linux_dsm_epyc7002/net/rxrpc/conn_event.c

507 lines
12 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/* connection-level event handling
*
* Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
#include <linux/net.h>
#include <linux/skbuff.h>
#include <linux/errqueue.h>
#include <net/sock.h>
#include <net/af_rxrpc.h>
#include <net/ip.h>
#include "ar-internal.h"
/*
* Retransmit terminal ACK or ABORT of the previous call.
*/
static void rxrpc_conn_retransmit_call(struct rxrpc_connection *conn,
struct sk_buff *skb,
unsigned int channel)
{
struct rxrpc_skb_priv *sp = skb ? rxrpc_skb(skb) : NULL;
struct rxrpc_channel *chan;
struct msghdr msg;
struct kvec iov[3];
struct {
struct rxrpc_wire_header whdr;
union {
__be32 abort_code;
struct rxrpc_ackpacket ack;
};
} __attribute__((packed)) pkt;
struct rxrpc_ackinfo ack_info;
size_t len;
int ret, ioc;
u32 serial, mtu, call_id, padding;
_enter("%d", conn->debug_id);
chan = &conn->channels[channel];
/* If the last call got moved on whilst we were waiting to run, just
* ignore this packet.
*/
call_id = READ_ONCE(chan->last_call);
/* Sync with __rxrpc_disconnect_call() */
smp_rmb();
if (skb && call_id != sp->hdr.callNumber)
return;
msg.msg_name = &conn->params.peer->srx.transport;
msg.msg_namelen = conn->params.peer->srx.transport_len;
msg.msg_control = NULL;
msg.msg_controllen = 0;
msg.msg_flags = 0;
iov[0].iov_base = &pkt;
iov[0].iov_len = sizeof(pkt.whdr);
iov[1].iov_base = &padding;
iov[1].iov_len = 3;
iov[2].iov_base = &ack_info;
iov[2].iov_len = sizeof(ack_info);
pkt.whdr.epoch = htonl(conn->proto.epoch);
pkt.whdr.cid = htonl(conn->proto.cid | channel);
pkt.whdr.callNumber = htonl(call_id);
pkt.whdr.seq = 0;
pkt.whdr.type = chan->last_type;
pkt.whdr.flags = conn->out_clientflag;
pkt.whdr.userStatus = 0;
pkt.whdr.securityIndex = conn->security_ix;
pkt.whdr._rsvd = 0;
pkt.whdr.serviceId = htons(conn->service_id);
len = sizeof(pkt.whdr);
switch (chan->last_type) {
case RXRPC_PACKET_TYPE_ABORT:
pkt.abort_code = htonl(chan->last_abort);
iov[0].iov_len += sizeof(pkt.abort_code);
len += sizeof(pkt.abort_code);
ioc = 1;
break;
case RXRPC_PACKET_TYPE_ACK:
mtu = conn->params.peer->if_mtu;
mtu -= conn->params.peer->hdrsize;
pkt.ack.bufferSpace = 0;
pkt.ack.maxSkew = htons(skb ? skb->priority : 0);
pkt.ack.firstPacket = htonl(chan->last_seq + 1);
pkt.ack.previousPacket = htonl(chan->last_seq);
pkt.ack.serial = htonl(skb ? sp->hdr.serial : 0);
pkt.ack.reason = skb ? RXRPC_ACK_DUPLICATE : RXRPC_ACK_IDLE;
pkt.ack.nAcks = 0;
ack_info.rxMTU = htonl(rxrpc_rx_mtu);
ack_info.maxMTU = htonl(mtu);
ack_info.rwind = htonl(rxrpc_rx_window_size);
ack_info.jumbo_max = htonl(rxrpc_rx_jumbo_max);
pkt.whdr.flags |= RXRPC_SLOW_START_OK;
padding = 0;
iov[0].iov_len += sizeof(pkt.ack);
len += sizeof(pkt.ack) + 3 + sizeof(ack_info);
ioc = 3;
break;
default:
return;
}
/* Resync with __rxrpc_disconnect_call() and check that the last call
* didn't get advanced whilst we were filling out the packets.
