linux_dsm_epyc7002/kernel/bpf/cgroup.c

1553 lines
40 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* Functions to manage eBPF programs attached to cgroups
*
* Copyright (c) 2016 Daniel Mack
*/
#include <linux/kernel.h>
#include <linux/atomic.h>
#include <linux/cgroup.h>
#include <linux/filter.h>
#include <linux/slab.h>
#include <linux/sysctl.h>
#include <linux/string.h>
#include <linux/bpf.h>
#include <linux/bpf-cgroup.h>
#include <net/sock.h>
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
#include <net/bpf_sk_storage.h>
#include "../cgroup/cgroup-internal.h"
DEFINE_STATIC_KEY_FALSE(cgroup_bpf_enabled_key);
EXPORT_SYMBOL(cgroup_bpf_enabled_key);
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
void cgroup_bpf_offline(struct cgroup *cgrp)
{
cgroup_get(cgrp);
percpu_ref_kill(&cgrp->bpf.refcnt);
}
/**
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
* cgroup_bpf_release() - put references of all bpf programs and
* release all cgroup bpf data
* @work: work structure embedded into the cgroup to modify
*/
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
static void cgroup_bpf_release(struct work_struct *work)
{
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
struct cgroup *cgrp = container_of(work, struct cgroup,
bpf.release_work);
enum bpf_cgroup_storage_type stype;
struct bpf_prog_array *old_array;
unsigned int type;
mutex_lock(&cgroup_mutex);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
for (type = 0; type < ARRAY_SIZE(cgrp->bpf.progs); type++) {
struct list_head *progs = &cgrp->bpf.progs[type];
struct bpf_prog_list *pl, *tmp;
list_for_each_entry_safe(pl, tmp, progs, node) {
list_del(&pl->node);
bpf_prog_put(pl->prog);
for_each_cgroup_storage_type(stype) {
bpf_cgroup_storage_unlink(pl->storage[stype]);
bpf_cgroup_storage_free(pl->storage[stype]);
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
kfree(pl);
static_branch_dec(&cgroup_bpf_enabled_key);
}
old_array = rcu_dereference_protected(
cgrp->bpf.effective[type],
lockdep_is_held(&cgroup_mutex));
bpf_prog_array_free(old_array);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
}
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
mutex_unlock(&cgroup_mutex);
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
percpu_ref_exit(&cgrp->bpf.refcnt);
cgroup_put(cgrp);
}
/**
* cgroup_bpf_release_fn() - callback used to schedule releasing
* of bpf cgroup data
* @ref: percpu ref counter structure
*/
static void cgroup_bpf_release_fn(struct percpu_ref *ref)
{
struct cgroup *cgrp = container_of(ref, struct cgroup, bpf.refcnt);
INIT_WORK(&cgrp->bpf.release_work, cgroup_bpf_release);
queue_work(system_wq, &cgrp->bpf.release_work);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
}
/* count number of elements in the list.
* it's slow but the list cannot be long
*/
static u32 prog_list_length(struct list_head *head)
{
struct bpf_prog_list *pl;
u32 cnt = 0;
list_for_each_entry(pl, head, node) {
if (!pl->prog)
continue;
cnt++;
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
return cnt;
}
/* if parent has non-overridable prog attached,
* disallow attaching new programs to the descendent cgroup.
* if parent has overridable or multi-prog, allow attaching
*/
static bool hierarchy_allows_attach(struct cgroup *cgrp,
enum bpf_attach_type type,
u32 new_flags)
{
struct cgroup *p;
p = cgroup_parent(cgrp);
if (!p)
return true;
do {
u32 flags = p->bpf.flags[type];
u32 cnt;
if (flags & BPF_F_ALLOW_MULTI)
return true;
cnt = prog_list_length(&p->bpf.progs[type]);
WARN_ON_ONCE(cnt > 1);
if (cnt == 1)
return !!(flags & BPF_F_ALLOW_OVERRIDE);
p = cgroup_parent(p);
} while (p);
return true;
}
/* compute a chain of effective programs for a given cgroup:
* start from the list of programs in this cgroup and add
* all parent programs.
* Note that parent's F_ALLOW_OVERRIDE-type program is yielding
* to programs in this cgroup
*/
static int compute_effective_progs(struct cgroup *cgrp,
enum bpf_attach_type type,
struct bpf_prog_array **array)
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
{
enum bpf_cgroup_storage_type stype;
struct bpf_prog_array *progs;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
struct bpf_prog_list *pl;
struct cgroup *p = cgrp;
int cnt = 0;
/* count number of effective programs by walking parents */
do {
if (cnt == 0 || (p->bpf.flags[type] & BPF_F_ALLOW_MULTI))
cnt += prog_list_length(&p->bpf.progs[type]);
p = cgroup_parent(p);
} while (p);
progs = bpf_prog_array_alloc(cnt, GFP_KERNEL);
if (!progs)
return -ENOMEM;
/* populate the array with effective progs */
cnt = 0;
p = cgrp;
do {
if (cnt > 0 && !(p->bpf.flags[type] & BPF_F_ALLOW_MULTI))
continue;
list_for_each_entry(pl, &p->bpf.progs[type], node) {
if (!pl->prog)
continue;
progs->items[cnt].prog = pl->prog;
for_each_cgroup_storage_type(stype)
progs->items[cnt].cgroup_storage[stype] =
pl->storage[stype];
cnt++;
}
} while ((p = cgroup_parent(p)));
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
*array = progs;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
return 0;
}
static void activate_effective_progs(struct cgroup *cgrp,
enum bpf_attach_type type,
struct bpf_prog_array *old_array)
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
{
old_array = rcu_replace_pointer(cgrp->bpf.effective[type], old_array,
lockdep_is_held(&cgroup_mutex));
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
/* free prog array after grace period, since __cgroup_bpf_run_*()
* might be still walking the array
*/
bpf_prog_array_free(old_array);
}
/**
* cgroup_bpf_inherit() - inherit effective programs from parent
* @cgrp: the cgroup to modify
*/
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
int cgroup_bpf_inherit(struct cgroup *cgrp)
{
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
/* has to use marco instead of const int, since compiler thinks
* that array below is variable length
*/
#define NR ARRAY_SIZE(cgrp->bpf.effective)
struct bpf_prog_array *arrays[NR] = {};
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
int ret, i;
ret = percpu_ref_init(&cgrp->bpf.refcnt, cgroup_bpf_release_fn, 0,
GFP_KERNEL);
if (ret)
return ret;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
for (i = 0; i < NR; i++)
INIT_LIST_HEAD(&cgrp->bpf.progs[i]);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
for (i = 0; i < NR; i++)
if (compute_effective_progs(cgrp, i, &arrays[i]))
goto cleanup;
for (i = 0; i < NR; i++)
activate_effective_progs(cgrp, i, arrays[i]);
return 0;
cleanup:
for (i = 0; i < NR; i++)
bpf_prog_array_free(arrays[i]);
bpf: decouple the lifetime of cgroup_bpf from cgroup itself Currently the lifetime of bpf programs attached to a cgroup is bound to the lifetime of the cgroup itself. It means that if a user forgets (or intentionally avoids) to detach a bpf program before removing the cgroup, it will stay attached up to the release of the cgroup. Since the cgroup can stay in the dying state (the state between being rmdir()'ed and being released) for a very long time, it leads to a waste of memory. Also, it blocks a possibility to implement the memcg-based memory accounting for bpf objects, because a circular reference dependency will occur. Charged memory pages are pinning the corresponding memory cgroup, and if the memory cgroup is pinning the attached bpf program, nothing will be ever released. A dying cgroup can not contain any processes, so the only chance for an attached bpf program to be executed is a live socket associated with the cgroup. So in order to release all bpf data early, let's count associated sockets using a new percpu refcounter. On cgroup removal the counter is transitioned to the atomic mode, and as soon as it reaches 0, all bpf programs are detached. Because cgroup_bpf_release() can block, it can't be called from the percpu ref counter callback directly, so instead an asynchronous work is scheduled. The reference counter is not socket specific, and can be used for any other types of programs, which can be executed from a cgroup-bpf hook outside of the process context, had such a need arise in the future. Signed-off-by: Roman Gushchin <guro@fb.com> Cc: jolsa@redhat.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-05-25 23:37:39 +07:00
percpu_ref_exit(&cgrp->bpf.refcnt);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
return -ENOMEM;
}
static int update_effective_progs(struct cgroup *cgrp,
enum bpf_attach_type type)
{
struct cgroup_subsys_state *css;
int err;
/* allocate and recompute effective prog arrays */
css_for_each_descendant_pre(css, &cgrp->self) {
struct cgroup *desc = container_of(css, struct cgroup, self);
if (percpu_ref_is_zero(&desc->bpf.refcnt))
continue;
err = compute_effective_progs(desc, type, &desc->bpf.inactive);
if (err)
goto cleanup;
}
/* all allocations were successful. Activate all prog arrays */
css_for_each_descendant_pre(css, &cgrp->self) {
struct cgroup *desc = container_of(css, struct cgroup, self);
if (percpu_ref_is_zero(&desc->bpf.refcnt)) {
if (unlikely(desc->bpf.inactive)) {
bpf_prog_array_free(desc->bpf.inactive);
desc->bpf.inactive = NULL;
}
continue;
}
activate_effective_progs(desc, type, desc->bpf.inactive);
desc->bpf.inactive = NULL;
}
return 0;
cleanup:
/* oom while computing effective. Free all computed effective arrays
* since they were not activated
*/
css_for_each_descendant_pre(css, &cgrp->self) {
struct cgroup *desc = container_of(css, struct cgroup, self);
bpf_prog_array_free(desc->bpf.inactive);
desc->bpf.inactive = NULL;
}
return err;
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
#define BPF_CGROUP_MAX_PROGS 64
/**
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
* __cgroup_bpf_attach() - Attach the program to a cgroup, and
* propagate the change to descendants
* @cgrp: The cgroup which descendants to traverse
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
* @prog: A program to attach
* @type: Type of attach operation
* @flags: Option flags
*
* Must be called with cgroup_mutex held.
*/
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
int __cgroup_bpf_attach(struct cgroup *cgrp, struct bpf_prog *prog,
enum bpf_attach_type type, u32 flags)
{
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
struct list_head *progs = &cgrp->bpf.progs[type];
struct bpf_prog *old_prog = NULL;
struct bpf_cgroup_storage *storage[MAX_BPF_CGROUP_STORAGE_TYPE],
*old_storage[MAX_BPF_CGROUP_STORAGE_TYPE] = {NULL};
enum bpf_cgroup_storage_type stype;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
struct bpf_prog_list *pl;
bool pl_was_allocated;
int err;
if ((flags & BPF_F_ALLOW_OVERRIDE) && (flags & BPF_F_ALLOW_MULTI))
/* invalid combination */
return -EINVAL;
if (!hierarchy_allows_attach(cgrp, type, flags))
return -EPERM;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
if (!list_empty(progs) && cgrp->bpf.flags[type] != flags)
/* Disallow attaching non-overridable on top
* of existing overridable in this cgroup.
