Skip to contentSkip to conversation

net/smc:Introduce SMC-D based loopback acceleration

RFC PATCH Superseded Wen Gu 5 participants 18 messages Dec 20, 2022 – Jan 30, 2023

Subsystems (4)

v2 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Wen Gu

author reviewer
#0 lore raw

v2 cover net/smc:Introduce SMC-D based loopback acceleration

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

 +------------------------------------------+
 |  +-----------+           +-----------+   |
 |  | process A |           | process B |   |
 |  +-----------+           +-----------+   |
 |       ^                        ^         |
 |       |    +---------------+   |         |
 |       |    |   SMC stack   |   |         |
 |       +--->| +-----------+ |<--|         |
 |            | |   dummy   | |             |
 |            | |   device  | |             |
 |            +-+-----------+-+             |
 |                   VM                     |
 +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

 +----------+                     +----------+
 | socket A |                     | socket B |
 +----------+                     +----------+
       |                               ^
       |         +---------+           |
  regard as      |         | ----------|
  local sndbuf   |  B's    |     regard as
       |         |  RMB    |     local RMB
       |-------> |         |
                 +---------+

# Benchmark Test

* Test environments:
- VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
- SMC sndbuf/RMB size 1MB.

* Test object:
- TCP: run on TCP loopback.
- domain: run on UNIX domain.
- SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
- SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

1. ipc-benchmark (see [3])

- ./<foo> -c 1000000 -s 100

TCP domain SMC-lo SMC-lo-nocpy
Message
rate (msg/s) 75140 129548(+72.41) 152266(+102.64%) 151914(+102.17%)

2. sockperf

- serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
- clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

TCP SMC-lo SMC-lo-nocpy
Bandwidth(MBps) 4943.359 4936.096(-0.15%) 8239.624(+66.68%)
Latency(us) 6.372 3.359(-47.28%) 3.25(-49.00%)

3. iperf3

- serv: <smc_run> taskset -c <cpu> iperf3 -s
- clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

TCP SMC-lo SMC-lo-nocpy
Bitrate(Gb/s) 40.5 41.4(+2.22%) 76.4(+88.64%)

4. nginx/wrk

- serv: <smc_run> nginx
- clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

TCP SMC-lo SMC-lo-nocpy
Requests/s 154643.22 220894.03(+42.84%) 226754.3(+46.63%)

# Discussion

1. API between SMC-D and ISM device

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1]) can provide some examples for new API design. And we are very glad to discuss this on the mail list.

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
1. Fix some build WARNINGs complained by kernel test rebot

2. Add iperf3 test data.

Wen Gu (5):
net/smc: introduce SMC-D loopback device
net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

 include/net/smc.h      |   3 +
 net/smc/Makefile       |   2 +-
 net/smc/af_smc.c       |  88 +++++++++++-
 net/smc/smc_cdc.c      |  59 ++++++--
 net/smc/smc_cdc.h      |   1 +
 net/smc/smc_clc.c      |   4 +-
 net/smc/smc_core.c     |  62 +++++++++
 net/smc/smc_core.h     |   2 +
 net/smc/smc_ism.c      |  39 +++++-
 net/smc/smc_ism.h      |   2 +
 net/smc/smc_loopback.c | 358 +++++++++++++++++++++++++++++++++++++++++++++++++
 net/smc/smc_loopback.h |  63 +++++++++
 12 files changed, 662 insertions(+), 21 deletions(-)
 create mode 100644 net/smc/smc_loopback.c
 create mode 100644 net/smc/smc_loopback.h
Signature · 1 line
1.8.3.1
author reviewer
#1 lore raw

RFC PATCH v2 1/5 net/smc: introduce SMC-D loopback device

This patch introduces a kind of loopback device for SMC-D, thus enabling the SMC communication between two local sockets in one kernel.

The loopback device supports basic capabilities defined by SMC-D, including registering DMB, unregistering DMB and moving data.

Considering that there is no ism device on other servers expect IBM z13, the loopback device can be used as a dummy device to test SMC-D logic for the broad community.

---
 include/net/smc.h      |   1 +
 net/smc/Makefile       |   2 +-
 net/smc/af_smc.c       |  12 ++-
 net/smc/smc_cdc.c      |   6 ++
 net/smc/smc_cdc.h      |   1 +
 net/smc/smc_loopback.c | 282 +++++++++++++++++++++++++++++++++++++++++++++++++
 net/smc/smc_loopback.h |  59 +++++++++++
 7 files changed, 361 insertions(+), 2 deletions(-)
 create mode 100644 net/smc/smc_loopback.c
 create mode 100644 net/smc/smc_loopback.h
include/net/smc.h +1 -0
index c926d33..7699f97 100644
@@ -93,6 +93,7 @@ struct smcd_dev {
 	atomic_t lgr_cnt;
 	wait_queue_head_t lgrs_deleted;
 	u8 going_away : 1;
+	u8 is_loopback : 1;
 };
 
 struct smcd_dev *smcd_alloc_dev(struct device *parent, const char *name,
net/smc/Makefile +1 -1
index 875efcd..a8c3711 100644
@@ -4,5 +4,5 @@ obj-$(CONFIG_SMC)	+= smc.o
 obj-$(CONFIG_SMC_DIAG)	+= smc_diag.o
 smc-y := af_smc.o smc_pnet.o smc_ib.o smc_clc.o smc_core.o smc_wr.o smc_llc.o
 smc-y += smc_cdc.o smc_tx.o smc_rx.o smc_close.o smc_ism.o smc_netlink.o smc_stats.o
-smc-y += smc_tracepoint.o
+smc-y += smc_tracepoint.o smc_loopback.o
 smc-$(CONFIG_SYSCTL) += smc_sysctl.o
net/smc/af_smc.c +11 -1
index e12d4fa..9546c02 100644
@@ -52,6 +52,7 @@
 #include "smc_stats.h"
 #include "smc_tracepoint.h"
 #include "smc_sysctl.h"
+#include "smc_loopback.h"
 
 static DEFINE_MUTEX(smc_server_lgr_pending);	/* serialize link group
 						 * creation on server
@@ -3451,15 +3452,23 @@ static int __init smc_init(void)
 		goto out_sock;
 	}
 
+	rc = smc_loopback_init();
+	if (rc) {
+		pr_err("%s: smc_loopback_init fails with %d\n", __func__, rc);
+		goto out_ib;
+	}
+
 	rc = tcp_register_ulp(&smc_ulp_ops);
 	if (rc) {
 		pr_err("%s: tcp_ulp_register fails with %d\n", __func__, rc);
-		goto out_ib;
+		goto out_lo;
 	}
 
 	static_branch_enable(&tcp_have_smc);
 	return 0;
 
+out_lo:
+	smc_loopback_exit();
 out_ib:
 	smc_ib_unregister_client();
 out_sock:
@@ -3494,6 +3503,7 @@ static void __exit smc_exit(void)
 	tcp_unregister_ulp(&smc_ulp_ops);
 	sock_unregister(PF_SMC);
 	smc_core_exit();
+	smc_loopback_exit();
 	smc_ib_unregister_client();
 	destroy_workqueue(smc_close_wq);
 	destroy_workqueue(smc_tcp_ls_wq);
net/smc/smc_cdc.c +6 -0
index 53f63bf..61f5ff7 100644
@@ -408,6 +408,12 @@ static void smc_cdc_msg_recv(struct smc_sock *smc, struct smc_cdc_msg *cdc)
 static void smcd_cdc_rx_tsklet(struct tasklet_struct *t)
 {
 	struct smc_connection *conn = from_tasklet(conn, t, rx_tsklet);
+
+	smcd_cdc_rx_handler(conn);
+}
+
+void smcd_cdc_rx_handler(struct smc_connection *conn)
+{
 	struct smcd_cdc_msg *data_cdc;
 	struct smcd_cdc_msg cdc;
 	struct smc_sock *smc;
net/smc/smc_cdc.h +1 -0
index 696cc11..11559d4 100644
@@ -301,5 +301,6 @@ int smcr_cdc_msg_send_validation(struct smc_connection *conn,
 				 struct smc_wr_buf *wr_buf);
 int smc_cdc_init(void) __init;
 void smcd_cdc_rx_init(struct smc_connection *conn);
+void smcd_cdc_rx_handler(struct smc_connection *conn);
 
 #endif /* SMC_CDC_H */
net/smc/smc_loopback.c new file +282 -0

282 lines not on this page · raw message

net/smc/smc_loopback.h new file +59 -0

59 lines not on this page · raw message

Signature · 1 line
1.8.3.1
#2 lore raw

Re: [RFC PATCH net-next v2 1/5] net/smc: introduce SMC-D loopback device

Quoted 9 lines · Wen Gu

This patch introduces a kind of loopback device for SMC-D, thus enabling the SMC communication between two local sockets in one kernel.

The loopback device supports basic capabilities defined by SMC-D, including registering DMB, unregistering DMB and moving data.

Considering that there is no ism device on other servers expect IBM z13,

Please use the wording 'on other architectures except s390'. That is how IBM Z is referred to in the Linux kernel.

Quoted 5 lines

the loopback device can be used as a dummy device to test SMC-D logic for the broad community.

---

Hello Wen Gu,

as the general design discussions are ongoing, I didn't do a thorough review. But here are some general remarks that you may want to consider for future versions.

I would propose to add a module parameter (default off) to enable SMC-D loopback.

Quoted 11 lines
 include/net/smc.h      |   1 +
 net/smc/Makefile       |   2 +-
 net/smc/af_smc.c       |  12 ++-
 net/smc/smc_cdc.c      |   6 ++
 net/smc/smc_cdc.h      |   1 +
 net/smc/smc_loopback.c | 282 +++++++++++++++++++++++++++++++++++++++++++++++++
 net/smc/smc_loopback.h |  59 +++++++++++
 7 files changed, 361 insertions(+), 2 deletions(-)
 create mode 100644 net/smc/smc_loopback.c
 create mode 100644 net/smc/smc_loopback.h

I am not convinced that this warrants a separate file.

[...]

Quoted 12 lines
net/smc/smc_loopback.c new file +11 -0
+}
+
+static int lo_add_vlan_id(struct smcd_dev *smcd, u64 vlan_id)
+{
+	return 0;
+}
+
+static int lo_del_vlan_id(struct smcd_dev *smcd, u64 vlan_id)
+{
+	return 0;
+}

net/smc/smc_loopback.c line 142

Quoted 10 lines
net/smc/smc_loopback.c new file +10 -0
+
+static int lo_set_vlan_required(struct smcd_dev *smcd)
+{
+	return 0;
+}
+
+static int lo_reset_vlan_required(struct smcd_dev *smcd)
+{
+	return 0;
+}

The VLAN functions are only required for SMC-Dv1 Seems you want to provide v1 support for loopback? May be nice for testing v1 VLAN support. But then you need proper VLAN support.

[...]

net/smc/smc_loopback.c line 185

Quoted 5 lines
net/smc/smc_loopback.c new file +5 -0
+
+static u8 *lo_get_system_eid(void)
+{
+	return &LO_SYSTEM_EID.seid_string[0];
+}

SEID is for the whole system not per device.
We probably need to register a different function for each architecture.

net/smc/smc_loopback.c line 191

Quoted 6 lines
net/smc/smc_loopback.c new file +6 -0
+
+static u16 lo_get_chid(struct smcd_dev *smcd)
+{
+	return 0;
+}
+

Shouldn't this return 0xFFFF in your current concept?

author reviewer
#3 lore raw

Re: [RFC PATCH net-next v2 1/5] net/smc: introduce SMC-D loopback device

Quoted 7 lines · Alexandra Winter
Quoted 4 lines · Wen Gu

This patch introduces a kind of loopback device for SMC-D, thus enabling the SMC communication between two local sockets in one kernel.

Quoted 9 lines
Quoted 5 lines

The loopback device supports basic capabilities defined by SMC-D, including registering DMB, unregistering DMB and moving data.

Considering that there is no ism device on other servers expect IBM z13,

Please use the wording 'on other architectures except s390'. That is how IBM Z is referred to in the Linux kernel.

Thanks, will use words consistent with this.

Quoted 7 lines
Quoted 6 lines

the loopback device can be used as a dummy device to test SMC-D logic for the broad community.

---

Quoted 8 lines

Hello Wen Gu,

as the general design discussions are ongoing, I didn't do a thorough review. But here are some general remarks that you may want to consider for future versions.
I would propose to add a module parameter (default off) to enable SMC-D loopback.

OK, will add a module parameter in the future version.

Quoted 12 lines
Quoted 11 lines
  include/net/smc.h      |   1 +
  net/smc/Makefile       |   2 +-
  net/smc/af_smc.c       |  12 ++-
  net/smc/smc_cdc.c      |   6 ++
  net/smc/smc_cdc.h      |   1 +
  net/smc/smc_loopback.c | 282 +++++++++++++++++++++++++++++++++++++++++++++++++
  net/smc/smc_loopback.h |  59 +++++++++++
  7 files changed, 361 insertions(+), 2 deletions(-)

create mode 100644 net/smc/smc_loopback.c
create mode 100644 net/smc/smc_loopback.h

Quoted 1 line

I am not convinced that this warrants a separate file.

IMHO, the dummy device used by smcd loopback is corresponding to ISM device. So I put the dummy device implementation in smc_loopback.c alone, imitating drivers/s390/net/ism_drv.c. I think it maybe clearer to do so.

Quoted 21 lines

[...]

Quoted 19 lines
net/smc/smc_loopback.c new file +18 -0
+}
+
+static int lo_add_vlan_id(struct smcd_dev *smcd, u64 vlan_id)
+{
+	return 0;
+}
+
+static int lo_del_vlan_id(struct smcd_dev *smcd, u64 vlan_id)
+{
+	return 0;
+}
+
+static int lo_set_vlan_required(struct smcd_dev *smcd)
+{
+	return 0;
+}
+
+static int lo_reset_vlan_required(struct smcd_dev *smcd)
Quoted 9 lines

net/smc/smc_loopback.c line 40

Quoted 3 lines
net/smc/smc_loopback.c new file +3 -0
+{
+	return 0;
+}

The VLAN functions are only required for SMC-Dv1 Seems you want to provide v1 support for loopback? May be nice for testing v1 VLAN support. But then you need proper VLAN support.

Based on the current discussion, I tend to only provide v2 support for loopback since v2 is the general trend. So I will fix this in the future version.

Quoted 10 lines

[...]

net/smc/smc_loopback.c line 185

Quoted 5 lines
net/smc/smc_loopback.c new file +5 -0
+
+static u8 *lo_get_system_eid(void)
+{
+	return &LO_SYSTEM_EID.seid_string[0];
+}

SEID is for the whole system not per device.
We probably need to register a different function for each architecture.

Yes, I agree.

Quoted 10 lines

net/smc/smc_loopback.c line 191

Quoted 6 lines
net/smc/smc_loopback.c new file +6 -0
+
+static u16 lo_get_chid(struct smcd_dev *smcd)
+{
+	return 0;
+}
+

Shouldn't this return 0xFFFF in your current concept?

Yes, this should return 0xFFFF.
I supplemented a patch as attachment in this earlier reply:
https://lore.kernel.org/netdev/[email protected]/
and have amended this.

author reviewer
#4 lore raw

RFC PATCH v2 2/5 net/smc: choose loopback device in SMC-D communication

In reply to Wen Gu

This patch allows SMC-D to use loopback device.

But noted that the implementation here is quiet simple and informal. Loopback device will always be firstly chosen, and fallback happens if loopback communication is impossible.

It needs to be discussed how client indicates to peer that multiple SMC-D devices are available and how server picks a suitable one.

