Added debug info at the place that seems to deadlock in SMP systems with > 2 CPUs
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@23209 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
parent
8c27b8623d
commit
7eb5430510
|
@ -216,6 +216,7 @@ scan_device_int(ide_device_info *device, bool atapi)
|
|||
// do a short wait first - if there's no device at all we could wait forever
|
||||
// ToDo: have a look at this; if it times out (when the time is too short),
|
||||
// the kernel seems to crash a little later)!
|
||||
TRACE("scan_device_int: waiting 100ms...\n");
|
||||
if (acquire_sem_etc(bus->sync_wait_sem, 1, B_RELATIVE_TIMEOUT, 100000) == B_TIMED_OUT) {
|
||||
bool cont;
|
||||
|
||||
|
@ -248,6 +249,7 @@ scan_device_int(ide_device_info *device, bool atapi)
|
|||
// there is something, so wait for it
|
||||
acquire_sem(bus->sync_wait_sem);
|
||||
}
|
||||
TRACE("scan_device_int: got a fast response\n");
|
||||
|
||||
// cancel the timeout manually; usually this is done by wait_for_sync(), but
|
||||
// we've used the wait semaphore directly
|
||||
|
|
Loading…
Reference in New Issue