rx: Don't clear the receive queue when out of packets

We can end up discarding a receive queue that's been soft acked,
effectively taking back soft acks we sent. Whilst the RX
documentation says that a client can drop soft acked packets at
will, our RX implementation assumes that if the final packet in
a call has been soft acked, we won't clear the queue. If a client
clears the queue in this situation, the call will hang.

What *should* happen is that we should take necessary locks,
confirm that we have not soft-acked all of the packets in a flow,
and then discard, or, if we're just going to discard, error the
call.

Change-Id: Ic8e358b8648c1a6f0154009093468531a9e3cf74
Reviewed-on: http://gerrit.openafs.org/5603
Reviewed-by: Derrick Brashear <shadow@dementix.org>
Tested-by: Derrick Brashear <shadow@dementix.org>
This commit is contained in:
Simon Wilkinson 2011-10-12 09:47:14 -04:00 committed by Derrick Brashear
parent 735873d515
commit f722a75fe2
2 changed files with 3 additions and 3 deletions

View File

@ -3803,8 +3803,9 @@ rxi_ReceiveDataPacket(struct rx_call *call,
call->rprev = np->header.serial;
rxi_calltrace(RX_TRACE_DROP, call);
dpf(("packet %"AFS_PTR_FMT" dropped on receipt - quota problems\n", np));
if (rxi_doreclaim)
rxi_ClearReceiveQueue(call);
/* We used to clear the receive queue here, in an attempt to free
* packets. However this is unsafe if the queue has received a
* soft ACK for the final packet */
clock_GetTime(&now);
when = now;
clock_Add(&when, &rx_softAckDelay);

View File

@ -481,7 +481,6 @@ EXT osi_socket rx_minSocketNumber GLOBALSINIT(0x7fffffff);
#define RX_MAX_QUOTA 15 /* part of min packet computation */
EXT int rx_packetQuota[RX_N_PACKET_CLASSES] GLOBALSINIT(RX_PACKET_QUOTAS);
EXT int meltdown_1pkt GLOBALSINIT(1); /* prefer to schedule single-packet calls */
EXT int rxi_doreclaim GLOBALSINIT(1); /* if discard one packet, discard all */
EXT int rxi_md2cnt GLOBALSINIT(0); /* counter of skipped calls */
EXT int rxi_2dchoice GLOBALSINIT(1); /* keep track of another call to schedule */