From: KP Singh <[email protected]>
# v4 -> v5
- Fixes to selftests as suggested by Martin.
- Added Martin's acks.
# v3 -> v4
- Move the patch that exposes spin lock helpers to LSM programs as the
first patch as some of the changes in the implementation are actually
for spin locks.
- Clarify the comment in the bpf_task_storage_{get, delete} helper as
discussed with Martin.
- Added Martin's ack and rebased.
# v2 -> v3
- Added bpf_spin_locks to the selftests for local storage, found that
these are not available for LSM programs.
- Made spin lock helpers available for LSM programs (except sleepable
programs which need more work).
- Minor fixes for includes and added short commit messages for patches
that were split up for libbpf and bpftool.
- Added Song's acks.
# v1 -> v2
- Updated the refcounting for task_struct and simplified conversion
of fd -> struct pid.
- Some fixes suggested by Martin and Andrii, notably:
* long return type for the bpf_task_storage_delete helper (update
for bpf_inode_storage_delete will be sent separately).
* Remove extra nullness check to task_storage_ptr in map syscall
ops.
* Changed the argument signature of the BPF helpers to use
task_struct pointer in uapi headers.
* Remove unnecessary verifier logic for the bpf_get_current_task_btf
helper.
* Split the changes for bpftool and libbpf.
- Exercised syscall operations for local storage (kept a simpler verison
in test_local_storage.c, the eventual goal will be to update
sk_storage_map.c for all local storage types).
- Formatting fixes + Rebase.
We already have socket and inode local storage since [1]
This patch series:
* Implements bpf_local_storage for task_struct.
* Implements the bpf_get_current_task_btf helper which returns a BTF
pointer to the current task. Not only is this generally cleaner
(reading from the task_struct currently requires BPF_CORE_READ), it
also allows the BTF pointer to be used in task_local_storage helpers.
* In order to implement this helper, a RET_PTR_TO_BTF_ID is introduced
which works similar to RET_PTR_TO_BTF_ID_OR_NULL but does not require
a nullness check.
* Implements a detection in selftests which uses the
task local storage to deny a running executable from unlinking itself.
[1]: https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git/commit/?id=f836a56e84ffc9f1a1cd73f77e10404ca46a4616
KP Singh (9):
bpf: Allow LSM programs to use bpf spin locks
bpf: Implement task local storage
libbpf: Add support for task local storage
bpftool: Add support for task local storage
bpf: Implement get_current_task_btf and RET_PTR_TO_BTF_ID
bpf: Fix tests for local_storage
bpf: Update selftests for local_storage to use vmlinux.h
bpf: Add tests for task_local_storage
bpf: Exercise syscall operations for inode and sk storage
include/linux/bpf.h | 1 +
include/linux/bpf_lsm.h | 23 ++
include/linux/bpf_types.h | 1 +
include/uapi/linux/bpf.h | 48 +++
kernel/bpf/Makefile | 1 +
kernel/bpf/bpf_lsm.c | 8 +
kernel/bpf/bpf_task_storage.c | 315 ++++++++++++++++++
kernel/bpf/syscall.c | 3 +-
kernel/bpf/verifier.c | 37 +-
kernel/trace/bpf_trace.c | 16 +
security/bpf/hooks.c | 2 +
.../bpf/bpftool/Documentation/bpftool-map.rst | 3 +-
tools/bpf/bpftool/bash-completion/bpftool | 2 +-
tools/bpf/bpftool/map.c | 4 +-
tools/include/uapi/linux/bpf.h | 48 +++
tools/lib/bpf/libbpf_probes.c | 1 +
.../bpf/prog_tests/test_local_storage.c | 195 ++++++++++-
.../selftests/bpf/progs/local_storage.c | 103 ++++--
18 files changed, 752 insertions(+), 59 deletions(-)
create mode 100644 kernel/bpf/bpf_task_storage.c
--
2.29.1.341.ge80a0c044ae-goog
From: KP Singh <[email protected]>
Updates the binary to handle the BPF_MAP_TYPE_TASK_STORAGE as
"task_storage" for printing and parsing. Also updates the documentation
and bash completion
Acked-by: Song Liu <[email protected]>
Acked-by: Martin KaFai Lau <[email protected]>
Signed-off-by: KP Singh <[email protected]>
---
tools/bpf/bpftool/Documentation/bpftool-map.rst | 3 ++-
tools/bpf/bpftool/bash-completion/bpftool | 2 +-
tools/bpf/bpftool/map.c | 4 +++-
3 files changed, 6 insertions(+), 3 deletions(-)
diff --git a/tools/bpf/bpftool/Documentation/bpftool-map.rst b/tools/bpf/bpftool/Documentation/bpftool-map.rst
index dade10cdf295..3d52256ba75f 100644
--- a/tools/bpf/bpftool/Documentation/bpftool-map.rst
+++ b/tools/bpf/bpftool/Documentation/bpftool-map.rst
@@ -50,7 +50,8 @@ MAP COMMANDS
| | **lru_percpu_hash** | **lpm_trie** | **array_of_maps** | **hash_of_maps**
| | **devmap** | **devmap_hash** | **sockmap** | **cpumap** | **xskmap** | **sockhash**
| | **cgroup_storage** | **reuseport_sockarray** | **percpu_cgroup_storage**
