Multiple codepaths were closing the fd they passed into
register_socket_transport on failure, which would close the fd itself.
Switch things over to unique_fd to make it clear that we don't actually
have to close on failure.
Test: mma
Change-Id: I2d9bdcb1142c24931d970f99ebdf9a8051daf05c
register_socket_transport takes ownership of the fd, leading to a
double close if the registration fails.
Test: mma
Change-Id: I9d699af1a979bedc1dc466427436d47f09d0a88f
Replace all usages of android::base::unique_fd with the unique_fd from
adb_unique_fd.h, which calls adb_close instead of close on Windows.
Most of these changes are no-ops, except for the ones to
create_service_thread, which was probably pretty broken on Windows.
Test: python test_device.py
Test: adb_test
Test: wine adb_test.exe
Change-Id: Ia4d03f9e2e7ddc88c97dc64172d721f93f3bf853
Also, fix error reporting when forward fails because we can't get a
transport, because there's either zero, or more than one transport that
matches the request.
Bug: http://b/111021517
Bug: http://b/111374366
Change-Id: Ia4c3bf6215c3ff4e9023ba1af556f5b10463bd6d
Test: manual
This change makes `adb remount` work again in some devices.
Bug: None
Test: `adb remount` in taimen
Change-Id: I04ec043fdb868f18a53a012c82e239c862ae7eff
This change gets rid of most malloc/calloc/free calls. The future is
now!
Bug: None
Test: test_device.py
Change-Id: Iccfe3bd4fe45a0319bd9f23b8cbff4c7070c9f4d
This change removes the void* argument passing and instead uses C++11
features to avoid having to handle memory manually.
Bug: None
Test: python ./system/core/adb/test_device.py
Change-Id: I6380245b2ca583591810e3e363c67c993a107621
This change preserves any additional flags mount flags (e.g. MS_NODEV)
that any of the mounts would have. This avoids unnecessarily opening up
permissions, and also allows kernels that have additional restrictions
about what mount flags can be used to be happy with the remounts.
Bug: 111618714
Test: `adb remount` works in Chrome OS
Test: `adb remount` works in sailfish_aosp
Change-Id: I20d9f2feaf3a47b93bfcdfb4164ee61546ec0b68
Fastboot protocol doesn't include zero packets,
so make it possible to configure that.
Allow fastbootd to see how many bytes the handle
can read/write at once.
Test: adb works
Bug: 78793464
Change-Id: I31e444f384d9d0cdf1ea936439b2028f8151c3b8
Previously, read and write would return 0
on success. Now it will return the number
of bytes read/write. This is more consistent
with other usb handles and is needed in order
to handle partial packets (for fastbootd).
Update usb_write in other usb handles
to return amount written.
Change transport_usb accordingly.
Test: adb works
Bug: 78793464
Change-Id: If07ff05fbc8120343f20661475d34f4e5ff805de
Fastbootd will reuses adb's functionfs transport
implementation.
Move it to daemon/include as well so it can be accessed
with "adbd/usb.h". Otherwise usb.h will conflict with
other imports.
Test: adb builds and works
Bug: 78793464
Change-Id: If3ba190d5c74b5f3633411f0484195e5a2a34d44
Builtin command "type" in zsh doesn't support -t option and this
is resolved by checking the shell type prior to sourcing them.
Instead of using "type -t" explicitly, check_type() knows the the
type of the current shell and behaves accordingly.
Bug: 110629292
Test: (zsh) source build/envsetup.sh && lunch aosp_arm
(bash) source build/envsetup.sh && lunch aosp_arm
Change-Id: Ie39e130e9cc87135cfd522abedba11ad5ac25261
escape_arg reuses the same index for the source (s) and the destination
(result), so it breaks on strings containing more than one quote, e.g:
* a'b'c ==> 'a'\''b'c'
* a'bcde'f ==> 'a'\''b'\'cde'f'
Also make the function more efficient by doing fewer string copies. This
code is based on the android::base::Split code.
Use EXPECT_EQ because the tests can keep going if one fails.
Bug: none
Test: adb_test --gtest_filter=adb_utils.escape_arg
Change-Id: I6ca6e905fa53cc61b9a87276cb7116a5df7e8017
We shipped (well, are about to ship) an adbd that spuriously fails to
create directories upon push. Work around this in the adb client by
running a mkdir on all of the directories we would have otherwise
created.
On devices where we perform the workaround, this coincidentally fixes
a historic bug where we failed to push empty directories.
Bug: http://b/25566053
Bug: http://b/110953234
Test: python test_device.py
Change-Id: I690ec356c206fed4e5ab2c681c5570c8b231e26b
When pushing to a path, we first try to ensure the directory path
exists and has the permissions expected by fs_config. Due to a change
that changed the fs_config check from a blacklist to a whitelist, we
started doing this for /data (which doesn't begin with /data/), and the
UID/GID for that path was accidentally being reused for following path
segments that didn't exist, leading to a failed attempt to chown
/data/local/tmp/foo to be owned by system.
Bug: http://b/110953234
Test: python test_device.py
Change-Id: Ie798eec48bcf54aea40f6d90cc03bb2170280ee8
* changes:
Do not customize __format__ for Windows/MinGW to gnu_printf
Adapt to switch to Clang for Windows host builds
Update cflags for building Windows modules with Clang
Bug: http://b/69933090
Bug: http://b/69933068
MinGW uses gnu_printf to force C99 printf analysis with GCC, but clang
does not support gnu_printf. So just use the default that's used for
other platforms. This also mirrors upstream commit
015e637b4b/.
Test: m native-host-cross with Clang.
Change-Id: I3deb266d70e25296c4ae1d58637afbabc3d949c1
Bug: http://b/69933068
- Add -Wexit-time-destructor globally after fixing one warning
- Enable -Wthread-safety only for non-windows targets. The version of
libstdc++ we use for Windows doesn't have the correct thread-safety
annotations.
Test: Run Clang-built adb_test.exe under wine
Change-Id: I76a43060342e7d1dc370ce3e117f3563c82b7751
* changes:
adb: detect some spin loops and abort.
adb: add an id field to fdevent.
adb: delete unused members in fdevent.
adb: add `adb raw` to connect to an arbitrary service.
Track pending fdevents, and abort if we see an fdevent ready for
read/write for 5 minutes continuously.
Also, add a service that intentionally starts spinning, to test this.
Test: manually changed timeout to a minute, `adb raw spin`
Change-Id: Ibfa5e7e654996587f745887cb04987b982d79bed
Tracking fdevents by pointer is dangerous because an fdevent can be
destroyed and recreated at the same address. Add a monotonically
increasing id field to fdevent for this purpose.
Test: treehugger
Change-Id: I706b57a9e669290ef2db258f85489a5155fc1152
Add a command to connect to an arbitrary service, for debugging
purposes.
Test: `adb raw shell:ls`
Change-Id: I69d4d585e5ecfa7cb8c7a543a2a27df7033b26c7
`adb shell logcat` works, and `adb logcat` should too.
Bug: http://b/110202648
Test: suspended the logd process and manually removed the socket
Change-Id: If14de7c2eb178c7b4210d6bdb44df084615db33a
Implement a nonblocking version of FdConnection. The initial
implementation will be somewhat slower than the blocking one for large
packet sizes, due to an extra copy when coalescing an IOVector into an
apacket, but is still substantially faster for small packets.
Test: adb_benchmark
Change-Id: I4900c9ddf685d3bd557b8cb43958452ecb23db53
adbd has been built as a static executable since the same binary was
copied to the recovery partition where shared library is not supported.
However, since we now support shared library in the recovery partition,
adbd is built as a dynamic executable.
In addition, the dependency from adbd to libdebuggerd_handler is removed
as debuggerd is handled by the dynamic linker.
A few more modules in /system/core are marked as recovery_available:
true as they are transitive dependencies of the dynamic linker.
This change also includes ld.config.recovery.txt which is the linker
config file for the recovery mode. It is installed to /etc/ld.config.txt
and contains linker namespace config for the dynamic binaries under
/sbin.
Bug: 63673171
Test: `adb reboot recovery; adb devices` shows the device ID
Test: Select 'mount /system' in the recovery mode, then `adb shell`.
$ lsof -p `pidof adbd` shows that libm.so, libc.so, etc. are loaded from
the /lib directory.
Change-Id: I363d5a787863f1677ee40afb5d5841321ddaae77
When using "adb remount" on a deduplicated filesystem, the current
response is a warning that the remount will not work. This patch
allows the user to specify an -R option. This option will reboot to recovery,
run e2fsck to undo deduplication, and then reboot the device where "adb
remount" will then succeed.
In addition, if verity needs to be disabled to remount, it will be disabled in
the same reboot cycle to minimize reboots.
Bug: 64109868
Test: adb remount -R on a deduplicated filesystem
Change-Id: I812407499b2df6f4d2509e8d51878117108a6849
adbd (and its dependencies) are marked as recovery_available:true so
that recovery version of the binary is built separately from the one for
system partition. This allows us to stop copying the system version to
the recovery partition and also opens up the way to enable shared
libraries in the recovery partition. Then we can also build adbd as a
dynamic executable.
Bug: 79146551
Test: m -j adbd.recovery
Change-Id: Ib95614c7435f9d0afc02a0c7d5ae1a94e439e32a
An IOVector is a collection of immutable reference counted blocks which
can have its head detached at an arbitrary index. This is extremely
useful for implementing packet-framed protocols like adb on top of a
stream protocol like TCP: a stream reader can read blocks, append them
to the end of the IOVector, and then pull packets off of the front.
This also lends itself naturally towards scatter/gather I/O, which will
enable us to read data from disk and send it across the wire with a
theoretical minimum number of copies in USB, and one extra copy over
TCP.
Since this is basically a generalization of std::deque<Range>, delete
Range and replace its uses with IOVector.
Test: adb_test
Test: wine adb_test.exe
Change-Id: I06561ad0bb25a3a51b378b61d257b5b04b41d9c4
This adds fdsan deallocation sanitization to all fds monitored by
fdevent, which is most of the ones in adb.
Bug: http://b/79786774
Test: python test_device.py
Change-Id: I465804fdb0fd0ac019445900a30ba3403f5bf711
Remove fdevent_install and fdevent_remove in favor of using
fdevent_create and fdevent_destroy, so that we can put RAII types (i.e.
unique_fd) into fdevent without worrying about -Wexit-time-destructors
or structs that are freed instead of deleted.
Bug: http://b/79786774
Test: python test_device.py
Change-Id: I8471cc00574ed492fe1b196944976cdaae8b7cff
This should prevent regressions in the future.
Bug: 78991667
Test: python system/core/adb/test_adb.py
Change-Id: I4d6da40da82c6d79797cec82ffaf071d4b56ddc7
This change:
* uses unittest.main(), which allows for a subset of the tests to be
selected.
* drops the requirement to have a device already connected since all the
tests that need a device now spin their own mock device.
* Splits the monolithic test class into more granular classes.
* Makes this file be pylint-compliant.
Bug: None
Test: python system/core/adb/test_adb.py
Test: pylint system/core/adb/test_adb.py
Change-Id: I91c7ced520c3c69f855d639e0dbf7e57bb690e97
This change adds a reconnect handler that tracks all TCP transports that
were connected at some point, but became disconnected. It does so by
attempting to reconnect every 10s for up to a minute.
