Skip to contentSkip to conversation

Syscall trouble

DISCUSSION Søren Nøhr Christensen 2 participants 2 messages Nov 30, 2004

lore
Share & tools
lore
#0 lore raw

Hi all!

I have implemented a systemcall against linux-2.6.8.1 using the following patches:

--- linux-2.6.8.1-clean/arch/i386/kernel/entry.S 2004-08-14
12:55:09.000000000 +0200
+++ linux-2.6.8.1/arch/i386/kernel/entry.S 2004-11-30 14:48:36.000000000
+0100
@@ -887,4 +887,7 @@
.long sys_mq_getsetattr
.long sys_ni_syscall /* reserved for kexec */

+ .long sys_umb_set_child_restrictions

syscall_table_size=(.-sys_call_table)

--- linux-2.6.8.1-clean/include/asm-i386/unistd.h 2004-08-14
12:55:35.000000000 +0200
+++ linux-2.6.8.1/include/asm-i386/unistd.h 2004-11-30 14:48:36.000000000
+0100
@@ -290,8 +290,13 @@
#define __NR_mq_getsetattr (__NR_mq_open+5)
#define __NR_sys_kexec_load 283

-#define NR_syscalls 284
-
+#define __NR_umb_set_child_restrictions 284
+
+#define NR_syscalls 285

/* user-visible error numbers are in the range -1 - -124: see
<asm-i386/errno.h> */

#define __syscall_return(type, res) \

This works like a charm. If i apply this to linux-2.6.9 it does not work anymore. The syscall is in kallsyms, but nothing happens when I call it.

Any suggestions?

What am I missing?

//snc

----------------
Søren Nøhr Christensen
Computer Science @ AAU
[email protected]

#1 lore raw

* Søren Nøhr Christensen ([email protected]) wrote:
<snip>

Quoted 1 line

+#define __NR_umb_set_child_restrictions 284

<snip>

Quoted 1 line

Any suggestions?

Don't do it this way. Use /proc/<pid>/attr/ interface if you only want to handle creating restrictions within a process. If you're using it to load your policy, then create a reasonable filesystem interface and do it that way. This will be portable across versions and architectures without allocating any syscalls.

thanks,
-chris

Signature · 1 line
Linux Security Modules     http://lsm.immunix.org     http://lsm.bkbits.net