*/
smp_rmb();
if (READ_ONCE(chan->last_call) != call_id)
return;
serial = atomic_inc_return(&conn->serial);
pkt.whdr.serial = htonl(serial);
switch (chan->last_type) {
case RXRPC_PACKET_TYPE_ABORT:
_proto("Tx ABORT %%%u { %d } [re]", serial, conn->abort_code);
break;
case RXRPC_PACKET_TYPE_ACK:
trace_rxrpc_tx_ack(chan->call_debug_id, serial,
ntohl(pkt.ack.firstPacket),
ntohl(pkt.ack.serial),
pkt.ack.reason, 0);
_proto("Tx ACK %%%u [re]", serial);
break;
}
ret = kernel_sendmsg(conn->params.local->socket, &msg, iov, ioc, len);
rxrpc: Fix the keepalive generator [ver #2] AF_RXRPC has a keepalive message generator that generates a message for a peer ~20s after the last transmission to that peer to keep firewall ports open. The implementation is incorrect in the following ways: (1) It mixes up ktime_t and time64_t types. (2) It uses ktime_get_real(), the output of which may jump forward or backward due to adjustments to the time of day. (3) If the current time jumps forward too much or jumps backwards, the generator function will crank the base of the time ring round one slot at a time (ie. a 1s period) until it catches up, spewing out VERSION packets as it goes. Fix the problem by: (1) Only using time64_t. There's no need for sub-second resolution. (2) Use ktime_get_seconds() rather than ktime_get_real() so that time isn't perceived to go backwards. (3) Simplifying rxrpc_peer_keepalive_worker() by splitting it into two parts: (a) The "worker" function that manages the buckets and the timer. (b) The "dispatch" function that takes the pending peers and potentially transmits a keepalive packet before putting them back in the ring into the slot appropriate to the revised last-Tx time. (4) Taking everything that's pending out of the ring and splicing it into a temporary collector list for processing. In the case that there's been a significant jump forward, the ring gets entirely emptied and then the time base can be warped forward before the peers are processed. The warping can't happen if the ring isn't empty because the slot a peer is in is keepalive-time dependent, relative to the base time. (5) Limit the number of iterations of the bucket array when scanning it. (6) Set the timer to skip any empty slots as there's no point waking up if there's nothing to do yet. This can be triggered by an incoming call from a server after a reboot with AF_RXRPC and AFS built into the kernel causing a peer record to be set up before userspace is started. The system clock is then adjusted by userspace, thereby potentially causing the keepalive generator to have a meltdown - which leads to a message like: watchdog: BUG: soft lockup - CPU#0 stuck for 23s! [kworker/0:1:23] ... Workqueue: krxrpcd rxrpc_peer_keepalive_worker EIP: lock_acquire+0x69/0x80 ... Call Trace: ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? _raw_spin_lock_bh+0x29/0x60 ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? __lock_acquire+0x3d3/0x870 ? process_one_work+0x110/0x340 ? process_one_work+0x166/0x340 ? process_one_work+0x110/0x340 ? worker_thread+0x39/0x3c0 ? kthread+0xdb/0x110 ? cancel_delayed_work+0x90/0x90 ? kthread_stop+0x70/0x70 ? ret_from_fork+0x19/0x24 Fixes: ace45bec6d77 ("rxrpc: Fix firewall route keepalive") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-08-08 17:30:02 +07:00
conn->params.peer->last_tx_at = ktime_get_seconds();
if (ret < 0)
trace_rxrpc_tx_fail(chan->call_debug_id, serial, ret,
rxrpc_tx_point_call_final_resend);
else
trace_rxrpc_tx_packet(chan->call_debug_id, &pkt.whdr,
rxrpc_tx_point_call_final_resend);
_leave("");
}
/*
* pass a connection-level abort onto all calls on that connection
*/
static void rxrpc_abort_calls(struct rxrpc_connection *conn,
enum rxrpc_call_completion compl,
rxrpc_serial_t serial)
{
struct rxrpc_call *call;
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
int i;
_enter("{%d},%x", conn->debug_id, conn->abort_code);
rxrpc: Call channels should have separate call number spaces Each channel on a connection has a separate, independent number space from which to allocate callNumber values. It is entirely possible, for example, to have a connection with four active calls, each with call number 1. Note that the callNumber values for any particular channel don't have to start at 1, but they are supposed to increment monotonically for that channel from a client's perspective and may not be reused once the call number is transmitted (until the epoch cycles all the way back round). Currently, however, call numbers are allocated on a per-connection basis and, further, are held in an rb-tree. The rb-tree is redundant as the four channel pointers in the rxrpc_connection struct are entirely capable of pointing to all the calls currently in progress on a connection. To this end, make the following changes: (1) Handle call number allocation independently per channel. (2) Get rid of the conn->calls rb-tree. This is overkill as a connection may have a maximum of four calls in progress at any one time. Use the pointers in the channels[] array instead, indexed by the channel number from the packet. (3) For each channel, save the result of the last call that was in progress on that channel in conn->channels[] so that the final ACK or ABORT packet can be replayed if necessary. Any call earlier than that is just ignored. If we've seen the next call number in a packet, the last one is most definitely defunct. (4) When generating a RESPONSE packet for a connection, the call number counter for each channel must be included in it. (5) When parsing a RESPONSE packet for a connection, the call number counters contained therein should be used to set the minimum expected call numbers on each channel. To do in future commits: (1) Replay terminal packets based on the last call stored in conn->channels[]. (2) Connections should be retired before the callNumber space on any channel runs out. (3) A server is expected to disregard or reject any new incoming call that has a call number less than the current call number counter. The call number counter for that channel must be advanced to the new call number. Note that the server cannot just require that the next call that it sees on a channel be exactly the call number counter + 1 because then there's a scenario that could cause a problem: The client transmits a packet to initiate a connection, the network goes out, the server sends an ACK (which gets lost), the client sends an ABORT (which also gets lost); the network then reconnects, the client then reuses the call number for the next call (it doesn't know the server already saw the call number), but the server thinks it already has the first packet of this call (it doesn't know that the client doesn't know that it saw the call number the first time). Signed-off-by: David Howells <dhowells@redhat.com>
2016-06-27 20:39:44 +07:00
spin_lock(&conn->channel_lock);
rxrpc: Call channels should have separate call number spaces Each channel on a connection has a separate, independent number space from which to allocate callNumber values. It is entirely possible, for example, to have a connection with four active calls, each with call number 1. Note that the callNumber values for any particular channel don't have to start at 1, but they are supposed to increment monotonically for that channel from a client's perspective and may not be reused once the call number is transmitted (until the epoch cycles all the way back round). Currently, however, call numbers are allocated on a per-connection basis and, further, are held in an rb-tree. The rb-tree is redundant as the four channel pointers in the rxrpc_connection struct are entirely capable of pointing to all the calls currently in progress on a connection. To this end, make the following changes: (1) Handle call number allocation independently per channel. (2) Get rid of the conn->calls rb-tree. This is overkill as a connection may have a maximum of four calls in progress at any one time. Use the pointers in the channels[] array instead, indexed by the channel number from the packet. (3) For each channel, save the result of the last call that was in progress on that channel in conn->channels[] so that the final ACK or ABORT packet can be replayed if necessary. Any call earlier than that is just ignored. If we've seen the next call number in a packet, the last one is most definitely defunct. (4) When generating a RESPONSE packet for a connection, the call number counter for each channel must be included in it. (5) When parsing a RESPONSE packet for a connection, the call number counters contained therein should be used to set the minimum expected call numbers on each channel. To do in future commits: (1) Replay terminal packets based on the last call stored in conn->channels[]. (2) Connections should be retired before the callNumber space on any channel runs out. (3) A server is expected to disregard or reject any new incoming call that has a call number less than the current call number counter. The call number counter for that channel must be advanced to the new call number. Note that the server cannot just require that the next call that it sees on a channel be exactly the call number counter + 1 because then there's a scenario that could cause a problem: The client transmits a packet to initiate a connection, the network goes out, the server sends an ACK (which gets lost), the client sends an ABORT (which also gets lost); the network then reconnects, the client then reuses the call number for the next call (it doesn't know the server already saw the call number), but the server thinks it already has the first packet of this call (it doesn't know that the client doesn't know that it saw the call number the first time). Signed-off-by: David Howells <dhowells@redhat.com>
2016-06-27 20:39:44 +07:00
for (i = 0; i < RXRPC_MAXCALLS; i++) {
call = rcu_dereference_protected(
conn->channels[i].call,
lockdep_is_held(&conn->channel_lock));
if (call) {
if (compl == RXRPC_CALL_LOCALLY_ABORTED)
trace_rxrpc_abort(call->debug_id,
"CON", call->cid,
call->call_id, 0,
conn->abort_code,
conn->error);
else
trace_rxrpc_rx_abort(call, serial,
conn->abort_code);
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
if (rxrpc_set_call_completion(call, compl,
conn->abort_code,
conn->error))
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
rxrpc_notify_socket(call);
}
}
rxrpc: Call channels should have separate call number spaces Each channel on a connection has a separate, independent number space from which to allocate callNumber values. It is entirely possible, for example, to have a connection with four active calls, each with call number 1. Note that the callNumber values for any particular channel don't have to start at 1, but they are supposed to increment monotonically for that channel from a client's perspective and may not be reused once the call number is transmitted (until the epoch cycles all the way back round). Currently, however, call numbers are allocated on a per-connection basis and, further, are held in an rb-tree. The rb-tree is redundant as the four channel pointers in the rxrpc_connection struct are entirely capable of pointing to all the calls currently in progress on a connection. To this end, make the following changes: (1) Handle call number allocation independently per channel. (2) Get rid of the conn->calls rb-tree. This is overkill as a connection may have a maximum of four calls in progress at any one time. Use the pointers in the channels[] array instead, indexed by the channel number from the packet. (3) For each channel, save the result of the last call that was in progress on that channel in conn->channels[] so that the final ACK or ABORT packet can be replayed if necessary. Any call earlier than that is just ignored. If we've seen the next call number in a packet, the last one is most definitely defunct. (4) When generating a RESPONSE packet for a connection, the call number counter for each channel must be included in it. (5) When parsing a RESPONSE packet for a connection, the call number counters contained therein should be used to set the minimum expected call numbers on each channel. To do in future commits: (1) Replay terminal packets based on the last call stored in conn->channels[]. (2) Connections should be retired before the callNumber space on any channel runs out. (3) A server is expected to disregard or reject any new incoming call that has a call number less than the current call number counter. The call number counter for that channel must be advanced to the new call number. Note that the server cannot just require that the next call that it sees on a channel be exactly the call number counter + 1 because then there's a scenario that could cause a problem: The client transmits a packet to initiate a connection, the network goes out, the server sends an ACK (which gets lost), the client sends an ABORT (which also gets lost); the network then reconnects, the client then reuses the call number for the next call (it doesn't know the server already saw the call number), but the server thinks it already has the first packet of this call (it doesn't know that the client doesn't know that it saw the call number the first time). Signed-off-by: David Howells <dhowells@redhat.com>
2016-06-27 20:39:44 +07:00
spin_unlock(&conn->channel_lock);
_leave("");
}
/*
* generate a connection-level abort
*/
static int rxrpc_abort_connection(struct rxrpc_connection *conn,
int error, u32 abort_code)
{
struct rxrpc_wire_header whdr;
struct msghdr msg;
struct kvec iov[2];
__be32 word;
size_t len;
u32 serial;
int ret;
_enter("%d,,%u,%u", conn->debug_id, error, abort_code);
/* generate a connection-level abort */
spin_lock_bh(&conn->state_lock);
if (conn->state >= RXRPC_CONN_REMOTELY_ABORTED) {
spin_unlock_bh(&conn->state_lock);
_leave(" = 0 [already dead]");
return 0;
}
conn->error = error;
conn->abort_code = abort_code;
conn->state = RXRPC_CONN_LOCALLY_ABORTED;
spin_unlock_bh(&conn->state_lock);
msg.msg_name = &conn->params.peer->srx.transport;
msg.msg_namelen = conn->params.peer->srx.transport_len;
msg.msg_control = NULL;
msg.msg_controllen = 0;
msg.msg_flags = 0;
whdr.epoch = htonl(conn->proto.epoch);
whdr.cid = htonl(conn->proto.cid);
whdr.callNumber = 0;
whdr.seq = 0;
whdr.type = RXRPC_PACKET_TYPE_ABORT;
whdr.flags = conn->out_clientflag;
whdr.userStatus = 0;
whdr.securityIndex = conn->security_ix;
whdr._rsvd = 0;
whdr.serviceId = htons(conn->service_id);
word = htonl(conn->abort_code);
iov[0].iov_base = &whdr;
iov[0].iov_len = sizeof(whdr);
iov[1].iov_base = &word;
iov[1].iov_len = sizeof(word);
len = iov[0].iov_len + iov[1].iov_len;
serial = atomic_inc_return(&conn->serial);
rxrpc_abort_calls(conn, RXRPC_CALL_LOCALLY_ABORTED, serial);
whdr.serial = htonl(serial);
_proto("Tx CONN ABORT %%%u { %d }", serial, conn->abort_code);
ret = kernel_sendmsg(conn->params.local->socket, &msg, iov, 2, len);
if (ret < 0) {
trace_rxrpc_tx_fail(conn->debug_id, serial, ret,
rxrpc_tx_point_conn_abort);
_debug("sendmsg failed: %d", ret);
return -EAGAIN;
}
trace_rxrpc_tx_packet(conn->debug_id, &whdr, rxrpc_tx_point_conn_abort);
rxrpc: Fix the keepalive generator [ver #2] AF_RXRPC has a keepalive message generator that generates a message for a peer ~20s after the last transmission to that peer to keep firewall ports open. The implementation is incorrect in the following ways: (1) It mixes up ktime_t and time64_t types. (2) It uses ktime_get_real(), the output of which may jump forward or backward due to adjustments to the time of day. (3) If the current time jumps forward too much or jumps backwards, the generator function will crank the base of the time ring round one slot at a time (ie. a 1s period) until it catches up, spewing out VERSION packets as it goes. Fix the problem by: (1) Only using time64_t. There's no need for sub-second resolution. (2) Use ktime_get_seconds() rather than ktime_get_real() so that time isn't perceived to go backwards. (3) Simplifying rxrpc_peer_keepalive_worker() by splitting it into two parts: (a) The "worker" function that manages the buckets and the timer. (b) The "dispatch" function that takes the pending peers and potentially transmits a keepalive packet before putting them back in the ring into the slot appropriate to the revised last-Tx time. (4) Taking everything that's pending out of the ring and splicing it into a temporary collector list for processing. In the case that there's been a significant jump forward, the ring gets entirely emptied and then the time base can be warped forward before the peers are processed. The warping can't happen if the ring isn't empty because the slot a peer is in is keepalive-time dependent, relative to the base time. (5) Limit the number of iterations of the bucket array when scanning it. (6) Set the timer to skip any empty slots as there's no point waking up if there's nothing to do yet. This can be triggered by an incoming call from a server after a reboot with AF_RXRPC and AFS built into the kernel causing a peer record to be set up before userspace is started. The system clock is then adjusted by userspace, thereby potentially causing the keepalive generator to have a meltdown - which leads to a message like: watchdog: BUG: soft lockup - CPU#0 stuck for 23s! [kworker/0:1:23] ... Workqueue: krxrpcd rxrpc_peer_keepalive_worker EIP: lock_acquire+0x69/0x80 ... Call Trace: ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? _raw_spin_lock_bh+0x29/0x60 ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? rxrpc_peer_keepalive_worker+0x5e/0x350 ? __lock_acquire+0x3d3/0x870 ? process_one_work+0x110/0x340 ? process_one_work+0x166/0x340 ? process_one_work+0x110/0x340 ? worker_thread+0x39/0x3c0 ? kthread+0xdb/0x110 ? cancel_delayed_work+0x90/0x90 ? kthread_stop+0x70/0x70 ? ret_from_fork+0x19/0x24 Fixes: ace45bec6d77 ("rxrpc: Fix firewall route keepalive") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-08-08 17:30:02 +07:00
conn->params.peer->last_tx_at = ktime_get_seconds();
_leave(" = 0");
return 0;
}
/*
* mark a call as being on a now-secured channel
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
* - must be called with BH's disabled.