* Disallow attaching multi-prog if overridable or none
*/
return -EPERM;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
if (prog_list_length(progs) >= BPF_CGROUP_MAX_PROGS)
return -E2BIG;
for_each_cgroup_storage_type(stype) {
storage[stype] = bpf_cgroup_storage_alloc(prog, stype);
if (IS_ERR(storage[stype])) {
storage[stype] = NULL;
for_each_cgroup_storage_type(stype)
bpf_cgroup_storage_free(storage[stype]);
return -ENOMEM;
}
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
if (flags & BPF_F_ALLOW_MULTI) {
list_for_each_entry(pl, progs, node) {
if (pl->prog == prog) {
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
/* disallow attaching the same prog twice */
for_each_cgroup_storage_type(stype)
bpf_cgroup_storage_free(storage[stype]);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
return -EINVAL;
}
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
pl = kmalloc(sizeof(*pl), GFP_KERNEL);
if (!pl) {
for_each_cgroup_storage_type(stype)
bpf_cgroup_storage_free(storage[stype]);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
return -ENOMEM;
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
pl_was_allocated = true;
pl->prog = prog;
for_each_cgroup_storage_type(stype)
pl->storage[stype] = storage[stype];
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
list_add_tail(&pl->node, progs);
} else {
if (list_empty(progs)) {
pl = kmalloc(sizeof(*pl), GFP_KERNEL);
if (!pl) {
for_each_cgroup_storage_type(stype)
bpf_cgroup_storage_free(storage[stype]);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
return -ENOMEM;
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
pl_was_allocated = true;
list_add_tail(&pl->node, progs);
} else {
pl = list_first_entry(progs, typeof(*pl), node);
old_prog = pl->prog;
for_each_cgroup_storage_type(stype) {
old_storage[stype] = pl->storage[stype];
bpf_cgroup_storage_unlink(old_storage[stype]);
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
pl_was_allocated = false;
}
pl->prog = prog;
for_each_cgroup_storage_type(stype)
pl->storage[stype] = storage[stype];
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
cgrp->bpf.flags[type] = flags;
err = update_effective_progs(cgrp, type);
if (err)
goto cleanup;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
static_branch_inc(&cgroup_bpf_enabled_key);
for_each_cgroup_storage_type(stype) {
if (!old_storage[stype])
continue;
bpf_cgroup_storage_free(old_storage[stype]);
}
if (old_prog) {
bpf_prog_put(old_prog);
static_branch_dec(&cgroup_bpf_enabled_key);
}
for_each_cgroup_storage_type(stype)
bpf_cgroup_storage_link(storage[stype], cgrp, type);
return 0;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
cleanup:
/* and cleanup the prog list */
pl->prog = old_prog;
for_each_cgroup_storage_type(stype) {
bpf_cgroup_storage_free(pl->storage[stype]);
pl->storage[stype] = old_storage[stype];
bpf_cgroup_storage_link(old_storage[stype], cgrp, type);
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
if (pl_was_allocated) {
list_del(&pl->node);
kfree(pl);
}
return err;
}
/**
* __cgroup_bpf_detach() - Detach the program from a cgroup, and
* propagate the change to descendants
* @cgrp: The cgroup which descendants to traverse
* @prog: A program to detach or NULL
* @type: Type of detach operation
*
* Must be called with cgroup_mutex held.
*/
int __cgroup_bpf_detach(struct cgroup *cgrp, struct bpf_prog *prog,
enum bpf_attach_type type)
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
{
struct list_head *progs = &cgrp->bpf.progs[type];
enum bpf_cgroup_storage_type stype;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
u32 flags = cgrp->bpf.flags[type];
struct bpf_prog *old_prog = NULL;
struct bpf_prog_list *pl;
int err;
if (flags & BPF_F_ALLOW_MULTI) {
if (!prog)
/* to detach MULTI prog the user has to specify valid FD
* of the program to be detached
*/
return -EINVAL;
} else {
if (list_empty(progs))
/* report error when trying to detach and nothing is attached */
return -ENOENT;
}
if (flags & BPF_F_ALLOW_MULTI) {
/* find the prog and detach it */
list_for_each_entry(pl, progs, node) {
if (pl->prog != prog)
continue;
old_prog = prog;
/* mark it deleted, so it's ignored while
* recomputing effective
*/
pl->prog = NULL;
break;
}
if (!old_prog)
return -ENOENT;
} else {
/* to maintain backward compatibility NONE and OVERRIDE cgroups
* allow detaching with invalid FD (prog==NULL)
*/
pl = list_first_entry(progs, typeof(*pl), node);
old_prog = pl->prog;
pl->prog = NULL;
}
err = update_effective_progs(cgrp, type);
if (err)
goto cleanup;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
/* now can actually delete it from this cgroup list */
list_del(&pl->node);
for_each_cgroup_storage_type(stype) {
bpf_cgroup_storage_unlink(pl->storage[stype]);
bpf_cgroup_storage_free(pl->storage[stype]);
}
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
kfree(pl);
if (list_empty(progs))
/* last program was detached, reset flags to zero */
cgrp->bpf.flags[type] = 0;
bpf_prog_put(old_prog);
static_branch_dec(&cgroup_bpf_enabled_key);
return 0;
cleanup:
/* and restore back old_prog */
pl->prog = old_prog;
return err;
}
/* Must be called with cgroup_mutex held to avoid races. */
int __cgroup_bpf_query(struct cgroup *cgrp, const union bpf_attr *attr,
union bpf_attr __user *uattr)
{
__u32 __user *prog_ids = u64_to_user_ptr(attr->query.prog_ids);
enum bpf_attach_type type = attr->query.attach_type;
struct list_head *progs = &cgrp->bpf.progs[type];
u32 flags = cgrp->bpf.flags[type];
struct bpf_prog_array *effective;
int cnt, ret = 0, i;
effective = rcu_dereference_protected(cgrp->bpf.effective[type],
lockdep_is_held(&cgroup_mutex));
if (attr->query.query_flags & BPF_F_QUERY_EFFECTIVE)
cnt = bpf_prog_array_length(effective);
else
cnt = prog_list_length(progs);
if (copy_to_user(&uattr->query.attach_flags, &flags, sizeof(flags)))
return -EFAULT;
if (copy_to_user(&uattr->query.prog_cnt, &cnt, sizeof(cnt)))
return -EFAULT;
if (attr->query.prog_cnt == 0 || !prog_ids || !cnt)
/* return early if user requested only program count + flags */
return 0;
if (attr->query.prog_cnt < cnt) {
cnt = attr->query.prog_cnt;
ret = -ENOSPC;
}
if (attr->query.query_flags & BPF_F_QUERY_EFFECTIVE) {
return bpf_prog_array_copy_to_user(effective, prog_ids, cnt);
} else {
struct bpf_prog_list *pl;
u32 id;
i = 0;
list_for_each_entry(pl, progs, node) {
id = pl->prog->aux->id;
if (copy_to_user(prog_ids + i, &id, sizeof(id)))
return -EFAULT;
if (++i == cnt)
break;
}
}
return ret;
}
int cgroup_bpf_prog_attach(const union bpf_attr *attr,
enum bpf_prog_type ptype, struct bpf_prog *prog)
{
struct cgroup *cgrp;
int ret;
cgrp = cgroup_get_from_fd(attr->target_fd);
if (IS_ERR(cgrp))
return PTR_ERR(cgrp);
ret = cgroup_bpf_attach(cgrp, prog, attr->attach_type,
attr->attach_flags);
cgroup_put(cgrp);
return ret;
}
int cgroup_bpf_prog_detach(const union bpf_attr *attr, enum bpf_prog_type ptype)
{
struct bpf_prog *prog;
struct cgroup *cgrp;
int ret;
cgrp = cgroup_get_from_fd(attr->target_fd);
if (IS_ERR(cgrp))
return PTR_ERR(cgrp);
prog = bpf_prog_get_type(attr->attach_bpf_fd, ptype);
if (IS_ERR(prog))
prog = NULL;
ret = cgroup_bpf_detach(cgrp, prog, attr->attach_type, 0);
if (prog)
bpf_prog_put(prog);
cgroup_put(cgrp);
return ret;
}
int cgroup_bpf_prog_query(const union bpf_attr *attr,
union bpf_attr __user *uattr)
{
struct cgroup *cgrp;
int ret;
cgrp = cgroup_get_from_fd(attr->query.target_fd);
if (IS_ERR(cgrp))
return PTR_ERR(cgrp);
ret = cgroup_bpf_query(cgrp, attr, uattr);
cgroup_put(cgrp);
return ret;
}
/**
* __cgroup_bpf_run_filter_skb() - Run a program for packet filtering
* @sk: The socket sending or receiving traffic
* @skb: The skb that is being sent or received
* @type: The type of program to be exectuted
*
* If no socket is passed, or the socket is not of type INET or INET6,
* this function does nothing and returns 0.
*
* The program type passed in via @type must be suitable for network
* filtering. No further check is performed to assert that.
*
* For egress packets, this function can return:
* NET_XMIT_SUCCESS (0) - continue with packet output
* NET_XMIT_DROP (1) - drop packet and notify TCP to call cwr
* NET_XMIT_CN (2) - continue with packet output and notify TCP
* to call cwr
* -EPERM - drop packet
*
* For ingress packets, this function will return -EPERM if any
* attached program was found and if it returned != 1 during execution.
* Otherwise 0 is returned.
*/
int __cgroup_bpf_run_filter_skb(struct sock *sk,
struct sk_buff *skb,
enum bpf_attach_type type)
{
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
unsigned int offset = skb->data - skb_network_header(skb);
struct sock *save_sk;
void *saved_data_end;
struct cgroup *cgrp;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
int ret;
if (!sk || !sk_fullsock(sk))
return 0;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
if (sk->sk_family != AF_INET && sk->sk_family != AF_INET6)
return 0;
cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
save_sk = skb->sk;
skb->sk = sk;
__skb_push(skb, offset);
/* compute pointers for the bpf prog */
bpf_compute_and_save_data_end(skb, &saved_data_end);
if (type == BPF_CGROUP_INET_EGRESS) {
ret = BPF_PROG_CGROUP_INET_EGRESS_RUN_ARRAY(
cgrp->bpf.effective[type], skb, __bpf_prog_run_save_cb);
} else {
ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], skb,
__bpf_prog_run_save_cb);
ret = (ret == 1 ? 0 : -EPERM);
}
bpf_restore_data_end(skb, saved_data_end);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
__skb_pull(skb, offset);
skb->sk = save_sk;
return ret;
}
EXPORT_SYMBOL(__cgroup_bpf_run_filter_skb);
/**
* __cgroup_bpf_run_filter_sk() - Run a program on a sock
* @sk: sock structure to manipulate
* @type: The type of program to be exectuted
*
* socket is passed is expected to be of type INET or INET6.
*
* The program type passed in via @type must be suitable for sock
* filtering. No further check is performed to assert that.
*
* This function will return %-EPERM if any if an attached program was found
* and if it returned != 1 during execution. In all other cases, 0 is returned.