Bug: 74411879
Test: system/core/adb/test_adb.py
Test: adb connect chromebook:22 # This runs with sslh
Test: CtsBootStatsTestCases
Test: emulator -show-kernel ; adb -s emulator-5554 shell
Change-Id: I7b9f6d181b71ccf5c26ff96c45d36aaf6409b992
Remove some redundancy by moving io_setup
into aio_block_init. The io_contexts will
now live till the end of the program, similar
to the rest of usb_handle
Test: adb works
Bug: 78793464
Change-Id: I910a39d4395645f1ff231f265c285a72893a0ac6
io_submit sleeps waiting for the gadget
to be enabled. Currently if that sleep
is interrupted it will shut down adb,
causing it to have to start back up again.
Rather than return EINTR if interrupted,
io_submit actually completes and the EINTR
is found later when looking through events.
Since an io that is interrupted will be
small anyway, add a loop to retry small ios.
Also upgrade aio logs in accordance with
their importance.
Fixes: 75981904
Test: adb works, logs show successful interrupt handling
Change-Id: I35973fce130ee849ce59fef80d15b65afb816ba4
(cherry picked from commit 6e9a327504)
A deduplicated filesystem might not have enough free space to undo its
block sharing. In this case "adb remount" should give a more precise
error message.
Bug: 64109868
Test: adb remount on deduplicated filesystem with no free space
Change-Id: I2beb67aa3dfc807cb8493b9c622e7b14174dc9c6
If BOARD_EXT4_SHARE_DUP_BLOCKS is true, "adb remount" will error saying
the filesystem is read-only. Instead, it will now list affected
partitions and explain why they can't be remounted.
Bug: 64109868
Test: adb remount with deduplicated system image
Change-Id: Ida737a6be8e9fe9ea26300897d82b1a149e72f6d
The tags property is (and has always been) unused by Soong. The property has
been defined as a list of strings, and the `androidmk` converted any
LOCAL_MODULE_TAGS entries over to it, but we've never done anything with it.
In preparation for removing the definition from Soong, I'm removing it from all
Android.bp files in the tree.
Since this has never done anything, this is a no-op, but if you really did want
the Android.mk behavior, the proper way to define a module to be installed in
userdebug / eng builds is to use PRODUCT_PACKAGES_DEBUG or PRODUCT_PACKAGES_ENG
in the appropriate product makefile.
Change-Id: Id519b2c0ec352e45c470a1734dfc633bbe39937e
Exempt-From-Owner-Approval: global no-op build change
Test: remove `tags` from Soong, see errors go away.
Add two states: connecting and authorizing, to disambiguate the offline
and unauthorized states, respectively.
Previously, devices would transition as follows:
offline -> unauthorized -> offline -> online
offline -> unauthorized (when actually unauthorized)
With this patch:
connecting -> authorizing -> online
connecting -> authorizing -> unauthorized (when actually unauthorized)
This allows test automation and the like to distinguish between offline
devices, unauthorized devices, and working devices without having to
do retry loops with arbitrary sleeps on their end.
Bug: http://b/79257434
Test: adb_test
Test: adbd_test
Test: manually plugging in a device with `while true; do adb shell echo foo; done`
Change-Id: I036d9b593b51a27a59ac3fc57da966fd52658567
This change exempts the emulator connections from
WaitableConnection.WaitForConnection(). This is because emulator
connections are a) more reliable and b) handled a bit differently than
normal TCP connections.
Bug: 78991667
Test: emulator -showkernel ; adb shell
Change-Id: I552946198889a82d6c265f45e8c3b38f6ac9d045
This change is in preparation to allow the TCP-based transports to be
able to reconnect. This is needed because multiple threads can access
the Connection object. It used to be safe to do because one instance of
atransport would have the same Connection instance throughout its
lifetime, but now it is possible to replace the Connection instance,
which could cause threads that were attempting to Write to an
atransport* to use-after-free the Connection instance.
Bug: 74411879
Test: system/core/adb/test_adb.py
Change-Id: I4f092be11b2095088a9a9de2c0386086814d37ce
This change adds a callback that is invoked exactly once, either when
the connection is fully established (i.e. CNXN packets have been sent
and received) or the atransport object is deleted before that (because
the connection failed).
This helps in distinguishing between successful and failing connections
for TCP. Especially when there is some kind of port
forwarding/multiplexing in between (like an SSH tunnel or SSLH proxy).
Bug: 74411879
Test: adb connect chromebook:22 (which runs an sslh tunnel to adbd).
either succeeds or fails, but not fake-succeeds.
Change-Id: I7e826c6f5d4c30338a03b2d376a857ac5d05672a
Host services are attempted after handle_host_request, which means that
failing to find a transport to give to handle_forward_request shouldn't
send an error over to the other end.
Bug: http://b/78294734
Test: `adb track-devices` with multiple devices connected
Change-Id: I46c89cc1894b51d48fea7d4e629b1d57f73e3fd6
This change uses a context manager to create the fake ADB servers (and
cleanly tear them down.
Bug: 74411879
Test: python system/core/adb/test_adb.py
Change-Id: I722d2c4135259b1b0ef00a1510aa8402e87ecf72
This change returns a different value (-EALREADY) when a connection has
already been established, as opposed to a real connection failure (which
still returns -1).
Bug: 74411879
Test: Opened a socket, tried to adb connect to it,
got "failed to connect to localhost:1337"
Change-Id: Ic216ddef7f28eb43ca750f9e51d068c077d54b07
Plumb the transport that we received the adb reverse request on through
to reverse_service, instead of trying to get a unique transport on
devices that have multiple active transports (e.g. a device with USB
(even unplugged) connected via TCP).