-| | **queue** | **stack** | **sk_storage** | **struct_ops** | **ringbuf** | **inode_storage** }
+| | **queue** | **stack** | **sk_storage** | **struct_ops** | **ringbuf** | **inode_storage**
+ | **task_storage** }
DESCRIPTION
===========
diff --git a/tools/bpf/bpftool/bash-completion/bpftool b/tools/bpf/bpftool/bash-completion/bpftool
index 3f1da30c4da6..fdffbc64c65c 100644
--- a/tools/bpf/bpftool/bash-completion/bpftool
+++ b/tools/bpf/bpftool/bash-completion/bpftool
@@ -705,7 +705,7 @@ _bpftool()
hash_of_maps devmap devmap_hash sockmap cpumap \
xskmap sockhash cgroup_storage reuseport_sockarray \
percpu_cgroup_storage queue stack sk_storage \
- struct_ops inode_storage' -- \
+ struct_ops inode_storage task_storage' -- \
"$cur" ) )
return 0
;;
diff --git a/tools/bpf/bpftool/map.c b/tools/bpf/bpftool/map.c
index a7efbd84fbcc..b400364ee054 100644
--- a/tools/bpf/bpftool/map.c
+++ b/tools/bpf/bpftool/map.c
@@ -51,6 +51,7 @@ const char * const map_type_name[] = {
[BPF_MAP_TYPE_STRUCT_OPS] = "struct_ops",
[BPF_MAP_TYPE_RINGBUF] = "ringbuf",
[BPF_MAP_TYPE_INODE_STORAGE] = "inode_storage",
+ [BPF_MAP_TYPE_TASK_STORAGE] = "task_storage",
};
const size_t map_type_name_size = ARRAY_SIZE(map_type_name);
@@ -1464,7 +1465,8 @@ static int do_help(int argc, char **argv)
" lru_percpu_hash | lpm_trie | array_of_maps | hash_of_maps |\n"
" devmap | devmap_hash | sockmap | cpumap | xskmap | sockhash |\n"
" cgroup_storage | reuseport_sockarray | percpu_cgroup_storage |\n"
- " queue | stack | sk_storage | struct_ops | ringbuf | inode_storage }\n"
+ " queue | stack | sk_storage | struct_ops | ringbuf | inode_storage |\n"
+ " task_storage }\n"
" " HELP_SPEC_OPTIONS "\n"
"",
bin_name, argv[-2]);
--
2.29.1.341.ge80a0c044ae-goog
From: KP Singh <[email protected]>
The currently available bpf_get_current_task returns an unsigned integer
which can be used along with BPF_CORE_READ to read data from
the task_struct but still cannot be used as an input argument to a
helper that accepts an ARG_PTR_TO_BTF_ID of type task_struct.
In order to implement this helper a new return type, RET_PTR_TO_BTF_ID,
is added. This is similar to RET_PTR_TO_BTF_ID_OR_NULL but does not
require checking the nullness of returned pointer.
Acked-by: Song Liu <[email protected]>
Signed-off-by: KP Singh <[email protected]>
---
include/linux/bpf.h | 1 +
include/uapi/linux/bpf.h | 9 +++++++++
kernel/bpf/verifier.c | 7 +++++--
kernel/trace/bpf_trace.c | 16 ++++++++++++++++
tools/include/uapi/linux/bpf.h | 9 +++++++++
5 files changed, 40 insertions(+), 2 deletions(-)
diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 2fffd30e13ac..73d5381a5d5c 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -310,6 +310,7 @@ enum bpf_return_type {
RET_PTR_TO_BTF_ID_OR_NULL, /* returns a pointer to a btf_id or NULL */
RET_PTR_TO_MEM_OR_BTF_ID_OR_NULL, /* returns a pointer to a valid memory or a btf_id or NULL */
RET_PTR_TO_MEM_OR_BTF_ID, /* returns a pointer to a valid memory or a btf_id */
+ RET_PTR_TO_BTF_ID, /* returns a pointer to a btf_id */
};
/* eBPF function prototype used by verifier to allow BPF_CALLs from eBPF programs
diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
index f4037b2161a6..9879d6793e90 100644
--- a/include/uapi/linux/bpf.h
+++ b/include/uapi/linux/bpf.h
@@ -3779,6 +3779,14 @@ union bpf_attr {
* 0 on success.
*
* **-ENOENT** if the bpf_local_storage cannot be found.
+ *
+ * struct task_struct *bpf_get_current_task_btf(void)
+ * Description
+ * Return a BTF pointer to the "current" task.
+ * This pointer can also be used in helpers that accept an
+ * *ARG_PTR_TO_BTF_ID* of type *task_struct*.
+ * Return
+ * Pointer to the current task.
*/
#define __BPF_FUNC_MAPPER(FN) \
FN(unspec), \
@@ -3939,6 +3947,7 @@ union bpf_attr {
FN(redirect_peer), \
FN(task_storage_get), \
FN(task_storage_delete), \
+ FN(get_current_task_btf), \
/* */
/* integer value in 'imm' field of BPF_CALL instruction selects which helper
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 00960f6a83ec..10da26e55130 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -5186,11 +5186,14 @@ static int check_helper_call(struct bpf_verifier_env *env, int func_id, int insn
PTR_TO_BTF_ID : PTR_TO_BTF_ID_OR_NULL;
regs[BPF_REG_0].btf_id = meta.ret_btf_id;
}
- } else if (fn->ret_type == RET_PTR_TO_BTF_ID_OR_NULL) {
+ } else if (fn->ret_type == RET_PTR_TO_BTF_ID_OR_NULL ||
+ fn->ret_type == RET_PTR_TO_BTF_ID) {
int ret_btf_id;
mark_reg_known_zero(env, regs, BPF_REG_0);
- regs[BPF_REG_0].type = PTR_TO_BTF_ID_OR_NULL;
+ regs[BPF_REG_0].type = fn->ret_type == RET_PTR_TO_BTF_ID ?
