[redhat-lspp] [PATCH 1/2] New system call unshare
Janak Desai
janak at us.ibm.com
Thu Oct 20 15:00:39 UTC 2005
Hi all, I am still working with Chris Wright on investigating
is clone() can be used in PAM, but in the meantime Steve Grubb
suggested that it would be a good idea to start including
unshare patches in the test kernel that is being built with
audit patches. This patch incorporates feedback that was
provided by Chris when it was posted on lkml about a week
ago.
The first patch implements the system call service routine
and the second patch register the system call number for i386.
-Janak
----------------------------------------------------------------
diff -Naurp 2.6.14-rc4-mm1/fs/namespace.c 2.6.14-rc4-mm1+unshare/fs/namespace.c
--- 2.6.14-rc4-mm1/fs/namespace.c 2005-10-17 18:07:40.000000000 +0000
+++ 2.6.14-rc4-mm1+unshare/fs/namespace.c 2005-10-20 13:47:08.000000000 +0000
@@ -1069,9 +1069,7 @@ int copy_namespace(int flags, struct tas
{
struct namespace *namespace = tsk->namespace;
struct namespace *new_ns;
- struct vfsmount *rootmnt = NULL, *pwdmnt = NULL, *altrootmnt = NULL;
- struct fs_struct *fs = tsk->fs;
- struct vfsmount *p, *q;
+ int err = 0;
if (!namespace)
return 0;
@@ -1082,10 +1080,35 @@ int copy_namespace(int flags, struct tas
return 0;
if (!capable(CAP_SYS_ADMIN)) {
- put_namespace(namespace);
- return -EPERM;
+ err = -EPERM;
+ goto out;
+ }
+
+ new_ns = dup_namespace(tsk);
+ if (!new_ns) {
+ err = -ENOMEM;
+ goto out;
}
+ tsk->namespace = new_ns;
+
+out:
+ put_namespace(namespace);
+ return err;
+}
+
+/*
+ * Allocate a new namespace structure and populate it with contents
+ * copied from the namespace of the passed in task structure.
+ */
+struct namespace *dup_namespace(struct task_struct *tsk)
+{
+ struct namespace *namespace = tsk->namespace;
+ struct namespace *new_ns;
+ struct vfsmount *rootmnt = NULL, *pwdmnt = NULL, *altrootmnt = NULL;
+ struct fs_struct *fs = tsk->fs;
+ struct vfsmount *p, *q;
+
new_ns = kmalloc(sizeof(struct namespace), GFP_KERNEL);
if (!new_ns)
goto out;
@@ -1134,8 +1157,6 @@ int copy_namespace(int flags, struct tas
}
up_write(&tsk->namespace->sem);
- tsk->namespace = new_ns;
-
if (rootmnt)
mntput(rootmnt);
if (pwdmnt)
@@ -1143,12 +1164,8 @@ int copy_namespace(int flags, struct tas
if (altrootmnt)
mntput(altrootmnt);
- put_namespace(namespace);
- return 0;
-
out:
- put_namespace(namespace);
- return -ENOMEM;
+ return new_ns;
}
asmlinkage long sys_mount(char __user * dev_name, char __user * dir_name,
diff -Naurp 2.6.14-rc4-mm1/include/linux/namespace.h 2.6.14-rc4-mm1+unshare/include/linux/namespace.h
--- 2.6.14-rc4-mm1/include/linux/namespace.h 2005-08-28 23:41:01.000000000 +0000
+++ 2.6.14-rc4-mm1+unshare/include/linux/namespace.h 2005-10-20 13:48:15.000000000 +0000
@@ -14,6 +14,7 @@ struct namespace {
extern int copy_namespace(int, struct task_struct *);
extern void __put_namespace(struct namespace *namespace);
+extern struct namespace *dup_namespace(struct task_struct *);
static inline void put_namespace(struct namespace *namespace)
{
diff -Naurp 2.6.14-rc4-mm1/kernel/fork.c 2.6.14-rc4-mm1+unshare/kernel/fork.c
--- 2.6.14-rc4-mm1/kernel/fork.c 2005-10-17 18:09:00.000000000 +0000
+++ 2.6.14-rc4-mm1+unshare/kernel/fork.c 2005-10-20 12:40:06.000000000 +0000
@@ -446,6 +446,55 @@ void mm_release(struct task_struct *tsk,
}
}
+/*
+ * Allocate a new mm structure and copy contents from the
+ * mm structure of the passed in task structure.
