linux/kernel/user_namespace.c

893 lines
23 KiB
C
Raw Normal View History

/*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation, version 2 of the
* License.
*/
#include <linux/export.h>
#include <linux/nsproxy.h>
#include <linux/slab.h>
#include <linux/user_namespace.h>
#include <linux/proc_ns.h>
#include <linux/highuid.h>
User namespaces: set of cleanups (v2) The user_ns is moved from nsproxy to user_struct, so that a struct cred by itself is sufficient to determine access (which it otherwise would not be). Corresponding ecryptfs fixes (by David Howells) are here as well. Fix refcounting. The following rules now apply: 1. The task pins the user struct. 2. The user struct pins its user namespace. 3. The user namespace pins the struct user which created it. User namespaces are cloned during copy_creds(). Unsharing a new user_ns is no longer possible. (We could re-add that, but it'll cause code duplication and doesn't seem useful if PAM doesn't need to clone user namespaces). When a user namespace is created, its first user (uid 0) gets empty keyrings and a clean group_info. This incorporates a previous patch by David Howells. Here is his original patch description: >I suggest adding the attached incremental patch. It makes the following >changes: > > (1) Provides a current_user_ns() macro to wrap accesses to current's user > namespace. > > (2) Fixes eCryptFS. > > (3) Renames create_new_userns() to create_user_ns() to be more consistent > with the other associated functions and because the 'new' in the name is > superfluous. > > (4) Moves the argument and permission checks made for CLONE_NEWUSER to the > beginning of do_fork() so that they're done prior to making any attempts > at allocation. > > (5) Calls create_user_ns() after prepare_creds(), and gives it the new creds > to fill in rather than have it return the new root user. I don't imagine > the new root user being used for anything other than filling in a cred > struct. > > This also permits me to get rid of a get_uid() and a free_uid(), as the > reference the creds were holding on the old user_struct can just be > transferred to the new namespace's creator pointer. > > (6) Makes create_user_ns() reset the UIDs and GIDs of the creds under > preparation rather than doing it in copy_creds(). > >David >Signed-off-by: David Howells <dhowells@redhat.com> Changelog: Oct 20: integrate dhowells comments 1. leave thread_keyring alone 2. use current_user_ns() in set_user() Signed-off-by: Serge Hallyn <serue@us.ibm.com>
2008-10-16 05:38:45 +08:00
#include <linux/cred.h>
#include <linux/securebits.h>
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
#include <linux/keyctl.h>
#include <linux/key-type.h>
#include <keys/user-type.h>
#include <linux/seq_file.h>
#include <linux/fs.h>
#include <linux/uaccess.h>
#include <linux/ctype.h>
#include <linux/projid.h>
#include <linux/fs_struct.h>
static struct kmem_cache *user_ns_cachep __read_mostly;
static bool new_idmap_permitted(const struct file *file,
struct user_namespace *ns, int cap_setid,
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
struct uid_gid_map *map);
static void set_cred_user_ns(struct cred *cred, struct user_namespace *user_ns)
{
/* Start with the same capabilities as init but useless for doing
* anything as the capabilities are bound to the new user namespace.
*/
cred->securebits = SECUREBITS_DEFAULT;
cred->cap_inheritable = CAP_EMPTY_SET;
cred->cap_permitted = CAP_FULL_SET;
cred->cap_effective = CAP_FULL_SET;
cred->cap_bset = CAP_FULL_SET;
#ifdef CONFIG_KEYS
key_put(cred->request_key_auth);
cred->request_key_auth = NULL;
#endif
/* tgcred will be cleared in our caller bc CLONE_THREAD won't be set */
cred->user_ns = user_ns;
}
/*
User namespaces: set of cleanups (v2) The user_ns is moved from nsproxy to user_struct, so that a struct cred by itself is sufficient to determine access (which it otherwise would not be). Corresponding ecryptfs fixes (by David Howells) are here as well. Fix refcounting. The following rules now apply: 1. The task pins the user struct. 2. The user struct pins its user namespace. 3. The user namespace pins the struct user which created it. User namespaces are cloned during copy_creds(). Unsharing a new user_ns is no longer possible. (We could re-add that, but it'll cause code duplication and doesn't seem useful if PAM doesn't need to clone user namespaces). When a user namespace is created, its first user (uid 0) gets empty keyrings and a clean group_info. This incorporates a previous patch by David Howells. Here is his original patch description: >I suggest adding the attached incremental patch. It makes the following >changes: > > (1) Provides a current_user_ns() macro to wrap accesses to current's user > namespace. > > (2) Fixes eCryptFS. > > (3) Renames create_new_userns() to create_user_ns() to be more consistent > with the other associated functions and because the 'new' in the name is > superfluous. > > (4) Moves the argument and permission checks made for CLONE_NEWUSER to the > beginning of do_fork() so that they're done prior to making any attempts > at allocation. > > (5) Calls create_user_ns() after prepare_creds(), and gives it the new creds > to fill in rather than have it return the new root user. I don't imagine > the new root user being used for anything other than filling in a cred > struct. > > This also permits me to get rid of a get_uid() and a free_uid(), as the > reference the creds were holding on the old user_struct can just be > transferred to the new namespace's creator pointer. > > (6) Makes create_user_ns() reset the UIDs and GIDs of the creds under > preparation rather than doing it in copy_creds(). > >David >Signed-off-by: David Howells <dhowells@redhat.com> Changelog: Oct 20: integrate dhowells comments 1. leave thread_keyring alone 2. use current_user_ns() in set_user() Signed-off-by: Serge Hallyn <serue@us.ibm.com>
2008-10-16 05:38:45 +08:00
* Create a new user namespace, deriving the creator from the user in the
* passed credentials, and replacing that user with the new root user for the
* new namespace.