+ PTR_TO_BTF_ID :
+ PTR_TO_BTF_ID_OR_NULL;
ret_btf_id = *fn->ret_btf_id;
if (ret_btf_id == 0) {
verbose(env, "invalid return type %d of func %s#%d\n",
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 4517c8b66518..e4515b0f62a8 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -1022,6 +1022,20 @@ const struct bpf_func_proto bpf_get_current_task_proto = {
.ret_type = RET_INTEGER,
};
+BPF_CALL_0(bpf_get_current_task_btf)
+{
+ return (unsigned long) current;
+}
+
+BTF_ID_LIST_SINGLE(bpf_get_current_btf_ids, struct, task_struct)
+
+static const struct bpf_func_proto bpf_get_current_task_btf_proto = {
+ .func = bpf_get_current_task_btf,
+ .gpl_only = true,
+ .ret_type = RET_PTR_TO_BTF_ID,
+ .ret_btf_id = &bpf_get_current_btf_ids[0],
+};
+
BPF_CALL_2(bpf_current_task_under_cgroup, struct bpf_map *, map, u32, idx)
{
struct bpf_array *array = container_of(map, struct bpf_array, map);
@@ -1265,6 +1279,8 @@ bpf_tracing_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
return &bpf_get_current_pid_tgid_proto;
case BPF_FUNC_get_current_task:
return &bpf_get_current_task_proto;
+ case BPF_FUNC_get_current_task_btf:
+ return &bpf_get_current_task_btf_proto;
case BPF_FUNC_get_current_uid_gid:
return &bpf_get_current_uid_gid_proto;
case BPF_FUNC_get_current_comm:
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
index f4037b2161a6..9879d6793e90 100644
--- a/tools/include/uapi/linux/bpf.h
+++ b/tools/include/uapi/linux/bpf.h
@@ -3779,6 +3779,14 @@ union bpf_attr {
* 0 on success.
*
* **-ENOENT** if the bpf_local_storage cannot be found.
+ *
+ * struct task_struct *bpf_get_current_task_btf(void)
+ * Description
+ * Return a BTF pointer to the "current" task.
+ * This pointer can also be used in helpers that accept an
+ * *ARG_PTR_TO_BTF_ID* of type *task_struct*.
+ * Return
+ * Pointer to the current task.
*/
#define __BPF_FUNC_MAPPER(FN) \
FN(unspec), \
@@ -3939,6 +3947,7 @@ union bpf_attr {
FN(redirect_peer), \
FN(task_storage_get), \
FN(task_storage_delete), \
+ FN(get_current_task_btf), \
/* */
/* integer value in 'imm' field of BPF_CALL instruction selects which helper
--
2.29.1.341.ge80a0c044ae-goog
From: KP Singh <[email protected]>
Use the check_syscall_operations added for task_local_storage to
exercise syscall operations for other local storage maps:
* Check the absence of an element for the given fd.
* Create a new element, retrieve and compare its value.
* Delete the element and check again for absence.
Acked-by: Martin KaFai Lau <[email protected]>
Signed-off-by: KP Singh <[email protected]>
---
.../bpf/prog_tests/test_local_storage.c | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/bpf/prog_tests/test_local_storage.c b/tools/testing/selftests/bpf/prog_tests/test_local_storage.c
index 5b4788b97d96..2c1b4e9c9a76 100644
--- a/tools/testing/selftests/bpf/prog_tests/test_local_storage.c
+++ b/tools/testing/selftests/bpf/prog_tests/test_local_storage.c
@@ -152,7 +152,7 @@ static bool check_syscall_operations(int map_fd, int obj_fd)
void test_test_local_storage(void)
{
char tmp_exec_path[PATH_MAX] = "/tmp/copy_of_rmXXXXXX";
- int err, serv_sk = -1, task_fd = -1;
+ int err, serv_sk = -1, task_fd = -1, rm_fd = -1;
struct local_storage *skel = NULL;
skel = local_storage__open_and_load();
@@ -176,6 +176,15 @@ void test_test_local_storage(void)
if (CHECK(err < 0, "copy_rm", "err %d errno %d\n", err, errno))
goto close_prog;
+ rm_fd = open(tmp_exec_path, O_RDONLY);
+ if (CHECK(rm_fd < 0, "open", "failed to open %s err:%d, errno:%d",
+ tmp_exec_path, rm_fd, errno))
+ goto close_prog;
+
+ if (!check_syscall_operations(bpf_map__fd(skel->maps.inode_storage_map),
+ rm_fd))
+ goto close_prog;
+
/* Sets skel->bss->monitored_pid to the pid of the forked child
* forks a child process that executes tmp_exec_path and tries to
* unlink its executable. This operation should be denied by the loaded
@@ -204,11 +213,15 @@ void test_test_local_storage(void)
CHECK(skel->data->sk_storage_result != 0, "sk_storage_result",
"sk_local_storage not set\n");
- close(serv_sk);
+ if (!check_syscall_operations(bpf_map__fd(skel->maps.sk_storage_map),
+ serv_sk))
+ goto close_prog;
close_prog_unlink:
unlink(tmp_exec_path);
close_prog:
+ close(serv_sk);
+ close(rm_fd);
close(task_fd);
local_storage__destroy(skel);
}
--
2.29.1.341.ge80a0c044ae-goog
From: KP Singh <[email protected]>
Similar to bpf_local_storage for sockets and inodes add local storage
for task_struct.
