Skip to content

K
Kernel

openeuler / Kernel
1 年多 前同步成功

通知 8
Star 0
Fork 0
K
Kernel
切换分支/标签
  1. 02 6月, 2020 1 次提交
    • A
      bpf: Implement BPF ring buffer and verifier support for it · 457f4436
      由 Andrii Nakryiko 提交于 
      This commit adds a new MPSC ring buffer implementation into BPF ecosystem,
which allows multiple CPUs to submit data to a single shared ring buffer. On
the consumption side, only single consumer is assumed.

Motivation
----------
There are two distinctive motivators for this work, which are not satisfied by
existing perf buffer, which prompted creation of a new ring buffer
implementation.
  - more efficient memory utilization by sharing ring buffer across CPUs;
  - preserving ordering of events that happen sequentially in time, even
  across multiple CPUs (e.g., fork/exec/exit events for a task).

These two problems are independent, but perf buffer fails to satisfy both.
Both are a result of a choice to have per-CPU perf ring buffer.  Both can be
also solved by having an MPSC implementation of ring buffer. The ordering
problem could technically be solved for perf buffer with some in-kernel
counting, but given the first one requires an MPSC buffer, the same solution
would solve the second problem automatically.

Semantics and APIs
------------------
Single ring buffer is presented to BPF programs as an instance of BPF map of
type BPF_MAP_TYPE_RINGBUF. Two other alternatives considered, but ultimately
rejected.

One way would be to, similar to BPF_MAP_TYPE_PERF_EVENT_ARRAY, make
BPF_MAP_TYPE_RINGBUF could represent an array of ring buffers, but not enforce
"same CPU only" rule. This would be more familiar interface compatible with
existing perf buffer use in BPF, but would fail if application needed more
advanced logic to lookup ring buffer by arbitrary key. HASH_OF_MAPS addresses
this with current approach. Additionally, given the performance of BPF
ringbuf, many use cases would just opt into a simple single ring buffer shared
among all CPUs, for which current approach would be an overkill.

Another approach could introduce a new concept, alongside BPF map, to
represent generic "container" object, which doesn't necessarily have key/value
interface with lookup/update/delete operations. This approach would add a lot
of extra infrastructure that has to be built for observability and verifier
support. It would also add another concept that BPF developers would have to
familiarize themselves with, new syntax in libbpf, etc. But then would really
provide no additional benefits over the approach of using a map.
BPF_MAP_TYPE_RINGBUF doesn't support lookup/update/delete operations, but so
doesn't few other map types (e.g., queue and stack; array doesn't support
delete, etc).

The approach chosen has an advantage of re-using existing BPF map
infrastructure (introspection APIs in kernel, libbpf support, etc), being
familiar concept (no need to teach users a new type of object in BPF program),
and utilizing existing tooling (bpftool). For common scenario of using
a single ring buffer for all CPUs, it's as simple and straightforward, as
would be with a dedicated "container" object. On the other hand, by being
a map, it can be combined with ARRAY_OF_MAPS and HASH_OF_MAPS map-in-maps to
implement a wide variety of topologies, from one ring buffer for each CPU
(e.g., as a replacement for perf buffer use cases), to a complicated
application hashing/sharding of ring buffers (e.g., having a small pool of
ring buffers with hashed task's tgid being a look up key to preserve order,
but reduce contention).

Key and value sizes are enforced to be zero. max_entries is used to specify
the size of ring buffer and has to be a power of 2 value.

There are a bunch of similarities between perf buffer
(BPF_MAP_TYPE_PERF_EVENT_ARRAY) and new BPF ring buffer semantics:
  - variable-length records;
  - if there is no more space left in ring buffer, reservation fails, no
    blocking;
  - memory-mappable data area for user-space applications for ease of
    consumption and high performance;
  - epoll notifications for new incoming data;
  - but still the ability to do busy polling for new data to achieve the
    lowest latency, if necessary.

