This patch set acts as the second part of the new version of [1] (The first
part can be referred from [2]), the updated things of this version are listed
at the end.
- Background
SMC-D is now used in IBM z with ISM function to optimize network interconnect
for intra-CPC communications. Inspired by this, we try to make SMC-D available
on the non-s390 architecture through a software-implemented Emulated-ISM device,
that is the loopback-ism device here, to accelerate inter-process or
inter-containers communication within the same OS instance.
- Design
This patch set includes 3 parts:
- Patch #1: some prepare work for loopback-ism.
- Patch #2-#7: implement loopback-ism device and adapt SMC-D for it.
loopback-ism now serves only SMC and no userspace interfaces exposed.
- Patch #8-#11: memory copy optimization for intra-OS scenario.
The loopback-ism device is designed as an ISMv2 device and not be limited to
a specific net namespace, ends of both inter-process connection (1/1' in diagram
below) or inter-container connection (2/2' in diagram below) can find the same
available loopback-ism and choose it during the CLC handshake.
Container 1 (ns1) Container 2 (ns2)
+-----------------------------------------+ +-------------------------+
| +-------+ +-------+ +-------+ | | +-------+ |
| | App A | | App B | | App C | | | | App D |<-+ |
| +-------+ +---^---+ +-------+ | | +-------+ |(2') |
| |127.0.0.1 (1')| |192.168.0.11 192.168.0.12| |
| (1)| +--------+ | +--------+ |(2) | | +--------+ +--------+ |
| `-->| lo |-` | eth0 |<-` | | | lo | | eth0 | |
+---------+--|---^-+---+-----|--+---------+ +-+--------+---+-^------+-+
| | | |
Kernel | | | |
+----+-------v---+-----------v----------------------------------+---+----+
| | TCP | |
| | | |
| +--------------------------------------------------------------+ |
| |
| +--------------+ |
| | smc loopback | |
+---------------------------+--------------+-----------------------------+
loopback-ism device creates DMBs (shared memory) for each connection peer.
Since data transfer occurs within the same kernel, the sndbuf of each peer
is only a descriptor and point to the same memory region as peer DMB, so that
the data copy from sndbuf to peer DMB can be avoided in loopback-ism case.
Container 1 (ns1) Container 2 (ns2)
+-----------------------------------------+ +-------------------------+
| +-------+ | | +-------+ |
| | App C |-----+ | | | App D | |
| +-------+ | | | +-^-----+ |
| | | | | |
| (2) | | | (2') | |
| | | | | |
+---------------|-------------------------+ +----------|--------------+
| |
Kernel | |
+---------------|-----------------------------------------|--------------+
| +--------+ +--v-----+ +--------+ +--------+ |
| |dmb_desc| |snd_desc| |dmb_desc| |snd_desc| |
| +-----|--+ +--|-----+ +-----|--+ +--------+ |
| +-----|--+ | +-----|--+ |
| | DMB C | +---------------------------------| DMB D | |
| +--------+ +--------+ |
| |
| +--------------+ |
| | smc loopback | |
+---------------------------+--------------+-----------------------------+
- Benchmark Test
* Test environments:
- VM with Intel Xeon Platinum 8 core 2.50GHz, 16 GiB mem.
- SMC sndbuf/DMB size 1MB.
* Test object:
- TCP: run on TCP loopback.
- SMC lo: run on SMC loopback-ism.