The life-cycle of storage is managed with the life-cycle of the
task_struct. i.e. the storage is destroyed along with the owning task
with a callback to the bpf_task_storage_free from the task_free LSM
hook.
The BPF LSM allocates an __rcu pointer to the bpf_local_storage in
the security blob which are now stackable and can co-exist with other
LSMs.
The userspace map operations can be done by using a pid fd as a key
passed to the lookup, update and delete operations.
Acked-by: Song Liu <[email protected]>
Acked-by: Martin KaFai Lau <[email protected]>
Signed-off-by: KP Singh <[email protected]>
---
include/linux/bpf_lsm.h | 23 +++
include/linux/bpf_types.h | 1 +
include/uapi/linux/bpf.h | 39 ++++
kernel/bpf/Makefile | 1 +
kernel/bpf/bpf_lsm.c | 4 +
kernel/bpf/bpf_task_storage.c | 315 +++++++++++++++++++++++++++++++++
kernel/bpf/syscall.c | 3 +-
kernel/bpf/verifier.c | 10 ++
security/bpf/hooks.c | 2 +
tools/include/uapi/linux/bpf.h | 39 ++++
10 files changed, 436 insertions(+), 1 deletion(-)
create mode 100644 kernel/bpf/bpf_task_storage.c
diff --git a/include/linux/bpf_lsm.h b/include/linux/bpf_lsm.h
index aaacb6aafc87..73226181b744 100644
--- a/include/linux/bpf_lsm.h
+++ b/include/linux/bpf_lsm.h
@@ -7,6 +7,7 @@
#ifndef _LINUX_BPF_LSM_H
#define _LINUX_BPF_LSM_H
+#include <linux/sched.h>
#include <linux/bpf.h>
#include <linux/lsm_hooks.h>
@@ -35,9 +36,21 @@ static inline struct bpf_storage_blob *bpf_inode(
return inode->i_security + bpf_lsm_blob_sizes.lbs_inode;
}
+static inline struct bpf_storage_blob *bpf_task(
+ const struct task_struct *task)
+{
+ if (unlikely(!task->security))
+ return NULL;
+
+ return task->security + bpf_lsm_blob_sizes.lbs_task;
+}
+
extern const struct bpf_func_proto bpf_inode_storage_get_proto;
extern const struct bpf_func_proto bpf_inode_storage_delete_proto;
+extern const struct bpf_func_proto bpf_task_storage_get_proto;
+extern const struct bpf_func_proto bpf_task_storage_delete_proto;
void bpf_inode_storage_free(struct inode *inode);
+void bpf_task_storage_free(struct task_struct *task);
#else /* !CONFIG_BPF_LSM */
@@ -53,10 +66,20 @@ static inline struct bpf_storage_blob *bpf_inode(
return NULL;
}
+static inline struct bpf_storage_blob *bpf_task(
+ const struct task_struct *task)
+{
+ return NULL;
+}
+
static inline void bpf_inode_storage_free(struct inode *inode)
{
}
+static inline void bpf_task_storage_free(struct task_struct *task)
+{
+}
+
#endif /* CONFIG_BPF_LSM */
#endif /* _LINUX_BPF_LSM_H */
diff --git a/include/linux/bpf_types.h b/include/linux/bpf_types.h
index 2e6f568377f1..99f7fd657d87 100644
--- a/include/linux/bpf_types.h
+++ b/include/linux/bpf_types.h
@@ -109,6 +109,7 @@ BPF_MAP_TYPE(BPF_MAP_TYPE_SOCKHASH, sock_hash_ops)
#endif
#ifdef CONFIG_BPF_LSM
BPF_MAP_TYPE(BPF_MAP_TYPE_INODE_STORAGE, inode_storage_map_ops)
+BPF_MAP_TYPE(BPF_MAP_TYPE_TASK_STORAGE, task_storage_map_ops)
#endif
BPF_MAP_TYPE(BPF_MAP_TYPE_CPUMAP, cpu_map_ops)
#if defined(CONFIG_XDP_SOCKETS)
diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
index e6ceac3f7d62..f4037b2161a6 100644
--- a/include/uapi/linux/bpf.h
+++ b/include/uapi/linux/bpf.h
@@ -157,6 +157,7 @@ enum bpf_map_type {
BPF_MAP_TYPE_STRUCT_OPS,
BPF_MAP_TYPE_RINGBUF,
BPF_MAP_TYPE_INODE_STORAGE,
+ BPF_MAP_TYPE_TASK_STORAGE,
};
/* Note that tracing related programs such as
@@ -3742,6 +3743,42 @@ union bpf_attr {
* Return
* The helper returns **TC_ACT_REDIRECT** on success or
* **TC_ACT_SHOT** on error.
+ *
+ * void *bpf_task_storage_get(struct bpf_map *map, struct task_struct *task, void *value, u64 flags)
+ * Description
+ * Get a bpf_local_storage from the *task*.