*/
int __cgroup_bpf_run_filter_sk(struct sock *sk,
enum bpf_attach_type type)
{
struct cgroup *cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
int ret;
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], sk, BPF_PROG_RUN);
return ret == 1 ? 0 : -EPERM;
}
EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
bpf: BPF support for sock_ops Created a new BPF program type, BPF_PROG_TYPE_SOCK_OPS, and a corresponding struct that allows BPF programs of this type to access some of the socket's fields (such as IP addresses, ports, etc.). It uses the existing bpf cgroups infrastructure so the programs can be attached per cgroup with full inheritance support. The program will be called at appropriate times to set relevant connections parameters such as buffer sizes, SYN and SYN-ACK RTOs, etc., based on connection information such as IP addresses, port numbers, etc. Alghough there are already 3 mechanisms to set parameters (sysctls, route metrics and setsockopts), this new mechanism provides some distinct advantages. Unlike sysctls, it can set parameters per connection. In contrast to route metrics, it can also use port numbers and information provided by a user level program. In addition, it could set parameters probabilistically for evaluation purposes (i.e. do something different on 10% of the flows and compare results with the other 90% of the flows). Also, in cases where IPv6 addresses contain geographic information, the rules to make changes based on the distance (or RTT) between the hosts are much easier than route metric rules and can be global. Finally, unlike setsockopt, it oes not require application changes and it can be updated easily at any time. Although the bpf cgroup framework already contains a sock related program type (BPF_PROG_TYPE_CGROUP_SOCK), I created the new type (BPF_PROG_TYPE_SOCK_OPS) beccause the existing type expects to be called only once during the connections's lifetime. In contrast, the new program type will be called multiple times from different places in the network stack code. For example, before sending SYN and SYN-ACKs to set an appropriate timeout, when the connection is established to set congestion control, etc. As a result it has "op" field to specify the type of operation requested. The purpose of this new program type is to simplify setting connection parameters, such as buffer sizes, TCP's SYN RTO, etc. For example, it is easy to use facebook's internal IPv6 addresses to determine if both hosts of a connection are in the same datacenter. Therefore, it is easy to write a BPF program to choose a small SYN RTO value when both hosts are in the same datacenter. This patch only contains the framework to support the new BPF program type, following patches add the functionality to set various connection parameters. This patch defines a new BPF program type: BPF_PROG_TYPE_SOCKET_OPS and a new bpf syscall command to load a new program of this type: BPF_PROG_LOAD_SOCKET_OPS. Two new corresponding structs (one for the kernel one for the user/BPF program): /* kernel version */ struct bpf_sock_ops_kern { struct sock *sk; __u32 op; union { __u32 reply; __u32 replylong[4]; }; }; /* user version * Some fields are in network byte order reflecting the sock struct * Use the bpf_ntohl helper macro in samples/bpf/bpf_endian.h to * convert them to host byte order. */ struct bpf_sock_ops { __u32 op; union { __u32 reply; __u32 replylong[4]; }; __u32 family; __u32 remote_ip4; /* In network byte order */ __u32 local_ip4; /* In network byte order */ __u32 remote_ip6[4]; /* In network byte order */ __u32 local_ip6[4]; /* In network byte order */ __u32 remote_port; /* In network byte order */ __u32 local_port; /* In host byte horder */ }; Currently there are two types of ops. The first type expects the BPF program to return a value which is then used by the caller (or a negative value to indicate the operation is not supported). The second type expects state changes to be done by the BPF program, for example through a setsockopt BPF helper function, and they ignore the return value. The reply fields of the bpf_sockt_ops struct are there in case a bpf program needs to return a value larger than an integer. Signed-off-by: Lawrence Brakmo <brakmo@fb.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-07-01 10:02:40 +07:00
bpf: Hooks for sys_bind == The problem == There is a use-case when all processes inside a cgroup should use one single IP address on a host that has multiple IP configured. Those processes should use the IP for both ingress and egress, for TCP and UDP traffic. So TCP/UDP servers should be bound to that IP to accept incoming connections on it, and TCP/UDP clients should make outgoing connections from that IP. It should not require changing application code since it's often not possible. Currently it's solved by intercepting glibc wrappers around syscalls such as `bind(2)` and `connect(2)`. It's done by a shared library that is preloaded for every process in a cgroup so that whenever TCP/UDP server calls `bind(2)`, the library replaces IP in sockaddr before passing arguments to syscall. When application calls `connect(2)` the library transparently binds the local end of connection to that IP (`bind(2)` with `IP_BIND_ADDRESS_NO_PORT` to avoid performance penalty). Shared library approach is fragile though, e.g.: * some applications clear env vars (incl. `LD_PRELOAD`); * `/etc/ld.so.preload` doesn't help since some applications are linked with option `-z nodefaultlib`; * other applications don't use glibc and there is nothing to intercept. == The solution == The patch provides much more reliable in-kernel solution for the 1st part of the problem: binding TCP/UDP servers on desired IP. It does not depend on application environment and implementation details (whether glibc is used or not). It adds new eBPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` and attach types `BPF_CGROUP_INET4_BIND` and `BPF_CGROUP_INET6_BIND` (similar to already existing `BPF_CGROUP_INET_SOCK_CREATE`). The new program type is intended to be used with sockets (`struct sock`) in a cgroup and provided by user `struct sockaddr`. Pointers to both of them are parts of the context passed to programs of newly added types. The new attach types provides hooks in `bind(2)` system call for both IPv4 and IPv6 so that one can write a program to override IP addresses and ports user program tries to bind to and apply such a program for whole cgroup. == Implementation notes == [1] Separate attach types for `AF_INET` and `AF_INET6` are added intentionally to prevent reading/writing to offsets that don't make sense for corresponding socket family. E.g. if user passes `sockaddr_in` it doesn't make sense to read from / write to `user_ip6[]` context fields. [2] The write access to `struct bpf_sock_addr_kern` is implemented using special field as an additional "register". There are just two registers in `sock_addr_convert_ctx_access`: `src` with value to write and `dst` with pointer to context that can't be changed not to break later instructions. But the fields, allowed to write to, are not available directly and to access them address of corresponding pointer has to be loaded first. To get additional register the 1st not used by `src` and `dst` one is taken, its content is saved to `bpf_sock_addr_kern.tmp_reg`, then the register is used to load address of pointer field, and finally the register's content is restored from the temporary field after writing `src` value. Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-31 05:08:02 +07:00
/**
* __cgroup_bpf_run_filter_sock_addr() - Run a program on a sock and
* provided by user sockaddr
* @sk: sock struct that will use sockaddr
* @uaddr: sockaddr struct provided by user
* @type: The type of program to be exectuted
bpf: Hooks for sys_sendmsg In addition to already existing BPF hooks for sys_bind and sys_connect, the patch provides new hooks for sys_sendmsg. It leverages existing BPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` that provides access to socket itlself (properties like family, type, protocol) and user-passed `struct sockaddr *` so that BPF program can override destination IP and port for system calls such as sendto(2) or sendmsg(2) and/or assign source IP to the socket. The hooks are implemented as two new attach types: `BPF_CGROUP_UDP4_SENDMSG` and `BPF_CGROUP_UDP6_SENDMSG` for UDPv4 and UDPv6 correspondingly. UDPv4 and UDPv6 separate attach types for same reason as sys_bind and sys_connect hooks, i.e. to prevent reading from / writing to e.g. user_ip6 fields when user passes sockaddr_in since it'd be out-of-bound. The difference with already existing hooks is sys_sendmsg are implemented only for unconnected UDP. For TCP it doesn't make sense to change user-provided `struct sockaddr *` at sendto(2)/sendmsg(2) time since socket either was already connected and has source/destination set or wasn't connected and call to sendto(2)/sendmsg(2) would lead to ENOTCONN anyway. Connected UDP is already handled by sys_connect hooks that can override source/destination at connect time and use fast-path later, i.e. these hooks don't affect UDP fast-path. Rewriting source IP is implemented differently than that in sys_connect hooks. When sys_sendmsg is used with unconnected UDP it doesn't work to just bind socket to desired local IP address since source IP can be set on per-packet basis by using ancillary data (cmsg(3)). So no matter if socket is bound or not, source IP has to be rewritten on every call to sys_sendmsg. To do so two new fields are added to UAPI `struct bpf_sock_addr`; * `msg_src_ip4` to set source IPv4 for UDPv4; * `msg_src_ip6` to set source IPv6 for UDPv6. Signed-off-by: Andrey Ignatov <rdna@fb.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-25 22:55:23 +07:00
* @t_ctx: Pointer to attach type specific context
bpf: Hooks for sys_bind == The problem == There is a use-case when all processes inside a cgroup should use one single IP address on a host that has multiple IP configured. Those processes should use the IP for both ingress and egress, for TCP and UDP traffic. So TCP/UDP servers should be bound to that IP to accept incoming connections on it, and TCP/UDP clients should make outgoing connections from that IP. It should not require changing application code since it's often not possible. Currently it's solved by intercepting glibc wrappers around syscalls such as `bind(2)` and `connect(2)`. It's done by a shared library that is preloaded for every process in a cgroup so that whenever TCP/UDP server calls `bind(2)`, the library replaces IP in sockaddr before passing arguments to syscall. When application calls `connect(2)` the library transparently binds the local end of connection to that IP (`bind(2)` with `IP_BIND_ADDRESS_NO_PORT` to avoid performance penalty). Shared library approach is fragile though, e.g.: * some applications clear env vars (incl. `LD_PRELOAD`); * `/etc/ld.so.preload` doesn't help since some applications are linked with option `-z nodefaultlib`; * other applications don't use glibc and there is nothing to intercept. == The solution == The patch provides much more reliable in-kernel solution for the 1st part of the problem: binding TCP/UDP servers on desired IP. It does not depend on application environment and implementation details (whether glibc is used or not). It adds new eBPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` and attach types `BPF_CGROUP_INET4_BIND` and `BPF_CGROUP_INET6_BIND` (similar to already existing `BPF_CGROUP_INET_SOCK_CREATE`). The new program type is intended to be used with sockets (`struct sock`) in a cgroup and provided by user `struct sockaddr`. Pointers to both of them are parts of the context passed to programs of newly added types. The new attach types provides hooks in `bind(2)` system call for both IPv4 and IPv6 so that one can write a program to override IP addresses and ports user program tries to bind to and apply such a program for whole cgroup. == Implementation notes == [1] Separate attach types for `AF_INET` and `AF_INET6` are added intentionally to prevent reading/writing to offsets that don't make sense for corresponding socket family. E.g. if user passes `sockaddr_in` it doesn't make sense to read from / write to `user_ip6[]` context fields. [2] The write access to `struct bpf_sock_addr_kern` is implemented using special field as an additional "register". There are just two registers in `sock_addr_convert_ctx_access`: `src` with value to write and `dst` with pointer to context that can't be changed not to break later instructions. But the fields, allowed to write to, are not available directly and to access them address of corresponding pointer has to be loaded first. To get additional register the 1st not used by `src` and `dst` one is taken, its content is saved to `bpf_sock_addr_kern.tmp_reg`, then the register is used to load address of pointer field, and finally the register's content is restored from the temporary field after writing `src` value. Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-31 05:08:02 +07:00
*
* socket is expected to be of type INET or INET6.
*
* This function will return %-EPERM if an attached program is found and
* returned value != 1 during execution. In all other cases, 0 is returned.
