| Message ID | 20240307095536.29648-1-guwen@linux.alibaba.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | net/smc: SMC intra-OS shortcut with loopback-ism | expand |
On 07/03/2024 10:55, Wen Gu wrote: > 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. Hi Wen Gu, thanks for the updated patch. Please givr me some time to review and test the latest version. Thanks - Jan > > - 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-#2: some prepare work for loopback-ism. > - Patch #3-#7: implement loopback-ism device. Noted that loopback-ism now > serves only SMC and no userspace interface exposed. > - Patch #10-#15: 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) 79287 148946(+87.86%) > > 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 > Bandwidth(MBps) 5053.47 8195.07(+62.17%) > Latency(us) 6.108 3.404(-44.27%) > > 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 178187.56 247970.92(+39.16%) > > 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) 89067.02 123877.36(+39.08%) > SET(Requests/s) 87700.07 131319.77(+49.73%) > > > Change log: > > v2->v1: > - 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: > - 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@linux.alibaba.com/ > - 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@linux.alibaba.com/ > [2] https://lore.kernel.org/netdev/20231219142616.80697-1-guwen@linux.alibaba.com/ > [3] https://github.com/goldsborough/ipc-bench > [4] https://lore.kernel.org/all/3189e342-c38f-6076-b730-19a6efd732a5@linux.alibaba.com/ > [5] https://lore.kernel.org/all/238e63cd-e0e8-4fbf-852f-bc4d5bc35d5a@linux.alibaba.com/ > [6] https://lore.kernel.org/all/20240102184633.748113-1-urezki@gmail.com/ > > Wen Gu (11): > net/smc: adapt SMC-D device dump for Emulated-ISM > 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 some unsupported operations of loopback-ism > net/smc: implement DMB-related operations of loopback-ism > net/smc: register loopback-ism into SMC-D device list > net/smc: add operations to merge sndbuf with peer DMB > net/smc: attach or detach ghost sndbuf to peer DMB > 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 | 7 +- > net/smc/Kconfig | 13 + > net/smc/Makefile | 2 +- > net/smc/af_smc.c | 28 ++- > net/smc/smc_cdc.c | 58 ++++- > net/smc/smc_cdc.h | 1 + > net/smc/smc_core.c | 61 ++++- > net/smc/smc_core.h | 1 + > net/smc/smc_ism.c | 71 +++++- > net/smc/smc_ism.h | 5 + > net/smc/smc_loopback.c | 469 +++++++++++++++++++++++++++++++++++++ > net/smc/smc_loopback.h | 52 ++++ > 13 files changed, 741 insertions(+), 29 deletions(-) > create mode 100644 net/smc/smc_loopback.c > create mode 100644 net/smc/smc_loopback.h >
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-#2: some prepare work for loopback-ism. - Patch #3-#7: implement loopback-ism device. Noted that loopback-ism now serves only SMC and no userspace interface exposed. - Patch #10-#15: 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) 79287 148946(+87.86%) 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 Bandwidth(MBps) 5053.47 8195.07(+62.17%) Latency(us) 6.108 3.404(-44.27%) 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 178187.56 247970.92(+39.16%) 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) 89067.02 123877.36(+39.08%) SET(Requests/s) 87700.07 131319.77(+49.73%) Change log: v2->v1: - 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: - 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@linux.alibaba.com/ - 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@linux.alibaba.com/ [2] https://lore.kernel.org/netdev/20231219142616.80697-1-guwen@linux.alibaba.com/ [3] https://github.com/goldsborough/ipc-bench [4] https://lore.kernel.org/all/3189e342-c38f-6076-b730-19a6efd732a5@linux.alibaba.com/ [5] https://lore.kernel.org/all/238e63cd-e0e8-4fbf-852f-bc4d5bc35d5a@linux.alibaba.com/ [6] https://lore.kernel.org/all/20240102184633.748113-1-urezki@gmail.com/ Wen Gu (11): net/smc: adapt SMC-D device dump for Emulated-ISM 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 some unsupported operations of loopback-ism net/smc: implement DMB-related operations of loopback-ism net/smc: register loopback-ism into SMC-D device list net/smc: add operations to merge sndbuf with peer DMB net/smc: attach or detach ghost sndbuf to peer DMB 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 | 7 +- net/smc/Kconfig | 13 + net/smc/Makefile | 2 +- net/smc/af_smc.c | 28 ++- net/smc/smc_cdc.c | 58 ++++- net/smc/smc_cdc.h | 1 + net/smc/smc_core.c | 61 ++++- net/smc/smc_core.h | 1 + net/smc/smc_ism.c | 71 +++++- net/smc/smc_ism.h | 5 + net/smc/smc_loopback.c | 469 +++++++++++++++++++++++++++++++++++++ net/smc/smc_loopback.h | 52 ++++ 13 files changed, 741 insertions(+), 29 deletions(-) create mode 100644 net/smc/smc_loopback.c create mode 100644 net/smc/smc_loopback.h