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/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2] https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
--
1.8.3.1
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.
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
---
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(-)
@@ -979,6 +979,28 @@ static int smc_find_ism_device(struct smc_sock *smc, struct smc_init_info *ini)return0;}+/* check if there is a lo device available for this connection. */+staticintsmc_find_lo_device(structsmc_sock*smc,structsmc_init_info*ini)+{+structsmcd_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])+returnSMC_CLC_DECL_NOSMCDDEV;+ini->ism_chid[0]=smc_ism_get_chid(ini->ism_dev[0]);+return0;+}+/* is chid unique for the ism devices that are already determined? */staticboolsmc_find_ism_v2_is_unique_chid(u16chid,structsmc_init_info*ini,intcnt)
@@ -1044,10 +1066,20 @@ static int smc_find_proposal_devices(struct smc_sock *smc,{intrc=0;-/* check if there is an ism device available */+/* TODO:+*Howtoindicatetopeerifismdeviceandloopback+*devicearebothavailable?+*+*TheRFCpatchhasn'tresolvedthis,justsimplyalways+*choosesloopbackdevicefirst,andfallbackifloopback+*communicationisimpossible.+*+*/+/* 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,gotonot_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)-gotonot_found;++/* TODO:+*Howtoknowthatpeerhasbothloopbackandismdevice?+*+*TheRFCpatchhasn'tresolvedthis,simplytriesloopback+*devicefirst,thenismdevice.+*/+/* find available loopback or ism device */+if(smc_find_lo_device(new_smc,ini)){+rc=smc_find_ism_device(new_smc,ini);+if(rc)+gotonot_found;+}+ini->ism_selected=0;rc=smc_listen_ism_init(new_smc,ini);if(!rc)
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.
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
---
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
@@ -0,0 +1,282 @@+// SPDX-License-Identifier: GPL-2.0+/*+*SharedMemoryCommunicationsDirectoverloopbackdevice.+*+*ProvideaSMC-Dloopbackdummydevice.+*+*Copyright(c)2022,AlibabaInc.+*+*Author:WenGu<guwen@linux.alibaba.com>+*TonyLu<tonylu@linux.alibaba.com>+*+*/++#include<linux/device.h>+#include<linux/types.h>+#include<net/smc.h>++#include"smc_cdc.h"+#include"smc_loopback.h"++#define DRV_NAME "smc_lodev"++structlo_dev*lo_dev;++staticstructlo_systemeidLO_SYSTEM_EID={+.seid_string="SMC-SYSZ-LOSEID000000000",+.serial_number="0000",+.type="0000",+};++staticintlo_query_rgid(structsmcd_dev*smcd,u64rgid,u32vid_valid,+u32vid)+{+structlo_dev*ldev=smcd->priv;++/* return local gid */+if(!ldev||rgid!