*/
int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
struct sockaddr *uaddr,
bpf: Hooks for sys_sendmsg In addition to already existing BPF hooks for sys_bind and sys_connect, the patch provides new hooks for sys_sendmsg. It leverages existing BPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` that provides access to socket itlself (properties like family, type, protocol) and user-passed `struct sockaddr *` so that BPF program can override destination IP and port for system calls such as sendto(2) or sendmsg(2) and/or assign source IP to the socket. The hooks are implemented as two new attach types: `BPF_CGROUP_UDP4_SENDMSG` and `BPF_CGROUP_UDP6_SENDMSG` for UDPv4 and UDPv6 correspondingly. UDPv4 and UDPv6 separate attach types for same reason as sys_bind and sys_connect hooks, i.e. to prevent reading from / writing to e.g. user_ip6 fields when user passes sockaddr_in since it'd be out-of-bound. The difference with already existing hooks is sys_sendmsg are implemented only for unconnected UDP. For TCP it doesn't make sense to change user-provided `struct sockaddr *` at sendto(2)/sendmsg(2) time since socket either was already connected and has source/destination set or wasn't connected and call to sendto(2)/sendmsg(2) would lead to ENOTCONN anyway. Connected UDP is already handled by sys_connect hooks that can override source/destination at connect time and use fast-path later, i.e. these hooks don't affect UDP fast-path. Rewriting source IP is implemented differently than that in sys_connect hooks. When sys_sendmsg is used with unconnected UDP it doesn't work to just bind socket to desired local IP address since source IP can be set on per-packet basis by using ancillary data (cmsg(3)). So no matter if socket is bound or not, source IP has to be rewritten on every call to sys_sendmsg. To do so two new fields are added to UAPI `struct bpf_sock_addr`; * `msg_src_ip4` to set source IPv4 for UDPv4; * `msg_src_ip6` to set source IPv6 for UDPv6. Signed-off-by: Andrey Ignatov <rdna@fb.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-25 22:55:23 +07:00
enum bpf_attach_type type,
void *t_ctx)
bpf: Hooks for sys_bind == The problem == There is a use-case when all processes inside a cgroup should use one single IP address on a host that has multiple IP configured. Those processes should use the IP for both ingress and egress, for TCP and UDP traffic. So TCP/UDP servers should be bound to that IP to accept incoming connections on it, and TCP/UDP clients should make outgoing connections from that IP. It should not require changing application code since it's often not possible. Currently it's solved by intercepting glibc wrappers around syscalls such as `bind(2)` and `connect(2)`. It's done by a shared library that is preloaded for every process in a cgroup so that whenever TCP/UDP server calls `bind(2)`, the library replaces IP in sockaddr before passing arguments to syscall. When application calls `connect(2)` the library transparently binds the local end of connection to that IP (`bind(2)` with `IP_BIND_ADDRESS_NO_PORT` to avoid performance penalty). Shared library approach is fragile though, e.g.: * some applications clear env vars (incl. `LD_PRELOAD`); * `/etc/ld.so.preload` doesn't help since some applications are linked with option `-z nodefaultlib`; * other applications don't use glibc and there is nothing to intercept. == The solution == The patch provides much more reliable in-kernel solution for the 1st part of the problem: binding TCP/UDP servers on desired IP. It does not depend on application environment and implementation details (whether glibc is used or not). It adds new eBPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` and attach types `BPF_CGROUP_INET4_BIND` and `BPF_CGROUP_INET6_BIND` (similar to already existing `BPF_CGROUP_INET_SOCK_CREATE`). The new program type is intended to be used with sockets (`struct sock`) in a cgroup and provided by user `struct sockaddr`. Pointers to both of them are parts of the context passed to programs of newly added types. The new attach types provides hooks in `bind(2)` system call for both IPv4 and IPv6 so that one can write a program to override IP addresses and ports user program tries to bind to and apply such a program for whole cgroup. == Implementation notes == [1] Separate attach types for `AF_INET` and `AF_INET6` are added intentionally to prevent reading/writing to offsets that don't make sense for corresponding socket family. E.g. if user passes `sockaddr_in` it doesn't make sense to read from / write to `user_ip6[]` context fields. [2] The write access to `struct bpf_sock_addr_kern` is implemented using special field as an additional "register". There are just two registers in `sock_addr_convert_ctx_access`: `src` with value to write and `dst` with pointer to context that can't be changed not to break later instructions. But the fields, allowed to write to, are not available directly and to access them address of corresponding pointer has to be loaded first. To get additional register the 1st not used by `src` and `dst` one is taken, its content is saved to `bpf_sock_addr_kern.tmp_reg`, then the register is used to load address of pointer field, and finally the register's content is restored from the temporary field after writing `src` value. Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-31 05:08:02 +07:00
{
struct bpf_sock_addr_kern ctx = {
.sk = sk,
.uaddr = uaddr,
bpf: Hooks for sys_sendmsg In addition to already existing BPF hooks for sys_bind and sys_connect, the patch provides new hooks for sys_sendmsg. It leverages existing BPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` that provides access to socket itlself (properties like family, type, protocol) and user-passed `struct sockaddr *` so that BPF program can override destination IP and port for system calls such as sendto(2) or sendmsg(2) and/or assign source IP to the socket. The hooks are implemented as two new attach types: `BPF_CGROUP_UDP4_SENDMSG` and `BPF_CGROUP_UDP6_SENDMSG` for UDPv4 and UDPv6 correspondingly. UDPv4 and UDPv6 separate attach types for same reason as sys_bind and sys_connect hooks, i.e. to prevent reading from / writing to e.g. user_ip6 fields when user passes sockaddr_in since it'd be out-of-bound. The difference with already existing hooks is sys_sendmsg are implemented only for unconnected UDP. For TCP it doesn't make sense to change user-provided `struct sockaddr *` at sendto(2)/sendmsg(2) time since socket either was already connected and has source/destination set or wasn't connected and call to sendto(2)/sendmsg(2) would lead to ENOTCONN anyway. Connected UDP is already handled by sys_connect hooks that can override source/destination at connect time and use fast-path later, i.e. these hooks don't affect UDP fast-path. Rewriting source IP is implemented differently than that in sys_connect hooks. When sys_sendmsg is used with unconnected UDP it doesn't work to just bind socket to desired local IP address since source IP can be set on per-packet basis by using ancillary data (cmsg(3)). So no matter if socket is bound or not, source IP has to be rewritten on every call to sys_sendmsg. To do so two new fields are added to UAPI `struct bpf_sock_addr`; * `msg_src_ip4` to set source IPv4 for UDPv4; * `msg_src_ip6` to set source IPv6 for UDPv6. Signed-off-by: Andrey Ignatov <rdna@fb.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-25 22:55:23 +07:00
.t_ctx = t_ctx,
bpf: Hooks for sys_bind == The problem == There is a use-case when all processes inside a cgroup should use one single IP address on a host that has multiple IP configured. Those processes should use the IP for both ingress and egress, for TCP and UDP traffic. So TCP/UDP servers should be bound to that IP to accept incoming connections on it, and TCP/UDP clients should make outgoing connections from that IP. It should not require changing application code since it's often not possible. Currently it's solved by intercepting glibc wrappers around syscalls such as `bind(2)` and `connect(2)`. It's done by a shared library that is preloaded for every process in a cgroup so that whenever TCP/UDP server calls `bind(2)`, the library replaces IP in sockaddr before passing arguments to syscall. When application calls `connect(2)` the library transparently binds the local end of connection to that IP (`bind(2)` with `IP_BIND_ADDRESS_NO_PORT` to avoid performance penalty). Shared library approach is fragile though, e.g.: * some applications clear env vars (incl. `LD_PRELOAD`); * `/etc/ld.so.preload` doesn't help since some applications are linked with option `-z nodefaultlib`; * other applications don't use glibc and there is nothing to intercept. == The solution == The patch provides much more reliable in-kernel solution for the 1st part of the problem: binding TCP/UDP servers on desired IP. It does not depend on application environment and implementation details (whether glibc is used or not). It adds new eBPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` and attach types `BPF_CGROUP_INET4_BIND` and `BPF_CGROUP_INET6_BIND` (similar to already existing `BPF_CGROUP_INET_SOCK_CREATE`). The new program type is intended to be used with sockets (`struct sock`) in a cgroup and provided by user `struct sockaddr`. Pointers to both of them are parts of the context passed to programs of newly added types. The new attach types provides hooks in `bind(2)` system call for both IPv4 and IPv6 so that one can write a program to override IP addresses and ports user program tries to bind to and apply such a program for whole cgroup. == Implementation notes == [1] Separate attach types for `AF_INET` and `AF_INET6` are added intentionally to prevent reading/writing to offsets that don't make sense for corresponding socket family. E.g. if user passes `sockaddr_in` it doesn't make sense to read from / write to `user_ip6[]` context fields. [2] The write access to `struct bpf_sock_addr_kern` is implemented using special field as an additional "register". There are just two registers in `sock_addr_convert_ctx_access`: `src` with value to write and `dst` with pointer to context that can't be changed not to break later instructions. But the fields, allowed to write to, are not available directly and to access them address of corresponding pointer has to be loaded first. To get additional register the 1st not used by `src` and `dst` one is taken, its content is saved to `bpf_sock_addr_kern.tmp_reg`, then the register is used to load address of pointer field, and finally the register's content is restored from the temporary field after writing `src` value. Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-31 05:08:02 +07:00
};
bpf: Hooks for sys_sendmsg In addition to already existing BPF hooks for sys_bind and sys_connect, the patch provides new hooks for sys_sendmsg. It leverages existing BPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` that provides access to socket itlself (properties like family, type, protocol) and user-passed `struct sockaddr *` so that BPF program can override destination IP and port for system calls such as sendto(2) or sendmsg(2) and/or assign source IP to the socket. The hooks are implemented as two new attach types: `BPF_CGROUP_UDP4_SENDMSG` and `BPF_CGROUP_UDP6_SENDMSG` for UDPv4 and UDPv6 correspondingly. UDPv4 and UDPv6 separate attach types for same reason as sys_bind and sys_connect hooks, i.e. to prevent reading from / writing to e.g. user_ip6 fields when user passes sockaddr_in since it'd be out-of-bound. The difference with already existing hooks is sys_sendmsg are implemented only for unconnected UDP. For TCP it doesn't make sense to change user-provided `struct sockaddr *` at sendto(2)/sendmsg(2) time since socket either was already connected and has source/destination set or wasn't connected and call to sendto(2)/sendmsg(2) would lead to ENOTCONN anyway. Connected UDP is already handled by sys_connect hooks that can override source/destination at connect time and use fast-path later, i.e. these hooks don't affect UDP fast-path. Rewriting source IP is implemented differently than that in sys_connect hooks. When sys_sendmsg is used with unconnected UDP it doesn't work to just bind socket to desired local IP address since source IP can be set on per-packet basis by using ancillary data (cmsg(3)). So no matter if socket is bound or not, source IP has to be rewritten on every call to sys_sendmsg. To do so two new fields are added to UAPI `struct bpf_sock_addr`; * `msg_src_ip4` to set source IPv4 for UDPv4; * `msg_src_ip6` to set source IPv6 for UDPv6. Signed-off-by: Andrey Ignatov <rdna@fb.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-25 22:55:23 +07:00
struct sockaddr_storage unspec;
bpf: Hooks for sys_bind == The problem == There is a use-case when all processes inside a cgroup should use one single IP address on a host that has multiple IP configured. Those processes should use the IP for both ingress and egress, for TCP and UDP traffic. So TCP/UDP servers should be bound to that IP to accept incoming connections on it, and TCP/UDP clients should make outgoing connections from that IP. It should not require changing application code since it's often not possible. Currently it's solved by intercepting glibc wrappers around syscalls such as `bind(2)` and `connect(2)`. It's done by a shared library that is preloaded for every process in a cgroup so that whenever TCP/UDP server calls `bind(2)`, the library replaces IP in sockaddr before passing arguments to syscall. When application calls `connect(2)` the library transparently binds the local end of connection to that IP (`bind(2)` with `IP_BIND_ADDRESS_NO_PORT` to avoid performance penalty). Shared library approach is fragile though, e.g.: * some applications clear env vars (incl. `LD_PRELOAD`); * `/etc/ld.so.preload` doesn't help since some applications are linked with option `-z nodefaultlib`; * other applications don't use glibc and there is nothing to intercept. == The solution == The patch provides much more reliable in-kernel solution for the 1st part of the problem: binding TCP/UDP servers on desired IP. It does not depend on application environment and implementation details (whether glibc is used or not). It adds new eBPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` and attach types `BPF_CGROUP_INET4_BIND` and `BPF_CGROUP_INET6_BIND` (similar to already existing `BPF_CGROUP_INET_SOCK_CREATE`). The new program type is intended to be used with sockets (`struct sock`) in a cgroup and provided by user `struct sockaddr`. Pointers to both of them are parts of the context passed to programs of newly added types. The new attach types provides hooks in `bind(2)` system call for both IPv4 and IPv6 so that one can write a program to override IP addresses and ports user program tries to bind to and apply such a program for whole cgroup. == Implementation notes == [1] Separate attach types for `AF_INET` and `AF_INET6` are added intentionally to prevent reading/writing to offsets that don't make sense for corresponding socket family. E.g. if user passes `sockaddr_in` it doesn't make sense to read from / write to `user_ip6[]` context fields. [2] The write access to `struct bpf_sock_addr_kern` is implemented using special field as an additional "register". There are just two registers in `sock_addr_convert_ctx_access`: `src` with value to write and `dst` with pointer to context that can't be changed not to break later instructions. But the fields, allowed to write to, are not available directly and to access them address of corresponding pointer has to be loaded first. To get additional register the 1st not used by `src` and `dst` one is taken, its content is saved to `bpf_sock_addr_kern.tmp_reg`, then the register is used to load address of pointer field, and finally the register's content is restored from the temporary field after writing `src` value. Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-31 05:08:02 +07:00
struct cgroup *cgrp;
int ret;
/* Check socket family since not all sockets represent network
* endpoint (e.g. AF_UNIX).
