platform_system_core/adb/usb_windows.cpp

669 lines
19 KiB
C++
Raw Normal View History

/*
* Copyright (C) 2007 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#define TRACE_TAG TRACE_USB
#include "sysdeps.h"
#include <winsock2.h> // winsock.h *must* be included before windows.h.
#include <adb_api.h>
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <usb100.h>
#include <windows.h>
#include <winerror.h>
#include "adb.h"
#include "transport.h"
/** Structure usb_handle describes our connection to the usb device via
AdbWinApi.dll. This structure is returned from usb_open() routine and
is expected in each subsequent call that is accessing the device.
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
Most members are protected by usb_lock, except for adb_{read,write}_pipe which
rely on AdbWinApi.dll's handle validation and AdbCloseHandle(endpoint)'s
ability to break a thread out of pipe IO.
*/
struct usb_handle {
/// Previous entry in the list of opened usb handles
usb_handle *prev;
/// Next entry in the list of opened usb handles
usb_handle *next;
/// Handle to USB interface
ADBAPIHANDLE adb_interface;
/// Handle to USB read pipe (endpoint)
ADBAPIHANDLE adb_read_pipe;
/// Handle to USB write pipe (endpoint)
ADBAPIHANDLE adb_write_pipe;
/// Interface name
char* interface_name;
/// Mask for determining when to use zero length packets
unsigned zero_mask;
};
/// Class ID assigned to the device by androidusb.sys
static const GUID usb_class_id = ANDROID_USB_CLASS_ID;
/// List of opened usb handles
static usb_handle handle_list = {
.prev = &handle_list,
.next = &handle_list,
};
/// Locker for the list of opened usb handles
ADB_MUTEX_DEFINE( usb_lock );
/// Checks if there is opened usb handle in handle_list for this device.
int known_device(const char* dev_name);
/// Checks if there is opened usb handle in handle_list for this device.
/// usb_lock mutex must be held before calling this routine.
int known_device_locked(const char* dev_name);
/// Registers opened usb handle (adds it to handle_list).
int register_new_device(usb_handle* handle);
/// Checks if interface (device) matches certain criteria
int recognized_device(usb_handle* handle);
/// Enumerates present and available interfaces (devices), opens new ones and
/// registers usb transport for them.
void find_devices();
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
/// Kicks all USB devices
static void kick_devices();
/// Entry point for thread that polls (every second) for new usb interfaces.
/// This routine calls find_devices in infinite loop.
void* device_poll_thread(void* unused);
/// Initializes this module
void usb_init();
/// Opens usb interface (device) by interface (device) name.
usb_handle* do_usb_open(const wchar_t* interface_name);
/// Writes data to the opened usb handle
int usb_write(usb_handle* handle, const void* data, int len);
/// Reads data using the opened usb handle
int usb_read(usb_handle *handle, void* data, int len);
/// Cleans up opened usb handle
void usb_cleanup_handle(usb_handle* handle);
/// Cleans up (but don't close) opened usb handle
void usb_kick(usb_handle* handle);
/// Closes opened usb handle
int usb_close(usb_handle* handle);
int known_device_locked(const char* dev_name) {
usb_handle* usb;
if (NULL != dev_name) {
// Iterate through the list looking for the name match.
for(usb = handle_list.next; usb != &handle_list; usb = usb->next) {
// In Windows names are not case sensetive!
if((NULL != usb->interface_name) &&
(0 == stricmp(usb->interface_name, dev_name))) {
return 1;
}
}
}
return 0;
}
int known_device(const char* dev_name) {
int ret = 0;
if (NULL != dev_name) {
adb_mutex_lock(&usb_lock);
ret = known_device_locked(dev_name);
adb_mutex_unlock(&usb_lock);
}
return ret;
}
int register_new_device(usb_handle* handle) {
if (NULL == handle)
return 0;
adb_mutex_lock(&usb_lock);
// Check if device is already in the list
if (known_device_locked(handle->interface_name)) {
adb_mutex_unlock(&usb_lock);
return 0;
}
// Not in the list. Add this handle to the list.