+ */
+static struct mm_struct *dup_mm(struct task_struct *tsk)
+{
+ struct mm_struct *mm, *oldmm = current->mm;
+ int err;
+
+ if (!oldmm)
+ return NULL;
+
+ mm = allocate_mm();
+ if (!mm)
+ goto fail_nomem;
+
+ memcpy(mm, oldmm, sizeof(*mm));
+
+ if (!mm_init(mm))
+ goto fail_nomem;
+
+ if (init_new_context(tsk, mm))
+ goto fail_nocontext;
+
+ err = dup_mmap(mm, oldmm);
+ if (err)
+ goto free_pt;
+
+ mm->hiwater_rss = get_mm_rss(mm);
+ mm->hiwater_vm = mm->total_vm;
+
+ return mm;
+
+free_pt:
+ mmput(mm);
+
+fail_nomem:
+ return NULL;
+
+fail_nocontext:
+ /*
+ * If init_new_context() failed, we cannot use mmput() to free the mm
+ * because it calls destroy_context()
+ */
+ mm_free_pgd(mm);
+ free_mm(mm);
+ return NULL;
+}
+
static int copy_mm(unsigned long clone_flags, struct task_struct * tsk)
{
struct mm_struct * mm, *oldmm;
@@ -480,43 +529,17 @@ static int copy_mm(unsigned long clone_f
}
retval = -ENOMEM;
- mm = allocate_mm();
+ mm = dup_mm(tsk);
if (!mm)
goto fail_nomem;
- /* Copy the current MM stuff.. */
- memcpy(mm, oldmm, sizeof(*mm));
- if (!mm_init(mm))
- goto fail_nomem;
-
- if (init_new_context(tsk,mm))
- goto fail_nocontext;
-
- retval = dup_mmap(mm, oldmm);
- if (retval)
- goto free_pt;
-
- mm->hiwater_rss = get_mm_rss(mm);
- mm->hiwater_vm = mm->total_vm;
-
good_mm:
tsk->mm = mm;
tsk->active_mm = mm;
return 0;
-free_pt:
- mmput(mm);
fail_nomem:
return retval;
-
-fail_nocontext:
- /*
- * If init_new_context() failed, we cannot use mmput() to free the mm
- * because it calls destroy_context()
- */
- mm_free_pgd(mm);
- free_mm(mm);
- return retval;
}
static inline struct fs_struct *__copy_fs_struct(struct fs_struct *old)
@@ -1319,3 +1342,101 @@ void __init proc_caches_init(void)
sizeof(struct mm_struct), 0,
SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL, NULL);
}
+
+/*
+ * Performs sanity checks on the flags passed to the unshare system
+ * call.
+ */
+static inline int check_unshare_flags(unsigned long unshare_flags)
+{
+ int err = -EINVAL;
+
+ if (unshare_flags & ~(CLONE_NEWNS | CLONE_VM))
+ goto errout;
+
+ /*
+ * Cannot unshare namespace if the fs structure is being shared
+ * through a previous call to clone()
+ */
+ if ((unshare_flags & CLONE_NEWNS) &&
+ (atomic_read(¤t->fs->count) > 1))
+ goto errout;
+
+ /*
+ * Cannot unshare vm if sighnal handlers are being shared through
+ * a previous call to clone()
+ */
+ if ((unshare_flags & CLONE_VM) &&
+ (atomic_read(¤t->sighand->count) > 1))
+ goto errout;
+
+ return 0;
+
+errout:
+ return err;
+
+}
+
+/*
+ * unshare allows a process to 'unshare' part of the process
+ * context which was originally shared using clone. copy_*
+ * functions used by do_fork() cannot be used here directly
+ * because they modify an inactive task_struct that is being
+ * constructed. Here we are modifying the current, active,
+ * task_struct.
+ */
+asmlinkage long sys_unshare(unsigned long unshare_flags)
+{
+ int err = 0;
+ struct namespace *new_ns = NULL, *ns = current->namespace;
+ struct mm_struct *new_mm = NULL, *active_mm = NULL, *mm = current->mm;
+
+ err = check_unshare_flags(unshare_flags);
+ if (err)
+ goto unshare_out;
+
+ if ((unshare_flags & CLONE_NEWNS) &&
+ (ns && atomic_read(&ns->count) > 1)) {
+ err = -EPERM;
+ if (!capable(CAP_SYS_ADMIN))
+ goto unshare_out;
+
+ err = -ENOMEM;
+ new_ns = dup_namespace(current);
+ if (!new_ns)
+ goto unshare_out;
+ }
+
+ if ((unshare_flags & CLONE_VM) && (atomic_read(&mm->mm_users) > 1)) {
+ err = -ENOMEM;
+ new_mm = dup_mm(current);
+ if (!new_mm)
+ goto unshare_cleanup_ns;
+ }
+
+ if (new_ns) {
+ task_lock(current);
+ current->namespace = new_ns;
+ task_unlock(current);
+ put_namespace(ns);
+ }
+
+ if (new_mm) {
+ task_lock(current);
+ active_mm = current->active_mm;
+ current->mm = new_mm;
+ current->active_mm = new_mm;
+ activate_mm(active_mm, new_mm);
+ task_unlock(current);
+ mmput(mm);
+ }
+
+ return 0;
+
+unshare_cleanup_ns:
+ if (new_ns)
+ put_namespace(new_ns);
+
+unshare_out:
+ return err;
+}
More information about the redhat-lspp
mailing list