*/
if (sk->sk_family != AF_INET && sk->sk_family != AF_INET6)
return 0;
bpf: Hooks for sys_sendmsg In addition to already existing BPF hooks for sys_bind and sys_connect, the patch provides new hooks for sys_sendmsg. It leverages existing BPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` that provides access to socket itlself (properties like family, type, protocol) and user-passed `struct sockaddr *` so that BPF program can override destination IP and port for system calls such as sendto(2) or sendmsg(2) and/or assign source IP to the socket. The hooks are implemented as two new attach types: `BPF_CGROUP_UDP4_SENDMSG` and `BPF_CGROUP_UDP6_SENDMSG` for UDPv4 and UDPv6 correspondingly. UDPv4 and UDPv6 separate attach types for same reason as sys_bind and sys_connect hooks, i.e. to prevent reading from / writing to e.g. user_ip6 fields when user passes sockaddr_in since it'd be out-of-bound. The difference with already existing hooks is sys_sendmsg are implemented only for unconnected UDP. For TCP it doesn't make sense to change user-provided `struct sockaddr *` at sendto(2)/sendmsg(2) time since socket either was already connected and has source/destination set or wasn't connected and call to sendto(2)/sendmsg(2) would lead to ENOTCONN anyway. Connected UDP is already handled by sys_connect hooks that can override source/destination at connect time and use fast-path later, i.e. these hooks don't affect UDP fast-path. Rewriting source IP is implemented differently than that in sys_connect hooks. When sys_sendmsg is used with unconnected UDP it doesn't work to just bind socket to desired local IP address since source IP can be set on per-packet basis by using ancillary data (cmsg(3)). So no matter if socket is bound or not, source IP has to be rewritten on every call to sys_sendmsg. To do so two new fields are added to UAPI `struct bpf_sock_addr`; * `msg_src_ip4` to set source IPv4 for UDPv4; * `msg_src_ip6` to set source IPv6 for UDPv6. Signed-off-by: Andrey Ignatov <rdna@fb.com> Acked-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Martin KaFai Lau <kafai@fb.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-05-25 22:55:23 +07:00
if (!ctx.uaddr) {
memset(&unspec, 0, sizeof(unspec));
ctx.uaddr = (struct sockaddr *)&unspec;
}
bpf: Hooks for sys_bind == The problem == There is a use-case when all processes inside a cgroup should use one single IP address on a host that has multiple IP configured. Those processes should use the IP for both ingress and egress, for TCP and UDP traffic. So TCP/UDP servers should be bound to that IP to accept incoming connections on it, and TCP/UDP clients should make outgoing connections from that IP. It should not require changing application code since it's often not possible. Currently it's solved by intercepting glibc wrappers around syscalls such as `bind(2)` and `connect(2)`. It's done by a shared library that is preloaded for every process in a cgroup so that whenever TCP/UDP server calls `bind(2)`, the library replaces IP in sockaddr before passing arguments to syscall. When application calls `connect(2)` the library transparently binds the local end of connection to that IP (`bind(2)` with `IP_BIND_ADDRESS_NO_PORT` to avoid performance penalty). Shared library approach is fragile though, e.g.: * some applications clear env vars (incl. `LD_PRELOAD`); * `/etc/ld.so.preload` doesn't help since some applications are linked with option `-z nodefaultlib`; * other applications don't use glibc and there is nothing to intercept. == The solution == The patch provides much more reliable in-kernel solution for the 1st part of the problem: binding TCP/UDP servers on desired IP. It does not depend on application environment and implementation details (whether glibc is used or not). It adds new eBPF program type `BPF_PROG_TYPE_CGROUP_SOCK_ADDR` and attach types `BPF_CGROUP_INET4_BIND` and `BPF_CGROUP_INET6_BIND` (similar to already existing `BPF_CGROUP_INET_SOCK_CREATE`). The new program type is intended to be used with sockets (`struct sock`) in a cgroup and provided by user `struct sockaddr`. Pointers to both of them are parts of the context passed to programs of newly added types. The new attach types provides hooks in `bind(2)` system call for both IPv4 and IPv6 so that one can write a program to override IP addresses and ports user program tries to bind to and apply such a program for whole cgroup. == Implementation notes == [1] Separate attach types for `AF_INET` and `AF_INET6` are added intentionally to prevent reading/writing to offsets that don't make sense for corresponding socket family. E.g. if user passes `sockaddr_in` it doesn't make sense to read from / write to `user_ip6[]` context fields. [2] The write access to `struct bpf_sock_addr_kern` is implemented using special field as an additional "register". There are just two registers in `sock_addr_convert_ctx_access`: `src` with value to write and `dst` with pointer to context that can't be changed not to break later instructions. But the fields, allowed to write to, are not available directly and to access them address of corresponding pointer has to be loaded first. To get additional register the 1st not used by `src` and `dst` one is taken, its content is saved to `bpf_sock_addr_kern.tmp_reg`, then the register is used to load address of pointer field, and finally the register's content is restored from the temporary field after writing `src` value. Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-31 05:08:02 +07:00
cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], &ctx, BPF_PROG_RUN);
return ret == 1 ? 0 : -EPERM;
}
EXPORT_SYMBOL(__cgroup_bpf_run_filter_sock_addr);
bpf: BPF support for sock_ops Created a new BPF program type, BPF_PROG_TYPE_SOCK_OPS, and a corresponding struct that allows BPF programs of this type to access some of the socket's fields (such as IP addresses, ports, etc.). It uses the existing bpf cgroups infrastructure so the programs can be attached per cgroup with full inheritance support. The program will be called at appropriate times to set relevant connections parameters such as buffer sizes, SYN and SYN-ACK RTOs, etc., based on connection information such as IP addresses, port numbers, etc. Alghough there are already 3 mechanisms to set parameters (sysctls, route metrics and setsockopts), this new mechanism provides some distinct advantages. Unlike sysctls, it can set parameters per connection. In contrast to route metrics, it can also use port numbers and information provided by a user level program. In addition, it could set parameters probabilistically for evaluation purposes (i.e. do something different on 10% of the flows and compare results with the other 90% of the flows). Also, in cases where IPv6 addresses contain geographic information, the rules to make changes based on the distance (or RTT) between the hosts are much easier than route metric rules and can be global. Finally, unlike setsockopt, it oes not require application changes and it can be updated easily at any time. Although the bpf cgroup framework already contains a sock related program type (BPF_PROG_TYPE_CGROUP_SOCK), I created the new type (BPF_PROG_TYPE_SOCK_OPS) beccause the existing type expects to be called only once during the connections's lifetime. In contrast, the new program type will be called multiple times from different places in the network stack code. For example, before sending SYN and SYN-ACKs to set an appropriate timeout, when the connection is established to set congestion control, etc. As a result it has "op" field to specify the type of operation requested. The purpose of this new program type is to simplify setting connection parameters, such as buffer sizes, TCP's SYN RTO, etc. For example, it is easy to use facebook's internal IPv6 addresses to determine if both hosts of a connection are in the same datacenter. Therefore, it is easy to write a BPF program to choose a small SYN RTO value when both hosts are in the same datacenter. This patch only contains the framework to support the new BPF program type, following patches add the functionality to set various connection parameters. This patch defines a new BPF program type: BPF_PROG_TYPE_SOCKET_OPS and a new bpf syscall command to load a new program of this type: BPF_PROG_LOAD_SOCKET_OPS. Two new corresponding structs (one for the kernel one for the user/BPF program): /* kernel version */ struct bpf_sock_ops_kern { struct sock *sk; __u32 op; union { __u32 reply; __u32 replylong[4]; }; }; /* user version * Some fields are in network byte order reflecting the sock struct * Use the bpf_ntohl helper macro in samples/bpf/bpf_endian.h to * convert them to host byte order. */ struct bpf_sock_ops { __u32 op; union { __u32 reply; __u32 replylong[4]; }; __u32 family; __u32 remote_ip4; /* In network byte order */ __u32 local_ip4; /* In network byte order */ __u32 remote_ip6[4]; /* In network byte order */ __u32 local_ip6[4]; /* In network byte order */ __u32 remote_port; /* In network byte order */ __u32 local_port; /* In host byte horder */ }; Currently there are two types of ops. The first type expects the BPF program to return a value which is then used by the caller (or a negative value to indicate the operation is not supported). The second type expects state changes to be done by the BPF program, for example through a setsockopt BPF helper function, and they ignore the return value. The reply fields of the bpf_sockt_ops struct are there in case a bpf program needs to return a value larger than an integer. Signed-off-by: Lawrence Brakmo <brakmo@fb.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-07-01 10:02:40 +07:00
/**
* __cgroup_bpf_run_filter_sock_ops() - Run a program on a sock
* @sk: socket to get cgroup from
* @sock_ops: bpf_sock_ops_kern struct to pass to program. Contains
* sk with connection information (IP addresses, etc.) May not contain
* cgroup info if it is a req sock.
* @type: The type of program to be exectuted
*
* socket passed is expected to be of type INET or INET6.
*
* The program type passed in via @type must be suitable for sock_ops
* filtering. No further check is performed to assert that.
*
* This function will return %-EPERM if any if an attached program was found
* and if it returned != 1 during execution. In all other cases, 0 is returned.
*/
int __cgroup_bpf_run_filter_sock_ops(struct sock *sk,
struct bpf_sock_ops_kern *sock_ops,
enum bpf_attach_type type)
{
struct cgroup *cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
int ret;
bpf: BPF support for sock_ops Created a new BPF program type, BPF_PROG_TYPE_SOCK_OPS, and a corresponding struct that allows BPF programs of this type to access some of the socket's fields (such as IP addresses, ports, etc.). It uses the existing bpf cgroups infrastructure so the programs can be attached per cgroup with full inheritance support. The program will be called at appropriate times to set relevant connections parameters such as buffer sizes, SYN and SYN-ACK RTOs, etc., based on connection information such as IP addresses, port numbers, etc. Alghough there are already 3 mechanisms to set parameters (sysctls, route metrics and setsockopts), this new mechanism provides some distinct advantages. Unlike sysctls, it can set parameters per connection. In contrast to route metrics, it can also use port numbers and information provided by a user level program. In addition, it could set parameters probabilistically for evaluation purposes (i.e. do something different on 10% of the flows and compare results with the other 90% of the flows). Also, in cases where IPv6 addresses contain geographic information, the rules to make changes based on the distance (or RTT) between the hosts are much easier than route metric rules and can be global. Finally, unlike setsockopt, it oes not require application changes and it can be updated easily at any time. Although the bpf cgroup framework already contains a sock related program type (BPF_PROG_TYPE_CGROUP_SOCK), I created the new type (BPF_PROG_TYPE_SOCK_OPS) beccause the existing type expects to be called only once during the connections's lifetime. In contrast, the new program type will be called multiple times from different places in the network stack code. For example, before sending SYN and SYN-ACKs to set an appropriate timeout, when the connection is established to set congestion control, etc. As a result it has "op" field to specify the type of operation requested. The purpose of this new program type is to simplify setting connection parameters, such as buffer sizes, TCP's SYN RTO, etc. For example, it is easy to use facebook's internal IPv6 addresses to determine if both hosts of a connection are in the same datacenter. Therefore, it is easy to write a BPF program to choose a small SYN RTO value when both hosts are in the same datacenter. This patch only contains the framework to support the new BPF program type, following patches add the functionality to set various connection parameters. This patch defines a new BPF program type: BPF_PROG_TYPE_SOCKET_OPS and a new bpf syscall command to load a new program of this type: BPF_PROG_LOAD_SOCKET_OPS. Two new corresponding structs (one for the kernel one for the user/BPF program): /* kernel version */ struct bpf_sock_ops_kern { struct sock *sk; __u32 op; union { __u32 reply; __u32 replylong[4]; }; }; /* user version * Some fields are in network byte order reflecting the sock struct * Use the bpf_ntohl helper macro in samples/bpf/bpf_endian.h to * convert them to host byte order. */ struct bpf_sock_ops { __u32 op; union { __u32 reply; __u32 replylong[4]; }; __u32 family; __u32 remote_ip4; /* In network byte order */ __u32 local_ip4; /* In network byte order */ __u32 remote_ip6[4]; /* In network byte order */ __u32 local_ip6[4]; /* In network byte order */ __u32 remote_port; /* In network byte order */ __u32 local_port; /* In host byte horder */ }; Currently there are two types of ops. The first type expects the BPF program to return a value which is then used by the caller (or a negative value to indicate the operation is not supported). The second type expects state changes to be done by the BPF program, for example through a setsockopt BPF helper function, and they ignore the return value. The reply fields of the bpf_sockt_ops struct are there in case a bpf program needs to return a value larger than an integer. Signed-off-by: Lawrence Brakmo <brakmo@fb.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-07-01 10:02:40 +07:00