=ldev->lgid)+return-ENETUNREACH;+return0;+}++staticintlo_register_dmb(structsmcd_dev*smcd,structsmcd_dmb*dmb)+{+structlo_dev*ldev=smcd->priv;+structlo_dmb_node*dmb_node;+intsba_idx,rc;++/* check space for new dmb */+for_each_clear_bit(sba_idx,ldev->sba_idx_mask,LODEV_MAX_DMBS){+if(!test_and_set_bit(sba_idx,ldev->sba_idx_mask))+break;+}+if(sba_idx==LODEV_MAX_DMBS)+return-ENOSPC;++dmb_node=kzalloc(sizeof(*dmb_node),GFP_KERNEL);+if(!dmb_node){+rc=-ENOMEM;+gotoerr_bit;+}++dmb_node->sba_idx=sba_idx;+dmb_node->cpu_addr=kzalloc(dmb->dmb_len,GFP_KERNEL|+__GFP_NOWARN|__GFP_NORETRY|+__GFP_NOMEMALLOC);+if(!dmb_node->cpu_addr){+rc=-ENOMEM;+gotoerr_node;+}+dmb_node->len=dmb->dmb_len;++/* TODO: token is random but not exclusive !+*supposetofindtokenindmbhasktable,ifhasthistoken+*already,thengenerateanotherone.+*/+/* add new dmb into hash table */+get_random_bytes(&dmb_node->token,sizeof(dmb_node->token));+write_lock(&ldev->dmb_ht_lock);+hash_add(ldev->dmb_ht,&dmb_node->list,dmb_node->token);+write_unlock(&ldev->dmb_ht_lock);++dmb->sba_idx=dmb_node->sba_idx;+dmb->dmb_tok=dmb_node->token;+dmb->cpu_addr=dmb_node->cpu_addr;+dmb->dma_addr=dmb_node->dma_addr;+dmb->dmb_len=dmb_node->len;++return0;++err_node:+kfree(dmb_node);+err_bit:+clear_bit(sba_idx,ldev->sba_idx_mask);+returnrc;+}++staticintlo_unregister_dmb(structsmcd_dev*smcd,structsmcd_dmb*dmb)+{+structlo_dmb_node*dmb_node=NULL,*tmp_node;+structlo_dev*ldev=smcd->priv;++/* remove dmb from hash table */+write_lock(&ldev->dmb_ht_lock);+hash_for_each_possible(ldev->dmb_ht,tmp_node,list,dmb->dmb_tok){+if(tmp_node->token==dmb->dmb_tok){+dmb_node=tmp_node;+break;+}+}+if(!dmb_node){+write_unlock(&ldev->dmb_ht_lock);+return-EINVAL;+}+hash_del(&dmb_node->list);+write_unlock(&ldev->dmb_ht_lock);++clear_bit(dmb_node->sba_idx,ldev->sba_idx_mask);+kfree(dmb_node->cpu_addr);+kfree(dmb_node);++return0;+}++staticintlo_add_vlan_id(structsmcd_dev*smcd,u64vlan_id)+{+return0;+}++staticintlo_del_vlan_id(structsmcd_dev*smcd,u64vlan_id)+{+return0;+}++staticintlo_set_vlan_required(structsmcd_dev*smcd)+{+return0;+}++staticintlo_reset_vlan_required(structsmcd_dev*smcd)+{+return0;+}++staticintlo_signal_ieq(structsmcd_dev*smcd,u64rgid,u32trigger_irq,+u32event_code,u64info)+{+return0;+}++staticintlo_move_data(structsmcd_dev*smcd,u64dmb_tok,unsignedintidx,+boolsf,unsignedintoffset,void*data,+unsignedintsize)+{+structlo_dmb_node*rmb_node=NULL,*tmp_node;+structlo_dev*ldev=smcd->priv;++read_lock(&ldev->dmb_ht_lock);+hash_for_each_possible(ldev->dmb_ht,tmp_node,list,dmb_tok){+if(tmp_node->token==dmb_tok){+rmb_node=tmp_node;+break;+}+}+if(!rmb_node){+read_unlock(&ldev->dmb_ht_lock);+return-EINVAL;+}+read_unlock(&ldev->dmb_ht_lock);++memcpy((char*)rmb_node->cpu_addr+offset,data,size);++if(sf){+structsmc_connection*conn=+smcd->conn[rmb_node->sba_idx];++if(conn&&!conn->killed)+smcd_cdc_rx_handler(conn);+}+return0;+}++staticu8*lo_get_system_eid(void)+{+return&LO_SYSTEM_EID.seid_string[0];+}++staticu16lo_get_chid(structsmcd_dev*smcd)+{+return0;+}++staticconststructsmcd_opslo_ops={+.query_remote_gid=lo_query_rgid,+.register_dmb=lo_register_dmb,+.unregister_dmb=lo_unregister_dmb,+.add_vlan_id=lo_add_vlan_id,+.del_vlan_id=lo_del_vlan_id,+.set_vlan_required=lo_set_vlan_required,+.reset_vlan_required=lo_reset_vlan_required,+.signal_event=lo_signal_ieq,+.move_data=lo_move_data,+.get_system_eid=lo_get_system_eid,+.get_chid=lo_get_chid,+};++staticintlo_dev_init(structlo_dev*ldev)+{+structsmcd_dev*smcd=ldev->smcd;++/* smcd