bpf_obj_get(path) is entirely equivalent to mapRetrieveRW(path)
See implementation in frameworks/libs/net common/native/bpf_syscall_wrappers/include/BpfSyscallWrappers.h:
inline int bpfFdGet(const char* pathname, uint32_t flag) {
return bpf(BPF_OBJ_GET, {
.pathname = ptr_to_u64(pathname),
.file_flags = flag,
});
}
inline int mapRetrieve(const char* pathname, uint32_t flag) { return bpfFdGet(pathname, flag); }
inline int mapRetrieveRW(const char* pathname) { return mapRetrieve(pathname, 0); }
inline int mapRetrieveRO(const char* pathname) { return mapRetrieve(pathname, BPF_F_RDONLY); }
inline int mapRetrieveWO(const char* pathname) { return mapRetrieve(pathname, BPF_F_WRONLY); }
inline int retrieveProgram(const char* pathname) { return bpfFdGet(pathname, BPF_F_RDONLY); }
However, this requires selinux file:write access which bpfloader
currently lacks, ie. we would need:
system/sepolicy private/bpfloader.te:
-allow bpfloader bpffs_type:file { create read rename setattr };
+allow bpfloader bpffs_type:file { create read rename setattr write };
Switching from mapRetrieveRW() to mapRetrieveRO() eliminates this problem.
BpfLoader itself does not need to write to the maps, while bpf program write
access is controlled by a different bit (it is independent of the r/w access
bits of the file descriptor itself).
Verified by re-opening the maps after pinning them.
Bug: 218408035
Bug: 237716689
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ica7ac8ee48d4a73e5f92dbf47cd441c3bfba38cf
Maciej Żenczykowskicommited 2022-07-01 15:56:21 +00:00
BTF support was added to bpfloader during Android T dev cycle.
As it causes bpfloader boot time process to shell out to a new
btfloader subprocess for every bpf.o file with BTF debugging
information compiled in, I'm worried this might have unforeseen
consequences - things like crashes or boot time bpfloader cpu
regressions.
However, BTF is exceedingly useful for debugging,
and it would be a huge pity if we were forced to disable it
and keep it disabled in mainline tethering module just to support
Android T devices for the next 5+ years.
It would also be a pity if the bpf.o files in prebuilt mainline
tethering apex in dev branches (like tm-dev) did not include
BTF debug information simply due to compatibility with older
kernels or OSes, since this would require rebuilding the module
everytime BTF information could be of use.
One of the things functioning BTF enables is 'cat /sys/fs/bpf/map_*'
as root on a userdebug build. Among other things this can be
used to verify that in kernel bpf map state matches that dumped
by the mainline module's pretty dump code.
Even if there's issues wrt. BTF in Android T (or on older kernels),
we can always fix them in Android U or later (for example build
btfloader into bpfloader to avoid exec overhead, etc...).
Bug: 218408035
Bug: 230585250
Bug: 235559605
Test: TreeHugger, cuttlefish devices boots, and:
adb root && adb shell cat /sys/fs/bpf/map_time_in_state_cpu_last_pid_map
continues to show information which is available due to BTF debug provided
in /system/etc/bpf/time_in_state.o file (due to "btf: true" in Android.bp)
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I553e90e0414453f8f3aaca8cf05d5decc8b911a2
(cherry picked from commit 37b3d06333)
Merged-In: I553e90e0414453f8f3aaca8cf05d5decc8b911a2
Maciej Żenczykowskicommited 2022-06-23 16:56:49 -07:00
This reverts commit 8e5e239dd0a7856e486a9e2d772ef3a9e38021eb.
No longer required now that selinux change has landed.
(while we're at it bump the bpfloader version to v0.19 just
to make sure we can tell these apart in the future)
Bug: 218408035
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I0b3349ef908bbfd225b8b7f83b2a4a8870c1e409
Maciej Żenczykowskicommited 2022-06-22 19:15:41 +00:00
(it requires bpfloader to be granted rename priv by selinux)
Bug: 218408035
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ic27d0e5f3c6b78db39c6742fe9e1955f65d2b1f5
Maciej Żenczykowskicommited 2022-06-21 09:42:48 -07:00
This is needed to make it safe to share maps across different .o's.
Bug: 218408035
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I9e8a5893ed4f91354f6544be587b10a97d179de6
Maciej Żenczykowskicommited 2022-06-18 00:33:31 -07:00
For use by:
- maps netd should have read but not write access to
(needed due to netd being root with DAC_OVERRIDE,
and thus not obeying standard unix permissions)
- programs that netd should have access to but
not netutils_wrappers (which due to being able to
run iptables, needs access to xt_bpf programs)
Bug: 218408035
Test: booted on cuttlefish
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I72b106692a25077ff54252fd93db81f46b52125d
Maciej Żenczykowskicommited 2022-06-17 18:08:00 -07:00
combined with content of /proc/${pid}/fd/${fd}
and /proc/${pid}/fdinfo/${fd} this allows figuring
out which programs/maps a process continues to
have open
See:
adbz logcat -d | egrep '(map|prog) .* id'
adbz root
adbz_bpf_progs_and_maps
Bug: 230880517
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I66c7fa12d079257486569105368bf1eea2d3ab0c
Maciej Żenczykowskicommited 2022-05-20 17:20:36 -07:00
(no selinux consequences to this in and of itself,
though required for follow up selinux changes)
Test: TreeHugger, manual
Bug: 218408035
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Id5ef7f3b119743ef26b062068756c4e0f754e694
Maciej Żenczykowskicommited 2022-04-21 06:45:25 -07:00
This also fixes a permissions issue if a non-root user is set. The read
permissions should be set before the file is set as non-root to ensure
that the permissions can be set without error.
Bump the BPF loader version.
Bug: 203462310
Test: Ensure that vendor skfilter bpf programs can load
Change-Id: Ib6b9a64d8652ff464c9d4d734bb8ae351673b6ce
Allow vendors to use bpf programs, but limit to tracepoints
for now (other types of programs, for instance skfilter, aren't
safe to expose, because the kernel gives us limited ways to
control which resources can have BPF programs attached, and
some shared resources only support a single BPF program at an
attach point).
Bug: 140330870
Bug: 162057235
Test: install bpf program to /vendor/etc/bpf/ and use it.
Test: atest libbpf_load_test
Change-Id: I6c876fe52739c38db73689ffd784167e7d35d58a
Steven Morelandcommited 2022-02-09 00:59:37 +00:00
readCodeSections implicitly assumes that a function's name will be the
first symbol in its section to appear in the symbol table, but this is
not guaranteed. In particular, when compiling with -g the section name
can appear before it, resulting in a failure to find the bpf_prog_def
information for that section's program.
Add a check for the STT_FUNC symbol type to narrow our search to the
function name symbol we actually want.
Also, add a test to check that a program with a minimum kernel version
of infinity will not be loaded, to verify that bpfloader successfully
finds bpf_prog_def information.
Test: boot & confirm progs all have correct owners & permissions
Test: build time_in_state.c with -g and confirm owner & permissions
set correctly
Test: libbpf_load_test now passes iff this fix is applied
Signed-off-by: Connor O'Brien <connoro@google.com>
Change-Id: I31db0c50d3adeed86ee5810b1a0fb972f56fdb47
btfloader uses upstream libbpf & libelf to parse the BTF sections of a
bpf .o file, load BTF information into the kernel, and identify the
BTF type ids corresponding to each map's keys and values.
Extend bpfloader to run btfloader on any .o file with a .BTF
section. We pass btfloader socket and pipe fds, which it will use to
send back a BTF fd and the key & value type ids, respectively.
To ensure a btfloader bug can't block bpfloader indefinitely, wait for
btfloader to run with a short timeout and use nonblocking socket &
pipe for communication.
Finally, add a check to BpfLoadTest to verify that map BTF is loaded
as expected.
Test: build & boot; existing programs load successfully
Test: libbpf_load_test
Signed-off-by: Connor O'Brien <connoro@google.com>
Change-Id: I7bac83a08c2dc452bdb9030f1e74781116c1dd5c
netd.c is moved from netd to tethering mainline module.