Bug: http://b/37066218
Bug: http://b/71898863
Test: `echo foo | nc -l 12345 & adb reverse tcp:12345 tcp:12345; adb shell nc localhost 12345` on a device connected via TCP
Change-Id: Iae199ae787f2e344126bbcacca8544cfc9844a4c
clang-format some functions that I'm about to touch, but do it in a
separate commit to avoid making the actual changes unreadable.
Test: treehugger
Change-Id: I7dc5258d9ecc1c091d42311149d9e18ae483715b
Switch from using std::string as the type we use to hold our payload in
apacket to a custom reimplementation that doesn't zero initialize. This
improves bulk transfer throughput in the adb_benchmark microbenchmark
on walleye by ~20%.
Test: adb shell taskset f0 /data/benchmarktest64/adb_benchmark/adb_benchmark
Change-Id: Ibad797701eb1460c9321b0400c5b167b89b2b4d0
OS X reports maxPacketSize as wMaxPacketSize * (bMaxBurst + 1) in the
deprecated GetPipeProperties function. Use the also-deprecated
GetPipePropetiesV2 API to get bMaxBurst and figure out what
wMaxPacketSize is. (This file is going to go away eventually, so don't
bother with switching to the recommended GetPipePropertiesV3, since
it would be a substantially larger charge.)
libusb is unaffected.
Bug: http://b/77733422
Test: python test_device.py
Change-Id: I66517d699a4f39b93ba5eb7882bd8ee6c70f3672
This change splits the capability-dropping step of adbd into two. This
is more robust when ambient capabilities are being used, since minijail
cannot currently handle that case.
Bug: 77146512
Test: grep Cap /proc/`pidof adbd`/status
CapInh: 0000000000000000
CapPrm: 0000000000000000
CapEff: 0000000000000000
CapBnd: 00000000000000c0
CapAmb: 0000000000000000
Change-Id: I0476a8d80f7a2497600196932542045f3bc87537
Profiling of adb_benchmark revealed that something like half of the CPU
time was being spent in malloc and free, which was odd because the
benchmark repeatedly mallocs and frees allocations of the same size. It
turns out that our default configuration of jemalloc will purge after
every free. Configure jemalloc to not do this, for gains of over 100%
on adb_benchmark, and up to 25% (on walleye USB3) in real-life.
Test: adb_benchmark
Change-Id: I602dd1645c3d21709c7f6a78903511ce4d576558
This reverts commit 2ef56a1174.
host_supported is broken in cc_benchmark on Mac, but we probably don't
actually too much about the host benchmarks, so just delete the line.
Bug: http://b/77585931
Test: treehugger
Change-Id: I8898b4037a3faf8e8dee2a99498a6f84e642b92b
When connecting to an address, we construct a transport first, and then
check whether we've already connected to that address. The consequent
destruction of the BlockingConnectionAdapter attempts to join threads
that haven't been started, which aborts.
Make it safe to destruct a BlockingConnectionAdapter without calling
Start on it first, to solve this.
Bug: http://b/69137547
Test: nc -l 12345 & (adb connect localhost:12345; adb connect localhost:12345)
Test: python test_adb.py
Change-Id: I6cb968a62dbac6332907e06575893d764905ee62
On platforms that implement sockets via underlying sockets (e.g. Wine),
a socket can appear to be full, and then become available for writes
again without read being called on the other end. Add a sleep after
each write to give the underlying implementation time to flush.
This doesn't help us if the buffer size is smaller than MAX_PAYLOAD,
but at least in the case of Wine, that doesn't seem to be the case.
Test: adb_test
Test: wine adb_test.exe
Change-Id: Iff1ec14bbf318b9742ec7e2fb72e34e3d6bbe6ad
Previously, each of the tests was spawning the fdevent thread manually,
in order to be able to set up listeners and such before running
fdevent_loop. Now that we have a way to run arbitrary code on the
fdevent thread, switch to having a generic fdevent thread and running
setup code via fdevent_run_on_main_thread.
Test: adb_test
Test: wine adb_test.exe
Change-Id: I517dbcbad31067b45087d9fbed67a75b75a75aec
Make it so that we handle run_on_main_thread calls after regular socket
events, so that we can use it as a way to ensure we've processed all
pending socket events.
Test: adb_test
Test: wine adb_test.exe
Change-Id: Ic215c7fed19a8e1699e759970658b3775aa08c45
Bail out immediately in our poll emulation if we have any invalid file
descriptors, instead of actually calling WSAPoll (which might block
forever).
Test: wine adb_test.exe
Change-Id: I06ccca305783ee7a1721b1585ddf73e022bd02d7
Make it easier to replace the sleep with something that actually waits
for the event loop to quiesce.
Test: adb_test
Change-Id: Iee86b058abbb607256dbfce43d8c70252b6bb0f6
Previously, processes started via `adb shell` have an oom_score_adj of
-1000, making them invisible to the oom killer. This makes running a
process that consumes all memory (e.g. by leaking in a loop) lead to
the entire rest of the system (including adbd, because of bad
heuristics in the kernel) being oom killed before getting to it.
Bug: http://b/63143027
Test: `adb shell cat /proc/self/oom_score_adj` with adb root
Change-Id: I59111134e36dc271adf4c1dd4bd4400d4fe6aee0
The switch over to soong unintentionally turned off building the
Windows tests (and there's a bug preventing them from being turned on).
Preemptively fix the build breakages that'll happen when we turn them
back on.
Test: mma with Android.bp 's/cc_test_host/cc_binary_host' hacks
Change-Id: I6582cfc61b96052537d50d8ba90533dbb66e3e1d
Match openssh's behavior with regards to login shells.
