Re: Unimplemented SPARC system calls
On Fri, Feb 01, 2002 at 09:00:02PM -0700, Mark Zimmerman wrote:
> I've recently installed woody on an IPX (and updated to the 2.2.20
> kernel and I've noticed the following oddities in dmesg:
>
> init[1]: Unimplemented SPARC system call 69
> portmap[104]: Unimplemented SPARC system call 87
> cp[123]: Unimplemented SPARC system call 69
> mv[142]: Unimplemented SPARC system call 69
> inetd[176]: Unimplemented SPARC system call 53
> lpd[182]: Unimplemented SPARC system call 69
You can ignore these. This is libc attempting to use LFS syscalls, which
don't exist on 2.2.x kernels. It falls back safely to non-LFS syscalls
though.
> /* #define __NR_mctl 53 SunOS specific */
> /* #define __NR_sbrk 69 SunOS Specific */
> /* #define __NR_gethostname 87 SunOS Specific */
>
> I also looked at the same file in a 2.4.14 source tree and found:
>
> #define __NR_getgid32 53 /* Linux sparc32 specific */
> #define __NR_geteuid32 69 /* Linux sparc32, sbrk under SunOS */
> #define __NR_setuid32 87 /* Linux sparc32, gethostname under SunOS*/
>
> This, combined with the problems with sudo make me wonder if
> sparc-woody is 2.4-centric and if I am mistaken in using a 2.2 kernel.
They should not cause any problems. Other architectures exhibit the same
features, but just don't have the printk for it. Believe me, if seteuid
was broken on sparc, more things than sudo would be broken (su for one).
Ben
--
.----------=======-=-======-=========-----------=====------------=-=-----.
/ Ben Collins -- Debian GNU/Linux -- WatchGuard.com \
` bcollins@debian.org -- Ben.Collins@watchguard.com '
`---=========------=======-------------=-=-----=-===-======-------=--=---'
Reply to: