This series improves the CPU cost of RX token management by adding an attribute to NETDEV_CMD_BIND_RX that configures sockets using the binding to avoid the xarray allocator and instead use a per-binding niov array and a uref field in niov.
Improvement is ~13% cpu util per RX user thread.
Using kperf, the following results were observed:
Before: Average RX worker idle %: 13.13, flows 4, test runs 11 After: Average RX worker idle %: 26.32, flows 4, test runs 11
Two other approaches were tested, but with no improvement. Namely, 1) using a hashmap for tokens and 2) keeping an xarray of atomic counters but using RCU so that the hotpath could be mostly lockless. Neither of these approaches proved better than the simple array in terms of CPU.
The attribute NETDEV_A_DMABUF_AUTORELEASE is added to toggle the optimization. It is an optional attribute and defaults to 0 (i.e., optimization on).
To: David S. Miller davem@davemloft.net To: Eric Dumazet edumazet@google.com To: Jakub Kicinski kuba@kernel.org To: Paolo Abeni pabeni@redhat.com To: Simon Horman horms@kernel.org To: Kuniyuki Iwashima kuniyu@google.com To: Willem de Bruijn willemb@google.com To: Neal Cardwell ncardwell@google.com To: David Ahern dsahern@kernel.org To: Mina Almasry almasrymina@google.com To: Arnd Bergmann arnd@arndb.de To: Jonathan Corbet corbet@lwn.net To: Andrew Lunn andrew+netdev@lunn.ch To: Shuah Khan shuah@kernel.org Cc: Stanislav Fomichev sdf@fomichev.me Cc: netdev@vger.kernel.org Cc: linux-kernel@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linux-doc@vger.kernel.org Cc: linux-kselftest@vger.kernel.org Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com
Changes in v7: - use netlink instead of sockopt (Stan) - restrict system to only one mode, dmabuf bindings can not co-exist with different modes (Stan) - use static branching to enforce single system-wide mode (Stan) - Link to v6: https://lore.kernel.org/r/20251104-scratch-bobbyeshleman-devmem-tcp-token-up...
Changes in v6: - renamed 'net: devmem: use niov array for token management' to refer to optionality of new config - added documentation and tests - make autorelease flag per-socket sockopt instead of binding field / sysctl - many per-patch changes (see Changes sections per-patch) - Link to v5: https://lore.kernel.org/r/20251023-scratch-bobbyeshleman-devmem-tcp-token-up...
Changes in v5: - add sysctl to opt-out of performance benefit, back to old token release - Link to v4: https://lore.kernel.org/all/20250926-scratch-bobbyeshleman-devmem-tcp-token-...
Changes in v4: - rebase to net-next - Link to v3: https://lore.kernel.org/r/20250926-scratch-bobbyeshleman-devmem-tcp-token-up...
Changes in v3: - make urefs per-binding instead of per-socket, reducing memory footprint - fallback to cleaning up references in dmabuf unbind if socket leaked tokens - drop ethtool patch - Link to v2: https://lore.kernel.org/r/20250911-scratch-bobbyeshleman-devmem-tcp-token-up...
Changes in v2: - net: ethtool: prevent user from breaking devmem single-binding rule (Mina) - pre-assign niovs in binding->vec for RX case (Mina) - remove WARNs on invalid user input (Mina) - remove extraneous binding ref get (Mina) - remove WARN for changed binding (Mina) - always use GFP_ZERO for binding->vec (Mina) - fix length of alloc for urefs - use atomic_set(, 0) to initialize sk_user_frags.urefs - Link to v1: https://lore.kernel.org/r/20250902-scratch-bobbyeshleman-devmem-tcp-token-up...
--- Bobby Eshleman (5): net: devmem: rename tx_vec to vec in dmabuf binding net: devmem: refactor sock_devmem_dontneed for autorelease split net: devmem: implement autorelease token management net: devmem: document NETDEV_A_DMABUF_AUTORELEASE netlink attribute selftests: drv-net: devmem: add autorelease tests
Documentation/netlink/specs/netdev.yaml | 12 +++ Documentation/networking/devmem.rst | 70 +++++++++++++ include/net/netmem.h | 1 + include/net/sock.h | 7 +- include/uapi/linux/netdev.h | 1 + net/core/devmem.c | 121 ++++++++++++++++++---- net/core/devmem.h | 13 ++- net/core/netdev-genl-gen.c | 5 +- net/core/netdev-genl.c | 13 ++- net/core/sock.c | 103 ++++++++++++++---- net/ipv4/tcp.c | 78 +++++++++++--- net/ipv4/tcp_ipv4.c | 13 ++- net/ipv4/tcp_minisocks.c | 3 +- tools/include/uapi/linux/netdev.h | 1 + tools/testing/selftests/drivers/net/hw/devmem.py | 22 +++- tools/testing/selftests/drivers/net/hw/ncdevmem.c | 19 ++-- 16 files changed, 401 insertions(+), 81 deletions(-) --- base-commit: 4c52142904b33b41c3ff7ee58670b4e3b3bf1120 change-id: 20250829-scratch-bobbyeshleman-devmem-tcp-token-upstream-292be174d503
Best regards,
From: Bobby Eshleman bobbyeshleman@meta.com
Rename the 'tx_vec' field in struct net_devmem_dmabuf_binding to 'vec'. This field holds pointers to net_iov structures. The rename prepares for reusing 'vec' for both TX and RX directions.
No functional change intended.
Reviewed-by: Mina Almasry almasrymina@google.com Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com --- net/core/devmem.c | 22 +++++++++++----------- net/core/devmem.h | 2 +- 2 files changed, 12 insertions(+), 12 deletions(-)
diff --git a/net/core/devmem.c b/net/core/devmem.c index 1d04754bc756..4dee2666dd07 100644 --- a/net/core/devmem.c +++ b/net/core/devmem.c @@ -75,7 +75,7 @@ void __net_devmem_dmabuf_binding_free(struct work_struct *wq) dma_buf_detach(binding->dmabuf, binding->attachment); dma_buf_put(binding->dmabuf); xa_destroy(&binding->bound_rxqs); - kvfree(binding->tx_vec); + kvfree(binding->vec); kfree(binding); }
@@ -232,10 +232,10 @@ net_devmem_bind_dmabuf(struct net_device *dev, }
if (direction == DMA_TO_DEVICE) { - binding->tx_vec = kvmalloc_array(dmabuf->size / PAGE_SIZE, - sizeof(struct net_iov *), - GFP_KERNEL); - if (!binding->tx_vec) { + binding->vec = kvmalloc_array(dmabuf->size / PAGE_SIZE, + sizeof(struct net_iov *), + GFP_KERNEL); + if (!binding->vec) { err = -ENOMEM; goto err_unmap; } @@ -249,7 +249,7 @@ net_devmem_bind_dmabuf(struct net_device *dev, dev_to_node(&dev->dev)); if (!binding->chunk_pool) { err = -ENOMEM; - goto err_tx_vec; + goto err_vec; }
virtual = 0; @@ -295,7 +295,7 @@ net_devmem_bind_dmabuf(struct net_device *dev, page_pool_set_dma_addr_netmem(net_iov_to_netmem(niov), net_devmem_get_dma_addr(niov)); if (direction == DMA_TO_DEVICE) - binding->tx_vec[owner->area.base_virtual / PAGE_SIZE + i] = niov; + binding->vec[owner->area.base_virtual / PAGE_SIZE + i] = niov; }
virtual += len; @@ -315,8 +315,8 @@ net_devmem_bind_dmabuf(struct net_device *dev, gen_pool_for_each_chunk(binding->chunk_pool, net_devmem_dmabuf_free_chunk_owner, NULL); gen_pool_destroy(binding->chunk_pool); -err_tx_vec: - kvfree(binding->tx_vec); +err_vec: + kvfree(binding->vec); err_unmap: dma_buf_unmap_attachment_unlocked(binding->attachment, binding->sgt, direction); @@ -363,7 +363,7 @@ struct net_devmem_dmabuf_binding *net_devmem_get_binding(struct sock *sk, int err = 0;
binding = net_devmem_lookup_dmabuf(dmabuf_id); - if (!binding || !binding->tx_vec) { + if (!binding || !binding->vec) { err = -EINVAL; goto out_err; } @@ -414,7 +414,7 @@ net_devmem_get_niov_at(struct net_devmem_dmabuf_binding *binding, *off = virt_addr % PAGE_SIZE; *size = PAGE_SIZE - *off;
- return binding->tx_vec[virt_addr / PAGE_SIZE]; + return binding->vec[virt_addr / PAGE_SIZE]; }
/*** "Dmabuf devmem memory provider" ***/ diff --git a/net/core/devmem.h b/net/core/devmem.h index 0b43a648cd2e..1ea6228e4f40 100644 --- a/net/core/devmem.h +++ b/net/core/devmem.h @@ -63,7 +63,7 @@ struct net_devmem_dmabuf_binding { * address. This array is convenient to map the virtual addresses to * net_iovs in the TX path. */ - struct net_iov **tx_vec; + struct net_iov **vec;
struct work_struct unbind_w; };
From: Bobby Eshleman bobbyeshleman@meta.com
Refactor sock_devmem_dontneed() in preparation for supporting both autorelease and manual token release modes.
Split the function into two parts: - sock_devmem_dontneed(): handles input validation, token allocation, and copying from userspace - sock_devmem_dontneed_autorelease(): performs the actual token release via xarray lookup and page pool put
This separation allows a future commit to add a parallel sock_devmem_dontneed_manual_release() function that uses a different token tracking mechanism (per-niov reference counting) without duplicating the input validation logic.
The refactoring is purely mechanical with no functional change. Only intended to minimize the noise in subsequent patches.
Reviewed-by: Mina Almasry almasrymina@google.com Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com --- net/core/sock.c | 52 ++++++++++++++++++++++++++++++++-------------------- 1 file changed, 32 insertions(+), 20 deletions(-)
diff --git a/net/core/sock.c b/net/core/sock.c index 3b74fc71f51c..41274bd0394e 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -1082,30 +1082,13 @@ static int sock_reserve_memory(struct sock *sk, int bytes) #define MAX_DONTNEED_FRAGS 1024
static noinline_for_stack int -sock_devmem_dontneed(struct sock *sk, sockptr_t optval, unsigned int optlen) +sock_devmem_dontneed_autorelease(struct sock *sk, struct dmabuf_token *tokens, + unsigned int num_tokens) { - unsigned int num_tokens, i, j, k, netmem_num = 0; - struct dmabuf_token *tokens; + unsigned int i, j, k, netmem_num = 0; int ret = 0, num_frags = 0; netmem_ref netmems[16];
- if (!sk_is_tcp(sk)) - return -EBADF; - - if (optlen % sizeof(*tokens) || - optlen > sizeof(*tokens) * MAX_DONTNEED_TOKENS) - return -EINVAL; - - num_tokens = optlen / sizeof(*tokens); - tokens = kvmalloc_array(num_tokens, sizeof(*tokens), GFP_KERNEL); - if (!tokens) - return -ENOMEM; - - if (copy_from_sockptr(tokens, optval, optlen)) { - kvfree(tokens); - return -EFAULT; - } - xa_lock_bh(&sk->sk_user_frags); for (i = 0; i < num_tokens; i++) { for (j = 0; j < tokens[i].token_count; j++) { @@ -1135,6 +1118,35 @@ sock_devmem_dontneed(struct sock *sk, sockptr_t optval, unsigned int optlen) for (k = 0; k < netmem_num; k++) WARN_ON_ONCE(!napi_pp_put_page(netmems[k]));
+ return ret; +} + +static noinline_for_stack int +sock_devmem_dontneed(struct sock *sk, sockptr_t optval, unsigned int optlen) +{ + struct dmabuf_token *tokens; + unsigned int num_tokens; + int ret; + + if (!sk_is_tcp(sk)) + return -EBADF; + + if (optlen % sizeof(*tokens) || + optlen > sizeof(*tokens) * MAX_DONTNEED_TOKENS) + return -EINVAL; + + num_tokens = optlen / sizeof(*tokens); + tokens = kvmalloc_array(num_tokens, sizeof(*tokens), GFP_KERNEL); + if (!tokens) + return -ENOMEM; + + if (copy_from_sockptr(tokens, optval, optlen)) { + kvfree(tokens); + return -EFAULT; + } + + ret = sock_devmem_dontneed_autorelease(sk, tokens, num_tokens); + kvfree(tokens); return ret; }
From: Bobby Eshleman bobbyeshleman@meta.com
Add support for autorelease toggling of tokens using a static branch to control system-wide behavior. This allows applications to choose between two memory management modes:
1. Autorelease on: Leaked tokens are automatically released when the socket closes.
2. Autorelease off: Leaked tokens are released during dmabuf unbind.
The autorelease mode is requested via the NETDEV_A_DMABUF_AUTORELEASE attribute of the NETDEV_CMD_BIND_RX message. Having separate modes per binding is disallowed and is rejected by netlink. The system will be "locked" into the mode that the first binding is set to. It can only be changed again once there are zero bindings on the system.
Disabling autorelease offers ~13% improvement in CPU utilization.
Static branching is used to limit the system to one mode or the other.
Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com --- Changes in v7: - implement autorelease with static branch (Stan) - use netlink instead of sockopt (Stan) - merge uAPI and implementation patches into one patch (seemed less confusing)
Changes in v6: - remove sk_devmem_info.autorelease, using binding->autorelease instead - move binding->autorelease check to outside of net_devmem_dmabuf_binding_put_urefs() (Mina) - remove overly defensive net_is_devmem_iov() (Mina) - add comment about multiple urefs mapping to a single netmem ref (Mina) - remove overly defense netmem NULL and netmem_is_net_iov checks (Mina) - use niov without casting back and forth with netmem (Mina) - move the autorelease flag from per-binding to per-socket (Mina) - remove the batching logic in sock_devmem_dontneed_manual_release() (Mina) - move autorelease check inside tcp_xa_pool_commit() (Mina) - remove single-binding restriction for autorelease mode (Mina) - unbind always checks for leaked urefs
Changes in v5: - remove unused variables - introduce autorelease flag, preparing for future patch toggle new behavior
Changes in v3: - make urefs per-binding instead of per-socket, reducing memory footprint - fallback to cleaning up references in dmabuf unbind if socket leaked tokens - drop ethtool patch
Changes in v2: - always use GFP_ZERO for binding->vec (Mina) - remove WARN for changed binding (Mina) - remove extraneous binding ref get (Mina) - remove WARNs on invalid user input (Mina) - pre-assign niovs in binding->vec for RX case (Mina) - use atomic_set(, 0) to initialize sk_user_frags.urefs - fix length of alloc for urefs --- Documentation/netlink/specs/netdev.yaml | 12 ++++ include/net/netmem.h | 1 + include/net/sock.h | 7 +- include/uapi/linux/netdev.h | 1 + net/core/devmem.c | 109 +++++++++++++++++++++++++++----- net/core/devmem.h | 11 +++- net/core/netdev-genl-gen.c | 5 +- net/core/netdev-genl.c | 13 +++- net/core/sock.c | 57 +++++++++++++++-- net/ipv4/tcp.c | 78 ++++++++++++++++++----- net/ipv4/tcp_ipv4.c | 13 +++- net/ipv4/tcp_minisocks.c | 3 +- tools/include/uapi/linux/netdev.h | 1 + 13 files changed, 262 insertions(+), 49 deletions(-)
diff --git a/Documentation/netlink/specs/netdev.yaml b/Documentation/netlink/specs/netdev.yaml index 82bf5cb2617d..913fccca4c4e 100644 --- a/Documentation/netlink/specs/netdev.yaml +++ b/Documentation/netlink/specs/netdev.yaml @@ -562,6 +562,17 @@ attribute-sets: type: u32 checks: min: 1 + - + name: autorelease + doc: | + Token autorelease mode. If true (1), leaked tokens are automatically + released when the socket closes. If false (0), leaked tokens are only + released when the dmabuf is unbound. Once a binding is created with a + specific mode, all subsequent bindings system-wide must use the same + mode. + + Optional. Defaults to false if not specified. + type: u8
operations: list: @@ -767,6 +778,7 @@ operations: - ifindex - fd - queues + - autorelease reply: attributes: - id diff --git a/include/net/netmem.h b/include/net/netmem.h index 9e10f4ac50c3..80d2263ba4ed 100644 --- a/include/net/netmem.h +++ b/include/net/netmem.h @@ -112,6 +112,7 @@ struct net_iov { }; struct net_iov_area *owner; enum net_iov_type type; + atomic_t uref; };
struct net_iov_area { diff --git a/include/net/sock.h b/include/net/sock.h index a5f36ea9d46f..797b21148945 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -350,7 +350,7 @@ struct sk_filter; * @sk_scm_rights: flagged by SO_PASSRIGHTS to recv SCM_RIGHTS * @sk_scm_unused: unused flags for scm_recv() * @ns_tracker: tracker for netns reference - * @sk_user_frags: xarray of pages the user is holding a reference on. + * @sk_devmem_info: the devmem binding information for the socket * @sk_owner: reference to the real owner of the socket that calls * sock_lock_init_class_and_name(). */ @@ -579,7 +579,10 @@ struct sock { struct numa_drop_counters *sk_drop_counters; struct rcu_head sk_rcu; netns_tracker ns_tracker; - struct xarray sk_user_frags; + struct { + struct xarray frags; + struct net_devmem_dmabuf_binding *binding; + } sk_devmem_info;
#if IS_ENABLED(CONFIG_PROVE_LOCKING) && IS_ENABLED(CONFIG_MODULES) struct module *sk_owner; diff --git a/include/uapi/linux/netdev.h b/include/uapi/linux/netdev.h index 048c8de1a130..dff0be8223a4 100644 --- a/include/uapi/linux/netdev.h +++ b/include/uapi/linux/netdev.h @@ -206,6 +206,7 @@ enum { NETDEV_A_DMABUF_QUEUES, NETDEV_A_DMABUF_FD, NETDEV_A_DMABUF_ID, + NETDEV_A_DMABUF_AUTORELEASE,
__NETDEV_A_DMABUF_MAX, NETDEV_A_DMABUF_MAX = (__NETDEV_A_DMABUF_MAX - 1) diff --git a/net/core/devmem.c b/net/core/devmem.c index 4dee2666dd07..bba21c6cb195 100644 --- a/net/core/devmem.c +++ b/net/core/devmem.c @@ -11,6 +11,7 @@ #include <linux/genalloc.h> #include <linux/mm.h> #include <linux/netdevice.h> +#include <linux/skbuff_ref.h> #include <linux/types.h> #include <net/netdev_queues.h> #include <net/netdev_rx_queue.h> @@ -28,6 +29,17 @@
static DEFINE_XARRAY_FLAGS(net_devmem_dmabuf_bindings, XA_FLAGS_ALLOC1);
+/* Static key to lock down autorelease to a single mode on a system. When + * enabled: autorelease mode (leaked tokens released on socket close). When + * disabled: manual mode (leaked tokens released on dmabuf unbind). Once the + * first binding is created, the mode is locked system-wide and can not be + * changed until the system has zero bindings again. + * + * Protected by xa_lock of net_devmem_dmabuf_bindings. + */ +DEFINE_STATIC_KEY_FALSE(tcp_devmem_ar_key); +EXPORT_SYMBOL(tcp_devmem_ar_key); + static const struct memory_provider_ops dmabuf_devmem_ops;
bool net_is_devmem_iov(struct net_iov *niov) @@ -116,6 +128,24 @@ void net_devmem_free_dmabuf(struct net_iov *niov) gen_pool_free(binding->chunk_pool, dma_addr, PAGE_SIZE); }
+static void +net_devmem_dmabuf_binding_put_urefs(struct net_devmem_dmabuf_binding *binding) +{ + int i; + + for (i = 0; i < binding->dmabuf->size / PAGE_SIZE; i++) { + struct net_iov *niov; + netmem_ref netmem; + + niov = binding->vec[i]; + netmem = net_iov_to_netmem(niov); + + /* Multiple urefs map to only a single netmem ref. */ + if (atomic_xchg(&niov->uref, 0) > 0) + WARN_ON_ONCE(!napi_pp_put_page(netmem)); + } +} + void net_devmem_unbind_dmabuf(struct net_devmem_dmabuf_binding *binding) { struct netdev_rx_queue *rxq; @@ -143,6 +173,10 @@ void net_devmem_unbind_dmabuf(struct net_devmem_dmabuf_binding *binding) __net_mp_close_rxq(binding->dev, rxq_idx, &mp_params); }
+ /* Clean up any lingering urefs from sockets that had autorelease + * disabled. + */ + net_devmem_dmabuf_binding_put_urefs(binding); net_devmem_dmabuf_binding_put(binding); }
@@ -179,8 +213,10 @@ struct net_devmem_dmabuf_binding * net_devmem_bind_dmabuf(struct net_device *dev, struct device *dma_dev, enum dma_data_direction direction, - unsigned int dmabuf_fd, struct netdev_nl_sock *priv, - struct netlink_ext_ack *extack) + unsigned int dmabuf_fd, + struct netdev_nl_sock *priv, + struct netlink_ext_ack *extack, + bool autorelease) { struct net_devmem_dmabuf_binding *binding; static u32 id_alloc_next; @@ -231,14 +267,13 @@ net_devmem_bind_dmabuf(struct net_device *dev, goto err_detach; }
- if (direction == DMA_TO_DEVICE) { - binding->vec = kvmalloc_array(dmabuf->size / PAGE_SIZE, - sizeof(struct net_iov *), - GFP_KERNEL); - if (!binding->vec) { - err = -ENOMEM; - goto err_unmap; - } + /* Used by TX and also by RX when socket has autorelease disabled */ + binding->vec = kvmalloc_array(dmabuf->size / PAGE_SIZE, + sizeof(struct net_iov *), + GFP_KERNEL | __GFP_ZERO); + if (!binding->vec) { + err = -ENOMEM; + goto err_unmap; }
/* For simplicity we expect to make PAGE_SIZE allocations, but the @@ -292,25 +327,67 @@ net_devmem_bind_dmabuf(struct net_device *dev, niov = &owner->area.niovs[i]; niov->type = NET_IOV_DMABUF; niov->owner = &owner->area; + atomic_set(&niov->uref, 0); page_pool_set_dma_addr_netmem(net_iov_to_netmem(niov), net_devmem_get_dma_addr(niov)); - if (direction == DMA_TO_DEVICE) - binding->vec[owner->area.base_virtual / PAGE_SIZE + i] = niov; + binding->vec[owner->area.base_virtual / PAGE_SIZE + i] = niov; }
virtual += len; }
- err = xa_alloc_cyclic(&net_devmem_dmabuf_bindings, &binding->id, - binding, xa_limit_32b, &id_alloc_next, - GFP_KERNEL); + /* Enforce system-wide autorelease mode consistency for RX bindings. + * TX bindings don't use autorelease (always false) since tokens aren't + * leaked in TX path. Only RX bindings must all have the same + * autorelease mode, never mixed. + * + * We use the xarray's lock to atomically check xa_empty() and toggle + * the static key, avoiding the race where two new bindings may try to + * set the static key to different states. + */ + xa_lock(&net_devmem_dmabuf_bindings); + + if (direction == DMA_FROM_DEVICE) { + if (!xa_empty(&net_devmem_dmabuf_bindings)) { + bool mode; + + mode = static_key_enabled(&tcp_devmem_ar_key); + + /* When bindings exist, enforce that the mode does not + * change. + */ + if (mode != autorelease) { + NL_SET_ERR_MSG_FMT(extack, + "System already configured with autorelease=%d", + mode); + err = -EINVAL; + goto err_unlock_xa; + } + } else { + /* First binding sets the mode for all subsequent + * bindings. + */ + if (autorelease) + static_branch_enable(&tcp_devmem_ar_key); + else + static_branch_disable(&tcp_devmem_ar_key); + } + } + + err = __xa_alloc_cyclic(&net_devmem_dmabuf_bindings, &binding->id, + binding, xa_limit_32b, &id_alloc_next, + GFP_KERNEL); if (err < 0) - goto err_free_chunks; + goto err_unlock_xa; + + xa_unlock(&net_devmem_dmabuf_bindings);
list_add(&binding->list, &priv->bindings);
return binding;
+err_unlock_xa: + xa_unlock(&net_devmem_dmabuf_bindings); err_free_chunks: gen_pool_for_each_chunk(binding->chunk_pool, net_devmem_dmabuf_free_chunk_owner, NULL); diff --git a/net/core/devmem.h b/net/core/devmem.h index 1ea6228e4f40..33e85ff5f35e 100644 --- a/net/core/devmem.h +++ b/net/core/devmem.h @@ -12,9 +12,13 @@
#include <net/netmem.h> #include <net/netdev_netlink.h> +#include <linux/jump_label.h>
struct netlink_ext_ack;
+/* static key for TCP devmem autorelease */ +extern struct static_key_false tcp_devmem_ar_key; + struct net_devmem_dmabuf_binding { struct dma_buf *dmabuf; struct dma_buf_attachment *attachment; @@ -61,7 +65,7 @@ struct net_devmem_dmabuf_binding {
/* Array of net_iov pointers for this binding, sorted by virtual * address. This array is convenient to map the virtual addresses to - * net_iovs in the TX path. + * net_iovs. */ struct net_iov **vec;
@@ -88,7 +92,7 @@ net_devmem_bind_dmabuf(struct net_device *dev, struct device *dma_dev, enum dma_data_direction direction, unsigned int dmabuf_fd, struct netdev_nl_sock *priv, - struct netlink_ext_ack *extack); + struct netlink_ext_ack *extack, bool autorelease); struct net_devmem_dmabuf_binding *net_devmem_lookup_dmabuf(u32 id); void net_devmem_unbind_dmabuf(struct net_devmem_dmabuf_binding *binding); int net_devmem_bind_dmabuf_to_queue(struct net_device *dev, u32 rxq_idx, @@ -174,7 +178,8 @@ net_devmem_bind_dmabuf(struct net_device *dev, enum dma_data_direction direction, unsigned int dmabuf_fd, struct netdev_nl_sock *priv, - struct netlink_ext_ack *extack) + struct netlink_ext_ack *extack, + bool autorelease) { return ERR_PTR(-EOPNOTSUPP); } diff --git a/net/core/netdev-genl-gen.c b/net/core/netdev-genl-gen.c index ff20435c45d2..ecbd8ae2a3fa 100644 --- a/net/core/netdev-genl-gen.c +++ b/net/core/netdev-genl-gen.c @@ -85,10 +85,11 @@ static const struct nla_policy netdev_qstats_get_nl_policy[NETDEV_A_QSTATS_SCOPE };
/* NETDEV_CMD_BIND_RX - do */ -static const struct nla_policy netdev_bind_rx_nl_policy[NETDEV_A_DMABUF_FD + 1] = { +static const struct nla_policy netdev_bind_rx_nl_policy[NETDEV_A_DMABUF_AUTORELEASE + 1] = { [NETDEV_A_DMABUF_IFINDEX] = NLA_POLICY_MIN(NLA_U32, 1), [NETDEV_A_DMABUF_FD] = { .type = NLA_U32, }, [NETDEV_A_DMABUF_QUEUES] = NLA_POLICY_NESTED(netdev_queue_id_nl_policy), + [NETDEV_A_DMABUF_AUTORELEASE] = { .type = NLA_U8, }, };
/* NETDEV_CMD_NAPI_SET - do */ @@ -187,7 +188,7 @@ static const struct genl_split_ops netdev_nl_ops[] = { .cmd = NETDEV_CMD_BIND_RX, .doit = netdev_nl_bind_rx_doit, .policy = netdev_bind_rx_nl_policy, - .maxattr = NETDEV_A_DMABUF_FD, + .maxattr = NETDEV_A_DMABUF_AUTORELEASE, .flags = GENL_ADMIN_PERM | GENL_CMD_CAP_DO, }, { diff --git a/net/core/netdev-genl.c b/net/core/netdev-genl.c index 470fabbeacd9..5f06a677f056 100644 --- a/net/core/netdev-genl.c +++ b/net/core/netdev-genl.c @@ -939,6 +939,7 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) struct netdev_nl_sock *priv; struct net_device *netdev; unsigned long *rxq_bitmap; + bool autorelease = false; struct device *dma_dev; struct sk_buff *rsp; int err = 0; @@ -952,6 +953,10 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) ifindex = nla_get_u32(info->attrs[NETDEV_A_DEV_IFINDEX]); dmabuf_fd = nla_get_u32(info->attrs[NETDEV_A_DMABUF_FD]);
+ if (info->attrs[NETDEV_A_DMABUF_AUTORELEASE]) + autorelease = + !!nla_get_u8(info->attrs[NETDEV_A_DMABUF_AUTORELEASE]); + priv = genl_sk_priv_get(&netdev_nl_family, NETLINK_CB(skb).sk); if (IS_ERR(priv)) return PTR_ERR(priv); @@ -1002,7 +1007,8 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) }
binding = net_devmem_bind_dmabuf(netdev, dma_dev, DMA_FROM_DEVICE, - dmabuf_fd, priv, info->extack); + dmabuf_fd, priv, info->extack, + autorelease); if (IS_ERR(binding)) { err = PTR_ERR(binding); goto err_rxq_bitmap; @@ -1096,8 +1102,11 @@ int netdev_nl_bind_tx_doit(struct sk_buff *skb, struct genl_info *info) }
dma_dev = netdev_queue_get_dma_dev(netdev, 0); + /* TX bindings don't use autorelease. Autorelease is only meaningful + * for RX where tokens may be leaked by userspace. + */ binding = net_devmem_bind_dmabuf(netdev, dma_dev, DMA_TO_DEVICE, - dmabuf_fd, priv, info->extack); + dmabuf_fd, priv, info->extack, false); if (IS_ERR(binding)) { err = PTR_ERR(binding); goto err_unlock_netdev; diff --git a/net/core/sock.c b/net/core/sock.c index 41274bd0394e..f945cdb5a337 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -87,6 +87,7 @@
#include <linux/unaligned.h> #include <linux/capability.h> +#include <linux/dma-buf.h> #include <linux/errno.h> #include <linux/errqueue.h> #include <linux/types.h> @@ -151,6 +152,7 @@ #include <uapi/linux/pidfd.h>
#include "dev.h" +#include "devmem.h"
static DEFINE_MUTEX(proto_list_mutex); static LIST_HEAD(proto_list); @@ -1081,6 +1083,44 @@ static int sock_reserve_memory(struct sock *sk, int bytes) #define MAX_DONTNEED_TOKENS 128 #define MAX_DONTNEED_FRAGS 1024
+static noinline_for_stack int +sock_devmem_dontneed_manual_release(struct sock *sk, + struct dmabuf_token *tokens, + unsigned int num_tokens) +{ + struct net_iov *niov; + unsigned int i, j; + netmem_ref netmem; + unsigned int token; + int num_frags = 0; + int ret = 0; + + if (!sk->sk_devmem_info.binding) + return -EINVAL; + + for (i = 0; i < num_tokens; i++) { + for (j = 0; j < tokens[i].token_count; j++) { + size_t size = sk->sk_devmem_info.binding->dmabuf->size; + + token = tokens[i].token_start + j; + if (token >= size / PAGE_SIZE) + break; + + if (++num_frags > MAX_DONTNEED_FRAGS) + return ret; + + niov = sk->sk_devmem_info.binding->vec[token]; + if (atomic_dec_and_test(&niov->uref)) { + netmem = net_iov_to_netmem(niov); + WARN_ON_ONCE(!napi_pp_put_page(netmem)); + } + ret++; + } + } + + return ret; +} + static noinline_for_stack int sock_devmem_dontneed_autorelease(struct sock *sk, struct dmabuf_token *tokens, unsigned int num_tokens) @@ -1089,32 +1129,33 @@ sock_devmem_dontneed_autorelease(struct sock *sk, struct dmabuf_token *tokens, int ret = 0, num_frags = 0; netmem_ref netmems[16];
- xa_lock_bh(&sk->sk_user_frags); + xa_lock_bh(&sk->sk_devmem_info.frags); for (i = 0; i < num_tokens; i++) { for (j = 0; j < tokens[i].token_count; j++) { if (++num_frags > MAX_DONTNEED_FRAGS) goto frag_limit_reached;
netmem_ref netmem = (__force netmem_ref)__xa_erase( - &sk->sk_user_frags, tokens[i].token_start + j); + &sk->sk_devmem_info.frags, + tokens[i].token_start + j);
if (!netmem || WARN_ON_ONCE(!netmem_is_net_iov(netmem))) continue;
netmems[netmem_num++] = netmem; if (netmem_num == ARRAY_SIZE(netmems)) { - xa_unlock_bh(&sk->sk_user_frags); + xa_unlock_bh(&sk->sk_devmem_info.frags); for (k = 0; k < netmem_num; k++) WARN_ON_ONCE(!napi_pp_put_page(netmems[k])); netmem_num = 0; - xa_lock_bh(&sk->sk_user_frags); + xa_lock_bh(&sk->sk_devmem_info.frags); } ret++; } }
frag_limit_reached: - xa_unlock_bh(&sk->sk_user_frags); + xa_unlock_bh(&sk->sk_devmem_info.frags); for (k = 0; k < netmem_num; k++) WARN_ON_ONCE(!napi_pp_put_page(netmems[k]));
@@ -1145,7 +1186,11 @@ sock_devmem_dontneed(struct sock *sk, sockptr_t optval, unsigned int optlen) return -EFAULT; }
- ret = sock_devmem_dontneed_autorelease(sk, tokens, num_tokens); + if (static_branch_unlikely(&tcp_devmem_ar_key)) + ret = sock_devmem_dontneed_autorelease(sk, tokens, num_tokens); + else + ret = sock_devmem_dontneed_manual_release(sk, tokens, + num_tokens);
kvfree(tokens); return ret; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index dee578aad690..e17b71244922 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -260,6 +260,7 @@ #include <linux/memblock.h> #include <linux/highmem.h> #include <linux/cache.h> +#include <linux/dma-buf.h> #include <linux/err.h> #include <linux/time.h> #include <linux/slab.h> @@ -492,7 +493,8 @@ void tcp_init_sock(struct sock *sk)
set_bit(SOCK_SUPPORT_ZC, &sk->sk_socket->flags); sk_sockets_allocated_inc(sk); - xa_init_flags(&sk->sk_user_frags, XA_FLAGS_ALLOC1); + xa_init_flags(&sk->sk_devmem_info.frags, XA_FLAGS_ALLOC1); + sk->sk_devmem_info.binding = NULL; } EXPORT_IPV6_MOD(tcp_init_sock);
@@ -2424,11 +2426,12 @@ static void tcp_xa_pool_commit_locked(struct sock *sk, struct tcp_xa_pool *p)
/* Commit part that has been copied to user space. */ for (i = 0; i < p->idx; i++) - __xa_cmpxchg(&sk->sk_user_frags, p->tokens[i], XA_ZERO_ENTRY, - (__force void *)p->netmems[i], GFP_KERNEL); + __xa_cmpxchg(&sk->sk_devmem_info.frags, p->tokens[i], + XA_ZERO_ENTRY, (__force void *)p->netmems[i], + GFP_KERNEL); /* Rollback what has been pre-allocated and is no longer needed. */ for (; i < p->max; i++) - __xa_erase(&sk->sk_user_frags, p->tokens[i]); + __xa_erase(&sk->sk_devmem_info.frags, p->tokens[i]);
p->max = 0; p->idx = 0; @@ -2436,14 +2439,18 @@ static void tcp_xa_pool_commit_locked(struct sock *sk, struct tcp_xa_pool *p)
static void tcp_xa_pool_commit(struct sock *sk, struct tcp_xa_pool *p) { + /* Skip xarray operations if autorelease is disabled (manual mode) */ + if (!static_branch_unlikely(&tcp_devmem_ar_key)) + return; + if (!p->max) return;
- xa_lock_bh(&sk->sk_user_frags); + xa_lock_bh(&sk->sk_devmem_info.frags);
tcp_xa_pool_commit_locked(sk, p);
- xa_unlock_bh(&sk->sk_user_frags); + xa_unlock_bh(&sk->sk_devmem_info.frags); }
static int tcp_xa_pool_refill(struct sock *sk, struct tcp_xa_pool *p, @@ -2454,24 +2461,42 @@ static int tcp_xa_pool_refill(struct sock *sk, struct tcp_xa_pool *p, if (p->idx < p->max) return 0;
- xa_lock_bh(&sk->sk_user_frags); + xa_lock_bh(&sk->sk_devmem_info.frags);
tcp_xa_pool_commit_locked(sk, p);
for (k = 0; k < max_frags; k++) { - err = __xa_alloc(&sk->sk_user_frags, &p->tokens[k], + err = __xa_alloc(&sk->sk_devmem_info.frags, &p->tokens[k], XA_ZERO_ENTRY, xa_limit_31b, GFP_KERNEL); if (err) break; }
- xa_unlock_bh(&sk->sk_user_frags); + xa_unlock_bh(&sk->sk_devmem_info.frags);
p->max = k; p->idx = 0; return k ? 0 : err; }
+static void tcp_xa_pool_inc_pp_ref_count(struct tcp_xa_pool *tcp_xa_pool, + skb_frag_t *frag, int *refs) +{ + struct net_iov *niov; + + niov = skb_frag_net_iov(frag); + + if (static_branch_unlikely(&tcp_devmem_ar_key)) { + atomic_long_inc(&niov->pp_ref_count); + tcp_xa_pool->netmems[tcp_xa_pool->idx++] = + skb_frag_netmem(frag); + } else { + if (atomic_inc_return(&niov->uref) == 1) + atomic_long_inc(&niov->pp_ref_count); + (*refs)++; + } +} + /* On error, returns the -errno. On success, returns number of bytes sent to the * user. May not consume all of @remaining_len. */ @@ -2479,10 +2504,12 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, unsigned int offset, struct msghdr *msg, int remaining_len) { + struct net_devmem_dmabuf_binding *binding = NULL; struct dmabuf_cmsg dmabuf_cmsg = { 0 }; struct tcp_xa_pool tcp_xa_pool; unsigned int start; int i, copy, n; + int refs = 0; int sent = 0; int err = 0;
@@ -2536,6 +2563,7 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; struct net_iov *niov; u64 frag_offset; + u32 token; int end;
/* !skb_frags_readable() should indicate that ALL the @@ -2568,13 +2596,32 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, start; dmabuf_cmsg.frag_offset = frag_offset; dmabuf_cmsg.frag_size = copy; - err = tcp_xa_pool_refill(sk, &tcp_xa_pool, - skb_shinfo(skb)->nr_frags - i); - if (err) + + binding = net_devmem_iov_binding(niov); + + if (!sk->sk_devmem_info.binding) + sk->sk_devmem_info.binding = binding; + + if (sk->sk_devmem_info.binding != binding) { + err = -EFAULT; goto out; + } + + if (static_branch_unlikely(&tcp_devmem_ar_key)) { + err = tcp_xa_pool_refill(sk, &tcp_xa_pool, + skb_shinfo(skb)->nr_frags - i); + if (err) + goto out; + + dmabuf_cmsg.frag_token = + tcp_xa_pool.tokens[tcp_xa_pool.idx]; + } else { + token = net_iov_virtual_addr(niov) >> PAGE_SHIFT; + dmabuf_cmsg.frag_token = token; + } +
/* Will perform the exchange later */ - dmabuf_cmsg.frag_token = tcp_xa_pool.tokens[tcp_xa_pool.idx]; dmabuf_cmsg.dmabuf_id = net_devmem_iov_binding_id(niov);
offset += copy; @@ -2587,8 +2634,8 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, if (err) goto out;
- atomic_long_inc(&niov->pp_ref_count); - tcp_xa_pool.netmems[tcp_xa_pool.idx++] = skb_frag_netmem(frag); + tcp_xa_pool_inc_pp_ref_count(&tcp_xa_pool, frag, + &refs);
sent += copy;
@@ -2617,6 +2664,7 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb,
out: tcp_xa_pool_commit(sk, &tcp_xa_pool); + if (!sent) sent = err;
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 6fcaecb67284..bfa19aeec6b5 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -89,6 +89,9 @@
#include <crypto/md5.h>
+#include <linux/dma-buf.h> +#include "../core/devmem.h" + #include <trace/events/tcp.h>
#ifdef CONFIG_TCP_MD5SIG @@ -2492,7 +2495,7 @@ static void tcp_release_user_frags(struct sock *sk) unsigned long index; void *netmem;
- xa_for_each(&sk->sk_user_frags, index, netmem) + xa_for_each(&sk->sk_devmem_info.frags, index, netmem) WARN_ON_ONCE(!napi_pp_put_page((__force netmem_ref)netmem)); #endif } @@ -2501,9 +2504,15 @@ void tcp_v4_destroy_sock(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk);
+ /* No static branch because sockets may outlive the binding, which + * opens the opportunity for static key state to change. In any + * scenario, if the xarray is non-empty then we need to free those + * frags. + */ tcp_release_user_frags(sk);
- xa_destroy(&sk->sk_user_frags); + xa_destroy(&sk->sk_devmem_info.frags); + sk->sk_devmem_info.binding = NULL;
trace_tcp_destroy_sock(sk);
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index bd5462154f97..2aec977f5c12 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -662,7 +662,8 @@ struct sock *tcp_create_openreq_child(const struct sock *sk,
__TCP_INC_STATS(sock_net(sk), TCP_MIB_PASSIVEOPENS);
- xa_init_flags(&newsk->sk_user_frags, XA_FLAGS_ALLOC1); + xa_init_flags(&newsk->sk_devmem_info.frags, XA_FLAGS_ALLOC1); + newsk->sk_devmem_info.binding = NULL;
return newsk; } diff --git a/tools/include/uapi/linux/netdev.h b/tools/include/uapi/linux/netdev.h index 048c8de1a130..dff0be8223a4 100644 --- a/tools/include/uapi/linux/netdev.h +++ b/tools/include/uapi/linux/netdev.h @@ -206,6 +206,7 @@ enum { NETDEV_A_DMABUF_QUEUES, NETDEV_A_DMABUF_FD, NETDEV_A_DMABUF_ID, + NETDEV_A_DMABUF_AUTORELEASE,
__NETDEV_A_DMABUF_MAX, NETDEV_A_DMABUF_MAX = (__NETDEV_A_DMABUF_MAX - 1)
On 11/20/25 4:37 AM, Bobby Eshleman wrote:
@@ -2479,10 +2504,12 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, unsigned int offset, struct msghdr *msg, int remaining_len) {
- struct net_devmem_dmabuf_binding *binding = NULL; struct dmabuf_cmsg dmabuf_cmsg = { 0 }; struct tcp_xa_pool tcp_xa_pool; unsigned int start; int i, copy, n;
- int refs = 0; int sent = 0; int err = 0;
@@ -2536,6 +2563,7 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; struct net_iov *niov; u64 frag_offset;
u32 token; int end;/* !skb_frags_readable() should indicate that ALL the @@ -2568,13 +2596,32 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, start; dmabuf_cmsg.frag_offset = frag_offset; dmabuf_cmsg.frag_size = copy;
err = tcp_xa_pool_refill(sk, &tcp_xa_pool,skb_shinfo(skb)->nr_frags - i);if (err)
binding = net_devmem_iov_binding(niov);if (!sk->sk_devmem_info.binding)sk->sk_devmem_info.binding = binding;if (sk->sk_devmem_info.binding != binding) {err = -EFAULT; goto out;}if (static_branch_unlikely(&tcp_devmem_ar_key)) {
Not a real/full review but the above is apparently causing kunit build failures:
ERROR:root:ld: vmlinux.o: in function `tcp_recvmsg_dmabuf': tcp.c:(.text+0x669b21): undefined reference to `tcp_devmem_ar_key' ld: tcp.c:(.text+0x669b68): undefined reference to `tcp_devmem_ar_key' ld: tcp.c:(.text+0x669c54): undefined reference to `tcp_devmem_ar_key' make[3]: *** [../scripts/Makefile.vmlinux:72: vmlinux.unstripped] Error 1 make[2]: *** [/home/kunit/testing/Makefile:1242: vmlinux] Error 2 make[1]: *** [/home/kunit/testing/Makefile:248: __sub-make] Error 2 make: *** [Makefile:248: __sub-make] Error 2
see:
https://netdev-3.bots.linux.dev/kunit/results/393664/
@@ -2617,6 +2664,7 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, out: tcp_xa_pool_commit(sk, &tcp_xa_pool);
[just because I stumbled upon the above while looking for the build issue]: please do not mix unrelated whitespace-change only with functional change.
/P
On Thu, Nov 20, 2025 at 01:19:54PM +0100, Paolo Abeni wrote:
On 11/20/25 4:37 AM, Bobby Eshleman wrote:
@@ -2479,10 +2504,12 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, unsigned int offset, struct msghdr *msg, int remaining_len) {
- struct net_devmem_dmabuf_binding *binding = NULL; struct dmabuf_cmsg dmabuf_cmsg = { 0 }; struct tcp_xa_pool tcp_xa_pool; unsigned int start; int i, copy, n;
- int refs = 0; int sent = 0; int err = 0;
@@ -2536,6 +2563,7 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, skb_frag_t *frag = &skb_shinfo(skb)->frags[i]; struct net_iov *niov; u64 frag_offset;
u32 token; int end;/* !skb_frags_readable() should indicate that ALL the @@ -2568,13 +2596,32 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, start; dmabuf_cmsg.frag_offset = frag_offset; dmabuf_cmsg.frag_size = copy;
err = tcp_xa_pool_refill(sk, &tcp_xa_pool,skb_shinfo(skb)->nr_frags - i);if (err)
binding = net_devmem_iov_binding(niov);if (!sk->sk_devmem_info.binding)sk->sk_devmem_info.binding = binding;if (sk->sk_devmem_info.binding != binding) {err = -EFAULT; goto out;}if (static_branch_unlikely(&tcp_devmem_ar_key)) {Not a real/full review but the above is apparently causing kunit build failures:
ERROR:root:ld: vmlinux.o: in function `tcp_recvmsg_dmabuf': tcp.c:(.text+0x669b21): undefined reference to `tcp_devmem_ar_key' ld: tcp.c:(.text+0x669b68): undefined reference to `tcp_devmem_ar_key' ld: tcp.c:(.text+0x669c54): undefined reference to `tcp_devmem_ar_key' make[3]: *** [../scripts/Makefile.vmlinux:72: vmlinux.unstripped] Error 1 make[2]: *** [/home/kunit/testing/Makefile:1242: vmlinux] Error 2 make[1]: *** [/home/kunit/testing/Makefile:248: __sub-make] Error 2 make: *** [Makefile:248: __sub-make] Error 2
see:
Thanks Paolo, I'll fix that for the next rev. And I'll have to add building kunit into my flow (currently using some custom stuff and nipa ingest_mdir, but don't think either has kunit).
@@ -2617,6 +2664,7 @@ static int tcp_recvmsg_dmabuf(struct sock *sk, const struct sk_buff *skb, out: tcp_xa_pool_commit(sk, &tcp_xa_pool); +
[just because I stumbled upon the above while looking for the build issue]: please do not mix unrelated whitespace-change only with functional change.
Don't know how I missed that line, will definitely remove.
Thanks again, Bobby
From: Bobby Eshleman bobbyeshleman@meta.com
Update devmem.rst documentation to describe the autorelease netlink attribute used during RX dmabuf binding.
The autorelease attribute is specified at bind-time via the netlink API (NETDEV_CMD_BIND_RX) and controls what happens to outstanding tokens when the socket closes.
Document the two token release modes (automatic vs manual), how to configure the binding for autorelease, the perf benefits, new caveats and restrictions, and the way the mode is enforced system-wide.
Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com --- Changes in v7: - Document netlink instead of sockopt - Mention system-wide locked to one mode --- Documentation/networking/devmem.rst | 70 +++++++++++++++++++++++++++++++++++++ 1 file changed, 70 insertions(+)
diff --git a/Documentation/networking/devmem.rst b/Documentation/networking/devmem.rst index a6cd7236bfbd..67c63bc5a7ae 100644 --- a/Documentation/networking/devmem.rst +++ b/Documentation/networking/devmem.rst @@ -235,6 +235,76 @@ can be less than the tokens provided by the user in case of: (a) an internal kernel leak bug. (b) the user passed more than 1024 frags.
+ +Autorelease Control +~~~~~~~~~~~~~~~~~~~ + +The autorelease mode controls what happens to outstanding tokens (tokens not +released via SO_DEVMEM_DONTNEED) when the socket closes. Autorelease is +configured per-binding at binding creation time via the netlink API:: + + struct netdev_bind_rx_req *req; + struct netdev_bind_rx_rsp *rsp; + struct ynl_sock *ys; + struct ynl_error yerr; + + ys = ynl_sock_create(&ynl_netdev_family, &yerr); + + req = netdev_bind_rx_req_alloc(); + netdev_bind_rx_req_set_ifindex(req, ifindex); + netdev_bind_rx_req_set_fd(req, dmabuf_fd); + netdev_bind_rx_req_set_autorelease(req, 0); /* 0 = manual, 1 = auto */ + __netdev_bind_rx_req_set_queues(req, queues, n_queues); + + rsp = netdev_bind_rx(ys, req); + + dmabuf_id = rsp->id; + +When autorelease is disabled (0): + +- Outstanding tokens are NOT released when the socket closes +- Outstanding tokens are only released when the dmabuf is unbound +- Provides better performance by eliminating xarray overhead (~13% CPU reduction) +- Kernel tracks tokens via atomic reference counters in net_iov structures + +When autorelease is enabled (1): + +- Outstanding tokens are automatically released when the socket closes +- Backwards compatible behavior +- Kernel tracks tokens in an xarray per socket + +The default is autorelease disabled. + +Important: In both modes, applications should call SO_DEVMEM_DONTNEED to +return tokens as soon as they are done processing. The autorelease setting only +affects what happens to tokens that are still outstanding when close() is called. + +The mode is enforced system-wide. Once a binding is created with a specific +autorelease mode, all subsequent bindings system-wide must use the same mode. + + +Performance Considerations +~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Disabling autorelease provides approximately ~13% CPU utilization improvement +in RX workloads. That said, applications must ensure all tokens are released +via SO_DEVMEM_DONTNEED before closing the socket, otherwise the backing pages +will remain pinned until the dmabuf is unbound. + + +Caveats +~~~~~~~ + +- Once a system-wide autorelease mode is selected (via the first binding), + all subsequent bindings must use the same mode. Attempts to create bindings + with a different mode will be rejected with -EINVAL. + +- Applications using manual release mode (autorelease=0) must ensure all tokens + are returned via SO_DEVMEM_DONTNEED before socket close to avoid resource + leaks during the lifetime of the dmabuf binding. Tokens not released before + close() will only be freed when the dmabuf is unbound. + + TX Interface ============
From: Bobby Eshleman bobbyeshleman@meta.com
Add tests cases that check the autorelease modes (on and off). The new tests are the same as the old ones, but just pass a flag to ncdevmem to select the autorelease mode.
Only for RX tests is autorelease checked, as the autorelease ncdevmem flag is unused in the TX case and doesn't apply to TX bind operations.
Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com --- Note: I tested successfully with kperf, but I'm troubleshooting some mlx5 issues with ncdevmem so this patch, though simple, is not fully validated. Will respond to this thread when solve the issue.
Changes in v7: - use autorelease netlink - remove sockopt tests --- tools/testing/selftests/drivers/net/hw/devmem.py | 22 ++++++++++++++++++++-- tools/testing/selftests/drivers/net/hw/ncdevmem.c | 19 +++++++++++++------ 2 files changed, 33 insertions(+), 8 deletions(-)
diff --git a/tools/testing/selftests/drivers/net/hw/devmem.py b/tools/testing/selftests/drivers/net/hw/devmem.py index 45c2d49d55b6..dddb9d77cb28 100755 --- a/tools/testing/selftests/drivers/net/hw/devmem.py +++ b/tools/testing/selftests/drivers/net/hw/devmem.py @@ -25,7 +25,23 @@ def check_rx(cfg) -> None:
port = rand_port() socat = f"socat -u - TCP{cfg.addr_ipver}:{cfg.baddr}:{port},bind={cfg.remote_baddr}:{port}" - listen_cmd = f"{cfg.bin_local} -l -f {cfg.ifname} -s {cfg.addr} -p {port} -c {cfg.remote_addr} -v 7" + listen_cmd = f"{cfg.bin_local} -l -f {cfg.ifname} -s {cfg.addr} -p {port} -c {cfg.remote_addr} -v 7 -a 0" + + with bkg(listen_cmd, exit_wait=True) as ncdevmem: + wait_port_listen(port) + cmd(f"yes $(echo -e \x01\x02\x03\x04\x05\x06) | \ + head -c 1K | {socat}", host=cfg.remote, shell=True) + + ksft_eq(ncdevmem.ret, 0) + + +@ksft_disruptive +def check_rx_autorelease(cfg) -> None: + require_devmem(cfg) + + port = rand_port() + socat = f"socat -u - TCP{cfg.addr_ipver}:{cfg.baddr}:{port},bind={cfg.remote_baddr}:{port}" + listen_cmd = f"{cfg.bin_local} -l -f {cfg.ifname} -s {cfg.addr} -p {port} -c {cfg.remote_addr} -v 7 -a 1"
with bkg(listen_cmd, exit_wait=True) as ncdevmem: wait_port_listen(port) @@ -68,7 +84,9 @@ def main() -> None: cfg.bin_local = path.abspath(path.dirname(__file__) + "/ncdevmem") cfg.bin_remote = cfg.remote.deploy(cfg.bin_local)
- ksft_run([check_rx, check_tx, check_tx_chunks], + ksft_run([check_rx, check_rx_autorelease, + check_tx, check_tx_autorelease, + check_tx_chunks, check_tx_chunks_autorelease], args=(cfg, )) ksft_exit()
diff --git a/tools/testing/selftests/drivers/net/hw/ncdevmem.c b/tools/testing/selftests/drivers/net/hw/ncdevmem.c index 3288ed04ce08..406f1771d9ec 100644 --- a/tools/testing/selftests/drivers/net/hw/ncdevmem.c +++ b/tools/testing/selftests/drivers/net/hw/ncdevmem.c @@ -92,6 +92,7 @@ static char *port; static size_t do_validation; static int start_queue = -1; static int num_queues = -1; +static int devmem_autorelease; static char *ifname; static unsigned int ifindex; static unsigned int dmabuf_id; @@ -679,7 +680,8 @@ static int configure_flow_steering(struct sockaddr_in6 *server_sin)
static int bind_rx_queue(unsigned int ifindex, unsigned int dmabuf_fd, struct netdev_queue_id *queues, - unsigned int n_queue_index, struct ynl_sock **ys) + unsigned int n_queue_index, struct ynl_sock **ys, + int autorelease) { struct netdev_bind_rx_req *req = NULL; struct netdev_bind_rx_rsp *rsp = NULL; @@ -695,6 +697,7 @@ static int bind_rx_queue(unsigned int ifindex, unsigned int dmabuf_fd, req = netdev_bind_rx_req_alloc(); netdev_bind_rx_req_set_ifindex(req, ifindex); netdev_bind_rx_req_set_fd(req, dmabuf_fd); + netdev_bind_rx_req_set_autorelease(req, autorelease); __netdev_bind_rx_req_set_queues(req, queues, n_queue_index);
rsp = netdev_bind_rx(*ys, req); @@ -872,7 +875,8 @@ static int do_server(struct memory_buffer *mem) goto err_reset_rss; }
- if (bind_rx_queue(ifindex, mem->fd, create_queues(), num_queues, &ys)) { + if (bind_rx_queue(ifindex, mem->fd, create_queues(), num_queues, &ys, + devmem_autorelease)) { pr_err("Failed to bind"); goto err_reset_flow_steering; } @@ -1092,7 +1096,7 @@ int run_devmem_tests(void) goto err_reset_headersplit; }
- if (!bind_rx_queue(ifindex, mem->fd, queues, num_queues, &ys)) { + if (!bind_rx_queue(ifindex, mem->fd, queues, num_queues, &ys, 0)) { pr_err("Binding empty queues array should have failed"); goto err_unbind; } @@ -1108,7 +1112,7 @@ int run_devmem_tests(void) goto err_reset_headersplit; }
- if (!bind_rx_queue(ifindex, mem->fd, queues, num_queues, &ys)) { + if (!bind_rx_queue(ifindex, mem->fd, queues, num_queues, &ys, 0)) { pr_err("Configure dmabuf with header split off should have failed"); goto err_unbind; } @@ -1124,7 +1128,7 @@ int run_devmem_tests(void) goto err_reset_headersplit; }
- if (bind_rx_queue(ifindex, mem->fd, queues, num_queues, &ys)) { + if (bind_rx_queue(ifindex, mem->fd, queues, num_queues, &ys, 0)) { pr_err("Failed to bind"); goto err_reset_headersplit; } @@ -1397,7 +1401,7 @@ int main(int argc, char *argv[]) int is_server = 0, opt; int ret, err = 1;
- while ((opt = getopt(argc, argv, "ls:c:p:v:q:t:f:z:")) != -1) { + while ((opt = getopt(argc, argv, "ls:c:p:v:q:t:f:z:a:")) != -1) { switch (opt) { case 'l': is_server = 1; @@ -1426,6 +1430,9 @@ int main(int argc, char *argv[]) case 'z': max_chunk = atoi(optarg); break; + case 'a': + devmem_autorelease = atoi(optarg); + break; case '?': fprintf(stderr, "unknown option: %c\n", optopt); break;
On Wed, Nov 19, 2025 at 07:37:07PM -0800, Bobby Eshleman wrote:
This series improves the CPU cost of RX token management by adding an attribute to NETDEV_CMD_BIND_RX that configures sockets using the binding to avoid the xarray allocator and instead use a per-binding niov array and a uref field in niov.