Bug: http://b/76116378
Test: `adb shell -T set -o` vs `ssh localhost -T set -o`
Test: `adb shell -tt set -o` vs `ssh localhost -tt set -o`
Test: `echo set -o | adb shell -T` vs `echo set -o | ssh localhost -T`
Test: `echo set -o | adb shell -tt` vs `echo set -o | ssh localhost -tt`
Change-Id: I6ed69b43e2ac79c947c242948b5617ece08a713e
The listen backlog seems to be more meaningful on Darwin than on Linux,
resulting in connections failing with ECONNRESET. Bump it up to the
maximum supported value to make this less likely. 128 pending
connections ought to be enough for anybody.
Bug: http://b/74616284
Test: python test_device.py
Change-Id: I5fe0205924188cf18ca1fc1204f923ab5523eeb2
This test is failing due to:
UnicodeWarning: Unicode equal comparison failed to convert both
arguments to Unicode - interpreting them as being unequal'
Both arguments are already unicode, so stop encoding one side.
Test: python -m unittest test_device.FileOperationsTest.test_unicode_paths
Change-Id: Iea8bf29845ab3008ccf4c7dbd6969196e57ea25d
There exists no path through which a value other than -1 can be written
to the SHELL_EXIT_NOTIFY_FD.
Test: adb_test
Test: adbd_test
Test: python test_device.py
Change-Id: I0630c302ba06bc76917f0445aea75d2dbe1dc865
When we fail to write to a local socket peer, we might still have data
queued up to send to the other side. Defer closing the socket until
we've failed to both read and write.
Bug: http://b/74616284
Test: python test_device.py
Change-Id: Ifc4b8fe95369b4872e475c2ae4ee611dd2d8b9d7
If we get a ton of fdevent_run_on_main_thread calls while running one
of the handlers, the socket might become full, which will result in a
deadlock in fdevent_run_on_main_thread when a write to the fd blocks
with the mutex taken. Resolve this by making the fd nonblocking, which
is safe because we always write after appending to the list, and read
before emptying the list, which guarantees that if the byte we write is
consumed, the std::function we appended will be run.
Bug: http://b/74616284
Test: adb_test
Test: python test_device.py
Change-Id: I29319bda2ad7b5a5cdcd91d1d0ddf39f7ab7d115
Add a test to hammer on `adb shell exit $n` for flakiness.
Bug: http://b/74616284
Test: python test_device.py
Change-Id: I6a842960f5b55ff739044698f5c9683992fc42f1
Make it so that the socket tests don't leak until your machine blows up
by switching an infinite loop into an assertion failure.
Bug: http://b/74616284
Test: adb_test
Change-Id: If618c26b224b660548454f542cab79bebe46f80e
Add an implementation of std::make_unique for Windows, where we're
currently stuck with C++11, and switch some uses of new over to it.
Test: treehugger
Change-Id: I99b85f07754adda7c525243480c3e0bce9a25ce7
All devices that previously used ENDPOINT_ALLOC
are on the new async io routines. None of the devices
using aio_compat have ENDPOINT_ALLOC so remove the
code to stop logging failures.
Fixes: 74213465
Test: adb works
Change-Id: I0c903eb76b006b6bcce48cec24f5353fa47cc128
(cherry picked from commit 9a1be3ad0e)
Add a currently failing test that's a reduced case of the 'concurrent
adb shell' flakiness reported in b/74616284.
Bug: http://b/74616284
Test: changed #if 0 to #if 1, test fails
Change-Id: Id170a185f7c5d2ff44f772f7849a73227f9df441
adb disable-verity will toggle a flag in /vbmeta and introduce AVB
verification error on next boot. If the device is LOCKED, it will make
the device unbootable because verification error isn't allowed when the
device is locked.
Also indicating 'adb root' when failed to get verity state.
Bug: 70969453
Test: adb disable-verity should pop-up warning if the device is locked.
Change-Id: I1ed705e34334ea2231c96b16ddb8d225067af2f0
This change opens the control, bulk-in, and bulk-out FunctionFS
endpoints with O_RDONLY or O_WRONLY.
Bug: 70349025
Test: Android on Chrome OS can still use ADB
Test: aosp_sailfish-userdebug can still use ADB
Change-Id: I3895b626cf7cec8437c36bdc6c6e6055aaf3777c
By default, the build system only sets ro.adb.secure (to 1) when
building user builds.
Bug: http://b/73993211
Bug: http://b/74059119
Test: treehugger
Change-Id: I2a683257c1027394ded90c8d2634def6003e0525
adbd is copied into the recovery image without the accompanying
linker/libraries, so it must be a static executable.
Bug: http://b/74237641
Test: `file $ANDROID_PRODUCT_OUT/system/bin/adbd`
Change-Id: Ie8d0710f6002458519d3eea3d696e9016faafee8
* changes:
adb: remove meaningless const.
adb: convert Connection to a nonblocking interface.
adb: use soong version stamping.
adb: switch over to Android.bp.
Rename the existing Connection to BlockingConnection, add a nonblocking
Connection, and add an adapter between the two, to enable future work
to reduce the impedance mismatch from implementing a blocking interface
on top of nonblocking primitives.
While we're here, delete A_SYNC, and remove one layer of pipes when
sending a packet (replacing it with a condition variable when using
BlockingConnectionAdapter).
Test: python test_device.py, manually plugging/unplugging devices
Change-Id: Ieac2bf937471d9d494075575f07e53b589aba20a
Added by I7255d45335fa009dc9e5de99dff67af52bd70e06
Bug: None
Test: Ran the analyzer. Complaint is gone.
