[PATCH v2] rtld: Enable MTE for stack when specified in .dynamic
Adhemerval Zanella Netto
adhemerval.zanella@linaro.org
Thu Sep 11 19:43:24 GMT 2025
On 01/09/25 18:10, Cupertino Miranda wrote:
> Hi everyone,
>
> This is v2 of MTE stack support patch series.
> This patch was adapted to conform with changes from:
>
> commit 3b2b88cceeb79f73a72367800d91599e2af4bb39
> elf: early conversion of elf p_flags to mprotect flags
>
> It still does not introduce any tunable options.
> They were initially discuss in v1 without any specific conclusions of
> what should be done in that regard.
>
> Looking forward to your reviews.
>
> Cheers,
> Cupertino
>
> Changes from v1:
> - dropped changes related to stack creation for threads. No longer
> needed.
> - Adaped code to changes introduced with:
> "elf: early conversion of elf p_flags to mprotect flags"
>
> MTE stack tagging enabled executables are flagged by setting the dynamic
> array tag DT_AARCH64_MEMTAG_STACK as described in section 6 from:
> https://github.com/ARM-software/abi-aa/blob/main/memtagabielf64/memtagabielf64.rst
>
> Also the mode in which MTE would react to MTE faults is set by the tag
> DT_AARCH64_MEMTAG_MODE.
>
> This patch verifies for the presence of those tags in the dynamic array
> and performs mprotect on the main application stack map with the extra
> PROT_MTE flag.
> ---
> elf/elf.h | 7 +-
> sysdeps/aarch64/Makefile | 3 +-
> sysdeps/aarch64/cpu-features.h | 14 ++++
> sysdeps/aarch64/dl-mte.c | 78 +++++++++++++++++++
> sysdeps/aarch64/dl-prop.h | 4 +
> sysdeps/unix/sysv/linux/aarch64/Makefile | 5 ++
> .../unix/sysv/linux/aarch64/dl-mte-stack.c | 42 ++++++++++
> 7 files changed, 151 insertions(+), 2 deletions(-)
> create mode 100644 sysdeps/aarch64/dl-mte.c
> create mode 100644 sysdeps/unix/sysv/linux/aarch64/dl-mte-stack.c
>
> diff --git a/elf/elf.h b/elf/elf.h
> index 2f29a47c0b..2c7b463769 100644
> --- a/elf/elf.h
> +++ b/elf/elf.h
> @@ -3039,7 +3039,12 @@ enum
> #define DT_AARCH64_BTI_PLT (DT_LOPROC + 1)
> #define DT_AARCH64_PAC_PLT (DT_LOPROC + 3)
> #define DT_AARCH64_VARIANT_PCS (DT_LOPROC + 5)
> -#define DT_AARCH64_NUM 6
> +#define DT_AARCH64_MEMTAG_MODE (DT_LOPROC + 9)
> +#define DT_AARCH64_MEMTAG_HEAP (DT_LOPROC + 11)
> +#define DT_AARCH64_MEMTAG_STACK (DT_LOPROC + 12)
> +#define DT_AARCH64_MEMTAG_GLOBALS (DT_LOPROC + 13)
> +#define DT_AARCH64_MEMTAG_GLOBALSSZ (DT_LOPROC + 15)
> +#define DT_AARCH64_NUM 16
>
> /* AArch64 specific values for the st_other field. */
> #define STO_AARCH64_VARIANT_PCS 0x80
> diff --git a/sysdeps/aarch64/Makefile b/sysdeps/aarch64/Makefile
> index bb97d31355..522813d6b9 100644
> --- a/sysdeps/aarch64/Makefile
> +++ b/sysdeps/aarch64/Makefile
> @@ -11,7 +11,8 @@ endif
> ifeq ($(subdir),elf)
> sysdep-dl-routines += \
> dl-bti \
> - dl-gcs
> + dl-gcs \
> + dl-mte
>
> tests += tst-audit26 \
> tst-audit27
> diff --git a/sysdeps/aarch64/cpu-features.h b/sysdeps/aarch64/cpu-features.h
> index ef4e947e8c..b60b804e57 100644
> --- a/sysdeps/aarch64/cpu-features.h
> +++ b/sysdeps/aarch64/cpu-features.h
> @@ -66,4 +66,18 @@ struct cpu_features
> bool mops;
> };
>
> +#define ARCH_MTE_MODE_SYNC (1 << 1)
> +#define ARCH_MTE_MODE_ASYNC (1 << 2)
> +#define AARCH64_CPU_FEATURE_MTE_STATE_STACK (1 << 3)
> +
> +#define AARCH64_CPU_FEATURE_MTE_STATE_MODE_MASK (0x3)
> +
> +#define AARCH64_CPU_FEATURE_MTE_STATE_MODE_MASK (0x3)
> +#define AARCH64_CPU_FEATURE_MTE_STATE_MODE_ASYNC (1)
> +#define AARCH64_CPU_FEATURE_MTE_STATE_MODE_SYNC (1 << 1)
> +
> +#ifndef ARCH_INIT_MEMORY_STACK_TAGGING
Why do you need to check if ARCH_INIT_MEMORY_STACK_TAGGING is already defined?
