I've been looking into this, and found that defining __KERNEL__ before
including unistd.h fix the FTBFS. However defining __KERNEL__ is usually
wrong, and not desirable so digging further I've found this thread which
explains the absence of the _syscall macros:
http://www.mail-archive.com/linux-arch@vger.kernel.org/msg01459.html
According to that thread the proper fix would be to change the way mbr
does the syscalls to the new api. An alternative, less desirable fix,
would be to include the macro definitions that David Woodhouse mentions
on that thread that implement the _syscall macros using the new libc
syscall api (it mentions that Fedora has them in unistd.h).
Also after fixing/working around that I've found that the test suite
fails miserably in i386 (probably not on amd68 since mosts tests aren't
run anyway) because what seems to be a kernel change (vm86.eflags have a
different default value or something like that, haven't dig too much in
this).
I haven't bothered to provide a patch since I'm not sure mbr is so
useful anymore. But if it is, please let me know and I'll try to come up
with a patch. If it's not, please consider a removal request.
Hope to help,
--
Damián Viano(Des) ¯ ¯ - _ _ - ¯ ¯
GPG: 0x6EB95A6F Debian ¯-_GNU_-¯ Linux
Web: http://damianv.com.ar/ ¯-¯
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
