[LTP] [PATCH v3 2/2] Add mseal01 test
Li Wang
liwang@redhat.com
Sat Aug 3 05:27:26 CEST 2024
Hi Andrea,
On Thu, Aug 1, 2024 at 8:44 PM Andrea Cervesato <andrea.cervesato@suse.de>
wrote:
> From: Andrea Cervesato <andrea.cervesato@suse.com>
>
> This is a smoke test that verifies if mseal() protects specific VMA
> portions of a process. According to documentation, the syscall should
> protect memory from the following actions:
>
> - unmapping, moving to another location, and shrinking the size, via
> munmap() and mremap()
> - moving or expanding a different VMA into the current location, via
> mremap()
> - modifying a VMA via mmap(MAP_FIXED)
> - mprotect() and pkey_mprotect()
> - destructive madvice() behaviors (e.g. MADV_DONTNEED) for anonymous
> memory, when users don’t have write permission to the memory
>
> Any of the described actions is recognized via EPERM errno.
>
> Signed-off-by: Andrea Cervesato <andrea.cervesato@suse.com>
> ---
> testcases/kernel/syscalls/mseal/.gitignore | 1 +
> testcases/kernel/syscalls/mseal/Makefile | 7 ++
> testcases/kernel/syscalls/mseal/mseal01.c | 156
> +++++++++++++++++++++++++++++
> 3 files changed, 164 insertions(+)
>
> diff --git a/testcases/kernel/syscalls/mseal/.gitignore
> b/testcases/kernel/syscalls/mseal/.gitignore
> new file mode 100644
> index 000000000..e13090994
> --- /dev/null
> +++ b/testcases/kernel/syscalls/mseal/.gitignore
> @@ -0,0 +1 @@
> +mseal01
> diff --git a/testcases/kernel/syscalls/mseal/Makefile
> b/testcases/kernel/syscalls/mseal/Makefile
> new file mode 100644
> index 000000000..35317f446
> --- /dev/null
> +++ b/testcases/kernel/syscalls/mseal/Makefile
> @@ -0,0 +1,7 @@
> +# SPDX-License-Identifier: GPL-2.0-or-later
> +# Copyright (C) 2023 SUSE LLC Andrea Cervesato <andrea.cervesato@suse.com
> >
> +
> +top_srcdir ?= ../../../..
> +
> +include $(top_srcdir)/include/mk/testcases.mk
> +include $(top_srcdir)/include/mk/generic_leaf_target.mk
> diff --git a/testcases/kernel/syscalls/mseal/mseal01.c
> b/testcases/kernel/syscalls/mseal/mseal01.c
> new file mode 100644
> index 000000000..eb2f4d588
> --- /dev/null
> +++ b/testcases/kernel/syscalls/mseal/mseal01.c
> @@ -0,0 +1,156 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (C) 2024 SUSE LLC Andrea Cervesato <
> andrea.cervesato@suse.com>
> + */
> +
> +/*\
> + * [Description]
> + *
> + * This is a smoke test that verifies if mseal() protects specific VMA
> portions
> + * of a process. According to documentation, the syscall should protect
> memory
> + * from the following actions:
> + *
> + * - unmapping, moving to another location, and shrinking the size, via
> munmap()
> + * and mremap()
> + * - moving or expanding a different VMA into the current location, via
> mremap()
> + * - modifying a VMA via mmap(MAP_FIXED)
> + * - mprotect() and pkey_mprotect()
> + * - destructive madvice() behaviors (e.g. MADV_DONTNEED) for anonymous
> memory,
> + * when users don’t have write permission to the memory
> + *
> + * Any of the described actions is recognized via EPERM errno.
> + */
> +
> +#define _GNU_SOURCE
> +
> +#include "tst_test.h"
> +#include "lapi/syscalls.h"
> +
> +#define MEMPAGES 8
> +#define MEMSEAL 2
> +
> +static void *mem_addr;
> +static int mem_size;
> +static int mem_offset;
> +static int mem_alignment;
> +
> +static inline int sys_mseal(void *start, size_t len)
> +{
> + return tst_syscall(__NR_mseal, start, len, 0);
> +}
> +
> +static void test_mprotect(void)
> +{
> + TST_EXP_FAIL(mprotect(mem_addr, mem_size, PROT_NONE), EPERM);
> +}
> +
> +static void test_pkey_mprotect(void)
> +{
> + int ret;
> + int pkey;
> +
> + pkey = tst_syscall(__NR_pkey_alloc, 0, 0);
> + if (pkey == -1) {
> + if (errno == EINVAL)
> + tst_brk(TCONF, "pku is not supported on this CPU");
>
Based on my limited knowledge of PKEY things, here might get three errnos:
ENOSYS, EINVAL and ENOSPC, so probably we can just do that by the exist
pkey support checker and uses the ltp_pkey_* series capsulated wrappers?
