Re: [PATCH v27 23/31] x86/cet/shstk: Add user-mode shadow stack support

["Edgecombe, Rick P" <rick.p.edgecombe@intel.com>]

On Fri, 2021-05-21 at 15:12 -0700, Yu-cheng Yu wrote:
> Introduce basic shadow stack enabling/disabling/allocation routines.
> A task's shadow stack is allocated from memory with VM_SHADOW_STACK
> flag
> and has a fixed size of min(RLIMIT_STACK, 4GB).
> 
> Signed-off-by: Yu-cheng Yu <yu-cheng.yu@intel.com>
> Cc: Kees Cook <keescook@chromium.org>
> ---
> v27:
> - Change 'struct cet_status' to 'struct thread_shstk', and change
> member
>   types from unsigned long to u64.
> - Re-order local variables in reverse order of length.
> - WARN_ON_ONCE() when vm_munmap() fails.
> 
>  arch/x86/include/asm/cet.h       |  30 +++++++
>  arch/x86/include/asm/processor.h |   5 ++
>  arch/x86/kernel/Makefile         |   1 +
>  arch/x86/kernel/shstk.c          | 130
> +++++++++++++++++++++++++++++++
>  4 files changed, 166 insertions(+)
>  create mode 100644 arch/x86/include/asm/cet.h
>  create mode 100644 arch/x86/kernel/shstk.c
> 
> diff --git a/arch/x86/include/asm/cet.h b/arch/x86/include/asm/cet.h
> new file mode 100644
> index 000000000000..6432baf4de1f
> --- /dev/null
> +++ b/arch/x86/include/asm/cet.h
> @@ -0,0 +1,30 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +#ifndef _ASM_X86_CET_H
> +#define _ASM_X86_CET_H
> +
> +#ifndef __ASSEMBLY__
> +#include <linux/types.h>
> +
> +struct task_struct;
> +
> +/*
> + * Per-thread CET status
> + */
> +struct thread_shstk {
> +       u64     base;
> +       u64     size;
> +};
> +
> +#ifdef CONFIG_X86_SHADOW_STACK
> +int shstk_setup(void);
> +void shstk_free(struct task_struct *p);
> +void shstk_disable(void);
> +#else
> +static inline int shstk_setup(void) { return 0; }
> +static inline void shstk_free(struct task_struct *p) {}
> +static inline void shstk_disable(void) {}
> +#endif
> +
> +#endif /* __ASSEMBLY__ */
> +
> +#endif /* _ASM_X86_CET_H */
> diff --git a/arch/x86/include/asm/processor.h
> b/arch/x86/include/asm/processor.h
> index 556b2b17c3e2..7eb56a837cfa 100644
> --- a/arch/x86/include/asm/processor.h
> +++ b/arch/x86/include/asm/processor.h
> @@ -27,6 +27,7 @@ struct vm86;
>  #include <asm/unwind_hints.h>
>  #include <asm/vmxfeatures.h>
>  #include <asm/vdso/processor.h>
> +#include <asm/cet.h>
>  
>  #include <linux/personality.h>
>  #include <linux/cache.h>
> @@ -518,6 +519,10 @@ struct thread_struct {
>  
>         unsigned int            sig_on_uaccess_err:1;
>  
> +#ifdef CONFIG_X86_SHADOW_STACK
> +       struct thread_shstk     shstk;
> +#endif
> +
>         /* Floating point and extended processor state */
>         struct fpu              fpu;
>         /*
> diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
> index 0f66682ac02a..b0e599102277 100644
> --- a/arch/x86/kernel/Makefile
> +++ b/arch/x86/kernel/Makefile
> @@ -149,6 +149,7 @@ obj-$(CONFIG_UNWINDER_FRAME_POINTER)        +=
> unwind_frame.o
>  obj-$(CONFIG_UNWINDER_GUESS)           += unwind_guess.o
>  
>  obj-$(CONFIG_AMD_MEM_ENCRYPT)          += sev.o
> +obj-$(CONFIG_X86_SHADOW_STACK)         += shstk.o
>  ###
>  # 64 bit specific files
>  ifeq ($(CONFIG_X86_64),y)
> diff --git a/arch/x86/kernel/shstk.c b/arch/x86/kernel/shstk.c
> new file mode 100644
> index 000000000000..5ea2b494e9f9
> --- /dev/null
> +++ b/arch/x86/kernel/shstk.c
> @@ -0,0 +1,130 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * shstk.c - Intel shadow stack support
> + *
> + * Copyright (c) 2021, Intel Corporation.
> + * Yu-cheng Yu <yu-cheng.yu@intel.com>
> + */
> +
> +#include <linux/types.h>
> +#include <linux/mm.h>
> +#include <linux/mman.h>
> +#include <linux/slab.h>
> +#include <linux/uaccess.h>
> +#include <linux/sched/signal.h>
> +#include <linux/compat.h>
> +#include <linux/sizes.h>
> +#include <linux/user.h>
> +#include <asm/msr.h>
> +#include <asm/fpu/internal.h>
> +#include <asm/fpu/xstate.h>
> +#include <asm/fpu/types.h>
> +#include <asm/cet.h>
> +
> +static void start_update_msrs(void)
> +{
> +       fpregs_lock();
> +       if (test_thread_flag(TIF_NEED_FPU_LOAD))
> +               __fpregs_load_activate();
> +}
> +
> +static void end_update_msrs(void)
> +{
> +       fpregs_unlock();
> +}
> +
> +static unsigned long alloc_shstk(unsigned long size)
> +{
> +       int flags = MAP_ANONYMOUS | MAP_PRIVATE;
> +       struct mm_struct *mm = current->mm;
> +       unsigned long addr, populate;
> +
> +       mmap_write_lock(mm);
> +       addr = do_mmap(NULL, 0, size, PROT_READ, flags,
> VM_SHADOW_STACK, 0,
> +                      &populate, NULL);
> +       mmap_write_unlock(mm);
> +
> +       return addr;
> +}
> +
> +int shstk_setup(void)
> +{
> +       struct thread_shstk *shstk = &current->thread.shstk;
> +       unsigned long addr, size;
> +
> +       if (!cpu_feature_enabled(X86_FEATURE_SHSTK))
> +               return -EOPNOTSUPP;
The only caller of this will skip it if
!cpu_feature_enabled(X86_FEATURE_SHSTK), so this is dead logic. Same
pattern in the IBT patch.