*
* This is called by copy_creds(), which will finish setting the target task's
* credentials.
*/
User namespaces: set of cleanups (v2) The user_ns is moved from nsproxy to user_struct, so that a struct cred by itself is sufficient to determine access (which it otherwise would not be). Corresponding ecryptfs fixes (by David Howells) are here as well. Fix refcounting. The following rules now apply: 1. The task pins the user struct. 2. The user struct pins its user namespace. 3. The user namespace pins the struct user which created it. User namespaces are cloned during copy_creds(). Unsharing a new user_ns is no longer possible. (We could re-add that, but it'll cause code duplication and doesn't seem useful if PAM doesn't need to clone user namespaces). When a user namespace is created, its first user (uid 0) gets empty keyrings and a clean group_info. This incorporates a previous patch by David Howells. Here is his original patch description: >I suggest adding the attached incremental patch. It makes the following >changes: > > (1) Provides a current_user_ns() macro to wrap accesses to current's user > namespace. > > (2) Fixes eCryptFS. > > (3) Renames create_new_userns() to create_user_ns() to be more consistent > with the other associated functions and because the 'new' in the name is > superfluous. > > (4) Moves the argument and permission checks made for CLONE_NEWUSER to the > beginning of do_fork() so that they're done prior to making any attempts > at allocation. > > (5) Calls create_user_ns() after prepare_creds(), and gives it the new creds > to fill in rather than have it return the new root user. I don't imagine > the new root user being used for anything other than filling in a cred > struct. > > This also permits me to get rid of a get_uid() and a free_uid(), as the > reference the creds were holding on the old user_struct can just be > transferred to the new namespace's creator pointer. > > (6) Makes create_user_ns() reset the UIDs and GIDs of the creds under > preparation rather than doing it in copy_creds(). > >David >Signed-off-by: David Howells <dhowells@redhat.com> Changelog: Oct 20: integrate dhowells comments 1. leave thread_keyring alone 2. use current_user_ns() in set_user() Signed-off-by: Serge Hallyn <serue@us.ibm.com>
2008-10-16 05:38:45 +08:00
int create_user_ns(struct cred *new)
{
struct user_namespace *ns, *parent_ns = new->user_ns;
kuid_t owner = new->euid;
kgid_t group = new->egid;
int ret;
/*
* Verify that we can not violate the policy of which files
* may be accessed that is specified by the root directory,
* by verifing that the root directory is at the root of the
* mount namespace which allows all files to be accessed.
*/
if (current_chrooted())
return -EPERM;
/* The creator needs a mapping in the parent user namespace
* or else we won't be able to reasonably tell userspace who
* created a user_namespace.
