Move BPF map definition and utilities to a common place that easy to be
referenced from both mainline module and platform code.
Bug: 202086915
Test: m; flash; boot
Test: cd system/netd/ && atest
Test: cd packages/modules/Connectivity && atest
Test: m gpuservice_unittest libtimeinstate_test bpf_module_test
CtsAppOpsTestCases libbpf_load_test VtsBootconfigTest
vts_test_binary_bpf_module bpf_benchmark libbpf_load_test
libbpf_android_test
Change-Id: Ib15cf78c2da97bff835fb406c866676eec77c013
Ken Chenauthored and Maciej Żenczykowskicommited 2021-12-04 03:42:55 +00:00
Move waitForProgsLoaded() from libbpf_android.h to a single header file.
A transitional commit of separating headers, utilities from bfploader.
Bug: 202086915
Test: build; flash; cd system/netd; atest
Change-Id: Ief1bcd09e973f952c29859a7a8fba92c8626e32d
Functions in BpfUtils.cpp are trivial, they can be static inlined
in BpfUtils.h.
Bug: 202086915
Test: m; flash; boot; cd system/netd && atest
Test: m gpuservice_unittest libtimeinstate_test bpf_module_test
CtsAppOpsTestCases libmeminfo_test VtsBootconfigTest
vts_test_binary_bpf_module
Change-Id: Ie1ece23b6fc9a4db5fc95930209a10da1e528cb5
Needs to be used by dns mainline module tests on R devices (Pixel 2).
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I2974a5368f4a9d744e30be9148aefac117d09acb
Maciej Żenczykowskicommited 2021-03-11 19:52:30 -08:00
Instead of doing 16/8/8 bits for major/minor/subver macro,
switch to 8/8/16 bits, since this gives plenty of space:
- major versions are bumped every few years, so 256 last forever
- minor versions are bumped every few months (~5-6 per year),
so 256 lasts for decades, additionally Linus Torvalds doesn't
like big numbers, and eventually bumps the major instead
- sub versions are the problematic ones, because they're bumped
on every LTS security release, however even at one release per day,
16 bits lasts for 180 years
Note: before this change 4.9.256 was treated as equivalent to 4.10.0.
Luckily all our tests were only ever (by chance) against the LTS release
bases (ie. 4.9.0, 4.14.0, 4.19.0, 5.4.0) and thus we wouldn't have hit
an actually failure until 4.9.(5*256) === 4.14.0 which would have caused
the bpfloader to attempt to load bpf code requiring 4.14 on a 4.9 kernel,
and resulted in failure of device to successfully boot.
Test: atest, TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I0535ce081967af092f3342c0abfd05a1e5028eb6
Maciej Żenczykowskicommited 2021-02-11 16:05:21 -08:00
This makes it harder for system and mainline bpf programs to
conflict with each other, thus enabling easier mainline module
updates.
Test: builds and boots, atest, TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ic1b69bb1ddc4a10bd4861dadbd6b97e2520c532d
Maciej Żenczykowskicommited 2021-01-20 05:17:14 +00:00
Cause that's what it actually does...
Bug: 173167302
Test: atest BpfMapTest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I8445cc4fe68f24bb3b26a678b7731b0c518e80f9
Maciej Żenczykowskicommited 2020-11-30 17:06:58 -08:00
This is useful for critical functions with fallbacks, but
may even be useful for non-critical functions, where a function
in the middle of the file may fail to load, but you still want
other (later) functions to be attempted.
Critical applies to the entire .c file (or to be more correct to
the entire resulting .o). Optional applies to a specific section
of that .o (ie. a specific individual function).
This new optional attribute is necessary to be able to declare
a .c/.o file critical even if *some* of the individual functions
might fail to load due to missing kernel patches.
(Note: we currently have no way to specify a map as optional)
Critical guarantees that all non-optional programs, and all maps,
have been created, pinned, chowned, and chmoded successfully
(or that they already existed).
For an example of use see:
system/netd/bpf_progs/offload.c
(while at it also add retrieveProgram() and mapRetrieve{RW,RO,WO}()
helpers to BpfUtils.h)
Test: builds, atest, see paired netd change for extra details
Bug: 150040815
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I50b292c061b05fc8f4b4b8574f128345c45c78db
Maciej Żenczykowskicommited 2020-06-17 10:44:19 +00:00
Reasoning: critical load failures for netd bpf programs will just result
in unpredictable behaviour later on. For example netd/systemserver
crash loops.
Test: builds
Bug: 150040815
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Iefa01e60cd0a9a223e96411726a199bfb4857a5a
Maciej Żenczykowskicommited 2020-06-15 08:01:50 +00:00
Fixes:
system/bpf/libbpf_android/include/bpf/BpfMap.h:132:10: warning: operator=() does not handle self-assignment properly [cert-oop54-cpp]
void operator=(const BpfMap<Key, Value>& other) {
^
Matches the self-assignment check from
https://clang.llvm.org/extra/clang-tidy/checks/bugprone-unhandled-self-assignment.html
This isn't needed here, since the code was already correct, but
it does prevent a pointless newfd = dup(fd); close(fd); sequence.
