lkml.org 
[lkml]   [1999]   [Jun]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRAMdisk revisited
Date

I had to tweak drivers/block/rd.c for a 2.88meg boot/root floppy a while
back. This lead to my first posts to this list and I never got a sensible
explanation of why rd_size is 4096 (k). I don't know if kernel
"commandline arguments" will override that. I don't see where it would,
and this is not a problem that's worth using a bootloader to solve, or
bothering the user about at boot time.

Well now I have 64meg of RAM and my rd_size is 40960 (k). I had some other
code at one time that would set the RAMdisk to 1/2 actual RAM, but I
didn't bother with that this time.

The comments about this are somewhat ambiguous, BTW. rd_size now seems to
be a MAXIMUM, and I think if the MAXIMUM is too big, i.e. bigger than your
RAM, booting fails. There's no need for it to, since the RAMdisk may be
smaller than the maximum.

I thought it would be cute to post to this list, and do a few IRCs and so
on at the same time, from a box with a total of zero mounted persistant
storage devices, but alas, that stunt has defeated me. I did get booted to
a 36 meg RAMdisk as /, and was able to copy most of my 32 meg core setup
to the RAMdisk. I had about 300 commands all in RAM, Lynx working out of
RAM, etc. But not PPP. This was also using the cc1obj-built kernel. I
didn't see anything that looked like it was definitely the kernel not
working, though. Between my non-standard filename tree and cp -u'ing the
whole world onto / on RAM, 2.2.9-obj can't be blamed for anything.

The decompression in the kernel does a nice job of decompressing a
compressed 36 meg ext2 filesystem on a floppy. The floppy read waits
for sections of the deflate to complete, and the light on the drive stays
on, and it chugs right along. I was worried about that in 64 meg, but
no problem.

In other words, the way I got a 36 meg "image" on a 1.44 meg floppy was
the usual method for making a root floppy, but do so with a real big
blank ext2 loopback file. A 36 meg empty ext2 filesystem compresses to
about 38k, i.e. 1000:1 .

Performance 100% in RAM seems slightly snappier for a few things, like
du, find, etc. on a 486. It may not be noticable on a current machine.
I suppose there may be other uses for a big RAMdisk. Demo'ing Linux,
for one thing. Or is that policy? ;o)

Rick Hohensee
http://linux01.gwdg.de/~rhohen
cyluphonics rootsy guitar xart



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:52    [W:0.042 / U:0.100 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site