handle->next = &handle_list;
handle->prev = handle_list.prev;
handle->prev->next = handle;
handle->next->prev = handle;
adb_mutex_unlock(&usb_lock);
return 1;
}
void* device_poll_thread(void* unused) {
adb_thread_setname("Device Poll");
D("Created device thread\n");
while(1) {
find_devices();
adb_sleep_ms(1000);
}
return NULL;
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
static LRESULT CALLBACK _power_window_proc(HWND hwnd, UINT uMsg, WPARAM wParam,
LPARAM lParam) {
switch (uMsg) {
case WM_POWERBROADCAST:
switch (wParam) {
case PBT_APMRESUMEAUTOMATIC:
// Resuming from sleep or hibernation, so kick all existing USB devices
// and then allow the device_poll_thread to redetect USB devices from
// scratch. If we don't do this, existing USB devices will never respond
// to us because they'll be waiting for the connect/auth handshake.
D("Received (WM_POWERBROADCAST, PBT_APMRESUMEAUTOMATIC) notification, "
"so kicking all USB devices\n");
kick_devices();
return TRUE;
}
}
return DefWindowProcW(hwnd, uMsg, wParam, lParam);
}
static void* _power_notification_thread(void* unused) {
// This uses a thread with its own window message pump to get power
// notifications. If adb runs from a non-interactive service account, this
// might not work (not sure). If that happens to not work, we could use
// heavyweight WMI APIs to get power notifications. But for the common case
// of a developer's interactive session, a window message pump is more
// appropriate.
D("Created power notification thread\n");
adb_thread_setname("Power Notifier");
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
// Window class names are process specific.
static const WCHAR kPowerNotificationWindowClassName[] =
L"PowerNotificationWindow";
// Get the HINSTANCE corresponding to the module that _power_window_proc
// is in (the main module).
const HINSTANCE instance = GetModuleHandleW(NULL);
if (!instance) {
// This is such a common API call that this should never fail.
fatal("GetModuleHandleW failed: %s",
SystemErrorCodeToString(GetLastError()).c_str());
}
WNDCLASSEXW wndclass;
memset(&wndclass, 0, sizeof(wndclass));
wndclass.cbSize = sizeof(wndclass);
wndclass.lpfnWndProc = _power_window_proc;
wndclass.hInstance = instance;
wndclass.lpszClassName = kPowerNotificationWindowClassName;
if (!RegisterClassExW(&wndclass)) {
fatal("RegisterClassExW failed: %s",
SystemErrorCodeToString(GetLastError()).c_str());
}
if (!CreateWindowExW(WS_EX_NOACTIVATE, kPowerNotificationWindowClassName,
L"ADB Power Notification Window", WS_POPUP, 0, 0, 0, 0,
NULL, NULL, instance, NULL)) {
fatal("CreateWindowExW failed: %s",
SystemErrorCodeToString(GetLastError()).c_str());
}
MSG msg;
while (GetMessageW(&msg, NULL, 0, 0)) {
TranslateMessage(&msg);
DispatchMessageW(&msg);
}
// GetMessageW() will return false if a quit message is posted. We don't
// do that, but it might be possible for that to occur when logging off or
// shutting down. Not a big deal since the whole process will be going away
// soon anyway.
D("Power notification thread exiting\n");
return NULL;
}
void usb_init() {
if (!adb_thread_create(device_poll_thread, nullptr)) {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
fatal_errno("cannot create device poll thread");
}
if (!adb_thread_create(_power_notification_thread, nullptr)) {
fatal_errno("cannot create power notification thread");
}
}
usb_handle* do_usb_open(const wchar_t* interface_name) {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
unsigned long name_len = 0;
// Allocate our handle
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
usb_handle* ret = (usb_handle*)calloc(1, sizeof(usb_handle));
if (NULL == ret) {
D("Could not allocate %u bytes for usb_handle: %s\n", sizeof(usb_handle),
strerror(errno));
goto fail;
}
// Set linkers back to the handle
ret->next = ret;
ret->prev = ret;
// Create interface.
ret->adb_interface = AdbCreateInterfaceByName(interface_name);
if (NULL == ret->adb_interface) {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("AdbCreateInterfaceByName failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
goto fail;
}
// Open read pipe (endpoint)
ret->adb_read_pipe =
AdbOpenDefaultBulkReadEndpoint(ret->adb_interface,
AdbOpenAccessTypeReadWrite,
AdbOpenSharingModeReadWrite);
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
if (NULL == ret->adb_read_pipe) {
D("AdbOpenDefaultBulkReadEndpoint failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
goto fail;
}
// Open write pipe (endpoint)
ret->adb_write_pipe =
AdbOpenDefaultBulkWriteEndpoint(ret->adb_interface,
AdbOpenAccessTypeReadWrite,
AdbOpenSharingModeReadWrite);
if (NULL == ret->adb_write_pipe) {
D("AdbOpenDefaultBulkWriteEndpoint failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
goto fail;
}
// Save interface name
// First get expected name length
AdbGetInterfaceName(ret->adb_interface,
NULL,
&name_len,
true);
if (0 == name_len) {
D("AdbGetInterfaceName returned name length of zero: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
goto fail;
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
ret->interface_name = (char*)malloc(name_len);
if (NULL == ret->interface_name) {
D("Could not allocate %lu bytes for interface_name: %s\n", name_len,
strerror(errno));
goto fail;
}
// Now save the name
if (!AdbGetInterfaceName(ret->adb_interface,
ret->interface_name,
&name_len,
true)) {
D("AdbGetInterfaceName failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
goto fail;
}
// We're done at this point
return ret;
fail:
if (NULL != ret) {
usb_cleanup_handle(ret);
free(ret);
}
return NULL;
}
int usb_write(usb_handle* handle, const void* data, int len) {
unsigned long time_out = 5000;
unsigned long written = 0;
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
int err = 0;
D("usb_write %d\n", len);
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
if (NULL == handle) {
D("usb_write was passed NULL handle\n");
err = EINVAL;
goto fail;
}
// Perform write
if (!AdbWriteEndpointSync(handle->adb_write_pipe,
(void*)data,
(unsigned long)len,
&written,
time_out)) {
D("AdbWriteEndpointSync failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
err = EIO;
goto fail;
}
// Make sure that we've written what we were asked to write
D("usb_write got: %ld, expected: %d\n", written, len);
if (written != (unsigned long)len) {
// If this occurs, this code should be changed to repeatedly call
// AdbWriteEndpointSync() until all bytes are written.
D("AdbWriteEndpointSync was supposed to write %d, but only wrote %ld\n",
len, written);
err = EIO;
goto fail;
}
if (handle->zero_mask && (len & handle->zero_mask) == 0) {
// Send a zero length packet
if (!AdbWriteEndpointSync(handle->adb_write_pipe,
(void*)data,
0,
&written,
time_out)) {
D("AdbWriteEndpointSync of zero length packet failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
err = EIO;
goto fail;
}
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
return 0;
fail:
// Any failure should cause us to kick the device instead of leaving it a
// zombie state with potential to hang.
if (NULL != handle) {
D("Kicking device due to error in usb_write\n");
usb_kick(handle);
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("usb_write failed\n");
errno = err;
return -1;
}
int usb_read(usb_handle *handle, void* data, int len) {
unsigned long time_out = 0;
unsigned long read = 0;
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
int err = 0;
D("usb_read %d\n", len);
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
if (NULL == handle) {
D("usb_read was passed NULL handle\n");
err = EINVAL;
goto fail;
}
while (len > 0) {
if (!AdbReadEndpointSync(handle->adb_read_pipe, data, len, &read,
time_out)) {
D("AdbReadEndpointSync failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
err = EIO;
goto fail;
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("usb_read got: %ld, expected: %d\n", read, len);
data = (char *)data + read;
len -= read;
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
return 0;
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
fail:
// Any failure should cause us to kick the device instead of leaving it a
// zombie state with potential to hang.
