kernel/git/stable/linux.git/drivers/net/ethernet/microsoft, branch master

net: mana: Cap MSI-X vectors to the device MSI-X table size

2026-08-24T19:00:31+00:00

mana_gd_query_max_resources() sizes gc->num_msix_usable from resp.max_msix and the CPU count, but never from the device MSI-X table. On a 1792 vCPU M-series VM that yields 1793 while the table has 1024 entries, and mana_gd_setup_remaining_irqs() then walks indices 1..1792, running off the end of the region mapped by msix_map_region(): BUG: unable to handle page fault for address: ff8e347f8b99800c RIP: 0010:msix_prepare_msi_desc+0x7a/0x90 RAX: 0000000000004000 RBX: ff4330cb164ea780 RCX: ff8e347f8b998000 Call Trace: __msi_domain_alloc_irqs+0x13a/0x440 msi_domain_alloc_irq_at+0x149/0x1b0 mana_gd_setup+0x351/0x890 mana_gd_probe+0x274/0x390 RAX is index 1024 * PCI_MSIX_ENTRY_SIZE, one entry past the table. msi_insert_desc() does range check the index, but only against the MSI domain hwsize, which matches the table only for devices on an MSI parent domain. With a global PCI/MSI domain hwsize is MSI_XA_DOMAIN_SIZE, so nothing bounds the request. Cap num_msix_usable with pci_msix_vec_count(). Fixes: 755391121038 ("net: mana: Allocate MSI-X vectors dynamically") Signed-off-by: Long Li Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260821183736.733296-1-longli@microsoft.com Signed-off-by: Jakub Kicinski

net: mana: Fall back to scattered pages for GDMA queues

2026-08-13T11:42:48+00:00

Each GDMA queue ring is one dma_alloc_coherent() of the whole ring size. Such high-order allocations fail first under memory fragmentation, so queue setup can fail with memory still free. The hardware does not need the ring physically contiguous: mana_gd_create_dma_region() already maps it as a list of MANA_PAGE_SIZE (4K) device addresses. Only the driver's linear CPU view needs contiguity, and it goes through mana_gd_ring_ptr() and mana_gd_ring_contig_avail(); change both to map offsets onto scattered pages. Add a fallback in mana_gd_alloc_memory(): data-path queues pass allow_scatter=true, so when the contiguous allocation fails the ring is backed by a vector of scattered PAGE_SIZE (order-0) coherent pages, presenting the same DMA page-list layout to the device. The HW channel bootstrap keeps allow_scatter=false, and the debugfs ring dumper reads scattered rings through the same helpers. Signed-off-by: Aditya Garg Link: https://patch.msgid.link/20260807210002.1695263-3-gargaditya@linux.microsoft.com Signed-off-by: Paolo Abeni

net: mana: Route ring-buffer access through offset-based helpers

2026-08-13T11:42:48+00:00

In preparation for backing GDMA queue memory with a vector of non-contiguous order-0 coherent pages, route CPU access to a queue's ring buffer through two new helpers: mana_gd_ring_ptr() returns the CPU address of a byte offset into the ring, and mana_gd_ring_contig_avail() the number of bytes left before the ring wraps, so a WQ write that runs past the end of the ring can be split at that point. Convert the EQ, CQ and work-request paths to use them. mana_gd_write_sgl() now takes a byte offset rather than a raw pointer, so mana_gd_post_work_request() derives the SGL position arithmetically. While queue memory is contiguous both helpers are simple arithmetic on the ring base and size, so there is no functional change. Signed-off-by: Aditya Garg Link: https://patch.msgid.link/20260807210002.1695263-2-gargaditya@linux.microsoft.com Signed-off-by: Paolo Abeni

net: mana: Extend RX CQE coalescing up to 8 packets

2026-08-08T01:29:21+00:00

To support up to 8 packets per CQE, update related CQE processing code and structures. Update ethtool handlers to set this feature. Update per queue stat to show the coalesced CQE counters. This feature is supported on NIC hardware showing the relevant PF flag. Signed-off-by: Haiyang Zhang Reviewed-by: Simon Horman Reviewed-by: Breno Leitao Link: https://patch.msgid.link/20260805185404.1052177-1-haiyangz@linux.microsoft.com Signed-off-by: Jakub Kicinski

