2020-05-31 16:28:37 +08:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
|
|
|
|
|
|
|
|
|
#include <linux/bpf.h>
|
|
|
|
|
#include <linux/filter.h>
|
|
|
|
|
#include <net/net_namespace.h>
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Functions to manage BPF programs attached to netns
|
|
|
|
|
*/
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
struct bpf_netns_link {
|
|
|
|
|
struct bpf_link link;
|
|
|
|
|
enum bpf_attach_type type;
|
|
|
|
|
enum netns_bpf_attach_type netns_type;
|
|
|
|
|
|
|
|
|
|
/* We don't hold a ref to net in order to auto-detach the link
|
|
|
|
|
* when netns is going away. Instead we rely on pernet
|
|
|
|
|
* pre_exit callback to clear this pointer. Must be accessed
|
|
|
|
|
* with netns_bpf_mutex held.
|
|
|
|
|
*/
|
|
|
|
|
struct net *net;
|
2020-06-25 22:13:56 +08:00
|
|
|
struct list_head node; /* node in list of links attached to net */
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
};
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
/* Protects updates to netns_bpf */
|
|
|
|
|
DEFINE_MUTEX(netns_bpf_mutex);
|
|
|
|
|
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
static void netns_bpf_attach_type_unneed(enum netns_bpf_attach_type type)
|
|
|
|
|
{
|
|
|
|
|
switch (type) {
|
2020-07-21 18:07:16 +08:00
|
|
|
#ifdef CONFIG_INET
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
case NETNS_BPF_SK_LOOKUP:
|
|
|
|
|
static_branch_dec(&bpf_sk_lookup_enabled);
|
|
|
|
|
break;
|
2020-07-21 18:07:16 +08:00
|
|
|
#endif
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void netns_bpf_attach_type_need(enum netns_bpf_attach_type type)
|
|
|
|
|
{
|
|
|
|
|
switch (type) {
|
2020-07-21 18:07:16 +08:00
|
|
|
#ifdef CONFIG_INET
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
case NETNS_BPF_SK_LOOKUP:
|
|
|
|
|
static_branch_inc(&bpf_sk_lookup_enabled);
|
|
|
|
|
break;
|
2020-07-21 18:07:16 +08:00
|
|
|
#endif
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2020-06-25 22:13:55 +08:00
|
|
|
/* Must be called with netns_bpf_mutex held. */
|
|
|
|
|
static void netns_bpf_run_array_detach(struct net *net,
|
|
|
|
|
enum netns_bpf_attach_type type)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_prog_array *run_array;
|
|
|
|
|
|
|
|
|
|
run_array = rcu_replace_pointer(net->bpf.run_array[type], NULL,
|
|
|
|
|
lockdep_is_held(&netns_bpf_mutex));
|
|
|
|
|
bpf_prog_array_free(run_array);
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-17 18:35:22 +08:00
|
|
|
static int link_index(struct net *net, enum netns_bpf_attach_type type,
|
|
|
|
|
struct bpf_netns_link *link)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_netns_link *pos;
|
|
|
|
|
int i = 0;
|
|
|
|
|
|
|
|
|
|
list_for_each_entry(pos, &net->bpf.links[type], node) {
|
|
|
|
|
if (pos == link)
|
|
|
|
|
return i;
|
|
|
|
|
i++;
|
|
|
|
|
}
|
|
|
|
|
return -ENOENT;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int link_count(struct net *net, enum netns_bpf_attach_type type)
|
|
|
|
|
{
|
|
|
|
|
struct list_head *pos;
|
|
|
|
|
int i = 0;
|
|
|
|
|
|
|
|
|
|
list_for_each(pos, &net->bpf.links[type])
|
|
|
|
|
i++;
|
|
|
|
|
return i;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void fill_prog_array(struct net *net, enum netns_bpf_attach_type type,
|
|
|
|
|
struct bpf_prog_array *prog_array)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_netns_link *pos;
|
|
|
|
|
unsigned int i = 0;
|
|
|
|
|
|
|
|
|
|
list_for_each_entry(pos, &net->bpf.links[type], node) {
|
|
|
|
|
prog_array->items[i].prog = pos->link.prog;
|
|
|
|
|
i++;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
static void bpf_netns_link_release(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_netns_link *net_link =
|
|
|
|
|
container_of(link, struct bpf_netns_link, link);
|
|
|
|
|
enum netns_bpf_attach_type type = net_link->netns_type;
|
2020-07-17 18:35:22 +08:00
|
|
|
struct bpf_prog_array *old_array, *new_array;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
struct net *net;
|
2020-07-17 18:35:22 +08:00
|
|
|
int cnt, idx;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
|
|
|
|
|
mutex_lock(&netns_bpf_mutex);
|
|
|
|
|
|
bpf, netns: Fix use-after-free in pernet pre_exit callback
Iterating over BPF links attached to network namespace in pre_exit hook is
not safe, even if there is just one. Once link gets auto-detached, that is
its back-pointer to net object is set to NULL, the link can be released and
freed without waiting on netns_bpf_mutex, effectively causing the list
element we are operating on to be freed.
