CVE-2024-26859

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> net/bnx2x: Prevent access to a freed page in page_pool<br /> <br /> Fix race condition leading to system crash during EEH error handling<br /> <br /> During EEH error recovery, the bnx2x driver&amp;#39;s transmit timeout logic<br /> could cause a race condition when handling reset tasks. The<br /> bnx2x_tx_timeout() schedules reset tasks via bnx2x_sp_rtnl_task(),<br /> which ultimately leads to bnx2x_nic_unload(). In bnx2x_nic_unload()<br /> SGEs are freed using bnx2x_free_rx_sge_range(). However, this could<br /> overlap with the EEH driver&amp;#39;s attempt to reset the device using<br /> bnx2x_io_slot_reset(), which also tries to free SGEs. This race<br /> condition can result in system crashes due to accessing freed memory<br /> locations in bnx2x_free_rx_sge()<br /> <br /> 799 static inline void bnx2x_free_rx_sge(struct bnx2x *bp,<br /> 800 struct bnx2x_fastpath *fp, u16 index)<br /> 801 {<br /> 802 struct sw_rx_page *sw_buf = &amp;fp-&gt;rx_page_ring[index];<br /> 803 struct page *page = sw_buf-&gt;page;<br /> ....<br /> where sw_buf was set to NULL after the call to dma_unmap_page()<br /> by the preceding thread.<br /> <br /> EEH: Beginning: &amp;#39;slot_reset&amp;#39;<br /> PCI 0011:01:00.0#10000: EEH: Invoking bnx2x-&gt;slot_reset()<br /> bnx2x: [bnx2x_io_slot_reset:14228(eth1)]IO slot reset initializing...<br /> bnx2x 0011:01:00.0: enabling device (0140 -&gt; 0142)<br /> bnx2x: [bnx2x_io_slot_reset:14244(eth1)]IO slot reset --&gt; driver unload<br /> Kernel attempted to read user page (0) - exploit attempt? (uid: 0)<br /> BUG: Kernel NULL pointer dereference on read at 0x00000000<br /> Faulting instruction address: 0xc0080000025065fc<br /> Oops: Kernel access of bad area, sig: 11 [#1]<br /> .....<br /> Call Trace:<br /> [c000000003c67a20] [c00800000250658c] bnx2x_io_slot_reset+0x204/0x610 [bnx2x] (unreliable)<br /> [c000000003c67af0] [c0000000000518a8] eeh_report_reset+0xb8/0xf0<br /> [c000000003c67b60] [c000000000052130] eeh_pe_report+0x180/0x550<br /> [c000000003c67c70] [c00000000005318c] eeh_handle_normal_event+0x84c/0xa60<br /> [c000000003c67d50] [c000000000053a84] eeh_event_handler+0xf4/0x170<br /> [c000000003c67da0] [c000000000194c58] kthread+0x1c8/0x1d0<br /> [c000000003c67e10] [c00000000000cf64] ret_from_kernel_thread+0x5c/0x64<br /> <br /> To solve this issue, we need to verify page pool allocations before<br /> freeing.