Improvement is ~13% cpu util per RX user thread. Using kperf, the following results were observed:
Before: Average RX worker idle %: 13.13, flows 4, test runs 11 After: Average RX worker idle %: 26.32, flows 4, test runs 11
Two other approaches were tested, but with no improvement. Namely, 1) using a hashmap for tokens and 2) keeping an xarray of atomic counters but using RCU so that the hotpath could be mostly lockless. Neither of these approaches proved better than the simple array in terms of CPU.
The attribute NETDEV_A_DMABUF_AUTORELEASE is added to toggle the optimization. It is an optional attribute and defaults to 0 (i.e., optimization on).
[...]
Changes in v7:
- use netlink instead of sockopt (Stan)
- restrict system to only one mode, dmabuf bindings can not co-exist with different modes (Stan)
- use static branching to enforce single system-wide mode (Stan)
- Link to v6: https://lore.kernel.org/r/20251104-scratch-bobbyeshleman-devmem-tcp-token-up...
Mina, I was wondering if you had any feedback on this approach?
Best, Bobby
On Tue, Dec 9, 2025 at 11:52 AM Bobby Eshleman bobbyeshleman@gmail.com wrote:
On Wed, Nov 19, 2025 at 07:37:07PM -0800, Bobby Eshleman wrote:
This series improves the CPU cost of RX token management by adding an attribute to NETDEV_CMD_BIND_RX that configures sockets using the binding to avoid the xarray allocator and instead use a per-binding niov array and a uref field in niov.
Improvement is ~13% cpu util per RX user thread.
Using kperf, the following results were observed:
Before: Average RX worker idle %: 13.13, flows 4, test runs 11 After: Average RX worker idle %: 26.32, flows 4, test runs 11
Two other approaches were tested, but with no improvement. Namely, 1) using a hashmap for tokens and 2) keeping an xarray of atomic counters but using RCU so that the hotpath could be mostly lockless. Neither of these approaches proved better than the simple array in terms of CPU.
The attribute NETDEV_A_DMABUF_AUTORELEASE is added to toggle the optimization. It is an optional attribute and defaults to 0 (i.e., optimization on).
[...]
Changes in v7:
- use netlink instead of sockopt (Stan)
- restrict system to only one mode, dmabuf bindings can not co-exist with different modes (Stan)
- use static branching to enforce single system-wide mode (Stan)
- Link to v6: https://lore.kernel.org/r/20251104-scratch-bobbyeshleman-devmem-tcp-token-up...
Mina, I was wondering if you had any feedback on this approach?
Sorry for the delay, I'll take a look shortly.
On 11/19, Bobby Eshleman wrote:
From: Bobby Eshleman bobbyeshleman@meta.com
Add support for autorelease toggling of tokens using a static branch to control system-wide behavior. This allows applications to choose between two memory management modes:
Autorelease on: Leaked tokens are automatically released when the socket closes.
Autorelease off: Leaked tokens are released during dmabuf unbind.
The autorelease mode is requested via the NETDEV_A_DMABUF_AUTORELEASE attribute of the NETDEV_CMD_BIND_RX message. Having separate modes per binding is disallowed and is rejected by netlink. The system will be "locked" into the mode that the first binding is set to. It can only be changed again once there are zero bindings on the system.
Disabling autorelease offers ~13% improvement in CPU utilization.
Static branching is used to limit the system to one mode or the other.
Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com
Changes in v7:
- implement autorelease with static branch (Stan)
- use netlink instead of sockopt (Stan)
- merge uAPI and implementation patches into one patch (seemed less confusing)
Changes in v6:
- remove sk_devmem_info.autorelease, using binding->autorelease instead
- move binding->autorelease check to outside of net_devmem_dmabuf_binding_put_urefs() (Mina)
- remove overly defensive net_is_devmem_iov() (Mina)
- add comment about multiple urefs mapping to a single netmem ref (Mina)
- remove overly defense netmem NULL and netmem_is_net_iov checks (Mina)
- use niov without casting back and forth with netmem (Mina)
- move the autorelease flag from per-binding to per-socket (Mina)
- remove the batching logic in sock_devmem_dontneed_manual_release() (Mina)
- move autorelease check inside tcp_xa_pool_commit() (Mina)
- remove single-binding restriction for autorelease mode (Mina)
- unbind always checks for leaked urefs
Changes in v5:
- remove unused variables
- introduce autorelease flag, preparing for future patch toggle new behavior
Changes in v3:
- make urefs per-binding instead of per-socket, reducing memory footprint
- fallback to cleaning up references in dmabuf unbind if socket leaked tokens
- drop ethtool patch
Changes in v2:
- always use GFP_ZERO for binding->vec (Mina)
- remove WARN for changed binding (Mina)
- remove extraneous binding ref get (Mina)
- remove WARNs on invalid user input (Mina)
- pre-assign niovs in binding->vec for RX case (Mina)
- use atomic_set(, 0) to initialize sk_user_frags.urefs
- fix length of alloc for urefs
Documentation/netlink/specs/netdev.yaml | 12 ++++ include/net/netmem.h | 1 + include/net/sock.h | 7 +- include/uapi/linux/netdev.h | 1 + net/core/devmem.c | 109 +++++++++++++++++++++++++++----- net/core/devmem.h | 11 +++- net/core/netdev-genl-gen.c | 5 +- net/core/netdev-genl.c | 13 +++- net/core/sock.c | 57 +++++++++++++++-- net/ipv4/tcp.c | 78 ++++++++++++++++++----- net/ipv4/tcp_ipv4.c | 13 +++- net/ipv4/tcp_minisocks.c | 3 +- tools/include/uapi/linux/netdev.h | 1 + 13 files changed, 262 insertions(+), 49 deletions(-)
diff --git a/Documentation/netlink/specs/netdev.yaml b/Documentation/netlink/specs/netdev.yaml index 82bf5cb2617d..913fccca4c4e 100644 --- a/Documentation/netlink/specs/netdev.yaml +++ b/Documentation/netlink/specs/netdev.yaml @@ -562,6 +562,17 @@ attribute-sets: type: u32 checks: min: 1
-name: autoreleasedoc: |Token autorelease mode. If true (1), leaked tokens are automaticallyreleased when the socket closes. If false (0), leaked tokens are onlyreleased when the dmabuf is unbound. Once a binding is created with aspecific mode, all subsequent bindings system-wide must use the samemode.Optional. Defaults to false if not specified.type: u8operations: list: @@ -767,6 +778,7 @@ operations: - ifindex - fd - queues
- autorelease reply: attributes: - iddiff --git a/include/net/netmem.h b/include/net/netmem.h index 9e10f4ac50c3..80d2263ba4ed 100644 --- a/include/net/netmem.h +++ b/include/net/netmem.h @@ -112,6 +112,7 @@ struct net_iov { }; struct net_iov_area *owner; enum net_iov_type type;
- atomic_t uref;
}; struct net_iov_area { diff --git a/include/net/sock.h b/include/net/sock.h index a5f36ea9d46f..797b21148945 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -350,7 +350,7 @@ struct sk_filter;
- @sk_scm_rights: flagged by SO_PASSRIGHTS to recv SCM_RIGHTS
- @sk_scm_unused: unused flags for scm_recv()
- @ns_tracker: tracker for netns reference
- @sk_user_frags: xarray of pages the user is holding a reference on.
*/
- @sk_devmem_info: the devmem binding information for the socket
- @sk_owner: reference to the real owner of the socket that calls
sock_lock_init_class_and_name().@@ -579,7 +579,10 @@ struct sock { struct numa_drop_counters *sk_drop_counters; struct rcu_head sk_rcu; netns_tracker ns_tracker;
- struct xarray sk_user_frags;
- struct {
struct xarray frags;struct net_devmem_dmabuf_binding *binding;- } sk_devmem_info;
#if IS_ENABLED(CONFIG_PROVE_LOCKING) && IS_ENABLED(CONFIG_MODULES) struct module *sk_owner; diff --git a/include/uapi/linux/netdev.h b/include/uapi/linux/netdev.h index 048c8de1a130..dff0be8223a4 100644 --- a/include/uapi/linux/netdev.h +++ b/include/uapi/linux/netdev.h @@ -206,6 +206,7 @@ enum { NETDEV_A_DMABUF_QUEUES, NETDEV_A_DMABUF_FD, NETDEV_A_DMABUF_ID,
- NETDEV_A_DMABUF_AUTORELEASE,
__NETDEV_A_DMABUF_MAX, NETDEV_A_DMABUF_MAX = (__NETDEV_A_DMABUF_MAX - 1) diff --git a/net/core/devmem.c b/net/core/devmem.c index 4dee2666dd07..bba21c6cb195 100644 --- a/net/core/devmem.c +++ b/net/core/devmem.c @@ -11,6 +11,7 @@ #include <linux/genalloc.h> #include <linux/mm.h> #include <linux/netdevice.h> +#include <linux/skbuff_ref.h> #include <linux/types.h> #include <net/netdev_queues.h> #include <net/netdev_rx_queue.h> @@ -28,6 +29,17 @@ static DEFINE_XARRAY_FLAGS(net_devmem_dmabuf_bindings, XA_FLAGS_ALLOC1); +/* Static key to lock down autorelease to a single mode on a system. When
- enabled: autorelease mode (leaked tokens released on socket close). When
- disabled: manual mode (leaked tokens released on dmabuf unbind). Once the
- first binding is created, the mode is locked system-wide and can not be
- changed until the system has zero bindings again.
- Protected by xa_lock of net_devmem_dmabuf_bindings.
