Date: Mon, 2 Jan 1995 04:55:04 +1100 From: Bruce Evans <bde@zeta.org.au> To: roberto@blaise.ibp.fr Cc: hackers@freebsd.org Subject: Re: Problem compiling current kernel (1st jan) Message-ID: <199501011755.EAA22724@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>> cc -c -pipe -W -Wreturn-type -Wcomment -Wredundant-decls -I. -I../.. -I../../sys -DKELTIA243 -DI486_CPU -DOPEN_MAX=128 -DCHILD_MAX=128 -DSYSVMSG -DSYSVSEM -DSYSVSHM -DTEST_LABELLING -DDISKSLICE -DDUMMY_NOPS -DAUTO_EOI_2 -DAUTO_EOI_1 -DARP_PROXYALL -DIPFI> REWALL_VERBOSE -DIPFIREWALL -DSCSI_2_DEF -DNCONS=8 -DUCONSOLE -DCOMPAT_43 -DMFS -DLFS -DFFS -DINET -DKERNEL -Di386 -DLOAD_ADDRESS=0xF0100000 ../../i386/isa/bt742a.c >> ./machine/cpufunc.h: In function `inbc': >> ./machine/cpufunc.h:125: inconsistent operand constraints in an `asm' >>... >I've found why : I have redefined COPTFLAGS in my /etc/make.conf file in order to >have a -pipe and forget to put the -O. >That's weird, the kernel should compile without -O... It never has in FreeBSD :-]. There is another problem with asms in npx.c. The problem with inbc() and outbc() is that the "i" (immediate mode) constraint probably isn't valid for args in inline functions. The args are constant literals before the call, but strictly speaking, function args are never constant literals. inb() and outb() are already ugly macros instead of inline functions to avoid the same problem with __builtin_constant_p(). Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199501011755.EAA22724>