]> git.baikalelectronics.ru Git - kernel.git/commitdiff
bpf: Add oversize check before call kvcalloc()
authorBixuan Cui <cuibixuan@huawei.com>
Sat, 11 Sep 2021 00:55:57 +0000 (08:55 +0800)
committerAlexei Starovoitov <ast@kernel.org>
Mon, 13 Sep 2021 23:28:15 +0000 (16:28 -0700)
Commit 7376f3330a82 ("mm: don't allow oversized kvmalloc() calls") add the
oversize check. When the allocation is larger than what kmalloc() supports,
the following warning triggered:

WARNING: CPU: 0 PID: 8408 at mm/util.c:597 kvmalloc_node+0x108/0x110 mm/util.c:597
Modules linked in:
CPU: 0 PID: 8408 Comm: syz-executor221 Not tainted 5.14.0-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:kvmalloc_node+0x108/0x110 mm/util.c:597
Call Trace:
 kvmalloc include/linux/mm.h:806 [inline]
 kvmalloc_array include/linux/mm.h:824 [inline]
 kvcalloc include/linux/mm.h:829 [inline]
 check_btf_line kernel/bpf/verifier.c:9925 [inline]
 check_btf_info kernel/bpf/verifier.c:10049 [inline]
 bpf_check+0xd634/0x150d0 kernel/bpf/verifier.c:13759
 bpf_prog_load kernel/bpf/syscall.c:2301 [inline]
 __sys_bpf+0x11181/0x126e0 kernel/bpf/syscall.c:4587
 __do_sys_bpf kernel/bpf/syscall.c:4691 [inline]
 __se_sys_bpf kernel/bpf/syscall.c:4689 [inline]
 __x64_sys_bpf+0x78/0x90 kernel/bpf/syscall.c:4689
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae

Reported-by: syzbot+f3e749d4c662818ae439@syzkaller.appspotmail.com
Signed-off-by: Bixuan Cui <cuibixuan@huawei.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20210911005557.45518-1-cuibixuan@huawei.com
kernel/bpf/verifier.c

index 047ac4b4703b08a72902f423a0a629803f2358cc..e76b559179054b93191b8959ddcad9ad01535e8c 100644 (file)
@@ -9912,6 +9912,8 @@ static int check_btf_line(struct bpf_verifier_env *env,
        nr_linfo = attr->line_info_cnt;
        if (!nr_linfo)
                return 0;
+       if (nr_linfo > INT_MAX / sizeof(struct bpf_line_info))
+               return -EINVAL;
 
        rec_size = attr->line_info_rec_size;
        if (rec_size < MIN_BPF_LINEINFO_SIZE ||