Re: (E)GLIBC: inline syscalls
There are two different macros INTERNAL_SYSCALL and INLINE_SYSCALL.
It remains to persuade Ulrich Drepper to distinct between them :-(
Unfortunately he his not someone easy to convince...
I know ...
But in fact, the HAVE_INLINED_SYSCALLS currently
means only HAVE_INTERNAL_SYSCALLS ...
The calls to INLINE_SYSCALL is not the part that worries me the most. It's
rather the code between #ifdef HAVE_INLINED_SYSCALLS in the elf/
directory. Not something we can workaround easily from our
ports/sysdeps/directory.
I would suggest declare HAVE_INLINED_SYSCALLS, but keep current
"#define INLINE_SYSCALL(name, nr, args...) __syscall_##name (args)"
and only provide INTERNAL_SYSCALL as
#define INTERNAL_SYSCALL(name, err, nr, args...) INTERNAL_SYSCALL_##name(err, args...)
Currently there is only one needed - INTERNAL_SYSCALL_writev due to
dl-misc.c: INTERNAL_SYSCALL (writev, err, 3, fd, &iov, niov)
It can be implemented similarly as in kfreebsd/i386/linuxthreads/tls.h
# define TLS_DO_SET_GSBASE(descr) \
({ \
long base = (long) descr; \
int result; \
asm volatile ( \
"pushl %3\n\t" \
"pushl %2\n\t" \
"pushl %1\n\t" \
"int $0x80\n\t" \
"popl %3\n\t" \
"popl %3\n\t" \
"popl %3\n\t" \
: "=a" (result) \
: "0" (SYS_sysarch), \
"i" (I386_SET_GSBASE), \
"d" (&base) \
: "memory", "cc" ); \
result; \
})
Petr
Reply to: