CVE-2022-50168

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> bpf, x86: fix freeing of not-finalized bpf_prog_pack<br /> <br /> syzbot reported a few issues with bpf_prog_pack [1], [2]. This only happens<br /> with multiple subprogs. In jit_subprogs(), we first call bpf_int_jit_compile()<br /> on each sub program. And then, we call it on each sub program again. jit_data<br /> is not freed in the first call of bpf_int_jit_compile(). Similarly we don&amp;#39;t<br /> call bpf_jit_binary_pack_finalize() in the first call of bpf_int_jit_compile().<br /> <br /> If bpf_int_jit_compile() failed for one sub program, we will call<br /> bpf_jit_binary_pack_finalize() for this sub program. However, we don&amp;#39;t have a<br /> chance to call it for other sub programs. Then we will hit "goto out_free" in<br /> jit_subprogs(), and call bpf_jit_free on some subprograms that haven&amp;#39;t got<br /> bpf_jit_binary_pack_finalize() yet.<br /> <br /> At this point, bpf_jit_binary_pack_free() is called and the whole 2MB page is<br /> freed erroneously.<br /> <br /> Fix this with a custom bpf_jit_free() for x86_64, which calls<br /> bpf_jit_binary_pack_finalize() if necessary. Also, with custom<br /> bpf_jit_free(), bpf_prog_aux-&gt;use_bpf_prog_pack is not needed any more,<br /> remove it.<br /> <br /> [1] https://syzkaller.appspot.com/bug?extid=2f649ec6d2eea1495a8f<br /> [2] https://syzkaller.appspot.com/bug?extid=87f65c75f4a72db05445