+ *
+ * Logically, it could be thought of as getting the value from
+ * a *map* with *task* as the **key**. From this
+ * perspective, the usage is not much different from
+ * **bpf_map_lookup_elem**\ (*map*, **&**\ *task*) except this
+ * helper enforces the key must be an task_struct and the map must also
+ * be a **BPF_MAP_TYPE_TASK_STORAGE**.
+ *
+ * Underneath, the value is stored locally at *task* instead of
+ * the *map*. The *map* is used as the bpf-local-storage
+ * "type". The bpf-local-storage "type" (i.e. the *map*) is
+ * searched against all bpf_local_storage residing at *task*.
+ *
+ * An optional *flags* (**BPF_LOCAL_STORAGE_GET_F_CREATE**) can be
+ * used such that a new bpf_local_storage will be
+ * created if one does not exist. *value* can be used
+ * together with **BPF_LOCAL_STORAGE_GET_F_CREATE** to specify
+ * the initial value of a bpf_local_storage. If *value* is
+ * **NULL**, the new bpf_local_storage will be zero initialized.
+ * Return
+ * A bpf_local_storage pointer is returned on success.
+ *
+ * **NULL** if not found or there was an error in adding
+ * a new bpf_local_storage.
+ *
+ * long bpf_task_storage_delete(struct bpf_map *map, struct task_struct *task)
+ * Description
+ * Delete a bpf_local_storage from a *task*.
+ * Return
+ * 0 on success.
+ *
+ * **-ENOENT** if the bpf_local_storage cannot be found.
*/
#define __BPF_FUNC_MAPPER(FN) \
FN(unspec), \
@@ -3900,6 +3937,8 @@ union bpf_attr {
FN(bpf_per_cpu_ptr), \
FN(bpf_this_cpu_ptr), \
FN(redirect_peer), \
+ FN(task_storage_get), \
+ FN(task_storage_delete), \
/* */
/* integer value in 'imm' field of BPF_CALL instruction selects which helper
diff --git a/kernel/bpf/Makefile b/kernel/bpf/Makefile
index bdc8cd1b6767..f0b93ced5a7f 100644
--- a/kernel/bpf/Makefile
+++ b/kernel/bpf/Makefile
@@ -6,6 +6,7 @@ obj-$(CONFIG_BPF_SYSCALL) += syscall.o verifier.o inode.o helpers.o tnum.o bpf_i
obj-$(CONFIG_BPF_SYSCALL) += hashtab.o arraymap.o percpu_freelist.o bpf_lru_list.o lpm_trie.o map_in_map.o
obj-$(CONFIG_BPF_SYSCALL) += local_storage.o queue_stack_maps.o ringbuf.o
obj-${CONFIG_BPF_LSM} += bpf_inode_storage.o
+obj-${CONFIG_BPF_LSM} += bpf_task_storage.o
obj-$(CONFIG_BPF_SYSCALL) += disasm.o
obj-$(CONFIG_BPF_JIT) += trampoline.o
obj-$(CONFIG_BPF_SYSCALL) += btf.o
diff --git a/kernel/bpf/bpf_lsm.c b/kernel/bpf/bpf_lsm.c
index cd8a617f2109..e92c51bebb47 100644
--- a/kernel/bpf/bpf_lsm.c
+++ b/kernel/bpf/bpf_lsm.c
@@ -63,6 +63,10 @@ bpf_lsm_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
return &bpf_spin_lock_proto;
case BPF_FUNC_spin_unlock:
return &bpf_spin_unlock_proto;
+ case BPF_FUNC_task_storage_get:
+ return &bpf_task_storage_get_proto;
+ case BPF_FUNC_task_storage_delete:
+ return &bpf_task_storage_delete_proto;
default:
return tracing_prog_func_proto(func_id, prog);
}
diff --git a/kernel/bpf/bpf_task_storage.c b/kernel/bpf/bpf_task_storage.c
new file mode 100644
index 000000000000..39a45fba4fb0
--- /dev/null
+++ b/kernel/bpf/bpf_task_storage.c
@@ -0,0 +1,315 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2020 Facebook
+ * Copyright 2020 Google LLC.
+ */
+
+#include <linux/pid.h>
+#include <linux/sched.h>
+#include <linux/rculist.h>
+#include <linux/list.h>
+#include <linux/hash.h>
+#include <linux/types.h>
+#include <linux/spinlock.h>
+#include <linux/bpf.h>
+#include <linux/bpf_local_storage.h>
+#include <linux/filter.h>
+#include <uapi/linux/btf.h>
+#include <linux/bpf_lsm.h>
+#include <linux/btf_ids.h>
+#include <linux/fdtable.h>
+
+DEFINE_BPF_STORAGE_CACHE(task_cache);
+
+static struct bpf_local_storage __rcu **task_storage_ptr(void *owner)
+{
+ struct task_struct *task = owner;
+ struct bpf_storage_blob *bsb;
+
+ bsb = bpf_task(task);
+ if (!bsb)
+ return NULL;
+ return &bsb->storage;
+}
+
+static struct bpf_local_storage_data *
+task_storage_lookup(struct task_struct *task, struct bpf_map *map,
+ bool cacheit_lockit)
+{
+ struct bpf_local_storage *task_storage;
+ struct bpf_local_storage_map *smap;
+ struct bpf_storage_blob *bsb;
+
+ bsb = bpf_task(task);
+ if (!bsb)
+ return NULL;
+
+ task_storage = rcu_dereference(bsb->storage);
+ if (!task_storage)
+ return NULL;
+
+ smap = (struct bpf_local_storage_map *)map;
+ return bpf_local_storage_lookup(task_storage, smap, cacheit_lockit);
+}
+
+void bpf_task_storage_free(struct task_struct *task)
+{
+ struct bpf_local_storage_elem *selem;
+ struct bpf_local_storage *local_storage;
+ bool free_task_storage = false;
+ struct bpf_storage_blob *bsb;
+ struct hlist_node *n;
+
+ bsb = bpf_task(task);
+ if (!bsb)
+ return;
+
+ rcu_read_lock();
+
+ local_storage = rcu_dereference(bsb->storage);
+ if (!local_storage) {
+ rcu_read_unlock();
+ return;
+ }
+
+ /* Neither the bpf_prog nor the bpf-map's syscall
+ * could be modifying the local_storage->list now.
