dibs - Direct Internal Buffer Sharing
From: | Alexandra Winter <wintera-AT-linux.ibm.com> | |
To: | David Miller <davem-AT-davemloft.net>, Jakub Kicinski <kuba-AT-kernel.org>, Paolo Abeni <pabeni-AT-redhat.com>, Eric Dumazet <edumazet-AT-google.com>, Andrew Lunn <andrew+netdev-AT-lunn.ch>, "D. Wythe" <alibuda-AT-linux.alibaba.com>, Dust Li <dust.li-AT-linux.alibaba.com>, Sidraya Jayagond <sidraya-AT-linux.ibm.com>, Wenjia Zhang <wenjia-AT-linux.ibm.com>, Julian Ruess <julianr-AT-linux.ibm.com> | |
Subject: | [RFC net-next 00/17] dibs - Direct Internal Buffer Sharing | |
Date: | Wed, 06 Aug 2025 17:41:05 +0200 | |
Message-ID: | <20250806154122.3413330-1-wintera@linux.ibm.com> | |
Cc: | netdev-AT-vger.kernel.org, linux-s390-AT-vger.kernel.org, Heiko Carstens <hca-AT-linux.ibm.com>, Vasily Gorbik <gor-AT-linux.ibm.com>, Alexander Gordeev <agordeev-AT-linux.ibm.com>, Christian Borntraeger <borntraeger-AT-linux.ibm.com>, Sven Schnelle <svens-AT-linux.ibm.com>, Thorsten Winkler <twinkler-AT-linux.ibm.com>, Simon Horman <horms-AT-kernel.org>, Mahanta Jambigi <mjambigi-AT-linux.ibm.com>, Tony Lu <tonylu-AT-linux.alibaba.com>, Wen Gu <guwen-AT-linux.alibaba.com>, Halil Pasic <pasic-AT-linux.ibm.com>, linux-rdma-AT-vger.kernel.org | |
Archive-link: | Article |
This series introduces a generic abstraction of existing components like: - the s390 specific ISM device (Internal Shared Memory), - the SMC-D loopback mechanism (Shared Memory Communication - Direct) - the client interface of the SMC-D module to the transport devices This generic shim layer can be extended with more devices, more clients and more features in the future. This layer is called 'dibs' for Direct Internal Buffer Sharing based on the common scheme that these mechanisms enable controlled sharing of memory buffers within some containing entity such as a hypervisor or a Linux instance. Benefits: - Cleaner separation of ISM and SMC-D functionality - simpler and less module dependencies - Clear interface definition. - Extendable for future devices and clients. An overview was given at the Netdev 0x19 conference, recordings and slides are available [1]. Background / Status quo: ------------------------ Currently s390 hardware provides virtual PCI ISM devices (Internal Shared Memory). Their driver is in drivers/s390/net/ism_drv.c. The main user is SMC-D (net/smc). The ism driver offers a client interface so other users/protocols can also use them, but it is still heavily intermingled with the smc code. Namely, the ism module cannot be used without the smc module, which feels artificial. There is ongoing work to extend the ISM concept of shared buffers that can be accessed directly by another instance on the same hardware: [2] proposed a loopback interface (ism_lo), that can be used on non-s390 architectures (e.g. between containers or to test SMC-D). A minimal implementation went upstream with [3]: ism_lo currently is a part of the smc protocol and rather hidden. [4] proposed a virtio definition of ism (ism_virtio) that can be used between kvm guests. We will shortly send an RFC for an dibs client that uses dibs as transport for TTY. Concept: -------- Create a shim layer in net/dibs that contains common definitions and code for all dibs devices and all dibs clients. Any device or client module only needs to depend on this dibs layer module and any device or client code only needs to include the definitions in include/linux/dibs.h. The name dibs was chosen to clearly distinguish it from the existing s390 ism devices. And to emphasize that it is not about sharing whole memory regions with anybody, but dedicating single buffers for another system. Implementation: --------------- The end result of this series is: A dibs shim layer with One dibs client: smc-d Two dibs device drivers: ism and dibs-loopback Everything prepared to add more clients and more device drivers. As