---
 net/smc/af_smc.c  | 55 +++++++++++++++++++++++++++++++++++++++++++++++++------
 net/smc/smc_clc.c |  4 +++-
 net/smc/smc_ism.c |  3 ++-
 3 files changed, 54 insertions(+), 8 deletions(-)
net/smc/af_smc.c +49 -6
index 9546c02..b9884c8 100644
@@ -979,6 +979,28 @@ static int smc_find_ism_device(struct smc_sock *smc, struct smc_init_info *ini)
 	return 0;
 }
 
+/* check if there is a lo device available for this connection. */
+static int smc_find_lo_device(struct smc_sock *smc, struct smc_init_info *ini)
+{
+	struct smcd_dev *sdev;
+
+	mutex_lock(&smcd_dev_list.mutex);
+	list_for_each_entry(sdev, &smcd_dev_list.list, list) {
+		if (sdev->is_loopback && !sdev->going_away &&
+		    (!ini->ism_peer_gid[0] ||
+		     !smc_ism_cantalk(ini->ism_peer_gid[0], ini->vlan_id,
+				      sdev))) {
+			ini->ism_dev[0] = sdev;
+			break;
+		}
+	}
+	mutex_unlock(&smcd_dev_list.mutex);
+	if (!ini->ism_dev[0])
+		return SMC_CLC_DECL_NOSMCDDEV;
+	ini->ism_chid[0] = smc_ism_get_chid(ini->ism_dev[0]);
+	return 0;
+}
+
 /* is chid unique for the ism devices that are already determined? */
 static bool smc_find_ism_v2_is_unique_chid(u16 chid, struct smc_init_info *ini,
 					   int cnt)
@@ -1044,10 +1066,20 @@ static int smc_find_proposal_devices(struct smc_sock *smc,
 {
 	int rc = 0;
 
-	/* check if there is an ism device available */
+	/* TODO:
+	 * How to indicate to peer if ism device and loopback
+	 * device are both available ?
+	 *
+	 * The RFC patch hasn't resolved this, just simply always
+	 * chooses loopback device first, and fallback if loopback
+	 * communication is impossible.
+	 *
+	 */
+	/* check if there is an ism or loopback device available */
 	if (!(ini->smcd_version & SMC_V1) ||
-	    smc_find_ism_device(smc, ini) ||
-	    smc_connect_ism_vlan_setup(smc, ini))
+	    (smc_find_lo_device(smc, ini) &&
+	    (smc_find_ism_device(smc, ini) ||
+	    smc_connect_ism_vlan_setup(smc, ini))))
 		ini->smcd_version &= ~SMC_V1;
 	/* else ISM V1 is supported for this connection */
 
@@ -2135,9 +2167,20 @@ static void smc_find_ism_v1_device_serv(struct smc_sock *new_smc,
 		goto not_found;
 	ini->is_smcd = true; /* prepare ISM check */
 	ini->ism_peer_gid[0] = ntohll(pclc_smcd->ism.gid);
-	rc = smc_find_ism_device(new_smc, ini);
-	if (rc)
-		goto not_found;
+
+	/* TODO:
+	 * How to know that peer has both loopback and ism device ?
+	 *
+	 * The RFC patch hasn't resolved this, simply tries loopback
+	 * device first, then ism device.
+	 */
+	/* find available loopback or ism device */
+	if (smc_find_lo_device(new_smc, ini)) {
+		rc = smc_find_ism_device(new_smc, ini);
+		if (rc)
+			goto not_found;
+	}
+
 	ini->ism_selected = 0;
 	rc = smc_listen_ism_init(new_smc, ini);
 	if (!rc)
net/smc/smc_clc.c +3 -1
index dfb9797..3887692 100644
@@ -486,7 +486,9 @@ static int smc_clc_prfx_set4_rcu(struct dst_entry *dst, __be32 ipv4,
 		return -ENODEV;
 
 	in_dev_for_each_ifa_rcu(ifa, in_dev) {
-		if (!inet_ifa_match(ipv4, ifa))
+		/* add loopback support */
+		if (inet_addr_type(dev_net(dst->dev), ipv4) != RTN_LOCAL &&
+		    !inet_ifa_match(ipv4, ifa))
 			continue;
 		prop->prefix_len = inet_mask_len(ifa->ifa_mask);
 		prop->outgoing_subnet = ifa->ifa_address & ifa->ifa_mask;
net/smc/smc_ism.c +2 -1
index 911fe08..1d10435 100644
@@ -227,7 +227,8 @@ static int smc_nl_handle_smcd_dev(struct smcd_dev *smcd,
 	if (nla_put_u8(skb, SMC_NLA_DEV_IS_CRIT, use_cnt > 0))
 		goto errattr;
 	memset(&smc_pci_dev, 0, sizeof(smc_pci_dev));
-	smc_set_pci_values(to_pci_dev(smcd->dev.parent), &smc_pci_dev);
+	if (!smcd->is_loopback)
+		smc_set_pci_values(to_pci_dev(smcd->dev.parent), &smc_pci_dev);
 	if (nla_put_u32(skb, SMC_NLA_DEV_PCI_FID, smc_pci_dev.pci_fid))
 		goto errattr;
 	if (nla_put_u16(skb, SMC_NLA_DEV_PCI_CHID, smc_pci_dev.pci_pchid))
Signature · 1 line
1.8.3.1
author reviewer
#5 lore raw

RFC PATCH v2 3/5 net/smc: add dmb attach and detach interface

In reply to Wen Gu

This patch extends smcd_ops, adding two more semantic for SMC-D device:

- attach_dmb:

Attach an already registered dmb to a specific buf_desc, so that we can refer to the dmb through this buf_desc.

- detach_dmb:

Reverse operation of attach_dmb. detach the dmb from the buf_desc.

This interface extension is to prepare for the reduction of data moving from sndbuf to RMB in SMC-D loopback device.

---
 include/net/smc.h      |  2 ++
 net/smc/smc_ism.c      | 36 ++++++++++++++++++++++++++
 net/smc/smc_ism.h      |  2 ++
 net/smc/smc_loopback.c | 69 ++++++++++++++++++++++++++++++++++++++++++++++++++
 net/smc/smc_loopback.h |  4 +++
 5 files changed, 113 insertions(+)
include/net/smc.h +2 -0
index 7699f97..60a96f7 100644
@@ -63,6 +63,8 @@ struct smcd_ops {
 				u32 vid);
 	int (*register_dmb)(struct smcd_dev *dev, struct smcd_dmb *dmb);
 	int (*unregister_dmb)(struct smcd_dev *dev, struct smcd_dmb *dmb);
+	int (*attach_dmb)(struct smcd_dev *dev, struct smcd_dmb *dmb);
+	int (*detach_dmb)(struct smcd_dev *dev, u64 token);
 	int (*add_vlan_id)(struct smcd_dev *dev, u64 vlan_id);
 	int (*del_vlan_id)(struct smcd_dev *dev, u64 vlan_id);
 	int (*set_vlan_required)(struct smcd_dev *dev);
net/smc/smc_ism.c +36 -0
index 1d10435..2049388 100644
@@ -202,6 +202,42 @@ int smc_ism_register_dmb(struct smc_link_group *lgr, int dmb_len,
 	return rc;
 }
 
+int smc_ism_attach_dmb(struct smcd_dev *dev, u64 token,
+		       struct smc_buf_desc *dmb_desc)
+{
+	struct smcd_dmb dmb;
+	int rc = 0;
+
+	memset(&dmb, 0, sizeof(dmb));
+	dmb.dmb_tok = token;
+
+	/* only support loopback device now */
+	if (!dev->is_loopback)
+		return -EINVAL;
+	if (!dev->ops->attach_dmb)
+		return -EINVAL;
+
+	rc = dev->ops->attach_dmb(dev, &dmb);
+	if (!rc) {
+		dmb_desc->sba_idx = dmb.sba_idx;
+		dmb_desc->token = dmb.dmb_tok;
+		dmb_desc->cpu_addr = dmb.cpu_addr;
+		dmb_desc->dma_addr = dmb.dma_addr;
+		dmb_desc->len = dmb.dmb_len;
+	}
+	return rc;
+}
+
+int smc_ism_detach_dmb(struct smcd_dev *dev, u64 token)
+{
+	if (!dev->is_loopback)
+		return -EINVAL;
+	if (!dev->ops->detach_dmb)
+		return -EINVAL;
+
+	return dev->ops->detach_dmb(dev, token);
+}
+
 static int smc_nl_handle_smcd_dev(struct smcd_dev *smcd,
 				  struct sk_buff *skb,
 				  struct netlink_callback *cb)
net/smc/smc_ism.h +2 -0
index d6b2db6..9022979 100644
@@ -38,6 +38,8 @@ struct smc_ism_vlanid {			/* VLAN id set on ISM device */
 int smc_ism_register_dmb(struct smc_link_group *lgr, int buf_size,
 			 struct smc_buf_desc *dmb_desc);
 int smc_ism_unregister_dmb(struct smcd_dev *dev, struct smc_buf_desc *dmb_desc);
+int smc_ism_attach_dmb(struct smcd_dev *dev, u64 token, struct smc_buf_desc *dmb_desc);
+int smc_ism_detach_dmb(struct smcd_dev *dev, u64 token);
 int smc_ism_signal_shutdown(struct smc_link_group *lgr);
 void smc_ism_get_system_eid(u8 **eid);
 u16 smc_ism_get_chid(struct smcd_dev *dev);
net/smc/smc_loopback.c +69 -0

108 lines not on this page · raw message

net/smc/smc_loopback.h +4 -0
index d7f7815..f4122be 100644
@@ -32,6 +32,7 @@ struct lo_dmb_node {
 	u32 sba_idx;
 	void *cpu_addr;
 	dma_addr_t dma_addr;
+	refcount_t refcnt;
 };
 
 struct lo_dev {
@@ -41,6 +42,9 @@ struct lo_dev {
 	DECLARE_BITMAP(sba_idx_mask, LODEV_MAX_DMBS);
 	rwlock_t dmb_ht_lock;
 	DECLARE_HASHTABLE(dmb_ht, LODEV_MAX_DMBS_BUCKETS);
+	atomic_t dmb_cnt;
+	wait_queue_head_t dmbs_release;
+	wait_queue_head_t ldev_release;
 };
 
 struct lo_systemeid {
Signature · 1 line
1.8.3.1
author reviewer
#6 lore raw

RFC PATCH v2 4/5 net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback

In reply to Wen Gu

This patch aims to improve SMC-D loopback performance by avoiding data copy from local sndbuf to peer RMB. The main idea is to let local sndbuf and peer RMB share the same physical memory.

 +----------+                     +----------+
 | socket A |                     | socket B |
 +----------+                     +----------+
       |                               ^
       |         +---------+           |
  regard as      |         | ----------|
  local sndbuf   |  B's    |     regard as
       |         |  RMB    |     local RMB
       |-------> |         |
                 +---------+

For connections using smcd loopback device:

1. Only create and maintain local RMB.

        a. Create or reuse RMB when create connection;
        b. Free RMB when lgr free;

2. Attach local sndbuf to peer RMB.
a. sndbuf_desc describes the same memory region as peer rmb_desc. b. sndbuf_desc is exclusive to specific connection and won't be
added to lgr buffer pool for reuse.
c. sndbuf is attached to peer RMB when receive remote token after
CLC accept/confirm message.
d. sndbuf is detached from peer RMB when connection is freed.

Therefore, the data copied from the userspace to local sndbuf directly reaches the peer RMB.

---
 net/smc/af_smc.c   | 23 +++++++++++++++++++-
 net/smc/smc_core.c | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 net/smc/smc_core.h |  2 ++
 3 files changed, 86 insertions(+), 1 deletion(-)
net/smc/af_smc.c +22 -1
index b9884c8..c7de566 100644
@@ -1073,7 +1073,6 @@ static int smc_find_proposal_devices(struct smc_sock *smc,
 	 * The RFC patch hasn't resolved this, just simply always
 	 * chooses loopback device first, and fallback if loopback
 	 * communication is impossible.
-	 *
 	 */
 	/* check if there is an ism or loopback device available */
 	if (!(ini->smcd_version & SMC_V1) ||
@@ -1397,6 +1396,17 @@ static int smc_connect_ism(struct smc_sock *smc,
 	}
 
 	smc_conn_save_peer_info(smc, aclc);
+
+	/* special for smcd loopback
+	 * conns above smcd loopback dev only create their rmbs.
+	 * their sndbufs are 'maps' of peer rmbs.
+	 */
+	if (smc->conn.lgr->smcd->is_loopback) {
+		rc = smcd_buf_attach(&smc->conn);
+		if (rc)
+			goto connect_abort;
+		smc->sk.sk_sndbuf = 2 * (smc->conn.sndbuf_desc->len);
+	}
 	smc_close_init(smc);
 	smc_rx_init(smc);
 	smc_tx_init(smc);
@@ -2464,6 +2474,17 @@ static void smc_listen_work(struct work_struct *work)
 		mutex_unlock(&smc_server_lgr_pending);
 	}
 	smc_conn_save_peer_info(new_smc, cclc);
+
+	/* special for smcd loopback
+	 * conns above smcd loopback dev only create their rmbs.
+	 * their sndbufs are 'maps' of peer rmbs.
+	 */
+	if (ini->is_smcd && new_smc->conn.lgr->smcd->is_loopback) {
+		rc = smcd_buf_attach(&new_smc->conn);
+		if (rc)
+			goto out_decl;
+		new_smc->sk.sk_sndbuf = 2 * (new_smc->conn.sndbuf_desc->len);
+	}
 	smc_listen_out_connected(new_smc);
 	SMC_STAT_SERV_SUCC_INC(sock_net(newclcsock->sk), ini);
 	goto out_free;
net/smc/smc_core.c +62 -0

80 lines not on this page · raw message

net/smc/smc_core.h +2 -0
index 285f9bd..b51b020 100644
@@ -518,6 +518,8 @@ void smc_smcd_terminate(struct smcd_dev *dev, u64 peer_gid,
 void smc_smcd_terminate_all(struct smcd_dev *dev);
 void smc_smcr_terminate_all(struct smc_ib_device *smcibdev);
 int smc_buf_create(struct smc_sock *smc, bool is_smcd);
+int smcd_buf_attach(struct smc_connection *conn);
+void smcd_buf_detach(struct smc_connection *conn);
 int smc_uncompress_bufsize(u8 compressed);
 int smc_rmb_rtoken_handling(struct smc_connection *conn, struct smc_link *link,
 			    struct smc_clc_msg_accept_confirm *clc);
Signature · 1 line
1.8.3.1
author reviewer
#7 lore raw

RFC PATCH v2 5/5 net/smc: logic of cursors update in SMC-D loopback connections

In reply to Wen Gu

Since local sndbuf of SMC-D loopback connection shares the same physical memory region with peer RMB, the logic of cursors update needs to be adapted.

The main difference from original implementation is need to ensure that the data copied to local sndbuf won't overwrite the unconsumed data of peer.

So, for SMC-D loopback connections:

1. TX
a. don't update fin_curs when send out cdc msg. b. fin_curs and sndbuf_space update will be deferred until
receiving peer cons_curs update.

2. RX
a. same as before. peer sndbuf is as large as local rmb,
which guarantees that prod_curs will behind prep_curs.

---
 net/smc/smc_cdc.c      | 53 +++++++++++++++++++++++++++++++++++++++-----------
 net/smc/smc_loopback.c |  7 +++++++
 2 files changed, 49 insertions(+), 11 deletions(-)
net/smc/smc_cdc.c +42 -11
index 61f5ff7..586472a 100644
@@ -253,17 +253,26 @@ int smcd_cdc_msg_send(struct smc_connection *conn)
 		return rc;
 	smc_curs_copy(&conn->rx_curs_confirmed, &curs, conn);
 	conn->local_rx_ctrl.prod_flags.cons_curs_upd_req = 0;
-	/* Calculate transmitted data and increment free send buffer space */
-	diff = smc_curs_diff(conn->sndbuf_desc->len, &conn->tx_curs_fin,
-			     &conn->tx_curs_sent);
-	/* increased by confirmed number of bytes */
-	smp_mb__before_atomic();
-	atomic_add(diff, &conn->sndbuf_space);
-	/* guarantee 0 <= sndbuf_space <= sndbuf_desc->len */
-	smp_mb__after_atomic();
-	smc_curs_copy(&conn->tx_curs_fin, &conn->tx_curs_sent, conn);
+	if (!conn->lgr->smcd->is_loopback) {
+		/* Note:
+		 * For smcd loopback device:
+		 *
+		 * Don't update the fin_curs and sndbuf_space here.
+		 * Update fin_curs when peer consumes the data in RMB.
+		 */
 
-	smc_tx_sndbuf_nonfull(smc);
+		/* Calculate transmitted data and increment free send buffer space */
+		diff = smc_curs_diff(conn->sndbuf_desc->len, &conn->tx_curs_fin,
+				     &conn->tx_curs_sent);
+		/* increased by confirmed number of bytes */
+		smp_mb__before_atomic();
+		atomic_add(diff, &conn->sndbuf_space);
+		/* guarantee 0 <= sndbuf_space <= sndbuf_desc->len */
+		smp_mb__after_atomic();
+		smc_curs_copy(&conn->tx_curs_fin, &conn->tx_curs_sent, conn);
+
+		smc_tx_sndbuf_nonfull(smc);
+	}
 	return rc;
 }
 
@@ -321,7 +330,7 @@ static void smc_cdc_msg_recv_action(struct smc_sock *smc,
 {
 	union smc_host_cursor cons_old, prod_old;
 	struct smc_connection *conn = &smc->conn;
-	int diff_cons, diff_prod;
+	int diff_cons, diff_prod, diff_tx;
 
 	smc_curs_copy(&prod_old, &conn->local_rx_ctrl.prod, conn);
 	smc_curs_copy(&cons_old, &conn->local_rx_ctrl.cons, conn);
@@ -337,6 +346,28 @@ static void smc_cdc_msg_recv_action(struct smc_sock *smc,
 		atomic_add(diff_cons, &conn->peer_rmbe_space);
 		/* guarantee 0 <= peer_rmbe_space <= peer_rmbe_size */
 		smp_mb__after_atomic();
+
+		/* For smcd loopback device:
+		 * Update of peer cons_curs indicates that
+		 * 1. peer rmbe space increases.
+		 * 2. local sndbuf space increases.
+		 *
+		 * So local sndbuf fin_curs should be equal to peer RMB cons_curs.
+		 */
+		if (conn->lgr->is_smcd &&
+		    conn->lgr->smcd->is_loopback) {
+			/* calculate peer rmb consumed data */
+			diff_tx = smc_curs_diff(conn->sndbuf_desc->len, &conn->tx_curs_fin,
+						&conn->local_rx_ctrl.cons);
+			/* increase local sndbuf space and fin_curs */
+			smp_mb__before_atomic();
+			atomic_add(diff_tx, &conn->sndbuf_space);
+			/* guarantee 0 <= sndbuf_space <= sndbuf_desc->len */
+			smp_mb__after_atomic();
+			smc_curs_copy(&conn->tx_curs_fin, &conn->local_rx_ctrl.cons, conn);
+
+			smc_tx_sndbuf_nonfull(smc);
+		}
 	}
 
 	diff_prod = smc_curs_diff(conn->rmb_desc->len, &prod_old,
net/smc/smc_loopback.c +7 -0
index bc3ff82..43f0287 100644
@@ -216,6 +216,13 @@ static int lo_move_data(struct smcd_dev *smcd, u64 dmb_tok, unsigned int idx,
 	struct lo_dmb_node *rmb_node = NULL, *tmp_node;
 	struct lo_dev *ldev = smcd->priv;
 
+	if (!sf) {
+		/* no need to move data.
+		 * sndbuf is equal to peer rmb.
+		 */
+		return 0;
+	}
+
 	read_lock(&ldev->dmb_ht_lock);
 	hash_for_each_possible(ldev->dmb_ht, tmp_node, list, dmb_tok) {
 		if (tmp_node->token == dmb_tok) {
Signature · 1 line
1.8.3.1
#8 lore raw
In reply to Wen Gu
Quoted 44 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

 +------------------------------------------+
 |  +-----------+           +-----------+   |
 |  | process A |           | process B |   |
 |  +-----------+           +-----------+   |
 |       ^                        ^         |
 |       |    +---------------+   |         |
 |       |    |   SMC stack   |   |         |
 |       +--->| +-----------+ |<--|         |
 |            | |   dummy   | |             |
 |            | |   device  | |             |
 |            +-+-----------+-+             |
 |                   VM                     |
 +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

Quoted 10 lines
 +----------+                     +----------+
 | socket A |                     | socket B |
 +----------+                     +----------+
       |                               ^
       |         +---------+           |
  regard as      |         | ----------|
  local sndbuf   |  B's    |     regard as
       |         |  RMB    |     local RMB
       |-------> |         |
                 +---------+

Hi Wen Gu,

I maintain the s390 specific PCI support in Linux and would like to provide a bit of background on this. You're surely wondering why we even have a copy in there for our ISM virtual PCI device. To understand why this copy operation exists and why we need to keep it working, one needs a bit of s390 aka mainframe background.

