Some versions of GCC are really nit-picky about strncpy() use. Use memcpy(),
as they are pretty much equivalent for the case of fixed length strings.

Fixes: e459f49b ("libbpf: Separate XDP program load with xsk socket creation")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203235440.2302137-1-andrii@kernel.org

Andrii Nakryiko says:

====================

This patch sets extends kernel and libbpf with support for attaching
BTF-powered raw tracepoint (tp_btf) and tracing (fentry/fexit/fmod_ret/lsm)
BPF programs to BPF hooks defined in kernel modules. As part of that, libbpf
now supports performing CO-RE relocations against types in kernel module BTFs,
in addition to existing vmlinux BTF support.

Kernel UAPI for BPF_PROG_LOAD now allows to specify kernel module (or vmlinux)
BTF object FD in attach_btf_obj_fd field, aliased to attach_prog_fd. This is
used to identify which BTF object needs to be used for finding BTF type by
provided attach_btf_id.

This patch set also sets up a convenient and fully-controlled custom kernel
module (called "bpf_testmod"), that is a predictable playground for all the
BPF selftests, that rely on module BTFs. Currently pahole doesn't generate
BTF_KIND_FUNC info for ftrace-able static functions in kernel modules, so
expose traced function in bpf_sidecar.ko. Once pahole is enhanced, we can go
back to static function.

From end user perspective there are no extra actions that need to happen.
Libbpf will continue searching across all kernel module BTFs, if desired
attach BTF type is not found in vmlinux. That way it doesn't matter if BPF
hook that user is trying to attach to is built into vmlinux image or is
loaded in kernel module.

v5->v6:
  - move btf_put() back to syscall.c (kernel test robot);
  - added close(fd) in patch #5 (John);
v4->v5:
  - use FD to specify BTF object (Alexei);
  - move prog->aux->attach_btf putting into bpf_prog_free() for consistency
    with putting prog->aux->dst_prog;
  - fix BTF FD leak(s) in libbpf;
v3->v4:
  - merge together patch sets [0] and [1];
  - avoid increasing bpf_reg_state by reordering fields (Alexei);
  - preserve btf_data_size in struct module;
v2->v3:
  - fix subtle uninitialized variable use in BTF ID iteration code;
v1->v2:
  - module_put() inside preempt_disable() region (Alexei);
  - bpf_sidecar -> bpf_testmod rename (Alexei);
  - test_progs more relaxed handling of bpf_testmod;
  - test_progs marks skipped sub-tests properly as SKIP now.

  [0] https://patchwork.kernel.org/project/netdevbpf/list/?series=393677&state=*
  [1] https://patchwork.kernel.org/project/netdevbpf/list/?series=393679&state=*
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Add new selftest checking attachment of fentry/fexit/fmod_ret (and raw
tracepoint ones for completeness) BPF programs to kernel module function.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-15-andrii@kernel.org

Add another CO-RE relocation test for kernel module relocations. This time for
tp_btf with direct memory reads.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-14-andrii@kernel.org

Teach libbpf to search for BTF types in kernel modules for tracing BPF
programs. This allows attachment of raw_tp/fentry/fexit/fmod_ret/etc BPF
program types to tracepoints and functions in kernel modules.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-13-andrii@kernel.org

Refactor low-level API for BPF program loading to not rely on public API
types. This allows painless extension without constant efforts to cleverly not
break backwards compatibility.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-12-andrii@kernel.org

Add ability for user-space programs to specify non-vmlinux BTF when attaching
BTF-powered BPF programs: raw_tp, fentry/fexit/fmod_ret, LSM, etc. For this,
attach_prog_fd (now with the alias name attach_btf_obj_fd) should specify FD
of a module or vmlinux BTF object. For backwards compatibility reasons,
0 denotes vmlinux BTF. Only kernel BTF (vmlinux or module) can be specified.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-11-andrii@kernel.org

Remove a permeating assumption thoughout BPF verifier of vmlinux BTF. Instead,
wherever BTF type IDs are involved, also track the instance of struct btf that
goes along with the type ID. This allows to gradually add support for kernel
module BTFs and using/tracking module types across BPF helper calls and
registers.