*/
static void rxrpc_call_is_secure(struct rxrpc_call *call)
{
_enter("%p", call);
if (call) {
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
write_lock_bh(&call->state_lock);
if (call->state == RXRPC_CALL_SERVER_SECURING) {
call->state = RXRPC_CALL_SERVER_ACCEPTING;
rxrpc_notify_socket(call);
}
write_unlock_bh(&call->state_lock);
}
}
/*
* connection-level Rx packet processor
*/
static int rxrpc_process_event(struct rxrpc_connection *conn,
struct sk_buff *skb,
u32 *_abort_code)
{
struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
__be32 wtmp;
u32 abort_code;
int loop, ret;
if (conn->state >= RXRPC_CONN_REMOTELY_ABORTED) {
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
_leave(" = -ECONNABORTED [%u]", conn->state);
return -ECONNABORTED;
}
_enter("{%d},{%u,%%%u},", conn->debug_id, sp->hdr.type, sp->hdr.serial);
switch (sp->hdr.type) {
case RXRPC_PACKET_TYPE_DATA:
case RXRPC_PACKET_TYPE_ACK:
rxrpc_conn_retransmit_call(conn, skb,
sp->hdr.cid & RXRPC_CHANNELMASK);
return 0;
rxrpc: Ignore BUSY packets on old calls If we receive a BUSY packet for a call we think we've just completed, the packet is handed off to the connection processor to deal with - but the connection processor doesn't expect a BUSY packet and so flags a protocol error. Fix this by simply ignoring the BUSY packet for the moment. The symptom of this may appear as a system call failing with EPROTO. This may be triggered by pressing ctrl-C under some circumstances. This comes about we abort calls due to interruption by a signal (which we shouldn't do, but that's going to be a large fix and mostly in fs/afs/). What happens is that we abort the call and may also abort follow up calls too (this needs offloading somehoe). So we see a transmission of something like the following sequence of packets: DATA for call N ABORT call N DATA for call N+1 ABORT call N+1 in very quick succession on the same channel. However, the peer may have deferred the processing of the ABORT from the call N to a background thread and thus sees the DATA message from the call N+1 coming in before it has cleared the channel. Thus it sends a BUSY packet[*]. [*] Note that some implementations (OpenAFS, for example) mark the BUSY packet with one plus the callNumber of the call prior to call N. Ordinarily, this would be call N, but there's no requirement for the calls on a channel to be numbered strictly sequentially (the number is required to increase). This is wrong and means that the callNumber in the BUSY packet should be ignored (it really ought to be N+1 since that's what it's in response to). Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-03-16 23:27:10 +07:00
case RXRPC_PACKET_TYPE_BUSY:
/* Just ignore BUSY packets for now. */
return 0;
case RXRPC_PACKET_TYPE_ABORT:
if (skb_copy_bits(skb, sizeof(struct rxrpc_wire_header),
&wtmp, sizeof(wtmp)) < 0) {
trace_rxrpc_rx_eproto(NULL, sp->hdr.serial,
tracepoint_string("bad_abort"));
return -EPROTO;
}
abort_code = ntohl(wtmp);
_proto("Rx ABORT %%%u { ac=%d }", sp->hdr.serial, abort_code);
conn->error = -ECONNABORTED;
conn->abort_code = abort_code;
conn->state = RXRPC_CONN_REMOTELY_ABORTED;
rxrpc_abort_calls(conn, RXRPC_CALL_REMOTELY_ABORTED, sp->hdr.serial);
return -ECONNABORTED;
case RXRPC_PACKET_TYPE_CHALLENGE:
return conn->security->respond_to_challenge(conn, skb,
_abort_code);
case RXRPC_PACKET_TYPE_RESPONSE:
ret = conn->security->verify_response(conn, skb, _abort_code);
if (ret < 0)
return ret;
ret = conn->security->init_connection_security(conn);
if (ret < 0)
return ret;
ret = conn->security->prime_packet_security(conn);
if (ret < 0)
return ret;
rxrpc: Call channels should have separate call number spaces Each channel on a connection has a separate, independent number space from which to allocate callNumber values. It is entirely possible, for example, to have a connection with four active calls, each with call number 1. Note that the callNumber values for any particular channel don't have to start at 1, but they are supposed to increment monotonically for that channel from a client's perspective and may not be reused once the call number is transmitted (until the epoch cycles all the way back round). Currently, however, call numbers are allocated on a per-connection basis and, further, are held in an rb-tree. The rb-tree is redundant as the four channel pointers in the rxrpc_connection struct are entirely capable of pointing to all the calls currently in progress on a connection. To this end, make the following changes: (1) Handle call number allocation independently per channel. (2) Get rid of the conn->calls rb-tree. This is overkill as a connection may have a maximum of four calls in progress at any one time. Use the pointers in the channels[] array instead, indexed by the channel number from the packet. (3) For each channel, save the result of the last call that was in progress on that channel in conn->channels[] so that the final ACK or ABORT packet can be replayed if necessary. Any call earlier than that is just ignored. If we've seen the next call number in a packet, the last one is most definitely defunct. (4) When generating a RESPONSE packet for a connection, the call number counter for each channel must be included in it. (5) When parsing a RESPONSE packet for a connection, the call number counters contained therein should be used to set the minimum expected call numbers on each channel. To do in future commits: (1) Replay terminal packets based on the last call stored in conn->channels[]. (2) Connections should be retired before the callNumber space on any channel runs out. (3) A server is expected to disregard or reject any new incoming call that has a call number less than the current call number counter. The call number counter for that channel must be advanced to the new call number. Note that the server cannot just require that the next call that it sees on a channel be exactly the call number counter + 1 because then there's a scenario that could cause a problem: The client transmits a packet to initiate a connection, the network goes out, the server sends an ACK (which gets lost), the client sends an ABORT (which also gets lost); the network then reconnects, the client then reuses the call number for the next call (it doesn't know the server already saw the call number), but the server thinks it already has the first packet of this call (it doesn't know that the client doesn't know that it saw the call number the first time). Signed-off-by: David Howells <dhowells@redhat.com>