On s390 all (currently supported) native machines have a mandatory machine level hypervisor. All OSs whether z/OS or Linux run either on this machine level hypervisor as so called Logical Partitions (LPARs) or as second/third/… level guests on e.g. a KVM or z/VM hypervisor that in turn runs in an LPAR. Now, in terms of memory this machine level hypervisor sometimes called PR/SM unlike KVM, z/VM, or VMWare is a partitioning hypervisor without paging. This is one of the main reasons for the very-near-native performance of the machine hypervisor as the memory of its guests acts just like native RAM on other systems. It is never paged out and always accessible to IOMMU translated DMA from devices without the need for pinning pages and besides a trivial offset/limit adjustment an LPAR's MMU does the same amount of work as an MMU on a bare metal x86_64/ARM64 box.

It also means however that when SMC-D is used to communicate between LPARs via an ISM device there is no way of mapping the DMBs to the same physical memory as there exists no MMU-like layer spanning partitions that could do such a mapping. Meanwhile for machine level firmware including the ISM virtual PCI device it is still possible to _copy_ memory between different memory partitions. So yeah while I do see the appeal of skipping the memcpy() for loopback or even between guests of a paging hypervisor such as KVM, which can map the DMBs on the same physical memory, we must keep in mind this original use case requiring a copy operation.

Thanks,
Niklas

Quoted 13 lines

# Benchmark Test

* Test environments:
- VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
- SMC sndbuf/RMB size 1MB.

* Test object:
- TCP: run on TCP loopback.
- domain: run on UNIX domain.
- SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
- SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

Quoted 7 lines

1. ipc-benchmark (see [3])

- ./<foo> -c 1000000 -s 100

TCP domain SMC-lo SMC-lo-nocpy
Message
rate (msg/s) 75140 129548(+72.41) 152266(+102.64%) 151914(+102.17%)

Interesting that it does beat UNIX domain sockets. Also, see my below comment for nginx/wrk as this seems very similar.

Quoted 18 lines

2. sockperf

- serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
- clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

TCP SMC-lo SMC-lo-nocpy
Bandwidth(MBps) 4943.359 4936.096(-0.15%) 8239.624(+66.68%)
Latency(us) 6.372 3.359(-47.28%) 3.25(-49.00%)

3. iperf3

- serv: <smc_run> taskset -c <cpu> iperf3 -s
- clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

TCP SMC-lo SMC-lo-nocpy
Bitrate(Gb/s) 40.5 41.4(+2.22%) 76.4(+88.64%)

Quoted 7 lines

4. nginx/wrk

- serv: <smc_run> nginx
- clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

TCP SMC-lo SMC-lo-nocpy
Requests/s 154643.22 220894.03(+42.84%) 226754.3(+46.63%)

This result is very interesting indeed. So with the much more realistic nginx/wrk workload it seems to copy hurts much less than the iperf3/sockperf would suggest while SMC-D itself seems to help more. I'd hope that this translates to actual applications as well. Maybe this makes SMC-D based loopback interesting even while keeping the copy, at least until we can come up with a sane way to work a no-copy variant into SMC-D?

Quoted 6 lines

# Discussion

1. API between SMC-D and ISM device

Quoted 8 lines

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

Due to the reasons explained above this behavior can't be emulated by ISM devices at least not when crossing partitions. Not sure if we can still incorporate it in the API and allow for both copying and remapping SMC-D like devices, it definitely needs careful consideration and I think also a better understanding of the benefit for real world workloads.

Quoted 6 lines

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1]) can provide some examples for new API design. And we are very glad to discuss this on the mail list.

Quoted 6 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

Agree on this part, though it is for the SMC maintainers to decide, I think we would definitely want to be able to use any upcoming inter-VM devices on s390 possibly also in conjunction with ISM devices for communication across partitions.

Quoted 36 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
1. Fix some build WARNINGs complained by kernel test rebot

2. Add iperf3 test data.

Wen Gu (5):
net/smc: introduce SMC-D loopback device
net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

 include/net/smc.h      |   3 +
 net/smc/Makefile       |   2 +-
 net/smc/af_smc.c       |  88 +++++++++++-
 net/smc/smc_cdc.c      |  59 ++++++--
 net/smc/smc_cdc.h      |   1 +
 net/smc/smc_clc.c      |   4 +-
 net/smc/smc_core.c     |  62 +++++++++
 net/smc/smc_core.h     |   2 +
 net/smc/smc_ism.c      |  39 +++++-
 net/smc/smc_ism.h      |   2 +
 net/smc/smc_loopback.c | 358 +++++++++++++++++++++++++++++++++++++++++++++++++
 net/smc/smc_loopback.h |  63 +++++++++
 12 files changed, 662 insertions(+), 21 deletions(-)
 create mode 100644 net/smc/smc_loopback.c
 create mode 100644 net/smc/smc_loopback.h
author reviewer
#9 lore raw
Quoted 89 lines · Niklas Schnelle
Quoted 44 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

  +------------------------------------------+
  |  +-----------+           +-----------+   |
  |  | process A |           | process B |   |
  |  +-----------+           +-----------+   |
  |       ^                        ^         |
  |       |    +---------------+   |         |
  |       |    |   SMC stack   |   |         |
  |       +--->| +-----------+ |<--|         |
  |            | |   dummy   | |             |
  |            | |   device  | |             |
  |            +-+-----------+-+             |
  |                   VM                     |
  +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

Quoted 10 lines
  +----------+                     +----------+
  | socket A |                     | socket B |
  +----------+                     +----------+
        |                               ^
        |         +---------+           |
   regard as      |         | ----------|
   local sndbuf   |  B's    |     regard as
        |         |  RMB    |     local RMB
        |-------> |         |
                  +---------+

Hi Wen Gu,

I maintain the s390 specific PCI support in Linux and would like to provide a bit of background on this. You're surely wondering why we even have a copy in there for our ISM virtual PCI device. To understand why this copy operation exists and why we need to keep it working, one needs a bit of s390 aka mainframe background.

On s390 all (currently supported) native machines have a mandatory machine level hypervisor. All OSs whether z/OS or Linux run either on this machine level hypervisor as so called Logical Partitions (LPARs) or as second/third/… level guests on e.g. a KVM or z/VM hypervisor that in turn runs in an LPAR. Now, in terms of memory this machine level hypervisor sometimes called PR/SM unlike KVM, z/VM, or VMWare is a partitioning hypervisor without paging. This is one of the main reasons for the very-near-native performance of the machine hypervisor as the memory of its guests acts just like native RAM on other systems. It is never paged out and always accessible to IOMMU translated DMA from devices without the need for pinning pages and besides a trivial offset/limit adjustment an LPAR's MMU does the same amount of work as an MMU on a bare metal x86_64/ARM64 box.

It also means however that when SMC-D is used to communicate between LPARs via an ISM device there is no way of mapping the DMBs to the same physical memory as there exists no MMU-like layer spanning partitions that could do such a mapping. Meanwhile for machine level firmware including the ISM virtual PCI device it is still possible to _copy_ memory between different memory partitions. So yeah while I do see the appeal of skipping the memcpy() for loopback or even between guests of a paging hypervisor such as KVM, which can map the DMBs on the same physical memory, we must keep in mind this original use case requiring a copy operation.

Quoted 3 lines

Thanks,
Niklas

Hi Niklas,

Thank you so much for the complete and detailed explanation! This provides me a brand new perspective of s390 device that we hadn't dabbled in before. Now I understand why shared memory is unavailable between different LPARs.

Our original intention of proposing loopback device and the incoming device (virtio-ism) for inter-VM is to use SMC-D to accelerate communication in the case with no existing s390 ISM devices. In our conception, s390 ISM device, loopback device and virtio-ism device are parallel and are abstracted by smcd_ops.

 +------------------------+
 |          SMC-D         |
 +------------------------+
 -------- smcd_ops ---------
 +------+ +------+ +------+
 | s390 | | loop | |virtio|
 | ISM  | | back | | -ism |
 | dev  | | dev  | | dev  |
 +------+ +------+ +------+

We also believe that keeping the existing design and behavior of s390 ISM device is unshaken. What we want to get support for is some smcd_ops extension for devices with optional beneficial capability, such as nocopy here (Let's call it this for now), which is really helpful for us in inter-process and inter-VM scenario.

And coincided with IBM's intention to add APIs between SMC-D and devices to support various devices for SMC-D, as mentioned in [2], we send out this RFC and the incoming virio-ism RFC, to provide some examples.

Quoted 51 lines
Quoted 13 lines

# Benchmark Test

* Test environments:
- VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
- SMC sndbuf/RMB size 1MB.

* Test object:
- TCP: run on TCP loopback.
- domain: run on UNIX domain.
- SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
- SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

Quoted 7 lines

1. ipc-benchmark (see [3])

- ./<foo> -c 1000000 -s 100

TCP domain SMC-lo SMC-lo-nocpy
Message
rate (msg/s) 75140 129548(+72.41) 152266(+102.64%) 151914(+102.17%)

Interesting that it does beat UNIX domain sockets. Also, see my below comment for nginx/wrk as this seems very similar.

Quoted 18 lines

2. sockperf

- serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
- clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

TCP SMC-lo SMC-lo-nocpy
Bandwidth(MBps) 4943.359 4936.096(-0.15%) 8239.624(+66.68%)
Latency(us) 6.372 3.359(-47.28%) 3.25(-49.00%)

3. iperf3

- serv: <smc_run> taskset -c <cpu> iperf3 -s
- clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

TCP SMC-lo SMC-lo-nocpy
Bitrate(Gb/s) 40.5 41.4(+2.22%) 76.4(+88.64%)

Quoted 7 lines

4. nginx/wrk

- serv: <smc_run> nginx
- clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

TCP SMC-lo SMC-lo-nocpy
Requests/s 154643.22 220894.03(+42.84%) 226754.3(+46.63%)

Quoted 8 lines

This result is very interesting indeed. So with the much more realistic nginx/wrk workload it seems to copy hurts much less than the iperf3/sockperf would suggest while SMC-D itself seems to help more. I'd hope that this translates to actual applications as well. Maybe this makes SMC-D based loopback interesting even while keeping the copy, at least until we can come up with a sane way to work a no-copy variant into SMC-D?

I agree, nginx/wrk workload is much more realistic for many applications.

But we also encounter many other cases similar to sockperf on the cloud, which requires high throughput, such as AI training and big data.

So avoidance of copying between DMBs can help these cases a lot :)

Quoted 6 lines
Quoted 6 lines

# Discussion

1. API between SMC-D and ISM device

Quoted 8 lines
Quoted 8 lines

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Quoted 8 lines

Due to the reasons explained above this behavior can't be emulated by ISM devices at least not when crossing partitions. Not sure if we can still incorporate it in the API and allow for both copying and remapping SMC-D like devices, it definitely needs careful consideration and I think also a better understanding of the benefit for real world workloads.

Here I am not rigorous.

Nocopy shouldn't be a standard ISM device behavior indeed. Actually we hope it be a standard optional _SMC-D_ device behavior and defined by smcd_ops.

For devices don't support these options, like ISM device on s390 architecture, .attach_dmb/.detach_dmb and other reasonable extensions (which will be proposed to discuss in incoming virtio-ism RFC) can be set to NULL or return invalid. And for devices do support, they may be used for improving performance in some cases.

In addition, can I know more latest news about the API design? :) , like its scale, will it be a almost refactor of existing interface or incremental patching? and its object, will it be tailored for exact ISM behavior or to reserve some options for other devices, like nocopy here? From my understanding of [2], it might be the latter?

Quoted 13 lines
Quoted 6 lines

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1]) can provide some examples for new API design. And we are very glad to discuss this on the mail list.

Quoted 6 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

Quoted 5 lines

Agree on this part, though it is for the SMC maintainers to decide, I think we would definitely want to be able to use any upcoming inter-VM devices on s390 possibly also in conjunction with ISM devices for communication across partitions.

Yes, this part needs to be discussed with SMC maintainers. And thank you, we are very glad if our devices can be applied on s390 through the efforts.

Best Regards,
Wen Gu

Quoted 36 lines
Quoted 36 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
1. Fix some build WARNINGs complained by kernel test rebot

2. Add iperf3 test data.

Wen Gu (5):
net/smc: introduce SMC-D loopback device
net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

  include/net/smc.h      |   3 +
  net/smc/Makefile       |   2 +-
  net/smc/af_smc.c       |  88 +++++++++++-
  net/smc/smc_cdc.c      |  59 ++++++--
  net/smc/smc_cdc.h      |   1 +
  net/smc/smc_clc.c      |   4 +-
  net/smc/smc_core.c     |  62 +++++++++
  net/smc/smc_core.h     |   2 +
  net/smc/smc_ism.c      |  39 +++++-
  net/smc/smc_ism.h      |   2 +
  net/smc/smc_loopback.c | 358 +++++++++++++++++++++++++++++++++++++++++++++++++
  net/smc/smc_loopback.h |  63 +++++++++
  12 files changed, 662 insertions(+), 21 deletions(-)

create mode 100644 net/smc/smc_loopback.c
create mode 100644 net/smc/smc_loopback.h

#10 lore raw
Quoted 253 lines · Wen Gu
Quoted 89 lines · Niklas Schnelle
Quoted 44 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

  +------------------------------------------+
  |  +-----------+           +-----------+   |
  |  | process A |           | process B |   |
  |  +-----------+           +-----------+   |
  |       ^                        ^         |
  |       |    +---------------+   |         |
  |       |    |   SMC stack   |   |         |
  |       +--->| +-----------+ |<--|         |
  |            | |   dummy   | |             |
  |            | |   device  | |             |
  |            +-+-----------+-+             |
  |                   VM                     |
  +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

Quoted 10 lines
  +----------+                     +----------+
  | socket A |                     | socket B |
  +----------+                     +----------+
        |                               ^
        |         +---------+           |
   regard as      |         | ----------|
   local sndbuf   |  B's    |     regard as
        |         |  RMB    |     local RMB
        |-------> |         |
                  +---------+

Hi Wen Gu,

I maintain the s390 specific PCI support in Linux and would like to provide a bit of background on this. You're surely wondering why we even have a copy in there for our ISM virtual PCI device. To understand why this copy operation exists and why we need to keep it working, one needs a bit of s390 aka mainframe background.

On s390 all (currently supported) native machines have a mandatory machine level hypervisor. All OSs whether z/OS or Linux run either on this machine level hypervisor as so called Logical Partitions (LPARs) or as second/third/… level guests on e.g. a KVM or z/VM hypervisor that in turn runs in an LPAR. Now, in terms of memory this machine level hypervisor sometimes called PR/SM unlike KVM, z/VM, or VMWare is a partitioning hypervisor without paging. This is one of the main reasons for the very-near-native performance of the machine hypervisor as the memory of its guests acts just like native RAM on other systems. It is never paged out and always accessible to IOMMU translated DMA from devices without the need for pinning pages and besides a trivial offset/limit adjustment an LPAR's MMU does the same amount of work as an MMU on a bare metal x86_64/ARM64 box.

It also means however that when SMC-D is used to communicate between LPARs via an ISM device there is  no way of mapping the DMBs to the same physical memory as there exists no MMU-like layer spanning partitions that could do such a mapping. Meanwhile for machine level firmware including the ISM virtual PCI device it is still possible to _copy_ memory between different memory partitions. So yeah while I do see the appeal of skipping the memcpy() for loopback or even between guests of a paging hypervisor such as KVM, which can map the DMBs on the same physical memory, we must keep in mind this original use case requiring a copy operation.

Quoted 3 lines

Thanks,
Niklas

Hi Niklas,

Thank you so much for the complete and detailed explanation! This provides me a brand new perspective of s390 device that we hadn't dabbled in before. Now I understand why shared memory is unavailable between different LPARs.

Our original intention of proposing loopback device and the incoming device (virtio-ism) for inter-VM is to use SMC-D to accelerate communication in the case with no existing s390 ISM devices. In our conception, s390 ISM device, loopback device and virtio-ism device are parallel and are abstracted by smcd_ops.

 +------------------------+
 |          SMC-D         |
 +------------------------+
 -------- smcd_ops ---------
 +------+ +------+ +------+
 | s390 | | loop | |virtio|
 | ISM  | | back | | -ism |
 | dev  | | dev  | | dev  |
 +------+ +------+ +------+

We also believe that keeping the existing design and behavior of s390 ISM device is unshaken. What we want to get support for is some smcd_ops extension for devices with optional beneficial capability, such as nocopy here (Let's call it this for now), which is really helpful for us in inter-process and inter-VM scenario.

And coincided with IBM's intention to add APIs between SMC-D and devices to support various devices for SMC-D, as mentioned in [2], we send out this RFC and the incoming virio-ism RFC, to provide some examples.

Quoted 51 lines
Quoted 13 lines

# Benchmark Test

  * Test environments:
       - VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
       - SMC sndbuf/RMB size 1MB.