bpf: multi program support for cgroup+bpf introduce BPF_F_ALLOW_MULTI flag that can be used to attach multiple bpf programs to a cgroup. The difference between three possible flags for BPF_PROG_ATTACH command: - NONE(default): No further bpf programs allowed in the subtree. - BPF_F_ALLOW_OVERRIDE: If a sub-cgroup installs some bpf program, the program in this cgroup yields to sub-cgroup program. - BPF_F_ALLOW_MULTI: If a sub-cgroup installs some bpf program, that cgroup program gets run in addition to the program in this cgroup. NONE and BPF_F_ALLOW_OVERRIDE existed before. This patch doesn't change their behavior. It only clarifies the semantics in relation to new flag. Only one program is allowed to be attached to a cgroup with NONE or BPF_F_ALLOW_OVERRIDE flag. Multiple programs are allowed to be attached to a cgroup with BPF_F_ALLOW_MULTI flag. They are executed in FIFO order (those that were attached first, run first) The programs of sub-cgroup are executed first, then programs of this cgroup and then programs of parent cgroup. All eligible programs are executed regardless of return code from earlier programs. To allow efficient execution of multiple programs attached to a cgroup and to avoid penalizing cgroups without any programs attached introduce 'struct bpf_prog_array' which is RCU protected array of pointers to bpf programs. Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Martin KaFai Lau <kafai@fb.com> for cgroup bits Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-10-03 12:50:21 +07:00
ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], sock_ops,
BPF_PROG_RUN);
return ret == 1 ? 0 : -EPERM;
bpf: BPF support for sock_ops Created a new BPF program type, BPF_PROG_TYPE_SOCK_OPS, and a corresponding struct that allows BPF programs of this type to access some of the socket's fields (such as IP addresses, ports, etc.). It uses the existing bpf cgroups infrastructure so the programs can be attached per cgroup with full inheritance support. The program will be called at appropriate times to set relevant connections parameters such as buffer sizes, SYN and SYN-ACK RTOs, etc., based on connection information such as IP addresses, port numbers, etc. Alghough there are already 3 mechanisms to set parameters (sysctls, route metrics and setsockopts), this new mechanism provides some distinct advantages. Unlike sysctls, it can set parameters per connection. In contrast to route metrics, it can also use port numbers and information provided by a user level program. In addition, it could set parameters probabilistically for evaluation purposes (i.e. do something different on 10% of the flows and compare results with the other 90% of the flows). Also, in cases where IPv6 addresses contain geographic information, the rules to make changes based on the distance (or RTT) between the hosts are much easier than route metric rules and can be global. Finally, unlike setsockopt, it oes not require application changes and it can be updated easily at any time. Although the bpf cgroup framework already contains a sock related program type (BPF_PROG_TYPE_CGROUP_SOCK), I created the new type (BPF_PROG_TYPE_SOCK_OPS) beccause the existing type expects to be called only once during the connections's lifetime. In contrast, the new program type will be called multiple times from different places in the network stack code. For example, before sending SYN and SYN-ACKs to set an appropriate timeout, when the connection is established to set congestion control, etc. As a result it has "op" field to specify the type of operation requested. The purpose of this new program type is to simplify setting connection parameters, such as buffer sizes, TCP's SYN RTO, etc. For example, it is easy to use facebook's internal IPv6 addresses to determine if both hosts of a connection are in the same datacenter. Therefore, it is easy to write a BPF program to choose a small SYN RTO value when both hosts are in the same datacenter. This patch only contains the framework to support the new BPF program type, following patches add the functionality to set various connection parameters. This patch defines a new BPF program type: BPF_PROG_TYPE_SOCKET_OPS and a new bpf syscall command to load a new program of this type: BPF_PROG_LOAD_SOCKET_OPS. Two new corresponding structs (one for the kernel one for the user/BPF program): /* kernel version */ struct bpf_sock_ops_kern { struct sock *sk; __u32 op; union { __u32 reply; __u32 replylong[4]; }; }; /* user version * Some fields are in network byte order reflecting the sock struct * Use the bpf_ntohl helper macro in samples/bpf/bpf_endian.h to * convert them to host byte order. */ struct bpf_sock_ops { __u32 op; union { __u32 reply; __u32 replylong[4]; }; __u32 family; __u32 remote_ip4; /* In network byte order */ __u32 local_ip4; /* In network byte order */ __u32 remote_ip6[4]; /* In network byte order */ __u32 local_ip6[4]; /* In network byte order */ __u32 remote_port; /* In network byte order */ __u32 local_port; /* In host byte horder */ }; Currently there are two types of ops. The first type expects the BPF program to return a value which is then used by the caller (or a negative value to indicate the operation is not supported). The second type expects state changes to be done by the BPF program, for example through a setsockopt BPF helper function, and they ignore the return value. The reply fields of the bpf_sockt_ops struct are there in case a bpf program needs to return a value larger than an integer. Signed-off-by: Lawrence Brakmo <brakmo@fb.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-07-01 10:02:40 +07:00
}
EXPORT_SYMBOL(__cgroup_bpf_run_filter_sock_ops);
int __cgroup_bpf_check_dev_permission(short dev_type, u32 major, u32 minor,
short access, enum bpf_attach_type type)
{
struct cgroup *cgrp;
struct bpf_cgroup_dev_ctx ctx = {
.access_type = (access << 16) | dev_type,
.major = major,
.minor = minor,
};
int allow = 1;
rcu_read_lock();
cgrp = task_dfl_cgroup(current);
allow = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], &ctx,
BPF_PROG_RUN);
rcu_read_unlock();
return !allow;
}
EXPORT_SYMBOL(__cgroup_bpf_check_dev_permission);
static const struct bpf_func_proto *
cgroup_base_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
{
switch (func_id) {
case BPF_FUNC_map_lookup_elem:
return &bpf_map_lookup_elem_proto;
case BPF_FUNC_map_update_elem:
return &bpf_map_update_elem_proto;
case BPF_FUNC_map_delete_elem:
return &bpf_map_delete_elem_proto;
case BPF_FUNC_map_push_elem:
return &bpf_map_push_elem_proto;
case BPF_FUNC_map_pop_elem:
return &bpf_map_pop_elem_proto;
case BPF_FUNC_map_peek_elem:
return &bpf_map_peek_elem_proto;
case BPF_FUNC_get_current_uid_gid:
return &bpf_get_current_uid_gid_proto;
case BPF_FUNC_get_local_storage:
return &bpf_get_local_storage_proto;
case BPF_FUNC_get_current_cgroup_id:
return &bpf_get_current_cgroup_id_proto;
case BPF_FUNC_trace_printk:
if (capable(CAP_SYS_ADMIN))
return bpf_get_trace_printk_proto();
/* fall through */
default:
return NULL;
}
}
static const struct bpf_func_proto *
cgroup_dev_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
{
return cgroup_base_func_proto(func_id, prog);
}
static bool cgroup_dev_is_valid_access(int off, int size,
enum bpf_access_type type,
bpf: Check attach type at prog load time == The problem == There are use-cases when a program of some type can be attached to multiple attach points and those attach points must have different permissions to access context or to call helpers. E.g. context structure may have fields for both IPv4 and IPv6 but it doesn't make sense to read from / write to IPv6 field when attach point is somewhere in IPv4 stack. Same applies to BPF-helpers: it may make sense to call some helper from some attach point, but not from other for same prog type. == The solution == Introduce `expected_attach_type` field in in `struct bpf_attr` for `BPF_PROG_LOAD` command. If scenario described in "The problem" section is the case for some prog type, the field will be checked twice: 1) At load time prog type is checked to see if attach type for it must be known to validate program permissions correctly. Prog will be rejected with EINVAL if it's the case and `expected_attach_type` is not specified or has invalid value. 2) At attach time `attach_type` is compared with `expected_attach_type`, if prog type requires to have one, and, if they differ, attach will be rejected with EINVAL. The `expected_attach_type` is now available as part of `struct bpf_prog` in both `bpf_verifier_ops->is_valid_access()` and `bpf_verifier_ops->get_func_proto()` () and can be used to check context accesses and calls to helpers correspondingly. Initially the idea was discussed by Alexei Starovoitov <ast@fb.com> and Daniel Borkmann <daniel@iogearbox.net> here: https://marc.info/?l=linux-netdev&m=152107378717201&w=2 Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-03-31 05:08:00 +07:00
const struct bpf_prog *prog,
struct bpf_insn_access_aux *info)
{
const int size_default = sizeof(__u32);
if (type == BPF_WRITE)
return false;
if (off < 0 || off + size > sizeof(struct bpf_cgroup_dev_ctx))
return false;
/* The verifier guarantees that size > 0. */
if (off % size != 0)
return false;
switch (off) {
case bpf_ctx_range(struct bpf_cgroup_dev_ctx, access_type):
bpf_ctx_record_field_size(info, size_default);
if (!bpf_ctx_narrow_access_ok(off, size, size_default))
return false;
break;
default:
if (size != size_default)
return false;
}
return true;
}
const struct bpf_prog_ops cg_dev_prog_ops = {
};
const struct bpf_verifier_ops cg_dev_verifier_ops = {
.get_func_proto = cgroup_dev_func_proto,
.is_valid_access = cgroup_dev_is_valid_access,
};
/**
* __cgroup_bpf_run_filter_sysctl - Run a program on sysctl
*
* @head: sysctl table header
* @table: sysctl table
* @write: sysctl is being read (= 0) or written (= 1)
* @buf: pointer to buffer passed by user space
* @pcount: value-result argument: value is size of buffer pointed to by @buf,
* result is size of @new_buf if program set new value, initial value
* otherwise
* @ppos: value-result argument: value is position at which read from or write
* to sysctl is happening, result is new position if program overrode it,
* initial value otherwise
* @new_buf: pointer to pointer to new buffer that will be allocated if program
* overrides new value provided by user space on sysctl write
* NOTE: it's caller responsibility to free *new_buf if it was set
* @type: type of program to be executed
*
* Program is run when sysctl is being accessed, either read or written, and
* can allow or deny such access.
*
* This function will return %-EPERM if an attached program is found and
* returned value != 1 during execution. In all other cases 0 is returned.
*/
int __cgroup_bpf_run_filter_sysctl(struct ctl_table_header *head,
struct ctl_table *table, int write,
void __user *buf, size_t *pcount,
loff_t *ppos, void **new_buf,
enum bpf_attach_type type)
{
struct bpf_sysctl_kern ctx = {
.head = head,
.table = table,
.write = write,
.ppos = ppos,
.cur_val = NULL,
.cur_len = PAGE_SIZE,
.new_val = NULL,
.new_len = 0,
.new_updated = 0,
};
struct cgroup *cgrp;
int ret;
ctx.cur_val = kmalloc_track_caller(ctx.cur_len, GFP_KERNEL);
if (ctx.cur_val) {
mm_segment_t old_fs;
loff_t pos = 0;
old_fs = get_fs();
set_fs(KERNEL_DS);
if (table->proc_handler(table, 0, (void __user *)ctx.cur_val,
&ctx.cur_len, &pos)) {
/* Let BPF program decide how to proceed. */
ctx.cur_len = 0;
}
set_fs(old_fs);
} else {
/* Let BPF program decide how to proceed. */
ctx.cur_len = 0;
}
if (write && buf && *pcount) {
/* BPF program should be able to override new value with a
* buffer bigger than provided by user.
*/
ctx.new_val = kmalloc_track_caller(PAGE_SIZE, GFP_KERNEL);
ctx.new_len = min_t(size_t, PAGE_SIZE, *pcount);
if (!ctx.new_val ||
copy_from_user(ctx.new_val, buf, ctx.new_len))
/* Let BPF program decide how to proceed. */
ctx.new_len = 0;
}
rcu_read_lock();
cgrp = task_dfl_cgroup(current);
ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[type], &ctx, BPF_PROG_RUN);
rcu_read_unlock();
kfree(ctx.cur_val);
if (ret == 1 && ctx.new_updated) {
*new_buf = ctx.new_val;
*pcount = ctx.new_len;
} else {
kfree(ctx.new_val);
}
return ret == 1 ? 0 : -EPERM;
}
EXPORT_SYMBOL(__cgroup_bpf_run_filter_sysctl);
#ifdef CONFIG_NET
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
static bool __cgroup_bpf_prog_array_is_empty(struct cgroup *cgrp,
enum bpf_attach_type attach_type)
{
struct bpf_prog_array *prog_array;
bool empty;
rcu_read_lock();
prog_array = rcu_dereference(cgrp->bpf.effective[attach_type]);
empty = bpf_prog_array_is_empty(prog_array);
rcu_read_unlock();
return empty;
}
static int sockopt_alloc_buf(struct bpf_sockopt_kern *ctx, int max_optlen)
{
if (unlikely(max_optlen > PAGE_SIZE) || max_optlen < 0)
return -EINVAL;
ctx->optval = kzalloc(max_optlen, GFP_USER);
if (!ctx->optval)
return -ENOMEM;
ctx->optval_end = ctx->optval + max_optlen;
return 0;
}
static void sockopt_free_buf(struct bpf_sockopt_kern *ctx)
{
kfree(ctx->optval);
}
int __cgroup_bpf_run_filter_setsockopt(struct sock *sk, int *level,
int *optname, char __user *optval,
int *optlen, char **kernel_optval)
{
struct cgroup *cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
struct bpf_sockopt_kern ctx = {
.sk = sk,
.level = *level,
.optname = *optname,
};
int ret, max_optlen;
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
/* Opportunistic check to see whether we have any BPF program
* attached to the hook so we don't waste time allocating
* memory and locking the socket.