This patch also renames btf_id() function to btf_obj_id() to minimize naming
clash with using btf_id to denote BTF *type* ID, rather than BTF *object*'s ID.

Also, altough btf_vmlinux can't get destructed and thus doesn't need
refcounting, module BTFs need that, so apply BTF refcounting universally when
BPF program is using BTF-powered attachment (tp_btf, fentry/fexit, etc). This
makes for simpler clean up code.

Now that BTF type ID is not enough to uniquely identify a BTF type, extend BPF
trampoline key to include BTF object ID. To differentiate that from target
program BPF ID, set 31st bit of type ID. BTF type IDs (at least currently) are
not allowed to take full 32 bits, so there is no danger of confusing that bit
with a valid BTF type ID.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-10-andrii@kernel.org

Add a self-tests validating libbpf is able to perform CO-RE relocations
against the type defined in kernel module BTF. if bpf_testmod.o is not
supported by the kernel (e.g., due to version mismatch), skip tests, instead
of failing.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-9-andrii@kernel.org

Previously skipped sub-tests would be counted as passing with ":OK" appened
in the log. Change that to be accounted as ":SKIP".
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-8-andrii@kernel.org

Add bpf_testmod module, which is conceptually out-of-tree module and provides
ways for selftests/bpf to test various kernel module-related functionality:
raw tracepoint, fentry/fexit/fmod_ret, etc. This module will be auto-loaded by
test_progs test runner and expected by some of selftests to be present and
loaded.

Pahole currently isn't able to generate BTF for static functions in kernel
modules, so make sure traced function is global.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-7-andrii@kernel.org

Teach libbpf to search for candidate types for CO-RE relocations across kernel
modules BTFs, in addition to vmlinux BTF. If at least one candidate type is
found in vmlinux BTF, kernel module BTFs are not iterated. If vmlinux BTF has
no matching candidates, then find all kernel module BTFs and search for all
matching candidates across all of them.

Kernel's support for module BTFs are inferred from the support for BTF name
pointer in BPF UAPI.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-6-andrii@kernel.org

Refactor CO-RE relocation candidate search to not expect a single BTF, rather
return all candidate types with their corresponding BTF objects. This will
allow to extend CO-RE relocations to accommodate kernel module BTFs.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-5-andrii@kernel.org

Add a btf_get_from_fd() helper, which constructs struct btf from in-kernel BTF
data by FD. This is used for loading module BTFs.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-4-andrii@kernel.org

Having real btf_data_size stored in struct module is benefitial to quickly
determine which kernel modules have associated BTF object and which don't.
There is no harm in keeping this info, as opposed to keeping invalid pointer.

Fixes: 607c543f ("bpf: Sanitize BTF data pointer after module is loaded")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-3-andrii@kernel.org

__module_address() needs to be called with preemption disabled or with
module_mutex taken. preempt_disable() is enough for read-only uses, which is
what this fix does. Also, module_put() does internal check for NULL, so drop
it as well.

Fixes: a38d1107 ("bpf: support raw tracepoints in modules")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20201203204634.1325171-2-andrii@kernel.org

Prankur gupta says:

====================

This patch contains support to set tcp window_field field from bpf setsockops.

v2: Used TCP_WINDOW_CLAMP setsockopt logic for bpf_setsockopt (review comment addressed)

v3: Created a common function for duplicated code (review comment addressed)

v4: Removing logic to pass struct sock and struct tcp_sock together (review comment addressed)
====================
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Adding selftests for new added functionality to set TCP_WINDOW_CLAMP
from bpf setsockopt.
Signed-off-by: Prankur gupta <prankgup@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201202213152.435886-3-prankgup@fb.com