related */+smcd->is_loopback=1;+smcd->priv=ldev;+get_random_bytes(&smcd->local_gid,sizeof(smcd->local_gid));++/* ldev related */+/* TODO: lgid is random but not exclusive !+*/+ldev->lgid=smcd->local_gid;+rwlock_init(&ldev->dmb_ht_lock);+hash_init(ldev->dmb_ht);++returnsmcd_register_dev(smcd);+}++staticintlo_dev_probe(void)+{+structlo_dev*ldev;+intret;++ldev=kzalloc(sizeof(*ldev),GFP_KERNEL);+if(!ldev)+return-ENOMEM;++ldev->smcd=smcd_alloc_dev(NULL,"smcd-loopback-dev",+&lo_ops,LODEV_MAX_DMBS);+if(!ldev->smcd){+ret=-ENOMEM;+gotoerr_ldev;+}++ret=lo_dev_init(ldev);+if(ret)+gotoerr_smcd;++lo_dev=ldev;+return0;++err_smcd:+smcd_free_dev(ldev->smcd);+err_ldev:+kfree(ldev);+returnret;+}++staticvoidlo_dev_exit(structlo_dev*ldev)+{+smcd_unregister_dev(ldev->smcd);+}++staticvoidlo_dev_remove(void)+{+if(!lo_dev)+return;++lo_dev_exit(lo_dev);+smcd_free_dev(lo_dev->smcd);+kfree(lo_dev);+}++intsmc_loopback_init(void)+{+/* TODO: now lo_dev is a global device shared by+*thewholekernel,andcan'tbereferredtoby+*smc-toolscommand'smcddev'.Needtobeimproved.+*/+returnlo_dev_probe();+}++voidsmc_loopback_exit(void)+{+lo_dev_remove();+}
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.
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
---
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(+)
@@ -202,6 +202,42 @@ int smc_ism_register_dmb(struct smc_link_group *lgr, int dmb_len,returnrc;}+intsmc_ism_attach_dmb(structsmcd_dev*dev,u64token,+structsmc_buf_desc*dmb_desc)+{+structsmcd_dmbdmb;+intrc=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;+}+returnrc;+}++intsmc_ism_detach_dmb(structsmcd_dev*dev,u64token)+{+if(!dev->is_loopback)+return-EINVAL;+if(!dev->ops->detach_dmb)+return-EINVAL;++returndev->ops->detach_dmb(dev,token);+}+staticintsmc_nl_handle_smcd_dev(structsmcd_dev*smcd,structsk_buff*skb,structnetlink_callback*cb)
@@ -38,6 +38,8 @@ struct smc_ism_vlanid { /* VLAN id set on ISM device */intsmc_ism_register_dmb(structsmc_link_group*lgr,intbuf_size,structsmc_buf_desc*dmb_desc);intsmc_ism_unregister_dmb(structsmcd_dev*dev,structsmc_buf_desc*dmb_desc);+intsmc_ism_attach_dmb(structsmcd_dev*dev,u64token,structsmc_buf_desc*dmb_desc);+intsmc_ism_detach_dmb(structsmcd_dev*dev,u64token);intsmc_ism_signal_shutdown(structsmc_link_group*lgr);voidsmc_ism_get_system_eid(u8**eid);u16smc_ism_get_chid(structsmcd_dev*dev);
