On 8/2/21 11:59 AM, Torge Matthies wrote:
Fixes a crash when shutting down a prefix.
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51479 Signed-off-by: Torge Matthies openglfreak@googlemail.com
dlls/wineusb.sys/wineusb.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/dlls/wineusb.sys/wineusb.c b/dlls/wineusb.sys/wineusb.c index fae297915fc..28a36b9f380 100644 --- a/dlls/wineusb.sys/wineusb.c +++ b/dlls/wineusb.sys/wineusb.c @@ -413,12 +413,15 @@ static NTSTATUS pdo_pnp(DEVICE_OBJECT *device_obj, IRP *irp) break;
case IRP_MN_REMOVE_DEVICE:
EnterCriticalSection(&wineusb_cs); remove_pending_irps(device); libusb_unref_device(device->libusb_device); libusb_close(device->handle);list_remove(&device->entry); IoDeleteDevice(device->device_obj);LeaveCriticalSection(&wineusb_cs); ret = STATUS_SUCCESS; break;
This looks wrong; the device could have been removed by remove_usb_device().
[1] implies that we actually shouldn't remove the device at all unless we reported it missing via IRP_MN_QUERY_DEVICE_RELATIONS, which probably implies that we shouldn't remove and delete the device here, but should instead remove (and delete?) it from the parent IRP_MN_SURPRISE_REMOVAL/IRP_MN_REMOVE_DEVICE. It's not particularly clear though. I'll have to run some tests and check native drivers, but I won't have access to those until next week.
[1] https://docs.microsoft.com/en-us/windows-hardware/drivers/kernel/removing-a-...