Re: [patch 5/8] Immediate Values - x86 Optimization (simplified)

[Rusty Russell <rusty@rustcorp.com.au>]

On Saturday 17 November 2007 01:03:35 Mathieu Desnoyers wrote:
> x86 optimization of the immediate values which uses a movl with code
> patching to set/unset the value used to populate the register used as
> variable source.

Since immediate values are by definition an optimization, I think it makes
sense to insist they be 1, 2, 4 or 8 bytes.  A BUILD_BUG_ON() in the right
place should ensure this (probably in generic code rather than x86).

> ===================================================================
> --- linux-2.6-lttng.orig/arch/x86/kernel/traps_32.c

I don't think you need any modification to this file now.

> + * Create the instruction in a discarded section to calculate its size. This is
> + * how we can align the beginning of the instruction on an address that will 
> + * permit atomic modificatino of the immediate value without knowing the size of 
> + * the opcode used by the compiler. The operand size is known in advance.
> + */ 

This alignment is also now unnecessary.

> +++ linux-2.6-lttng/arch/x86/kernel/immediate.c	2007-11-16
> 08:56:22.000000000 -0500 @@ -0,0 +1,143 @@
> +/*
> + * Immediate Value - x86 architecture specific code.

This is now almost entirely generic code, but I suppose we can let the next
architecture hoist it out.

> +/**
> + * arch_immediate_update_early - update one immediate value at boot time
> + * @immediate: pointer of type const struct __immediate to update
> + *
> + * Update one immediate value at boot time.
> + */
> +void arch_immediate_update_early(const struct __immediate *immediate)

I think it would be easier to just fast-path the num_online_cpus == 1 case,
even if you want to keep this "update_early" interface.

But I like your IPI algorithm: very tight.

Thanks,
Rusty.