{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2024-35968","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2024-05-17T13:50:33.140Z","datePublished":"2024-05-20T09:41:57.162Z","dateUpdated":"2026-08-05T11:31:17.459Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-05T11:31:17.459Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\npds_core: Fix pdsc_check_pci_health function to use work thread\n\nWhen the driver notices fw_status == 0xff it tries to perform a PCI\nreset on itself via pci_reset_function() in the context of the driver's\nhealth thread. However, pdsc_reset_prepare calls\npdsc_stop_health_thread(), which attempts to stop/flush the health\nthread. This results in a deadlock because the stop/flush will never\ncomplete since the driver called pci_reset_function() from the health\nthread context. Fix by changing the pdsc_check_pci_health_function()\nto queue a newly introduced pdsc_pci_reset_thread() on the pdsc's\nwork queue.\n\nUnloading the driver in the fw_down/dead state uncovered another issue,\nwhich can be seen in the following trace:\n\nWARNING: CPU: 51 PID: 6914 at kernel/workqueue.c:1450 __queue_work+0x358/0x440\n[...]\nRIP: 0010:__queue_work+0x358/0x440\n[...]\nCall Trace:\n <TASK>\n ? __warn+0x85/0x140\n ? __queue_work+0x358/0x440\n ? report_bug+0xfc/0x1e0\n ? handle_bug+0x3f/0x70\n ? exc_invalid_op+0x17/0x70\n ? asm_exc_invalid_op+0x1a/0x20\n ? __queue_work+0x358/0x440\n queue_work_on+0x28/0x30\n pdsc_devcmd_locked+0x96/0xe0 [pds_core]\n pdsc_devcmd_reset+0x71/0xb0 [pds_core]\n pdsc_teardown+0x51/0xe0 [pds_core]\n pdsc_remove+0x106/0x200 [pds_core]\n pci_device_remove+0x37/0xc0\n device_release_driver_internal+0xae/0x140\n driver_detach+0x48/0x90\n bus_remove_driver+0x6d/0xf0\n pci_unregister_driver+0x2e/0xa0\n pdsc_cleanup_module+0x10/0x780 [pds_core]\n __x64_sys_delete_module+0x142/0x2b0\n ? syscall_trace_enter.isra.18+0x126/0x1a0\n do_syscall_64+0x3b/0x90\n entry_SYSCALL_64_after_hwframe+0x72/0xdc\nRIP: 0033:0x7fbd9d03a14b\n[...]\n\nFix this by preventing the devcmd reset if the FW is not running."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","baseScore":7.8,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:L - The vulnerable code is a PCI device driver health/watchdog path driven by a kernel timer and MMIO register reads, with trigger surfaces limited to local interfaces (sysfs PCI reset/unbind, devlink, module unload, adminq/aux-bus clients such as a pds_vdpa guest). No remote or adjacent network data reaches this path.\nAC:L - Once the driver observes the bad-PCI status the deadlock is unconditional — `cancel_work_sync()` on the running work item is a deterministic self-deadlock, with no race to win and no memory-layout dependency; the unload-with-dead-FW path that reaches `queue_work()` on the freed workqueue is likewise deterministic because the devcmd always times out in that state.\nPR:L - The health thread executes autonomously from a 5-second kernel timer with no privilege involved, and low-privileged local actors that can drive the device — an aux-bus/adminq client or a guest using a passed-through PDS VF — can force the adminq timeouts that immediately queue the health work.\nUI:N - No victim action is needed; the watchdog timer queues the health work automatically and the deadlock or the post-`destroy_workqueue()` `queue_work()` occurs without any interaction from another user.\nS:U - The deadlock, lock retention and use-after-free are all confined to the host kernel's own security authority; no VM, IOMMU or sandbox boundary is crossed.\nC:H - `queue_work()` on the freed `workqueue_struct` after `destroy_workqueue()` is a use-after-free that reads freed heap contents and follows a stale `wq->cpu_pwq` pointer, which under heap grooming yields an attacker-influenced read primitive into kernel memory.\nI:H - The same use-after-free proceeds to lock and perform a list insertion through the stale `pwq->pool` pointer, writing into a reallocated slab object — a controllable write primitive suitable for control-flow corruption.\nA:H - The self-deadlock permanently hangs the driver's single-threaded workqueue while holding `config_lock` plus the pdev and upstream-bridge `device_lock`s, blocking further resets, unbind, hotplug and removal on that bridge and leaving unkillable D-state tasks that only a reboot clears; the UAF additionally causes kernel crashes."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/net/ethernet/amd/pds_core/core.c","drivers/net/ethernet/amd/pds_core/core.h","drivers/net/ethernet/amd/pds_core/dev.c","drivers/net/ethernet/amd/pds_core/main.c"],"versions":[{"version":"1e18ec3e9d46e4ad2b6507c3bfc7f59e2ab449a2","lessThan":"38407914d48273d7f8ab765b9243658afe1c3ab6","status":"affected","versionType":"git"},{"version":"1e18ec3e9d46e4ad2b6507c3bfc7f59e2ab449a2","lessThan":"81665adf25d28a00a986533f1d3a5df76b79cad9","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/net/ethernet/amd/pds_core/core.c","drivers/net/ethernet/amd/pds_core/core.h","drivers/net/ethernet/amd/pds_core/dev.c","drivers/net/ethernet/amd/pds_core/main.c"],"versions":[{"version":"6.7","status":"affected"},{"version":"0","lessThan":"6.7","status":"unaffected","versionType":"semver"},{"version":"6.8.7","lessThanOrEqual":"6.8.*","status":"unaffected","versionType":"semver"},{"version":"6.9","lessThanOrEqual":"*","status":"unaffected","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"operator":"OR","negate":false,"cpeMatch":[{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.7","versionEndExcluding":"6.8.7"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.7","versionEndExcluding":"6.9"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/38407914d48273d7f8ab765b9243658afe1c3ab6"},{"url":"https://git.kernel.org/stable/c/81665adf25d28a00a986533f1d3a5df76b79cad9"}],"title":"pds_core: Fix pdsc_check_pci_health function to use work thread","x_generator":{"engine":"bippy-1.2.0"}},"adp":[{"title":"CISA ADP Vulnrichment","metrics":[{"other":{"type":"ssvc","content":{"id":"CVE-2024-35968","role":"CISA Coordinator","options":[{"Exploitation":"none"},{"Automatable":"no"},{"Technical Impact":"partial"}],"version":"2.0.3","timestamp":"2024-05-21T15:11:50.638119Z"}}}],"providerMetadata":{"orgId":"134c704f-9b21-4f2e-91b3-4a467353bcc0","shortName":"CISA-ADP","dateUpdated":"2024-06-04T17:34:48.818Z"}},{"providerMetadata":{"orgId":"af854a3a-2127-422b-91ae-364da2661108","shortName":"CVE","dateUpdated":"2024-08-02T03:21:49.041Z"},"title":"CVE Program Container","references":[{"url":"https://git.kernel.org/stable/c/38407914d48273d7f8ab765b9243658afe1c3ab6","tags":["x_transferred"]},{"url":"https://git.kernel.org/stable/c/81665adf25d28a00a986533f1d3a5df76b79cad9","tags":["x_transferred"]}]}]}}