kernel/linux.git/drivers/net/ethernet/microsoft, branch v6.19.11

net: mana: fix use-after-free in mana_hwc_destroy_channel() by reordering teardown

2026-03-25T10:13:26+00:00

[ Upstream commit fa103fc8f56954a60699a29215cb713448a39e87 ] A potential race condition exists in mana_hwc_destroy_channel() where hwc->caller_ctx is freed before the HWC's Completion Queue (CQ) and Event Queue (EQ) are destroyed. This allows an in-flight CQ interrupt handler to dereference freed memory, leading to a use-after-free or NULL pointer dereference in mana_hwc_handle_resp(). mana_smc_teardown_hwc() signals the hardware to stop but does not synchronize against IRQ handlers already executing on other CPUs. The IRQ synchronization only happens in mana_hwc_destroy_cq() via mana_gd_destroy_eq() -> mana_gd_deregister_irq(). Since this runs after kfree(hwc->caller_ctx), a concurrent mana_hwc_rx_event_handler() can dereference freed caller_ctx (and rxq->msg_buf) in mana_hwc_handle_resp(). Fix this by reordering teardown to reverse-of-creation order: destroy the TX/RX work queues and CQ/EQ before freeing hwc->caller_ctx. This ensures all in-flight interrupt handlers complete before the memory they access is freed. Fixes: ca9c54d2d6a5 ("net: mana: Add a driver for Microsoft Azure Network Adapter (MANA)") Reviewed-by: Haiyang Zhang Signed-off-by: Dipayaan Roy Reviewed-by: Simon Horman Link: https://patch.msgid.link/abHA3AjNtqa1nx9k@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

net: mana: Ring doorbell at 4 CQ wraparounds

2026-03-19T15:15:18+00:00

commit dabffd08545ffa1d7183bc45e387860984025291 upstream. MANA hardware requires at least one doorbell ring every 8 wraparounds of the CQ. The driver rings the doorbell as a form of flow control to inform hardware that CQEs have been consumed. The NAPI poll functions mana_poll_tx_cq() and mana_poll_rx_cq() can poll up to CQE_POLLING_BUFFER (512) completions per call. If the CQ has fewer than 512 entries, a single poll call can process more than 4 wraparounds without ringing the doorbell. The doorbell threshold check also uses ">" instead of ">=", delaying the ring by one extra CQE beyond 4 wraparounds. Combined, these issues can cause the driver to exceed the 8-wraparound hardware limit, leading to missed completions and stalled queues. Fix this by capping the number of CQEs polled per call to 4 wraparounds of the CQ in both TX and RX paths. Also change the doorbell threshold from ">" to ">=" so the doorbell is rung as soon as 4 wraparounds are reached. Cc: stable@vger.kernel.org Fixes: 58a63729c957 ("net: mana: Fix doorbell out of order violation and avoid unnecessary doorbell rings") Signed-off-by: Long Li Reviewed-by: Haiyang Zhang Reviewed-by: Vadim Fedorenko Link: https://patch.msgid.link/20260226192833.1050807-1-longli@microsoft.com Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman

net/mana: Null service_wq on setup error to prevent double destroy

2026-03-19T15:14:58+00:00

[ Upstream commit 87c2302813abc55c46485711a678e3c312b00666 ] In mana_gd_setup() error path, set gc->service_wq to NULL after destroy_workqueue() to match the cleanup in mana_gd_cleanup(). This prevents a use-after-free if the workqueue pointer is checked after a failed setup. Fixes: f975a0955276 ("net: mana: Fix double destroy_workqueue on service rescan PCI path") Signed-off-by: Shiraz Saleem Signed-off-by: Konstantin Taranov Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260309172443.688392-1-kotaranov@linux.microsoft.com Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