This leads to use-after-free when trying to access the next element on the
list, as reported by KASAN. Bug can be triggered by destroying a network
namespace, while also releasing a link attached to this network namespace.
| ==================================================================
| BUG: KASAN: use-after-free in netns_bpf_pernet_pre_exit+0xd9/0x130
| Read of size 8 at addr ffff888119e0d778 by task kworker/u8:2/177
|
| CPU: 3 PID: 177 Comm: kworker/u8:2 Not tainted 5.8.0-rc1-00197-ga0c04c9d1008-dirty #776
| Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014
| Workqueue: netns cleanup_net
| Call Trace:
| dump_stack+0x9e/0xe0
| print_address_description.constprop.0+0x3a/0x60
| ? netns_bpf_pernet_pre_exit+0xd9/0x130
| kasan_report.cold+0x1f/0x40
| ? netns_bpf_pernet_pre_exit+0xd9/0x130
| netns_bpf_pernet_pre_exit+0xd9/0x130
| cleanup_net+0x30b/0x5b0
| ? unregister_pernet_device+0x50/0x50
| ? rcu_read_lock_bh_held+0xb0/0xb0
| ? _raw_spin_unlock_irq+0x24/0x50
| process_one_work+0x4d1/0xa10
| ? lock_release+0x3e0/0x3e0
| ? pwq_dec_nr_in_flight+0x110/0x110
| ? rwlock_bug.part.0+0x60/0x60
| worker_thread+0x7a/0x5c0
| ? process_one_work+0xa10/0xa10
| kthread+0x1e3/0x240
| ? kthread_create_on_node+0xd0/0xd0
| ret_from_fork+0x1f/0x30
|
| Allocated by task 280:
| save_stack+0x1b/0x40
| __kasan_kmalloc.constprop.0+0xc2/0xd0
| netns_bpf_link_create+0xfe/0x650
| __do_sys_bpf+0x153a/0x2a50
| do_syscall_64+0x59/0x300
| entry_SYSCALL_64_after_hwframe+0x44/0xa9
|
| Freed by task 198:
| save_stack+0x1b/0x40
| __kasan_slab_free+0x12f/0x180
| kfree+0xed/0x350
| process_one_work+0x4d1/0xa10
| worker_thread+0x7a/0x5c0
| kthread+0x1e3/0x240
| ret_from_fork+0x1f/0x30
|
| The buggy address belongs to the object at ffff888119e0d700
| which belongs to the cache kmalloc-192 of size 192
| The buggy address is located 120 bytes inside of
| 192-byte region [ffff888119e0d700, ffff888119e0d7c0)
| The buggy address belongs to the page:
| page:ffffea0004678340 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0
| flags: 0x2fffe0000000200(slab)
| raw: 02fffe0000000200 ffffea00045ba8c0 0000000600000006 ffff88811a80ea80
| raw: 0000000000000000 0000000000100010 00000001ffffffff 0000000000000000
| page dumped because: kasan: bad access detected
|
| Memory state around the buggy address:
| ffff888119e0d600: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
| ffff888119e0d680: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
| >ffff888119e0d700: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
| ^
| ffff888119e0d780: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
| ffff888119e0d800: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
| ==================================================================
Remove the "fast-path" for releasing a link that got auto-detached by a
dying network namespace to fix it. This way as long as link is on the list
and netns_bpf mutex is held, we have a guarantee that link memory can be
accessed.
An alternative way to fix this issue would be to safely iterate over the
list of links and ensure there is no access to link object after detaching
it. But, at the moment, optimizing synchronization overhead on link release
without a workload in mind seems like an overkill.
Fixes: ab53cad90eb1 ("bpf, netns: Keep a list of attached bpf_link's")
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200630164541.1329993-1-jakub@cloudflare.com
2020-07-01 00:45:41 +08:00
|
|
|
/* We can race with cleanup_net, but if we see a non-NULL
|
|
|
|
|
* struct net pointer, pre_exit has not run yet and wait for
|
|
|
|
|
* netns_bpf_mutex.