2016-06-27 20:39:44 +07:00
spin_lock(&conn->channel_lock);
spin_lock(&conn->state_lock);
if (conn->state == RXRPC_CONN_SERVICE_CHALLENGING) {
conn->state = RXRPC_CONN_SERVICE;
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
spin_unlock(&conn->state_lock);
for (loop = 0; loop < RXRPC_MAXCALLS; loop++)
rxrpc_call_is_secure(
rcu_dereference_protected(
rxrpc: Call channels should have separate call number spaces Each channel on a connection has a separate, independent number space from which to allocate callNumber values. It is entirely possible, for example, to have a connection with four active calls, each with call number 1. Note that the callNumber values for any particular channel don't have to start at 1, but they are supposed to increment monotonically for that channel from a client's perspective and may not be reused once the call number is transmitted (until the epoch cycles all the way back round). Currently, however, call numbers are allocated on a per-connection basis and, further, are held in an rb-tree. The rb-tree is redundant as the four channel pointers in the rxrpc_connection struct are entirely capable of pointing to all the calls currently in progress on a connection. To this end, make the following changes: (1) Handle call number allocation independently per channel. (2) Get rid of the conn->calls rb-tree. This is overkill as a connection may have a maximum of four calls in progress at any one time. Use the pointers in the channels[] array instead, indexed by the channel number from the packet. (3) For each channel, save the result of the last call that was in progress on that channel in conn->channels[] so that the final ACK or ABORT packet can be replayed if necessary. Any call earlier than that is just ignored. If we've seen the next call number in a packet, the last one is most definitely defunct. (4) When generating a RESPONSE packet for a connection, the call number counter for each channel must be included in it. (5) When parsing a RESPONSE packet for a connection, the call number counters contained therein should be used to set the minimum expected call numbers on each channel. To do in future commits: (1) Replay terminal packets based on the last call stored in conn->channels[]. (2) Connections should be retired before the callNumber space on any channel runs out. (3) A server is expected to disregard or reject any new incoming call that has a call number less than the current call number counter. The call number counter for that channel must be advanced to the new call number. Note that the server cannot just require that the next call that it sees on a channel be exactly the call number counter + 1 because then there's a scenario that could cause a problem: The client transmits a packet to initiate a connection, the network goes out, the server sends an ACK (which gets lost), the client sends an ABORT (which also gets lost); the network then reconnects, the client then reuses the call number for the next call (it doesn't know the server already saw the call number), but the server thinks it already has the first packet of this call (it doesn't know that the client doesn't know that it saw the call number the first time). Signed-off-by: David Howells <dhowells@redhat.com>
2016-06-27 20:39:44 +07:00
conn->channels[loop].call,
lockdep_is_held(&conn->channel_lock)));
rxrpc: Rewrite the data and ack handling code Rewrite the data and ack handling code such that: (1) Parsing of received ACK and ABORT packets and the distribution and the filing of DATA packets happens entirely within the data_ready context called from the UDP socket. This allows us to process and discard ACK and ABORT packets much more quickly (they're no longer stashed on a queue for a background thread to process). (2) We avoid calling skb_clone(), pskb_pull() and pskb_trim(). We instead keep track of the offset and length of the content of each packet in the sk_buff metadata. This means we don't do any allocation in the receive path. (3) Jumbo DATA packet parsing is now done in data_ready context. Rather than cloning the packet once for each subpacket and pulling/trimming it, we file the packet multiple times with an annotation for each indicating which subpacket is there. From that we can directly calculate the offset and length. (4) A call's receive queue can be accessed without taking locks (memory barriers do have to be used, though). (5) Incoming calls are set up from preallocated resources and immediately made live. They can than have packets queued upon them and ACKs generated. If insufficient resources exist, DATA packet #1 is given a BUSY reply and other DATA packets are discarded). (6) sk_buffs no longer take a ref on their parent call. To make this work, the following changes are made: (1) Each call's receive buffer is now a circular buffer of sk_buff pointers (rxtx_buffer) rather than a number of sk_buff_heads spread between the call and the socket. This permits each sk_buff to be in the buffer multiple times. The receive buffer is reused for the transmit buffer. (2) A circular buffer of annotations (rxtx_annotations) is kept parallel to the data buffer. Transmission phase annotations indicate whether a buffered packet has been ACK'd or not and whether it needs retransmission. Receive phase annotations indicate whether a slot holds a whole packet or a jumbo subpacket and, if the latter, which subpacket. They also note whether the packet has been decrypted in place. (3) DATA packet window tracking is much simplified. Each phase has just two numbers representing the window (rx_hard_ack/rx_top and tx_hard_ack/tx_top). The hard_ack number is the sequence number before base of the window, representing the last packet the other side says it has consumed. hard_ack starts from 0 and the first packet is sequence number 1. The top number is the sequence number of the highest-numbered packet residing in the buffer. Packets between hard_ack+1 and top are soft-ACK'd to indicate they've been received, but not yet consumed. Four macros, before(), before_eq(), after() and after_eq() are added to compare sequence numbers within the