net: mana: Fix double destroy_workqueue on service rescan PCI path

2026-03-04T12:20:57+00:00

[ Upstream commit f975a0955276579e2176a134366ed586071c7c6a ] While testing corner cases in the driver, a use-after-free crash was found on the service rescan PCI path. When mana_serv_reset() calls mana_gd_suspend(), mana_gd_cleanup() destroys gc->service_wq. If the subsequent mana_gd_resume() fails with -ETIMEDOUT or -EPROTO, the code falls through to mana_serv_rescan() which triggers pci_stop_and_remove_bus_device(). This invokes the PCI .remove callback (mana_gd_remove), which calls mana_gd_cleanup() a second time, attempting to destroy the already- freed workqueue. Fix this by NULL-checking gc->service_wq in mana_gd_cleanup() and setting it to NULL after destruction. Call stack of issue for reference: [Sat Feb 21 18:53:48 2026] Call Trace: [Sat Feb 21 18:53:48 2026] [Sat Feb 21 18:53:48 2026] mana_gd_cleanup+0x33/0x70 [mana] [Sat Feb 21 18:53:48 2026] mana_gd_remove+0x3a/0xc0 [mana] [Sat Feb 21 18:53:48 2026] pci_device_remove+0x41/0xb0 [Sat Feb 21 18:53:48 2026] device_remove+0x46/0x70 [Sat Feb 21 18:53:48 2026] device_release_driver_internal+0x1e3/0x250 [Sat Feb 21 18:53:48 2026] device_release_driver+0x12/0x20 [Sat Feb 21 18:53:48 2026] pci_stop_bus_device+0x6a/0x90 [Sat Feb 21 18:53:48 2026] pci_stop_and_remove_bus_device+0x13/0x30 [Sat Feb 21 18:53:48 2026] mana_do_service+0x180/0x290 [mana] [Sat Feb 21 18:53:48 2026] mana_serv_func+0x24/0x50 [mana] [Sat Feb 21 18:53:48 2026] process_one_work+0x190/0x3d0 [Sat Feb 21 18:53:48 2026] worker_thread+0x16e/0x2e0 [Sat Feb 21 18:53:48 2026] kthread+0xf7/0x130 [Sat Feb 21 18:53:48 2026] ? __pfx_worker_thread+0x10/0x10 [Sat Feb 21 18:53:48 2026] ? __pfx_kthread+0x10/0x10 [Sat Feb 21 18:53:48 2026] ret_from_fork+0x269/0x350 [Sat Feb 21 18:53:48 2026] ? __pfx_kthread+0x10/0x10 [Sat Feb 21 18:53:48 2026] ret_from_fork_asm+0x1a/0x30 [Sat Feb 21 18:53:48 2026] Fixes: 505cc26bcae0 ("net: mana: Add support for auxiliary device servicing events") Reviewed-by: Haiyang Zhang Signed-off-by: Dipayaan Roy Reviewed-by: Simon Horman Link: https://patch.msgid.link/aZ2bzL64NagfyHpg@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

net: mana: Fix use-after-free in reset service rescan path

2025-12-28T09:34:00+00:00

When mana_serv_reset() encounters -ETIMEDOUT or -EPROTO from mana_gd_resume(), it performs a PCI rescan via mana_serv_rescan(). mana_serv_rescan() calls pci_stop_and_remove_bus_device(), which can invoke the driver's remove path and free the gdma_context associated with the device. After returning, mana_serv_reset() currently jumps to the out label and attempts to clear gc->in_service, dereferencing a freed gdma_context. The issue was observed with the following call logs: [ 698.942636] BUG: unable to handle page fault for address: ff6c2b638088508d [ 698.943121] #PF: supervisor write access in kernel mode [ 698.943423] #PF: error_code(0x0002) - not-present page [S[ 698.943793] Pat Dec 6 07:GD5 100000067 P4D 1002f7067 PUD 1002f8067 PMD 101bef067 PTE 0 0:56 2025] hv_[n e 698.944283] Oops: Oops: 0002 [#1] SMP NOPTI tvsc f8615163-00[ 698.944611] CPU: 28 UID: 0 PID: 249 Comm: kworker/28:1 ... [Sat Dec 6 07:50:56 2025] R10: [ 699.121594] mana 7870:00:00.0 enP30832s1: Configured vPort 0 PD 18 DB 16 000000000000001b R11: 0000000000000000 R12: ff44cf3f40270000 [Sat Dec 6 07:50:56 2025] R13: 0000000000000001 R14: ff44cf3f402700c8 R15: ff44cf3f4021b405 [Sat Dec 6 07:50:56 2025] FS: 0000000000000000(0000) GS:ff44cf7e9fcf9000(0000) knlGS:0000000000000000 [Sat Dec 6 07:50:56 2025] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [Sat Dec 6 07:50:56 2025] CR2: ff6c2b638088508d CR3: 000000011fe43001 CR4: 0000000000b73ef0 [Sat Dec 6 07:50:56 2025] Call Trace: [Sat Dec 6 07:50:56 2025] [Sat Dec 6 07:50:56 2025] mana_serv_func+0x24/0x50 [mana] [Sat Dec 6 07:50:56 2025] process_one_work+0x190/0x350 [Sat Dec 6 07:50:56 2025] worker_thread+0x2b7/0x3d0 [Sat Dec 6 07:50:56 2025] kthread+0xf3/0x200 [Sat Dec 6 07:50:56 2025] ? __pfx_worker_thread+0x10/0x10 [Sat Dec 6 07:50:56 2025] ? __pfx_kthread+0x10/0x10 [Sat Dec 6 07:50:56 2025] ret_from_fork+0x21a/0x250 [Sat Dec 6 07:50:56 2025] ? __pfx_kthread+0x10/0x10 [Sat Dec 6 07:50:56 2025] ret_from_fork_asm+0x1a/0x30 [Sat Dec 6 07:50:56 2025] Fix this by returning immediately after mana_serv_rescan() to avoid accessing GC state that may no longer be valid. Fixes: 9bf66036d686 ("net: mana: Handle hardware recovery events when probing the device") Reviewed-by: Simon Horman Reviewed-by: Long Li Signed-off-by: Dipayaan Roy Link: https://patch.msgid.link/20251218131054.GA3173@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Signed-off-by: Paolo Abeni