+ * Thus, no elem can be added-to or deleted-from the
+ * local_storage->list by the bpf_prog or by the bpf-map's syscall.
+ *
+ * It is racing with bpf_local_storage_map_free() alone
+ * when unlinking elem from the local_storage->list and
+ * the map's bucket->list.
+ */
+ raw_spin_lock_bh(&local_storage->lock);
+ hlist_for_each_entry_safe(selem, n, &local_storage->list, snode) {
+ /* Always unlink from map before unlinking from
+ * local_storage.
+ */
+ bpf_selem_unlink_map(selem);
+ free_task_storage = bpf_selem_unlink_storage_nolock(
+ local_storage, selem, false);
+ }
+ raw_spin_unlock_bh(&local_storage->lock);
+ rcu_read_unlock();
+
+ /* free_task_storage should always be true as long as
+ * local_storage->list was non-empty.
+ */
+ if (free_task_storage)
+ kfree_rcu(local_storage, rcu);
+}
+
+static void *bpf_pid_task_storage_lookup_elem(struct bpf_map *map, void *key)
+{
+ struct bpf_local_storage_data *sdata;
+ struct task_struct *task;
+ unsigned int f_flags;
+ struct pid *pid;
+ int fd, err;
+
+ fd = *(int *)key;
+ pid = pidfd_get_pid(fd, &f_flags);
+ if (IS_ERR(pid))
+ return ERR_CAST(pid);
+
+ /* We should be in an RCU read side critical section, it should be safe
+ * to call pid_task.
+ */
+ WARN_ON_ONCE(!rcu_read_lock_held());
+ task = pid_task(pid, PIDTYPE_PID);
+ if (!task) {
+ err = -ENOENT;
+ goto out;
+ }
+
+ sdata = task_storage_lookup(task, map, true);
+ put_pid(pid);
+ return sdata ? sdata->data : NULL;
+out:
+ put_pid(pid);
+ return ERR_PTR(err);
+}
+
+static int bpf_pid_task_storage_update_elem(struct bpf_map *map, void *key,
+ void *value, u64 map_flags)
+{
+ struct bpf_local_storage_data *sdata;
+ struct task_struct *task;
+ unsigned int f_flags;
+ struct pid *pid;
+ int fd, err;
+
+ fd = *(int *)key;
+ pid = pidfd_get_pid(fd, &f_flags);
+ if (IS_ERR(pid))
+ return PTR_ERR(pid);
+
+ /* We should be in an RCU read side critical section, it should be safe
+ * to call pid_task.
+ */
+ WARN_ON_ONCE(!rcu_read_lock_held());
+ task = pid_task(pid, PIDTYPE_PID);
+ if (!task) {
+ err = -ENOENT;
+ goto out;
+ }
+
+ sdata = bpf_local_storage_update(
+ task, (struct bpf_local_storage_map *)map, value, map_flags);
+
+ err = PTR_ERR_OR_ZERO(sdata);
+out:
+ put_pid(pid);
+ return err;
+}
+
+static int task_storage_delete(struct task_struct *task, struct bpf_map *map)
+{
+ struct bpf_local_storage_data *sdata;
+
+ sdata = task_storage_lookup(task, map, false);
+ if (!sdata)
+ return -ENOENT;
+
+ bpf_selem_unlink(SELEM(sdata));
+
+ return 0;
+}
+
+static int bpf_pid_task_storage_delete_elem(struct bpf_map *map, void *key)
+{
+ struct task_struct *task;
+ unsigned int f_flags;
+ struct pid *pid;
+ int fd, err;
+
+ fd = *(int *)key;
+ pid = pidfd_get_pid(fd, &f_flags);
+ if (IS_ERR(pid))
+ return PTR_ERR(pid);
+
+ /* We should be in an RCU read side critical section, it should be safe
+ * to call pid_task.
+ */
+ WARN_ON_ONCE(!rcu_read_lock_held());
+ task = pid_task(pid, PIDTYPE_PID);
+ if (!task) {
+ err = -ENOENT;
+ goto out;
+ }
+
+ err = task_storage_delete(task, map);
+out:
+ put_pid(pid);
+ return err;
+}
+
+BPF_CALL_4(bpf_task_storage_get, struct bpf_map *, map, struct task_struct *,
+ task, void *, value, u64, flags)
+{
+ struct bpf_local_storage_data *sdata;
+
+ if (flags & ~(BPF_LOCAL_STORAGE_GET_F_CREATE))
+ return (unsigned long)NULL;
+
+ /* explicitly check that the task_storage_ptr is not
+ * NULL as task_storage_lookup returns NULL in this case and
+ * bpf_local_storage_update expects the owner to have a
+ * valid storage pointer.