Test: builds
Bug: 153035880
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ia57f95d4ab180783c97db7e6f8d929f822c8958d
Maciej Żenczykowskicommited 2020-04-24 08:15:00 -07:00
Some ebpf code cannot be loaded on too old kernels.
Sometimes we want a different - more advanced - version of an ebpf program
to be loaded on a newer kernel.
Test: build, atest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I9e93e7246951916e6d60544575337a7a19c82886
Maciej Żenczykowskicommited 2020-02-19 15:46:23 -08:00
(while at it also remove 3 spurious this-> prefixes)
Test: build, atest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I2419926b555734d7f5e052fb9e0686c52bdb32d0
Maciej Żenczykowskicommited 2020-02-19 13:29:02 -08:00
This is to allow fixing a broken OffloadUtils netd test, but this will
also be useful for future support to make programs per kernel version.
We make the string versions of the enum more readable, because there
is exactly one user:
adb shell dumpsys netd | egrep -i bpf
BPF module status: enabled
BPF support level: Extended [4.14]
...
Test: build, atest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I843bd7355703a78acc8343ca7f647ce9366927eb
Maciej Żenczykowskicommited 2020-02-18 04:46:58 +00:00
(being moved from //system/netd/tests/bpf_base_test.cpp)
Test: build, atest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I7a38eff6d86a1bcd6dde193fd360d198bd57a900
Maciej Żenczykowskicommited 2020-02-12 04:15:29 -08:00
Per Bernie: "in C++, inline implies internal linkage, you can drop static"
Test: build, atest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I139af1d826d4982a1480fbee433fd9a743509786
Maciej Żenczykowskicommited 2020-02-11 16:43:15 -08:00
Long term we don't even want reset() to exist.
It's simply not useful since only the bpfloader can create maps,
and thus all we can do is construct from pre-existing pinned bpf
path - and we should never destroy/unpin/delete these.
Similarly the move constructor is probably spurious.
Basically the only real use for reset() is in tests,
and even there the overall utility is not clear.
Hence, in the mean time, make it harder to use incorrectly,
and harder to use in general.
To do that we get rid of reset() -> you must call reset(-1)
and we eliminate the ability to call reset(unique_fd)
which has unclear fd-leaking semantics.
Also remove a spot where it is called spuriously.
Test: build, atest
Bug: 129773125
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ia018224aca0005ab68809b350595485ccd3f1bd3
Maciej Żenczykowskicommited 2020-01-30 08:59:53 +00:00
(and remove the old straight from fd constructor)
Note that BpfMapRO is not yet truly compile time read-only,
since it still has methods that can modify things (although those
modifications will of course fail due to the map fd being r/o).
Test: build, atest with followup fixes
Bug: 129773125
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I70b511e3f0a62cd4d9d5a923e658b1086337ec4c
Maciej Żenczykowskicommited 2020-01-30 08:50:58 +00:00
Remove libnetdutils dependency because bpf is being used for things other
than networking these days, and we don't want to make libnetdutils
vendor-available in the future. libbase provides an alternative type now.
Bug: 140330870
Test: atest libbpf_android_test netd_integration_test netd_unit_test
libnetdbpf_test bpf_module_test
Change-Id: I72ae8cd7f58a49bfc7dcb914a332a4c4bad5dea5
Steven Morelandcommited 2020-01-21 10:47:33 -08:00
When accessing maps keys are always const/read-only.
Similarly for the value when it is being stored in the map.
This eliminates four const_casts.
Test: build, atest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Icc57b46a8e9e81b5b06264260bfbf34a4ba9105d
Maciej Żenczykowskicommited 2020-01-18 19:32:09 -08:00
(due to switch to bpfloader from loading via netd)
Test: build, atest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ic95cdfde26b8e7b905551a7fd2b465e0cf99bbce
Maciej Żenczykowskiauthored and Steven Morelandcommited 2020-01-14 09:10:32 -08:00
Creating a function for this rather than having people copy/paste a loop
there.
It's very intentional that there is no timeout here. If bpfloader is
down, there is a serious problem, and we shouldn't let code move forward
in a state where it may not have been tested.
Bug: 140330870
Test: boot & use bpf/netd
Change-Id: I358d1fcf77ad5406e294dd57307bc411e2c8ad6a
Steven Morelandcommited 2020-01-10 15:37:53 -08:00
This is part of 3 commits in 3 diff git repos, the main one is:
https://android-review.googlesource.com/c/platform/system/netd/+/1200479
"share eBPF struct definitions between ebpf and C++ netd"
The struct definitions are being moved to system/netd in:
libnetdbpf/include/netdbpf/bpf_shared.h
(they also become typedefs and more consistent naming is used)
Test: builds
Bug: 146787904
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I0200bce35f910414b96d802cd316717b4307b950
Maciej Żenczykowskicommited 2020-01-06 23:28:06 +00:00
Ken Chen
authored and
Maciej Żenczykowski
commited 
Treehugger Robot
authored and
Gerrit Code Review
commited 
Hungming Chen
commited 
markchien
commited 
Bernie Innocenti
commited 
Tom Cherry
authored and
Steven Moreland
commited