Change-Id: Id9b8debd8b9690210f5a8e19154ee3a5095b8fbd
Rearrange some files while we're doing this.
Bug: http://b/71721338
Test: manually ran adb on windows
Change-Id: Ie47bda82279e4b9521505ad0353bf9ef649fc7d7
Add a way to test LeakSanitizer with the server by adding an
environment variable that intentionally leaks.
Test: ASAN_OPTIONS=detect_leaks=1:leak_check_at_exit=1 ADB_LEAK=1 adb server nodaemon
Change-Id: I7255d45335fa009dc9e5de99dff67af52bd70e06
We don't actually need to use quick_exit to avoid calling static
destructors, since we have -Wexit-time-destructors to guarantee we
don't actually have any, and this precludes the use of asan's exit time
leak checking, so switch back to atexit/exit.
Test: ASAN_OPTIONS=detect_leaks=1:leak_check_at_exit=1 adb server nodaemon with a manually inserted leak
Change-Id: Id8178913f64cb02c820c5073351369a9e4d8c74d
Passing in MS_REMOUNT | MS_BIND tells the kernel to only remove the
MS_RDONLY flag and keep all the other flags the same. This is also the
only kind of remount that is allowed in Linux containers (to be more
precise, within user namespaces).
This change also attempts to always run the remount command when dealing
with /, since containers will almost always run a loop device, and since
the number of the device changes, it is not convenient to put it in the
fstab. Plus, the container won't have permission to modify it, but might
be able to perform the remount.
Bug: 72178046
Test: `adb remount` works in both sailfish and Chrome OS
Change-Id: I9e8ec8afcd57f67875a312824667768b3aa89faa
If create_service_thread fails, we'll leak `arg`.
This fixes a static analyzer complaint:
system/core/adb/services.cpp:298:13: warning: Potential leak of memory
pointed to by 'arg'
Bug: None
Test: Reran the static analyzer. No more complaints about this leak.
Change-Id: I5aec7fd78f2cc775b650501b02bdf0039d1647ca
asocket has a destructor now, so we have to delete it, or leak the
data pointed to by its std::string.
Bug: http://b/73257049
Test: manual testing with asan
Change-Id: Ia88199292cc74e10032a9a16226d3afc61c3e0be
Don't go through a pointless conversion to hex that reduces our maximum
pid to 64k.
Bug: http://b/17661822
Test: `adb jdwp`
Change-Id: Idd572eac827f3ed3825ac2cf55fc4db109e70854
When persist.adb.trace_mask is used to enable verbose logging, avoid
exponential logging when a user runs `adb logcat`. As a side-effect,
set the minimum logging level to include everything when ADB_TRACE is
used on the client (which is something we want anyway).
Bug: http://b/72971016
Test: `adb shell 'setprop persist.adb.trace_mask; killall adbd'; adb logcat`
Change-Id: Id4bca1f3933a920345499dbeaeb103c8a2e220cd
Switch asocket over to taking a std::string instead of apacket* for
data. This allows us to remove asocket specific fields from apacket*.
Test: python test_device.py with x86_64 emulator, walleye
Test: adb_test on host
Change-Id: I9d157ff331a75ba49a54fdd4194e3f6cdff722f4
These checks were moved to after the read of the payload, which is too
late. Add a check before each read to avoid a heap buffer overflow.
Test: python test_device.py with x86_64 emulator, walleye
Change-Id: I86bcfaaa9004951cc52ad89af74680cf748e717d
As step one of refactoring atransport to separate out protocol handling
from its underlying connection, extract atransport's existing
hand-rolled connection vtable out to its own abstract interface.
This should not change behavior except in one case: emulators are
now treated as TCP devices for the purposes of `adb disconnect`.
Test: python test_device.py, with walleye over USB + TCP
Test: manually connecting and disconnecting devices/emulators
Change-Id: I877b8027e567cc6a7461749432b49f6cb2c2f0d7
We should have done this from the beginning. Thanks to Windows, we're not
going to be able to switch libbase over to std::string_view any time soon.
Bug: N/A
Test: ran tests
Change-Id: Iff2f56986e39de53f3ac484415378af17dacf26b
Makes copy of the transport list and sorts it by type and serial before
printing.
Bug: 70748433
Test: adb devices displays all connected devices
Change-Id: I917728a102972f2f38f2e370a0c6011c1eb883c7
The checksum is unnecessary. Improves adb performance by 40% on USB2.
Test: new adb works with new + old adbd, old adb works with new adbd
bug 67327728
Change-Id: I761d8a5a62deaea9bbb092ea9926b2d6d312f00d
This still doesn't work yet because libusb doesn't support hotplug on
Windows yet, but it's in the milestones for the next libusb release,
and this is enough to start poking around with the WinUSB device-side
stuff, so leave it broken for now.
Bug: http://b/68993980
Test: set ADB_LIBUSB=1; adb.exe server nodaemon aborts in the expected place
Change-Id: Icef7d46e31c847d6a8e724c6f58ae5db43673c16
OS X appears to be misreporting the maximum packet size on some
hardware. Disable this on OS X until we figure out what's going on.
Bug: http://b/65063965
Bug: https://issuetracker.google.com/70244520
Test: python test_device.py on linux with CHECK_PACKET_OVERFLOW disabled
Test: python test_device.py on darwin (but on hardware that wasn't reproducing this)
Change-Id: I57e1adfa162e40ed79f71f97af552b3f0519324e
In https://github.com/torvalds/linux/commit/8db6c34f1dbc8 , namespaced
file capabilities were introduced. That change updated VFS_CAP_REVISION
from VFS_CAP_REVISION_2 to VFS_CAP_REVISION_3.
Android code is written assuming v2 capabilities, and the code will
break if we naively try to treat a v2 structure as a v3 structure.