@@ -68,6 +68,7 @@ static int lo_register_dmb(struct smcd_dev *smcd, struct smcd_dmb *dmb)gotoerr_node;}dmb_node->len=dmb->dmb_len;+refcount_set(&dmb_node->refcnt,1);/* TODO: token is random but not exclusive !*supposetofindtokenindmbhasktable,ifhasthistoken
@@ -115,9 +117,69 @@ static int lo_unregister_dmb(struct smcd_dev *smcd, struct smcd_dmb *dmb)write_unlock(&ldev->dmb_ht_lock);clear_bit(dmb_node->sba_idx,ldev->sba_idx_mask);++/* wait for dmb refcnt equal to 0 */+if(!refcount_dec_and_test(&dmb_node->refcnt))+wait_event(ldev->dmbs_release,!refcount_read(&dmb_node->refcnt));kfree(dmb_node->cpu_addr);kfree(dmb_node);+if(atomic_dec_and_test(&ldev->dmb_cnt))+wake_up(&ldev->ldev_release);++return0;+}++staticintlo_attach_dmb(structsmcd_dev*smcd,structsmcd_dmb*dmb)+{+structlo_dmb_node*dmb_node=NULL,*tmp_node;+structlo_dev*ldev=smcd->priv;++/* find dmb_node according to dmb->dmb_tok */+read_lock(&ldev->dmb_ht_lock);+hash_for_each_possible(ldev->dmb_ht,tmp_node,list,dmb->dmb_tok){+if(tmp_node->token==dmb->dmb_tok){+dmb_node=tmp_node;+break;+}+}+if(!dmb_node){+read_unlock(&ldev->dmb_ht_lock);+return-EINVAL;+}+read_unlock(&ldev->dmb_ht_lock);+refcount_inc(&dmb_node->refcnt);++/* provide dmb information */+dmb->sba_idx=dmb_node->sba_idx;+dmb->dmb_tok=dmb_node->token;+dmb->cpu_addr=dmb_node->cpu_addr;+dmb->dma_addr=dmb_node->dma_addr;+dmb->dmb_len=dmb_node->len;+return0;+}++staticintlo_detach_dmb(structsmcd_dev*smcd,u64token)+{+structlo_dmb_node*dmb_node=NULL,*tmp_node;+structlo_dev*ldev=smcd->priv;++/* find dmb_node according to dmb->dmb_tok */+read_lock(&ldev->dmb_ht_lock);+hash_for_each_possible(ldev->dmb_ht,tmp_node,list,token){+if(tmp_node->token==token){+dmb_node=tmp_node;+break;+}+}+if(!dmb_node){+read_unlock(&ldev->dmb_ht_lock);+return-EINVAL;+}+read_unlock(&ldev->dmb_ht_lock);++if(refcount_dec_and_test(&dmb_node->refcnt))+wake_up_all(&ldev->dmbs_release);return0;}
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.
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
---
net/smc/smc_cdc.c | 53 +++++++++++++++++++++++++++++++++++++++-----------
net/smc/smc_loopback.c | 7 +++++++
2 files changed, 49 insertions(+), 11 deletions(-)
@@ -253,17 +253,26 @@ int smcd_cdc_msg_send(struct smc_connection *conn)returnrc;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:+*Forsmcdloopbackdevice:+*+*Don'tupdatethefin_cursandsndbuf_spacehere.+*Updatefin_curswhenpeerconsumesthedatainRMB.+*/-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);+}returnrc;}
@@ -216,6 +216,13 @@ static int lo_move_data(struct smcd_dev *smcd, u64 dmb_tok, unsigned int idx,structlo_dmb_node*rmb_node=NULL,*tmp_node;structlo_dev*ldev=smcd->priv;+if(!sf){+/* no need to move data.+*sndbufisequaltopeerrmb.+*/+return0;+}+read_lock(&ldev->dmb_ht_lock);hash_for_each_possible(ldev->dmb_ht,tmp_node,list,dmb_tok){if(tmp_node->token==dmb_tok){
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.