*/
if (!kuid_has_mapping(parent_ns, owner) ||
!kgid_has_mapping(parent_ns, group))
return -EPERM;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
ns = kmem_cache_zalloc(user_ns_cachep, GFP_KERNEL);
if (!ns)
User namespaces: set of cleanups (v2) The user_ns is moved from nsproxy to user_struct, so that a struct cred by itself is sufficient to determine access (which it otherwise would not be). Corresponding ecryptfs fixes (by David Howells) are here as well. Fix refcounting. The following rules now apply: 1. The task pins the user struct. 2. The user struct pins its user namespace. 3. The user namespace pins the struct user which created it. User namespaces are cloned during copy_creds(). Unsharing a new user_ns is no longer possible. (We could re-add that, but it'll cause code duplication and doesn't seem useful if PAM doesn't need to clone user namespaces). When a user namespace is created, its first user (uid 0) gets empty keyrings and a clean group_info. This incorporates a previous patch by David Howells. Here is his original patch description: >I suggest adding the attached incremental patch. It makes the following >changes: > > (1) Provides a current_user_ns() macro to wrap accesses to current's user > namespace. > > (2) Fixes eCryptFS. > > (3) Renames create_new_userns() to create_user_ns() to be more consistent > with the other associated functions and because the 'new' in the name is > superfluous. > > (4) Moves the argument and permission checks made for CLONE_NEWUSER to the > beginning of do_fork() so that they're done prior to making any attempts > at allocation. > > (5) Calls create_user_ns() after prepare_creds(), and gives it the new creds > to fill in rather than have it return the new root user. I don't imagine > the new root user being used for anything other than filling in a cred > struct. > > This also permits me to get rid of a get_uid() and a free_uid(), as the > reference the creds were holding on the old user_struct can just be > transferred to the new namespace's creator pointer. > > (6) Makes create_user_ns() reset the UIDs and GIDs of the creds under > preparation rather than doing it in copy_creds(). > >David >Signed-off-by: David Howells <dhowells@redhat.com> Changelog: Oct 20: integrate dhowells comments 1. leave thread_keyring alone 2. use current_user_ns() in set_user() Signed-off-by: Serge Hallyn <serue@us.ibm.com>
2008-10-16 05:38:45 +08:00
return -ENOMEM;
ret = proc_alloc_inum(&ns->proc_inum);
if (ret) {
kmem_cache_free(user_ns_cachep, ns);
return ret;
}
atomic_set(&ns->count, 1);
/* Leave the new->user_ns reference with the new user namespace. */
ns->parent = parent_ns;
ns->owner = owner;
ns->group = group;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
set_cred_user_ns(new, ns);
update_mnt_policy(ns);
User namespaces: set of cleanups (v2) The user_ns is moved from nsproxy to user_struct, so that a struct cred by itself is sufficient to determine access (which it otherwise would not be). Corresponding ecryptfs fixes (by David Howells) are here as well. Fix refcounting. The following rules now apply: 1. The task pins the user struct. 2. The user struct pins its user namespace. 3. The user namespace pins the struct user which created it. User namespaces are cloned during copy_creds(). Unsharing a new user_ns is no longer possible. (We could re-add that, but it'll cause code duplication and doesn't seem useful if PAM doesn't need to clone user namespaces). When a user namespace is created, its first user (uid 0) gets empty keyrings and a clean group_info. This incorporates a previous patch by David Howells. Here is his original patch description: >I suggest adding the attached incremental patch. It makes the following >changes: > > (1) Provides a current_user_ns() macro to wrap accesses to current's user > namespace. > > (2) Fixes eCryptFS. > > (3) Renames create_new_userns() to create_user_ns() to be more consistent > with the other associated functions and because the 'new' in the name is > superfluous. > > (4) Moves the argument and permission checks made for CLONE_NEWUSER to the > beginning of do_fork() so that they're done prior to making any attempts > at allocation. > > (5) Calls create_user_ns() after prepare_creds(), and gives it the new creds > to fill in rather than have it return the new root user. I don't imagine > the new root user being used for anything other than filling in a cred > struct. > > This also permits me to get rid of a get_uid() and a free_uid(), as the > reference the creds were holding on the old user_struct can just be > transferred to the new namespace's creator pointer. > > (6) Makes create_user_ns() reset the UIDs and GIDs of the creds under > preparation rather than doing it in copy_creds(). > >David >Signed-off-by: David Howells <dhowells@redhat.com> Changelog: Oct 20: integrate dhowells comments 1. leave thread_keyring alone 2. use current_user_ns() in set_user() Signed-off-by: Serge Hallyn <serue@us.ibm.com>
2008-10-16 05:38:45 +08:00
return 0;
}
int unshare_userns(unsigned long unshare_flags, struct cred **new_cred)
{
struct cred *cred;
if (!(unshare_flags & CLONE_NEWUSER))
return 0;
cred = prepare_creds();
if (!cred)
return -ENOMEM;
*new_cred = cred;
return create_user_ns(cred);
}
void free_user_ns(struct user_namespace *ns)
{
struct user_namespace *parent;
do {
parent = ns->parent;
proc_free_inum(ns->proc_inum);
kmem_cache_free(user_ns_cachep, ns);
ns = parent;
} while (atomic_dec_and_test(&parent->count));
}
EXPORT_SYMBOL(free_user_ns);
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
static u32 map_id_range_down(struct uid_gid_map *map, u32 id, u32 count)
{
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
unsigned idx, extents;
u32 first, last, id2;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
id2 = id + count - 1;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
/* Find the matching extent */
extents = map->nr_extents;
smp_read_barrier_depends();
for (idx = 0; idx < extents; idx++) {
first = map->extent[idx].first;
last = first + map->extent[idx].count - 1;
if (id >= first && id <= last &&
(id2 >= first && id2 <= last))
break;
}
/* Map the id or note failure */
if (idx < extents)
id = (id - first) + map->extent[idx].lower_first;
else
id = (u32) -1;
return id;
}
static u32 map_id_down(struct uid_gid_map *map, u32 id)
{
unsigned idx, extents;
u32 first, last;
/* Find the matching extent */
extents = map->nr_extents;
smp_read_barrier_depends();
for (idx = 0; idx < extents; idx++) {
first = map->extent[idx].first;
last = first + map->extent[idx].count - 1;
if (id >= first && id <= last)
break;
}
/* Map the id or note failure */
if (idx < extents)
id = (id - first) + map->extent[idx].lower_first;
else
id = (u32) -1;
return id;
}
static u32 map_id_up(struct uid_gid_map *map, u32 id)
{
unsigned idx, extents;
u32 first, last;
/* Find the matching extent */
extents = map->nr_extents;
smp_read_barrier_depends();
for (idx = 0; idx < extents; idx++) {
first = map->extent[idx].lower_first;
last = first + map->extent[idx].count - 1;
if (id >= first && id <= last)
break;
}
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
/* Map the id or note failure */
if (idx < extents)
id = (id - first) + map->extent[idx].first;
else
id = (u32) -1;
return id;
}
/**
* make_kuid - Map a user-namespace uid pair into a kuid.