So don't even try.
Android kernels prior to v4.14 will not support this extended capability
structure, so attempting to set such capabilities will ultimately fail.
With 8db6c34f1dbc8, it appears that attempting to read a v3 capabilities
xattr will always downgrade the capability to a v2 capability, so it
really doesn't make sense to look for a v3 capability.
Android capabilities are only created at /system and /vendor filesystem
creation time by host tools. Android processes, within or outside a
namespace, are not permitted CAP_SETFCAP
(https://android-review.googlesource.com/c/platform/system/sepolicy/+/547801/1/public/domain.te
line 1101). So we should never have to deal with a v3 capability other
than those that might appear on the /system / /vendor partition at a
future date by a future author.
Bug: 69617725
Test: build/test/boot/CTS passes
Change-Id: I0378b3f1195dc62dbeb771944ab378c881441118
This test actually belongs with the adb python wrapper, since it's not
affected by changes to adb itself.
Bug: http://b/64260633
Test: python test_device.py
Change-Id: I58e5dff760f35923f8dbbdc7de6ffd75254880dc
We want repeatable builds for minimal OTAs.
(cherrypick of 2e93eda69cf20d9c04d56ea6bf1e2ea62ad62eb4.)
Bug: http://b/64634017
Bug: 64709603 (presubmit balking at the line above)
Test: repeated builds after `touch daemon/main.cpp` keep same md5sum
Change-Id: Iedeb44377ea45bbf20e2212760802198a57dc14a
This probably doesn't have any actual effect, because there should only
be one smartsocket listener so this will happen to work, but this looks
sketchy.
Test: mma
Test: adb server nodaemon & (sleep 1; adb kill-server)
Change-Id: Id00ab1fc1fc029949c675ae6e3b1a77f3e9e4c69
Future adjustments triggered cleanup and transition of adb/adb.h and
adb/client/usb_windows.cpp to be moved to Android coding standard.
Test: build
Bug: 63736262
Bug: 38446744
Bug: 66912053
Change-Id: I6eb3f0665b9670b9b3d5f5397f271605b48f4ff0
bash sets this itself, but mksh doesn't.
This also makes interactive and non-interactive adb shells more similar,
because when we set this in mkshrc that wouldn't affect non-interactive
shells. (I'm not sure whether this is an improvement or not.)
This is certainly _cheaper_ than doing this in mkshrc by calling out to
getprop though.
Bug: N/A
Test: `adb shell printenv` and `adb shell`, `printenv`
Change-Id: I3ff724f19a5098313df83836253f1f7e7872d6a4
This logic appears to be racy, and it shouldn't actually be needed, if
our devices follow the USB spec. Use libusb_set_interface_alt_setting
on device initialization as well, to add one more thing that should
reset the data toggles.
Bug: http://b/32952319
Test: python test_device.py
Change-Id: I392198af3d72c524b893e5056afa2b4617cea49c
Name each service thread specifically to improve debuggability.
Bug: http://b/65648324
Test: adb shell debuggerd -b `adb shell pidof adbd` during a sync
Change-Id: I644e25fc2f14a26eafd19f596e4e8f3c2bb79967
Revert the write_msg_lock part of commit b5e11415. A write which hangs
will hold onto the mutex, preventing the device kick from ever
happening, which also causes lots of other stuff to hang, due to Kick
being called with the transport lock taken.
Test: python test_devices.py
Change-Id: Ie7c958799c93cad287c32d6bbef30c07f40c2d51
This reverts commit 7e197ef833.
The mutex lock in transport_unref hides a race that seems otherwise
hard to fix. Specifically, there's no synchronization between acquiring
a transport and attaching it to an asocket*, leading to badness if the
transport is closed in between the two operations.
Fix the original problem the reverted patch addressed by manually
unlocking before calling unregister_usb_transport.
Bug: http://b/65419665
Test: python test_device.py
Change-Id: I0ed0044129b1671b2c5dd1b9fa2e70a9b4475dc5
The comment that was previously here says that local_socket_list_lock
must be taken, but this function is exposed to external callers that
can't possibly take the lock.
Bug: http://b/65419665
Test: python test_device.py
Change-Id: I12d464933936b2a210a827ccf19ea201020d8d78
For reboot [reboot_arg] requests via either reboot or adb reboot,
if reboot_arg is empty then report "shell" or "adb" respectively.
Test: boot_reason_test.sh shell_reboot adb_reboot
Bug: 63736262
Change-Id: Ie613d9e62db6a705885e4e7520aede27af3aa1b9
Use CHECK_LT(foo, 1234) instead of CHECK(foo < 1234) so that failing
checks have the values emitted.
Bug: http://b/65063965
Test: adb server nodaemon
Change-Id: I094287f7c4d2d177ea216568d06e9e425a28dd96
The test was assuming we still output the full help for every syntax error.
While I'm here, make the diagnostics suck less.
Bug: N/A
Test: ran tests
Change-Id: Idc28616f20c66391f32046cf4216f122998a84bd
Extend device selection to allow selecting a specific transport via
monotonically increasing identifier (visible in devices -l).
This is useful when using multiple devices (like hikey960...) that
have identical bogus serial numbers like 0123456789ABCDEF.
Bug: http://b/37043226
Test: adb -t {1, 2, 9999999} {get-serialno, shell, features}
Change-Id: I55e5dc5a406a4eeee0012e39b52e8cd232e608a6
Previously, kick_all_transports would deadlock if there were any
inaccessible transports, because the transport kick function would call
unregister_usb_transport, which attempts to take the already-held
transport lock.
Fix this by switching the transport lock over to a recursive mutex.