- */
+DEFINE_STATIC_KEY_FALSE(tcp_devmem_ar_key); +EXPORT_SYMBOL(tcp_devmem_ar_key);
static const struct memory_provider_ops dmabuf_devmem_ops; bool net_is_devmem_iov(struct net_iov *niov) @@ -116,6 +128,24 @@ void net_devmem_free_dmabuf(struct net_iov *niov) gen_pool_free(binding->chunk_pool, dma_addr, PAGE_SIZE); } +static void +net_devmem_dmabuf_binding_put_urefs(struct net_devmem_dmabuf_binding *binding) +{
- int i;
- for (i = 0; i < binding->dmabuf->size / PAGE_SIZE; i++) {
struct net_iov *niov;netmem_ref netmem;niov = binding->vec[i];netmem = net_iov_to_netmem(niov);/* Multiple urefs map to only a single netmem ref. */if (atomic_xchg(&niov->uref, 0) > 0)WARN_ON_ONCE(!napi_pp_put_page(netmem));- }
+}
void net_devmem_unbind_dmabuf(struct net_devmem_dmabuf_binding *binding) { struct netdev_rx_queue *rxq; @@ -143,6 +173,10 @@ void net_devmem_unbind_dmabuf(struct net_devmem_dmabuf_binding *binding) __net_mp_close_rxq(binding->dev, rxq_idx, &mp_params); }
- /* Clean up any lingering urefs from sockets that had autorelease
* disabled.*/- net_devmem_dmabuf_binding_put_urefs(binding); net_devmem_dmabuf_binding_put(binding);
} @@ -179,8 +213,10 @@ struct net_devmem_dmabuf_binding * net_devmem_bind_dmabuf(struct net_device *dev, struct device *dma_dev, enum dma_data_direction direction,
unsigned int dmabuf_fd, struct netdev_nl_sock *priv,struct netlink_ext_ack *extack)
unsigned int dmabuf_fd,struct netdev_nl_sock *priv,struct netlink_ext_ack *extack,bool autorelease){ struct net_devmem_dmabuf_binding *binding; static u32 id_alloc_next; @@ -231,14 +267,13 @@ net_devmem_bind_dmabuf(struct net_device *dev, goto err_detach; }
- if (direction == DMA_TO_DEVICE) {
binding->vec = kvmalloc_array(dmabuf->size / PAGE_SIZE,sizeof(struct net_iov *),GFP_KERNEL);if (!binding->vec) {err = -ENOMEM;goto err_unmap;}
- /* Used by TX and also by RX when socket has autorelease disabled */
- binding->vec = kvmalloc_array(dmabuf->size / PAGE_SIZE,
sizeof(struct net_iov *),GFP_KERNEL | __GFP_ZERO);- if (!binding->vec) {
err = -ENOMEM;goto err_unmap;/* For simplicity we expect to make PAGE_SIZE allocations, but the @@ -292,25 +327,67 @@ net_devmem_bind_dmabuf(struct net_device *dev, niov = &owner->area.niovs[i]; niov->type = NET_IOV_DMABUF; niov->owner = &owner->area;
atomic_set(&niov->uref, 0); page_pool_set_dma_addr_netmem(net_iov_to_netmem(niov), net_devmem_get_dma_addr(niov));
if (direction == DMA_TO_DEVICE)binding->vec[owner->area.base_virtual / PAGE_SIZE + i] = niov;
binding->vec[owner->area.base_virtual / PAGE_SIZE + i] = niov;virtual += len; }
- err = xa_alloc_cyclic(&net_devmem_dmabuf_bindings, &binding->id,
binding, xa_limit_32b, &id_alloc_next,GFP_KERNEL);
- /* Enforce system-wide autorelease mode consistency for RX bindings.
* TX bindings don't use autorelease (always false) since tokens aren't* leaked in TX path. Only RX bindings must all have the same* autorelease mode, never mixed.** We use the xarray's lock to atomically check xa_empty() and toggle* the static key, avoiding the race where two new bindings may try to* set the static key to different states.*/- xa_lock(&net_devmem_dmabuf_bindings);
- if (direction == DMA_FROM_DEVICE) {
if (!xa_empty(&net_devmem_dmabuf_bindings)) {bool mode;mode = static_key_enabled(&tcp_devmem_ar_key);/* When bindings exist, enforce that the mode does not* change.*/if (mode != autorelease) {NL_SET_ERR_MSG_FMT(extack,"System already configured with autorelease=%d",mode);err = -EINVAL;goto err_unlock_xa;}} else {/* First binding sets the mode for all subsequent* bindings.*/if (autorelease)static_branch_enable(&tcp_devmem_ar_key);
[..]
elsestatic_branch_disable(&tcp_devmem_ar_key);
nit: don't think this is needed? DEFINE_STATIC_KEY_FALSE should already have it disabled by default iiuc.
I was also expecting to see something in net_devmem_unbind_dmabuf to undo this static_branch_enable(tcp_devmem_ar_key) and bring the system back into the default state, but I can't find it. Am I missing something?
The rest looks good to me, thanks! Let's see if Mina sees any problems with this approach. IMO, it's a bit easier to reason about two separate code paths now.
On Fri, Nov 21, 2025 at 08:39:50AM -0800, Stanislav Fomichev wrote:
On 11/19, Bobby Eshleman wrote:
From: Bobby Eshleman bobbyeshleman@meta.com
Add support for autorelease toggling of tokens using a static branch to control system-wide behavior. This allows applications to choose between two memory management modes:
Autorelease on: Leaked tokens are automatically released when the socket closes.
Autorelease off: Leaked tokens are released during dmabuf unbind.
The autorelease mode is requested via the NETDEV_A_DMABUF_AUTORELEASE attribute of the NETDEV_CMD_BIND_RX message. Having separate modes per binding is disallowed and is rejected by netlink. The system will be "locked" into the mode that the first binding is set to. It can only be changed again once there are zero bindings on the system.
Disabling autorelease offers ~13% improvement in CPU utilization.
Static branching is used to limit the system to one mode or the other.
Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com
Changes in v7:
- implement autorelease with static branch (Stan)
- use netlink instead of sockopt (Stan)
- merge uAPI and implementation patches into one patch (seemed less confusing)
Changes in v6:
- remove sk_devmem_info.autorelease, using binding->autorelease instead
- move binding->autorelease check to outside of net_devmem_dmabuf_binding_put_urefs() (Mina)
- remove overly defensive net_is_devmem_iov() (Mina)
- add comment about multiple urefs mapping to a single netmem ref (Mina)
- remove overly defense netmem NULL and netmem_is_net_iov checks (Mina)
- use niov without casting back and forth with netmem (Mina)
- move the autorelease flag from per-binding to per-socket (Mina)
- remove the batching logic in sock_devmem_dontneed_manual_release() (Mina)
- move autorelease check inside tcp_xa_pool_commit() (Mina)
- remove single-binding restriction for autorelease mode (Mina)
- unbind always checks for leaked urefs
Changes in v5:
- remove unused variables
- introduce autorelease flag, preparing for future patch toggle new behavior
Changes in v3:
- make urefs per-binding instead of per-socket, reducing memory footprint
- fallback to cleaning up references in dmabuf unbind if socket leaked tokens
- drop ethtool patch
Changes in v2:
- always use GFP_ZERO for binding->vec (Mina)
- remove WARN for changed binding (Mina)
- remove extraneous binding ref get (Mina)
- remove WARNs on invalid user input (Mina)
- pre-assign niovs in binding->vec for RX case (Mina)
- use atomic_set(, 0) to initialize sk_user_frags.urefs
- fix length of alloc for urefs
Documentation/netlink/specs/netdev.yaml | 12 ++++ include/net/netmem.h | 1 + include/net/sock.h | 7 +- include/uapi/linux/netdev.h | 1 + net/core/devmem.c | 109 +++++++++++++++++++++++++++----- net/core/devmem.h | 11 +++- net/core/netdev-genl-gen.c | 5 +- net/core/netdev-genl.c | 13 +++- net/core/sock.c | 57 +++++++++++++++-- net/ipv4/tcp.c | 78 ++++++++++++++++++----- net/ipv4/tcp_ipv4.c | 13 +++- net/ipv4/tcp_minisocks.c | 3 +- tools/include/uapi/linux/netdev.h | 1 + 13 files changed, 262 insertions(+), 49 deletions(-)
diff --git a/Documentation/netlink/specs/netdev.yaml b/Documentation/netlink/specs/netdev.yaml index 82bf5cb2617d..913fccca4c4e 100644 --- a/Documentation/netlink/specs/netdev.yaml +++ b/Documentation/netlink/specs/netdev.yaml @@ -562,6 +562,17 @@ attribute-sets: type: u32 checks: min: 1
-name: autoreleasedoc: |Token autorelease mode. If true (1), leaked tokens are automaticallyreleased when the socket closes. If false (0), leaked tokens are onlyreleased when the dmabuf is unbound. Once a binding is created with aspecific mode, all subsequent bindings system-wide must use the samemode.Optional. Defaults to false if not specified.type: u8operations: list: @@ -767,6 +778,7 @@ operations: - ifindex - fd - queues
- autorelease reply: attributes: - iddiff --git a/include/net/netmem.h b/include/net/netmem.h index 9e10f4ac50c3..80d2263ba4ed 100644 --- a/include/net/netmem.h +++ b/include/net/netmem.h @@ -112,6 +112,7 @@ struct net_iov { }; struct net_iov_area *owner; enum net_iov_type type;
- atomic_t uref;
}; struct net_iov_area { diff --git a/include/net/sock.h b/include/net/sock.h index a5f36ea9d46f..797b21148945 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -350,7 +350,7 @@ struct sk_filter;
- @sk_scm_rights: flagged by SO_PASSRIGHTS to recv SCM_RIGHTS
- @sk_scm_unused: unused flags for scm_recv()
- @ns_tracker: tracker for netns reference
- @sk_user_frags: xarray of pages the user is holding a reference on.
*/
- @sk_devmem_info: the devmem binding information for the socket
- @sk_owner: reference to the real owner of the socket that calls
sock_lock_init_class_and_name().@@ -579,7 +579,10 @@ struct sock { struct numa_drop_counters *sk_drop_counters; struct rcu_head sk_rcu; netns_tracker ns_tracker;
- struct xarray sk_user_frags;
- struct {
struct xarray frags;struct net_devmem_dmabuf_binding *binding;- } sk_devmem_info;
#if IS_ENABLED(CONFIG_PROVE_LOCKING) && IS_ENABLED(CONFIG_MODULES) struct module *sk_owner; diff --git a/include/uapi/linux/netdev.h b/include/uapi/linux/netdev.h index 048c8de1a130..dff0be8223a4 100644 --- a/include/uapi/linux/netdev.h +++ b/include/uapi/linux/netdev.h @@ -206,6 +206,7 @@ enum { NETDEV_A_DMABUF_QUEUES, NETDEV_A_DMABUF_FD, NETDEV_A_DMABUF_ID,
- NETDEV_A_DMABUF_AUTORELEASE,
__NETDEV_A_DMABUF_MAX, NETDEV_A_DMABUF_MAX = (__NETDEV_A_DMABUF_MAX - 1) diff --git a/net/core/devmem.c b/net/core/devmem.c index 4dee2666dd07..bba21c6cb195 100644 --- a/net/core/devmem.c +++ b/net/core/devmem.c @@ -11,6 +11,7 @@ #include <linux/genalloc.h> #include <linux/mm.h> #include <linux/netdevice.h> +#include <linux/skbuff_ref.h> #include <linux/types.h> #include <net/netdev_queues.h> #include <net/netdev_rx_queue.h> @@ -28,6 +29,17 @@ static DEFINE_XARRAY_FLAGS(net_devmem_dmabuf_bindings, XA_FLAGS_ALLOC1); +/* Static key to lock down autorelease to a single mode on a system. When
- enabled: autorelease mode (leaked tokens released on socket close). When
- disabled: manual mode (leaked tokens released on dmabuf unbind). Once the
- first binding is created, the mode is locked system-wide and can not be
- changed until the system has zero bindings again.
- Protected by xa_lock of net_devmem_dmabuf_bindings.
