Skip to contentSkip to conversation

i386: mark cpu_dev structures as __cpuinitdata

PATCH Chuck Ebbert 3 participants 3 messages Aug 11, 2006 – Aug 14, 2006

lore
Share & tools
lore
#0 lore raw

In-Reply-To: <[email protected]>

Quoted 6 lines · Andi Kleen

From: Magnus Damm <[email protected]>

The different cpu_dev structures are all used from __cpuinit callers what I can tell. So mark them as __cpuinitdata instead of __initdata. I am a little bit unsure about arch/i386/common.c:default_cpu, especially when it comes to the purpose of this_cpu.

But none of these CPUs supports hotplug and only one (AMD) does SMP. So this is just wasting space in the kernel at runtime.

If anything I would only do this for AMD.

Same for the other patch that does more of this kind of change.

(IIRC I tried to do this a while ago and was told not to.)

Signature · 1 line
Chuck
#1 lore raw
Quoted 9 lines

But none of these CPUs supports hotplug and only one (AMD) does SMP. So this is just wasting space in the kernel at runtime.

If anything I would only do this for AMD.

Same for the other patch that does more of this kind of change.

(IIRC I tried to do this a while ago and was told not to.)

But wouldn't the reference check during build always warn if that wasn't fixed?

-Andi

#2 lore raw
In reply to Chuck Ebbert
Quoted 6 lines · Chuck Ebbert

In-Reply-To: <[email protected]>

Quoted 2 lines · Andi Kleen

From: Magnus Damm <[email protected]>

Quoted 7 lines
Quoted 4 lines

The different cpu_dev structures are all used from __cpuinit callers what I can tell. So mark them as __cpuinitdata instead of __initdata. I am a little bit unsure about arch/i386/common.c:default_cpu, especially when it comes to the purpose of this_cpu.

But none of these CPUs supports hotplug and only one (AMD) does SMP. So this is just wasting space in the kernel at runtime.

How could this be wasting space? If you compile with CONFIG_HOTPLUG_CPU disabled then __cpuinitdata will become __initdata - ie the same as before. Not a single byte wasted what I can tell.

The first version of this patch simply added a missing __init to some function, but I was then corrected by akpm that __cpuinit should be used instead.

Thanks,

/ magnus