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
*/
|
|
|
|
|
net = net_link->net;
|
|
|
|
|
if (!net)
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
/* Mark attach point as unused */
|
|
|
|
|
netns_bpf_attach_type_unneed(type);
|
|
|
|
|
|
2020-07-17 18:35:22 +08:00
|
|
|
/* Remember link position in case of safe delete */
|
|
|
|
|
idx = link_index(net, type, net_link);
|
2020-06-25 22:13:56 +08:00
|
|
|
list_del(&net_link->node);
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
|
2020-07-17 18:35:22 +08:00
|
|
|
cnt = link_count(net, type);
|
|
|
|
|
if (!cnt) {
|
|
|
|
|
netns_bpf_run_array_detach(net, type);
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
old_array = rcu_dereference_protected(net->bpf.run_array[type],
|
|
|
|
|
lockdep_is_held(&netns_bpf_mutex));
|
|
|
|
|
new_array = bpf_prog_array_alloc(cnt, GFP_KERNEL);
|
|
|
|
|
if (!new_array) {
|
|
|
|
|
WARN_ON(bpf_prog_array_delete_safe_at(old_array, idx));
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
fill_prog_array(net, type, new_array);
|
|
|
|
|
rcu_assign_pointer(net->bpf.run_array[type], new_array);
|
|
|
|
|
bpf_prog_array_free(old_array);
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
out_unlock:
|
2020-08-01 02:28:26 +08:00
|
|
|
net_link->net = NULL;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
|
|
|
|
}
|
|
|
|
|
|
2020-08-01 02:28:26 +08:00
|
|
|
static int bpf_netns_link_detach(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
bpf_netns_link_release(link);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
static void bpf_netns_link_dealloc(struct bpf_link *link)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_netns_link *net_link =
|
|
|
|
|
container_of(link, struct bpf_netns_link, link);
|
|
|
|
|
|
|
|
|
|
kfree(net_link);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_netns_link_update_prog(struct bpf_link *link,
|
|
|
|
|
struct bpf_prog *new_prog,
|
|
|
|
|
struct bpf_prog *old_prog)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_netns_link *net_link =
|
|
|
|
|
container_of(link, struct bpf_netns_link, link);
|
|
|
|
|
enum netns_bpf_attach_type type = net_link->netns_type;
|
2020-06-25 22:13:55 +08:00
|
|
|
struct bpf_prog_array *run_array;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
struct net *net;
|
2020-07-17 18:35:22 +08:00
|
|
|
int idx, ret;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
|
|
|
|
|
if (old_prog && old_prog != link->prog)
|
|
|
|
|
return -EPERM;
|
|
|
|
|
if (new_prog->type != link->prog->type)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
mutex_lock(&netns_bpf_mutex);
|
|
|
|
|
|
|
|
|
|
net = net_link->net;
|
|
|
|
|
if (!net || !check_net(net)) {
|
|
|
|
|
/* Link auto-detached or netns dying */
|
|
|
|
|
ret = -ENOLINK;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
|
2020-06-25 22:13:55 +08:00
|
|
|
run_array = rcu_dereference_protected(net->bpf.run_array[type],
|
|
|
|
|
lockdep_is_held(&netns_bpf_mutex));
|
2020-07-17 18:35:22 +08:00
|
|
|
idx = link_index(net, type, net_link);
|
|
|
|
|
ret = bpf_prog_array_update_at(run_array, idx, new_prog);
|
|
|
|
|
if (ret)
|
|
|
|
|
goto out_unlock;
|
2020-06-25 22:13:55 +08:00
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
old_prog = xchg(&link->prog, new_prog);
|
|
|
|
|
bpf_prog_put(old_prog);
|
|
|
|
|
|
|
|
|
|
out_unlock:
|
|
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int bpf_netns_link_fill_info(const struct bpf_link *link,
|
|
|
|
|
struct bpf_link_info *info)
|
|
|
|
|
{
|
|
|
|
|
const struct bpf_netns_link *net_link =
|
|
|
|
|
container_of(link, struct bpf_netns_link, link);
|
|
|
|
|
unsigned int inum = 0;
|
|
|
|
|
struct net *net;
|
|
|
|
|
|
|
|
|
|
mutex_lock(&netns_bpf_mutex);
|
|
|
|
|
net = net_link->net;
|
|
|
|
|
if (net && check_net(net))
|
|
|
|
|
inum = net->ns.inum;
|
|
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
|
|
|
|
|
|
|
|
|
info->netns.netns_ino = inum;
|
|
|
|
|
info->netns.attach_type = net_link->type;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void bpf_netns_link_show_fdinfo(const struct bpf_link *link,
