On 2009-07-06 11:39, Ruben de Groot wrote:
>>> r...@amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC amd64
>>>
>>> atom# du -hs /boot/kernel*
>>> 205M /boot/kernel
>>
>> Right, so it's a lot bigger on amd64. I guess those 64-bit pointers
>> aren't entirely free. :)
>
> I'm not sure where the size difference comes from. I have some sparc64
> systems running -current with symbols and the size of /boot/kernel is
> more comparable to i386, even with the 8-byte pointer size:
>
>> uname -p
> sparc64
>> du -sk /boot/kernel
> 137918 /boot/kernel
It looks like on amd64, the kernel is compiled with optimization flags:
-O2 -frename-registers -pipe -fno-strict-aliasing
by default, while on i386, they are just:
-O -pipe
Maybe this accounts for the huge difference? Does -O2 do a lot more
inlining?
In any case, it's a weird inconstency, if you ask me. But it's
intentional, as /usr/src/sys/conf/kern.pre.mk says:
. if ${MACHINE_ARCH} == "amd64"
COPTFLAGS?=-O2 -frename-registers -pipe
. else
COPTFLAGS?=${_MINUS_O} -pipe
. endif
where ${_MINUS_O} is by default just "-O", since DEBUG is enabled...
_______________________________________________
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
