Re: linux-kernel-digest V1 #103

Alan Cox (
Mon, 3 Jul 1995 08:39:35 +0100 (BST)

> But I do also agree with Chad's points, a true filesystem ramdisk has
> it's advantages (not least being simplicity) including the mmap stuff,
> mmap on a userfs (or even just a tmpfs) would be ... interesting ...

And its vital for a real ramdisk and for XIP mapped ROM disks one day. The
point of tmpfs is speed, the point of userfs is flexibility. userfs is not
fast enough for the job.

> From: Linus Torvalds <>
> Hmm... It would be reasonably easy to change the current array of
> interrupt handlers to an array of a linked list of interrupt handlers:
> essentially just add a "struct irqaction * next" to the definition of
> irqaction, and make the irqactions be allocated dynamically when adding
> a new irq.

What about the times when IRQ lines are allocated before kmalloc/kfree are
available during boot ?

This would also be a good time to make the irq allocation more generic. At
the moment it basically assumes single CPU PC bus style interrupts. Some
systems need to have (bus,irq) pairs. [eg Intel SMP mode on Pentiums]

> - everything that shares an irq has the same "flags" (the irq_request()
> function could check this easily), especially as you can't reasonably
> mix slow and fast interrupts on the same line.

You can. The fast ones run first then the slow one. The end result is like a
slow one. It still works.