*/
if (!cgroup_bpf_enabled ||
__cgroup_bpf_prog_array_is_empty(cgrp, BPF_CGROUP_SETSOCKOPT))
return 0;
/* Allocate a bit more than the initial user buffer for
* BPF program. The canonical use case is overriding
* TCP_CONGESTION(nv) to TCP_CONGESTION(cubic).
*/
max_optlen = max_t(int, 16, *optlen);
ret = sockopt_alloc_buf(&ctx, max_optlen);
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
if (ret)
return ret;
ctx.optlen = *optlen;
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
if (copy_from_user(ctx.optval, optval, *optlen) != 0) {
ret = -EFAULT;
goto out;
}
lock_sock(sk);
ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[BPF_CGROUP_SETSOCKOPT],
&ctx, BPF_PROG_RUN);
release_sock(sk);
if (!ret) {
ret = -EPERM;
goto out;
}
if (ctx.optlen == -1) {
/* optlen set to -1, bypass kernel */
ret = 1;
} else if (ctx.optlen > max_optlen || ctx.optlen < -1) {
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
/* optlen is out of bounds */
ret = -EFAULT;
} else {
/* optlen within bounds, run kernel handler */
ret = 0;
/* export any potential modifications */
*level = ctx.level;
*optname = ctx.optname;
*optlen = ctx.optlen;
*kernel_optval = ctx.optval;
}
out:
if (ret)
sockopt_free_buf(&ctx);
return ret;
}
EXPORT_SYMBOL(__cgroup_bpf_run_filter_setsockopt);
int __cgroup_bpf_run_filter_getsockopt(struct sock *sk, int level,
int optname, char __user *optval,
int __user *optlen, int max_optlen,
int retval)
{
struct cgroup *cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
struct bpf_sockopt_kern ctx = {
.sk = sk,
.level = level,
.optname = optname,
.retval = retval,
};
int ret;
/* Opportunistic check to see whether we have any BPF program
* attached to the hook so we don't waste time allocating
* memory and locking the socket.
*/
if (!cgroup_bpf_enabled ||
__cgroup_bpf_prog_array_is_empty(cgrp, BPF_CGROUP_GETSOCKOPT))
return retval;
ret = sockopt_alloc_buf(&ctx, max_optlen);
if (ret)
return ret;
ctx.optlen = max_optlen;
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
if (!retval) {
/* If kernel getsockopt finished successfully,
* copy whatever was returned to the user back
* into our temporary buffer. Set optlen to the
* one that kernel returned as well to let
* BPF programs inspect the value.
*/
if (get_user(ctx.optlen, optlen)) {
ret = -EFAULT;
goto out;
}
if (ctx.optlen > max_optlen)
ctx.optlen = max_optlen;
if (copy_from_user(ctx.optval, optval, ctx.optlen) != 0) {
ret = -EFAULT;
goto out;
}
}
lock_sock(sk);
ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[BPF_CGROUP_GETSOCKOPT],
&ctx, BPF_PROG_RUN);
release_sock(sk);
if (!ret) {
ret = -EPERM;
goto out;
}
if (ctx.optlen > max_optlen) {
ret = -EFAULT;
goto out;
}
/* BPF programs only allowed to set retval to 0, not some
* arbitrary value.
*/
if (ctx.retval != 0 && ctx.retval != retval) {
ret = -EFAULT;
goto out;
}
if (copy_to_user(optval, ctx.optval, ctx.optlen) ||
put_user(ctx.optlen, optlen)) {
ret = -EFAULT;
goto out;
}
ret = ctx.retval;
out:
sockopt_free_buf(&ctx);
return ret;
}
EXPORT_SYMBOL(__cgroup_bpf_run_filter_getsockopt);
#endif
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
static ssize_t sysctl_cpy_dir(const struct ctl_dir *dir, char **bufp,
size_t *lenp)
{
ssize_t tmp_ret = 0, ret;
if (dir->header.parent) {
tmp_ret = sysctl_cpy_dir(dir->header.parent, bufp, lenp);
if (tmp_ret < 0)
return tmp_ret;
}
ret = strscpy(*bufp, dir->header.ctl_table[0].procname, *lenp);
if (ret < 0)
return ret;
*bufp += ret;
*lenp -= ret;
ret += tmp_ret;
/* Avoid leading slash. */
if (!ret)
return ret;
tmp_ret = strscpy(*bufp, "/", *lenp);
if (tmp_ret < 0)
return tmp_ret;
*bufp += tmp_ret;
*lenp -= tmp_ret;
return ret + tmp_ret;
}
BPF_CALL_4(bpf_sysctl_get_name, struct bpf_sysctl_kern *, ctx, char *, buf,
size_t, buf_len, u64, flags)
{
ssize_t tmp_ret = 0, ret;
if (!buf)
return -EINVAL;
if (!(flags & BPF_F_SYSCTL_BASE_NAME)) {
if (!ctx->head)
return -EINVAL;
tmp_ret = sysctl_cpy_dir(ctx->head->parent, &buf, &buf_len);
if (tmp_ret < 0)
return tmp_ret;
}
ret = strscpy(buf, ctx->table->procname, buf_len);
return ret < 0 ? ret : tmp_ret + ret;
}
static const struct bpf_func_proto bpf_sysctl_get_name_proto = {
.func = bpf_sysctl_get_name,
.gpl_only = false,
.ret_type = RET_INTEGER,
.arg1_type = ARG_PTR_TO_CTX,
.arg2_type = ARG_PTR_TO_MEM,
.arg3_type = ARG_CONST_SIZE,
.arg4_type = ARG_ANYTHING,
};
static int copy_sysctl_value(char *dst, size_t dst_len, char *src,
size_t src_len)
{
if (!dst)
return -EINVAL;
if (!dst_len)
return -E2BIG;
if (!src || !src_len) {
memset(dst, 0, dst_len);
return -EINVAL;
}
memcpy(dst, src, min(dst_len, src_len));
if (dst_len > src_len) {
memset(dst + src_len, '\0', dst_len - src_len);
return src_len;
}
dst[dst_len - 1] = '\0';
return -E2BIG;
}
BPF_CALL_3(bpf_sysctl_get_current_value, struct bpf_sysctl_kern *, ctx,
char *, buf, size_t, buf_len)
{
return copy_sysctl_value(buf, buf_len, ctx->cur_val, ctx->cur_len);
}
static const struct bpf_func_proto bpf_sysctl_get_current_value_proto = {
.func = bpf_sysctl_get_current_value,
.gpl_only = false,
.ret_type = RET_INTEGER,
.arg1_type = ARG_PTR_TO_CTX,
.arg2_type = ARG_PTR_TO_UNINIT_MEM,
.arg3_type = ARG_CONST_SIZE,
};
BPF_CALL_3(bpf_sysctl_get_new_value, struct bpf_sysctl_kern *, ctx, char *, buf,
size_t, buf_len)
{
if (!ctx->write) {
if (buf && buf_len)
memset(buf, '\0', buf_len);
return -EINVAL;
}
return copy_sysctl_value(buf, buf_len, ctx->new_val, ctx->new_len);
}
static const struct bpf_func_proto bpf_sysctl_get_new_value_proto = {
.func = bpf_sysctl_get_new_value,
.gpl_only = false,
.ret_type = RET_INTEGER,
.arg1_type = ARG_PTR_TO_CTX,
.arg2_type = ARG_PTR_TO_UNINIT_MEM,
.arg3_type = ARG_CONST_SIZE,
};
BPF_CALL_3(bpf_sysctl_set_new_value, struct bpf_sysctl_kern *, ctx,
const char *, buf, size_t, buf_len)
{
if (!ctx->write || !ctx->new_val || !ctx->new_len || !buf || !buf_len)
return -EINVAL;
if (buf_len > PAGE_SIZE - 1)
return -E2BIG;
memcpy(ctx->new_val, buf, buf_len);
ctx->new_len = buf_len;
ctx->new_updated = 1;
return 0;
}
static const struct bpf_func_proto bpf_sysctl_set_new_value_proto = {
.func = bpf_sysctl_set_new_value,
.gpl_only = false,
.ret_type = RET_INTEGER,
.arg1_type = ARG_PTR_TO_CTX,
.arg2_type = ARG_PTR_TO_MEM,
.arg3_type = ARG_CONST_SIZE,
};
static const struct bpf_func_proto *
sysctl_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
{
switch (func_id) {
bpf: Introduce bpf_strtol and bpf_strtoul helpers Add bpf_strtol and bpf_strtoul to convert a string to long and unsigned long correspondingly. It's similar to user space strtol(3) and strtoul(3) with a few changes to the API: * instead of NUL-terminated C string the helpers expect buffer and buffer length; * resulting long or unsigned long is returned in a separate result-argument; * return value is used to indicate success or failure, on success number of consumed bytes is returned that can be used to identify position to read next if the buffer is expected to contain multiple integers; * instead of *base* argument, *flags* is used that provides base in 5 LSB, other bits are reserved for future use; * number of supported bases is limited. Documentation for the new helpers is provided in bpf.h UAPI. The helpers are made available to BPF_PROG_TYPE_CGROUP_SYSCTL programs to be able to convert string input to e.g. "ulongvec" output. E.g. "net/ipv4/tcp_mem" consists of three ulong integers. They can be parsed by calling to bpf_strtoul three times. Implementation notes: Implementation includes "../../lib/kstrtox.h" to reuse integer parsing functions. It's done exactly same way as fs/proc/base.c already does. Unfortunately existing kstrtoX function can't be used directly since they fail if any invalid character is present right after integer in the string. Existing simple_strtoX functions can't be used either since they're obsolete and don't handle overflow properly. Signed-off-by: Andrey Ignatov <rdna@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-03-19 07:55:26 +07:00
case BPF_FUNC_strtol:
return &bpf_strtol_proto;
case BPF_FUNC_strtoul:
return &bpf_strtoul_proto;
case BPF_FUNC_sysctl_get_name:
return &bpf_sysctl_get_name_proto;
case BPF_FUNC_sysctl_get_current_value:
return &bpf_sysctl_get_current_value_proto;
case BPF_FUNC_sysctl_get_new_value:
return &bpf_sysctl_get_new_value_proto;
case BPF_FUNC_sysctl_set_new_value:
return &bpf_sysctl_set_new_value_proto;
default:
return cgroup_base_func_proto(func_id, prog);
}
}
static bool sysctl_is_valid_access(int off, int size, enum bpf_access_type type,
const struct bpf_prog *prog,
struct bpf_insn_access_aux *info)
{
const int size_default = sizeof(__u32);
if (off < 0 || off + size > sizeof(struct bpf_sysctl) || off % size)
return false;
switch (off) {
bpf: Allow narrow loads of bpf_sysctl fields with offset > 0 "ctx:file_pos sysctl:read read ok narrow" works on s390 by accident: it reads the wrong byte, which happens to have the expected value of 0. Improve the test by seeking to the 4th byte and expecting 4 instead of 0. This makes the latent problem apparent: the test attempts to read the first byte of bpf_sysctl.file_pos, assuming this is the least-significant byte, which is not the case on big-endian machines: a non-zero offset is needed. The point of the test is to verify narrow loads, so we cannot cheat our way out by simply using BPF_W. The existence of the test means that such loads have to be supported, most likely because llvm can generate them. Fix the test by adding a big-endian variant, which uses an offset to access the least-significant byte of bpf_sysctl.file_pos. This reveals the final problem: verifier rejects accesses to bpf_sysctl fields with offset > 0. Such accesses are already allowed for a wide range of structs: __sk_buff, bpf_sock_addr and sk_msg_md to name a few. Extend this support to bpf_sysctl by using bpf_ctx_range instead of offsetof when matching field offsets. Fixes: 7b146cebe30c ("bpf: Sysctl hook") Fixes: e1550bfe0de4 ("bpf: Add file_pos field to bpf_sysctl ctx") Fixes: 9a1027e52535 ("selftests/bpf: Test file_pos field in bpf_sysctl ctx") Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrey Ignatov <rdna@fb.com> Acked-by: Andrii Nakryiko <andriin@fb.com> Link: https://lore.kernel.org/bpf/20191028122902.9763-1-iii@linux.ibm.com
2019-10-28 19:29:02 +07:00
case bpf_ctx_range(struct bpf_sysctl, write):
if (type != BPF_READ)
return false;
bpf_ctx_record_field_size(info, size_default);
return bpf_ctx_narrow_access_ok(off, size, size_default);
bpf: Allow narrow loads of bpf_sysctl fields with offset > 0 "ctx:file_pos sysctl:read read ok narrow" works on s390 by accident: it reads the wrong byte, which happens to have the expected value of 0. Improve the test by seeking to the 4th byte and expecting 4 instead of 0. This makes the latent problem apparent: the test attempts to read the first byte of bpf_sysctl.file_pos, assuming this is the least-significant byte, which is not the case on big-endian machines: a non-zero offset is needed. The point of the test is to verify narrow loads, so we cannot cheat our way out by simply using BPF_W. The existence of the test means that such loads have to be supported, most likely because llvm can generate them. Fix the test by adding a big-endian variant, which uses an offset to access the least-significant byte of bpf_sysctl.file_pos. This reveals the final problem: verifier rejects accesses to bpf_sysctl fields with offset > 0. Such accesses are already allowed for a wide range of structs: __sk_buff, bpf_sock_addr and sk_msg_md to name a few. Extend this support to bpf_sysctl by using bpf_ctx_range instead of offsetof when matching field offsets. Fixes: 7b146cebe30c ("bpf: Sysctl hook") Fixes: e1550bfe0de4 ("bpf: Add file_pos field to bpf_sysctl ctx") Fixes: 9a1027e52535 ("selftests/bpf: Test file_pos field in bpf_sysctl ctx") Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Acked-by: Andrey Ignatov <rdna@fb.com> Acked-by: Andrii Nakryiko <andriin@fb.com> Link: https://lore.kernel.org/bpf/20191028122902.9763-1-iii@linux.ibm.com
2019-10-28 19:29:02 +07:00
case bpf_ctx_range(struct bpf_sysctl, file_pos):
if (type == BPF_READ) {
bpf_ctx_record_field_size(info, size_default);
return bpf_ctx_narrow_access_ok(off, size, size_default);
} else {
return size == size_default;
}
default:
return false;
}
}
static u32 sysctl_convert_ctx_access(enum bpf_access_type type,
const struct bpf_insn *si,
struct bpf_insn *insn_buf,
struct bpf_prog *prog, u32 *target_size)
{
struct bpf_insn *insn = insn_buf;
u32 read_size;
switch (si->off) {
case offsetof(struct bpf_sysctl, write):
*insn++ = BPF_LDX_MEM(
BPF_SIZE(si->code), si->dst_reg, si->src_reg,
bpf_target_off(struct bpf_sysctl_kern, write,
sizeof_field(struct bpf_sysctl_kern,
write),
target_size));
break;
case offsetof(struct bpf_sysctl, file_pos):
/* ppos is a pointer so it should be accessed via indirect
* loads and stores. Also for stores additional temporary
* register is used since neither src_reg nor dst_reg can be
* overridden.
