Skip to contentSkip to conversation

Still no ramfs usage limits in 2.5.9 or 2.4.19-pre8

DISCUSSION Dale Farnsworth 3 participants 4 messages May 16, 2002 – May 17, 2002

lore
Share & tools
lore
#0 lore raw

David Gibson produced a patch about 18 months ago that added usage limits to ramfs. <http://ozlabs.org/people/dgibson/dwg-ramfs.patch>

Through 2.4.10, the AC kernel series carried this patch.

Linus' kernel from 2.4.11 onward adopted only one line from the patch, the following attribution comment:
"Usage limits added by David Gibson, Linuxcare Australia."

The usage limit code is missing, however.

Was this a simple mistake that has been overlooked? Or are there drawbacks to applying the patch that keep it out of the mainstream kernels?

-Dale Farnsworth

#1 lore raw
Quoted 10 lines

David Gibson produced a patch about 18 months ago that added usage limits to ramfs. <http://ozlabs.org/people/dgibson/dwg-ramfs.patch>

Through 2.4.10, the AC kernel series carried this patch.

Linus' kernel from 2.4.11 onward adopted only one line from the patch, the following attribution comment:
"Usage limits added by David Gibson, Linuxcare Australia."

The usage limit code is missing, however.

Linus certainly never added it, or the address space hooks it needed. It should have been added though definitely

#2 lore raw
In reply to Dale Farnsworth
Quoted 9 lines · Dale Farnsworth

David Gibson produced a patch about 18 months ago that added usage limits to ramfs. <http://ozlabs.org/people/dgibson/dwg-ramfs.patch>

Through 2.4.10, the AC kernel series carried this patch.

Linus' kernel from 2.4.11 onward adopted only one line from the patch, the following attribution comment:
"Usage limits added by David Gibson, Linuxcare Australia."

Quoted 7 lines

The usage limit code is missing, however.

Was this a simple mistake that has been overlooked? Or are there drawbacks to applying the patch that keep it out of the mainstream kernels?

-Dale Farnsworth

Linus said the code should be kept as simple
as possible because it's only an example file system. However simple should also not mean practically useless. Also as of 2.5 it's mandatory, so it should get the limit code IMHO. There are also a few other patches floating around for ramfs:

http://marc.theaimsgroup.com/?l=linux-kernel&m=100553363224376&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=98701101629272&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=100753902321780&w=2

Padraig.

#3 lore raw
Quoted 3 lines

Linus said the code should be kept as simple
as possible because it's only an example file system. However simple should also not mean practically useless.

Its useless as an example too - because it has no examples of error handling