adb: libusb: switch from polling for accessibility to a sleep.

For unclear reasons, it seems that for some people, devices are
becoming accessible, inaccessible, and then inaccessible again.
Switch to a sleep for now.

Bug: http://b/62200735
Test: manual
Change-Id: Id51dcb188c3534a57c35cbfd30a181c99115a23c
diff --git a/adb/client/usb_libusb.cpp b/adb/client/usb_libusb.cpp
index dc1a6b4..e7f44c6 100644
--- a/adb/client/usb_libusb.cpp
+++ b/adb/client/usb_libusb.cpp
@@ -180,10 +180,6 @@
     if (port_count < 0) return "";
     return StringPrintf("/dev/bus/usb/%03u/%03u", libusb_get_bus_number(device), ports[0]);
 }
-
-static bool is_device_accessible(libusb_device* device) {
-    return access(get_device_dev_path(device).c_str(), R_OK | W_OK) == 0;
-}
 #endif
 
 static bool endpoint_is_output(uint8_t endpoint) {
@@ -389,18 +385,12 @@
 static void device_connected(libusb_device* device) {
 #if defined(__linux__)
     // Android's host linux libusb uses netlink instead of udev for device hotplug notification,
-    // which means we can get hotplug notifications before udev has updated ownership/perms on
-    // the device. Since we're not going to be able to link against the system's libudev any
-    // time soon, hack around this by checking for accessibility in a loop.
+    // which means we can get hotplug notifications before udev has updated ownership/perms on the
+    // device. Since we're not going to be able to link against the system's libudev any time soon,
+    // hack around this by inserting a sleep.
     auto thread = std::thread([device]() {
         std::string device_path = get_device_dev_path(device);
-        auto start = std::chrono::steady_clock::now();
-        while (std::chrono::steady_clock::now() - start < 500ms) {
-            if (is_device_accessible(device)) {
-                break;
-            }
-            std::this_thread::sleep_for(10ms);
-        }
+        std::this_thread::sleep_for(1s);
 
         process_device(device);
         if (--connecting_devices == 0) {