  * Test object:
       - TCP: run on TCP loopback.
       - domain: run on UNIX domain.
       - SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
       - SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

Quoted 7 lines

1. ipc-benchmark (see [3])

  - ./<foo> -c 1000000 -s 100

                        TCP              domain              SMC-lo             SMC-lo-nocpy Message
rate (msg/s)         75140      129548(+72.41)    152266(+102.64%)         151914(+102.17%)

Interesting that it does beat UNIX domain sockets. Also, see my below comment for nginx/wrk as this seems very similar.

Quoted 18 lines

2. sockperf

  - serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
  - clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

                        TCP                  SMC-lo             SMC-lo-nocpy Bandwidth(MBps)   4943.359        4936.096(-0.15%)        8239.624(+66.68%) Latency(us)          6.372          3.359(-47.28%)            3.25(-49.00%)

3. iperf3

  - serv: <smc_run> taskset -c <cpu> iperf3 -s
  - clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

                        TCP                  SMC-lo             SMC-lo-nocpy Bitrate(Gb/s)         40.5            41.4(+2.22%)            76.4(+88.64%)

Quoted 7 lines

4. nginx/wrk

  - serv: <smc_run> nginx
  - clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

                        TCP                  SMC-lo             SMC-lo-nocpy Requests/s       154643.22      220894.03(+42.84%)        226754.3(+46.63%)

Quoted 8 lines

This result is very interesting indeed. So with the much more realistic nginx/wrk workload it seems to copy hurts much less than the iperf3/sockperf would suggest while SMC-D itself seems to help more. I'd hope that this translates to actual applications as well. Maybe this makes SMC-D based loopback interesting even while keeping the copy, at least until we can come up with a sane way to work a no-copy variant into SMC-D?

I agree, nginx/wrk workload is much more realistic for many applications.

But we also encounter many other cases similar to sockperf on the cloud, which requires high throughput, such as AI training and big data.

So avoidance of copying between DMBs can help these cases a lot :)

Quoted 6 lines
Quoted 6 lines

# Discussion

1. API between SMC-D and ISM device

Quoted 8 lines
Quoted 8 lines

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Quoted 8 lines

Due to the reasons explained above this behavior can't be emulated by ISM devices at least not when crossing partitions. Not sure if we can still incorporate it in the API and allow for both copying and remapping SMC-D like devices, it definitely needs careful consideration and I think also a better understanding of the benefit for real world workloads.

Here I am not rigorous.

Nocopy shouldn't be a standard ISM device behavior indeed. Actually we hope it be a standard optional _SMC-D_ device behavior and defined by smcd_ops.

For devices don't support these options, like ISM device on s390 architecture, .attach_dmb/.detach_dmb and other reasonable extensions (which will be proposed to discuss in incoming virtio-ism RFC) can be set to NULL or return invalid. And for devices do support, they may be used for improving performance in some cases.

In addition, can I know more latest news about the API design? :) , like its scale, will it be a almost refactor of existing interface or incremental patching? and its object, will it be tailored for exact ISM behavior or to reserve some options for other devices, like nocopy here? From my understanding of [2], it might be the latter?

Quoted 13 lines
Quoted 6 lines

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1]) can provide some examples for new API design. And we are very glad to discuss this on the mail list.

Quoted 6 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

Quoted 5 lines

Agree on this part, though it is for the SMC maintainers to decide, I think we would definitely want to be able to use any upcoming inter-VM devices on s390 possibly also in conjunction with ISM devices for communication across partitions.

Quoted 10 lines

Yes, this part needs to be discussed with SMC maintainers. And thank you, we are very glad if our devices can be applied on s390 through the efforts.

Best Regards,
Wen Gu

Quoted 3 lines
Quoted 3 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

As described in
SMC protocol (including SMC-D): https://www.ibm.com/support/pages/system/files/inline-files/IBM%20Shared%20Memory%20Communications%20Version%202_2.pdf
the CLC messages provide a list of up to 8 ISM devices to chose from.
So I would hope that we can use the existing protocol.

The challenge will be to define GID (Global Interface ID) and CHID (a fabric ID) in a meaningful way for the new devices.
There is always smcd_ops->query_remote_gid() as a safety net. But the idea is that a CHID mismatch is a fast way to tell that these 2 interfaces do match.

Quoted 33 lines
Quoted 33 lines
Quoted 33 lines

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
  1. Fix some build WARNINGs complained by kernel test rebot

  2. Add iperf3 test data.

Wen Gu (5):
   net/smc: introduce SMC-D loopback device
   net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
   net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

  include/net/smc.h      |   3 +
  net/smc/Makefile       |   2 +-
  net/smc/af_smc.c       |  88 +++++++++++-

  net/smc/smc_cdc.c      |  59 ++++++--
  net/smc/smc_cdc.h      |   1 +
  net/smc/smc_clc.c      |   4 +-

  net/smc/smc_core.c     |  62 +++++++++
  net/smc/smc_core.h     |   2 +
  net/smc/smc_ism.c      |  39 +++++-
  net/smc/smc_ism.h      |   2 +
  net/smc/smc_loopback.c | 358 +++++++++++++++++++++++++++++++++++++++++++++++++ net/smc/smc_loopback.h |  63 +++++++++

  12 files changed, 662 insertions(+), 21 deletions(-)

  create mode 100644 net/smc/smc_loopback.c
  create mode 100644 net/smc/smc_loopback.h

author reviewer
#11 lore raw
Quoted 272 lines · Alexandra Winter
Quoted 253 lines · Wen Gu
Quoted 89 lines · Niklas Schnelle
Quoted 44 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

   +------------------------------------------+
   |  +-----------+           +-----------+   |
   |  | process A |           | process B |   |
   |  +-----------+           +-----------+   |
   |       ^                        ^         |
   |       |    +---------------+   |         |
   |       |    |   SMC stack   |   |         |
   |       +--->| +-----------+ |<--|         |
   |            | |   dummy   | |             |
   |            | |   device  | |             |
   |            +-+-----------+-+             |
   |                   VM                     |
   +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

Quoted 10 lines
   +----------+                     +----------+
   | socket A |                     | socket B |
   +----------+                     +----------+
         |                               ^
         |         +---------+           |
    regard as      |         | ----------|
    local sndbuf   |  B's    |     regard as
         |         |  RMB    |     local RMB
         |-------> |         |
                   +---------+

Hi Wen Gu,

I maintain the s390 specific PCI support in Linux and would like to provide a bit of background on this. You're surely wondering why we even have a copy in there for our ISM virtual PCI device. To understand why this copy operation exists and why we need to keep it working, one needs a bit of s390 aka mainframe background.

On s390 all (currently supported) native machines have a mandatory machine level hypervisor. All OSs whether z/OS or Linux run either on this machine level hypervisor as so called Logical Partitions (LPARs) or as second/third/… level guests on e.g. a KVM or z/VM hypervisor that in turn runs in an LPAR. Now, in terms of memory this machine level hypervisor sometimes called PR/SM unlike KVM, z/VM, or VMWare is a partitioning hypervisor without paging. This is one of the main reasons for the very-near-native performance of the machine hypervisor as the memory of its guests acts just like native RAM on other systems. It is never paged out and always accessible to IOMMU translated DMA from devices without the need for pinning pages and besides a trivial offset/limit adjustment an LPAR's MMU does the same amount of work as an MMU on a bare metal x86_64/ARM64 box.

It also means however that when SMC-D is used to communicate between LPARs via an ISM device there is  no way of mapping the DMBs to the same physical memory as there exists no MMU-like layer spanning partitions that could do such a mapping. Meanwhile for machine level firmware including the ISM virtual PCI device it is still possible to _copy_ memory between different memory partitions. So yeah while I do see the appeal of skipping the memcpy() for loopback or even between guests of a paging hypervisor such as KVM, which can map the DMBs on the same physical memory, we must keep in mind this original use case requiring a copy operation.

Quoted 3 lines

Thanks,
Niklas

Hi Niklas,

Thank you so much for the complete and detailed explanation! This provides me a brand new perspective of s390 device that we hadn't dabbled in before. Now I understand why shared memory is unavailable between different LPARs.

Our original intention of proposing loopback device and the incoming device (virtio-ism) for inter-VM is to use SMC-D to accelerate communication in the case with no existing s390 ISM devices. In our conception, s390 ISM device, loopback device and virtio-ism device are parallel and are abstracted by smcd_ops.

  +------------------------+
  |          SMC-D         |
  +------------------------+
  -------- smcd_ops ---------
  +------+ +------+ +------+
  | s390 | | loop | |virtio|
  | ISM  | | back | | -ism |
  | dev  | | dev  | | dev  |
  +------+ +------+ +------+

We also believe that keeping the existing design and behavior of s390 ISM device is unshaken. What we want to get support for is some smcd_ops extension for devices with optional beneficial capability, such as nocopy here (Let's call it this for now), which is really helpful for us in inter-process and inter-VM scenario.

And coincided with IBM's intention to add APIs between SMC-D and devices to support various devices for SMC-D, as mentioned in [2], we send out this RFC and the incoming virio-ism RFC, to provide some examples.

Quoted 51 lines
Quoted 13 lines

# Benchmark Test

  * Test environments:
       - VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
       - SMC sndbuf/RMB size 1MB.

  * Test object:
       - TCP: run on TCP loopback.
       - domain: run on UNIX domain.
       - SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
       - SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

Quoted 7 lines

1. ipc-benchmark (see [3])

  - ./<foo> -c 1000000 -s 100

                        TCP              domain              SMC-lo             SMC-lo-nocpy
Message
rate (msg/s)         75140      129548(+72.41)    152266(+102.64%)         151914(+102.17%)

Interesting that it does beat UNIX domain sockets. Also, see my below comment for nginx/wrk as this seems very similar.

Quoted 18 lines

2. sockperf

  - serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
  - clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

                        TCP                  SMC-lo             SMC-lo-nocpy
Bandwidth(MBps)   4943.359        4936.096(-0.15%)        8239.624(+66.68%) Latency(us)          6.372          3.359(-47.28%)            3.25(-49.00%)

3. iperf3

  - serv: <smc_run> taskset -c <cpu> iperf3 -s
  - clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

                        TCP                  SMC-lo             SMC-lo-nocpy
Bitrate(Gb/s)         40.5            41.4(+2.22%)            76.4(+88.64%)

Quoted 7 lines

4. nginx/wrk

  - serv: <smc_run> nginx
  - clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

                        TCP                  SMC-lo             SMC-lo-nocpy
Requests/s       154643.22      220894.03(+42.84%)        226754.3(+46.63%)

Quoted 8 lines

This result is very interesting indeed. So with the much more realistic nginx/wrk workload it seems to copy hurts much less than the iperf3/sockperf would suggest while SMC-D itself seems to help more. I'd hope that this translates to actual applications as well. Maybe this makes SMC-D based loopback interesting even while keeping the copy, at least until we can come up with a sane way to work a no-copy variant into SMC-D?

I agree, nginx/wrk workload is much more realistic for many applications.

But we also encounter many other cases similar to sockperf on the cloud, which requires high throughput, such as AI training and big data.

So avoidance of copying between DMBs can help these cases a lot :)

Quoted 6 lines
Quoted 6 lines

# Discussion

1. API between SMC-D and ISM device

Quoted 8 lines
Quoted 8 lines

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Quoted 8 lines

Due to the reasons explained above this behavior can't be emulated by ISM devices at least not when crossing partitions. Not sure if we can still incorporate it in the API and allow for both copying and remapping SMC-D like devices, it definitely needs careful consideration and I think also a better understanding of the benefit for real world workloads.

Here I am not rigorous.

Nocopy shouldn't be a standard ISM device behavior indeed. Actually we hope it be a standard optional _SMC-D_ device behavior and defined by smcd_ops.

For devices don't support these options, like ISM device on s390 architecture, .attach_dmb/.detach_dmb and other reasonable extensions (which will be proposed to discuss in incoming virtio-ism RFC) can be set to NULL or return invalid. And for devices do support, they may be used for improving performance in some cases.

In addition, can I know more latest news about the API design? :) , like its scale, will it be a almost refactor of existing interface or incremental patching? and its object, will it be tailored for exact ISM behavior or to reserve some options for other devices, like nocopy here? From my understanding of [2], it might be the latter?

Quoted 13 lines
Quoted 6 lines

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1]) can provide some examples for new API design. And we are very glad to discuss this on the mail list.

Quoted 6 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

Quoted 5 lines

Agree on this part, though it is for the SMC maintainers to decide, I think we would definitely want to be able to use any upcoming inter-VM devices on s390 possibly also in conjunction with ISM devices for communication across partitions.

Quoted 10 lines

Yes, this part needs to be discussed with SMC maintainers. And thank you, we are very glad if our devices can be applied on s390 through the efforts.

Best Regards,
Wen Gu

Quoted 3 lines
Quoted 3 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

As described in
SMC protocol (including SMC-D): https://www.ibm.com/support/pages/system/files/inline-files/IBM%20Shared%20Memory%20Communications%20Version%202_2.pdf
the CLC messages provide a list of up to 8 ISM devices to chose from.
So I would hope that we can use the existing protocol.

Quoted 6 lines

The challenge will be to define GID (Global Interface ID) and CHID (a fabric ID) in a meaningful way for the new devices.
There is always smcd_ops->query_remote_gid() as a safety net. But the idea is that a CHID mismatch is a fast way to tell that these 2 interfaces do match.

Hi Winter and all,

Thanks for your reply and suggestions! And sorry for my late reply because it took me some time to understand SMC-Dv2 protocol and implementation.

I agree with your opinion. The existing SMC-Dv2 protocol whose CLC messages include ism_dev[] list can solve the devices negotiation problem. And I am very willing to use the existing protocol, because we all know that the protocol update is a long and complex process.

If I understand correctly, SMC-D loopback(dummy) device can coordinate with existing SMC-Dv2 protocol as follows. If there is any mistake, please point out.

# Initialization

- Initialize the loopback device with unique GID [Q-1].

- Register the loopback device as SMC-Dv2-capable device with a system_eid whose 24th or 28th byte is non-zero [Q-2], so that this system's smc_ism_v2_capable will be set to TRUE and SMC-Dv2 is available.

# Proposal

- Find the loopback device from the smcd_dev_list in smc_find_ism_v2_device_clnt();

- Record the SEID, GID and CHID[Q-3] of loopback device in the v2 extension part of CLC proposal message.

# Accept

- Check the GID/CHID list and SEID in CLC proposal message, and find local matched ISM device from smcd_dev_list in smc_find_ism_v2_device_serv(). If both sides of the communication are in the same VM and share the same loopback device, the SEID, GID and CHID will match and loopback device will be chosen [Q-4].

- Record the loopback device's GID/CHID and matched SEID into CLC accept message.

# Confirm

- Confirm the server-selected device (loopback device) accordingto CLC accept messages.

- Record the loopback device's GID/CHID and server-selected SEID in CLC confirm message.

Follow the above process, I supplement a patch based on this RFC in the email attachment. With the attachment patch, SMC-D loopback will switch to use SMC-Dv2 protocol.

And in the above process, there are something I want to consult and discuss, which is marked with '[Q-*]' in the above description.

# [Q-1]:

The GID of loopback device is randomly generated in this RFC patch set, but I will find a way to unique the GID in formal patches. Any suggestions are welcome.

# [Q-2]:

In Linux implementation, the system_eid of the first registered smcd device will determinate system's smc_ism_v2_capable (see smcd_register_dev()).

And I wonder that

1) How to define the system_eid? It can be inferred from the code that the 24th and 28th byte

    are special for SMC-Dv2. So in attachment patch, I define the loopback device SEID as

    static struct smc_lo_systemeid LO_SYSTEM_EID = {
            .seid_string = "SMC-SYSZ-LOSEID000000000",
            .serial_number = "1000",
            .type = "1000",
    };

    Is there anything else I need to pay attention to?

2) Seems only the first added smcd device determinate the system smc_ism_v2_capable? If two different smcd devices respectively with v1-indicated and v2-indicated system_eid, will the order in which they are registered affects the result of smc_ism_v2_capable ?

# [Q-3]:

In attachment patch, I define a special CHID (0xFFFF) for loopback device, as a kind of 'unassociated ISM CHID' that not associated with any IP (OSA or HiperSockets) interfaces.

What's your opinion about this?

# [Q-4]:

In current Linux implementation, server will select the first successfully initialized device from the candidates as the final selected one in smc_find_ism_v2_device_serv().

for (i = 0; i < matches; i++) {
	ini->smcd_version = SMC_V2;
	ini->is_smcd = true;
	ini->ism_selected = i;
	rc = smc_listen_ism_init(new_smc, ini);
	if (rc) {
		smc_find_ism_store_rc(rc, ini);
		/* try next active ISM device */
		continue;
	}
	return; /* matching and usable V2 ISM device found */
}

IMHO, maybe candidate devices should have different priorities? For example, the loopback device may be preferred to use if loopback is available.

Best Regards,
Wen Gu

Quoted 17 lines
Quoted 17 lines
Quoted 17 lines
Quoted 17 lines

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
  1. Fix some build WARNINGs complained by kernel test rebot

  2. Add iperf3 test data.

Wen Gu (5):
   net/smc: introduce SMC-D loopback device
   net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
   net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

Quoted 16 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
   include/net/smc.h      |   3 +
   net/smc/Makefile       |   2 +-
   net/smc/af_smc.c       |  88 +++++++++++-
   net/smc/smc_cdc.c      |  59 ++++++--
   net/smc/smc_cdc.h      |   1 +
   net/smc/smc_clc.c      |   4 +-
   net/smc/smc_core.c     |  62 +++++++++
   net/smc/smc_core.h     |   2 +
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
   net/smc/smc_ism.c      |  39 +++++-
   net/smc/smc_ism.h      |   2 +
   net/smc/smc_loopback.c | 358 +++++++++++++++++++++++++++++++++++++++++++++++++
   net/smc/smc_loopback.h |  63 +++++++++
   12 files changed, 662 insertions(+), 21 deletions(-)

  create mode 100644 net/smc/smc_loopback.c
  create mode 100644 net/smc/smc_loopback.h

From bc94984d599e2e8cbc408c42896973745c533bb7 Mon Sep 17 00:00:00 2001
From: Wen Gu <[email protected]>
Date: Sat, 7 Jan 2023 16:58:37 +0800
Subject: [PATCH] net/smc: define SEID and CHID of loopback device

This patch defines SEID and CHID of loopback device and take it as SMC-Dv2 device.