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
---
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(-)
@@ -1073,7 +1073,6 @@ static int smc_find_proposal_devices(struct smc_sock *smc,*TheRFCpatchhasn'tresolvedthis,justsimplyalways*choosesloopbackdevicefirst,andfallbackifloopback*communicationisimpossible.-**//* 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+*connsabovesmcdloopbackdevonlycreatetheirrmbs.+*theirsndbufsare'maps'ofpeerrmbs.+*/+if(smc->conn.lgr->smcd->is_loopback){+rc=smcd_buf_attach(&smc->conn);+if(rc)+gotoconnect_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+*connsabovesmcdloopbackdevonlycreatetheirrmbs.+*theirsndbufsare'maps'ofpeerrmbs.+*/+if(ini->is_smcd&&new_smc->conn.lgr->smcd->is_loopback){+rc=smcd_buf_attach(&new_smc->conn);+if(rc)+gotoout_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);gotoout_free;
@@ -2423,6 +2427,14 @@ int smc_buf_create(struct smc_sock *smc, bool is_smcd){intrc;+if(is_smcd&&smc->conn.lgr->smcd->is_loopback){+/* Conns above smcd loopback device only create and maintain+*theirRMBs.ThesndbufswillbeattachedtopeerRMBsonce+*gettingthetokens.+*/+return__smc_buf_create(smc,is_smcd,true);+}+/* create send buffer */rc=__smc_buf_create(smc,is_smcd,false);if(rc)
@@ -2439,6 +2451,56 @@ int smc_buf_create(struct smc_sock *smc, bool is_smcd)returnrc;}+/* for smcd loopback conns, attach local sndbuf to peer RMB.+*ThedatacopytosndbufisequaltodatacopytopeerRMB.+*/+intsmcd_buf_attach(structsmc_connection*conn)+{+structsmcd_dev*smcd=conn->lgr->smcd;+u64peer_token=conn->peer_token;+structsmc_buf_desc*buf_desc;+intrc;++buf_desc=kzalloc(sizeof(*buf_desc),GFP_KERNEL);+if(!buf_desc)+return-ENOMEM;+rc=smc_ism_attach_dmb(smcd,peer_token,buf_desc);+if(rc){+rc=SMC_CLC_DECL_ERR_RTOK;+gotofree;+}++/* attach local sndbuf to peer RMB.+*refertolocalsndbufisequaltorefertopeerRMB.+*/+/* align with peer rmb */+buf_desc->cpu_addr=(u8*)buf_desc->cpu_addr+sizeof(structsmcd_cdc_msg);+buf_desc->len-=sizeof(structsmcd_cdc_msg);+conn->sndbuf_desc=buf_desc;+conn->sndbuf_desc->used=1;+//smc->sk.sk_sndbuf = 2 * (smc->conn->sndbuf_desc->len);+atomic_set(&conn->sndbuf_space,conn->sndbuf_desc->len);+return0;++free:+kfree(buf_desc);+returnrc;+}++voidsmcd_buf_detach(structsmc_connection*conn)+{+structsmcd_dev*smcd=conn->lgr->smcd;+u64peer_token=conn->peer_token;++if(!conn->sndbuf_desc)+return;++smc_ism_detach_dmb(smcd,peer_token);++kfree(conn->sndbuf_desc);+conn->sndbuf_desc=NULL;+}+staticinlineintsmc_rmb_reserve_rtoken_idx(structsmc_link_group*lgr){inti;
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
|-------> | |
+---------+
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
# 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%)
Interesting that it does beat UNIX domain sockets. Also, see my below
comment for nginx/wrk as this seems very similar.
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?
# 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?
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.
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...
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.
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/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2] https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
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
|-------> | |
+---------+
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
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
# 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%)
Interesting that it does beat UNIX domain sockets. Also, see my below
comment for nginx/wrk as this seems very similar.
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
# 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?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
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
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...
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
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/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2] https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
From: Dust Li <dust.li@linux.alibaba.com> Date: 2022-12-26 10:46:26
On Tue, Dec 20, 2022 at 03:02:45PM +0100, Niklas Schnelle wrote:
On Tue, 2022-12-20 at 11:21 +0800, Wen Gu wrote:
quoted
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
|-------> | |
+---------+
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
# 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%)
Interesting that it does beat UNIX domain sockets. Also, see my below
comment for nginx/wrk as this seems very similar.
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!
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/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2] https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
From: Alexandra Winter <wintera@linux.ibm.com> Date: 2023-01-04 16:10:54
On 21.12.22 14:14, Wen Gu wrote:
On 2022/12/20 22:02, Niklas Schnelle wrote:
quoted
On Tue, 2022-12-20 at 11:21 +0800, Wen Gu wrote:
quoted
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
|-------> | |
+---------+
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
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
quoted
# 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%)
Interesting that it does beat UNIX domain sockets. Also, see my below
comment for nginx/wrk as this seems very similar.
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
quoted
# 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?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
quoted
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
quoted
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...
