I'm compiling 2.5.45 for 486.
For example, exit.c got compiled with this command:
gcc -v -Wp,-MD,kernel/.exit.o.d -D__KERNEL__ -Iinclude -Wall \
-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer \
-fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 \
-march=i486 -Iarch/i386/mach-generic -nostdinc -iwithprefix include \
-DKBUILD_BASENAME=exit -c -o zz25.o kernel/exit.c
Whereas for 2.4, command is different:
gcc -v -D__KERNEL__ -Iinclude -Iarch/i386/mach-generic \
-Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing \
-fno-common -pipe -mpreferred-stack-boundary=2 -march=i386 -nostdinc \
-iwithprefix include -DKBUILD_BASENAME=exit -c -o zz24.o kernel/exit.c
(commands slightly adapted to actually execute in 2.5 tree and
don't destroy exit.o)
As you can see, the difference is in -march=.
2.5 compiles for 486. I have nothing against 486, but
gcc 3.x got nasty habit of aligning functions and sometimes
*even loops* to 16 bytes. Result:
objdump -D exit.o, abridged
=================================
exit.o: file format elf32-i386
Disassembly of section .text:
00000000 <__unhash_process>:
00000130 <release_task>:
000002f0 <unhash_process>:
348: 90 nop
349: 8d b4 26 00 00 00 00 lea 0x0(%esi,1),%esi
00000350 <session_of_pgrp>:
3d7: 90 nop
3d8: 90 nop
3d9: 8d b4 26 00 00 00 00 lea 0x0(%esi,1),%esi
000003e0 <__will_become_orphaned_pgrp>:
46b: 90 nop
46c: 8d 74 26 00 lea 0x0(%esi,1),%esi
00000470 <will_become_orphaned_pgrp>:
4ac: 8d 74 26 00 lea 0x0(%esi,1),%esi
000004b0 <is_orphaned_pgrp>:
4bf: 90 nop
Tons of wasted space.
my .config, oldconfig'ed from 2.4, is below. What's that, marked <=== ?
#
# Processor type and features
#
# CONFIG_M386 is not set
CONFIG_M486=y
# CONFIG_M586 is not set
# CONFIG_M586TSC is not set
# CONFIG_M586MMX is not set
# CONFIG_M686 is not set
# CONFIG_MPENTIUMIII is not set
# CONFIG_MPENTIUM4 is not set
# CONFIG_MK6 is not set
# CONFIG_MK7 is not set
# CONFIG_MELAN is not set
# CONFIG_MCRUSOE is not set
# CONFIG_MWINCHIPC6 is not set
# CONFIG_MWINCHIP2 is not set
# CONFIG_MWINCHIP3D is not set
# CONFIG_MCYRIXIII is not set
CONFIG_X86_CMPXCHG=y
CONFIG_X86_XADD=y
CONFIG_X86_L1_CACHE_SHIFT=4
CONFIG_RWSEM_XCHGADD_ALGORITHM=y
CONFIG_X86_PPRO_FENCE=y
CONFIG_X86_F00F_BUG=y
CONFIG_X86_WP_WORKS_OK=y
CONFIG_X86_INVLPG=y
CONFIG_X86_BSWAP=y
CONFIG_X86_POPAD_OK=y
CONFIG_X86_USE_STRING_486=y
CONFIG_X86_ALIGNMENT_16=y <============= ??!
# CONFIG_HUGETLB_PAGE is not set
CONFIG_SMP=y
# CONFIG_PREEMPT is not set
CONFIG_X86_UP_APIC=y
CONFIG_X86_LOCAL_APIC=y
CONFIG_X86_IO_APIC=y
CONFIG_NR_CPUS=32
# CONFIG_X86_NUMA is not set
CONFIG_X86_MCE=y
CONFIG_X86_MCE_NONFATAL=y
CONFIG_X86_MCE_P4THERMAL=y
# CONFIG_CPU_FREQ is not set
# CONFIG_TOSHIBA is not set
# CONFIG_I8K is not set
CONFIG_MICROCODE=m
CONFIG_X86_MSR=m
CONFIG_X86_CPUID=m
# CONFIG_EDD is not set
CONFIG_NOHIGHMEM=y
# CONFIG_HIGHMEM4G is not set
# CONFIG_HIGHMEM64G is not set
# CONFIG_MATH_EMULATION is not set
CONFIG_MTRR=y
CONFIG_HAVE_DEC_LOCK=y
On Sun, 3 Nov 2002, Denis Vlasenko wrote:
> I'm compiling 2.5.45 for 486.
>
> For example, exit.c got compiled with this command:
>
> gcc -v -Wp,-MD,kernel/.exit.o.d -D__KERNEL__ -Iinclude -Wall \
> -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer \
> -fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 \
> -march=i486 -Iarch/i386/mach-generic -nostdinc -iwithprefix include \
> -DKBUILD_BASENAME=exit -c -o zz25.o kernel/exit.c
>
> Whereas for 2.4, command is different:
>
> gcc -v -D__KERNEL__ -Iinclude -Iarch/i386/mach-generic \
> -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing \
> -fno-common -pipe -mpreferred-stack-boundary=2 -march=i386 -nostdinc \
> -iwithprefix include -DKBUILD_BASENAME=exit -c -o zz24.o kernel/exit.c
>
> (commands slightly adapted to actually execute in 2.5 tree and
> don't destroy exit.o)
>
> As you can see, the difference is in -march=.
> 2.5 compiles for 486. I have nothing against 486, but
> gcc 3.x got nasty habit of aligning functions and sometimes
> *even loops* to 16 bytes. Result:
>From memory that's to handle cache line length. I seem to remember back in
the 486 days setting my flags explicitly for anything which might be the
target of a branch. However, my recollection is that things were put on a
quadbyte, not position 16.
You can set the placement of loops, functions, and something else to be
mod any power of two, I believe.
I still run routers on one 486 and a boatload of Cyrix and similar clones.
--
bill davidsen <[email protected]>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.