i386: mark cpu_dev structures as __cpuinitdata
In-Reply-To: <1155518783.5764.10.camel@localhost>
Quoted 11 lines · Magnus Damm
Quoted 7 lines
Quoted 4 lines
The different
cpu_devstructures are all used from__cpuinitcallers what I can tell. So mark them as__cpuinitdatainstead of__initdata. I am a little bit unsure about arch/i386/common.c:default_cpu, especially when it comes to the purpose ofthis_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_CPUdisabled then__cpuinitdatawill become__initdata- ie the same as before. Not a single byte wasted what I can tell.
I was talking about wasted space with HOTPLUG_CPU enabled, of course. Nobody is ever going to hotplug a VIA, Cyrix, Geode, etc. CPU, yet your patch makes the kernel carry that code and data anyway.
Yes, the checking scripts will complain. But we know it's OK.
Signature · 1 line
Chuck
Quoted 12 lines · Chuck Ebbert
In-Reply-To: <1155518783.5764.10.camel@localhost>
Quoted 8 lines · Magnus Damm
Quoted 7 lines
Quoted 4 lines
The different
cpu_devstructures are all used from__cpuinitcallers what I can tell. So mark them as__cpuinitdatainstead of__initdata. I am a little bit unsure about arch/i386/common.c:default_cpu, especially when it comes to the purpose ofthis_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.
Quoted 7 lines
Quoted 3 lines
How could this be wasting space? If you compile with
CONFIG_HOTPLUG_CPUdisabled then__cpuinitdatawill become__initdata- ie the same as before. Not a single byte wasted what I can tell.I was talking about wasted space with
HOTPLUG_CPUenabled, of course. Nobody is ever going to hotplug a VIA, Cyrix, Geode, etc. CPU, yet your patch makes the kernel carry that code and data anyway.
remember that suspend uses software hot(un)plug as well...
Quoted 17 lines · Arjan van de Ven
Quoted 16 lines · Chuck Ebbert
In-Reply-To: <1155518783.5764.10.camel@localhost>
Quoted 11 lines · Magnus Damm
Quoted 7 lines
Quoted 4 lines
The different
cpu_devstructures are all used from__cpuinitcallers what I can tell. So mark them as__cpuinitdatainstead of__initdata. I am a little bit unsure about arch/i386/common.c:default_cpu, especially when it comes to the purpose ofthis_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_CPUdisabled then__cpuinitdatawill become__initdata- ie the same as before. Not a single byte wasted what I can tell.
Quoted 5 lines
Quoted 3 lines
I was talking about wasted space with
HOTPLUG_CPUenabled, of course. Nobody is ever going to hotplug a VIA, Cyrix, Geode, etc. CPU, yet your patch makes the kernel carry that code and data anyway.remember that suspend uses software hot(un)plug as well...
Only for non-boot CPUs. The vendors above (with exception of VIA) never made SMP systems.
Dave
Signature · 1 line
http://www.codemonkey.org.uk