> +
> +       size = round_up(min_t(unsigned long long,
> rlimit(RLIMIT_STACK), SZ_4G), PAGE_SIZE);
> +       addr = alloc_shstk(size);
> +       if (IS_ERR_VALUE(addr))
> +               return PTR_ERR((void *)addr);
> +
> +       shstk->base = addr;
> +       shstk->size = size;
> +
> +       start_update_msrs();
> +       wrmsrl(MSR_IA32_PL3_SSP, addr + size);
> +       wrmsrl(MSR_IA32_U_CET, CET_SHSTK_EN);
> +       end_update_msrs();
> +
> +       return 0;
> +}
> +
> +void shstk_free(struct task_struct *tsk)
> +{
> +       struct thread_shstk *shstk = &tsk->thread.shstk;
> +
> +       if (!cpu_feature_enabled(X86_FEATURE_SHSTK) ||
> +           !shstk->size ||
> +           !shstk->base)
> +               return;
> +
> +       if (!tsk->mm)
> +               return;
> +
> +       while (1) {
> +               int r;
> +
> +               r = vm_munmap(shstk->base, shstk->size);
> +
> +               /*
> +                * vm_munmap() returns -EINTR when mmap_lock is held
> by
> +                * something else, and that lock should not be held
> for a
> +                * long time.  Retry it for the case.
> +                */
> +               if (r == -EINTR) {
> +                       cond_resched();
> +                       continue;
> +               }
> +
> +               /*
> +                * For all other types of vm_munmap() failure, either
> the
> +                * system is out of memory or there is bug.
> +                */
> +               WARN_ON_ONCE(r);
> +               break;
> +       }
> +
> +       shstk->base = 0;
> +       shstk->size = 0;
> +}
> +
> +void shstk_disable(void)
> +{
> +       struct thread_shstk *shstk = &current->thread.shstk;
> +       u64 msr_val;
> +
> +       if (!cpu_feature_enabled(X86_FEATURE_SHSTK) ||
> +           !shstk->size ||
> +           !shstk->base)
> +               return;
> +
> +       start_update_msrs();
> +       rdmsrl(MSR_IA32_U_CET, msr_val);
> +       wrmsrl(MSR_IA32_U_CET, msr_val & ~CET_SHSTK_EN);
> +       wrmsrl(MSR_IA32_PL3_SSP, 0);
> +       end_update_msrs();
> +
> +       shstk_free(current);
> +}