+ */
+ if (!task_storage_ptr(task))
+ return (unsigned long)NULL;
+
+ sdata = task_storage_lookup(task, map, true);
+ if (sdata)
+ return (unsigned long)sdata->data;
+
+ /* This helper must only be called from places where the lifetime of the task
+ * is guaranteed. Either by being refcounted or by being protected
+ * by an RCU read-side critical section.
+ */
+ if (flags & BPF_LOCAL_STORAGE_GET_F_CREATE) {
+ sdata = bpf_local_storage_update(
+ task, (struct bpf_local_storage_map *)map, value,
+ BPF_NOEXIST);
+ return IS_ERR(sdata) ? (unsigned long)NULL :
+ (unsigned long)sdata->data;
+ }
+
+ return (unsigned long)NULL;
+}
+
+BPF_CALL_2(bpf_task_storage_delete, struct bpf_map *, map, struct task_struct *,
+ task)
+{
+ /* This helper must only be called from places where the lifetime of the task
+ * is guaranteed. Either by being refcounted or by being protected
+ * by an RCU read-side critical section.
+ */
+ return task_storage_delete(task, map);
+}
+
+static int notsupp_get_next_key(struct bpf_map *map, void *key, void *next_key)
+{
+ return -ENOTSUPP;
+}
+
+static struct bpf_map *task_storage_map_alloc(union bpf_attr *attr)
+{
+ struct bpf_local_storage_map *smap;
+
+ smap = bpf_local_storage_map_alloc(attr);
+ if (IS_ERR(smap))
+ return ERR_CAST(smap);
+
+ smap->cache_idx = bpf_local_storage_cache_idx_get(&task_cache);
+ return &smap->map;
+}
+
+static void task_storage_map_free(struct bpf_map *map)
+{
+ struct bpf_local_storage_map *smap;
+
+ smap = (struct bpf_local_storage_map *)map;
+ bpf_local_storage_cache_idx_free(&task_cache, smap->cache_idx);
+ bpf_local_storage_map_free(smap);
+}
+
+static int task_storage_map_btf_id;
+const struct bpf_map_ops task_storage_map_ops = {
+ .map_meta_equal = bpf_map_meta_equal,
+ .map_alloc_check = bpf_local_storage_map_alloc_check,
+ .map_alloc = task_storage_map_alloc,
+ .map_free = task_storage_map_free,
+ .map_get_next_key = notsupp_get_next_key,
+ .map_lookup_elem = bpf_pid_task_storage_lookup_elem,
+ .map_update_elem = bpf_pid_task_storage_update_elem,
+ .map_delete_elem = bpf_pid_task_storage_delete_elem,
+ .map_check_btf = bpf_local_storage_map_check_btf,
+ .map_btf_name = "bpf_local_storage_map",
+ .map_btf_id = &task_storage_map_btf_id,
+ .map_owner_storage_ptr = task_storage_ptr,
+};
+
+BTF_ID_LIST_SINGLE(bpf_task_storage_btf_ids, struct, task_struct)
+
+const struct bpf_func_proto bpf_task_storage_get_proto = {
+ .func = bpf_task_storage_get,
+ .gpl_only = false,
+ .ret_type = RET_PTR_TO_MAP_VALUE_OR_NULL,
+ .arg1_type = ARG_CONST_MAP_PTR,
+ .arg2_type = ARG_PTR_TO_BTF_ID,
+ .arg2_btf_id = &bpf_task_storage_btf_ids[0],
+ .arg3_type = ARG_PTR_TO_MAP_VALUE_OR_NULL,
+ .arg4_type = ARG_ANYTHING,
+};
+
+const struct bpf_func_proto bpf_task_storage_delete_proto = {
+ .func = bpf_task_storage_delete,
+ .gpl_only = false,
+ .ret_type = RET_INTEGER,
+ .arg1_type = ARG_CONST_MAP_PTR,
+ .arg2_type = ARG_PTR_TO_BTF_ID,
+ .arg2_btf_id = &bpf_task_storage_btf_ids[0],
+};
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 8f50c9c19f1b..f3fe9f53f93c 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -773,7 +773,8 @@ static int map_check_btf(struct bpf_map *map, const struct btf *btf,
map->map_type != BPF_MAP_TYPE_ARRAY &&
map->map_type != BPF_MAP_TYPE_CGROUP_STORAGE &&
map->map_type != BPF_MAP_TYPE_SK_STORAGE &&
- map->map_type != BPF_MAP_TYPE_INODE_STORAGE)
+ map->map_type != BPF_MAP_TYPE_INODE_STORAGE &&
+ map->map_type != BPF_MAP_TYPE_TASK_STORAGE)
return -ENOTSUPP;
if (map->spin_lock_off + sizeof(struct bpf_spin_lock) >
map->value_size) {
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index f863aa84d0a2..00960f6a83ec 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -4469,6 +4469,11 @@ static int check_map_func_compatibility(struct bpf_verifier_env *env,
func_id != BPF_FUNC_inode_storage_delete)
goto error;
break;
+ case BPF_MAP_TYPE_TASK_STORAGE:
+ if (func_id != BPF_FUNC_task_storage_get &&
+ func_id != BPF_FUNC_task_storage_delete)
+ goto error;
+ break;
default:
break;
}
