"value":"In the Linux kernel, the following vulnerability has been resolved:\n\nusb: gadget: f_ncm: fix potential NULL ptr deref in ncm_bitrate()\n\nIn Google internal bug 265639009 we've received an (as yet) unreproducible\ncrash report from an aarch64 GKI 5.10.149-android13 running device.\n\nAFAICT the source code is at:\n https://android.googlesource.com/kernel/common/+/refs/tags/ASB-2022-12-05_13-5.10\n\nThe call stack is:\n ncm_close() -> ncm_notify() -> ncm_do_notify()\nwith the crash at:\n ncm_do_notify+0x98/0x270\nCode: 79000d0b b9000a6c f940012a f9400269 (b9405d4b)\n\nWhich I believe disassembles to (I don't know ARM assembly, but it looks sane enough to me...):\n\n // halfword (16-bit) store presumably to event->wLength (at offset 6 of struct usb_cdc_notification)\n 0B 0D 00 79 strh w11, [x8, #6]\n\n // word (32-bit) store presumably to req->Length (at offset 8 of struct usb_request)\n 6C 0A 00 B9 str w12, [x19, #8]\n\n // x10 (NULL) was read here from offset 0 of valid pointer x9\n // IMHO we're reading 'cdev->gadget' and getting NULL\n // gadget is indeed at offset 0 of struct usb_composite_dev\n 2A 01 40 F9 ldr x10, [x9]\n\n // loading req->buf pointer, which is at offset 0 of struct usb_request\n 69 02 40 F9 ldr x9, [x19]\n\n // x10 is null, crash, appears to be attempt to read cdev->gadget->max_speed\n 4B 5D 40 B9 ldr w11, [x10, #0x5c]\n\nwhich seems to line up with ncm_do_notify() case NCM_NOTIFY_SPEED code fragment:\n\n event->wLength = cpu_to_le16(8);\n req->length = NCM_STATUS_BYTECOUNT;\n\n /* SPEED_CHANGE data is up/down speeds in bits/sec */\n data = req->buf + sizeof *event;\n data[0] = cpu_to_le32(ncm_bitrate(cdev->gadget));\n\nMy analysis of registers and NULL ptr deref crash offset\n (Unable to handle kernel NULL pointer dereference at virtual address 000000000000005c)\nheavily suggests that the crash is due to 'cdev->gadget' being NULL when executing:\n data[0] = cpu_to_le32(ncm_bitrate(cdev->gadget));\nwhich calls:\n ncm_bitrate(NULL)\nwhich then calls:\n gadget_is_superspeed(NULL)\nwhich reads\n ((struct usb_gadget *)NULL)->max_speed\nand hits a panic.\n\nAFAICT, if I'm counting right, the offset of max_speed is indeed 0x5C.\n(remember there's a GKI KABI reservation of 16 bytes in struct work_struct)\n\nIt's not at all clear to me how this is all supposed to work...\nbut returning 0 seems much better than panic-ing..."
