CVE-2022-49837

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> bpf: Fix memory leaks in __check_func_call<br /> <br /> kmemleak reports this issue:<br /> <br /> unreferenced object 0xffff88817139d000 (size 2048):<br /> comm "test_progs", pid 33246, jiffies 4307381979 (age 45851.820s)<br /> hex dump (first 32 bytes):<br /> 01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................<br /> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................<br /> backtrace:<br /> [] kmalloc_trace+0x27/0xa0<br /> [] __check_func_call+0x316/0x1230<br /> [] check_helper_call+0x172e/0x4700<br /> [] do_check+0x21d8/0x45e0<br /> [] do_check_common+0x767/0xaf0<br /> [] bpf_check+0x43e3/0x5bc0<br /> [] bpf_prog_load+0xf26/0x1940<br /> [] __sys_bpf+0xd2c/0x3650<br /> [] __x64_sys_bpf+0x75/0xc0<br /> [] do_syscall_64+0x3b/0x90<br /> [] entry_SYSCALL_64_after_hwframe+0x63/0xcd<br /> <br /> The root case here is: In function prepare_func_exit(), the callee is<br /> not released in the abnormal scenario after "state-&gt;curframe--;". To<br /> fix, move "state-&gt;curframe--;" to the very bottom of the function,<br /> right when we free callee and reset frame[] pointer to NULL, as Andrii<br /> suggested.<br /> <br /> In addition, function __check_func_call() has a similar problem. In<br /> the abnormal scenario before "state-&gt;curframe++;", the callee also<br /> should be released by free_func_state().