Besides, this patch delete the most logic of RFC patch 2/5 as well because device selection will be covered by SMC-Dv2 protocol.

---
 net/smc/af_smc.c       | 50 +++++---------------------------------------------
 net/smc/smc_clc.c      |  4 +---
 net/smc/smc_loopback.c | 11 +++++++----
 3 files changed, 13 insertions(+), 52 deletions(-)
net/smc/af_smc.c +5 -45
index c7de566..4396392 100644
@@ -979,28 +979,6 @@ static int smc_find_ism_device(struct smc_sock *smc, struct smc_init_info *ini)
 	return 0;
 }
 
-/* check if there is a lo device available for this connection. */
-static int smc_find_lo_device(struct smc_sock *smc, struct smc_init_info *ini)
-{
-	struct smcd_dev *sdev;
-
-	mutex_lock(&smcd_dev_list.mutex);
-	list_for_each_entry(sdev, &smcd_dev_list.list, list) {
-		if (sdev->is_loopback && !sdev->going_away &&
-		    (!ini->ism_peer_gid[0] ||
-		     !smc_ism_cantalk(ini->ism_peer_gid[0], ini->vlan_id,
-				      sdev))) {
-			ini->ism_dev[0] = sdev;
-			break;
-		}
-	}
-	mutex_unlock(&smcd_dev_list.mutex);
-	if (!ini->ism_dev[0])
-		return SMC_CLC_DECL_NOSMCDDEV;
-	ini->ism_chid[0] = smc_ism_get_chid(ini->ism_dev[0]);
-	return 0;
-}
-
 /* is chid unique for the ism devices that are already determined? */
 static bool smc_find_ism_v2_is_unique_chid(u16 chid, struct smc_init_info *ini,
 					   int cnt)
@@ -1066,19 +1044,10 @@ static int smc_find_proposal_devices(struct smc_sock *smc,
 {
 	int rc = 0;
 
-	/* TODO:
-	 * How to indicate to peer if ism device and loopback
-	 * device are both available ?
-	 *
-	 * The RFC patch hasn't resolved this, just simply always
-	 * chooses loopback device first, and fallback if loopback
-	 * communication is impossible.
-	 */
 	/* check if there is an ism or loopback device available */
 	if (!(ini->smcd_version & SMC_V1) ||
-	    (smc_find_lo_device(smc, ini) &&
-	    (smc_find_ism_device(smc, ini) ||
-	    smc_connect_ism_vlan_setup(smc, ini))))
+	    smc_find_ism_device(smc, ini) ||
+	    smc_connect_ism_vlan_setup(smc, ini))
 		ini->smcd_version &= ~SMC_V1;
 	/* else ISM V1 is supported for this connection */
 
@@ -2178,18 +2147,9 @@ static void smc_find_ism_v1_device_serv(struct smc_sock *new_smc,
 	ini->is_smcd = true; /* prepare ISM check */
 	ini->ism_peer_gid[0] = ntohll(pclc_smcd->ism.gid);
 
-	/* TODO:
-	 * How to know that peer has both loopback and ism device ?
-	 *
-	 * The RFC patch hasn't resolved this, simply tries loopback
-	 * device first, then ism device.
-	 */
-	/* find available loopback or ism device */
-	if (smc_find_lo_device(new_smc, ini)) {
-		rc = smc_find_ism_device(new_smc, ini);
-		if (rc)
-			goto not_found;
-	}
+	rc = smc_find_ism_device(new_smc, ini);
+	if (rc)
+		goto not_found;
 
 	ini->ism_selected = 0;
 	rc = smc_listen_ism_init(new_smc, ini);
net/smc/smc_clc.c +1 -3
index 3887692..dfb9797 100644
@@ -486,9 +486,7 @@ static int smc_clc_prfx_set4_rcu(struct dst_entry *dst, __be32 ipv4,
 		return -ENODEV;
 
 	in_dev_for_each_ifa_rcu(ifa, in_dev) {
-		/* add loopback support */
-		if (inet_addr_type(dev_net(dst->dev), ipv4) != RTN_LOCAL &&
-		    !inet_ifa_match(ipv4, ifa))
+		if (!inet_ifa_match(ipv4, ifa))
 			continue;
 		prop->prefix_len = inet_mask_len(ifa->ifa_mask);
 		prop->outgoing_subnet = ifa->ifa_address & ifa->ifa_mask;
net/smc/smc_loopback.c +7 -4
index 3dedcc4..642b241 100644
@@ -19,13 +19,14 @@
 #include "smc_loopback.h"
 
 #define DRV_NAME "smc_lodev"
+#define LO_CHID	0xFFFF	/* specific for lo dev */
 
 struct smc_lo_dev *lo_dev;
 
 static struct smc_lo_systemeid LO_SYSTEM_EID = {
 	.seid_string = "SMC-SYSZ-LOSEID000000000",
-	.serial_number = "0000",
-	.type = "0000",
+	.serial_number = "1000",
+	.type = "1000",
 };
 
 static int smc_lo_query_rgid(struct smcd_dev *smcd, u64 rgid, u32 vid_valid,
@@ -33,7 +34,9 @@ static int smc_lo_query_rgid(struct smcd_dev *smcd, u64 rgid, u32 vid_valid,
 {
 	struct smc_lo_dev *ldev = smcd->priv;
 
-	/* return local gid */
+	if (!vid_valid || vid != ISM_RESERVED_VLANID)
+		return -EINVAL;
+	/* rgid should be equal to lgid */
 	if (!ldev || rgid != ldev->lgid)
 		return -ENETUNREACH;
 	return 0;
@@ -255,7 +258,7 @@ static u8 *smc_lo_get_system_eid(void)
 
 static u16 smc_lo_get_chid(struct smcd_dev *smcd)
 {
-	return 0;
+	return LO_CHID;
 }
 
 static const struct smcd_ops lo_ops = {
Signature · 1 line
1.8.3.1
#12 lore raw
Quoted 276 lines · Wen Gu
Quoted 273 lines · Alexandra Winter
Quoted 254 lines · Wen Gu
Quoted 89 lines · Niklas Schnelle
Quoted 44 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

   +------------------------------------------+
   |  +-----------+           +-----------+   |
   |  | process A |           | process B |   |
   |  +-----------+           +-----------+   |
   |       ^                        ^         |
   |       |    +---------------+   |         |
   |       |    |   SMC stack   |   |         |
   |       +--->| +-----------+ |<--|         |
   |            | |   dummy   | |             |
   |            | |   device  | |             |
   |            +-+-----------+-+             |
   |                   VM                     |
   +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

Quoted 10 lines
   +----------+                     +----------+
   | socket A |                     | socket B |
   +----------+                     +----------+
         |                               ^
         |         +---------+           |
    regard as      |         | ----------|
    local sndbuf   |  B's    |     regard as
         |         |  RMB    |     local RMB
         |-------> |         |
                   +---------+

Hi Wen Gu,

I maintain the s390 specific PCI support in Linux and would like to provide a bit of background on this. You're surely wondering why we even have a copy in there for our ISM virtual PCI device. To understand why this copy operation exists and why we need to keep it working, one needs a bit of s390 aka mainframe background.

On s390 all (currently supported) native machines have a mandatory machine level hypervisor. All OSs whether z/OS or Linux run either on this machine level hypervisor as so called Logical Partitions (LPARs) or as second/third/… level guests on e.g. a KVM or z/VM hypervisor that in turn runs in an LPAR. Now, in terms of memory this machine level hypervisor sometimes called PR/SM unlike KVM, z/VM, or VMWare is a partitioning hypervisor without paging. This is one of the main reasons for the very-near-native performance of the machine hypervisor as the memory of its guests acts just like native RAM on other systems. It is never paged out and always accessible to IOMMU translated DMA from devices without the need for pinning pages and besides a trivial offset/limit adjustment an LPAR's MMU does the same amount of work as an MMU on a bare metal x86_64/ARM64 box.

It also means however that when SMC-D is used to communicate between LPARs via an ISM device there is  no way of mapping the DMBs to the same physical memory as there exists no MMU-like layer spanning partitions that could do such a mapping. Meanwhile for machine level firmware including the ISM virtual PCI device it is still possible to _copy_ memory between different memory partitions. So yeah while I do see the appeal of skipping the memcpy() for loopback or even between guests of a paging hypervisor such as KVM, which can map the DMBs on the same physical memory, we must keep in mind this original use case requiring a copy operation.

Quoted 3 lines

Thanks,
Niklas

Hi Niklas,

Thank you so much for the complete and detailed explanation! This provides me a brand new perspective of s390 device that we hadn't dabbled in before. Now I understand why shared memory is unavailable between different LPARs.

Our original intention of proposing loopback device and the incoming device (virtio-ism) for inter-VM is to use SMC-D to accelerate communication in the case with no existing s390 ISM devices. In our conception, s390 ISM device, loopback device and virtio-ism device are parallel and are abstracted by smcd_ops.

  +------------------------+
  |          SMC-D         |
  +------------------------+
  -------- smcd_ops ---------
  +------+ +------+ +------+
  | s390 | | loop | |virtio|
  | ISM  | | back | | -ism |
  | dev  | | dev  | | dev  |
  +------+ +------+ +------+

We also believe that keeping the existing design and behavior of s390 ISM device is unshaken. What we want to get support for is some smcd_ops extension for devices with optional beneficial capability, such as nocopy here (Let's call it this for now), which is really helpful for us in inter-process and inter-VM scenario.

And coincided with IBM's intention to add APIs between SMC-D and devices to support various devices for SMC-D, as mentioned in [2], we send out this RFC and the incoming virio-ism RFC, to provide some examples.

Quoted 52 lines
Quoted 13 lines

# Benchmark Test

   * Test environments:
        - VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
        - SMC sndbuf/RMB size 1MB.

   * Test object:
        - TCP: run on TCP loopback.
        - domain: run on UNIX domain.
        - SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
        - SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

Quoted 7 lines

1. ipc-benchmark (see [3])

   - ./<foo> -c 1000000 -s 100

                         TCP              domain SMC-lo             SMC-lo-nocpy
Message
rate (msg/s)         75140      129548(+72.41) 152266(+102.64%)         151914(+102.17%)

Interesting that it does beat UNIX domain sockets. Also, see my below comment for nginx/wrk as this seems very similar.

Quoted 19 lines

2. sockperf

   - serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
   - clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

                         TCP                  SMC-lo SMC-lo-nocpy
Bandwidth(MBps)   4943.359        4936.096(-0.15%) 8239.624(+66.68%)
Latency(us)          6.372          3.359(-47.28%) 3.25(-49.00%)

3. iperf3

   - serv: <smc_run> taskset -c <cpu> iperf3 -s
   - clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

                         TCP                  SMC-lo SMC-lo-nocpy
Bitrate(Gb/s)         40.5            41.4(+2.22%) 76.4(+88.64%)

Quoted 7 lines

4. nginx/wrk

   - serv: <smc_run> nginx
   - clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

                         TCP                  SMC-lo SMC-lo-nocpy
Requests/s       154643.22      220894.03(+42.84%) 226754.3(+46.63%)

Quoted 8 lines

This result is very interesting indeed. So with the much more realistic nginx/wrk workload it seems to copy hurts much less than the iperf3/sockperf would suggest while SMC-D itself seems to help more. I'd hope that this translates to actual applications as well. Maybe this makes SMC-D based loopback interesting even while keeping the copy, at least until we can come up with a sane way to work a no-copy variant into SMC-D?

I agree, nginx/wrk workload is much more realistic for many applications.

But we also encounter many other cases similar to sockperf on the cloud, which requires high throughput, such as AI training and big data.

So avoidance of copying between DMBs can help these cases a lot :)

Quoted 6 lines
Quoted 6 lines

# Discussion

1. API between SMC-D and ISM device

Quoted 8 lines
Quoted 8 lines

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Quoted 8 lines

Due to the reasons explained above this behavior can't be emulated by ISM devices at least not when crossing partitions. Not sure if we can still incorporate it in the API and allow for both copying and remapping SMC-D like devices, it definitely needs careful consideration and I think also a better understanding of the benefit for real world workloads.

Here I am not rigorous.

Nocopy shouldn't be a standard ISM device behavior indeed. Actually we hope it be a standard optional _SMC-D_ device behavior and defined by smcd_ops.

For devices don't support these options, like ISM device on s390 architecture, .attach_dmb/.detach_dmb and other reasonable extensions (which will be proposed to discuss in incoming virtio-ism RFC) can be set to NULL or return invalid. And for devices do support, they may be used for improving performance in some cases.

In addition, can I know more latest news about the API design? :) , like its scale, will it be a almost refactor of existing interface or incremental patching? and its object, will it be tailored for exact ISM behavior or to reserve some options for other devices, like nocopy here? From my understanding of [2], it might be the latter?

Quoted 13 lines
Quoted 6 lines

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1]) can provide some examples for new API design. And we are very glad to discuss this on the mail list.

Quoted 6 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

Quoted 5 lines

Agree on this part, though it is for the SMC maintainers to decide, I think we would definitely want to be able to use any upcoming inter-VM devices on s390 possibly also in conjunction with ISM devices for communication across partitions.

Quoted 10 lines

Yes, this part needs to be discussed with SMC maintainers. And thank you, we are very glad if our devices can be applied on s390 through the efforts.

Best Regards,
Wen Gu

Quoted 3 lines
Quoted 3 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

As described in
SMC protocol (including SMC-D): https://www.ibm.com/support/pages/system/files/inline-files/IBM%20Shared%20Memory%20Communications%20Version%202_2.pdf
the CLC messages provide a list of up to 8 ISM devices to chose from.
So I would hope that we can use the existing protocol.

Quoted 7 lines
Quoted 6 lines

The challenge will be to define GID (Global Interface ID) and CHID (a fabric ID) in a meaningful way for the new devices.
There is always smcd_ops->query_remote_gid()  as a safety net. But the idea is that a CHID mismatch is a fast way to tell that these 2 interfaces do match.

FYI, we just sent the rest part of the API to the net-next https://lore.kernel.org/netdev/[email protected]/T/#t,
which should answer some questions in your patch series.

Quoted 16 lines

Hi Winter and all,

Thanks for your reply and suggestions! And sorry for my late reply because it took me some time to understand SMC-Dv2 protocol and implementation.

I agree with your opinion. The existing SMC-Dv2 protocol whose CLC messages include ism_dev[] list can solve the devices negotiation problem. And I am very willing to use the existing protocol, because we all know that the protocol update is a long and complex process.

If I understand correctly, SMC-D loopback(dummy) device can coordinate with existing SMC-Dv2 protocol as follows. If there is any mistake, please point out.

# Initialization

Quoted 8 lines

- Initialize the loopback device with unique GID [Q-1].

- Register the loopback device as SMC-Dv2-capable device with a system_eid whose 24th
   or 28th byte is non-zero [Q-2], so that this system's
smc_ism_v2_capable will be set
   to TRUE and SMC-Dv2 is available.

The decision point is the VLAN_ID, if it is x1FFF, the device will support V2. i.e. If you can have subnet with VLAN_ID x1FFF, then the SEID is necessary, so that the series or types is non-zero. (*1)

Quoted 39 lines

# Proposal

- Find the loopback device from the smcd_dev_list in smc_find_ism_v2_device_clnt();

- Record the SEID, GID and CHID[Q-3] of loopback device in the v2 extension part of CLC
   proposal message.

# Accept

- Check the GID/CHID list and SEID in CLC proposal message, and find local matched ISM
   device from smcd_dev_list in smc_find_ism_v2_device_serv(). If both
sides of the
   communication are in the same VM and share the same loopback device,
the SEID, GID and
   CHID will match and loopback device will be chosen [Q-4].

- Record the loopback device's GID/CHID and matched SEID into CLC accept message.

# Confirm

- Confirm the server-selected device (loopback device) accordingto CLC accept messages.

- Record the loopback device's GID/CHID and server-selected SEID in CLC confirm message.

Follow the above process, I supplement a patch based on this RFC in the email attachment. With the attachment patch, SMC-D loopback will switch to use SMC-Dv2 protocol.

Quoted 8 lines

And in the above process, there are something I want to consult and discuss, which is marked with '[Q-*]' in the above description.

# [Q-1]:

The GID of loopback device is randomly generated in this RFC patch set, but I will find a way to unique the GID in formal patches. Any suggestions are welcome.

I think the randowmly generated GID is fine in your case, which is equivalent to the IP address.

Quoted 12 lines

# [Q-2]:

In Linux implementation, the system_eid of the first registered smcd device will determinate system's smc_ism_v2_capable (see smcd_register_dev()).

And I wonder that

1) How to define the system_eid? It can be inferred from the code that the 24th and 28th byte are special for SMC-Dv2. So in attachment patch, I define the
loopback device SEID as

Quoted 8 lines

    static struct smc_lo_systemeid LO_SYSTEM_EID = { .seid_string = "SMC-SYSZ-LOSEID000000000", .serial_number = "1000",
            .type = "1000",
    };

    Is there anything else I need to pay attention to?

If you just want to use V2, such defination looks good. e.g. you can use some unique information from "lshw"

Quoted 7 lines

2) Seems only the first added smcd device determinate the system smc_ism_v2_capable? If two different smcd devices respectively with v1-indicated and
v2-indicated system_eid, will
    the order in which they are registered affects the result of
smc_ism_v2_capable ?

see (*1)

Quoted 8 lines

# [Q-3]:

In attachment patch, I define a special CHID (0xFFFF) for loopback device, as a kind of 'unassociated ISM CHID' that not associated with any IP (OSA or HiperSockets) interfaces.