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
quoted
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
quoted
[1] https://lore.kernel.org/netdev/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2] https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
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
|-------> | |
+---------+
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
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
quoted
# 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%)
Interesting that it does beat UNIX domain sockets. Also, see my below
comment for nginx/wrk as this seems very similar.
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
quoted
# 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?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
quoted
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
quoted
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...
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
quoted
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.
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
quoted
quoted
[1] https://lore.kernel.org/netdev/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2] https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
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
|-------> | |
+---------+
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
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
quoted
# 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%)
Interesting that it does beat UNIX domain sockets. Also, see my below
comment for nginx/wrk as this seems very similar.
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
quoted
# 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?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
quoted
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
quoted
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...
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
quoted
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.
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.
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)
# 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.
I think the randowmly generated GID is fine in your case, which is
equivalent to the IP address.
# [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?
If you just want to use V2, such defination looks good.
e.g. you can use some unique information from "lshw"
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)
# [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
# [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.
IMO, I'd prefer such a order: ISM -> loopback -> RoCE
Because ISM for SMC-D is our standard user case, not loopback.
Best Regards,
Wen Gu
quoted
quoted
quoted
quoted
[1]
https://lore.kernel.org/netdev/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2]
https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
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
|-------> | |
+---------+
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
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
quoted
# 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%)
Interesting that it does beat UNIX domain sockets. Also, see my below
comment for nginx/wrk as this seems very similar.
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
quoted
# 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?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
quoted
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
quoted
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...
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
quoted
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.
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
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.
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
# 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.
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
# [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?
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
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
# [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
# [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.
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
Best Regards,
Wen Gu
quoted
quoted
quoted
quoted
[1] https://lore.kernel.org/netdev/20220720170048.20806-1-tonylu@linux.alibaba.com/
[2] https://lore.kernel.org/netdev/35d14144-28f7-6129-d6d3-ba16dae7a646@linux.ibm.com/
[3] https://github.com/goldsborough/ipc-bench
v1->v2
1. Fix some build WARNINGs complained by kernel test rebot
Reported-by: kernel test robot [off-list ref]
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
From: Alexandra Winter <wintera@linux.ibm.com> Date: 2023-01-19 13:17:41
On 18.01.23 13:15, Wen Gu wrote:
On 2023/1/16 19:01, Wenjia Zhang wrote:
quoted
On 12.01.23 13:12, Wen Gu wrote:
quoted
On 2023/1/5 00:09, Alexandra Winter wrote:
quoted
On 21.12.22 14:14, Wen Gu wrote:
quoted
On 2022/12/20 22:02, Niklas Schnelle wrote:
quoted
On Tue, 2022-12-20 at 11:21 +0800, Wen Gu wrote:
[...]
quoted
quoted
quoted
quoted
quoted
quoted
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
quoted
quoted
quoted
quoted
quoted
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.
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].
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
quoted
- 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)
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 ?
This was a rather indirect way to determine smc_ism_v2_capable,
which is improved in the ism patches currently under review.
quoted
quoted
# 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.
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
quoted
# [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
quoted
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?
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
quoted
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
quoted
# [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
quoted
# [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.
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?
From: Alexandra Winter <wintera@linux.ibm.com> Date: 2023-01-19 16:25:30
On 20.12.22 04:21, Wen Gu wrote:
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.
the loopback device can be used as a dummy device to
test SMC-D logic for the broad community.
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
---
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.
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.
[...]
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
quoted
quoted
quoted
quoted
quoted
quoted
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.
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].
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?
Hi Winter,
Thanks a lot for your suggestions, and forgive me for the delay in replying due
to the vacation.
My opinions are below.
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.
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
quoted
quoted
- 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)
Thanks for explanation, now I understand.
quoted
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 ?
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
quoted
quoted
# 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.
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
quoted
quoted
# [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
quoted
quoted
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?
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.
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
quoted
quoted
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
quoted
quoted
# [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.
Thanks, that's a good news.
quoted
quoted
quoted
# [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.
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?
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.
Thanks, will use words consistent with this.
quoted
the loopback device can be used as a dummy device to
test SMC-D logic for the broad community.
Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
---
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.
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.
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.