[patch] big IRQ lock removal, 2.5.27-F8

Ingo Molnar (mingo@elte.hu)
Tue, 23 Jul 2002 14:21:35 +0200 (CEST)


the latest irqlock patch, against 2.5.27-BK-current can be found at:

http://redhat.com/~mingo/remove-irqlock-patches/remove-irqlock-2.5.27-F8

the big change since -F3 is the introduction of a separate preemption,
hardirq and softirq bitrange within preempt_count:

* PREEMPT_MASK: 0x000000ff
* HARDIRQ_MASK: 0x0000ff00
* SOFTIRQ_MASK: 0x00ff0000
* IRQ_MASK: 0x00ffff00

with this it was possible to have a functional in_irq() again, so the
skbuff.c hack could be reverted.

The definition of the bitrange depends on the following base defines:

#define NR_PREEMPT 256
#define NR_HARDIRQ 256
#define NR_SOFTIRQ 256

NOTE: other, non-x86 architectures might want to define a tighter
bitrange, to make the preempt-only check cheaper. On x86 the above
bitrange is the most optimal.

compiles, boots, works just fine.

Changes in -F8:

- preempt/hardirq/softirq count separation, cleanups.

- skbuff.c fix.

- use irq_count() in scheduler_tick()

Changes in -F3:

- the entry.S cleanups/speedups by Oleg Nesterov.

- a rather critical synchronize_irq() bugfix: if a driver frees an
interrupt that is still being probed then synchronize_irq() locks up.
This bug has caused a spurious boot-lockup on one of my testsystems,
ifconfig would lock up trying to close eth0.

- remove duplicate definitions from asm-i386/system.h, this fixes
compiler warnings.

Ingo

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