Robert Millan a écrit :
On Sat, Jan 21, 2006 at 12:11:52AM +0100, Aurelien Jarno wrote:This is the comment in the FreeBSD kernel, file sys/kern/kern_xxx.c /* * This is the FreeBSD-1.1 compatable uname(2) interface. These * days it is done in libc as a wrapper around a bunch of sysctl's. * This must maintain the old 1.1 binary ABI. */Then if you look at the code corresponding to the syscall, you will see that the syscall correspond to a call to the corresponding sysctl, buttruncated: ... name[0] = CTL_KERN; name[1] = KERN_OSTYPE; len = sizeof (uap->name->sysname); mtx_lock(&Giant); error = userland_sysctl(td, name, 2, uap->name->sysname, &len, 1, 0, 0, 0); if (error) goto done2; subyte( uap->name->sysname + sizeof(uap->name->sysname) - 1, 0); ... So, in short the current implementation in our glibc is the correct one.Weird. How about disabling SYS_uname macro, then?
I don't see why we should disable it here. This syscall does its job correclty, ie it returns the requested data, but in a format different from our userland structure (32 bytes field instead of 256 bytes).
There is a lot of unused syscalls in sys/syscalls.h. Maybe we could decide not to export this file, but I don't think it is a good idea to remove some syscalls just because they are not suitable for the corresponding glibc functions.
Aurelien -- .''`. Aurelien Jarno | GPG: 1024D/F1BCDB73 : :' : Debian developer | Electrical Engineer `. `' aurel32@debian.org | aurelien@aurel32.net `- people.debian.org/~aurel32 | www.aurel32.net