CVE-2025-38168

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> perf: arm-ni: Unregister PMUs on probe failure<br /> <br /> When a resource allocation fails in one clock domain of an NI device,<br /> we need to properly roll back all previously registered perf PMUs in<br /> other clock domains of the same device.<br /> <br /> Otherwise, it can lead to kernel panics.<br /> <br /> Calling arm_ni_init+0x0/0xff8 [arm_ni] @ 2374<br /> arm-ni ARMHCB70:00: Failed to request PMU region 0x1f3c13000<br /> arm-ni ARMHCB70:00: probe with driver arm-ni failed with error -16<br /> list_add corruption: next-&gt;prev should be prev (fffffd01e9698a18),<br /> but was 0000000000000000. (next=ffff10001a0decc8).<br /> pstate: 6340009 (nZCv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)<br /> pc : list_add_valid_or_report+0x7c/0xb8<br /> lr : list_add_valid_or_report+0x7c/0xb8<br /> Call trace:<br /> __list_add_valid_or_report+0x7c/0xb8<br /> perf_pmu_register+0x22c/0x3a0<br /> arm_ni_probe+0x554/0x70c [arm_ni]<br /> platform_probe+0x70/0xe8<br /> really_probe+0xc6/0x4d8<br /> driver_probe_device+0x48/0x170<br /> __driver_attach+0x8e/0x1c0<br /> bus_for_each_dev+0x64/0xf0<br /> driver_add+0x138/0x260<br /> bus_add_driver+0x68/0x138<br /> __platform_driver_register+0x2c/0x40<br /> arm_ni_init+0x14/0x2a [arm_ni]<br /> do_init_module+0x36/0x298<br /> ---[ end trace 0000000000000000 ]---<br /> Kernel panic - not syncing: Oops - BUG: Fatal exception<br /> SMP: stopping secondary CPUs