BPF ringbuf provides two sets of APIs to BPF programs:
  - bpf_ringbuf_output() allows to *copy* data from one place to a ring
    buffer, similarly to bpf_perf_event_output();
  - bpf_ringbuf_reserve()/bpf_ringbuf_commit()/bpf_ringbuf_discard() APIs
    split the whole process into two steps. First, a fixed amount of space is
    reserved. If successful, a pointer to a data inside ring buffer data area
    is returned, which BPF programs can use similarly to a data inside
    array/hash maps. Once ready, this piece of memory is either committed or
    discarded. Discard is similar to commit, but makes consumer ignore the
    record.

bpf_ringbuf_output() has disadvantage of incurring extra memory copy, because
record has to be prepared in some other place first. But it allows to submit
records of the length that's not known to verifier beforehand. It also closely
matches bpf_perf_event_output(), so will simplify migration significantly.

bpf_ringbuf_reserve() avoids the extra copy of memory by providing a memory
pointer directly to ring buffer memory. In a lot of cases records are larger
than BPF stack space allows, so many programs have use extra per-CPU array as
a temporary heap for preparing sample. bpf_ringbuf_reserve() avoid this needs
completely. But in exchange, it only allows a known constant size of memory to
be reserved, such that verifier can verify that BPF program can't access
memory outside its reserved record space. bpf_ringbuf_output(), while slightly
slower due to extra memory copy, covers some use cases that are not suitable
for bpf_ringbuf_reserve().

The difference between commit and discard is very small. Discard just marks
a record as discarded, and such records are supposed to be ignored by consumer
code. Discard is useful for some advanced use-cases, such as ensuring
all-or-nothing multi-record submission, or emulating temporary malloc()/free()
within single BPF program invocation.

Each reserved record is tracked by verifier through existing
reference-tracking logic, similar to socket ref-tracking. It is thus
impossible to reserve a record, but forget to submit (or discard) it.

bpf_ringbuf_query() helper allows to query various properties of ring buffer.
Currently 4 are supported:
  - BPF_RB_AVAIL_DATA returns amount of unconsumed data in ring buffer;
  - BPF_RB_RING_SIZE returns the size of ring buffer;
  - BPF_RB_CONS_POS/BPF_RB_PROD_POS returns current logical possition of
    consumer/producer, respectively.
Returned values are momentarily snapshots of ring buffer state and could be
off by the time helper returns, so this should be used only for
debugging/reporting reasons or for implementing various heuristics, that take
into account highly-changeable nature of some of those characteristics.

One such heuristic might involve more fine-grained control over poll/epoll
notifications about new data availability in ring buffer. Together with
BPF_RB_NO_WAKEUP/BPF_RB_FORCE_WAKEUP flags for output/commit/discard helpers,
it allows BPF program a high degree of control and, e.g., more efficient
batched notifications. Default self-balancing strategy, though, should be
adequate for most applications and will work reliable and efficiently already.

Design and implementation
-------------------------
This reserve/commit schema allows a natural way for multiple producers, either
on different CPUs or even on the same CPU/in the same BPF program, to reserve
independent records and work with them without blocking other producers. This
means that if BPF program was interruped by another BPF program sharing the
same ring buffer, they will both get a record reserved (provided there is
enough space left) and can work with it and submit it independently. This
applies to NMI context as well, except that due to using a spinlock during
reservation, in NMI context, bpf_ringbuf_reserve() might fail to get a lock,
in which case reservation will fail even if ring buffer is not full.

The ring buffer itself internally is implemented as a power-of-2 sized
circular buffer, with two logical and ever-increasing counters (which might
wrap around on 32-bit architectures, that's not a problem):
  - consumer counter shows up to which logical position consumer consumed the
    data;
  - producer counter denotes amount of data reserved by all producers.

Each time a record is reserved, producer that "owns" the record will
successfully advance producer counter. At that point, data is still not yet
ready to be consumed, though. Each record has 8 byte header, which contains
the length of reserved record, as well as two extra bits: busy bit to denote
that record is still being worked on, and discard bit, which might be set at
commit time if record is discarded. In the latter case, consumer is supposed
to skip the record and move on to the next one. Record header also encodes
record's relative offset from the beginning of ring buffer data area (in
pages). This allows bpf_ringbuf_commit()/bpf_ringbuf_discard() to accept only
the pointer to the record itself, without requiring also the pointer to ring
buffer itself. Ring buffer memory location will be restored from record
metadata header. This significantly simplifies verifier, as well as improving
API usability.