@@ -4547,6 +4552,11 @@ static int check_map_func_compatibility(struct bpf_verifier_env *env,
if (map->map_type != BPF_MAP_TYPE_INODE_STORAGE)
goto error;
break;
+ case BPF_FUNC_task_storage_get:
+ case BPF_FUNC_task_storage_delete:
+ if (map->map_type != BPF_MAP_TYPE_TASK_STORAGE)
+ goto error;
+ break;
default:
break;
}
diff --git a/security/bpf/hooks.c b/security/bpf/hooks.c
index 788667d582ae..e5971fa74fd7 100644
--- a/security/bpf/hooks.c
+++ b/security/bpf/hooks.c
@@ -12,6 +12,7 @@ static struct security_hook_list bpf_lsm_hooks[] __lsm_ro_after_init = {
#include <linux/lsm_hook_defs.h>
#undef LSM_HOOK
LSM_HOOK_INIT(inode_free_security, bpf_inode_storage_free),
+ LSM_HOOK_INIT(task_free, bpf_task_storage_free),
};
static int __init bpf_lsm_init(void)
@@ -23,6 +24,7 @@ static int __init bpf_lsm_init(void)
struct lsm_blob_sizes bpf_lsm_blob_sizes __lsm_ro_after_init = {
.lbs_inode = sizeof(struct bpf_storage_blob),
+ .lbs_task = sizeof(struct bpf_storage_blob),
};
DEFINE_LSM(bpf) = {
diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
index e6ceac3f7d62..f4037b2161a6 100644
--- a/tools/include/uapi/linux/bpf.h
+++ b/tools/include/uapi/linux/bpf.h
@@ -157,6 +157,7 @@ enum bpf_map_type {
BPF_MAP_TYPE_STRUCT_OPS,
BPF_MAP_TYPE_RINGBUF,
BPF_MAP_TYPE_INODE_STORAGE,
+ BPF_MAP_TYPE_TASK_STORAGE,
};
/* Note that tracing related programs such as
@@ -3742,6 +3743,42 @@ union bpf_attr {
* Return
* The helper returns **TC_ACT_REDIRECT** on success or
* **TC_ACT_SHOT** on error.
+ *
+ * void *bpf_task_storage_get(struct bpf_map *map, struct task_struct *task, void *value, u64 flags)
+ * Description
+ * Get a bpf_local_storage from the *task*.
+ *
+ * Logically, it could be thought of as getting the value from
+ * a *map* with *task* as the **key**. From this
+ * perspective, the usage is not much different from
+ * **bpf_map_lookup_elem**\ (*map*, **&**\ *task*) except this
+ * helper enforces the key must be an task_struct and the map must also
+ * be a **BPF_MAP_TYPE_TASK_STORAGE**.
+ *
+ * Underneath, the value is stored locally at *task* instead of
+ * the *map*. The *map* is used as the bpf-local-storage
+ * "type". The bpf-local-storage "type" (i.e. the *map*) is
+ * searched against all bpf_local_storage residing at *task*.
+ *
+ * An optional *flags* (**BPF_LOCAL_STORAGE_GET_F_CREATE**) can be
+ * used such that a new bpf_local_storage will be
+ * created if one does not exist. *value* can be used
+ * together with **BPF_LOCAL_STORAGE_GET_F_CREATE** to specify
+ * the initial value of a bpf_local_storage. If *value* is
+ * **NULL**, the new bpf_local_storage will be zero initialized.
+ * Return
+ * A bpf_local_storage pointer is returned on success.
+ *
+ * **NULL** if not found or there was an error in adding
+ * a new bpf_local_storage.
+ *
+ * long bpf_task_storage_delete(struct bpf_map *map, struct task_struct *task)
+ * Description
+ * Delete a bpf_local_storage from a *task*.
+ * Return
+ * 0 on success.
+ *
+ * **-ENOENT** if the bpf_local_storage cannot be found.
*/
#define __BPF_FUNC_MAPPER(FN) \
FN(unspec), \
@@ -3900,6 +3937,8 @@ union bpf_attr {
FN(bpf_per_cpu_ptr), \
FN(bpf_this_cpu_ptr), \
FN(redirect_peer), \
+ FN(task_storage_get), \
+ FN(task_storage_delete), \
/* */
/* integer value in 'imm' field of BPF_CALL instruction selects which helper
--
2.29.1.341.ge80a0c044ae-goog
On Thu, Nov 05, 2020 at 10:58:23PM +0000, KP Singh wrote:
> From: KP Singh <[email protected]>
>
> The currently available bpf_get_current_task returns an unsigned integer
> which can be used along with BPF_CORE_READ to read data from
> the task_struct but still cannot be used as an input argument to a
> helper that accepts an ARG_PTR_TO_BTF_ID of type task_struct.
>
> In order to implement this helper a new return type, RET_PTR_TO_BTF_ID,
> is added. This is similar to RET_PTR_TO_BTF_ID_OR_NULL but does not
> require checking the nullness of returned pointer.
>
> Acked-by: Song Liu <[email protected]>
> Signed-off-by: KP Singh <[email protected]>
Acked-by: Martin KaFai Lau <[email protected]>