evin/ms-nfs41-client
3
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

making all but CLOSE interruptable

Browse source 
leaving CLOSE upcall non-interruptable as it leads to issues with security context.

making all other upcalls interruptable so that when something goes wrong we can ctrl-c out of a user application. otherwise, the machine requires a reboot (ie caz the wait we made the wait non-interrutable so nothing can kill it).
This commit is contained in:
Olga Kornievskaia 2011-02-03 11:46:51 -05:00
parent 4411d3d807
commit 67ae1eddaf
1 changed files with 19 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

19
sys/nfs41_driver.c
View file

@ -1245,7 +1245,26 @@ NTSTATUS nfs41_UpcallWaitForReply(
KeSetEvent(&upcallEvent, 0, FALSE);
DbgP("@@@ Creating %s upcall entry=%p xid=%d\n", opstring, entry, entry->xid);
if (!entry->async_op) {
/* 02/03/2011 AGLO: it is not clear what the "right" waiting design should be.
* Having non-interruptable waiting seems to be the right approach. However,
* when things go wrong, the only wait to proceed is a reboot (since "waits"
* are not interruptable we can't stop a hung task.
* Having interruptable wait causes issues with security context.
* For now, I'm making CLOSE non-interruptable but keeping the rest interruptable
* so that we don't have to reboot all the time
*/
#define MAKE_WAITONCLOSE_NONITERRUPTABLE
#ifdef MAKE_WAITONCLOSE_NONITERRUPTABLE
if (entry->opcode == NFS41_CLOSE)
status = KeWaitForSingleObject(&entry->cond, Executive,
KernelMode, FALSE, NULL);
else
status = KeWaitForSingleObject(&entry->cond, Executive,
UserMode, TRUE, NULL);
#else
status = KeWaitForSingleObject(&entry->cond, Executive, KernelMode, FALSE, NULL);
#endif
print_wait_status(1, "[downcall]", status, opcode2string(entry->opcode),
entry, entry->xid);
} else