Adds a new bpf_setsockopt for TCP sockets, TCP_BPF_WINDOW_CLAMP,
which sets the maximum receiver window size. It will be useful for
limiting receiver window based on RTT.
Signed-off-by: Prankur gupta <prankgup@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201202213152.435886-2-prankgup@fb.com

There is a spelling mistake in an error message. Fix it.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Björn Töpel <bjorn.topel@intel.com>
Link: https://lore.kernel.org/bpf/20201203114452.1060017-1-colin.king@canonical.com

This object lives inside the trunner output dir,
i.e. tools/testing/selftests/bpf/no_alu32/btf_data.o

At some point it gets copied into the parent directory during another
part of the build, but that doesn't happen when building
test_progs-no_alu32 from clean.
Signed-off-by: Brendan Jackman <jackmanb@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Link: https://lore.kernel.org/bpf/20201203120850.859170-1-jackmanb@google.com

I've seen a situation, where a process that's under pprof constantly
generates SIGPROF which prevents program loading indefinitely.
The right thing to do probably is to disable signals in the upper
layers while loading, but it still would be nice to get some error from
libbpf instead of an endless loop.

Let's add some small retry limit to the program loading:
try loading the program 5 (arbitrary) times and give up.

v2:
* 10 -> 5 retires (Andrii Nakryiko)
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20201202231332.3923644-1-sdf@google.com

When we added sanitising of map names before loading programs to libbpf, we
still allowed periods in the name. While the kernel will accept these for
the map names themselves, they are not allowed in file names when pinning
maps. This means that bpf_object__pin_maps() will fail if called on an
object that contains internal maps (such as sections .rodata).

Fix this by replacing periods with underscores when constructing map pin
paths. This only affects the paths generated by libbpf when
bpf_object__pin_maps() is called with a path argument. Any pin paths set
by bpf_map__set_pin_path() are unaffected, and it will still be up to the
caller to avoid invalid characters in those.

Fixes: 113e6b7e ("libbpf: Sanitise internal map names so they are not rejected by the kernel")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20201203093306.107676-1-toke@redhat.com

Before this patch, a program with unspecified type
(BPF_PROG_TYPE_UNSPEC) would be passed to the BPF syscall, only to have
the kernel reject it with an opaque invalid argument error. This patch
makes libbpf reject such programs with a nicer error message - in
particular libbpf now tries to diagnose bad ELF section names at both
open time and load time.
Signed-off-by: Andrei Matei <andreimatei1@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20201203043410.59699-1-andreimatei1@gmail.com

KP Singh says:

====================

From: KP Singh <kpsingh@google.com>

# v3 -> v4

* Fix typos.
* Update commit message for the indentation patch.
* Added Andrii's acks.

# v2 -> v3

* Added missing tags.
* Indentation fixes + some other fixes suggested by Andrii.
* Re-indent file to tabs.

The selftest for the bpf_ima_inode_hash helper uses a shell script to
setup the system for ima. While this worked without an issue on recent
desktop distros, it failed on environments with stripped out shells like
busybox which is also used by the bpf CI.

This series fixes the assumptions made on the availablity of certain
command line switches and the expectation that securityfs being mounted
by default.

It also adds the missing kernel config dependencies in
tools/testing/selftests/bpf and, lastly, changes the indentation of
ima_setup.sh to use tabs.
====================
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>

The file was formatted with spaces instead of tabs and went unnoticed
as checkpatch.pl did not complain (probably because this is a shell
script). Re-indent it with tabs to be consistent with other scripts.
Signed-off-by: KP Singh <kpsingh@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20201203191437.666737-5-kpsingh@chromium.org

The ima selftest restricts its scope to a test filesystem image
mounted on a loop device and prevents permanent ima policy changes for
the whole system.