* @ns: User namespace that the uid is in
* @uid: User identifier
*
* Maps a user-namespace uid pair into a kernel internal kuid,
* and returns that kuid.
*
* When there is no mapping defined for the user-namespace uid
* pair INVALID_UID is returned. Callers are expected to test
* for and handle handle INVALID_UID being returned. INVALID_UID
* may be tested for using uid_valid().
*/
kuid_t make_kuid(struct user_namespace *ns, uid_t uid)
{
/* Map the uid to a global kernel uid */
return KUIDT_INIT(map_id_down(&ns->uid_map, uid));
}
EXPORT_SYMBOL(make_kuid);
/**
* from_kuid - Create a uid from a kuid user-namespace pair.
* @targ: The user namespace we want a uid in.
* @kuid: The kernel internal uid to start with.
*
* Map @kuid into the user-namespace specified by @targ and
* return the resulting uid.
*
* There is always a mapping into the initial user_namespace.
*
* If @kuid has no mapping in @targ (uid_t)-1 is returned.
*/
uid_t from_kuid(struct user_namespace *targ, kuid_t kuid)
{
/* Map the uid from a global kernel uid */
return map_id_up(&targ->uid_map, __kuid_val(kuid));
}
EXPORT_SYMBOL(from_kuid);
/**
* from_kuid_munged - Create a uid from a kuid user-namespace pair.
* @targ: The user namespace we want a uid in.
* @kuid: The kernel internal uid to start with.
*
* Map @kuid into the user-namespace specified by @targ and
* return the resulting uid.
*
* There is always a mapping into the initial user_namespace.
*
* Unlike from_kuid from_kuid_munged never fails and always
* returns a valid uid. This makes from_kuid_munged appropriate
* for use in syscalls like stat and getuid where failing the
* system call and failing to provide a valid uid are not an
* options.
*
* If @kuid has no mapping in @targ overflowuid is returned.
*/
uid_t from_kuid_munged(struct user_namespace *targ, kuid_t kuid)
{
uid_t uid;
uid = from_kuid(targ, kuid);
if (uid == (uid_t) -1)
uid = overflowuid;
return uid;
}
EXPORT_SYMBOL(from_kuid_munged);
/**
* make_kgid - Map a user-namespace gid pair into a kgid.
* @ns: User namespace that the gid is in
* @uid: group identifier
*
* Maps a user-namespace gid pair into a kernel internal kgid,
* and returns that kgid.
*
* When there is no mapping defined for the user-namespace gid
* pair INVALID_GID is returned. Callers are expected to test
* for and handle INVALID_GID being returned. INVALID_GID may be
* tested for using gid_valid().
*/
kgid_t make_kgid(struct user_namespace *ns, gid_t gid)
{
/* Map the gid to a global kernel gid */
return KGIDT_INIT(map_id_down(&ns->gid_map, gid));
}
EXPORT_SYMBOL(make_kgid);
/**
* from_kgid - Create a gid from a kgid user-namespace pair.
* @targ: The user namespace we want a gid in.
* @kgid: The kernel internal gid to start with.
*
* Map @kgid into the user-namespace specified by @targ and
* return the resulting gid.
*
* There is always a mapping into the initial user_namespace.
*
* If @kgid has no mapping in @targ (gid_t)-1 is returned.
*/
gid_t from_kgid(struct user_namespace *targ, kgid_t kgid)
{
/* Map the gid from a global kernel gid */
return map_id_up(&targ->gid_map, __kgid_val(kgid));
}
EXPORT_SYMBOL(from_kgid);
/**
* from_kgid_munged - Create a gid from a kgid user-namespace pair.