Producer counter increments are serialized under spinlock, so there is
a strict ordering between reservations. Commits, on the other hand, are
completely lockless and independent. All records become available to consumer
in the order of reservations, but only after all previous records where
already committed. It is thus possible for slow producers to temporarily hold
off submitted records, that were reserved later.

Reservation/commit/consumer protocol is verified by litmus tests in
Documentation/litmus-test/bpf-rb.

One interesting implementation bit, that significantly simplifies (and thus
speeds up as well) implementation of both producers and consumers is how data
area is mapped twice contiguously back-to-back in the virtual memory. This
allows to not take any special measures for samples that have to wrap around
at the end of the circular buffer data area, because the next page after the
last data page would be first data page again, and thus the sample will still
appear completely contiguous in virtual memory. See comment and a simple ASCII
diagram showing this visually in bpf_ringbuf_area_alloc().

Another feature that distinguishes BPF ringbuf from perf ring buffer is
a self-pacing notifications of new data being availability.
bpf_ringbuf_commit() implementation will send a notification of new record
being available after commit only if consumer has already caught up right up
to the record being committed. If not, consumer still has to catch up and thus
will see new data anyways without needing an extra poll notification.
Benchmarks (see tools/testing/selftests/bpf/benchs/bench_ringbuf.c) show that
this allows to achieve a very high throughput without having to resort to
tricks like "notify only every Nth sample", which are necessary with perf
buffer. For extreme cases, when BPF program wants more manual control of
notifications, commit/discard/output helpers accept BPF_RB_NO_WAKEUP and
BPF_RB_FORCE_WAKEUP flags, which give full control over notifications of data
availability, but require extra caution and diligence in using this API.

Comparison to alternatives
--------------------------
Before considering implementing BPF ring buffer from scratch existing
alternatives in kernel were evaluated, but didn't seem to meet the needs. They
largely fell into few categores:
  - per-CPU buffers (perf, ftrace, etc), which don't satisfy two motivations
    outlined above (ordering and memory consumption);
  - linked list-based implementations; while some were multi-producer designs,
    consuming these from user-space would be very complicated and most
    probably not performant; memory-mapping contiguous piece of memory is
    simpler and more performant for user-space consumers;
  - io_uring is SPSC, but also requires fixed-sized elements. Naively turning
    SPSC queue into MPSC w/ lock would have subpar performance compared to
    locked reserve + lockless commit, as with BPF ring buffer. Fixed sized
    elements would be too limiting for BPF programs, given existing BPF
    programs heavily rely on variable-sized perf buffer already;
  - specialized implementations (like a new printk ring buffer, [0]) with lots
    of printk-specific limitations and implications, that didn't seem to fit
    well for intended use with BPF programs.

  [0] https://lwn.net/Articles/779550/Signed-off-by: NAndrii Nakryiko <andriin@fb.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200529075424.3139988-2-andriin@fb.comSigned-off-by: NAlexei Starovoitov <ast@kernel.org>
      457f4436
  3. 15 5月, 2020 1 次提交
    • A
      bpf: Implement CAP_BPF · 2c78ee89
      由 Alexei Starovoitov 提交于 
      Implement permissions as stated in uapi/linux/capability.h
In order to do that the verifier allow_ptr_leaks flag is split
into four flags and they are set as:
  env->allow_ptr_leaks = bpf_allow_ptr_leaks();
  env->bypass_spec_v1 = bpf_bypass_spec_v1();
  env->bypass_spec_v4 = bpf_bypass_spec_v4();
  env->bpf_capable = bpf_capable();

The first three currently equivalent to perfmon_capable(), since leaking kernel
pointers and reading kernel memory via side channel attacks is roughly
equivalent to reading kernel memory with cap_perfmon.

'bpf_capable' enables bounded loops, precision tracking, bpf to bpf calls and
other verifier features. 'allow_ptr_leaks' enable ptr leaks, ptr conversions,
subtraction of pointers. 'bypass_spec_v1' disables speculative analysis in the
verifier, run time mitigations in bpf array, and enables indirect variable
access in bpf programs. 'bypass_spec_v4' disables emission of sanitation code
by the verifier.