|
|
|
|
|
struct seq_file *seq)
|
|
|
|
|
{
|
|
|
|
|
struct bpf_link_info info = {};
|
|
|
|
|
|
|
|
|
|
bpf_netns_link_fill_info(link, &info);
|
|
|
|
|
seq_printf(seq,
|
|
|
|
|
"netns_ino:\t%u\n"
|
|
|
|
|
"attach_type:\t%u\n",
|
|
|
|
|
info.netns.netns_ino,
|
|
|
|
|
info.netns.attach_type);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static const struct bpf_link_ops bpf_netns_link_ops = {
|
|
|
|
|
.release = bpf_netns_link_release,
|
|
|
|
|
.dealloc = bpf_netns_link_dealloc,
|
2020-08-01 02:28:26 +08:00
|
|
|
.detach = bpf_netns_link_detach,
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
.update_prog = bpf_netns_link_update_prog,
|
|
|
|
|
.fill_link_info = bpf_netns_link_fill_info,
|
|
|
|
|
.show_fdinfo = bpf_netns_link_show_fdinfo,
|
|
|
|
|
};
|
|
|
|
|
|
2020-06-25 22:13:55 +08:00
|
|
|
/* Must be called with netns_bpf_mutex held. */
|
|
|
|
|
static int __netns_bpf_prog_query(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr,
|
|
|
|
|
struct net *net,
|
|
|
|
|
enum netns_bpf_attach_type type)
|
|
|
|
|
{
|
|
|
|
|
__u32 __user *prog_ids = u64_to_user_ptr(attr->query.prog_ids);
|
|
|
|
|
struct bpf_prog_array *run_array;
|
|
|
|
|
u32 prog_cnt = 0, flags = 0;
|
|
|
|
|
|
|
|
|
|
run_array = rcu_dereference_protected(net->bpf.run_array[type],
|
|
|
|
|
lockdep_is_held(&netns_bpf_mutex));
|
|
|
|
|
if (run_array)
|
|
|
|
|
prog_cnt = bpf_prog_array_length(run_array);
|
|
|
|
|
|
|
|
|
|
if (copy_to_user(&uattr->query.attach_flags, &flags, sizeof(flags)))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
if (copy_to_user(&uattr->query.prog_cnt, &prog_cnt, sizeof(prog_cnt)))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
if (!attr->query.prog_cnt || !prog_ids || !prog_cnt)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
return bpf_prog_array_copy_to_user(run_array, prog_ids,
|
|
|
|
|
attr->query.prog_cnt);
|
|
|
|
|
}
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
int netns_bpf_prog_query(const union bpf_attr *attr,
|
|
|
|
|
union bpf_attr __user *uattr)
|
|
|
|
|
{
|
|
|
|
|
enum netns_bpf_attach_type type;
|
|
|
|
|
struct net *net;
|
2020-06-25 22:13:55 +08:00
|
|
|
int ret;
|
2020-05-31 16:28:37 +08:00
|
|
|
|
|
|
|
|
if (attr->query.query_flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
type = to_netns_bpf_attach_type(attr->query.attach_type);
|
|
|
|
|
if (type < 0)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
net = get_net_ns_by_fd(attr->query.target_fd);
|
|
|
|
|
if (IS_ERR(net))
|
|
|
|
|
return PTR_ERR(net);
|
|
|
|
|
|
2020-06-25 22:13:55 +08:00
|
|
|
mutex_lock(&netns_bpf_mutex);
|
|
|
|
|
ret = __netns_bpf_prog_query(attr, uattr, net, type);
|
|
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
2020-05-31 16:28:37 +08:00
|
|
|
|
|
|
|
|
put_net(net);
|
2020-06-25 22:13:55 +08:00
|
|
|
return ret;
|
2020-05-31 16:28:37 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int netns_bpf_prog_attach(const union bpf_attr *attr, struct bpf_prog *prog)
|
|
|
|
|
{
|
2020-06-25 22:13:55 +08:00
|
|
|
struct bpf_prog_array *run_array;
|
2020-05-31 16:28:37 +08:00
|
|
|
enum netns_bpf_attach_type type;
|
2020-06-25 22:13:54 +08:00
|
|
|
struct bpf_prog *attached;
|
2020-05-31 16:28:37 +08:00
|
|
|
struct net *net;
|
|
|
|
|
int ret;
|
|
|
|
|
|
2020-06-29 17:56:25 +08:00
|
|
|
if (attr->target_fd || attr->attach_flags || attr->replace_bpf_fd)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
type = to_netns_bpf_attach_type(attr->attach_type);
|
|
|
|
|
if (type < 0)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
net = current->nsproxy->net_ns;
|
|
|
|
|
mutex_lock(&netns_bpf_mutex);
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
|
|
|
|
|
/* Attaching prog directly is not compatible with links */
|
2020-06-25 22:13:56 +08:00
|
|
|
if (!list_empty(&net->bpf.links[type])) {
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
ret = -EEXIST;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
switch (type) {
|
|
|
|
|
case NETNS_BPF_FLOW_DISSECTOR:
|
2020-06-25 22:13:54 +08:00
|
|
|
