Use SO_SNDBUF which doesn't require CAP_NET_ADMIN instead.
Change the value of kFuseMaxWrite to 128KB.
In the kernel code, there is a constant FUSE_MAX_PAGES_PER_REQ which
limits the size of requests to 128KB.
Bug: 74725300
Test: atest android.os.storage.cts.StorageManagerTest
Change-Id: Ic3a8f1a7378d027a6c0ee054cedc2c9f4b7509ad
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>
The ScopedLogSeverity was added to observe APCT failures, which turned
out to be a compiler optimization error.
Bug: 62429763
Test: None
Change-Id: Ibb45d018d8eaf4b29cb417da80ae5f0b000dda8e
FuseBuffer::HandleNotImpl save the value of |request.header.unique| to the
temporary variable, clear the buffer which is a union of |request| and
|response|, then write back the unique value to response.header.unique.
Before the CL, the temporary variable was wrongly removed by the compiler
optimization, and response.header.unique was always 0. The CL adds
volatile modifier as workaround to prevent the compiler optimization
from removing the temporary value.
Bug: 62429763
Test: libappfuse_tests
Change-Id: Ia853f805633f646f316f585a35c7b018000b6eb3
(cherry picked from commit a6dee5e279)
FuseBuffer::HandleNotImpl save the value of |request.header.unique| to the
temporary variable, clear the buffer which is a union of |request| and
|response|, then write back the unique value to response.header.unique.
Before the CL, the temporary variable was wrongly removed by the compiler
optimization, and response.header.unique was always 0. The CL adds
volatile modifier as workaround to prevent the compiler optimization
from removing the temporary value.
Bug: 62429763
Test: libappfuse_tests
Change-Id: Ia853f805633f646f316f585a35c7b018000b6eb3
APCT log shows that we got FUSE request unique=0 and replying to such
request causes a EINVAL.
The possible reasons of getting unique=0 here are:
* /dev/fuse actually submits such requests. In this case, not replying
to such request probabbly safe as the kernel cannot wait corresponding
response without a unique number. We can observing the kernel code to
find out what unique=0 actually means.
* Memory corruption happens and unique number are cleared with zero.
In this case, if we skip unique=0 request, libappfuse does not reply
to the kernel request and APCT result will become timeout .
To see which case happens, the CL ScopedLogSeverity to output
verbose logs and lets FuseBridgeLoop skip a request from /dev/fuse if unique=0.
Bug: 62429763
Test: libappfuse_tests
Change-Id: I8c4d532564b690d55573b92260170b0cd68150ab
Previously we have CHECK in WriteInternal function to observe short
writing. However it turns out short write can happen according to the
bug report.
To prevent app from crashing due to CHECK failure, the CL removes the
CHECK and let WriteInternal return a failure value.
Bug: 37561460
Test: libappfuse_tests, manually re-wrote the return value of write()
and checked logcat.
Change-Id: I6a1e233c3ddb8eb68f59e7c606ad0459b5ca2c6e
When /proc/sys/net/core/wmem_max is smaller than kMaxMessageSize, we
need to override the limitation.
Bug: 37561460
Test: libappfuse_tests
Change-Id: Ibaac8db61290d661459fdc46f0ae8416f7db1d9e
This change allows TradeFederation to run the test directly.
Refer to b/35882476 for design and discussion of this change.
Bug: 35882476
Test: local test
tradefed.sh run template/local --template:map test=libappfuse_test
Change-Id: Ic3a23d9d609036658fd37fc72571e2fc7db0e88d
Previously libappfuse set SO_SNDBUF to the maximum message size. However
it does not prevent ENOBUF and it made AppFusePerfTest#testReadWriteFile
flaky.
The CL let FuseBuffer retry write operation when getting ENOBUFS.