Fixes: 34b82d3a ("bpf: Add a selftest for bpf_ima_inode_hash")
Reported-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: KP Singh <kpsingh@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20201203191437.666737-4-kpsingh@chromium.org

SecurityFS may not be mounted even if it is enabled in the kernel
config. So, check if the mount exists in /proc/mounts by parsing the
file and, if not, mount it on /sys/kernel/security.

Fixes: 34b82d3a ("bpf: Add a selftest for bpf_ima_inode_hash")
Reported-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: KP Singh <kpsingh@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20201203191437.666737-3-kpsingh@chromium.org

losetup on busybox does not output the name of loop device on using
-f with --show. It also doesn't support -j to find the loop devices
for a given backing file. losetup is updated to use "-a" which is
available on busybox.

blkid does not support options (-s and -o) to only display the uuid, so
parse the output instead.

Not all environments have mkfs.ext4, the test requires a loop device
with a backing image file which could formatted with any filesystem.
Update to using mkfs.ext2 which is available on busybox.

Fixes: 34b82d3a ("bpf: Add a selftest for bpf_ima_inode_hash")
Reported-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: KP Singh <kpsingh@google.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20201203191437.666737-2-kpsingh@chromium.org

Mariusz Dudek says:

====================

From: Mariusz Dudek <mariuszx.dudek@intel.com>

This patch series adds support for separation of eBPF program
load and xsk socket creation. In for example a Kubernetes
environment you can have an AF_XDP CNI or daemonset that is
responsible for launching pods that execute an application
using AF_XDP sockets. It is desirable that the pod runs with
as low privileges as possible, CAP_NET_RAW in this case,
and that all operations that require privileges are contained
in the CNI or daemonset.

In this case, you have to be able separate ePBF program load from
xsk socket creation.

Currently, this will not work with the xsk_socket__create APIs
because you need to have CAP_NET_ADMIN privileges to load eBPF
program and CAP_SYS_ADMIN privileges to create update xsk_bpf_maps.
To be exact xsk_set_bpf_maps does not need those privileges but
it takes the prog_fd and xsks_map_fd and those are known only to
process that was loading eBPF program. The api bpf_prog_get_fd_by_id
that looks up the fd of the prog using an prog_id and
bpf_map_get_fd_by_id that looks for xsks_map_fd usinb map_id both
requires CAP_SYS_ADMIN.

With this patch, the pod can be run with CAP_NET_RAW capability
only. In case your umem is larger or equal process limit for
MEMLOCK you need either increase the limit or CAP_IPC_LOCK capability.
Without this patch in case of insufficient rights ENOPERM is
returned by xsk_socket__create.

To resolve this privileges issue two new APIs are introduced:
- xsk_setup_xdp_prog - loads the built in XDP program. It can
also return xsks_map_fd which is needed by unprivileged
process to update xsks_map with AF_XDP socket "fd"
- xsk_sokcet__update_xskmap - inserts an AF_XDP socket into an
xskmap for a particular xsk_socket

Usage example:
int xsk_setup_xdp_prog(int ifindex, int *xsks_map_fd)

int xsk_socket__update_xskmap(struct xsk_socket *xsk, int xsks_map_fd);

Inserts AF_XDP socket "fd" into the xskmap.

The first patch introduces the new APIs. The second patch provides
a new sample applications working as control and modification to
existing xdpsock application to work with less privileges.

This patch set is based on bpf-next commit 97306be4
(Merge branch 'switch to memcg-based memory accounting')

Since v6
- rebase on 97306be4 to resolve RLIMIT conflicts

Since v5
- fixed sample/bpf/xdpsock_user.c to resolve merge conflicts

Since v4
- sample/bpf/Makefile issues fixed

Since v3:
- force_set_map flag removed
- leaking of xsk struct fixed
- unified function error returning policy implemented

Since v2:
- new APIs moved itto LIBBPF_0.3.0 section
- struct bpf_prog_cfg_opts removed
- loading own eBPF program via xsk_setup_xdp_prog functionality removed

Since v1:
- struct bpf_prog_cfg improved for backward/forward compatibility
- API xsk_update_xskmap renamed to xsk_socket__update_xskmap
- commit message formatting fixed
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Introduce a sample program to demonstrate the control and data
plane split. For the control plane part a new program called
xdpsock_ctrl_proc is introduced. For the data plane part, some code
was added to xdpsock_user.c to act as the data plane entity.