* @targ: The user namespace we want a gid in.
* @kgid: The kernel internal gid to start with.
*
* Map @kgid into the user-namespace specified by @targ and
* return the resulting gid.
*
* There is always a mapping into the initial user_namespace.
*
* Unlike from_kgid from_kgid_munged never fails and always
* returns a valid gid. This makes from_kgid_munged appropriate
* for use in syscalls like stat and getgid where failing the
* system call and failing to provide a valid gid are not options.
*
* If @kgid has no mapping in @targ overflowgid is returned.
*/
gid_t from_kgid_munged(struct user_namespace *targ, kgid_t kgid)
{
gid_t gid;
gid = from_kgid(targ, kgid);
if (gid == (gid_t) -1)
gid = overflowgid;
return gid;
}
EXPORT_SYMBOL(from_kgid_munged);
/**
* make_kprojid - Map a user-namespace projid pair into a kprojid.
* @ns: User namespace that the projid is in
* @projid: Project identifier
*
* Maps a user-namespace uid pair into a kernel internal kuid,
* and returns that kuid.
*
* When there is no mapping defined for the user-namespace projid
* pair INVALID_PROJID is returned. Callers are expected to test
* for and handle handle INVALID_PROJID being returned. INVALID_PROJID
* may be tested for using projid_valid().
*/
kprojid_t make_kprojid(struct user_namespace *ns, projid_t projid)
{
/* Map the uid to a global kernel uid */
return KPROJIDT_INIT(map_id_down(&ns->projid_map, projid));
}
EXPORT_SYMBOL(make_kprojid);
/**
* from_kprojid - Create a projid from a kprojid user-namespace pair.
* @targ: The user namespace we want a projid in.
* @kprojid: The kernel internal project identifier to start with.
*
* Map @kprojid into the user-namespace specified by @targ and
* return the resulting projid.
*
* There is always a mapping into the initial user_namespace.
*
* If @kprojid has no mapping in @targ (projid_t)-1 is returned.
*/
projid_t from_kprojid(struct user_namespace *targ, kprojid_t kprojid)
{
/* Map the uid from a global kernel uid */
return map_id_up(&targ->projid_map, __kprojid_val(kprojid));
}
EXPORT_SYMBOL(from_kprojid);
/**
* from_kprojid_munged - Create a projiid from a kprojid user-namespace pair.
* @targ: The user namespace we want a projid in.
* @kprojid: The kernel internal projid to start with.
*
* Map @kprojid into the user-namespace specified by @targ and
* return the resulting projid.
*
* There is always a mapping into the initial user_namespace.
*
* Unlike from_kprojid from_kprojid_munged never fails and always
* returns a valid projid. This makes from_kprojid_munged
* appropriate for use in syscalls like stat and where
* failing the system call and failing to provide a valid projid are
* not an options.
*
* If @kprojid has no mapping in @targ OVERFLOW_PROJID is returned.
*/
projid_t from_kprojid_munged(struct user_namespace *targ, kprojid_t kprojid)
{
projid_t projid;
projid = from_kprojid(targ, kprojid);
if (projid == (projid_t) -1)
projid = OVERFLOW_PROJID;
return projid;
}
EXPORT_SYMBOL(from_kprojid_munged);
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
static int uid_m_show(struct seq_file *seq, void *v)
{
struct user_namespace *ns = seq->private;
struct uid_gid_extent *extent = v;
struct user_namespace *lower_ns;
uid_t lower;
lower_ns = seq_user_ns(seq);
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
if ((lower_ns == ns) && lower_ns->parent)
lower_ns = lower_ns->parent;
lower = from_kuid(lower_ns, KUIDT_INIT(extent->lower_first));
seq_printf(seq, "%10u %10u %10u\n",
extent->first,
lower,
extent->count);
return 0;
}
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
static int gid_m_show(struct seq_file *seq, void *v)
{
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
struct user_namespace *ns = seq->private;
struct uid_gid_extent *extent = v;
struct user_namespace *lower_ns;
gid_t lower;
lower_ns = seq_user_ns(seq);
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
if ((lower_ns == ns) && lower_ns->parent)
lower_ns = lower_ns->parent;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
lower = from_kgid(lower_ns, KGIDT_INIT(extent->lower_first));
seq_printf(seq, "%10u %10u %10u\n",
extent->first,
lower,
extent->count);
return 0;
}