if (NULL != handle) {
D("Kicking device due to error in usb_read\n");
usb_kick(handle);
}
D("usb_read failed\n");
errno = err;
return -1;
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
// Wrapper around AdbCloseHandle() that logs diagnostics.
static void _adb_close_handle(ADBAPIHANDLE adb_handle) {
if (!AdbCloseHandle(adb_handle)) {
D("AdbCloseHandle(%p) failed: %s\n", adb_handle,
SystemErrorCodeToString(GetLastError()).c_str());
}
}
void usb_cleanup_handle(usb_handle* handle) {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("usb_cleanup_handle\n");
if (NULL != handle) {
if (NULL != handle->interface_name)
free(handle->interface_name);
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
// AdbCloseHandle(pipe) will break any threads out of pending IO calls and
// wait until the pipe no longer uses the interface. Then we can
// AdbCloseHandle() the interface.
if (NULL != handle->adb_write_pipe)
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
_adb_close_handle(handle->adb_write_pipe);
if (NULL != handle->adb_read_pipe)
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
_adb_close_handle(handle->adb_read_pipe);
if (NULL != handle->adb_interface)
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
_adb_close_handle(handle->adb_interface);
handle->interface_name = NULL;
handle->adb_write_pipe = NULL;
handle->adb_read_pipe = NULL;
handle->adb_interface = NULL;
}
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
static void usb_kick_locked(usb_handle* handle) {
// The reason the lock must be acquired before calling this function is in
// case multiple threads are trying to kick the same device at the same time.
usb_cleanup_handle(handle);
}
void usb_kick(usb_handle* handle) {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("usb_kick\n");
if (NULL != handle) {
adb_mutex_lock(&usb_lock);
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
usb_kick_locked(handle);
adb_mutex_unlock(&usb_lock);
} else {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
errno = EINVAL;
}
}
int usb_close(usb_handle* handle) {
D("usb_close\n");
if (NULL != handle) {
// Remove handle from the list
adb_mutex_lock(&usb_lock);
if ((handle->next != handle) && (handle->prev != handle)) {
handle->next->prev = handle->prev;
handle->prev->next = handle->next;
handle->prev = handle;
handle->next = handle;
}
adb_mutex_unlock(&usb_lock);
// Cleanup handle
usb_cleanup_handle(handle);
free(handle);
}
return 0;
}
int recognized_device(usb_handle* handle) {
if (NULL == handle)
return 0;
// Check vendor and product id first
USB_DEVICE_DESCRIPTOR device_desc;
if (!AdbGetUsbDeviceDescriptor(handle->adb_interface,
&device_desc)) {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("AdbGetUsbDeviceDescriptor failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
return 0;
}
// Then check interface properties
USB_INTERFACE_DESCRIPTOR interf_desc;
if (!AdbGetUsbInterfaceDescriptor(handle->adb_interface,
&interf_desc)) {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("AdbGetUsbInterfaceDescriptor failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
return 0;
}
// Must have two endpoints
if (2 != interf_desc.bNumEndpoints) {
return 0;
}
if (is_adb_interface(device_desc.idVendor, device_desc.idProduct,
interf_desc.bInterfaceClass, interf_desc.bInterfaceSubClass, interf_desc.bInterfaceProtocol)) {
if(interf_desc.bInterfaceProtocol == 0x01) {
AdbEndpointInformation endpoint_info;
// assuming zero is a valid bulk endpoint ID
if (AdbGetEndpointInformation(handle->adb_interface, 0, &endpoint_info)) {
handle->zero_mask = endpoint_info.max_packet_size - 1;
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("device zero_mask: 0x%x\n", handle->zero_mask);
} else {
D("AdbGetEndpointInformation failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
}
}
return 1;
}
return 0;
}
void find_devices() {
usb_handle* handle = NULL;
char entry_buffer[2048];
char interf_name[2048];
AdbInterfaceInfo* next_interface = (AdbInterfaceInfo*)(&entry_buffer[0]);
unsigned long entry_buffer_size = sizeof(entry_buffer);
char* copy_name;
// Enumerate all present and active interfaces.
