Hi!
Rik van Riel wrote:
I just started reading the vserver patch again and am now officially scared to hell:
.long SYMBOL_NAME(sys_ni_syscall) /* sys_free_hugepages */ .long SYMBOL_NAME(sys_ni_syscall) /* sys_exit_group */ + .long SYMBOL_NAME(sys_new_s_context) + .long SYMBOL_NAME(sys_set_ipv4root) .long SYMBOL_NAME(sys_ni_syscall) /* sys_lookup_dcookie */ .long SYMBOL_NAME(sys_ni_syscall) /* sys_epoll_create */
This patch fragment to entry.S adds two syscalls into the middle of the table, shifting the syscalls below it to different numbers!
I always edit and and move my patches to the end :)
The userland programs inspect /proc/self/status and search for: __NR_new_s_context: 260 __NR_set_ipv4root: 261 rev2
...so the syscall number can be dinamic. It would be nice to get a fixed one, though.
Does vserver have reserved syscall numbers already ? If so, are there any plans to get them ?
I don't think so... Jacques is the man to ask :)
How about a sys_vserver multiplexer so we can easily add things like setting a new ipv6 root to the interface, without needing yet another syscall ?
Even nicer! :)
Herbert has a wiki with some wishes for vserver-2.6 at http://vserver.13thfloor.at/Linux2.6/
This would be a good adition to it ;)
Regards, Nuno Silva
Rik