1. ipc-benchmark (see [3])
- ./<foo> -c 1000000 -s 100
TCP SMC-lo
Message
rate (msg/s) 79693 148236(+86.01%)
2. sockperf
- serv: <smc_run> sockperf sr --tcp
- clnt: <smc_run> sockperf { tp | pp } --tcp --msg-size={ 64000 for tp | 14 for pp } -i 127.0.0.1 -t 30
TCP SMC-lo
Bandwidth(MBps) 4815.18 8061.77(+67.42%)
Latency(us) 6.176 3.449(-44.15%)
3. nginx/wrk
- serv: <smc_run> nginx
- clnt: <smc_run> wrk -t 8 -c 1000 -d 30 http://127.0.0.1:80
TCP SMC-lo
Requests/s 196555.02 263270.95(+33.94%)
4. redis-benchmark
- serv: <smc_run> redis-server
- clnt: <smc_run> redis-benchmark -h 127.0.0.1 -q -t set,get -n 400000 -c 200 -d 1024
TCP SMC-lo
GET(Requests/s) 88711.47 120048.02(+35.32%)
SET(Requests/s) 89465.44 123152.71(+37.65%)
Change log:
v6->RFC v5
- Patch #2: make the use of CONFIG_SMC_LO cleaner.
- Patch #5: mark some smcd_ops that loopback-ism doesn't support as
optional and check for the support when they are called.
- Patch #7: keep loopback-ism at the beginning of the SMC-D device list.
- Some expression changes in commit logs and comments.
RFC v5->RFC v4:
Link: https://lore.kernel.org/netdev/20240324135522.108564-1-guwen@xxxxxxxxxxxxxxxxx/
- Patch #2: minor changes in description of config SMC_LO and comments.
- Patch #10: minor changes in comments and if(smc_ism_support_dmb_nocopy())
check in smcd_cdc_msg_send().
- Patch #3: change smc_lo_generate_id() to smc_lo_generate_ids() and SMC_LO_CHID
to SMC_LO_RESERVED_CHID.
- Patch #5: memcpy while holding the ldev->dmb_ht_lock.
- Some expression changes in commit logs.
RFC v4->v3:
Link: https://lore.kernel.org/netdev/20240317100545.96663-1-guwen@xxxxxxxxxxxxxxxxx/
- The merge window of v6.9 is open, so post this series as an RFC.
- Patch #6: since some information fed back by smc_nl_handle_smcd_dev() dose
not apply to Emulated-ISM (including loopback-ism here), loopback-ism is
not exposed through smc netlink for the time being. we may refactor this
part when smc netlink interface is updated.
v3->v2:
Link: https://lore.kernel.org/netdev/20240312142743.41406-1-guwen@xxxxxxxxxxxxxxxxx/
- Patch #11: use tasklet_schedule(&conn->rx_tsklet) instead of smcd_cdc_rx_handler()
to avoid possible recursive locking of conn->send_lock and use {read|write}_lock_bh()
to acquire dmb_ht_lock.
v2->v1:
Link: https://lore.kernel.org/netdev/20240307095536.29648-1-guwen@xxxxxxxxxxxxxxxxx/
- All the patches: changed the term virtual-ISM to Emulated-ISM as defined by SMCv2.1.
- Patch #3: optimized the description of SMC_LO config. Avoid exposing loopback-ism
to sysfs and remove all the knobs until future definition clear.
- Patch #3: try to make lockdep happy by using read_lock_bh() in smc_lo_move_data().
- Patch #6: defaultly use physical contiguous DMB buffers.
- Patch #11: defaultly enable DMB no-copy for loopback-ism and free the DMB in
unregister_dmb or detach_dmb when dmb_node->refcnt reaches 0, instead of using
wait_event to keep waiting in unregister_dmb.
v1->RFC:
Link: https://lore.kernel.org/netdev/20240111120036.109903-1-guwen@xxxxxxxxxxxxxxxxx/
- Patch #9: merge rx_bytes and tx_bytes as xfer_bytes statistics:
/sys/devices/virtual/smc/loopback-ism/xfer_bytes
- Patch #10: add support_dmb_nocopy operation to check if SMC-D device supports
merging sndbuf with peer DMB.