*/
if (type == BPF_WRITE) {
int treg = BPF_REG_9;
if (si->src_reg == treg || si->dst_reg == treg)
--treg;
if (si->src_reg == treg || si->dst_reg == treg)
--treg;
*insn++ = BPF_STX_MEM(
BPF_DW, si->dst_reg, treg,
offsetof(struct bpf_sysctl_kern, tmp_reg));
*insn++ = BPF_LDX_MEM(
BPF_FIELD_SIZEOF(struct bpf_sysctl_kern, ppos),
treg, si->dst_reg,
offsetof(struct bpf_sysctl_kern, ppos));
*insn++ = BPF_STX_MEM(
BPF_SIZEOF(u32), treg, si->src_reg,
bpf_ctx_narrow_access_offset(
0, sizeof(u32), sizeof(loff_t)));
*insn++ = BPF_LDX_MEM(
BPF_DW, treg, si->dst_reg,
offsetof(struct bpf_sysctl_kern, tmp_reg));
} else {
*insn++ = BPF_LDX_MEM(
BPF_FIELD_SIZEOF(struct bpf_sysctl_kern, ppos),
si->dst_reg, si->src_reg,
offsetof(struct bpf_sysctl_kern, ppos));
read_size = bpf_size_to_bytes(BPF_SIZE(si->code));
*insn++ = BPF_LDX_MEM(
BPF_SIZE(si->code), si->dst_reg, si->dst_reg,
bpf_ctx_narrow_access_offset(
0, read_size, sizeof(loff_t)));
}
*target_size = sizeof(u32);
break;
}
return insn - insn_buf;
}
const struct bpf_verifier_ops cg_sysctl_verifier_ops = {
.get_func_proto = sysctl_func_proto,
.is_valid_access = sysctl_is_valid_access,
.convert_ctx_access = sysctl_convert_ctx_access,
};
const struct bpf_prog_ops cg_sysctl_prog_ops = {
};
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
static const struct bpf_func_proto *
cg_sockopt_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
{
switch (func_id) {
#ifdef CONFIG_NET
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
case BPF_FUNC_sk_storage_get:
return &bpf_sk_storage_get_proto;
case BPF_FUNC_sk_storage_delete:
return &bpf_sk_storage_delete_proto;
#endif
bpf: implement getsockopt and setsockopt hooks Implement new BPF_PROG_TYPE_CGROUP_SOCKOPT program type and BPF_CGROUP_{G,S}ETSOCKOPT cgroup hooks. BPF_CGROUP_SETSOCKOPT can modify user setsockopt arguments before passing them down to the kernel or bypass kernel completely. BPF_CGROUP_GETSOCKOPT can can inspect/modify getsockopt arguments that kernel returns. Both hooks reuse existing PTR_TO_PACKET{,_END} infrastructure. The buffer memory is pre-allocated (because I don't think there is a precedent for working with __user memory from bpf). This might be slow to do for each {s,g}etsockopt call, that's why I've added __cgroup_bpf_prog_array_is_empty that exits early if there is nothing attached to a cgroup. Note, however, that there is a race between __cgroup_bpf_prog_array_is_empty and BPF_PROG_RUN_ARRAY where cgroup program layout might have changed; this should not be a problem because in general there is a race between multiple calls to {s,g}etsocktop and user adding/removing bpf progs from a cgroup. The return code of the BPF program is handled as follows: * 0: EPERM * 1: success, continue with next BPF program in the cgroup chain v9: * allow overwriting setsockopt arguments (Alexei Starovoitov): * use set_fs (same as kernel_setsockopt) * buffer is always kzalloc'd (no small on-stack buffer) v8: * use s32 for optlen (Andrii Nakryiko) v7: * return only 0 or 1 (Alexei Starovoitov) * always run all progs (Alexei Starovoitov) * use optval=0 as kernel bypass in setsockopt (Alexei Starovoitov) (decided to use optval=-1 instead, optval=0 might be a valid input) * call getsockopt hook after kernel handlers (Alexei Starovoitov) v6: * rework cgroup chaining; stop as soon as bpf program returns 0 or 2; see patch with the documentation for the details * drop Andrii's and Martin's Acked-by (not sure they are comfortable with the new state of things) v5: * skip copy_to_user() and put_user() when ret == 0 (Martin Lau) v4: * don't export bpf_sk_fullsock helper (Martin Lau) * size != sizeof(__u64) for uapi pointers (Martin Lau) * offsetof instead of bpf_ctx_range when checking ctx access (Martin Lau) v3: * typos in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY comments (Andrii Nakryiko) * reverse christmas tree in BPF_PROG_CGROUP_SOCKOPT_RUN_ARRAY (Andrii Nakryiko) * use __bpf_md_ptr instead of __u32 for optval{,_end} (Martin Lau) * use BPF_FIELD_SIZEOF() for consistency (Martin Lau) * new CG_SOCKOPT_ACCESS macro to wrap repeated parts v2: * moved bpf_sockopt_kern fields around to remove a hole (Martin Lau) * aligned bpf_sockopt_kern->buf to 8 bytes (Martin Lau) * bpf_prog_array_is_empty instead of bpf_prog_array_length (Martin Lau) * added [0,2] return code check to verifier (Martin Lau) * dropped unused buf[64] from the stack (Martin Lau) * use PTR_TO_SOCKET for bpf_sockopt->sk (Martin Lau) * dropped bpf_target_off from ctx rewrites (Martin Lau) * use return code for kernel bypass (Martin Lau & Andrii Nakryiko) Cc: Andrii Nakryiko <andriin@fb.com> Cc: Martin Lau <kafai@fb.com> Signed-off-by: Stanislav Fomichev <sdf@google.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2019-06-28 03:38:47 +07:00
#ifdef CONFIG_INET
case BPF_FUNC_tcp_sock:
return &bpf_tcp_sock_proto;
#endif
default:
return cgroup_base_func_proto(func_id, prog);
}
}
static bool cg_sockopt_is_valid_access(int off, int size,
enum bpf_access_type type,
const struct bpf_prog *prog,
struct bpf_insn_access_aux *info)
{
const int size_default = sizeof(__u32);
if (off < 0 || off >= sizeof(struct bpf_sockopt))
return false;
if (off % size != 0)
return false;
if (type == BPF_WRITE) {
switch (off) {
case offsetof(struct bpf_sockopt, retval):
if (size != size_default)
return false;
return prog->expected_attach_type ==
BPF_CGROUP_GETSOCKOPT;
case offsetof(struct bpf_sockopt, optname):
/* fallthrough */
case offsetof(struct bpf_sockopt, level):
if (size != size_default)
return false;
return prog->expected_attach_type ==
BPF_CGROUP_SETSOCKOPT;
case offsetof(struct bpf_sockopt, optlen):
return size == size_default;
default:
return false;
}
}
switch (off) {
case offsetof(struct bpf_sockopt, sk):
if (size != sizeof(__u64))
return false;
info->reg_type = PTR_TO_SOCKET;
break;
case offsetof(struct bpf_sockopt, optval):
if (size != sizeof(__u64))
return false;
info->reg_type = PTR_TO_PACKET;
break;
case offsetof(struct bpf_sockopt, optval_end):
if (size != sizeof(__u64))
return false;
info->reg_type = PTR_TO_PACKET_END;
break;
case offsetof(struct bpf_sockopt, retval):
if (size != size_default)
return false;
return prog->expected_attach_type == BPF_CGROUP_GETSOCKOPT;
default:
if (size != size_default)
return false;
break;
}
return true;
}
#define CG_SOCKOPT_ACCESS_FIELD(T, F) \
T(BPF_FIELD_SIZEOF(struct bpf_sockopt_kern, F), \
si->dst_reg, si->src_reg, \
offsetof(struct bpf_sockopt_kern, F))
static u32 cg_sockopt_convert_ctx_access(enum bpf_access_type type,
const struct bpf_insn *si,
struct bpf_insn *insn_buf,
struct bpf_prog *prog,
u32 *target_size)
{
struct bpf_insn *insn = insn_buf;
switch (si->off) {
case offsetof(struct bpf_sockopt, sk):
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_LDX_MEM, sk);
break;
case offsetof(struct bpf_sockopt, level):
if (type == BPF_WRITE)
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_STX_MEM, level);
else
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_LDX_MEM, level);
break;
case offsetof(struct bpf_sockopt, optname):
if (type == BPF_WRITE)
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_STX_MEM, optname);
else
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_LDX_MEM, optname);
break;
case offsetof(struct bpf_sockopt, optlen):
if (type == BPF_WRITE)
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_STX_MEM, optlen);
else
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_LDX_MEM, optlen);
break;
case offsetof(struct bpf_sockopt, retval):
if (type == BPF_WRITE)
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_STX_MEM, retval);
else
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_LDX_MEM, retval);
break;
case offsetof(struct bpf_sockopt, optval):
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_LDX_MEM, optval);
break;
case offsetof(struct bpf_sockopt, optval_end):
*insn++ = CG_SOCKOPT_ACCESS_FIELD(BPF_LDX_MEM, optval_end);
break;
}
return insn - insn_buf;
}
static int cg_sockopt_get_prologue(struct bpf_insn *insn_buf,
bool direct_write,
const struct bpf_prog *prog)
{
/* Nothing to do for sockopt argument. The data is kzalloc'ated.
*/
return 0;
}
const struct bpf_verifier_ops cg_sockopt_verifier_ops = {
.get_func_proto = cg_sockopt_func_proto,
.is_valid_access = cg_sockopt_is_valid_access,
.convert_ctx_access = cg_sockopt_convert_ctx_access,
.gen_prologue = cg_sockopt_get_prologue,
};
const struct bpf_prog_ops cg_sockopt_prog_ops = {
};