window. This allows for the top of the window to wrap when the hard-ack sequence number gets close to the limit. Two flags, RXRPC_CALL_RX_LAST and RXRPC_CALL_TX_LAST, are added also to indicate when rx_top and tx_top point at the packets with the LAST_PACKET bit set, indicating the end of the phase. (4) Calls are queued on the socket 'receive queue' rather than packets. This means that we don't need have to invent dummy packets to queue to indicate abnormal/terminal states and we don't have to keep metadata packets (such as ABORTs) around (5) The offset and length of a (sub)packet's content are now passed to the verify_packet security op. This is currently expected to decrypt the packet in place and validate it. However, there's now nowhere to store the revised offset and length of the actual data within the decrypted blob (there may be a header and padding to skip) because an sk_buff may represent multiple packets, so a locate_data security op is added to retrieve these details from the sk_buff content when needed. (6) recvmsg() now has to handle jumbo subpackets, where each subpacket is individually secured and needs to be individually decrypted. The code to do this is broken out into rxrpc_recvmsg_data() and shared with the kernel API. It now iterates over the call's receive buffer rather than walking the socket receive queue. Additional changes: (1) The timers are condensed to a single timer that is set for the soonest of three timeouts (delayed ACK generation, DATA retransmission and call lifespan). (2) Transmission of ACK and ABORT packets is effected immediately from process-context socket ops/kernel API calls that cause them instead of them being punted off to a background work item. The data_ready handler still has to defer to the background, though. (3) A shutdown op is added to the AF_RXRPC socket so that the AFS filesystem can shut down the socket and flush its own work items before closing the socket to deal with any in-progress service calls. Future additional changes that will need to be considered: (1) Make sure that a call doesn't hog the front of the queue by receiving data from the network as fast as userspace is consuming it to the exclusion of other calls. (2) Transmit delayed ACKs from within recvmsg() when we've consumed sufficiently more packets to avoid the background work item needing to run. Signed-off-by: David Howells <dhowells@redhat.com>
2016-09-08 17:10:12 +07:00
} else {
spin_unlock(&conn->state_lock);
}
rxrpc: Call channels should have separate call number spaces Each channel on a connection has a separate, independent number space from which to allocate callNumber values. It is entirely possible, for example, to have a connection with four active calls, each with call number 1. Note that the callNumber values for any particular channel don't have to start at 1, but they are supposed to increment monotonically for that channel from a client's perspective and may not be reused once the call number is transmitted (until the epoch cycles all the way back round). Currently, however, call numbers are allocated on a per-connection basis and, further, are held in an rb-tree. The rb-tree is redundant as the four channel pointers in the rxrpc_connection struct are entirely capable of pointing to all the calls currently in progress on a connection. To this end, make the following changes: (1) Handle call number allocation independently per channel. (2) Get rid of the conn->calls rb-tree. This is overkill as a connection may have a maximum of four calls in progress at any one time. Use the pointers in the channels[] array instead, indexed by the channel number from the packet. (3) For each channel, save the result of the last call that was in progress on that channel in conn->channels[] so that the final ACK or ABORT packet can be replayed if necessary. Any call earlier than that is just ignored. If we've seen the next call number in a packet, the last one is most definitely defunct. (4) When generating a RESPONSE packet for a connection, the call number counter for each channel must be included in it. (5) When parsing a RESPONSE packet for a connection, the call number counters contained therein should be used to set the minimum expected call numbers on each channel. To do in future commits: (1) Replay terminal packets based on the last call stored in conn->channels[]. (2) Connections should be retired before the callNumber space on any channel runs out. (3) A server is expected to disregard or reject any new incoming call that has a call number less than the current call number counter. The call number counter for that channel must be advanced to the new call number. Note that the server cannot just require that the next call that it sees on a channel be exactly the call number counter + 1 because then there's a scenario that could cause a problem: The client transmits a packet to initiate a connection, the network goes out, the server sends an ACK (which gets lost), the client sends an ABORT (which also gets lost); the network then reconnects, the client then reuses the call number for the next call (it doesn't know the server already saw the call number), but the server thinks it already has the first packet of this call (it doesn't know that the client doesn't know that it saw the call number the first time). Signed-off-by: David Howells <dhowells@redhat.com>
2016-06-27 20:39:44 +07:00
spin_unlock(&conn->channel_lock);
return 0;
default:
trace_rxrpc_rx_eproto(NULL, sp->hdr.serial,
tracepoint_string("bad_conn_pkt"));
return -EPROTO;
}
}
/*
* set up security and issue a challenge
*/
static void rxrpc_secure_connection(struct rxrpc_connection *conn)
{
u32 abort_code;
int ret;
_enter("{%d}", conn->debug_id);
ASSERT(conn->security_ix != 0);
if (!conn->params.key) {
_debug("set up security");
ret = rxrpc_init_server_conn_security(conn);
switch (ret) {
case 0:
break;
case -ENOENT:
abort_code = RX_CALL_DEAD;
goto abort;
default:
abort_code = RXKADNOAUTH;
goto abort;
}
}
if (conn->security->issue_challenge(conn) < 0) {
abort_code = RX_CALL_DEAD;
ret = -ENOMEM;
goto abort;
}
_leave("");
return;
abort:
_debug("abort %d, %d", ret, abort_code);
rxrpc_abort_connection(conn, ret, abort_code);
_leave(" [aborted]");
}
/*
* Process delayed final ACKs that we haven't subsumed into a subsequent call.