- */
+DEFINE_STATIC_KEY_FALSE(tcp_devmem_ar_key); +EXPORT_SYMBOL(tcp_devmem_ar_key);
static const struct memory_provider_ops dmabuf_devmem_ops; bool net_is_devmem_iov(struct net_iov *niov) @@ -116,6 +128,24 @@ void net_devmem_free_dmabuf(struct net_iov *niov) gen_pool_free(binding->chunk_pool, dma_addr, PAGE_SIZE); } +static void +net_devmem_dmabuf_binding_put_urefs(struct net_devmem_dmabuf_binding *binding) +{
- int i;
- for (i = 0; i < binding->dmabuf->size / PAGE_SIZE; i++) {
struct net_iov *niov;netmem_ref netmem;niov = binding->vec[i];netmem = net_iov_to_netmem(niov);/* Multiple urefs map to only a single netmem ref. */if (atomic_xchg(&niov->uref, 0) > 0)WARN_ON_ONCE(!napi_pp_put_page(netmem));- }
+}
void net_devmem_unbind_dmabuf(struct net_devmem_dmabuf_binding *binding) { struct netdev_rx_queue *rxq; @@ -143,6 +173,10 @@ void net_devmem_unbind_dmabuf(struct net_devmem_dmabuf_binding *binding) __net_mp_close_rxq(binding->dev, rxq_idx, &mp_params); }
- /* Clean up any lingering urefs from sockets that had autorelease
* disabled.*/- net_devmem_dmabuf_binding_put_urefs(binding); net_devmem_dmabuf_binding_put(binding);
} @@ -179,8 +213,10 @@ struct net_devmem_dmabuf_binding * net_devmem_bind_dmabuf(struct net_device *dev, struct device *dma_dev, enum dma_data_direction direction,
unsigned int dmabuf_fd, struct netdev_nl_sock *priv,struct netlink_ext_ack *extack)
unsigned int dmabuf_fd,struct netdev_nl_sock *priv,struct netlink_ext_ack *extack,bool autorelease){ struct net_devmem_dmabuf_binding *binding; static u32 id_alloc_next; @@ -231,14 +267,13 @@ net_devmem_bind_dmabuf(struct net_device *dev, goto err_detach; }
- if (direction == DMA_TO_DEVICE) {
binding->vec = kvmalloc_array(dmabuf->size / PAGE_SIZE,sizeof(struct net_iov *),GFP_KERNEL);if (!binding->vec) {err = -ENOMEM;goto err_unmap;}
- /* Used by TX and also by RX when socket has autorelease disabled */
- binding->vec = kvmalloc_array(dmabuf->size / PAGE_SIZE,
sizeof(struct net_iov *),GFP_KERNEL | __GFP_ZERO);- if (!binding->vec) {
err = -ENOMEM;goto err_unmap;/* For simplicity we expect to make PAGE_SIZE allocations, but the @@ -292,25 +327,67 @@ net_devmem_bind_dmabuf(struct net_device *dev, niov = &owner->area.niovs[i]; niov->type = NET_IOV_DMABUF; niov->owner = &owner->area;
atomic_set(&niov->uref, 0); page_pool_set_dma_addr_netmem(net_iov_to_netmem(niov), net_devmem_get_dma_addr(niov));
if (direction == DMA_TO_DEVICE)binding->vec[owner->area.base_virtual / PAGE_SIZE + i] = niov;
binding->vec[owner->area.base_virtual / PAGE_SIZE + i] = niov;virtual += len; }
- err = xa_alloc_cyclic(&net_devmem_dmabuf_bindings, &binding->id,
binding, xa_limit_32b, &id_alloc_next,GFP_KERNEL);
- /* Enforce system-wide autorelease mode consistency for RX bindings.
* TX bindings don't use autorelease (always false) since tokens aren't* leaked in TX path. Only RX bindings must all have the same* autorelease mode, never mixed.** We use the xarray's lock to atomically check xa_empty() and toggle* the static key, avoiding the race where two new bindings may try to* set the static key to different states.*/- xa_lock(&net_devmem_dmabuf_bindings);
- if (direction == DMA_FROM_DEVICE) {
if (!xa_empty(&net_devmem_dmabuf_bindings)) {bool mode;mode = static_key_enabled(&tcp_devmem_ar_key);/* When bindings exist, enforce that the mode does not* change.*/if (mode != autorelease) {NL_SET_ERR_MSG_FMT(extack,"System already configured with autorelease=%d",mode);err = -EINVAL;goto err_unlock_xa;}} else {/* First binding sets the mode for all subsequent* bindings.*/if (autorelease)static_branch_enable(&tcp_devmem_ar_key);[..]
elsestatic_branch_disable(&tcp_devmem_ar_key);nit: don't think this is needed? DEFINE_STATIC_KEY_FALSE should already have it disabled by default iiuc.
I was also expecting to see something in net_devmem_unbind_dmabuf to undo this static_branch_enable(tcp_devmem_ar_key) and bring the system back into the default state, but I can't find it. Am I missing something?
So the way I've written it here is that the branch state always updates according to the mode of the first binding, so we can just ignore the branch state when unbind reaches zero bindings (it will update the next time we see a first binding anyway).
... but I think your mental model is a bit simpler and probably fewer lines, and we can move to it for the next rev if Mina is okay with the overall approach.
In that, we enforce that when no bindings are alive the branch is disabled, and only updates when the first binding mode is autorelease=1, the autorelease=0 case is a no-op when binding. Unbind restores to disabled when it sees no more bindings.
The rest looks good to me, thanks! Let's see if Mina sees any problems with this approach. IMO, it's a bit easier to reason about two separate code paths now.
Thanks!
Best, Bobby
On Wed, Nov 19, 2025 at 07:37:12PM -0800, Bobby Eshleman wrote:
From: Bobby Eshleman bobbyeshleman@meta.com
Add tests cases that check the autorelease modes (on and off). The new tests are the same as the old ones, but just pass a flag to ncdevmem to select the autorelease mode.
Only for RX tests is autorelease checked, as the autorelease ncdevmem flag is unused in the TX case and doesn't apply to TX bind operations.
Signed-off-by: Bobby Eshleman bobbyeshleman@meta.com
Note: I tested successfully with kperf, but I'm troubleshooting some mlx5 issues with ncdevmem so this patch, though simple, is not fully validated. Will respond to this thread when solve the issue.
Changes in v7:
- use autorelease netlink
- remove sockopt tests
tools/testing/selftests/drivers/net/hw/devmem.py | 22 ++++++++++++++++++++-- tools/testing/selftests/drivers/net/hw/ncdevmem.c | 19 +++++++++++++------ 2 files changed, 33 insertions(+), 8 deletions(-)
diff --git a/tools/testing/selftests/drivers/net/hw/devmem.py b/tools/testing/selftests/drivers/net/hw/devmem.py index 45c2d49d55b6..dddb9d77cb28 100755 --- a/tools/testing/selftests/drivers/net/hw/devmem.py +++ b/tools/testing/selftests/drivers/net/hw/devmem.py @@ -25,7 +25,23 @@ def check_rx(cfg) -> None: port = rand_port() socat = f"socat -u - TCP{cfg.addr_ipver}:{cfg.baddr}:{port},bind={cfg.remote_baddr}:{port}"
- listen_cmd = f"{cfg.bin_local} -l -f {cfg.ifname} -s {cfg.addr} -p {port} -c {cfg.remote_addr} -v 7"
- listen_cmd = f"{cfg.bin_local} -l -f {cfg.ifname} -s {cfg.addr} -p {port} -c {cfg.remote_addr} -v 7 -a 0"
- with bkg(listen_cmd, exit_wait=True) as ncdevmem:
wait_port_listen(port)cmd(f"yes $(echo -e \x01\x02\x03\x04\x05\x06) | \head -c 1K | {socat}", host=cfg.remote, shell=True)- ksft_eq(ncdevmem.ret, 0)
+@ksft_disruptive +def check_rx_autorelease(cfg) -> None:
- require_devmem(cfg)
- port = rand_port()
- socat = f"socat -u - TCP{cfg.addr_ipver}:{cfg.baddr}:{port},bind={cfg.remote_baddr}:{port}"
- listen_cmd = f"{cfg.bin_local} -l -f {cfg.ifname} -s {cfg.addr} -p {port} -c {cfg.remote_addr} -v 7 -a 1"
with bkg(listen_cmd, exit_wait=True) as ncdevmem: wait_port_listen(port) @@ -68,7 +84,9 @@ def main() -> None: cfg.bin_local = path.abspath(path.dirname(__file__) + "/ncdevmem") cfg.bin_remote = cfg.remote.deploy(cfg.bin_local)
ksft_run([check_rx, check_tx, check_tx_chunks],
ksft_run([check_rx, check_rx_autorelease,check_tx, check_tx_autorelease,check_tx_chunks, check_tx_chunks_autorelease], args=(cfg, ))
Hi Bobby,
Sorry if this is a duplicate: I drafted this before bit AFAIKT I accidently hit the worng key and deleted it.
I see that check_rx_autorelease is implemented above, but I don't see an implementatoin of check_tx_autorelease or check_tx_chunks_autorelease. Perhaps they should be added too?
Flagged by pylint and ruff. Also confirmed by running devmem.py using vng (possibly incorrectly).
This patch also adds some other pylint warnings, about missing documentation and lines that are too long. Although there are already instances of the same warning in this file. So I'm not sure how we should feel about that.
...
On Wed, Nov 19, 2025 at 07:37:10PM -0800, Bobby Eshleman wrote:
...
@@ -292,25 +327,67 @@ net_devmem_bind_dmabuf(struct net_device *dev,
...
- /* Enforce system-wide autorelease mode consistency for RX bindings.
* TX bindings don't use autorelease (always false) since tokens aren't* leaked in TX path. Only RX bindings must all have the same* autorelease mode, never mixed.** We use the xarray's lock to atomically check xa_empty() and toggle* the static key, avoiding the race where two new bindings may try to* set the static key to different states.*/- xa_lock(&net_devmem_dmabuf_bindings);
- if (direction == DMA_FROM_DEVICE) {
if (!xa_empty(&net_devmem_dmabuf_bindings)) {bool mode;mode = static_key_enabled(&tcp_devmem_ar_key);/* When bindings exist, enforce that the mode does not* change.*/if (mode != autorelease) {NL_SET_ERR_MSG_FMT(extack,"System already configured with autorelease=%d",mode);err = -EINVAL;goto err_unlock_xa;}} else {/* First binding sets the mode for all subsequent* bindings.*/if (autorelease)static_branch_enable(&tcp_devmem_ar_key);elsestatic_branch_disable(&tcp_devmem_ar_key);
Hi Bobby,
This code runs inside xa_lock, which is a spinlock.
But static_branch_enable() and static_branch_disable() may sleep due to some combination of taking a mutex and cpu_read_lock.
Flagged by Claude Code with https://github.com/masoncl/review-prompts/
}- }
- err = __xa_alloc_cyclic(&net_devmem_dmabuf_bindings, &binding->id,
binding, xa_limit_32b, &id_alloc_next,GFP_KERNEL);
goto err_free_chunks;
goto err_unlock_xa;- xa_unlock(&net_devmem_dmabuf_bindings);
list_add(&binding->list, &priv->bindings); return binding; +err_unlock_xa:
- xa_unlock(&net_devmem_dmabuf_bindings);
...
linux-kselftest-mirror@lists.linaro.org
-
Bobby Eshleman -
Mina Almasry -
Paolo Abeni -
Simon Horman -
Stanislav Fomichev