What's your opinion about this?

It looks good to me

Quoted 19 lines

# [Q-4]:

In current Linux implementation, server will select the first successfully initialized device from the candidates as the final selected one in smc_find_ism_v2_device_serv().

for (i = 0; i < matches; i++) {
    ini->smcd_version = SMC_V2;
    ini->is_smcd = true;
    ini->ism_selected = i;
    rc = smc_listen_ism_init(new_smc, ini);
    if (rc) {
        smc_find_ism_store_rc(rc, ini);
        /* try next active ISM device */ continue;
    }
    return; /* matching and usable V2 ISM device found */
}

Quoted 3 lines

IMHO, maybe candidate devices should have different priorities? For example, the loopback device may be preferred to use if loopback is available.

IMO, I'd prefer such a order: ISM -> loopback -> RoCE Because ISM for SMC-D is our standard user case, not loopback.

Quoted 38 lines

Best Regards,
Wen Gu

Quoted 34 lines
Quoted 34 lines
Quoted 34 lines
Quoted 34 lines

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
   1. Fix some build WARNINGs complained by kernel test rebot

   2. Add iperf3 test data.

Wen Gu (5):
    net/smc: introduce SMC-D loopback device
    net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
    net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

   include/net/smc.h      |   3 +
   net/smc/Makefile       |   2 +-
   net/smc/af_smc.c       |  88 +++++++++++-

   net/smc/smc_cdc.c      |  59 ++++++--
   net/smc/smc_cdc.h      |   1 +
   net/smc/smc_clc.c      |   4 +-

   net/smc/smc_core.c     |  62 +++++++++
   net/smc/smc_core.h     |   2 +
   net/smc/smc_ism.c      |  39 +++++-
   net/smc/smc_ism.h      |   2 +
   net/smc/smc_loopback.c | 358
+++++++++++++++++++++++++++++++++++++++++++++++++
   net/smc/smc_loopback.h |  63 +++++++++

   12 files changed, 662 insertions(+), 21 deletions(-)

   create mode 100644 net/smc/smc_loopback.c
   create mode 100644 net/smc/smc_loopback.h

author reviewer
#13 lore raw
Quoted 286 lines · Wenjia Zhang
Quoted 275 lines · Wen Gu
Quoted 272 lines · Alexandra Winter
Quoted 253 lines · Wen Gu
Quoted 89 lines · Niklas Schnelle
Quoted 44 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

   +------------------------------------------+
   |  +-----------+           +-----------+   |
   |  | process A |           | process B |   |
   |  +-----------+           +-----------+   |
   |       ^                        ^         |
   |       |    +---------------+   |         |
   |       |    |   SMC stack   |   |         |
   |       +--->| +-----------+ |<--|         |
   |            | |   dummy   | |             |
   |            | |   device  | |             |
   |            +-+-----------+-+             |
   |                   VM                     |
   +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

Quoted 10 lines
   +----------+                     +----------+
   | socket A |                     | socket B |
   +----------+                     +----------+
         |                               ^
         |         +---------+           |
    regard as      |         | ----------|
    local sndbuf   |  B's    |     regard as
         |         |  RMB    |     local RMB
         |-------> |         |
                   +---------+

Hi Wen Gu,

I maintain the s390 specific PCI support in Linux and would like to provide a bit of background on this. You're surely wondering why we even have a copy in there for our ISM virtual PCI device. To understand why this copy operation exists and why we need to keep it working, one needs a bit of s390 aka mainframe background.

On s390 all (currently supported) native machines have a mandatory machine level hypervisor. All OSs whether z/OS or Linux run either on this machine level hypervisor as so called Logical Partitions (LPARs) or as second/third/… level guests on e.g. a KVM or z/VM hypervisor that in turn runs in an LPAR. Now, in terms of memory this machine level hypervisor sometimes called PR/SM unlike KVM, z/VM, or VMWare is a partitioning hypervisor without paging. This is one of the main reasons for the very-near-native performance of the machine hypervisor as the memory of its guests acts just like native RAM on other systems. It is never paged out and always accessible to IOMMU translated DMA from devices without the need for pinning pages and besides a trivial offset/limit adjustment an LPAR's MMU does the same amount of work as an MMU on a bare metal x86_64/ARM64 box.

It also means however that when SMC-D is used to communicate between LPARs via an ISM device there is  no way of mapping the DMBs to the same physical memory as there exists no MMU-like layer spanning partitions that could do such a mapping. Meanwhile for machine level firmware including the ISM virtual PCI device it is still possible to _copy_ memory between different memory partitions. So yeah while I do see the appeal of skipping the memcpy() for loopback or even between guests of a paging hypervisor such as KVM, which can map the DMBs on the same physical memory, we must keep in mind this original use case requiring a copy operation.

Quoted 3 lines

Thanks,
Niklas

Hi Niklas,

Thank you so much for the complete and detailed explanation! This provides me a brand new perspective of s390 device that we hadn't dabbled in before. Now I understand why shared memory is unavailable between different LPARs.

Our original intention of proposing loopback device and the incoming device (virtio-ism) for inter-VM is to use SMC-D to accelerate communication in the case with no existing s390 ISM devices. In our conception, s390 ISM device, loopback device and virtio-ism device are parallel and are abstracted by smcd_ops.

  +------------------------+
  |          SMC-D         |
  +------------------------+
  -------- smcd_ops ---------
  +------+ +------+ +------+
  | s390 | | loop | |virtio|
  | ISM  | | back | | -ism |
  | dev  | | dev  | | dev  |
  +------+ +------+ +------+

We also believe that keeping the existing design and behavior of s390 ISM device is unshaken. What we want to get support for is some smcd_ops extension for devices with optional beneficial capability, such as nocopy here (Let's call it this for now), which is really helpful for us in inter-process and inter-VM scenario.

And coincided with IBM's intention to add APIs between SMC-D and devices to support various devices for SMC-D, as mentioned in [2], we send out this RFC and the incoming virio-ism RFC, to provide some examples.

Quoted 51 lines
Quoted 13 lines

# Benchmark Test

   * Test environments:
        - VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
        - SMC sndbuf/RMB size 1MB.

   * Test object:
        - TCP: run on TCP loopback.
        - domain: run on UNIX domain.
        - SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
        - SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

Quoted 7 lines

1. ipc-benchmark (see [3])

   - ./<foo> -c 1000000 -s 100

                         TCP              domain SMC-lo             SMC-lo-nocpy Message
rate (msg/s)         75140      129548(+72.41) 152266(+102.64%)         151914(+102.17%)

Interesting that it does beat UNIX domain sockets. Also, see my below comment for nginx/wrk as this seems very similar.

Quoted 18 lines

2. sockperf

   - serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
   - clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

                         TCP                  SMC-lo SMC-lo-nocpy Bandwidth(MBps)   4943.359        4936.096(-0.15%) 8239.624(+66.68%) Latency(us)          6.372          3.359(-47.28%) 3.25(-49.00%)

3. iperf3

   - serv: <smc_run> taskset -c <cpu> iperf3 -s
   - clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

                         TCP                  SMC-lo SMC-lo-nocpy Bitrate(Gb/s)         40.5            41.4(+2.22%) 76.4(+88.64%)

Quoted 7 lines

4. nginx/wrk

   - serv: <smc_run> nginx
   - clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

                         TCP                  SMC-lo SMC-lo-nocpy Requests/s       154643.22      220894.03(+42.84%) 226754.3(+46.63%)

Quoted 8 lines

This result is very interesting indeed. So with the much more realistic nginx/wrk workload it seems to copy hurts much less than the iperf3/sockperf would suggest while SMC-D itself seems to help more. I'd hope that this translates to actual applications as well. Maybe this makes SMC-D based loopback interesting even while keeping the copy, at least until we can come up with a sane way to work a no-copy variant into SMC-D?

I agree, nginx/wrk workload is much more realistic for many applications.

But we also encounter many other cases similar to sockperf on the cloud, which requires high throughput, such as AI training and big data.

So avoidance of copying between DMBs can help these cases a lot :)

Quoted 6 lines
Quoted 6 lines

# Discussion

1. API between SMC-D and ISM device

Quoted 8 lines
Quoted 8 lines

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Quoted 8 lines

Due to the reasons explained above this behavior can't be emulated by ISM devices at least not when crossing partitions. Not sure if we can still incorporate it in the API and allow for both copying and remapping SMC-D like devices, it definitely needs careful consideration and I think also a better understanding of the benefit for real world workloads.

Here I am not rigorous.

Nocopy shouldn't be a standard ISM device behavior indeed. Actually we hope it be a standard optional _SMC-D_ device behavior and defined by smcd_ops.

For devices don't support these options, like ISM device on s390 architecture, .attach_dmb/.detach_dmb and other reasonable extensions (which will be proposed to discuss in incoming virtio-ism RFC) can be set to NULL or return invalid. And for devices do support, they may be used for improving performance in some cases.

In addition, can I know more latest news about the API design? :) , like its scale, will it be a almost refactor of existing interface or incremental patching? and its object, will it be tailored for exact ISM behavior or to reserve some options for other devices, like nocopy here? From my understanding of [2], it might be the latter?

Quoted 13 lines
Quoted 6 lines

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1]) can provide some examples for new API design. And we are very glad to discuss this on the mail list.

Quoted 6 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

Quoted 5 lines

Agree on this part, though it is for the SMC maintainers to decide, I think we would definitely want to be able to use any upcoming inter-VM devices on s390 possibly also in conjunction with ISM devices for communication across partitions.

Quoted 10 lines

Yes, this part needs to be discussed with SMC maintainers. And thank you, we are very glad if our devices can be applied on s390 through the efforts.

Best Regards,
Wen Gu

Quoted 3 lines
Quoted 3 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

As described in
SMC protocol (including SMC-D): https://www.ibm.com/support/pages/system/files/inline-files/IBM%20Shared%20Memory%20Communications%20Version%202_2.pdf
the CLC messages provide a list of up to 8 ISM devices to chose from.
So I would hope that we can use the existing protocol.

Quoted 7 lines
Quoted 6 lines

The challenge will be to define GID (Global Interface ID) and CHID (a fabric ID) in a meaningful way for the new devices.
There is always smcd_ops->query_remote_gid()  as a safety net. But the idea is that a CHID mismatch is a fast way to tell that these 2 interfaces do match.

Quoted 3 lines

FYI, we just sent the rest part of the API to the net-next https://lore.kernel.org/netdev/[email protected]/T/#t,
which should answer some questions in your patch series.

Hi Wenjia,

Thanks for notification and your excellent work!

I will learn about the new API and find if it fits our needs. If all goes well I will refactor my RFC based on the new API and send a v2, most likely after the Chinese New Year.

Quoted 17 lines
Quoted 17 lines

Hi Winter and all,

Thanks for your reply and suggestions! And sorry for my late reply because it took me some time to understand SMC-Dv2 protocol and implementation.

I agree with your opinion. The existing SMC-Dv2 protocol whose CLC messages include ism_dev[] list can solve the devices negotiation problem. And I am very willing to use the existing protocol, because we all know that the protocol update is a long and complex process.

If I understand correctly, SMC-D loopback(dummy) device can coordinate with existing SMC-Dv2 protocol as follows. If there is any mistake, please point out.

# Initialization

Quoted 7 lines
Quoted 6 lines

- Initialize the loopback device with unique GID [Q-1].

- Register the loopback device as SMC-Dv2-capable device with a system_eid whose 24th or 28th byte is non-zero [Q-2], so that this system's smc_ism_v2_capable will be set to TRUE and SMC-Dv2 is available.

The decision point is the VLAN_ID, if it is x1FFF, the device will support V2. i.e. If you can have subnet with VLAN_ID x1FFF, then the SEID is necessary, so that the series or types is non-zero. (*1)

In case there is any misunderstanding between us, I would like to rephrase my [Q-2] question:

int smcd_register_dev(struct smcd_dev *smcd)

{
	<...>
	mutex_lock(&smcd_dev_list.mutex);
	if (list_empty(&smcd_dev_list.list)) {
		u8 *system_eid = NULL;

		smcd->ops->get_system_eid(smcd, &system_eid);
		if (system_eid[24] != '0' || system_eid[28] != '0') {
			smc_ism_v2_capable = true;
			memcpy(smc_ism_v2_system_eid, system_eid,
			       SMC_MAX_EID_LEN);
		}
	}
	<...>
}

It can be inferred from smcd_register_dev() that:

1) The 24th and 28th byte are special and determinate whether smc_ism_v2_capable is true. Besides these, do other bytes of system_eid have hidden meanings that need attention ?

2) Only when smcd_dev_list is empty, the added smcd_dev will be checked, and its system_eid determinates whether smc_ism_v2_capable is true. Why only the first added device will be checked ?

If the first added smcd_dev has an system_eid whose 24th and 28th bytes are zero, and the second added smcd_dev has an system_eid whose 24th and 28th bytes are non-zero. Should smc_ism_v2_capable be true, since the second smcd_dev has v2-indicated system_eid ?

Quoted 34 lines
Quoted 34 lines

# Proposal

- Find the loopback device from the smcd_dev_list in smc_find_ism_v2_device_clnt();

- Record the SEID, GID and CHID[Q-3] of loopback device in the v2 extension part of CLC proposal message.

# Accept

- Check the GID/CHID list and SEID in CLC proposal message, and find local matched ISM device from smcd_dev_list in smc_find_ism_v2_device_serv(). If both sides of the communication are in the same VM and share the same loopback device, the SEID, GID and CHID will match and loopback device will be chosen [Q-4].

- Record the loopback device's GID/CHID and matched SEID into CLC accept message.

# Confirm

- Confirm the server-selected device (loopback device) accordingto CLC accept messages.

- Record the loopback device's GID/CHID and server-selected SEID in CLC confirm message.

Follow the above process, I supplement a patch based on this RFC in the email attachment. With the attachment patch, SMC-D loopback will switch to use SMC-Dv2 protocol.

And in the above process, there are something I want to consult and discuss, which is marked with '[Q-*]' in the above description.

Quoted 6 lines
Quoted 5 lines

# [Q-1]:

The GID of loopback device is randomly generated in this RFC patch set, but I will find a way to unique the GID in formal patches. Any suggestions are welcome.

I think the randowmly generated GID is fine in your case, which is equivalent to the IP address.

Since whether the two sides can communicate through the loopback will be judged by whether the gid of their loopback device is equal, the random GID may bring the risk of misjudgment because it may not be unique. But considering this is an RFC, I simply used random GIDs.

Quoted 17 lines
Quoted 17 lines

# [Q-2]:

In Linux implementation, the system_eid of the first registered smcd device will determinate system's smc_ism_v2_capable (see smcd_register_dev()).

And I wonder that

1) How to define the system_eid? It can be inferred from the code that the 24th and 28th byte are special for SMC-Dv2. So in attachment patch, I define the loopback device SEID as

     static struct smc_lo_systemeid LO_SYSTEM_EID = { .seid_string = "SMC-SYSZ-LOSEID000000000", .serial_number = "1000",
             .type = "1000",
     };

Quoted 4 lines
Quoted 2 lines

     Is there anything else I need to pay attention to?

If you just want to use V2, such defination looks good. e.g. you can use some unique information from "lshw"

OK, thank you.

Quoted 7 lines
Quoted 5 lines

2) Seems only the first added smcd device determinate the system smc_ism_v2_capable? If two different smcd devices respectively with v1-indicated and v2-indicated system_eid, will the order in which they are registered affects the result of smc_ism_v2_capable ?

see (*1)

Quoted 1 line
Quoted 8 lines
Quoted 7 lines

# [Q-3]:

In attachment patch, I define a special CHID (0xFFFF) for loopback device, as a kind of 'unassociated ISM CHID' that not associated with any IP (OSA or HiperSockets) interfaces.

What's your opinion about this?

It looks good to me

OK.

Quoted 19 lines
Quoted 19 lines

# [Q-4]:

In current Linux implementation, server will select the first successfully initialized device from the candidates as the final selected one in smc_find_ism_v2_device_serv().

for (i = 0; i < matches; i++) {
     ini->smcd_version = SMC_V2;
     ini->is_smcd = true;
     ini->ism_selected = i;
     rc = smc_listen_ism_init(new_smc, ini);
     if (rc) {
         smc_find_ism_store_rc(rc, ini);
         /* try next active ISM device */ continue;
     }
     return; /* matching and usable V2 ISM device found */ }

Quoted 5 lines
Quoted 3 lines

IMHO, maybe candidate devices should have different priorities? For example, the loopback device may be preferred to use if loopback is available.

IMO, I'd prefer such a order: ISM -> loopback -> RoCE Because ISM for SMC-D is our standard user case, not loopback.

OK, will follow this order.

Quoted 37 lines
Quoted 37 lines

Best Regards,
Wen Gu

Quoted 33 lines
Quoted 33 lines
Quoted 33 lines
Quoted 33 lines

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
   1. Fix some build WARNINGs complained by kernel test rebot

   2. Add iperf3 test data.

Wen Gu (5):
    net/smc: introduce SMC-D loopback device
    net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
    net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

   include/net/smc.h      |   3 +
   net/smc/Makefile       |   2 +-
   net/smc/af_smc.c       |  88 +++++++++++-

   net/smc/smc_cdc.c      |  59 ++++++--
   net/smc/smc_cdc.h      |   1 +
   net/smc/smc_clc.c      |   4 +-

   net/smc/smc_core.c     |  62 +++++++++
   net/smc/smc_core.h     |   2 +
   net/smc/smc_ism.c      |  39 +++++-
   net/smc/smc_ism.h      |   2 +
   net/smc/smc_loopback.c | 358 +++++++++++++++++++++++++++++++++++++++++++++++++ net/smc/smc_loopback.h |  63 +++++++++

   12 files changed, 662 insertions(+), 21 deletions(-)

   create mode 100644 net/smc/smc_loopback.c
   create mode 100644 net/smc/smc_loopback.h

#14 lore raw
Quoted 17 lines · Wen Gu
Quoted 14 lines · Wenjia Zhang
Quoted 11 lines · Wen Gu
Quoted 8 lines · Alexandra Winter
Quoted 5 lines · Wen Gu
Quoted 1 line · Niklas Schnelle

On Tue, 2022-12-20 at 11:21 +0800, Wen Gu wrote:

[...]

