Revision 8339f0008c47cdd921c73f6d53d5588b5484f93c authored by Eric W. Biederman on 29 January 2007, 20:19:05 UTC, committed by Linus Torvalds on 30 January 2007, 16:29:58 UTC
When the world was a simple and static place setting up irqs was easy.
It sufficed to allocate a linux irq number and a find a free cpu
vector we could receive that linux irq on.  In those days it was
a safe assumption that any allocated vector was actually in use
so after one global pass through all of the vectors we would have
none left.

These days things are much more dynamic with interrupt controllers
(in the form of MSI or MSI-X) appearing on plug in cards and linux
irqs appearing and disappearing.  As these irqs come and go vectors
are allocated and freed,  invalidating the ancient assumption that all
allocated vectors stayed in use forever.

So this patch modifies the vector allocator to walk through every
possible vector before giving up, and to check to see if a vector
is in use before assigning it.  With these changes we stop leaking
freed vectors and it becomes possible to allocate and free irq vectors
all day long.

This changed was modeled after the vector allocator on x86_64 where
this limitation has already been removed.  In essence we don't update
the static variables that hold the position of the last vector we
allocated until have successfully allocated another vector.  This
allows us to detect if we have completed one complete scan through
all of the possible vectors.

Acked-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent c9cc8e7
Raw File
Kconfig.preempt

choice
	prompt "Preemption Model"
	default PREEMPT_NONE

config PREEMPT_NONE
	bool "No Forced Preemption (Server)"
	help
	  This is the traditional Linux preemption model, geared towards
	  throughput. It will still provide good latencies most of the
	  time, but there are no guarantees and occasional longer delays
	  are possible.

	  Select this option if you are building a kernel for a server or
	  scientific/computation system, or if you want to maximize the
	  raw processing power of the kernel, irrespective of scheduling
	  latencies.

config PREEMPT_VOLUNTARY
	bool "Voluntary Kernel Preemption (Desktop)"
	help
	  This option reduces the latency of the kernel by adding more
	  "explicit preemption points" to the kernel code. These new
	  preemption points have been selected to reduce the maximum
	  latency of rescheduling, providing faster application reactions,
	  at the cost of slighly lower throughput.

	  This allows reaction to interactive events by allowing a
	  low priority process to voluntarily preempt itself even if it
	  is in kernel mode executing a system call. This allows
	  applications to run more 'smoothly' even when the system is
	  under load.

	  Select this if you are building a kernel for a desktop system.

config PREEMPT
	bool "Preemptible Kernel (Low-Latency Desktop)"
	help
	  This option reduces the latency of the kernel by making
	  all kernel code (that is not executing in a critical section)
	  preemptible.  This allows reaction to interactive events by
	  permitting a low priority process to be preempted involuntarily
	  even if it is in kernel mode executing a system call and would
	  otherwise not be about to reach a natural preemption point.
	  This allows applications to run more 'smoothly' even when the
	  system is under load, at the cost of slighly lower throughput
	  and a slight runtime overhead to kernel code.

	  Select this if you are building a kernel for a desktop or
	  embedded system with latency requirements in the milliseconds
	  range.

endchoice

config PREEMPT_BKL
	bool "Preempt The Big Kernel Lock"
	depends on SMP || PREEMPT
	default y
	help
	  This option reduces the latency of the kernel by making the
	  big kernel lock preemptible.

	  Say Y here if you are building a kernel for a desktop system.
	  Say N if you are unsure.

back to top