static int projid_m_show(struct seq_file *seq, void *v)
{
struct user_namespace *ns = seq->private;
struct uid_gid_extent *extent = v;
struct user_namespace *lower_ns;
projid_t lower;
lower_ns = seq_user_ns(seq);
if ((lower_ns == ns) && lower_ns->parent)
lower_ns = lower_ns->parent;
lower = from_kprojid(lower_ns, KPROJIDT_INIT(extent->lower_first));
seq_printf(seq, "%10u %10u %10u\n",
extent->first,
lower,
extent->count);
return 0;
}
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
static void *m_start(struct seq_file *seq, loff_t *ppos, struct uid_gid_map *map)
{
struct uid_gid_extent *extent = NULL;
loff_t pos = *ppos;
if (pos < map->nr_extents)
extent = &map->extent[pos];
return extent;
}
static void *uid_m_start(struct seq_file *seq, loff_t *ppos)
{
struct user_namespace *ns = seq->private;
return m_start(seq, ppos, &ns->uid_map);
}
static void *gid_m_start(struct seq_file *seq, loff_t *ppos)
{
struct user_namespace *ns = seq->private;
return m_start(seq, ppos, &ns->gid_map);
}
static void *projid_m_start(struct seq_file *seq, loff_t *ppos)
{
struct user_namespace *ns = seq->private;
return m_start(seq, ppos, &ns->projid_map);
}
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
static void *m_next(struct seq_file *seq, void *v, loff_t *pos)
{
(*pos)++;
return seq->op->start(seq, pos);
}
static void m_stop(struct seq_file *seq, void *v)
{
return;
}
struct seq_operations proc_uid_seq_operations = {
.start = uid_m_start,
.stop = m_stop,
.next = m_next,
.show = uid_m_show,
};
struct seq_operations proc_gid_seq_operations = {
.start = gid_m_start,
.stop = m_stop,
.next = m_next,
.show = gid_m_show,
};
struct seq_operations proc_projid_seq_operations = {
.start = projid_m_start,
.stop = m_stop,
.next = m_next,
.show = projid_m_show,
};
static bool mappings_overlap(struct uid_gid_map *new_map, struct uid_gid_extent *extent)
{
u32 upper_first, lower_first, upper_last, lower_last;
unsigned idx;
upper_first = extent->first;
lower_first = extent->lower_first;
upper_last = upper_first + extent->count - 1;
lower_last = lower_first + extent->count - 1;
for (idx = 0; idx < new_map->nr_extents; idx++) {
u32 prev_upper_first, prev_lower_first;
u32 prev_upper_last, prev_lower_last;
struct uid_gid_extent *prev;
prev = &new_map->extent[idx];
prev_upper_first = prev->first;
prev_lower_first = prev->lower_first;
prev_upper_last = prev_upper_first + prev->count - 1;
prev_lower_last = prev_lower_first + prev->count - 1;
/* Does the upper range intersect a previous extent? */
if ((prev_upper_first <= upper_last) &&
(prev_upper_last >= upper_first))
return true;
/* Does the lower range intersect a previous extent? */
if ((prev_lower_first <= lower_last) &&
(prev_lower_last >= lower_first))
return true;
}
return false;
}
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
static DEFINE_MUTEX(id_map_mutex);
static ssize_t map_write(struct file *file, const char __user *buf,
size_t count, loff_t *ppos,
int cap_setid,
struct uid_gid_map *map,
struct uid_gid_map *parent_map)
{
struct seq_file *seq = file->private_data;
struct user_namespace *ns = seq->private;
struct uid_gid_map new_map;
unsigned idx;
struct uid_gid_extent *extent = NULL;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
unsigned long page = 0;
char *kbuf, *pos, *next_line;
ssize_t ret = -EINVAL;
/*
* The id_map_mutex serializes all writes to any given map.
*
* Any map is only ever written once.
*
* An id map fits within 1 cache line on most architectures.
*
* On read nothing needs to be done unless you are on an
* architecture with a crazy cache coherency model like alpha.
*
* There is a one time data dependency between reading the
* count of the extents and the values of the extents. The
* desired behavior is to see the values of the extents that
* were written before the count of the extents.
*
* To achieve this smp_wmb() is used on guarantee the write
* order and smp_read_barrier_depends() is guaranteed that we
* don't have crazy architectures returning stale data.
*
*/
mutex_lock(&id_map_mutex);
ret = -EPERM;
/* Only allow one successful write to the map */
if (map->nr_extents != 0)
goto out;
/*
* Adjusting namespace settings requires capabilities on the target.