That means that the networking BPF program loaded with CAP_BPF + CAP_NET_ADMIN
will have speculative checks done by the verifier and other spectre mitigation
applied. Such networking BPF program will not be able to leak kernel pointers
and will not be able to access arbitrary kernel memory.
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20200513230355.7858-3-alexei.starovoitov@gmail.com
      2c78ee89
  5. 14 5月, 2020 4 次提交
  7. 10 5月, 2020 8 次提交
  9. 02 5月, 2020 1 次提交
    • S
      bpf: Sharing bpf runtime stats with BPF_ENABLE_STATS · d46edd67
      由 Song Liu 提交于 
      Currently, sysctl kernel.bpf_stats_enabled controls BPF runtime stats.
Typical userspace tools use kernel.bpf_stats_enabled as follows:

  1. Enable kernel.bpf_stats_enabled;
  2. Check program run_time_ns;
  3. Sleep for the monitoring period;
  4. Check program run_time_ns again, calculate the difference;
  5. Disable kernel.bpf_stats_enabled.

The problem with this approach is that only one userspace tool can toggle
this sysctl. If multiple tools toggle the sysctl at the same time, the
measurement may be inaccurate.

To fix this problem while keep backward compatibility, introduce a new
bpf command BPF_ENABLE_STATS. On success, this command enables stats and
returns a valid fd. BPF_ENABLE_STATS takes argument "type". Currently,
only one type, BPF_STATS_RUN_TIME, is supported. We can extend the
command to support other types of stats in the future.

With BPF_ENABLE_STATS, user space tool would have the following flow:

  1. Get a fd with BPF_ENABLE_STATS, and make sure it is valid;
  2. Check program run_time_ns;
  3. Sleep for the monitoring period;
  4. Check program run_time_ns again, calculate the difference;
  5. Close the fd.
Signed-off-by: NSong Liu <songliubraving@fb.com>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200430071506.1408910-2-songliubraving@fb.com
      d46edd67
  11. 29 4月, 2020 3 次提交
  13. 27 4月, 2020 1 次提交
  15. 26 4月, 2020 2 次提交
  17. 31 3月, 2020 1 次提交
    • A
      bpf: Implement bpf_link-based cgroup BPF program attachment · af6eea57
      由 Andrii Nakryiko 提交于 
      Implement new sub-command to attach cgroup BPF programs and return FD-based
bpf_link back on success. bpf_link, once attached to cgroup, cannot be
replaced, except by owner having its FD. Cgroup bpf_link supports only
BPF_F_ALLOW_MULTI semantics. Both link-based and prog-based BPF_F_ALLOW_MULTI
attachments can be freely intermixed.

To prevent bpf_cgroup_link from keeping cgroup alive past the point when no
BPF program can be executed, implement auto-detachment of link. When
cgroup_bpf_release() is called, all attached bpf_links are forced to release
cgroup refcounts, but they leave bpf_link otherwise active and allocated, as
well as still owning underlying bpf_prog. This is because user-space might
still have FDs open and active, so bpf_link as a user-referenced object can't
be freed yet. Once last active FD is closed, bpf_link will be freed and
underlying bpf_prog refcount will be dropped. But cgroup refcount won't be
touched, because cgroup is released already.

The inherent race between bpf_cgroup_link release (from closing last FD) and
cgroup_bpf_release() is resolved by both operations taking cgroup_mutex. So
the only additional check required is when bpf_cgroup_link attempts to detach
itself from cgroup. At that time we need to check whether there is still
cgroup associated with that link. And if not, exit with success, because
bpf_cgroup_link was already successfully detached.
Signed-off-by: NAndrii Nakryiko <andriin@fb.com>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>
Acked-by: NRoman Gushchin <guro@fb.com>
Link: https://lore.kernel.org/bpf/20200330030001.2312810-2-andriin@fb.com
      af6eea57
  19. 30 3月, 2020 1 次提交
  21. 28 3月, 2020 2 次提交
    • D
      bpf: Enable bpf cgroup hooks to retrieve cgroup v2 and ancestor id · 0f09abd1
      由 Daniel Borkmann 提交于 
      Enable the bpf_get_current_cgroup_id() helper for connect(), sendmsg(),
recvmsg() and bind-related hooks in order to retrieve the cgroup v2
context which can then be used as part of the key for BPF map lookups,
for example. Given these hooks operate in process context 'current' is
always valid and pointing to the app that is performing mentioned
syscalls if it's subject to a v2 cgroup. Also with same motivation of
commit 77236281 ("bpf: Introduce bpf_skb_ancestor_cgroup_id helper")
enable retrieval of ancestor from current so the cgroup id can be used
for policy lookups which can then forbid connect() / bind(), for example.
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/d2a7ef42530ad299e3cbb245e6c12374b72145ef.1585323121.git.daniel@iogearbox.net
      0f09abd1
    • D
      bpf: Add netns cookie and enable it for bpf cgroup hooks · f318903c
      由 Daniel Borkmann 提交于 
      In Cilium we're mainly using BPF cgroup hooks today in order to implement
kube-proxy free Kubernetes service translation for ClusterIP, NodePort (*),
ExternalIP, and LoadBalancer as well as HostPort mapping [0] for all traffic
between Cilium managed nodes. While this works in its current shape and avoids
packet-level NAT for inter Cilium managed node traffic, there is one major
limitation we're facing today, that is, lack of netns awareness.

