We've seen USB writes failing due to inability to allocate contiguous
chunks of memory in the kernel on devices, but it looks like the same
problem can occur on the host, as well. It's a mild performance
regression (90->80 MB/s on a blueline) to split the writes always, so
attempt the full write first, and fall back to splitting it up if that
fails with ENOMEM. Once we switch over the the asynchronous transport
API, we'll be able to submit multiple writes cheaply, like on devices,
so we won't need to retry at that point.
Bug: http://b/140985544
Test: test_device.py
Change-Id: I1517c348375b829dfff6796c4e9d394802b02d5b
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
Previously, adb was assuming a fixed maximum packet size of 1024 bytes
(the value for an endpoint connected via USB 3.0). When connected to an
endpoint that has an actual maximum packet size of 512 bytes (i.e.
every single device over USB 2.0), the following could occur:
device sends amessage with 512 byte payload
client reads amessage
client tries to read payload with a length of 1024
In this scenario, the kernel will block, waiting for an additional
packet which won't arrive until something else gets sent across the
wire, which will result in the previous read failing, and the new
packet being dropped.
Bug: http://b/37783561
Test: python test_device.py on linux/darwin, with native/libusb
Change-Id: I556f5344945e22dd1533b076f662a97eea24628e
When device goes offline, user usually has to manually replug the
usb device. This patch tries to solve two offline situations, all
because when adb on host is killed, the adbd on device is not notified.
1. When adb server is killed while pushing a large file to device,
the device is still reading the unfinished large message. So the
device thinks of the CNXN message as part of the previous unfinished
message, so it doesn't reply and the device is in offline state.
The solution is to add a write_msg_lock in atransport struct. And it
kicks the transport only after sending a whole message. By kicking
all transports before exit, we ensure that we don't write part of
a message to any device. So next time we start adb server, the device
should be waiting for a new message.
2. When adb server is killed while pulling a large file from device,
the device is still trying to send the unfinished large message. So
adb on host usually reads data with EOVERFLOW error. This is because
adb on host is reading less than one packet sent from device.
The solution is to use buffered read on host. The max packet size
of bulk transactions in USB 3.0 is 1024 bytes. By preparing an at least
1024 bytes buffer when reading, EOVERFLOW no longer occurs. And teach
adb host to ignore wrong messages.
To be safe, this patch doesn't change any logic on device.
Bug: http://b/32952319
Test: run python -m unittest -q test_device.DeviceOfflineTest
Test: on linux/mac/windows with bullhead, ryu.
Change-Id: Ib149d30028a62a6f03857b8a95ab5a1d6e9b9c4e
We have std::thread now, so we can delete this cruft.
Test: python test_device.py
Test: adb_test
Test: wine adb_test.exe
Test: /data/nativetest/adbd_test/adbd_test
Change-Id: Ie1c1792547b20dec45e2a62ce6515fcb981c3ef8
Add a libusb-based implementation alongside the existing native
implementations, controlled by the ADB_LIBUSB environment variable.
Windows will need more work for the usb driver.
Bug: http://b/31321337
Test: python test_device.py on linux/darwin, with ADB_LIBUSB=0 and 1
Change-Id: Ib68fb2c6c05475eae3ff4cc19f55802a6f489bb7
Josh Gao
Ivan Afonichev
Jerry Zhang
Yi Kong
Mark Salyzyn
Yabin Cui