*/
if (cap_valid(cap_setid) && !file_ns_capable(file, ns, CAP_SYS_ADMIN))
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
goto out;
/* Get a buffer */
ret = -ENOMEM;
page = __get_free_page(GFP_TEMPORARY);
kbuf = (char *) page;
if (!page)
goto out;
/* Only allow <= page size writes at the beginning of the file */
ret = -EINVAL;
if ((*ppos != 0) || (count >= PAGE_SIZE))
goto out;
/* Slurp in the user data */
ret = -EFAULT;
if (copy_from_user(kbuf, buf, count))
goto out;
kbuf[count] = '\0';
/* Parse the user data */
ret = -EINVAL;
pos = kbuf;
new_map.nr_extents = 0;
for (;pos; pos = next_line) {
extent = &new_map.extent[new_map.nr_extents];
/* Find the end of line and ensure I don't look past it */
next_line = strchr(pos, '\n');
if (next_line) {
*next_line = '\0';
next_line++;
if (*next_line == '\0')
next_line = NULL;
}
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
pos = skip_spaces(pos);
extent->first = simple_strtoul(pos, &pos, 10);
if (!isspace(*pos))
goto out;
pos = skip_spaces(pos);
extent->lower_first = simple_strtoul(pos, &pos, 10);
if (!isspace(*pos))
goto out;
pos = skip_spaces(pos);
extent->count = simple_strtoul(pos, &pos, 10);
if (*pos && !isspace(*pos))
goto out;
/* Verify there is not trailing junk on the line */
pos = skip_spaces(pos);
if (*pos != '\0')
goto out;
/* Verify we have been given valid starting values */
if ((extent->first == (u32) -1) ||
(extent->lower_first == (u32) -1 ))
goto out;
/* Verify count is not zero and does not cause the extent to wrap */
if ((extent->first + extent->count) <= extent->first)
goto out;
if ((extent->lower_first + extent->count) <= extent->lower_first)
goto out;
/* Do the ranges in extent overlap any previous extents? */
if (mappings_overlap(&new_map, extent))
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
goto out;
new_map.nr_extents++;
/* Fail if the file contains too many extents */
if ((new_map.nr_extents == UID_GID_MAP_MAX_EXTENTS) &&
(next_line != NULL))
goto out;
}
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
/* Be very certaint the new map actually exists */
if (new_map.nr_extents == 0)
goto out;
ret = -EPERM;
/* Validate the user is allowed to use user id's mapped to. */
if (!new_idmap_permitted(file, ns, cap_setid, &new_map))
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
goto out;
/* Map the lower ids from the parent user namespace to the
* kernel global id space.
*/
for (idx = 0; idx < new_map.nr_extents; idx++) {
u32 lower_first;
extent = &new_map.extent[idx];
lower_first = map_id_range_down(parent_map,
extent->lower_first,
extent->count);
/* Fail if we can not map the specified extent to
* the kernel global id space.
*/
if (lower_first == (u32) -1)
goto out;
extent->lower_first = lower_first;
}
/* Install the map */
memcpy(map->extent, new_map.extent,
new_map.nr_extents*sizeof(new_map.extent[0]));
smp_wmb();
map->nr_extents = new_map.nr_extents;
*ppos = count;
ret = count;
out:
mutex_unlock(&id_map_mutex);
if (page)
free_page(page);
return ret;
}
ssize_t proc_uid_map_write(struct file *file, const char __user *buf, size_t size, loff_t *ppos)
{
struct seq_file *seq = file->private_data;
struct user_namespace *ns = seq->private;
struct user_namespace *seq_ns = seq_user_ns(seq);
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
if (!ns->parent)
return -EPERM;
if ((seq_ns != ns) && (seq_ns != ns->parent))
return -EPERM;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
return map_write(file, buf, size, ppos, CAP_SETUID,
&ns->uid_map, &ns->parent->uid_map);
}
ssize_t proc_gid_map_write(struct file *file, const char __user *buf, size_t size, loff_t *ppos)
{
struct seq_file *seq = file->private_data;
struct user_namespace *ns = seq->private;
struct user_namespace *seq_ns = seq_user_ns(seq);
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
if (!ns->parent)
return -EPERM;
if ((seq_ns != ns) && (seq_ns != ns->parent))
return -EPERM;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
return map_write(file, buf, size, ppos, CAP_SETGID,
&ns->gid_map, &ns->parent->gid_map);
}
ssize_t proc_projid_map_write(struct file *file, const char __user *buf, size_t size, loff_t *ppos)
{
struct seq_file *seq = file->private_data;
struct user_namespace *ns = seq->private;
struct user_namespace *seq_ns = seq_user_ns(seq);
if (!ns->parent)
return -EPERM;
if ((seq_ns != ns) && (seq_ns != ns->parent))
return -EPERM;
/* Anyone can set any valid project id no capability needed */
return map_write(file, buf, size, ppos, -1,
&ns->projid_map, &ns->parent->projid_map);
}
static bool new_idmap_permitted(const struct file *file,
struct user_namespace *ns, int cap_setid,
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
struct uid_gid_map *new_map)
{
/* Allow mapping to your own filesystem ids */
if ((new_map->nr_extents == 1) && (new_map->extent[0].count == 1)) {
u32 id = new_map->extent[0].lower_first;
if (cap_setid == CAP_SETUID) {
kuid_t uid = make_kuid(ns->parent, id);
if (uid_eq(uid, file->f_cred->fsuid))
return true;
}
else if (cap_setid == CAP_SETGID) {
kgid_t gid = make_kgid(ns->parent, id);
if (gid_eq(gid, file->f_cred->fsgid))
return true;
}
}
/* Allow anyone to set a mapping that doesn't require privilege */
if (!cap_valid(cap_setid))
return true;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
/* Allow the specified ids if we have the appropriate capability
* (CAP_SETUID or CAP_SETGID) over the parent user namespace.
