system/systemd
system/systemd
1
0
Fork 0
mirror of https://github.com/systemd/systemd synced 2024-10-07 08:40:44 +00:00
Code Issues Packages Projects Releases Wiki Activity

bpf: fix restrict_fs on s390x

Browse source 
Linux kernel's bpf-next contains BPF LSM support for s390x. systemd's
test-bpf-lsm currently fails with this kernel.

This is an endianness issue: in the restrict_fs bpf program,
magic_number has type unsigned long (64 bits on s390x), but magic_map
keys are uint32_t (32 bits). Accessing magic_map using 64-bit keys may
work by accident on little-endian systems, but fails hard on big-endian
ones.

Fix by casting magic_number to uint32_t.
This commit is contained in:
Ilya Leoshkevich 2023-01-30 21:21:48 +01:00 committed by Lennart Poettering
parent 181eea677d
commit 907046282c
1 changed files with 7 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
src/core/bpf/restrict_fs/restrict-fs.bpf.c
View file

@ -39,16 +39,20 @@ struct {
SEC("lsm/file_open")
int BPF_PROG(restrict_filesystems, struct file *file, int ret)
{
unsigned long magic_number;
unsigned long raw_magic_number;
uint64_t cgroup_id;
uint32_t *value, *magic_map, zero = 0, *is_allow;
uint32_t *value, *magic_map, magic_number, zero = 0, *is_allow;
/* ret is the return value from the previous BPF program or 0 if it's
* the first hook */
if (ret != 0)
return ret;
BPF_CORE_READ_INTO(&magic_number, file, f_inode, i_sb, s_magic);
BPF_CORE_READ_INTO(&raw_magic_number, file, f_inode, i_sb, s_magic);
/* super_block.s_magic is unsigned long, but magic_map keys are
* uint32_t. Using s_magic as-is would fail on big-endian systems,
* which have 64-bit unsigned long. So cast it. */
magic_number = (uint32_t)raw_magic_number;
cgroup_id = bpf_get_current_cgroup_id();