net: mana: Handle hardware recovery events when probing the device

2025-12-01T21:53:53+00:00

When MANA is being probed, it's possible that hardware is in recovery mode and the device may get GDMA_EQE_HWC_RESET_REQUEST over HWC in the middle of the probe. Detect such condition and go through the recovery service procedure. Signed-off-by: Long Li Reviewed-by: Haiyang Zhang Link: https://patch.msgid.link/1764193552-9712-1-git-send-email-longli@linux.microsoft.com Signed-off-by: Jakub Kicinski

net: mana: Drop TX skb on post_work_request failure and unmap resources

2025-11-20T04:11:57+00:00

Drop TX packets when posting the work request fails and ensure DMA mappings are always cleaned up. Signed-off-by: Aditya Garg Reviewed-by: Haiyang Zhang Link: https://patch.msgid.link/1763464269-10431-3-git-send-email-gargaditya@linux.microsoft.com Signed-off-by: Jakub Kicinski

net: mana: Handle SKB if TX SGEs exceed hardware limit

2025-11-20T04:11:57+00:00

The MANA hardware supports a maximum of 30 scatter-gather entries (SGEs) per TX WQE. Exceeding this limit can cause TX failures. Add ndo_features_check() callback to validate SKB layout before transmission. For GSO SKBs that would exceed the hardware SGE limit, clear NETIF_F_GSO_MASK to enforce software segmentation in the stack. Add a fallback in mana_start_xmit() to linearize non-GSO SKBs that still exceed the SGE limit. Also, Add ethtool counter for SKBs linearized Co-developed-by: Dipayaan Roy Signed-off-by: Dipayaan Roy Signed-off-by: Aditya Garg Reviewed-by: Eric Dumazet Reviewed-by: Haiyang Zhang Link: https://patch.msgid.link/1763464269-10431-2-git-send-email-gargaditya@linux.microsoft.com Signed-off-by: Jakub Kicinski

net: mana: Add standard counter rx_missed_errors

2025-11-18T03:52:30+00:00

Report standard counter stats->rx_missed_errors using hc_rx_discards_no_wqe from the hardware. Add a global workqueue to periodically run mana_query_gf_stats every 2 seconds to get the latest info in eth_stats and define a driver capability flag to notify hardware of the periodic queries. To avoid repeated failures and log flooding, the workqueue is not rescheduled if mana_query_gf_stats fails on HWC timeout error and the stats are reset to 0. Other errors are transient which will not need a VF reset for recovery. Signed-off-by: Erni Sri Satya Vennela Reviewed-by: Haiyang Zhang Link: https://patch.msgid.link/1763120599-6331-3-git-send-email-ernis@linux.microsoft.com Signed-off-by: Jakub Kicinski

net: mana: Move hardware counter stats from per-port to per-VF context

2025-11-18T03:52:30+00:00

Move hardware counter (HC) statistics from mana_port_context to mana_context to enable sharing stats across multiple network ports on the same MANA VF. Previously, each network port queried hardware counters independently using MANA_QUERY_GF_STAT command (GF = Generic Function stats from GDMA hardware), resulting in redundant queries when multiple ports existed on the same device. Isolate hardware counter stats by introducing mana_ethtool_hc_stats in mana_context and update the code to ensure all stats are properly reported via ethtool -S , maintaining consistency with previous behavior. Signed-off-by: Erni Sri Satya Vennela Reviewed-by: Haiyang Zhang Link: https://patch.msgid.link/1763120599-6331-2-git-send-email-ernis@linux.microsoft.com Signed-off-by: Jakub Kicinski