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

[...]

Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

As described in
SMC protocol (including SMC-D): https://www.ibm.com/support/pages/system/files/inline-files/IBM%20Shared%20Memory%20Communications%20Version%202_2.pdf
the CLC messages provide a list of up to 8 ISM devices to chose from.
So I would hope that we can use the existing protocol.

The challenge will be to define GID (Global Interface ID) and CHID (a fabric ID) in a meaningful way for the new devices.

[...]

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

I agree with your opinion. The existing SMC-Dv2 protocol whose CLC messages include ism_dev[] list can solve the devices negotiation problem. And I am very willing to use the existing protocol, because we all know that the protocol update is a long and complex process.

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

If I understand correctly, SMC-D loopback(dummy) device can coordinate with existing SMC-Dv2 protocol as follows. If there is any mistake, please point out.

# Initialization

- Initialize the loopback device with unique GID [Q-1].

As you point out below [Q-1], the issue here is the uniqueness of the 8 byte GID across all possible SMC peers (they could all offer loopback as a choice).

I was pondering some more about this and I am wondering, whether there is another way to detect that source and target of the CLC proposal are the same Linux instance (without using the content of the CLC message).
One idea could be:
If the netdev used for the CLC proposal is a local interface, then use smcd-loopback. What do you think?

Let me try to summarize my pondering as a base for future discussion:

SMC CLC handshake as described in SMC protocol:
------------------------------------------------
The purpose of the handshake is to efficiently find out whether and how SMC can be used to communicate with a TCP peer (happens over a standard netdevice). For this purpose the following fields are exchanged:

+ UEID – (32 byte) Defined per system (OS instance).
User defined group of systems that should use SMC between each other. If the peer has only different UEIDs: Don’t try to use SMC with this peer.

+ SEID – (32 byte) Defined per system.
Maximum space of systems that are able to use SMC-D between each other. Equals to the machine hardware / first level hypervisor.
(in s390 KVM is at least the second level hipervisor)
Is unique per machine, derived from unique machine ID.
If the peer has a different SEIDs: Don’t try to use SMC-D with this peer.

+ CHID – (2 Bytes) Defined per SMC-D interface.
Fabric ID of this interface. Unique against other fabrics on this machine. Try to find a pair of SMC-D interfaces on the same fabric for the 2 peers to use for this SMC-D connection.

+ GID – (8 bytes) Defined per SMC-D interface.
ID of this interface.
For s390 ISM is actually globally unique. But for SMC-D protocol purpose just needs to be unique within the CHID (fabric).
Use it to identify the chosen interfaces of the 2 peers.

Important usecases:
As the CLC handshake is exchanged via standard netdevice, the 2 peers can be:
- on different machines
- on same machine but different first level guest
- on same KVM host but different KVM guest,
- on same KVM guest (loopback case)

and the proposal list can include any combination of:
- one or more RDMA devices (to use with SMC-R)
- one or more s390 ISM devices on one or more CHIDs
- future: virtio-smcd interface(s)

So for loopback there are 2 options to think about, that can be implemented with today’s SMC-D protocol definition:

(1) Smc-d Loopback is listed in the SMC-D proposal list

Loopback could be one interface in the SMC-Dv2 list of up to 8 CHID/GID pairs proposed. We could use CHID 0xFFFF to point out this is a loopback.
And then only use this GID, if it belongs to both peers. (this is may be a small add-on to today's protocol)
CON: This requires a unique loopback-GID for every OS instance (on this SEID). Must also be unique against any other OS that would ever implement SMC-D loopback, because this could be a handshake with any OS on this SEID.
PRO: It works in all cases where SMC works today.

(2)Find out that both peers of the CLC handshake are actually the same Linux instance, without using the values of the proposal

One idea is that if a local netdev is used for the handshake, then we are in the same instance.
CON: I guess this should work, but may not cover all usecases, where smcd-loopback could be desired. (namespaces, etc..)
I still have some hope that there is another way to detect this somehow… ideas are very welcome.
PRO: This is independent of the SMC protocol. It would be a Linux-only solution, actually even *this* Linux only, future implementations could differ.

Quoted 7 lines
Quoted 6 lines
Quoted 5 lines

- Register the loopback device as SMC-Dv2-capable device with a system_eid whose 24th or 28th byte is non-zero [Q-2], so that this system's smc_ism_v2_capable will be set to TRUE and SMC-Dv2 is available.

The decision point is the VLAN_ID, if it is x1FFF, the device will support V2. i.e. If you can have subnet with VLAN_ID x1FFF, then the SEID is necessary, so that the series or types is non-zero. (*1)

I guess there is some misunderstanding of today's code.
The invalid VLAN_ID 0x1FFF is used to signal to s390 ISM hardware that SMC-Dv2 is used, where the CLC handshake does not have to be between peers on the same IP subnet.
If we cannot set x1FFF, then this is old hardware and we have to use SMC-Dv1 on this machine. (all ISM interfaces are same hardware level).
If we can use SMC-Dv2 then we need to determine the SEID of this machine, so we can use it in the CLC handshake. (if we run on v1 hardware, there is no need to determine the SEID)

Quoted 24 lines

In case there is any misunderstanding between us, I would like to rephrase my [Q-2] question:

int smcd_register_dev(struct smcd_dev *smcd)
{
    <...>
    mutex_lock(&smcd_dev_list.mutex);
    if (list_empty(&smcd_dev_list.list)) {
        u8 *system_eid = NULL;

        smcd->ops->get_system_eid(smcd, &system_eid); if (system_eid[24] != '0' || system_eid[28] != '0') { smc_ism_v2_capable = true; memcpy(smc_ism_v2_system_eid, system_eid, SMC_MAX_EID_LEN);
        }
    }
    <...>
}

It can be inferred from smcd_register_dev() that:

1) The 24th and 28th byte are special and determinate whether smc_ism_v2_capable is true. Besides these, do other bytes of system_eid have hidden meanings that need attention ?

Quoted 8 lines

2) Only when smcd_dev_list is empty, the added smcd_dev will be checked, and its system_eid determinates whether smc_ism_v2_capable is true. Why only the first added device will be checked ?

   If the first added smcd_dev has an system_eid whose 24th and 28th bytes are zero, and the second added smcd_dev has an system_eid whose 24th and 28th bytes are non-zero. Should smc_ism_v2_capable be true, since the second smcd_dev has v2-indicated system_eid ?

This was a rather indirect way to determine smc_ism_v2_capable, which is improved in the ism patches currently under review.

Quoted 36 lines
Quoted 36 lines
Quoted 36 lines

# Proposal

- Find the loopback device from the smcd_dev_list in smc_find_ism_v2_device_clnt();

- Record the SEID, GID and CHID[Q-3] of loopback device in the v2 extension part of CLC proposal message.

# Accept

- Check the GID/CHID list and SEID in CLC proposal message, and find local matched ISM device from smcd_dev_list in smc_find_ism_v2_device_serv(). If both sides of the communication are in the same VM and share the same loopback device, the SEID, GID and CHID will match and loopback device will be chosen [Q-4].

- Record the loopback device's GID/CHID and matched SEID into CLC accept message.

# Confirm

- Confirm the server-selected device (loopback device) accordingto CLC accept messages.

- Record the loopback device's GID/CHID and server-selected SEID in CLC confirm message.

Follow the above process, I supplement a patch based on this RFC in the email attachment. With the attachment patch, SMC-D loopback will switch to use SMC-Dv2 protocol.

And in the above process, there are something I want to consult and discuss, which is marked with '[Q-*]' in the above description.

# [Q-1]:

Quoted 8 lines
Quoted 4 lines
Quoted 3 lines

The GID of loopback device is randomly generated in this RFC patch set, but I will find a way to unique the GID in formal patches. Any suggestions are welcome.

I think the randowmly generated GID is fine in your case, which is equivalent to the IP address.

Since whether the two sides can communicate through the loopback will be judged by whether the gid of their loopback device is equal, the random GID may bring the risk of misjudgment because it may not be unique. But considering this is an RFC, I simply used random GIDs.

I share your concerns about using random 8 byte numbers for all possible instances. Collisions may be unlikely, but if they happen, they cannot be detected and have nasty effects.

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

# [Q-2]:

In Linux implementation, the system_eid of the first registered smcd device will determinate system's smc_ism_v2_capable (see smcd_register_dev()).

See above, this is a rather indirect correlation.

Quoted 12 lines
Quoted 12 lines
Quoted 12 lines

And I wonder that

1) How to define the system_eid? It can be inferred from the code that the 24th and 28th byte are special for SMC-Dv2. So in attachment patch, I define the loopback device SEID as

     static struct smc_lo_systemeid LO_SYSTEM_EID = { .seid_string = "SMC-SYSZ-LOSEID000000000", .serial_number = "1000",
             .type = "1000",
     };

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

     Is there anything else I need to pay attention to?

If you just want to use V2, such defination looks good. e.g. you can use some unique information from "lshw"

OK, thank you.

As mentioned above:
+ SEID – (32 byte) Defined per system.
Maximum space of systems that are able to use SMC-D between each other. Equals to the machine hardware / first level hypervisor.
(in s390 KVM is at least the second level hipervisor)
Is unique per machine, derived from unique machine ID.
If the peer has a different SEIDs: Don’t try to use SMC-D with this peer.

We need to continue to use today's values on s390 architecture for backward compatibility! Other architectures need to also use values that uniquely identify the machine it is running on.

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

2) Seems only the first added smcd device determinate the system smc_ism_v2_capable? If two different smcd devices respectively with v1-indicated and v2-indicated system_eid, will the order in which they are registered affects the result of smc_ism_v2_capable ?

see (*1)

see above: all s390 ISM interfaces on a machine are same hardware level.

Quoted 11 lines
Quoted 9 lines
Quoted 8 lines

# [Q-3]:

In attachment patch, I define a special CHID (0xFFFF) for loopback device, as a kind of 'unassociated ISM CHID' that not associated with any IP (OSA or HiperSockets) interfaces.

What's your opinion about this?

It looks good to me

OK.

This maybe a small add-on to today's protocol, as a special CHID number that is evaluated differently. But IMHO it would fit with the purpose of the VHCID/GID pairs. 0xFFFF cannot appear on today's s390ISM CHIDs. So this could be backwards compatible.

Quoted 20 lines
Quoted 20 lines
Quoted 20 lines

# [Q-4]:

In current Linux implementation, server will select the first successfully initialized device from the candidates as the final selected one in smc_find_ism_v2_device_serv().

for (i = 0; i < matches; i++) {
     ini->smcd_version = SMC_V2;
     ini->is_smcd = true;
     ini->ism_selected = i;
     rc = smc_listen_ism_init(new_smc, ini);
     if (rc) {
         smc_find_ism_store_rc(rc, ini);
         /* try next active ISM device */ continue;
     }
     return; /* matching and usable V2 ISM device found */ }

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

IMHO, maybe candidate devices should have different priorities? For example, the loopback device may be preferred to use if loopback is available.

IMO, I'd prefer such a order: ISM -> loopback -> RoCE Because ISM for SMC-D is our standard user case, not loopback.

OK, will follow this order.

My initial thought would be: loopback -> ISM -> RoCE, just as it is for netdev loopback.
But that only makes sense if loopback performs better than ISM. Can we postpone that decision until we have measurements?

author reviewer
#15 lore raw
Quoted 59 lines · Alexandra Winter
Quoted 17 lines · Wen Gu
Quoted 14 lines · Wenjia Zhang
Quoted 11 lines · Wen Gu
Quoted 8 lines · Alexandra Winter
Quoted 5 lines · Wen Gu
Quoted 1 line · Niklas Schnelle

On Tue, 2022-12-20 at 11:21 +0800, Wen Gu wrote:

[...]

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

[...]

Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

As described in
SMC protocol (including SMC-D): https://www.ibm.com/support/pages/system/files/inline-files/IBM%20Shared%20Memory%20Communications%20Version%202_2.pdf
the CLC messages provide a list of up to 8 ISM devices to chose from.
So I would hope that we can use the existing protocol.

The challenge will be to define GID (Global Interface ID) and CHID (a fabric ID) in a meaningful way for the new devices.

[...]

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

I agree with your opinion. The existing SMC-Dv2 protocol whose CLC messages include ism_dev[] list can solve the devices negotiation problem. And I am very willing to use the existing protocol, because we all know that the protocol update is a long and complex process.

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

If I understand correctly, SMC-D loopback(dummy) device can coordinate with existing SMC-Dv2 protocol as follows. If there is any mistake, please point out.

# Initialization

- Initialize the loopback device with unique GID [Q-1].

As you point out below [Q-1], the issue here is the uniqueness of the 8 byte GID across all possible SMC peers (they could all offer loopback as a choice).

Quoted 7 lines

I was pondering some more about this and I am wondering, whether there is another way to detect that source and target of the CLC proposal are the same Linux instance (without using the content of the CLC message).
One idea could be:
If the netdev used for the CLC proposal is a local interface, then use smcd-loopback. What do you think?

Hi Winter,

Thanks a lot for your suggestions, and forgive me for the delay in replying due to the vacation.

My opinions are below.

Quoted 60 lines

Let me try to summarize my pondering as a base for future discussion:

SMC CLC handshake as described in SMC protocol:
------------------------------------------------
The purpose of the handshake is to efficiently find out whether and how SMC can be used to communicate with a TCP peer (happens over a standard netdevice). For this purpose the following fields are exchanged:

+ UEID – (32 byte) Defined per system (OS instance).
User defined group of systems that should use SMC between each other. If the peer has only different UEIDs: Don’t try to use SMC with this peer.

+ SEID – (32 byte) Defined per system.
Maximum space of systems that are able to use SMC-D between each other. Equals to the machine hardware / first level hypervisor.
(in s390 KVM is at least the second level hipervisor)
Is unique per machine, derived from unique machine ID.
If the peer has a different SEIDs: Don’t try to use SMC-D with this peer.

+ CHID – (2 Bytes) Defined per SMC-D interface.
Fabric ID of this interface. Unique against other fabrics on this machine. Try to find a pair of SMC-D interfaces on the same fabric for the 2 peers to use for this SMC-D connection.

+ GID – (8 bytes) Defined per SMC-D interface.
ID of this interface.
For s390 ISM is actually globally unique. But for SMC-D protocol purpose just needs to be unique within the CHID (fabric).
Use it to identify the chosen interfaces of the 2 peers.

Important usecases:
As the CLC handshake is exchanged via standard netdevice, the 2 peers can be:
- on different machines
- on same machine but different first level guest
- on same KVM host but different KVM guest,
- on same KVM guest (loopback case)

and the proposal list can include any combination of:
- one or more RDMA devices (to use with SMC-R)
- one or more s390 ISM devices on one or more CHIDs
- future: virtio-smcd interface(s)

So for loopback there are 2 options to think about, that can be implemented with today’s SMC-D protocol definition:

(1) Smc-d Loopback is listed in the SMC-D proposal list

Loopback could be one interface in the SMC-Dv2 list of up to 8 CHID/GID pairs proposed. We could use CHID 0xFFFF to point out this is a loopback.
And then only use this GID, if it belongs to both peers. (this is may be a small add-on to today's protocol)
CON: This requires a unique loopback-GID for every OS instance (on this SEID). Must also be unique against any other OS that would ever implement SMC-D loopback, because this could be a handshake with any OS on this SEID.
PRO: It works in all cases where SMC works today.

(2)Find out that both peers of the CLC handshake are actually the same Linux instance, without using the values of the proposal

Quoted 9 lines

One idea is that if a local netdev is used for the handshake, then we are in the same instance.
CON: I guess this should work, but may not cover all usecases, where smcd-loopback could be desired. (namespaces, etc..)
I still have some hope that there is another way to detect this somehow… ideas are very welcome.
PRO: This is independent of the SMC protocol. It would be a Linux-only solution, actually even *this* Linux only, future implementations could differ.

I totally agree with your summary.

I also considered option #2 at the beginning, like through established clcsock's fib to judge whether the CLC proposal target is local. But as you mentioned, it is hard to cover all possible usecases, a particular example is that two containers in the same VM communicate with each other through physical NICs.

So IMHO, it seems hardly to avoid exchanging information (like GID) between each side to confirm both sides are on the same Linux instance. Please correct me if I'm wrong.

For now, I perfer the option #1. But option #2 is still one of our options. All ideas are welcome.

Quoted 8 lines
Quoted 7 lines
Quoted 6 lines
Quoted 5 lines

- Register the loopback device as SMC-Dv2-capable device with a system_eid whose 24th or 28th byte is non-zero [Q-2], so that this system's smc_ism_v2_capable will be set to TRUE and SMC-Dv2 is available.

The decision point is the VLAN_ID, if it is x1FFF, the device will support V2. i.e. If you can have subnet with VLAN_ID x1FFF, then the SEID is necessary, so that the series or types is non-zero. (*1)

Quoted 8 lines

I guess there is some misunderstanding of today's code.
The invalid VLAN_ID 0x1FFF is used to signal to s390 ISM hardware that SMC-Dv2 is used, where the CLC handshake does not have to be between peers on the same IP subnet.
If we cannot set x1FFF, then this is old hardware and we have to use SMC-Dv1 on this machine. (all ISM interfaces are same hardware level).
If we can use SMC-Dv2 then we need to determine the SEID of this machine, so we can use it in the CLC handshake. (if we run on v1 hardware, there is no need to determine the SEID)

Thanks for explanation, now I understand.