net: mana: force full-page RX buffers via ethtool private flag

2026-08-04T02:09:07+00:00

On some ARM64 platforms with 4K PAGE_SIZE, page_pool fragment allocation in the RX refill path can cause 15-20% throughput regression under high connection counts (>16 TCP streams). Add an ethtool private flag "full-page-rx" that allows the user to force one RX buffer per page, bypassing the page_pool fragment path. This restores line-rate (180+ Gbps) performance on affected platforms. Usage: ethtool --set-priv-flags eth0 full-page-rx on There is no behavioral change by default. The flag must be explicitly enabled by the user or udev rule. The existing single-buffer-per-page logic for XDP and jumbo frames is consolidated into a new helper mana_use_single_rxbuf_per_page() which is now the single decision point for both the automatic and user-controlled paths. Reviewed-by: Jacob Keller Reviewed-by: Haiyang Zhang Signed-off-by: Dipayaan Roy Link: https://patch.msgid.link/20260729063347.3388035-3-dipayanroy@linux.microsoft.com Signed-off-by: Jakub Kicinski

net: mana: refactor mana_get_strings() and mana_get_sset_count() to use switch

2026-08-04T02:09:07+00:00

Refactor mana_get_strings() and mana_get_sset_count() from if/else to switch statements in preparation for adding ethtool private flags support which requires handling ETH_SS_PRIV_FLAGS. No functional change. Reviewed-by: Haiyang Zhang Signed-off-by: Dipayaan Roy Link: https://patch.msgid.link/20260729063347.3388035-2-dipayanroy@linux.microsoft.com Signed-off-by: Jakub Kicinski

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2026-07-30T19:53:19+00:00

Cross-merge networking fixes after downstream PR (net-7.2-rc6). No conflicts. Adjacent changes: net/ipv4/route.c dbc3791e3b24 ("net: do not send ICMP/NDISC Redirects when peer allocation fails") 7804eaa057fe ("ipv4: snapshot dst.dev in ip_rt_send_redirect() and ip_rt_get_source()") drivers/net/tun.c 23dad2d088df ("tun: no longer rely on RTNL in tun_fill_info()") c3da92af07ea ("Revert "tun/tap: add ptr_ring consume helper with netdev queue wakeup"") drivers/net/ethernet/marvell/octeontx2/af/rvu_npc.c 3bd438a58e91 ("octeontx2-af: Block VFs from clobbering special CGX PKIND state") 5ba5611ef946 ("octeontx2-af: reserve 4 PKINDs for skip-size custom use") drivers/net/wireless/ath/ath12k/core.h drivers/net/wireless/ath/ath12k/mac.c drivers/net/wireless/ath/ath12k/peer.c 469d7e6077c1 ("wifi: ath12k: resolve PENDING ML peer ID from MLO_PEER_MAP HTT event") 378e659029d5 ("wifi: ath12k: introduce host_alloc_ml_id hardware parameter") c42b27336eef ("wifi: ath12k: fix survey indexing across bands") Signed-off-by: Jakub Kicinski

net: mana: Return error code from mana_create_rxq()

2026-07-30T00:46:35+00:00

mana_create_rxq() returns a struct mana_rxq pointer and returns NULL on any failure. The caller, mana_add_rx_queues(), cannot tell what went wrong and hardcodes the error as -ENOMEM. As a result the actual failure reported by the lower layers (for example -EPROTO from a failed HW request) is masked and every RX queue creation failure looks like an out-of-memory error. Return an ERR_PTR() encoded error code from mana_create_rxq() on failure instead of NULL. The caller now propagates the returned error code directly instead of substituting -ENOMEM. Fixes: ca9c54d2d6a5 ("net: mana: Add a driver for Microsoft Azure Network Adapter (MANA)") Signed-off-by: Aditya Garg Reviewed-by: Joe Damato Link: https://patch.msgid.link/20260727113759.2881500-1-gargaditya@linux.microsoft.com Signed-off-by: Jakub Kicinski

net: mana: Add handler for sriov configure

2026-07-17T13:31:35+00:00

Add callback function for the pci_driver / sriov_configure. It asks the NIC to provide certain number of VFs, or disable VFs if the request is zero. Signed-off-by: Haiyang Zhang Link: https://patch.msgid.link/20260710192735.2921300-1-haiyangz@linux.microsoft.com Signed-off-by: Paolo Abeni

net: mana: Add debug knob to skip TX timeout recovery reset

2026-07-17T11:25:20+00:00

Add a per-port debugfs boolean "tx_timeout_skip_reset" that, when enabled, makes mana_tx_timeout() log the TX timeout and return without queueing the per-port detach/attach recovery work. This is a debug-only aid for bringup and qualification: skipping the recovery reset keeps the device and queue state intact so a TX timeout can be correlated with hardware telemetry. The knob defaults to false, so production recovery behaviour is unchanged. Signed-off-by: Aditya Garg Reviewed-by: Haiyang Zhang Reviewed-by: Dipayaan Roy Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260710132229.2851441-1-gargaditya@linux.microsoft.com Signed-off-by: Paolo Abeni