net-next is still closed, I am sending this series as RFC to give reviewers an early start (vacation season). This series is based on net-next tag 'bpf-next-6.17' and contains everything I want to include for the initial implementation. Patches 1-5 contain some issues that were found along the way. They make sense on their own, but also enable a better structured dibs series. There are three components that exist today: a) smc module (especially SMC-D functionality, which is an ism client today) b) ism device driver (supports multiple ism clients today) c) smc-loopback (integrated with smc today) In order to preserve existing functionality at each step, these are not moved to dibs layer by component, instead: - the dibs layer is established in parallel to existing code [patches 6-9] - then some service functions are moved to the dibs layer [patches 10-15] - the actual data movement is moved to the dibs layer [patch 16] - and last event handling is moved to the dibs layer [patch 17] Future: ------- Items that are not part of this patchset but could be added later: - dynamically add or remove dibs_loopback. That will be allow for simple testing of add_dev()/del_dev() - handle_irq(): Call clients without interrupt context. e.g using threaded interrupts. I left this for a follow-on, because it includes conceptual changes for the smcd receive code. - Any improvements of locking scopes. I mainly moved some of the the existing locks to dibs layer. I have the feeling there is room for improvements. Link: [1] https://netdevconf.info/0x19/sessions/talk/communication-... Link: [2] https://lore.kernel.org/netdev/1695568613-125057-1-git-se... Link: [3] https://lore.kernel.org/linux-kernel//20240428060738.6084... Link: [4] https://groups.oasis-open.org/communities/community-home/... Alexandra Winter (14): net/smc: Remove __init marker from smc_core_init() s390/ism: Log module load/unload net/smc: Remove error handling of unregister_dmb() net/smc: Decouple sf and attached send_buf in smc_loopback net/smc: Improve log message for devices w/o pnetid net/dibs: Create net/dibs net/dibs: Register smc as dibs_client net/dibs: Register ism as dibs device net/dibs: Define dibs loopback net/dibs: Define dibs_client_ops and dibs_dev_ops net/dibs: Local gid for dibs devices net/dibs: Move vlan support to dibs_dev_ops net/dibs: Move query_remote_gid() to dibs_dev_ops net/dibs: Move data path to dibs layer Julian Ruess (3): net/dibs: Move struct device to dibs_dev net/dibs: Create class dibs net/dibs: Move event handling to dibs layer MAINTAINERS | 9 +- drivers/s390/net/Kconfig | 3 +- drivers/s390/net/ism.h | 53 +++- drivers/s390/net/ism_drv.c | 580 ++++++++++++++----------------------- include/linux/dibs.h | 479 ++++++++++++++++++++++++++++++ include/linux/ism.h | 93 ------ include/net/smc.h | 51 +--- net/Kconfig | 1 + net/Makefile | 1 + net/dibs/Kconfig | 27 ++ net/dibs/Makefile | 8 + net/dibs/dibs_loopback.c | 355 +++++++++++++++++++++++ net/dibs/dibs_loopback.h | 57 ++++ net/dibs/dibs_main.c | 280 ++++++++++++++++++ net/smc/Kconfig | 16 +- net/smc/Makefile | 1 - net/smc/af_smc.c | 12 +- net/smc/smc_clc.c | 6 +- net/smc/smc_core.c | 8 +- net/smc/smc_core.h | 5 + net/smc/smc_diag.c | 2 +- net/smc/smc_ib.c | 18 +- net/smc/smc_ism.c | 233 ++++++++------- net/smc/smc_ism.h | 36 ++- net/smc/smc_loopback.c | 421 --------------------------- net/smc/smc_loopback.h | 60 ---- net/smc/smc_pnet.c | 8 +- net/smc/smc_tx.c | 3 + 28 files changed, 1672 insertions(+), 1154 deletions(-) create mode 100644 include/linux/dibs.h delete mode 100644 include/linux/ism.h create mode 100644 net/dibs/Kconfig create mode 100644 net/dibs/Makefile create mode 100644 net/dibs/dibs_loopback.c create mode 100644 net/dibs/dibs_loopback.h create mode 100644 net/dibs/dibs_main.c delete mode 100644 net/smc/smc_loopback.c delete mode 100644 net/smc/smc_loopback.h -- 2.48.1