ret = flow_dissector_bpf_prog_attach_check(net, prog);
|
2020-05-31 16:28:37 +08:00
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
ret = -EINVAL;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2020-06-25 22:13:54 +08:00
|
|
|
if (ret)
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
2020-06-25 22:13:55 +08:00
|
|
|
attached = net->bpf.progs[type];
|
2020-06-25 22:13:54 +08:00
|
|
|
if (attached == prog) {
|
|
|
|
|
/* The same program cannot be attached twice */
|
|
|
|
|
ret = -EINVAL;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
2020-06-25 22:13:55 +08:00
|
|
|
|
|
|
|
|
run_array = rcu_dereference_protected(net->bpf.run_array[type],
|
|
|
|
|
lockdep_is_held(&netns_bpf_mutex));
|
|
|
|
|
if (run_array) {
|
|
|
|
|
WRITE_ONCE(run_array->items[0].prog, prog);
|
|
|
|
|
} else {
|
|
|
|
|
run_array = bpf_prog_array_alloc(1, GFP_KERNEL);
|
|
|
|
|
if (!run_array) {
|
|
|
|
|
ret = -ENOMEM;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
run_array->items[0].prog = prog;
|
|
|
|
|
rcu_assign_pointer(net->bpf.run_array[type], run_array);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
net->bpf.progs[type] = prog;
|
2020-06-25 22:13:54 +08:00
|
|
|
if (attached)
|
|
|
|
|
bpf_prog_put(attached);
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
out_unlock:
|
2020-05-31 16:28:37 +08:00
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
|
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Must be called with netns_bpf_mutex held. */
|
|
|
|
|
static int __netns_bpf_prog_detach(struct net *net,
|
2020-06-29 17:56:26 +08:00
|
|
|
enum netns_bpf_attach_type type,
|
|
|
|
|
struct bpf_prog *old)
|
2020-05-31 16:28:37 +08:00
|
|
|
{
|
|
|
|
|
struct bpf_prog *attached;
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
/* Progs attached via links cannot be detached */
|
2020-06-25 22:13:56 +08:00
|
|
|
if (!list_empty(&net->bpf.links[type]))
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
return -EINVAL;
|
|
|
|
|
|
2020-06-25 22:13:55 +08:00
|
|
|
attached = net->bpf.progs[type];
|
2020-06-29 17:56:26 +08:00
|
|
|
if (!attached || attached != old)
|
2020-05-31 16:28:37 +08:00
|
|
|
return -ENOENT;
|
2020-06-25 22:13:55 +08:00
|
|
|
netns_bpf_run_array_detach(net, type);
|
|
|
|
|
net->bpf.progs[type] = NULL;
|
2020-05-31 16:28:37 +08:00
|
|
|
bpf_prog_put(attached);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2020-06-29 17:56:26 +08:00
|
|
|
int netns_bpf_prog_detach(const union bpf_attr *attr, enum bpf_prog_type ptype)
|
2020-05-31 16:28:37 +08:00
|
|
|
{
|
|
|
|
|
enum netns_bpf_attach_type type;
|
2020-06-29 17:56:26 +08:00
|
|
|
struct bpf_prog *prog;
|
2020-05-31 16:28:37 +08:00
|
|
|
int ret;
|
|
|
|
|
|
2020-06-29 17:56:26 +08:00
|
|
|
if (attr->target_fd)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
type = to_netns_bpf_attach_type(attr->attach_type);
|
|
|
|
|
if (type < 0)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
2020-06-29 17:56:26 +08:00
|
|
|
prog = bpf_prog_get_type(attr->attach_bpf_fd, ptype);
|
|
|
|
|
if (IS_ERR(prog))
|
|
|
|
|
return PTR_ERR(prog);
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
mutex_lock(&netns_bpf_mutex);
|
2020-06-29 17:56:26 +08:00
|
|
|
ret = __netns_bpf_prog_detach(current->nsproxy->net_ns, type, prog);
|
2020-05-31 16:28:37 +08:00
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
|
|
|
|
|
2020-06-29 17:56:26 +08:00
|
|
|
bpf_prog_put(prog);
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2020-07-17 18:35:22 +08:00
|
|
|
static int netns_bpf_max_progs(enum netns_bpf_attach_type type)
|
|
|
|
|
{
|
|
|
|
|
switch (type) {
|
|
|
|
|
case NETNS_BPF_FLOW_DISSECTOR:
|
|
|
|
|
return 1;
|
bpf: Introduce SK_LOOKUP program type with a dedicated attach point
Add a new program type BPF_PROG_TYPE_SK_LOOKUP with a dedicated attach type
BPF_SK_LOOKUP. The new program kind is to be invoked by the transport layer
when looking up a listening socket for a new connection request for
connection oriented protocols, or when looking up an unconnected socket for
a packet for connection-less protocols.