Application xdpsock_ctrl_proc works as control entity with sudo
privileges (CAP_SYS_ADMIN and CAP_NET_ADMIN are sufficient) and the
extended xdpsock as data plane entity with CAP_NET_RAW capability
only.

Usage example:

sudo ./samples/bpf/xdpsock_ctrl_proc -i <interface>

sudo ./samples/bpf/xdpsock -i <interface> -q <queue_id>
	-n <interval> -N -l -R
Signed-off-by: Mariusz Dudek <mariuszx.dudek@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/20201203090546.11976-3-mariuszx.dudek@intel.com

Add support for separation of eBPF program load and xsk socket
creation.

This is needed for use-case when you want to privide as little
privileges as possible to the data plane application that will
handle xsk socket creation and incoming traffic.

With this patch the data entity container can be run with only
CAP_NET_RAW capability to fulfill its purpose of creating xsk
socket and handling packages. In case your umem is larger or
equal process limit for MEMLOCK you need either increase the
limit or CAP_IPC_LOCK capability.

To resolve privileges issue two APIs are introduced:

- xsk_setup_xdp_prog - loads the built in XDP program. It can
also return xsks_map_fd which is needed by unprivileged process
to update xsks_map with AF_XDP socket "fd"

- xsk_socket__update_xskmap - inserts an AF_XDP socket into an xskmap
for a particular xsk_socket
Signed-off-by: Mariusz Dudek <mariuszx.dudek@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/20201203090546.11976-2-mariuszx.dudek@intel.com

Add missing newlines and fix polarity of strerror argument.
Signed-off-by: Brendan Jackman <jackmanb@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Link: https://lore.kernel.org/bpf/20201203102234.648540-1-jackmanb@google.com

Splice (copy_file_range) doesn't work on all filesystems. I'm running
test kernels on top of my read-only disk image and it uses plan9 under the
hood. This prevents test_local_storage from successfully passing.

There is really no technical reason to use splice, so lets do
old-school read/write to copy file; this should work in all
environments.
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201202174947.3621989-1-sdf@google.com

Andrii Nakryiko says:

====================

Few follow up improvements to bpftool for split BTF support:
  - emit "name <anon>" for non-named BTFs in `bpftool btf show` command;
  - when dumping /sys/kernel/btf/<module> use /sys/kernel/btf/vmlinux as the
    base BTF, unless base BTF is explicitly specified with -B flag.

This patch set also adds btf__base_btf() getter to access base BTF of the
struct btf.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

In case of working with module's split BTF from /sys/kernel/btf/*,
auto-substitute /sys/kernel/btf/vmlinux as the base BTF. This makes using
bpftool with module BTFs faster and simpler.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201202065244.530571-4-andrii@kernel.org

Add ability to get base BTF. It can be also used to check if BTF is split BTF.
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201202065244.530571-3-andrii@kernel.org

For consistency of output, emit "name <anon>" for BTFs without the name. This
keeps output more consistent and obvious.
Suggested-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20201202065244.530571-2-andrii@kernel.org

Roman Gushchin says:

====================

Currently bpf is using the memlock rlimit for the memory accounting.
This approach has its downsides and over time has created a significant
amount of problems:

1) The limit is per-user, but because most bpf operations are performed
   as root, the limit has a little value.

2) It's hard to come up with a specific maximum value. Especially because
   the counter is shared with non-bpf use cases (e.g. memlock()).
   Any specific value is either too low and creates false failures
   or is too high and useless.