ADBAPIHANDLE enum_handle =
AdbEnumInterfaces(usb_class_id, true, true, true);
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
if (NULL == enum_handle) {
D("AdbEnumInterfaces failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
return;
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
}
while (AdbNextInterface(enum_handle, next_interface, &entry_buffer_size)) {
// TODO: FIXME - temp hack converting wchar_t into char.
// It would be better to change AdbNextInterface so it will return
// interface name as single char string.
const wchar_t* wchar_name = next_interface->device_name;
for(copy_name = interf_name;
L'\0' != *wchar_name;
wchar_name++, copy_name++) {
*copy_name = (char)(*wchar_name);
}
*copy_name = '\0';
// Lets see if we already have this device in the list
if (!known_device(interf_name)) {
// This seems to be a new device. Open it!
handle = do_usb_open(next_interface->device_name);
if (NULL != handle) {
// Lets see if this interface (device) belongs to us
if (recognized_device(handle)) {
D("adding a new device %s\n", interf_name);
char serial_number[512];
unsigned long serial_number_len = sizeof(serial_number);
if (AdbGetSerialNumber(handle->adb_interface,
serial_number,
&serial_number_len,
true)) {
// Lets make sure that we don't duplicate this device
if (register_new_device(handle)) {
register_usb_transport(handle, serial_number, NULL, 1);
} else {
D("register_new_device failed for %s\n", interf_name);
usb_cleanup_handle(handle);
free(handle);
}
} else {
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
D("cannot get serial number: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
usb_cleanup_handle(handle);
free(handle);
}
} else {
usb_cleanup_handle(handle);
free(handle);
}
}
}
entry_buffer_size = sizeof(entry_buffer);
}
adb: win32: fix USB device hang when resuming from sleep/hibernation After resuming Windows from sleep or hibernation, USB connections are not automatically disconnected. Writing to the USB connections does not return any errors, but read never returns. My theory is that the device saw the host sleep/hibernation as a disconnect, so the device is waiting for re-auth from the host as if the host was just connected. To solve this, detect resume from sleep/hibernation, disconnect all USB connections and let the device poll thread re-detect the USB devices in 1 sec. This is done by using a hidden window that receives power notifications. The hidden window code is based on Chromium's similar code (platform-tools already includes the Chromium Authors license). This depends on a change to AdbWinUsbApi.dll that makes AdbCloseHandle(endpoint) abort any pending IOs and wait for those IOs to be aborted. I expect that this should solve many adb and Android Studio related bugs regarding hangs or errors. Also in this change: - Add D() logging for any errors from AdbWinApi.dll API calls. - Stop setting errno to Win32 error values which the caller can't really do anything with. Stop calling SetLastError() because the callers don't check GetLastError() anyway. - Check the return value from writing zero length packets. - If the full amount of data isn't written, return an error. - Upon any usb_read/usb_write error, kick the connection instead of only kicking when ERROR_INVALID_HANDLE is encountered. - Restructure some code from nested-if-trees to goto-fail to make it easier to follow. - Delete usb_name() since it isn't thread-safe and it isn't used. Change-Id: Iffcf5315ad8593d0c7e93012afaabe6fae354ac1 Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
2015-07-23 01:17:07 +02:00
if (GetLastError() != ERROR_NO_MORE_ITEMS) {
// Only ERROR_NO_MORE_ITEMS is expected at the end of enumeration.
D("AdbNextInterface failed: %s\n",
SystemErrorCodeToString(GetLastError()).c_str());
}
_adb_close_handle(enum_handle);
}
static void kick_devices() {
// Need to acquire lock to safely walk the list which might be modified
// by another thread.
adb_mutex_lock(&usb_lock);
for (usb_handle* usb = handle_list.next; usb != &handle_list; usb = usb->next) {
usb_kick_locked(usb);
}
adb_mutex_unlock(&usb_lock);
}