*/
static void rxrpc_process_delayed_final_acks(struct rxrpc_connection *conn)
{
unsigned long j = jiffies, next_j;
unsigned int channel;
bool set;
again:
next_j = j + LONG_MAX;
set = false;
for (channel = 0; channel < RXRPC_MAXCALLS; channel++) {
struct rxrpc_channel *chan = &conn->channels[channel];
unsigned long ack_at;
if (!test_bit(RXRPC_CONN_FINAL_ACK_0 + channel, &conn->flags))
continue;
smp_rmb(); /* vs rxrpc_disconnect_client_call */
ack_at = READ_ONCE(chan->final_ack_at);
if (time_before(j, ack_at)) {
if (time_before(ack_at, next_j)) {
next_j = ack_at;
set = true;
}
continue;
}
if (test_and_clear_bit(RXRPC_CONN_FINAL_ACK_0 + channel,
&conn->flags))
rxrpc_conn_retransmit_call(conn, NULL, channel);
}
j = jiffies;
if (time_before_eq(next_j, j))
goto again;
if (set)
rxrpc_reduce_conn_timer(conn, next_j);
}
/*
* connection-level event processor
*/
void rxrpc_process_connection(struct work_struct *work)
{
struct rxrpc_connection *conn =
container_of(work, struct rxrpc_connection, processor);
struct sk_buff *skb;
u32 abort_code = RX_PROTOCOL_ERROR;
int ret;
rxrpc_see_connection(conn);
if (test_and_clear_bit(RXRPC_CONN_EV_CHALLENGE, &conn->events))
rxrpc_secure_connection(conn);
/* Process delayed ACKs whose time has come. */
if (conn->flags & RXRPC_CONN_FINAL_ACK_MASK)
rxrpc_process_delayed_final_acks(conn);
/* go through the conn-level event packets, releasing the ref on this
* connection that each one has when we've finished with it */
while ((skb = skb_dequeue(&conn->rx_queue))) {
rxrpc_see_skb(skb, rxrpc_skb_seen);
ret = rxrpc_process_event(conn, skb, &abort_code);
switch (ret) {
case -EPROTO:
case -EKEYEXPIRED:
case -EKEYREJECTED:
goto protocol_error;
rxrpc: Don't put crypto buffers on the stack Don't put buffers of data to be handed to crypto on the stack as this may cause an assertion failure in the kernel (see below). Fix this by using an kmalloc'd buffer instead. kernel BUG at ./include/linux/scatterlist.h:147! ... RIP: 0010:rxkad_encrypt_response.isra.6+0x191/0x1b0 [rxrpc] RSP: 0018:ffffbe2fc06cfca8 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff989277d59900 RCX: 0000000000000028 RDX: 0000259dc06cfd88 RSI: 0000000000000025 RDI: ffffbe30406cfd88 RBP: ffffbe2fc06cfd60 R08: ffffbe2fc06cfd08 R09: ffffbe2fc06cfd08 R10: 0000000000000000 R11: 0000000000000000 R12: 1ffff7c5f80d9f95 R13: ffffbe2fc06cfd88 R14: ffff98927a3f7aa0 R15: ffffbe2fc06cfd08 FS: 0000000000000000(0000) GS:ffff98927fc00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 000055b1ff28f0f8 CR3: 000000001b412003 CR4: 00000000003606f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: rxkad_respond_to_challenge+0x297/0x330 [rxrpc] rxrpc_process_connection+0xd1/0x690 [rxrpc] ? process_one_work+0x1c3/0x680 ? __lock_is_held+0x59/0xa0 process_one_work+0x249/0x680 worker_thread+0x3a/0x390 ? process_one_work+0x680/0x680 kthread+0x121/0x140 ? kthread_create_worker_on_cpu+0x70/0x70 ret_from_fork+0x3a/0x50 Reported-by: Jonathan Billings <jsbillings@jsbillings.org> Reported-by: Marc Dionne <marc.dionne@auristor.com> Signed-off-by: David Howells <dhowells@redhat.com> Tested-by: Jonathan Billings <jsbillings@jsbillings.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-08 22:59:07 +07:00
case -ENOMEM:
case -EAGAIN:
goto requeue_and_leave;
case -ECONNABORTED:
default:
rxrpc_free_skb(skb, rxrpc_skb_freed);
break;
}
}
out:
rxrpc_put_connection(conn);
_leave("");
return;
requeue_and_leave:
skb_queue_head(&conn->rx_queue, skb);
goto out;
protocol_error:
if (rxrpc_abort_connection(conn, ret, abort_code) < 0)
goto requeue_and_leave;
rxrpc_free_skb(skb, rxrpc_skb_freed);
goto out;
}