allow 
> for supporting initramfs on a disk, but only in memory?  I could see a 
> use for the former, if we every wish to allow "mounting" initramfs files 
> that are not in memory, but it seems a tad more complicated.

We can do the block cache abstraction, or even block driver abstraction,
and just point it into memory.


More information about the Syslinux mailing list