Quoted 28 lines
Quoted 28 lines

In case there is any misunderstanding between us, I would like to rephrase my [Q-2] question:

int smcd_register_dev(struct smcd_dev *smcd)
{
    <...>
    mutex_lock(&smcd_dev_list.mutex);
    if (list_empty(&smcd_dev_list.list)) {
        u8 *system_eid = NULL;

        smcd->ops->get_system_eid(smcd, &system_eid); if (system_eid[24] != '0' || system_eid[28] != '0') { smc_ism_v2_capable = true; memcpy(smc_ism_v2_system_eid, system_eid, SMC_MAX_EID_LEN);
        }
    }
    <...>
}

It can be inferred from smcd_register_dev() that:

1) The 24th and 28th byte are special and determinate whether smc_ism_v2_capable is true. Besides these, do other bytes of system_eid have hidden meanings that need attention ?

2) Only when smcd_dev_list is empty, the added smcd_dev will be checked, and its system_eid determinates whether smc_ism_v2_capable is true. Why only the first added device will be checked ?

Quoted 9 lines
Quoted 4 lines

   If the first added smcd_dev has an system_eid whose 24th and 28th bytes are zero, and the second added smcd_dev has an system_eid whose 24th and 28th bytes are non-zero. Should smc_ism_v2_capable be true, since the second smcd_dev has v2-indicated system_eid ?

This was a rather indirect way to determine smc_ism_v2_capable, which is improved in the ism patches currently under review.

OK, thanks.

Quoted 39 lines
Quoted 39 lines
Quoted 39 lines
Quoted 39 lines

# Proposal

- Find the loopback device from the smcd_dev_list in smc_find_ism_v2_device_clnt();

- Record the SEID, GID and CHID[Q-3] of loopback device in the v2 extension part of CLC proposal message.

# Accept

- Check the GID/CHID list and SEID in CLC proposal message, and find local matched ISM device from smcd_dev_list in smc_find_ism_v2_device_serv(). If both sides of the communication are in the same VM and share the same loopback device, the SEID, GID and CHID will match and loopback device will be chosen [Q-4].

- Record the loopback device's GID/CHID and matched SEID into CLC accept message.

# Confirm

- Confirm the server-selected device (loopback device) accordingto CLC accept messages.

- Record the loopback device's GID/CHID and server-selected SEID in CLC confirm message.

Follow the above process, I supplement a patch based on this RFC in the email attachment. With the attachment patch, SMC-D loopback will switch to use SMC-Dv2 protocol.

And in the above process, there are something I want to consult and discuss, which is marked with '[Q-*]' in the above description.

# [Q-1]:

The GID of loopback device is randomly generated in this RFC patch set, but I will find a way to unique the GID in formal patches. Any suggestions are welcome.

Quoted 9 lines
Quoted 5 lines
Quoted 1 line

I think the randowmly generated GID is fine in your case, which is equivalent to the IP address.

Since whether the two sides can communicate through the loopback will be judged by whether the gid of their loopback device is equal, the random GID may bring the risk of misjudgment because it may not be unique. But considering this is an RFC, I simply used random GIDs.

I share your concerns about using random 8 byte numbers for all possible instances. Collisions may be unlikely, but if they happen, they cannot be detected and have nasty effects.

Yes, I'm thinking about it and trying to find a way to avoid collision, or safely fallback if collision happens.

Quoted 37 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

# [Q-2]:

In Linux implementation, the system_eid of the first registered smcd device will determinate system's smc_ism_v2_capable (see smcd_register_dev()).

See above, this is a rather indirect correlation.

Quoted 12 lines
Quoted 12 lines
Quoted 12 lines

And I wonder that

1) How to define the system_eid? It can be inferred from the code that the 24th and 28th byte are special for SMC-Dv2. So in attachment patch, I define the loopback device SEID as

     static struct smc_lo_systemeid LO_SYSTEM_EID = { .seid_string = "SMC-SYSZ-LOSEID000000000", .serial_number = "1000",
             .type = "1000",
     };

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

     Is there anything else I need to pay attention to?

If you just want to use V2, such defination looks good. e.g. you can use some unique information from "lshw"

OK, thank you.

As mentioned above:
+ SEID – (32 byte) Defined per system.
Maximum space of systems that are able to use SMC-D between each other. Equals to the machine hardware / first level hypervisor.
(in s390 KVM is at least the second level hipervisor)
Is unique per machine, derived from unique machine ID.
If the peer has a different SEIDs: Don’t try to use SMC-D with this peer.

Quoted 4 lines

We need to continue to use today's values on s390 architecture for backward compatibility! Other architectures need to also use values that uniquely identify the machine it is running on.

Yes, I agree.

IIUC, on s390 architecture, if machine's ISM hardware is V2 capable (can set VLAN_ID 0x1FFF), then SEID will be set according to unique machine ID (for s390 architecture, it will be the ID returned from get_cpu_id() in arch/s390/include/asm/processor.h). So,

- On s390 architecture, SMC-D loopback device should use the same SEID as ISM devices on the same machine. If SMC-D loopback only supports V2 and machine is not V2 capable, then SMC-D loopback should not be used.
- On the other architecture except s390, we need to find a similar way to generate a machine-unique ID as part of SEID. (perhaps need a unified helper to do so ?)

Quoted 9 lines
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

2) Seems only the first added smcd device determinate the system smc_ism_v2_capable? If two different smcd devices respectively with v1-indicated and v2-indicated system_eid, will the order in which they are registered affects the result of smc_ism_v2_capable ?

see (*1)

see above: all s390 ISM interfaces on a machine are same hardware level.

OK.

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

# [Q-3]:

In attachment patch, I define a special CHID (0xFFFF) for loopback device, as a kind of 'unassociated ISM CHID' that not associated with any IP (OSA or HiperSockets) interfaces.

What's your opinion about this?

Quoted 8 lines
Quoted 3 lines
Quoted 1 line

It looks good to me

OK.

This maybe a small add-on to today's protocol, as a special CHID number that is evaluated differently. But IMHO it would fit with the purpose of the VHCID/GID pairs. 0xFFFF cannot appear on today's s390ISM CHIDs. So this could be backwards compatible.

Thanks, that's a good news.

Quoted 26 lines
Quoted 26 lines
Quoted 25 lines
Quoted 20 lines

# [Q-4]:

In current Linux implementation, server will select the first successfully initialized device from the candidates as the final selected one in smc_find_ism_v2_device_serv().

for (i = 0; i < matches; i++) {
     ini->smcd_version = SMC_V2;
     ini->is_smcd = true;
     ini->ism_selected = i;
     rc = smc_listen_ism_init(new_smc, ini);
     if (rc) {
         smc_find_ism_store_rc(rc, ini);
         /* try next active ISM device */ continue;
     }
     return; /* matching and usable V2 ISM device found */
}

Quoted 3 lines

IMHO, maybe candidate devices should have different priorities? For example, the loopback device may be preferred to use if loopback is available.

IMO, I'd prefer such a order: ISM -> loopback -> RoCE Because ISM for SMC-D is our standard user case, not loopback.

Quoted 7 lines
Quoted 1 line

OK, will follow this order.

My initial thought would be: loopback -> ISM -> RoCE, just as it is for netdev loopback.
But that only makes sense if loopback performs better than ISM. Can we postpone that decision until we have measurements?

Sure, it's very reasonable.

maintainer
#16 lore raw
In reply to Niklas Schnelle
Quoted 143 lines · Niklas Schnelle
Quoted 44 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

 +------------------------------------------+
 |  +-----------+           +-----------+   |
 |  | process A |           | process B |   |
 |  +-----------+           +-----------+   |
 |       ^                        ^         |
 |       |    +---------------+   |         |
 |       |    |   SMC stack   |   |         |
 |       +--->| +-----------+ |<--|         |
 |            | |   dummy   | |             |
 |            | |   device  | |             |
 |            +-+-----------+-+             |
 |                   VM                     |
 +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

Quoted 10 lines
 +----------+                     +----------+
 | socket A |                     | socket B |
 +----------+                     +----------+
       |                               ^
       |         +---------+           |
  regard as      |         | ----------|
  local sndbuf   |  B's    |     regard as
       |         |  RMB    |     local RMB
       |-------> |         |
                 +---------+

Hi Wen Gu,

I maintain the s390 specific PCI support in Linux and would like to provide a bit of background on this. You're surely wondering why we even have a copy in there for our ISM virtual PCI device. To understand why this copy operation exists and why we need to keep it working, one needs a bit of s390 aka mainframe background.

On s390 all (currently supported) native machines have a mandatory machine level hypervisor. All OSs whether z/OS or Linux run either on this machine level hypervisor as so called Logical Partitions (LPARs) or as second/third/… level guests on e.g. a KVM or z/VM hypervisor that in turn runs in an LPAR. Now, in terms of memory this machine level hypervisor sometimes called PR/SM unlike KVM, z/VM, or VMWare is a partitioning hypervisor without paging. This is one of the main reasons for the very-near-native performance of the machine hypervisor as the memory of its guests acts just like native RAM on other systems. It is never paged out and always accessible to IOMMU translated DMA from devices without the need for pinning pages and besides a trivial offset/limit adjustment an LPAR's MMU does the same amount of work as an MMU on a bare metal x86_64/ARM64 box.

It also means however that when SMC-D is used to communicate between LPARs via an ISM device there is no way of mapping the DMBs to the same physical memory as there exists no MMU-like layer spanning partitions that could do such a mapping. Meanwhile for machine level firmware including the ISM virtual PCI device it is still possible to _copy_ memory between different memory partitions. So yeah while I do see the appeal of skipping the memcpy() for loopback or even between guests of a paging hypervisor such as KVM, which can map the DMBs on the same physical memory, we must keep in mind this original use case requiring a copy operation.

Thanks,
Niklas

Quoted 13 lines

# Benchmark Test

* Test environments:
- VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
- SMC sndbuf/RMB size 1MB.

* Test object:
- TCP: run on TCP loopback.
- domain: run on UNIX domain.
- SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
- SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

Quoted 7 lines

1. ipc-benchmark (see [3])

- ./<foo> -c 1000000 -s 100

TCP domain SMC-lo SMC-lo-nocpy
Message
rate (msg/s) 75140 129548(+72.41) 152266(+102.64%) 151914(+102.17%)

Interesting that it does beat UNIX domain sockets. Also, see my below comment for nginx/wrk as this seems very similar.

Quoted 18 lines

2. sockperf

- serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
- clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

TCP SMC-lo SMC-lo-nocpy
Bandwidth(MBps) 4943.359 4936.096(-0.15%) 8239.624(+66.68%)
Latency(us) 6.372 3.359(-47.28%) 3.25(-49.00%)

3. iperf3

- serv: <smc_run> taskset -c <cpu> iperf3 -s
- clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

TCP SMC-lo SMC-lo-nocpy
Bitrate(Gb/s) 40.5 41.4(+2.22%) 76.4(+88.64%)

Quoted 7 lines

4. nginx/wrk

- serv: <smc_run> nginx
- clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

TCP SMC-lo SMC-lo-nocpy
Requests/s 154643.22 220894.03(+42.84%) 226754.3(+46.63%)

Quoted 7 lines

This result is very interesting indeed. So with the much more realistic nginx/wrk workload it seems to copy hurts much less than the iperf3/sockperf would suggest while SMC-D itself seems to help more. I'd hope that this translates to actual applications as well. Maybe this makes SMC-D based loopback interesting even while keeping the copy, at least until we can come up with a sane way to work a no-copy variant into SMC-D?

Yes, SMC-D based loopback shows great advantages over TCP loopback, with or without copy.

The advantage of zero-copy should be observed when we need to transfer a large mount of data. But here in this wrk/nginx case, the test file transferred from server to client is a small file. So we didn't see much gain. If we use a large file(e.g >=1MB file), I think we should observe a much different result.

Thinks!

author reviewer
#17 lore raw
In reply to Wen Gu
Quoted 108 lines · Wen Gu

Hi, all

# Background

As previously mentioned in [1], we (Alibaba Cloud) are trying to use SMC to accelerate TCP applications in cloud environment, improving inter-host or inter-VM communication.

In addition of these, we also found the value of SMC-D in scenario of local inter-process communication, such as accelerate communication between containers within the same host. So this RFC tries to provide a SMC-D loopback solution in such scenario, to bring a significant improvement in latency and throughput compared to TCP loopback.

# Design

This patch set provides a kind of SMC-D loopback solution.

Patch #1/5 and #2/5 provide an SMC-D based dummy device, preparing for the inter-process communication acceleration. Except for loopback acceleration, the dummy device can also meet the requirements mentioned in [2], which is providing a way to test SMC-D logic for broad community without ISM device.

  +------------------------------------------+
  |  +-----------+           +-----------+   |
  |  | process A |           | process B |   |
  |  +-----------+           +-----------+   |
  |       ^                        ^         |
  |       |    +---------------+   |         |
  |       |    |   SMC stack   |   |         |
  |       +--->| +-----------+ |<--|         |
  |            | |   dummy   | |             |
  |            | |   device  | |             |
  |            +-+-----------+-+             |
  |                   VM                     |
  +------------------------------------------+

Patch #3/5, #4/5, #5/5 provides a way to avoid data copy from sndbuf to RMB and improve SMC-D loopback performance. Through extending smcd_ops with two new semantic: attach_dmb and detach_dmb, sender's sndbuf shares the same physical memory region with receiver's RMB. The data copied from userspace to sender's sndbuf directly reaches the receiver's RMB without unnecessary memory copy in the same kernel.

  +----------+                     +----------+
  | socket A |                     | socket B |
  +----------+                     +----------+
        |                               ^
        |         +---------+           |
   regard as      |         | ----------|
   local sndbuf   |  B's    |     regard as
        |         |  RMB    |     local RMB
        |-------> |         |
                  +---------+

# Benchmark Test

* Test environments:
- VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
- SMC sndbuf/RMB size 1MB.

* Test object:
- TCP: run on TCP loopback.
- domain: run on UNIX domain.
- SMC lo: run on SMC loopback device with patch #1/5 ~ #2/5.
- SMC lo-nocpy: run on SMC loopback device with patch #1/5 ~ #5/5.

1. ipc-benchmark (see [3])

- ./<foo> -c 1000000 -s 100

TCP domain SMC-lo SMC-lo-nocpy
Message
rate (msg/s) 75140 129548(+72.41) 152266(+102.64%) 151914(+102.17%)

2. sockperf

- serv: <smc_run> taskset -c <cpu> sockperf sr --tcp
- clnt: <smc_run> taskset -c <cpu> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30

TCP SMC-lo SMC-lo-nocpy
Bandwidth(MBps) 4943.359 4936.096(-0.15%) 8239.624(+66.68%)
Latency(us) 6.372 3.359(-47.28%) 3.25(-49.00%)

3. iperf3

- serv: <smc_run> taskset -c <cpu> iperf3 -s
- clnt: <smc_run> taskset -c <cpu> iperf3 -c 127.0.0.1 -t 15

TCP SMC-lo SMC-lo-nocpy
Bitrate(Gb/s) 40.5 41.4(+2.22%) 76.4(+88.64%)

4. nginx/wrk

- serv: <smc_run> nginx
- clnt: <smc_run> wrk -t 8 -c 500 -d 30 http://127.0.0.1:80

TCP SMC-lo SMC-lo-nocpy
Requests/s 154643.22 220894.03(+42.84%) 226754.3(+46.63%)

# Discussion

1. API between SMC-D and ISM device

As Jan mentioned in [2], IBM are working on placing an API between SMC-D and the ISM device for easier use of different "devices" for SMC-D.

Quoted 8 lines

So, considering that the introduction of attach_dmb or detach_dmb can effectively avoid data copying from sndbuf to RMB and brings obvious throughput advantages in inter-VM or inter-process scenarios, can the attach/detach semantics be taken into consideration when designing the API to make it a standard ISM device behavior?

Maybe our RFC of SMC-D based inter-process acceleration (this one) and inter-VM acceleration (will coming soon, which is the update of [1])

The patch of SMC-D + virtio-ism device is now discussed in virtio community:

https://lists.oasis-open.org/archives/virtio-comment/202212/msg00030.html

Quoted 45 lines

can provide some examples for new API design. And we are very glad to discuss this on the mail list.

2. Way to select different ISM-like devices

With the proposal of SMC-D loopback 'device' (this RFC) and incoming device used for inter-VM acceleration as update of [1], SMC-D has more options to choose from. So we need to consider that how to indicate supported devices, how to determine which one to use, and their priority...

IMHO, this may require an update of CLC message and negotiation mechanism. Again, we are very glad to discuss this with you on the mailing list.

[1] https://lore.kernel.org/netdev/[email protected]/
[2] https://lore.kernel.org/netdev/[email protected]/
[3] https://github.com/goldsborough/ipc-bench

v1->v2
1. Fix some build WARNINGs complained by kernel test rebot

2. Add iperf3 test data.

Wen Gu (5):
net/smc: introduce SMC-D loopback device
net/smc: choose loopback device in SMC-D communication net/smc: add dmb attach and detach interface
net/smc: avoid data copy from sndbuf to peer RMB in SMC-D loopback net/smc: logic of cursors update in SMC-D loopback connections

  include/net/smc.h      |   3 +
  net/smc/Makefile       |   2 +-
  net/smc/af_smc.c       |  88 +++++++++++-
  net/smc/smc_cdc.c      |  59 ++++++--
  net/smc/smc_cdc.h      |   1 +
  net/smc/smc_clc.c      |   4 +-
  net/smc/smc_core.c     |  62 +++++++++
  net/smc/smc_core.h     |   2 +
  net/smc/smc_ism.c      |  39 +++++-
  net/smc/smc_ism.h      |   2 +
  net/smc/smc_loopback.c | 358 +++++++++++++++++++++++++++++++++++++++++++++++++
  net/smc/smc_loopback.h |  63 +++++++++
  12 files changed, 662 insertions(+), 21 deletions(-)

create mode 100644 net/smc/smc_loopback.c
create mode 100644 net/smc/smc_loopback.h