When called, SK_LOOKUP BPF program can select a socket that will receive
the packet. This serves as a mechanism to overcome the limits of what
bind() API allows to express. Two use-cases driving this work are:
(1) steer packets destined to an IP range, on fixed port to a socket
192.0.2.0/24, port 80 -> NGINX socket
(2) steer packets destined to an IP address, on any port to a socket
198.51.100.1, any port -> L7 proxy socket
In its run-time context program receives information about the packet that
triggered the socket lookup. Namely IP version, L4 protocol identifier, and
address 4-tuple. Context can be further extended to include ingress
interface identifier.
To select a socket BPF program fetches it from a map holding socket
references, like SOCKMAP or SOCKHASH, and calls bpf_sk_assign(ctx, sk, ...)
helper to record the selection. Transport layer then uses the selected
socket as a result of socket lookup.
In its basic form, SK_LOOKUP acts as a filter and hence must return either
SK_PASS or SK_DROP. If the program returns with SK_PASS, transport should
look for a socket to receive the packet, or use the one selected by the
program if available, while SK_DROP informs the transport layer that the
lookup should fail.
This patch only enables the user to attach an SK_LOOKUP program to a
network namespace. Subsequent patches hook it up to run on local delivery
path in ipv4 and ipv6 stacks.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-3-jakub@cloudflare.com
2020-07-17 18:35:23 +08:00
|
|
|
case NETNS_BPF_SK_LOOKUP:
|
|
|
|
|
return 64;
|
2020-07-17 18:35:22 +08:00
|
|
|
default:
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
static int netns_bpf_link_attach(struct net *net, struct bpf_link *link,
|
|
|
|
|
enum netns_bpf_attach_type type)
|
|
|
|
|
{
|
2020-06-25 22:13:56 +08:00
|
|
|
struct bpf_netns_link *net_link =
|
|
|
|
|
container_of(link, struct bpf_netns_link, link);
|
2020-06-25 22:13:55 +08:00
|
|
|
struct bpf_prog_array *run_array;
|
2020-07-17 18:35:22 +08:00
|
|
|
int cnt, err;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
|
|
|
|
|
mutex_lock(&netns_bpf_mutex);
|
|
|
|
|
|
2020-07-17 18:35:22 +08:00
|
|
|
cnt = link_count(net, type);
|
|
|
|
|
if (cnt >= netns_bpf_max_progs(type)) {
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
err = -E2BIG;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
/* Links are not compatible with attaching prog directly */
|
2020-06-25 22:13:55 +08:00
|
|
|
if (net->bpf.progs[type]) {
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
err = -EEXIST;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
|
case NETNS_BPF_FLOW_DISSECTOR:
|
2020-06-25 22:13:54 +08:00
|
|
|
err = flow_dissector_bpf_prog_attach_check(net, link->prog);
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
break;
|
bpf: Introduce SK_LOOKUP program type with a dedicated attach point
Add a new program type BPF_PROG_TYPE_SK_LOOKUP with a dedicated attach type
BPF_SK_LOOKUP. The new program kind is to be invoked by the transport layer
when looking up a listening socket for a new connection request for
connection oriented protocols, or when looking up an unconnected socket for
a packet for connection-less protocols.
When called, SK_LOOKUP BPF program can select a socket that will receive
the packet. This serves as a mechanism to overcome the limits of what
bind() API allows to express. Two use-cases driving this work are:
(1) steer packets destined to an IP range, on fixed port to a socket
192.0.2.0/24, port 80 -> NGINX socket
(2) steer packets destined to an IP address, on any port to a socket
198.51.100.1, any port -> L7 proxy socket
In its run-time context program receives information about the packet that
triggered the socket lookup. Namely IP version, L4 protocol identifier, and
address 4-tuple. Context can be further extended to include ingress
interface identifier.
To select a socket BPF program fetches it from a map holding socket
references, like SOCKMAP or SOCKHASH, and calls bpf_sk_assign(ctx, sk, ...)
helper to record the selection. Transport layer then uses the selected
socket as a result of socket lookup.