Bump BpfLoader to v0.9
Bug: 202086915
Test: cd system/netd; atest
Test: atest TrafficStatsTest NetworkUsageStatsTest
Change-Id: I0b8ba90bbbe4e16d979c4a40d6fdbbaac0d14d27
Ken Chenauthored and Maciej Żenczykowskicommited 2022-01-19 04:18:53 +00:00
bpf_create_map and bpf_prog_load have been renamed, so update those
calls. bcc now depends on upstream libbpf and has deleted its
redundant implementations of bpf_obj_pin and bpf_obj_get, so add
upstream libbpf as a dependency.
Test: m libbpf_android
Bug: 203823368
Signed-off-by: Connor O'Brien <connoro@google.com>
Change-Id: I324cab44227b0e2f73873f37a28d640ab2373204
BPF headers used by both platform and mainline modules should be in
frameworks/libs/net.
Bug: 202086915
Test: build and boot
Test: cd packages/modules/Connectivity/netd; atest
Change-Id: Id9ac888d5519b2a8663232610d36386cabfe4e94
When fuse-bpf lands in the kernel, support reading its bpf program type
from /sys/fs/fuse/bpf_prog_type_fuse and load such programs.
This is to allow development of fuse-bpf to continue, and is needed
until such time as fuse-bpf is upstreamed and this value is fixed.
Bug: 202785178
Test: Along with change to bpfprogs, prog_fuse_media_fuse_media appears
in /sys/fs/bpf
Change-Id: I83636f25b0a0c678b3ee48c6450d9c8dc157387b
Paul Lawrenceauthored and Maciej Żenczykowskicommited 2021-11-23 00:28:12 +00:00
... it does prefix matching - and first match wins.
By appending / to the strings this no longer matters.
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I54f41fbccd2632334adb9d1ae4fcc44cd1b0141f
Maciej Żenczykowskicommited 2021-11-19 11:18:57 +00:00
This will allow shipping .o's targetting only T and not S devices.
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I651b4442625df9011ff7c5ad283943caae579565
Maciej Żenczykowskicommited 2021-11-18 05:47:24 +00:00
The only purpose of this is to break cut-and-paste of examples
that use SEC() causing people to think things through, look
at Android bpf program examples and use a better macro.
In particular programs with SEC("license") / SEC("maps") need to be
changed to use LICENSE("license") and the map helpers or the map
section is wrong and cannot be correctly parsed by the bpfloader.
Generated via:
git grep 'SEC\(' | cut -d: -f1-2 | while read i; do mcedit $i; done
and manually editting found locations
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Idb333967e054e096fe74f910a5f8aaf1d6c5dc81
Maciej Żenczykowskicommited 2021-11-18 05:45:21 +00:00
This also bumps the bpfloader version to 0.3 to signify support
for loading schedact (ie. tc scheduler action) programs.
Test: load netd.o
Change-Id: Ic3fdc80d3f6ea2b42a4f1d6829b1ff12dd4d52c0
Patrick Rohrauthored and Maciej Żenczykowskicommited 2021-10-13 19:28:53 +00:00
This is also bpfloader v0.2.
Some newer map types (for example DEVMAP) are unusable
on older kernel versions.
Bug: 190519702
Test: atest, TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I085cc723ff1c19d8acc8972a391f894e16dd1875
Maciej Żenczykowskicommited 2021-07-05 18:02:16 -07:00
This is bpfLoader v0.1, previously we had no version number,
the version immediately preceding this commit we'll call v0.0.
Versions older than that are either pre-S and don't matter, since
loading mainline eBpf code was only added in S, or are from early
during the S development cycle (ie. pre-March 5th 2021 or earlier)
and simply no longer supported (no need to maintain compatibility).
Bug: 190519702
Test: atest, TreeHugger - existing bpf programs load
examination of bpfloader logs
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I36fa5b917540be7ea3ecfddc5fe7834e9eb18d88
Maciej Żenczykowskicommited 2021-07-02 03:36:49 +00:00
Basically:
<4.14: DEVMAP -> ARRAY
<5.4: DEVMAP_HASH -> HASH
See added comments for explanation of why, but basically:
This allows our bpf program .o files to load maps on all kernel versions,
even if those map types are not supported by the kernel.
This avoids the need for code that conditionally creates maps based on
kernel version. Any program that actually attempts to use one of these
maps will fail to load, but programs are already loaded conditionally
based on kernel version so this is not a problem.