> +#define ARCH_INIT_MEMORY_STACK_TAGGING _dl_mte_stack_protect
> +#endif
> +
> #endif /* _CPU_FEATURES_AARCH64_H */
> diff --git a/sysdeps/aarch64/dl-mte.c b/sysdeps/aarch64/dl-mte.c
> new file mode 100644
> index 0000000000..b7d435bc53
> --- /dev/null
> +++ b/sysdeps/aarch64/dl-mte.c
> @@ -0,0 +1,78 @@
> +/* AArch64 MTE Stack functions.
> + Copyright (C) 2020-2025 Free Software Foundation, Inc.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <https://www.gnu.org/licenses/>. */
> +
> +#include <unistd.h>
> +#include <ldsodefs.h>
> +#include <assert.h>
> +#include <libintl.h>
> +
> +extern int
> +_dl_mte_mode (void);
I think you will need to add a attribute_hidden here otherwise might
either create a PLT or a GOT entry (in stack-pie).
> +
> +void
> +_dl_mte_stack_check (struct link_map *l, const char *program)
> +{
> + ElfW(Dyn) *d;
> + bool mte_enabled = false;
> + bool mte_mode_selected = false;
> + for (d = l->l_ld; d->d_tag != DT_NULL; ++d)
> + {
> + if (d->d_tag == DT_AARCH64_MEMTAG_STACK)
> + {
> + GLRO(dl_aarch64_cpu_features).mte_state |= AARCH64_CPU_FEATURE_MTE_STATE_STACK;
> + mte_enabled = true;
> + }
> + if (d->d_tag == DT_AARCH64_MEMTAG_MODE)
> + {
> + if (d->d_un.d_val == 1)
> + {
> + GLRO(dl_aarch64_cpu_features).mte_state &= ~AARCH64_CPU_FEATURE_MTE_STATE_MODE_MASK;
> + GLRO(dl_aarch64_cpu_features).mte_state |= AARCH64_CPU_FEATURE_MTE_STATE_MODE_ASYNC;
> + }
> + else
> + {
> + GLRO(dl_aarch64_cpu_features).mte_state &= ~AARCH64_CPU_FEATURE_MTE_STATE_MODE_MASK;
> + GLRO(dl_aarch64_cpu_features).mte_state |= AARCH64_CPU_FEATURE_MTE_STATE_MODE_SYNC;
> + }
Maybe always clear the mte_mode mask if DT_AARCH64_MEMTAG_MODE is present?
> + mte_mode_selected = true;
> + }
> + }
> +
> + if (mte_enabled && mte_mode_selected)
> + {
> + int errval = 0;
> + _dl_mte_mode ();
It is not clear to me why _dl_mte_mode returns an error mode if it not checked.
If the prctl can indeed fail if HWCAP2_MTE is enabled, I think we need to proper
handle it and abort the process statup.
Also, MTE is enabled also on init_cpu_features, called earlier in dynamic case at:
_dl_start
|_ _dl_start_final
|_ _dl_sysdep_start
|_ dl_platform_init
|_ init_cpu_features
Although only for USE_MTAG (--enable-memory-tagging) and iff the glibc.mem.tagging
if set.
To enable MTE in this case it requires the loader to first setup the main_map
link_map. So I think maybe it would be better to move the MTE enablement for both
case on _dl_mte_mode, there is no need to have the dynamic memory tagging support
as early as before link_map setup.
> + GL(dl_stack_prot_flags) |= PROT_MTE;
> +
> + uintptr_t page = ((uintptr_t) __libc_stack_end
> + & -(intptr_t) GLRO(dl_pagesize));
Maybe user ALIGN_DOWN or even PTR_ALIGN_DOWN to avoid the extra cast in the
__mprotect below.
> +
> + if (__mprotect ((void *) page, GLRO(dl_pagesize),
> + GL(dl_stack_prot_flags)) != 0)
> + errval = errno;
> +
> + if (errval)> + {
> + const char *errstring = NULL;
> + errstring = N_("cannot set stack with PROT_MTE");
> + const char *name = "MTE Stack";
> + struct dl_exception exception;
> + _dl_exception_create (&exception, name, errstring);
This will create a error message with 'error while loading shared libraries:' even
for the main_map; which is misleading. Maybe _dl_signal_error here.