In its basic form, SK_LOOKUP acts as a filter and hence must return either
SK_PASS or SK_DROP. If the program returns with SK_PASS, transport should
look for a socket to receive the packet, or use the one selected by the
program if available, while SK_DROP informs the transport layer that the
lookup should fail.
This patch only enables the user to attach an SK_LOOKUP program to a
network namespace. Subsequent patches hook it up to run on local delivery
path in ipv4 and ipv6 stacks.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-3-jakub@cloudflare.com
2020-07-17 18:35:23 +08:00
|
|
|
case NETNS_BPF_SK_LOOKUP:
|
|
|
|
|
err = 0; /* nothing to check */
|
|
|
|
|
break;
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
default:
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
if (err)
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
2020-07-17 18:35:22 +08:00
|
|
|
run_array = bpf_prog_array_alloc(cnt + 1, GFP_KERNEL);
|
2020-06-25 22:13:55 +08:00
|
|
|
if (!run_array) {
|
|
|
|
|
err = -ENOMEM;
|
|
|
|
|
goto out_unlock;
|
|
|
|
|
}
|
|
|
|
|
|
2020-06-25 22:13:56 +08:00
|
|
|
list_add_tail(&net_link->node, &net->bpf.links[type]);
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
|
2020-07-17 18:35:22 +08:00
|
|
|
fill_prog_array(net, type, run_array);
|
|
|
|
|
run_array = rcu_replace_pointer(net->bpf.run_array[type], run_array,
|
|
|
|
|
lockdep_is_held(&netns_bpf_mutex));
|
|
|
|
|
bpf_prog_array_free(run_array);
|
|
|
|
|
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
/* Mark attach point as used */
|
|
|
|
|
netns_bpf_attach_type_need(type);
|
|
|
|
|
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
out_unlock:
|
|
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int netns_bpf_link_create(const union bpf_attr *attr, struct bpf_prog *prog)
|
|
|
|
|
{
|
|
|
|
|
enum netns_bpf_attach_type netns_type;
|
|
|
|
|
struct bpf_link_primer link_primer;
|
|
|
|
|
struct bpf_netns_link *net_link;
|
|
|
|
|
enum bpf_attach_type type;
|
|
|
|
|
struct net *net;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (attr->link_create.flags)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
type = attr->link_create.attach_type;
|
|
|
|
|
netns_type = to_netns_bpf_attach_type(type);
|
|
|
|
|
if (netns_type < 0)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
net = get_net_ns_by_fd(attr->link_create.target_fd);
|
|
|
|
|
if (IS_ERR(net))
|
|
|
|
|
return PTR_ERR(net);
|
|
|
|
|
|
|
|
|
|
net_link = kzalloc(sizeof(*net_link), GFP_USER);
|
|
|
|
|
if (!net_link) {
|
|
|
|
|
err = -ENOMEM;
|
|
|
|
|
goto out_put_net;
|
|
|
|
|
}
|
|
|
|
|
bpf_link_init(&net_link->link, BPF_LINK_TYPE_NETNS,
|
|
|
|
|
&bpf_netns_link_ops, prog);
|
|
|
|
|
net_link->net = net;
|
|
|
|
|
net_link->type = type;
|
|
|
|
|
net_link->netns_type = netns_type;
|
|
|
|
|
|
|
|
|
|
err = bpf_link_prime(&net_link->link, &link_primer);
|
|
|
|
|
if (err) {
|
|
|
|
|
kfree(net_link);
|
|
|
|
|
goto out_put_net;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
err = netns_bpf_link_attach(net, &net_link->link, netns_type);
|
|
|
|
|
if (err) {
|
|
|
|
|
bpf_link_cleanup(&link_primer);
|
|
|
|
|
goto out_put_net;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
put_net(net);
|
|
|
|
|
return bpf_link_settle(&link_primer);
|
|
|
|
|
|
|
|
|
|
out_put_net:
|
|
|
|
|
put_net(net);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2020-06-25 22:13:56 +08:00
|
|
|
static int __net_init netns_bpf_pernet_init(struct net *net)
|
|
|
|
|
{
|
|
|
|
|
int type;
|
|
|
|
|
|
|
|
|
|
for (type = 0; type < MAX_NETNS_BPF_ATTACH_TYPE; type++)
|
|
|
|
|
INIT_LIST_HEAD(&net->bpf.links[type]);
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2020-05-31 16:28:37 +08:00
|
|
|
static void __net_exit netns_bpf_pernet_pre_exit(struct net *net)
|
|
|
|
|
{
|
|
|
|
|
enum netns_bpf_attach_type type;
|
2020-06-25 22:13:56 +08:00
|
|
|
struct bpf_netns_link *net_link;
|
2020-05-31 16:28:37 +08:00
|
|
|
|
|
|
|
|
mutex_lock(&netns_bpf_mutex);
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
for (type = 0; type < MAX_NETNS_BPF_ATTACH_TYPE; type++) {
|
2020-06-25 22:13:55 +08:00
|
|
|
netns_bpf_run_array_detach(net, type);
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
list_for_each_entry(net_link, &net->bpf.links[type], node) {
|
2020-06-25 22:13:56 +08:00
|
|
|
net_link->net = NULL; /* auto-detach link */
|
inet: Run SK_LOOKUP BPF program on socket lookup
Run a BPF program before looking up a listening socket on the receive path.