In Kubernetes, the concept of Pods (which hold one or multiple containers)
has been built around network namespaces, so while we can use the global scope
of attaching to root BPF cgroup hooks also to our advantage (e.g. for exposing
NodePort ports on loopback addresses), we also have the need to differentiate
between initial network namespaces and non-initial one. For example, ExternalIP
services mandate that non-local service IPs are not to be translated from the
host (initial) network namespace as one example. Right now, we have an ugly
work-around in place where non-local service IPs for ExternalIP services are
not xlated from connect() and friends BPF hooks but instead via less efficient
packet-level NAT on the veth tc ingress hook for Pod traffic.

On top of determining whether we're in initial or non-initial network namespace
we also have a need for a socket-cookie like mechanism for network namespaces
scope. Socket cookies have the nice property that they can be combined as part
of the key structure e.g. for BPF LRU maps without having to worry that the
cookie could be recycled. We are planning to use this for our sessionAffinity
implementation for services. Therefore, add a new bpf_get_netns_cookie() helper
which would resolve both use cases at once: bpf_get_netns_cookie(NULL) would
provide the cookie for the initial network namespace while passing the context
instead of NULL would provide the cookie from the application's network namespace.
We're using a hole, so no size increase; the assignment happens only once.
Therefore this allows for a comparison on initial namespace as well as regular
cookie usage as we have today with socket cookies. We could later on enable
this helper for other program types as well as we would see need.

  (*) Both externalTrafficPolicy={Local|Cluster} types
  [0] https://github.com/cilium/cilium/blob/master/bpf/bpf_sock.cSigned-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/c47d2346982693a9cf9da0e12690453aded4c788.1585323121.git.daniel@iogearbox.net
      f318903c
  23. 18 3月, 2020 1 次提交
  25. 14 3月, 2020 11 次提交
  27. 13 3月, 2020 1 次提交
  29. 11 3月, 2020 1 次提交
    • A
      bpf: Add bpf_link_new_file that doesn't install FD · babf3164
      由 Andrii Nakryiko 提交于 
      Add bpf_link_new_file() API for cases when we need to ensure anon_inode is
successfully created before we proceed with expensive BPF program attachment
procedure, which will require equally (if not more so) expensive and
potentially failing compensation detachment procedure just because anon_inode
creation failed. This API allows to simplify code by ensuring first that
anon_inode is created and after BPF program is attached proceed with
fd_install() that can't fail.

After anon_inode file is created, link can't be just kfree()'d anymore,
because its destruction will be performed by deferred file_operations->release
call. For this, bpf_link API required specifying two separate operations:
release() and dealloc(), former performing detachment only, while the latter
frees memory used by bpf_link itself. dealloc() needs to be specified, because
struct bpf_link is frequently embedded into link type-specific container
struct (e.g., struct bpf_raw_tp_link), so bpf_link itself doesn't know how to
properly free the memory. In case when anon_inode file was successfully
created, but subsequent BPF attachment failed, bpf_link needs to be marked as
"defunct", so that file's release() callback will perform only memory
deallocation, but no detachment.

Convert raw tracepoint and tracing attachment to new API and eliminate
detachment from error handling path.
Signed-off-by: NAndrii Nakryiko <andriin@fb.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>
Acked-by: NJohn Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200309231051.1270337-1-andriin@fb.com
      babf3164
  31. 10 3月, 2020 1 次提交