"value":"In the Linux kernel, the following vulnerability has been resolved:\n\nfirmware_loader: Block path traversal\n\nMost firmware names are hardcoded strings, or are constructed from fairly\nconstrained format strings where the dynamic parts are just some hex\nnumbers or such.\n\nHowever, there are a couple codepaths in the kernel where firmware file\nnames contain string components that are passed through from a device or\nsemi-privileged userspace; the ones I could find (not counting interfaces\nthat require root privileges) are:\n\n - lpfc_sli4_request_firmware_update() seems to construct the firmware\n filename from \"ModelName\", a string that was previously parsed out of\n some descriptor (\"Vital Product Data\") in lpfc_fill_vpd()\n - nfp_net_fw_find() seems to construct a firmware filename from a model\n name coming from nfp_hwinfo_lookup(pf->hwinfo, \"nffw.partno\"), which I\n think parses some descriptor that was read from the device.\n (But this case likely isn't exploitable because the format string looks\n like \"netronome/nic_%s\", and there shouldn't be any *folders* starting\n with \"netronome/nic_\". The previous case was different because there,\n the \"%s\" is *at the start* of the format string.)\n - module_flash_fw_schedule() is reachable from the\n ETHTOOL_MSG_MODULE_FW_FLASH_ACT netlink command, which is marked as\n GENL_UNS_ADMIN_PERM (meaning CAP_NET_ADMIN inside a user namespace is\n enough to pass the privilege check), and takes a userspace-provided\n firmware name.\n (But I think to reach this case, you need to have CAP_NET_ADMIN over a\n network namespace that a special kind of ethernet device is mapped into,\n so I think this is not a viable attack path in practice.)\n\nFix it by rejecting any firmware names containing \"..\" path components.\n\nFor what it's worth, I went looking and haven't found any USB device\ndrivers that use the firmware loader dangerously."
