kernel/drivers/net/ethernet/pensando/ionic/ionic_txrx.c, branch linux-6.18.y

net: ionic: map SKB after pseudo-header checksum prep

2025-11-04T01:19:21Z

The TSO path called ionic_tx_map_skb() before preparing the TCP pseudo checksum (ionic_tx_tcp_[inner_]pseudo_csum()), which may perform skb_cow_head() and might modifies bytes in the linear header area. Mapping first and then mutating the header risks: - Using a stale DMA address if skb_cow_head() relocates the head, and/or - Device reading stale header bytes on weakly-ordered systems (CPU writes after mapping are not guaranteed visible without an explicit dma_sync_single_for_device()). Reorder the TX path to perform all header mutations (including skb_cow_head()) *before* DMA mapping. Mapping is now done only after the skb layout and header contents are final. This removes the need for any post-mapping dma_sync and prevents on-wire corruption observed under VLAN+TSO load after repeated runs. This change is purely an ordering fix; no functional behavior change otherwise. Fixes: 0f3154e6bcb3 ("ionic: Add Tx and Rx handling") Signed-off-by: Mohammad Heib Reviewed-by: Brett Creeley Link: https://patch.msgid.link/20251031155203.203031-2-mheib@redhat.com Signed-off-by: Jakub Kicinski

net: ionic: add dma_wmb() before ringing TX doorbell

2025-11-04T01:19:21Z

The TX path currently writes descriptors and then immediately writes to the MMIO doorbell register to notify the NIC. On weakly ordered architectures, descriptor writes may still be pending in CPU or DMA write buffers when the doorbell is issued, leading to the device fetching stale or incomplete descriptors. Add a dma_wmb() in ionic_txq_post() to ensure all descriptor writes are visible to the device before the doorbell MMIO write. Fixes: 0f3154e6bcb3 ("ionic: Add Tx and Rx handling") Signed-off-by: Mohammad Heib Link: https://patch.msgid.link/20251031155203.203031-1-mheib@redhat.com Signed-off-by: Jakub Kicinski

ethernet: ionic: Fix DMA mapping tests

2025-06-23T23:42:39Z

Change error values of `ionic_tx_map_single()` and `ionic_tx_map_frag()` from 0 to `DMA_MAPPING_ERROR` to prevent collision with 0 as a valid address. This also fixes the use of `dma_mapping_error()` to test against 0 in `ionic_xdp_post_frame()` Fixes: 0f3154e6bcb3 ("ionic: Add Tx and Rx handling") Fixes: 56e41ee12d2d ("ionic: better dma-map error handling") Fixes: ac8813c0ab7d ("ionic: convert Rx queue buffers to use page_pool") Signed-off-by: Thomas Fourier Reviewed-by: Brett Creeley Link: https://patch.msgid.link/20250619094538.283723-2-fourier.thomas@gmail.com Signed-off-by: Jakub Kicinski

dim: pass dim_sample to net_dim() by reference

2024-11-03T20:36:54Z

net_dim() is currently passed a struct dim_sample argument by value. struct dim_sample is 24 bytes. Since this is greater 16 bytes, x86-64 passes it on the stack. All callers have already initialized dim_sample on the stack, so passing it by value requires pushing a duplicated copy to the stack. Either witing to the stack and immediately reading it, or perhaps dereferencing addresses relative to the stack pointer in a chain of push instructions, seems to perform quite poorly. In a heavy TCP workload, mlx5e_handle_rx_dim() consumes 3% of CPU time, 94% of which is attributed to the first push instruction to copy dim_sample on the stack for the call to net_dim(): // Call ktime_get() 0.26 |4ead2: call 4ead7 // Pass the address of struct dim in %rdi |4ead7: lea 0x3d0(%rbx),%rdi // Set dim_sample.pkt_ctr |4eade: mov %r13d,0x8(%rsp) // Set dim_sample.byte_ctr |4eae3: mov %r12d,0xc(%rsp) // Set dim_sample.event_ctr 0.15 |4eae8: mov %bp,0x10(%rsp) // Duplicate dim_sample on the stack 94.16 |4eaed: push 0x10(%rsp) 2.79 |4eaf1: push 0x10(%rsp) 0.07 |4eaf5: push %rax // Call net_dim() 0.21 |4eaf6: call 4eafb To allow the caller to reuse the struct dim_sample already on the stack, pass the struct dim_sample by reference to net_dim(). Signed-off-by: Caleb Sander Mateos Reviewed-by: Vladimir Oltean Reviewed-by: Shannon Nelson Reviewed-by: Florian Fainelli Reviewed-by: Arthur Kiyanovski Reviewed-by: Louis Peens Link: https://patch.msgid.link/20241031002326.3426181-2-csander@purestorage.com Signed-off-by: Jakub Kicinski