3) Charging is not connected to the actual memory allocation. Bpf code
   should manually calculate the estimated cost and charge the counter,
   and then take care of uncharging, including all fail paths.
   It adds to the code complexity and makes it easy to leak a charge.

4) There is no simple way of getting the current value of the counter.
   We've used drgn for it, but it's far from being convenient.

5) Cryptic -EPERM is returned on exceeding the limit. Libbpf even had
   a function to "explain" this case for users.

6) rlimits are generally considered as (at least partially) obsolete.
   They do not provide a comprehensive system for the control of physical
   resources: memory, cpu, io etc. All resource control developments
   in the recent years were related to cgroups.

In order to overcome these problems let's switch to the memory cgroup-based
memory accounting of bpf objects. With the recent addition of the percpu
memory accounting, now it's possible to provide a comprehensive accounting
of the memory used by bpf programs and maps.

This approach has the following advantages:
1) The limit is per-cgroup and hierarchical. It's way more flexible and allows
   a better control over memory usage by different workloads.

2) The actual memory consumption is taken into account. It happens automatically
   on the allocation time if __GFP_ACCOUNT flags is passed. Uncharging is also
   performed automatically on releasing the memory. So the code on the bpf side
   becomes simpler and safer.

3) There is a simple way to get the current value and statistics.

Cgroup-based accounting adds new requirements:
1) The kernel config should have CONFIG_CGROUPS and CONFIG_MEMCG_KMEM enabled.
   These options are usually enabled, maybe excluding tiny builds for embedded
   devices.
2) The system should have a configured cgroup hierarchy, including reasonable
   memory limits and/or guarantees. Modern systems usually delegate this task
   to systemd or similar task managers.

Without meeting these requirements there are no limits on how much memory bpf
can use and a non-root user is able to hurt the system by allocating too much.
But because per-user rlimits do not provide a functional system to protect
and manage physical resources anyway, anyone who seriously depends on it,
should use cgroups.

When a bpf map is created, the memory cgroup of the process which creates
the map is recorded. Subsequently all memory allocation related to the bpf map
are charged to the same cgroup. It includes allocations made from interrupts
and by any processes. Bpf program memory is charged to the memory cgroup of
a process which loads the program.

The patchset consists of the following parts:
1) 4 mm patches are required on the mm side, otherwise vmallocs cannot be mapped
   to userspace
2) memcg-based accounting for various bpf objects: progs and maps
3) removal of the rlimit-based accounting
4) removal of rlimit adjustments in userspace samples

v9:
  - always charge the saved memory cgroup, by Daniel, Toke and Alexei
  - added bpf_map_kzalloc()
  - rebase and minor fixes

v8:
  - extended the cover letter to be more clear on new requirements, by Daniel
  - an approximate value is provided by map memlock info, by Alexei

v7:
  - introduced bpf_map_kmalloc_node() and bpf_map_alloc_percpu(), by Alexei
  - switched allocations made from an interrupt context to new helpers,
    by Daniel
  - rebase and minor fixes

v6:
  - rebased to the latest version of the remote charging API
  - fixed signatures, added acks

v5:
  - rebased to the latest version of the remote charging API
  - implemented kmem accounting from an interrupt context, by Shakeel
  - rebased to latest changes in mm allowed to map vmallocs to userspace
  - fixed a build issue in kselftests, by Alexei
  - fixed a use-after-free bug in bpf_map_free_deferred()
  - added bpf line info coverage, by Shakeel
  - split bpf map charging preparations into a separate patch

v4:
  - covered allocations made from an interrupt context, by Daniel
  - added some clarifications to the cover letter

v3:
  - droped the userspace part for further discussions/refinements,
    by Andrii and Song

v2:
  - fixed build issue, caused by the remaining rlimit-based accounting
    for sockhash maps
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Since bpf is not using rlimit memlock for the memory accounting
and control, do not change the limit in sample applications.
Signed-off-by: Roman Gushchin <guro@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20201201215900.3569844-35-guro@fb.com