Fine-grained locking for POSIX local sockets (UNIX domain sockets)

[Andrew Gallatin]

Daniel Eischen writes:
 > On Wed, 10 May 2006, David Xu wrote:
 > > One of the problems to implement it is that atomic operations,
 > > if there are multiple integer needs to be updated by kernel,
 > > userland maybe gets an inconsistent result, the way to avoid the
 > > problem is using two generation numbers.
 > >
 > > http://gsu.linux.org.tr/~mpekmezci/kernelapi/unitedlinux/arch/x86_64/kernel/vsyscall.c.html
 > > check do_vgettimeofday(struct timeval * tv):
 > >
 > > Another problem is how you tell userland the address of the kernel
 > > page ? do you use fixed address or tell it via program headers like
 > > the PT_TLS set by kernel, check /usr/src/lib/libc/gen/tls.c.
 > 
 > Can you not make a simple pseudo device driver and mmap the page?

In addition to this linux vsyscall, there is the MacOSX/Darwin
commpage.  The map machine-specific implementations of atomic
operations, bcopy, bzero, spinlocks, pthread_getspecific, etc into a
common page mapped into userspace applications. The also do a (mostly)
syscall-free gettimeoday this way.

See http://fxr.watson.org/fxr/source/osfmk/ppc/commpage/?v=xnu-792

Obviously, we could not take the code due to APSL infection (unless
Apple were to donate it), but it is something else to look at.

Drew