2012/5/29 Reinhard Tartler <siret...@tauware.de>:
> What I find a bit irritating is that while using the same glibc version,
> Debian/kFreeBSD ships a different /usr/include/bits/fnctl.h to the other
> Debian architectures.
>
> The kFreeBSD Version reads:
>
> struct flock
> {
> __off_t l_start; /* Offset where the lock begins. */
> __off_t l_len; /* Size of the locked area; zero means until EOF. */
> __pid_t l_pid; /* Process holding the lock. */
> short int l_type; /* Type of lock: F_RDLCK, F_WRLCK, or F_UNLCK. */
> short int l_whence; /* Where `l_start' is relative to (like `lseek'). */
> int __l_sysid; /* remote system id or zero for local */
> };
>
> While the Linux version reads:
>
> struct flock
> {
> short int l_type; /* Type of lock: F_RDLCK, F_WRLCK, or F_UNLCK. */
> short int l_whence; /* Where `l_start' is relative to (like `lseek'). */
> #ifndef __USE_FILE_OFFSET64
> __off_t l_start; /* Offset where the lock begins. */
> __off_t l_len; /* Size of the locked area; zero means until EOF. */
> #else
> __off64_t l_start; /* Offset where the lock begins. */
> __off64_t l_len; /* Size of the locked area; zero means until EOF. */
> #endif
> __pid_t l_pid; /* Process holding the lock. */
> };
>
>
> Why on earth has the kFreeBSD version of glibc decided to shuffle around
> the members of the flock struct?
Kernel ABI is different. Not all of glibc are user structures....
> While discussing the order in detail,
> the POSIX spec [1] does mention 'struct flock' having a order of members
> that matches the Linux version.
>
> [1] http://pubs.opengroup.org/onlinepubs/009695399/basedefs/fcntl.h.html
>
> The FreeBSD *kernel* on the other hand does not seem to obey this
> ordering [2]. Obviously, glibc has to model the flock structure after
> what is being passed down from the kernel. *sigh*
>
> [2] http://fxr.watson.org/fxr/source/sys/fcntl.h
...but it seems you already knew that. Why did you ask? :-)
If kernel ABI violates POSIX in some way, this should be discussed
with upstream (freebsd-a...@freebsd.org). Even if we agreed with you,
we couldn't change it on our own.
--
Robert Millan
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