* And the opener of the id file also had the approprpiate capability.
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
*/
if (ns_capable(ns->parent, cap_setid) &&
file_ns_capable(file, ns->parent, cap_setid))
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
return true;
userns: Rework the user_namespace adding uid/gid mapping support - Convert the old uid mapping functions into compatibility wrappers - Add a uid/gid mapping layer from user space uid and gids to kernel internal uids and gids that is extent based for simplicty and speed. * Working with number space after mapping uids/gids into their kernel internal version adds only mapping complexity over what we have today, leaving the kernel code easy to understand and test. - Add proc files /proc/self/uid_map /proc/self/gid_map These files display the mapping and allow a mapping to be added if a mapping does not exist. - Allow entering the user namespace without a uid or gid mapping. Since we are starting with an existing user our uids and gids still have global mappings so are still valid and useful they just don't have local mappings. The requirement for things to work are global uid and gid so it is odd but perfectly fine not to have a local uid and gid mapping. Not requiring global uid and gid mappings greatly simplifies the logic of setting up the uid and gid mappings by allowing the mappings to be set after the namespace is created which makes the slight weirdness worth it. - Make the mappings in the initial user namespace to the global uid/gid space explicit. Today it is an identity mapping but in the future we may want to twist this for debugging, similar to what we do with jiffies. - Document the memory ordering requirements of setting the uid and gid mappings. We only allow the mappings to be set once and there are no pointers involved so the requirments are trivial but a little atypical. Performance: In this scheme for the permission checks the performance is expected to stay the same as the actuall machine instructions should remain the same. The worst case I could think of is ls -l on a large directory where all of the stat results need to be translated with from kuids and kgids to uids and gids. So I benchmarked that case on my laptop with a dual core hyperthread Intel i5-2520M cpu with 3M of cpu cache. My benchmark consisted of going to single user mode where nothing else was running. On an ext4 filesystem opening 1,000,000 files and looping through all of the files 1000 times and calling fstat on the individuals files. This was to ensure I was benchmarking stat times where the inodes were in the kernels cache, but the inode values were not in the processors cache. My results: v3.4-rc1: ~= 156ns (unmodified v3.4-rc1 with user namespace support disabled) v3.4-rc1-userns-: ~= 155ns (v3.4-rc1 with my user namespace patches and user namespace support disabled) v3.4-rc1-userns+: ~= 164ns (v3.4-rc1 with my user namespace patches and user namespace support enabled) All of the configurations ran in roughly 120ns when I performed tests that ran in the cpu cache. So in summary the performance impact is: 1ns improvement in the worst case with user namespace support compiled out. 8ns aka 5% slowdown in the worst case with user namespace support compiled in. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
2011-11-17 16:11:58 +08:00
return false;
}
static void *userns_get(struct task_struct *task)
{
struct user_namespace *user_ns;
rcu_read_lock();
user_ns = get_user_ns(__task_cred(task)->user_ns);
rcu_read_unlock();
return user_ns;
}
static void userns_put(void *ns)
{
put_user_ns(ns);
}
static int userns_install(struct nsproxy *nsproxy, void *ns)
{
struct user_namespace *user_ns = ns;
struct cred *cred;
/* Don't allow gaining capabilities by reentering
* the same user namespace.
*/
if (user_ns == current_user_ns())
return -EINVAL;
/* Threaded processes may not enter a different user namespace */
if (atomic_read(&current->mm->mm_users) > 1)
return -EINVAL;
if (current->fs->users != 1)
return -EINVAL;
if (!ns_capable(user_ns, CAP_SYS_ADMIN))
return -EPERM;
cred = prepare_creds();
if (!cred)
return -ENOMEM;
put_user_ns(cred->user_ns);
set_cred_user_ns(cred, get_user_ns(user_ns));
return commit_creds(cred);
}
static unsigned int userns_inum(void *ns)
{
struct user_namespace *user_ns = ns;
return user_ns->proc_inum;
}
const struct proc_ns_operations userns_operations = {
.name = "user",
.type = CLONE_NEWUSER,
.get = userns_get,
.put = userns_put,
.install = userns_install,
.inum = userns_inum,
};
static __init int user_namespaces_init(void)
{
user_ns_cachep = KMEM_CACHE(user_namespace, SLAB_PANIC);
return 0;
}
module_init(user_namespaces_init);