Kconfig (83c2da2e605c73aafcc02df04b2dbf1ccbfc24c0) Kconfig (0d8315dddd2899f519fe1ca3d4d5cdaf44ea421e)
1# SPDX-License-Identifier: GPL-2.0
2#
3# General architecture dependent options
4#
5
6#
7# Note: arch/$(SRCARCH)/Kconfig needs to be included first so that it can
8# override the default values in this file.

--- 472 unchanged lines hidden (view full) ---

481 - syscall_get_arguments()
482 - syscall_rollback()
483 - syscall_set_return_value()
484 - SIGSYS siginfo_t support
485 - secure_computing is called from a ptrace_event()-safe context
486 - secure_computing return value is checked and a return value of -1
487 results in the system call being skipped immediately.
488 - seccomp syscall wired up
1# SPDX-License-Identifier: GPL-2.0
2#
3# General architecture dependent options
4#
5
6#
7# Note: arch/$(SRCARCH)/Kconfig needs to be included first so that it can
8# override the default values in this file.

--- 472 unchanged lines hidden (view full) ---

481 - syscall_get_arguments()
482 - syscall_rollback()
483 - syscall_set_return_value()
484 - SIGSYS siginfo_t support
485 - secure_computing is called from a ptrace_event()-safe context
486 - secure_computing return value is checked and a return value of -1
487 results in the system call being skipped immediately.
488 - seccomp syscall wired up
489 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE,
490 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If
491 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too.
489
490config SECCOMP
491 prompt "Enable seccomp to safely execute untrusted bytecode"
492 def_bool y
493 depends on HAVE_ARCH_SECCOMP
494 help
495 This kernel feature is useful for number crunching applications
496 that may need to handle untrusted bytecode during their

--- 12 unchanged lines hidden (view full) ---

509 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
510 help
511 Enable tasks to build secure computing environments defined
512 in terms of Berkeley Packet Filter programs which implement
513 task-defined system call filtering polices.
514
515 See Documentation/userspace-api/seccomp_filter.rst for details.
516
492
493config SECCOMP
494 prompt "Enable seccomp to safely execute untrusted bytecode"
495 def_bool y
496 depends on HAVE_ARCH_SECCOMP
497 help
498 This kernel feature is useful for number crunching applications
499 that may need to handle untrusted bytecode during their

--- 12 unchanged lines hidden (view full) ---

512 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
513 help
514 Enable tasks to build secure computing environments defined
515 in terms of Berkeley Packet Filter programs which implement
516 task-defined system call filtering polices.
517
518 See Documentation/userspace-api/seccomp_filter.rst for details.
519
520config SECCOMP_CACHE_DEBUG
521 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache"
522 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR
523 depends on PROC_FS
524 help
525 This enables the /proc/pid/seccomp_cache interface to monitor
526 seccomp cache data. The file format is subject to change. Reading
527 the file requires CAP_SYS_ADMIN.
528
529 This option is for debugging only. Enabling presents the risk that
530 an adversary may be able to infer the seccomp filter logic.
531
532 If unsure, say N.
533
517config HAVE_ARCH_STACKLEAK
518 bool
519 help
520 An architecture should select this if it has the code which
521 fills the used part of the kernel stack with the STACKLEAK_POISON
522 value before returning from system calls.
523
524config HAVE_STACKPROTECTOR

--- 88 unchanged lines hidden (view full) ---

613 Provide kernel/user boundaries probes necessary for subsystems
614 that need it, such as userspace RCU extended quiescent state.
615 Syscalls need to be wrapped inside user_exit()-user_enter(), either
616 optimized behind static key or through the slow path using TIF_NOHZ
617 flag. Exceptions handlers must be wrapped as well. Irqs are already
618 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal
619 handling on irq exit still need to be protected.
620
534config HAVE_ARCH_STACKLEAK
535 bool
536 help
537 An architecture should select this if it has the code which
538 fills the used part of the kernel stack with the STACKLEAK_POISON
539 value before returning from system calls.
540
541config HAVE_STACKPROTECTOR

--- 88 unchanged lines hidden (view full) ---

630 Provide kernel/user boundaries probes necessary for subsystems
631 that need it, such as userspace RCU extended quiescent state.
632 Syscalls need to be wrapped inside user_exit()-user_enter(), either
633 optimized behind static key or through the slow path using TIF_NOHZ
634 flag. Exceptions handlers must be wrapped as well. Irqs are already
635 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal
636 handling on irq exit still need to be protected.
637
621config HAVE_CONTEXT_TRACKING_OFFSTACK
622 bool
623 help
624 Architecture neither relies on exception_enter()/exception_exit()
625 nor on schedule_user(). Also preempt_schedule_notrace() and
626 preempt_schedule_irq() can't be called in a preemptible section
627 while context tracking is CONTEXT_USER. This feature reflects a sane
628 entry implementation where the following requirements are met on
629 critical entry code, ie: before user_exit() or after user_enter():
630
631 - Critical entry code isn't preemptible (or better yet:
632 not interruptible).
633 - No use of RCU read side critical sections, unless rcu_nmi_enter()
634 got called.
635 - No use of instrumentation, unless instrumentation_begin() got
636 called.
637
638config HAVE_TIF_NOHZ
639 bool
640 help
641 Arch relies on TIF_NOHZ and syscall slow path to implement context
642 tracking calls to user_enter()/user_exit().
643
644config HAVE_VIRT_CPU_ACCOUNTING
645 bool

--- 407 unchanged lines hidden ---
638config HAVE_TIF_NOHZ
639 bool
640 help
641 Arch relies on TIF_NOHZ and syscall slow path to implement context
642 tracking calls to user_enter()/user_exit().
643
644config HAVE_VIRT_CPU_ACCOUNTING
645 bool

--- 407 unchanged lines hidden ---