Filter PCI_address_space bit out of bus_master_base in
ide_adapter_probe_controller(). ide_adapter_detect_controller() does that itself, but ide_adapter_detect_channel() doesn't. Fixes off-by-one bus master status register access. Shouldn't have caused any serious problems, though. git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22950 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
parent
aaa10d6056
commit
e0fa65f7bd
@ -716,6 +716,8 @@ ide_adapter_probe_controller(device_node_handle parent, const char *controller_d
|
||||
if (res != B_OK || controller_node == NULL)
|
||||
goto err;
|
||||
|
||||
bus_master_base &= ~PCI_address_space;
|
||||
|
||||
// ignore errors during registration of channels - could be a simple rescan collision
|
||||
ide_adapter_detect_channel(pci, device, controller_node, channel_module_name,
|
||||
can_dma, command_block_base[0], control_block_base[0], bus_master_base,
|
||||
|
Loading…
Reference in New Issue
Block a user