CVE-2021-46931

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> net/mlx5e: Wrap the tx reporter dump callback to extract the sq<br /> <br /> Function mlx5e_tx_reporter_dump_sq() casts its void * argument to struct<br /> mlx5e_txqsq *, but in TX-timeout-recovery flow the argument is actually<br /> of type struct mlx5e_tx_timeout_ctx *.<br /> <br /> mlx5_core 0000:08:00.1 enp8s0f1: TX timeout detected<br /> mlx5_core 0000:08:00.1 enp8s0f1: TX timeout on queue: 1, SQ: 0x11ec, CQ: 0x146d, SQ Cons: 0x0 SQ Prod: 0x1, usecs since last trans: 21565000<br /> BUG: stack guard page was hit at 0000000093f1a2de (stack is 00000000b66ea0dc..000000004d932dae)<br /> kernel stack overflow (page fault): 0000 [#1] SMP NOPTI<br /> CPU: 5 PID: 95 Comm: kworker/u20:1 Tainted: G W OE 5.13.0_mlnx #1<br /> Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014<br /> Workqueue: mlx5e mlx5e_tx_timeout_work [mlx5_core]<br /> RIP: 0010:mlx5e_tx_reporter_dump_sq+0xd3/0x180<br /> [mlx5_core]<br /> Call Trace:<br /> mlx5e_tx_reporter_dump+0x43/0x1c0 [mlx5_core]<br /> devlink_health_do_dump.part.91+0x71/0xd0<br /> devlink_health_report+0x157/0x1b0<br /> mlx5e_reporter_tx_timeout+0xb9/0xf0 [mlx5_core]<br /> ? mlx5e_tx_reporter_err_cqe_recover+0x1d0/0x1d0<br /> [mlx5_core]<br /> ? mlx5e_health_queue_dump+0xd0/0xd0 [mlx5_core]<br /> ? update_load_avg+0x19b/0x550<br /> ? set_next_entity+0x72/0x80<br /> ? pick_next_task_fair+0x227/0x340<br /> ? finish_task_switch+0xa2/0x280<br /> mlx5e_tx_timeout_work+0x83/0xb0 [mlx5_core]<br /> process_one_work+0x1de/0x3a0<br /> worker_thread+0x2d/0x3c0<br /> ? process_one_work+0x3a0/0x3a0<br /> kthread+0x115/0x130<br /> ? kthread_park+0x90/0x90<br /> ret_from_fork+0x1f/0x30<br /> --[ end trace 51ccabea504edaff ]---<br /> RIP: 0010:mlx5e_tx_reporter_dump_sq+0xd3/0x180<br /> PKRU: 55555554<br /> Kernel panic - not syncing: Fatal exception<br /> Kernel Offset: disabled<br /> end Kernel panic - not syncing: Fatal exception<br /> <br /> To fix this bug add a wrapper for mlx5e_tx_reporter_dump_sq() which<br /> extracts the sq from struct mlx5e_tx_timeout_ctx and set it as the<br /> TX-timeout-recovery flow dump callback.