https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/syscalls/pkeys/pkey.h
> +
> + tst_brk(TBROK | TERRNO, "pkey_alloc() error");
> + }
> +
> + TST_EXP_FAIL(tst_syscall(__NR_pkey_mprotect,
> + mem_addr, mem_size,
> + PROT_NONE,
> + pkey),
> + EPERM);
> +
> + ret = tst_syscall(__NR_pkey_free, pkey);
> + if (ret == -1)
> + tst_brk(TBROK | TERRNO, "pkey_free() error");
> +}
> +
> +static void test_madvise(void)
> +{
> + TST_EXP_FAIL(madvise(mem_addr, mem_size, MADV_DONTNEED), EPERM);
> +}
> +
> +static void test_munmap(void)
> +{
> + TST_EXP_FAIL(munmap(mem_addr, mem_size), EPERM);
> +}
> +
> +static void test_mremap_resize(void)
> +{
> + void *new_addr;
> + size_t new_size = 2 * mem_alignment;
> +
> + new_addr = SAFE_MMAP(NULL, mem_size,
> + PROT_READ,
> + MAP_ANONYMOUS | MAP_PRIVATE,
> + -1, 0);
> +
> + TST_EXP_FAIL_PTR_VOID(mremap(mem_addr, mem_size, new_size,
> + MREMAP_MAYMOVE | MREMAP_FIXED,
> + new_addr),
> + EPERM);
> +
> + SAFE_MUNMAP(new_addr, new_size);
> +}
> +
> +static void test_mmap_change_prot(void)
> +{
> + TST_EXP_FAIL_PTR_VOID(mmap(mem_addr, mem_size,
> + PROT_READ,
> + MAP_ANONYMOUS | MAP_PRIVATE | MAP_FIXED,
> + -1, 0), EPERM);
> +}
> +
> +static struct tcase {
> + void (*func_test)(void);
> + int prot;
> + char *message;
> +} tcases[] = {
> + {test_mprotect, PROT_READ | PROT_WRITE, "mprotect() availability"},
> + {test_pkey_mprotect, PROT_READ | PROT_WRITE, "pkey_mprotect()
> availability"},
> + {test_madvise, PROT_READ, "madvise() availability"},
> + {test_munmap, PROT_READ | PROT_WRITE, "munmap() availability from
> child"},
> + {test_mremap_resize, PROT_READ | PROT_WRITE, "mremap() address
> move/resize"},
> + {test_mmap_change_prot, PROT_READ | PROT_WRITE, "mmap() protection
> change"},
> +};
> +
> +static void run(unsigned int n)
> +{
> + /* the reason why we spawn a child is that mseal() will
> + * protect VMA until process will call _exit()
> + */
> + if (!SAFE_FORK()) {
> + struct tcase *tc = &tcases[n];
> +
> + mem_addr = SAFE_MMAP(NULL, mem_size,
> + tc->prot,
> + MAP_ANONYMOUS | MAP_PRIVATE,
> + -1, 0);
> +
> + tst_res(TINFO, "Testing %s", tc->message);
> +
> + TST_EXP_PASS(sys_mseal(mem_addr + mem_offset,
> mem_alignment));
> +
> + tc->func_test();
> + _exit(0);
> + }
> +}
> +
> +static void setup(void)
> +{
> + mem_alignment = getpagesize();
> + mem_size = mem_alignment * MEMPAGES;
> + mem_offset = mem_alignment * MEMSEAL;
> +}
> +
> +static struct tst_test test = {
> + .test = run,
> + .tcnt = ARRAY_SIZE(tcases),
> + .setup = setup,
>
> + .min_kver = "6.10",
>
The minimal kernel version check is unnecessary.
> + .forks_child = 1,
> +};
> +
>
> --
> 2.43.0
>
>
> --
> Mailing list info: https://lists.linux.it/listinfo/ltp
>
--
Regards,
Li Wang
More information about the ltp
mailing list