Test: manual
Change-Id: If61296ff4745e1699f3e216811c1383582627604
...otherwise it could have weird dates when the device clock is wrong.
Test: m -j32 adb && ./out/host/linux-x86/bin/adb bugreport test.zip && ls -l test.zip
Test: m -j32 adb_test && ./out/host/linux-x86/nativetest64/adb_test/adb_test
Fixes: 63927360
Change-Id: I3f87277f1acb97212a87b89639e42b6d7d300faa
This method works around the downsides of
ENDPOINT_ALLOC, namely that it is not affected
by memory fragmentation and it uses an upstream
interface.
Also add libasyncio to provide the necessary syscalls
to both adb and mtp.
Add some small optimizations to file_sync.
Bug: 37916658
Test: run adb push/pull
Change-Id: If3b3be02b5e2d4f9cffec1b8ddc02a5768a51a1f
clang is the default compiler since Android nougat
Test: mma & verified it´s still build with clang
Change-Id: I34adaeef2f6558a09f26027271222bad94780507
Signed-off-by: Lennart Wieboldt <lennart.1997@gmx.de>
ALLOW_ADBD_ROOT ('adb root') and ALLOW_ADBD_NO_AUTH (ro.adb.secure = 0)
are false in user build. This prevents a non-A/B device from running
Treble VTS because it requires 'adb root'. Without ALLOW_ADBD_NO_AUTH,
adb still can work if ro.adb.secure = 1. However, allowing it to be 0
is better for test automation.
The image combination in VTS is:
- system.img (userdebug): provided by Googlg
- boot.img (user): provided by the OEM <-- adbd is here
- vendor.img (user): provided by the OEM
This change allows 'adb root' and 'ro.adb.secure = 0' when the device is
unlocked in user build. No changes for userdebug/eng builds.
Note that the device must be unlocked when running VTS. Otherwise,
verified boot will prevent it from booting the system.img provided by
Google (no verity metadata).
Bug: 63313955
Bug: 63381692
Test: use the above image combination, check 'adb root' and
'ro.adb.secure = 0' can work
Change-Id: I109d96c950e54c4fb0ac0c98b989a20593681e52
This reverts commit f2f0b31850.
libusb seems to be causing flakiness on some machines. Disable it for
now.
Bug: http://b/62962248
Test: python test_device.py (with DeviceOfflineTest commented out)
Change-Id: Ia9de78ab772c22574cf5ca7facb78f22af6d7a71
Relax the shell command length limits when talking to an adbd with the
shell protocol.
shell is pretty much the only service that takes an arbitrarily long
string, so this is somewhat safe.
Bug: http://b/37716055
Test: `adb shell $(python -c 'print "echo " + "f" * (32*1024)') | wc` on L and master
Change-Id: I0737fd2244530ef8080f300cd3a3549a1ab93465
adbd currently decides to use the QEMUD pipe if ro.kernel.qemu=1, as set
for ranchu. The Android container in Chrome OS doesn't have that
property set and it can't be set to 1 because it's used as equivalent to
"runs inside an emulator" throughout Android and changes the way
graphics are handled, whether Bluetooth is supported, etc., behaviour
that we do not want to trigger in Chrome OS.
adbd now also checks service.adb.transport to decide whether to use the
QEMUD (goldfish) pipe.
adbd still first checks for ro.kernel.qemu to preserve existing
behaviour and will still fallback to TCP if it can't use Goldfish.
Bug: 38497992
Test: tested by jmgao@ -thanks!- on aosp_angler, adb still works.
Change-Id: I8370704145ae7301ac7aeef81c5cbd94cfcb7fd7
Fix a deadlock that happened when a reader/writer thread released a
transport while the hotplug thread attempted to handle a device
disconnection. Decrementing a transport refcount to zero would hold the
global transport mutex and attempt to take the usb handles mutex, while
the hotplug thread would hold the usb handles mutex and try to call
unregister_usb_transport, which would attempt to take the global
transport mutex.
Resolve this by making transport_unref not take the global transport
mutex.
Bug: http://b/62423753
Test: python test_device.py
Change-Id: Ib48b80a2091a254527f3a7d945b6a11fae61f937
For unclear reasons, it seems that for some people, devices are
becoming accessible, inaccessible, and then inaccessible again.
Switch to a sleep for now.
Bug: http://b/62200735
Test: manual
Change-Id: Id51dcb188c3534a57c35cbfd30a181c99115a23c
If a device is unplugged while it's being processed, we can be inside
register_usb_transport still when we call unregister_usb_transport,
leading to bad things happening. Move the call of
register_usb_transport into the scope guarded by usb_handles_mutex.
Bug: http://b/62200735
Test: manual
Change-Id: I2517dcb9d30d9a08cebbb5f715ae3fd30c510109
Previously, we would attempt to delete a usb_handle in both
device_disconnected and usb_close. If the one in device_disconnected
happened to happen first, usb_close would abort when it failed to find
the handle it was supposed to own.
Bug: http://b/62020217
Test: unplugging device on darwin
Change-Id: I6c6bf61bf89a4d9a23458c00b457080d3d6cc744
Version 2 of the framebuffer protocol in adb can now be used to read
the color space information of the screenshot: unknown (native),
sRGB or Display P3.
Bug: 29940137
Test: Manual
Change-Id: I037da58e1c4f6e9e874d72348140c067640c6931
Move the invocation of adb_notify_device_scan_complete to
the end of device_connected, where we decrement connecting_devices.
Also, create a dedicated thread for handling hotplug events, instead of
reusing the main thread for this, since the main thread blocks until
device scan is complete.
Test: `adb kill-server; adb devices`
Change-Id: Ia73b1a57538174282a48ef73ab0a3e58152d6f83