Test: atest, TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I1a1d73b68de3606423de078fddb224402621e154
Maciej Żenczykowskicommited 2021-03-05 06:00:59 +00:00
This slightly improves error logging.
(also fd == 0 is not an error condition)
Test: atest, TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I135e7405c508951fba632c634bc4e2d2161fc940
Maciej Żenczykowskicommited 2021-03-02 07:46:32 +00:00
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
//system/bpf/libbpf_android:libbpf_android clang-tidy Loader.cpp
system/bpf/libbpf_android/Loader.cpp:133:5: warning: Value stored to 'entries' is never read [clang-analyzer-deadcode.DeadStores]
entries = shTable.size();
^
system/bpf/libbpf_android/Loader.cpp:133:5: note: Value stored to 'entries' is never read
Test: builds without warnings
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I548976b59d44a51e992c9520e730fc9dae82fce2
Maciej Żenczykowskicommited 2021-01-19 21:18:20 +00:00
Tested via:
packages/modules/Connectivity/Tethering/Tethering/bpf_progs/offload.c:
+DEFINE_BPF_PROG_KVER("xdp/test", AID_ROOT, AID_ROOT,
+ xdp_test, KVER(5, 10, 0))
+(struct xdp_md *ctx) {
+ return XDP_PASS;
+}
vsoc_x86:/ # ip link show dev lo
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
vsoc_x86:/ # ip link set dev lo xdp pinned /sys/fs/bpf/prog_offload_xdp_test
vsoc_x86:/ # ip link show dev lo
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 xdpgeneric qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
prog/xdp id 3 tag 3b185187f1855c4c jited
vsoc_x86:/ # ip link set dev lo xdp off
vsoc_x86:/ # ip link show dev lo
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
Test: atest, TreeHugger
Bug: 167540099
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I7472d6d8120344abde9ea0f44c59a46f30da77f1
Maciej Żenczykowskicommited 2021-01-07 14:51:20 -08:00
Logging every 5s isn't actually useful and it just gets the important stuff out
of log buffers.
Test: adb logcat -s 'bpfloader:D' 'BpfUtils:D' 'LibBpfLoader:D'
...
06-17 01:02:34.205 433 433 E bpfloader: --- DO NOT EXPECT SYSTEM TO BOOT SUCCESSFULLY ---
06-17 01:02:38.103 335 335 W LibBpfLoader: Waited 5s for bpf.progs_loaded, still waiting...
06-17 01:02:48.104 335 335 W LibBpfLoader: Waited 10s for bpf.progs_loaded, still waiting...
06-17 01:03:08.104 335 335 W LibBpfLoader: Waited 20s for bpf.progs_loaded, still waiting...
06-17 01:04:12.226 335 335 W LibBpfLoader: Waited 60s for bpf.progs_loaded, still waiting...
06-17 01:04:17.628 1324 1324 W LibBpfLoader: Waited 5s for bpf.progs_loaded, still waiting...
06-17 01:04:27.629 1324 1324 W LibBpfLoader: Waited 10s for bpf.progs_loaded, still waiting...
06-17 01:04:47.629 1324 1324 W LibBpfLoader: Waited 20s for bpf.progs_loaded, still waiting...
06-17 01:05:50.163 1324 1324 W LibBpfLoader: Waited 60s for bpf.progs_loaded, still waiting...
06-17 01:05:55.551 1741 1741 W LibBpfLoader: Waited 5s for bpf.progs_loaded, still waiting...
Bug: 150040815
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I77b82bf44141df8ffab1d11d01a89c3927fbe827
Maciej Żenczykowskicommited 2020-06-17 21:24:55 +00:00
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
... which also means waitForProgsLoaded() has to be a no-op on pre-bpf devices.
This is more consistent: it's weird to have bpf.progs_loaded set on a 4.4 device.
Test: builds, atest
Bug: 150040815
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Iddabcf56c187ea518dd2e4d8db451b2f6902aa8e
Maciej Żenczykowskicommited 2020-06-13 22:41:48 -07:00
Maciej Żenczykowski
authored and
Gerrit Code Review
commited 
Stephane Lee
commited 
Steven Moreland
commited 
Tyler Wear
commited 
Connor O'Brien
commited 
Ken Chen
authored and
Paul Lawrence
authored and
Tyler Wear
commited 
Patrick Rohr
authored and