ionic: Allow XDP program to be hot swapped

2024-09-10T02:18:15Z

Using examples of other driver(s), add the ability to hot-swap an XDP program without having to reconfigure the queues. To prevent the q->xdp_prog to be read/written more than once use READ_ONCE() and WRITE_ONCE() on the q->xdp_prog. The q->xdp_prog was being checked in multiple different for loops in the hot path. The change to allow xdp_prog hot swapping created the possibility for many READ_ONCE(q->xdp_prog) calls during a single napi callback. Refactor the Rx napi handling to allow a previous READ_ONCE(q->xdp_prog) (or NULL for hwstamp_rxq) to be passed into the relevant functions. Also, move other Rx related hotpath handling into the newly created ionic_rx_cq_service() function to reduce the scope of the xdp_prog local variable and put all Rx handling in one function similar to Tx. Signed-off-by: Brett Creeley Signed-off-by: Shannon Nelson Link: https://patch.msgid.link/20240906232623.39651-8-brett.creeley@amd.com Signed-off-by: Jakub Kicinski

ionic: convert Rx queue buffers to use page_pool

2024-09-10T02:18:15Z

Our home-grown buffer management needs to go away and we need to be playing nicely with the page_pool infrastructure. This converts the Rx traffic queues to use page_pool. Also, since ionic_rx_buf_size() was removed, redefine IONIC_PAGE_SIZE to account for IONIC_MAX_BUF_LEN being the largest allowed buffer to prevent overflowing u16 variables, which could happen when PAGE_SIZE is defined as >= 64KB. include/linux/minmax.h:93:37: warning: conversion from 'long unsigned int' to 'u16' {aka 'short unsigned int'} changes value from '65536' to '0' [-Woverflow] Signed-off-by: Shannon Nelson Signed-off-by: Brett Creeley Link: https://patch.msgid.link/20240906232623.39651-7-brett.creeley@amd.com Signed-off-by: Jakub Kicinski

ionic: use per-queue xdp_prog

2024-09-10T02:18:14Z

We originally were using a per-interface xdp_prog variable to track a loaded XDP program since we knew there would never be support for a per-queue XDP program. With that, we only built the per queue rxq_info struct when an XDP program was loaded and removed it on XDP program unload, and used the pointer as an indicator in the Rx hotpath to know to how build the buffers. However, that's really not the model generally used, and makes a conversion to page_pool Rx buffer cacheing a little problematic. This patch converts the driver to use the more common approach of using a per-queue xdp_prog pointer to work out buffer allocations and need for bpf_prog_run_xdp(). We jostle a couple of fields in the queue struct in order to keep the new xdp_prog pointer in a warm cacheline. Signed-off-by: Shannon Nelson Signed-off-by: Brett Creeley Link: https://patch.msgid.link/20240906232623.39651-4-brett.creeley@amd.com Signed-off-by: Jakub Kicinski

ionic: rename ionic_xdp_rx_put_bufs

2024-09-10T02:18:14Z

We aren't "putting" buf, we're just unlinking them from our tracking in order to let the XDP_TX and XDP_REDIRECT tx clean paths take care of the pages when they are done with them. This rename clears up the intent. Signed-off-by: Shannon Nelson Signed-off-by: Brett Creeley Link: https://patch.msgid.link/20240906232623.39651-3-brett.creeley@amd.com Signed-off-by: Jakub Kicinski

ionic: debug line for Tx completion errors

2024-09-10T02:18:14Z

Here's a little debugging aid in case the device starts throwing Tx completion errors. Signed-off-by: Shannon Nelson Signed-off-by: Brett Creeley Link: https://patch.msgid.link/20240906232623.39651-2-brett.creeley@amd.com Signed-off-by: Jakub Kicinski

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

2024-06-27T19:14:11Z

Cross-merge networking fixes after downstream PR. No conflicts. Adjacent changes: e3f02f32a050 ("ionic: fix kernel panic due to multi-buffer handling") d9c04209990b ("ionic: Mark error paths in the data path as unlikely") Signed-off-by: Jakub Kicinski