> + _dl_signal_exception (errval, &exception, NULL);
> + }
> + }
> +}
> diff --git a/sysdeps/aarch64/dl-prop.h b/sysdeps/aarch64/dl-prop.h
> index abca2be7fa..66b05c1b75 100644
> --- a/sysdeps/aarch64/dl-prop.h
> +++ b/sysdeps/aarch64/dl-prop.h
> @@ -27,11 +27,15 @@ extern void _dl_bti_check (struct link_map *, const char *)
> extern void _dl_gcs_check (struct link_map *, const char *)
> attribute_hidden;
>
> +extern void _dl_mte_stack_check (struct link_map *, const char *)
> + attribute_hidden;
> +
> static inline void __attribute__ ((always_inline))
> _rtld_main_check (struct link_map *m, const char *program)
> {
> _dl_bti_check (m, program);
> _dl_gcs_check (m, program);
> + _dl_mte_stack_check (m, program);
> }
>
> static inline void __attribute__ ((always_inline))
> diff --git a/sysdeps/unix/sysv/linux/aarch64/Makefile b/sysdeps/unix/sysv/linux/aarch64/Makefile
> index 15a2b4471d..32bc043810 100644
> --- a/sysdeps/unix/sysv/linux/aarch64/Makefile
> +++ b/sysdeps/unix/sysv/linux/aarch64/Makefile
> @@ -145,6 +145,11 @@ endif # ifeq ($(have-test-cc-gcs),yes)
>
> endif # ifeq ($(subdir),misc)
>
> +ifeq ($(subdir),elf)
> +sysdep-dl-routines += \
> + dl-mte-stack
> +endif
> +
> ifeq ($(subdir),stdlib)
> gen-as-const-headers += ucontext_i.sym
> endif
> diff --git a/sysdeps/unix/sysv/linux/aarch64/dl-mte-stack.c b/sysdeps/unix/sysv/linux/aarch64/dl-mte-stack.c
> new file mode 100644
> index 0000000000..98fae85cfd
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/aarch64/dl-mte-stack.c
> @@ -0,0 +1,42 @@
> +/* Memory tagging handling for GNU dynamic linker. Stub version.
> + Copyright (C) 2003-2025 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <https://www.gnu.org/licenses/>. */
> +
> +#include <ldsodefs.h>
> +#include <libintl.h>
> +#include <not-cancel.h>
> +#include <sys/prctl.h>
> +
> +#define MTE_ALLOWED_TAGS (0xfffe << PR_MTE_TAG_SHIFT)
> +
> +int
> +_dl_mte_mode (void)
> +{
> + int err = 0;
The init_cpu_features also check for HWCAP2_MTE before trying the __prctl,
I think it should do the same.
> + if (GLRO(dl_aarch64_cpu_features).mte_state & AARCH64_CPU_FEATURE_MTE_STATE_MODE_SYNC)
> + err = __prctl (PR_SET_TAGGED_ADDR_CTRL,
> + (PR_TAGGED_ADDR_ENABLE | PR_MTE_TCF_SYNC | MTE_ALLOWED_TAGS),
> + 0, 0, 0);
> + else if (GLRO(dl_aarch64_cpu_features).mte_state & AARCH64_CPU_FEATURE_MTE_STATE_MODE_ASYNC)
> + err = __prctl (PR_SET_TAGGED_ADDR_CTRL,
> + (PR_TAGGED_ADDR_ENABLE | PR_MTE_TCF_ASYNC | MTE_ALLOWED_TAGS),
> + 0, 0, 0);
> +
> + return (err != 0
> + && (GLRO(dl_aarch64_cpu_features).mte_state
> + & AARCH64_CPU_FEATURE_MTE_STATE_MODE_MASK) != 0);
> +}
We need proper testcases, along with configure checks for:
* Check if static linker support the new tags and enable the tests accordingly.
* Adds test for AARCH64_CPU_FEATURE_MTE_STATE_STACK and check after process
execution whether the stack is correctly setup. I take we can check the procfs
maps or smaps for some special flags.
* Same tests for AARCH64_CPU_FEATURE_MTE_STATE_MODE_MASK and
AARCH64_CPU_FEATURE_MTE_STATE_MODE_SYNC.
More information about the Libc-alpha
mailing list