- Patch #13 & #14: introduce loopback-ism device control of DMB memory type and
control of whether to merge sndbuf and DMB. They can be respectively set by:
/sys/devices/virtual/smc/loopback-ism/dmb_type
/sys/devices/virtual/smc/loopback-ism/dmb_copy
The motivation for these two control is that a performance bottleneck was
found when using vzalloced DMB and sndbuf is merged with DMB, and there are
many CPUs and CONFIG_HARDENED_USERCOPY is set [4]. The bottleneck is caused
by the lock contention in vmap_area_lock [5] which is involved in memcpy_from_msg()
or memcpy_to_msg(). Currently, Uladzislau Rezki is working on mitigating the
vmap lock contention [6]. It has significant effects, but using virtual memory
still has additional overhead compared to using physical memory.
So this new version provides controls of dmb_type and dmb_copy to suit
different scenarios.
- Some minor changes and comments improvements.
RFC->old version([1]):
Link: https://lore.kernel.org/netdev/1702214654-32069-1-git-send-email-guwen@xxxxxxxxxxxxxxxxx/
- Patch #1: improve the loopback-ism dump, it shows as follows now:
# smcd d
FID Type PCI-ID PCHID InUse #LGs PNET-ID
0000 0 loopback-ism ffff No 0
- Patch #3: introduce the smc_ism_set_v2_capable() helper and set
smc_ism_v2_capable when ISMv2 or virtual ISM is registered,
regardless of whether there is already a device in smcd device list.
- Patch #3: loopback-ism will be added into /sys/devices/virtual/smc/loopback-ism/.
- Patch #8: introduce the runtime switch /sys/devices/virtual/smc/loopback-ism/active
to activate or deactivate the loopback-ism.
- Patch #9: introduce the statistics of loopback-ism by
/sys/devices/virtual/smc/loopback-ism/{{tx|rx}_tytes|dmbs_cnt}.
- Some minor changes and comments improvements.
[1] https://lore.kernel.org/netdev/1695568613-125057-1-git-send-email-guwen@xxxxxxxxxxxxxxxxx/
[2] https://lore.kernel.org/netdev/20231219142616.80697-1-guwen@xxxxxxxxxxxxxxxxx/
[3] https://github.com/goldsborough/ipc-bench
[4] https://lore.kernel.org/all/3189e342-c38f-6076-b730-19a6efd732a5@xxxxxxxxxxxxxxxxx/
[5] https://lore.kernel.org/all/238e63cd-e0e8-4fbf-852f-bc4d5bc35d5a@xxxxxxxxxxxxxxxxx/
[6] https://lore.kernel.org/all/20240102184633.748113-1-urezki@xxxxxxxxx/
Wen Gu (11):
net/smc: decouple ism_client from SMC-D DMB registration
net/smc: introduce loopback-ism for SMC intra-OS shortcut
net/smc: implement ID-related operations of loopback-ism
net/smc: implement DMB-related operations of loopback-ism
net/smc: mark optional smcd_ops and check for support when called
net/smc: ignore loopback-ism when dumping SMC-D devices
net/smc: register loopback-ism into SMC-D device list
net/smc: add operations to merge sndbuf with peer DMB
net/smc: {at|de}tach sndbuf to peer DMB if supported
net/smc: adapt cursor update when sndbuf and peer DMB are merged
net/smc: implement DMB-merged operations of loopback-ism
drivers/s390/net/ism_drv.c | 2 +-
include/net/smc.h | 21 +-
net/smc/Kconfig | 13 ++
net/smc/Makefile | 1 +
net/smc/af_smc.c | 28 ++-
net/smc/smc_cdc.c | 34 ++-
net/smc/smc_core.c | 61 +++++-
net/smc/smc_core.h | 1 +
net/smc/smc_ism.c | 88 ++++++--
net/smc/smc_ism.h | 10 +
net/smc/smc_loopback.c | 427 +++++++++++++++++++++++++++++++++++++
net/smc/smc_loopback.h | 62 ++++++
12 files changed, 721 insertions(+), 27 deletions(-)
create mode 100644 net/smc/smc_loopback.c
create mode 100644 net/smc/smc_loopback.h
--
2.32.0.3.g01195cf9f