Program selects a listening socket to yield as result of socket lookup by
calling bpf_sk_assign() helper and returning SK_PASS code. Program can
revert its decision by assigning a NULL socket with bpf_sk_assign().
Alternatively, BPF program can also fail the lookup by returning with
SK_DROP, or let the lookup continue as usual with SK_PASS on return, when
no socket has been selected with bpf_sk_assign().
This lets the user match packets with listening sockets freely at the last
possible point on the receive path, where we know that packets are destined
for local delivery after undergoing policing, filtering, and routing.
With BPF code selecting the socket, directing packets destined to an IP
range or to a port range to a single socket becomes possible.
In case multiple programs are attached, they are run in series in the order
in which they were attached. The end result is determined from return codes
of all the programs according to following rules:
1. If any program returned SK_PASS and selected a valid socket, the socket
is used as result of socket lookup.
2. If more than one program returned SK_PASS and selected a socket,
last selection takes effect.
3. If any program returned SK_DROP, and no program returned SK_PASS and
selected a socket, socket lookup fails with -ECONNREFUSED.
4. If all programs returned SK_PASS and none of them selected a socket,
socket lookup continues to htable-based lookup.
Suggested-by: Marek Majkowski <marek@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200717103536.397595-5-jakub@cloudflare.com
2020-07-17 18:35:25 +08:00
|
|
|
netns_bpf_attach_type_unneed(type);
|
|
|
|
|
}
|
2020-06-25 22:13:56 +08:00
|
|
|
if (net->bpf.progs[type])
|
2020-06-25 22:13:55 +08:00
|
|
|
bpf_prog_put(net->bpf.progs[type]);
|
bpf: Add link-based BPF program attachment to network namespace
Extend bpf() syscall subcommands that operate on bpf_link, that is
LINK_CREATE, LINK_UPDATE, OBJ_GET_INFO, to accept attach types tied to
network namespaces (only flow dissector at the moment).
Link-based and prog-based attachment can be used interchangeably, but only
one can exist at a time. Attempts to attach a link when a prog is already
attached directly, and the other way around, will be met with -EEXIST.
Attempts to detach a program when link exists result in -EINVAL.
Attachment of multiple links of same attach type to one netns is not
supported with the intention to lift the restriction when a use-case
presents itself. Because of that link create returns -E2BIG when trying to
create another netns link, when one already exists.
Link-based attachments to netns don't keep a netns alive by holding a ref
to it. Instead links get auto-detached from netns when the latter is being
destroyed, using a pernet pre_exit callback.
When auto-detached, link lives in defunct state as long there are open FDs
for it. -ENOLINK is returned if a user tries to update a defunct link.
Because bpf_link to netns doesn't hold a ref to struct net, special care is
taken when releasing, updating, or filling link info. The netns might be
getting torn down when any of these link operations are in progress. That
is why auto-detach and update/release/fill_info are synchronized by the
same mutex. Also, link ops have to always check if auto-detach has not
happened yet and if netns is still alive (refcnt > 0).
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200531082846.2117903-5-jakub@cloudflare.com
2020-05-31 16:28:38 +08:00
|
|
|
}
|
2020-05-31 16:28:37 +08:00
|
|
|
mutex_unlock(&netns_bpf_mutex);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct pernet_operations netns_bpf_pernet_ops __net_initdata = {
|
2020-06-25 22:13:56 +08:00
|
|
|
.init = netns_bpf_pernet_init,
|
2020-05-31 16:28:37 +08:00
|
|
|
.pre_exit = netns_bpf_pernet_pre_exit,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static int __init netns_bpf_init(void)
|
|
|
|
|
{
|
|
|
|
|
return register_pernet_subsys(&netns_bpf_pernet_ops);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
subsys_initcall(netns_bpf_init);
|