Bug: 34903085
Test: libappfuse
Change-Id: I1602474d852e1599f6e69103bcf6f18277a5644b
Previously FuseAppLoopCallback needs to return values in a synchrnous
manner. The CL changes it to asynchronous mannger so that apps can
process FUSE message asynchrnously.
Bug: 35229514
Test: FuseAppLoopTest
Change-Id: I8edcfdb003a25cfd5e9c490ec871140220b21e35
(cherry picked from commit f5d15f9fc4)
The CL turns StartFuseBridgeLoop function into FuseBridgeLoop class, and
adds a method adding new appfuse mount to the loop.
After doing this, one FuseBridgeLoop can process FUSE commands from
multiple AppFuse mounts.
Bug: 34903085
Test: libappfuse_test
Change-Id: I54f11f54bc26c551281b9c32e9bb91f8f043774c
The funciton is going to be used to write FUSE header with external
body.
Bug: 35229514
Test: libappfuse_tests
Change-Id: I303022b555deca960b8e08f26140a5ef10133efe
The class is a thin wrapper for C epoll functions.
Bug: 34903085
Test: Build EpollController.cc and libappfuse_test after applying
future changes locally.
Change-Id: Iedce7f35e4397f80cde1054d53261ad94f9e58a8
These methods return kAgain if operation cannot be done without blocking
the current thread.
The CL also introduecs new helper function SetupMessageSockets so that
FuseMessages are always transfered via sockets that save message
boundaries.
Bug: 34903085
Test: libappfuse_test
Change-Id: I34544372cc1b0c7bc9622e581ae16c018a123fa9
It's an error to pass open O_CREAT without giving it mode bits:
https://linux.die.net/man/3/open
Bug: 32073964
Test: Compiles with clang FORTIFY.
Change-Id: I6b2a3694f85565afdeb782585c6af36e8c4d1557
Previously AppFuse use SOCK_SEQPACKET for sockets communicating system
and app. However SOCK_SEQPACKET requires the buffer of message size in
the kernel and sometimes failed to write with ENOBUF.
The CL updates libappfuse so that it can use SOCK_STREAM instead of
SOCK_SEQPACKET.
Bug: 33279206
Test: libappfuse_test
Change-Id: I622ada9ac1d71d0c57b6cfff0904c7829cea7995
Previously FuseMessage were checking result of read/write operation
after checking header.len value is valid. This was wrong because
header.len does not contain correct value when read function does not
read any bytes and returns zero.
Bug: 33278098
Test: libappfuse_test
Change-Id: Icf998ca6c3eeee20cbc4aa2f65195a87e59ffc27
The CL changes FuseBridgeLoop so that it exits when all files opened on
the AppFuse mount point are closed. Note that the client code will
unmount the FUSE mount point after the loop exits.
Bug: 32260320
Test: libappfuse_test
Change-Id: I4965fbb48de8a280c6306e70757a07376b1956a7
We return the minor version number 15 to FUSE_INIT since we don't handle
BATCH_FORGET. Thus the kernel does not accept the latest size of
fuse_init_out. Instead we need to use FUSE_COMPAT_22_INIT_OUT_SIZE.
Bug: 32779923
Test: libappfuse_test
Change-Id: I5c979d0e45344ca8adfe3ad3f4a9561442abcb3a
The class is used at the app side (StorageManager) to parse FUSE
commands.
Bug: 32260320
Test: libappfuse_test
Change-Id: I1ae2904d3290a041f1efbf8fc10ba032eda5449c
The CL adds FuseBridgeLoop class to libappfuse, which is used in the
system service to proxy fuse commands to applications.
Bug: 29970149
Test: libappfuse_test
Change-Id: I0708f608b3868721ab16ba4028fd2c17a6735af7
The CL adds utility functions to framework to parse FUSE messages
from the kernel. The library will be used from framework JNI and service
JNI.
Bug: 32260320
Test: libappfuse_test
Change-Id: Ib89b26d34789e6c26a3288beceb3ea145c1ae780
