This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
libseccomp (userspace) was rebased: https://github.com/seccomp/libseccomp/pull/134
Fully passes libseccomp regression testing (simulation and live).
There is one failing kernel selftest: global.user_notification_signal
v1 -> v2: - return immediatly if secure_computing(NULL) returns -1 - fixed whitespace issues - add missing seccomp.h - remove patch #2 (solved now) - add riscv to seccomp kernel selftest
Cc: keescook@chromium.org Cc: me@carlosedp.com
Signed-off-by: David Abdurachmanov david.abdurachmanov@sifive.com --- arch/riscv/Kconfig | 14 ++++++++++ arch/riscv/include/asm/seccomp.h | 10 +++++++ arch/riscv/include/asm/thread_info.h | 5 +++- arch/riscv/kernel/entry.S | 27 +++++++++++++++++-- arch/riscv/kernel/ptrace.c | 10 +++++++ tools/testing/selftests/seccomp/seccomp_bpf.c | 8 +++++- 6 files changed, 70 insertions(+), 4 deletions(-) create mode 100644 arch/riscv/include/asm/seccomp.h
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig index 59a4727ecd6c..441e63ff5adc 100644 --- a/arch/riscv/Kconfig +++ b/arch/riscv/Kconfig @@ -31,6 +31,7 @@ config RISCV select GENERIC_SMP_IDLE_THREAD select GENERIC_ATOMIC64 if !64BIT select HAVE_ARCH_AUDITSYSCALL + select HAVE_ARCH_SECCOMP_FILTER select HAVE_MEMBLOCK_NODE_MAP select HAVE_DMA_CONTIGUOUS select HAVE_FUTEX_CMPXCHG if FUTEX @@ -235,6 +236,19 @@ menu "Kernel features"
source "kernel/Kconfig.hz"
+config SECCOMP + bool "Enable seccomp to safely compute untrusted bytecode" + help + This kernel feature is useful for number crunching applications + that may need to compute untrusted bytecode during their + execution. By using pipes or other transports made available to + the process as file descriptors supporting the read/write + syscalls, it's possible to isolate those applications in + their own address space using seccomp. Once seccomp is + enabled via prctl(PR_SET_SECCOMP), it cannot be disabled + and the task is only allowed to execute a few safe syscalls + defined by each seccomp mode. + endmenu
menu "Boot options" diff --git a/arch/riscv/include/asm/seccomp.h b/arch/riscv/include/asm/seccomp.h new file mode 100644 index 000000000000..bf7744ee3b3d --- /dev/null +++ b/arch/riscv/include/asm/seccomp.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#ifndef _ASM_SECCOMP_H +#define _ASM_SECCOMP_H + +#include <asm/unistd.h> + +#include <asm-generic/seccomp.h> + +#endif /* _ASM_SECCOMP_H */ diff --git a/arch/riscv/include/asm/thread_info.h b/arch/riscv/include/asm/thread_info.h index 905372d7eeb8..a0b2a29a0da1 100644 --- a/arch/riscv/include/asm/thread_info.h +++ b/arch/riscv/include/asm/thread_info.h @@ -75,6 +75,7 @@ struct thread_info { #define TIF_MEMDIE 5 /* is terminating due to OOM killer */ #define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */ #define TIF_SYSCALL_AUDIT 7 /* syscall auditing */ +#define TIF_SECCOMP 8 /* syscall secure computing */
#define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -82,11 +83,13 @@ struct thread_info { #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) +#define _TIF_SECCOMP (1 << TIF_SECCOMP)
#define _TIF_WORK_MASK \ (_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | _TIF_NEED_RESCHED)
#define _TIF_SYSCALL_WORK \ - (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT) + (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT | \ + _TIF_SECCOMP )
#endif /* _ASM_RISCV_THREAD_INFO_H */ diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S index bc7a56e1ca6f..0bbedfa3e47d 100644 --- a/arch/riscv/kernel/entry.S +++ b/arch/riscv/kernel/entry.S @@ -203,8 +203,25 @@ check_syscall_nr: /* Check to make sure we don't jump to a bogus syscall number. */ li t0, __NR_syscalls la s0, sys_ni_syscall - /* Syscall number held in a7 */ - bgeu a7, t0, 1f + /* + * The tracer can change syscall number to valid/invalid value. + * We use syscall_set_nr helper in syscall_trace_enter thus we + * cannot trust the current value in a7 and have to reload from + * the current task pt_regs. + */ + REG_L a7, PT_A7(sp) + /* + * Syscall number held in a7. + * If syscall number is above allowed value, redirect to ni_syscall. + */ + bge a7, t0, 1f + /* + * Check if syscall is rejected by tracer or seccomp, i.e., a7 == -1. + * If yes, we pretend it was executed. + */ + li t1, -1 + beq a7, t1, ret_from_syscall_rejected + /* Call syscall */ la s0, sys_call_table slli t0, a7, RISCV_LGPTR add s0, s0, t0 @@ -215,6 +232,12 @@ check_syscall_nr: ret_from_syscall: /* Set user a0 to kernel a0 */ REG_S a0, PT_A0(sp) + /* + * We didn't execute the actual syscall. + * Seccomp already set return value for the current task pt_regs. + * (If it was configured with SECCOMP_RET_ERRNO/TRACE) + */ +ret_from_syscall_rejected: /* Trace syscalls, but only if requested by the user. */ REG_L t0, TASK_TI_FLAGS(tp) andi t0, t0, _TIF_SYSCALL_WORK diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c index 368751438366..63e47c9f85f0 100644 --- a/arch/riscv/kernel/ptrace.c +++ b/arch/riscv/kernel/ptrace.c @@ -154,6 +154,16 @@ void do_syscall_trace_enter(struct pt_regs *regs) if (tracehook_report_syscall_entry(regs)) syscall_set_nr(current, regs, -1);
+ /* + * Do the secure computing after ptrace; failures should be fast. + * If this fails we might have return value in a0 from seccomp + * (via SECCOMP_RET_ERRNO/TRACE). + */ + if (secure_computing(NULL) == -1) { + syscall_set_nr(current, regs, -1); + return; + } + #ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS if (test_thread_flag(TIF_SYSCALL_TRACEPOINT)) trace_sys_enter(regs, syscall_get_nr(current, regs)); diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c index 6ef7f16c4cf5..492e0adad9d3 100644 --- a/tools/testing/selftests/seccomp/seccomp_bpf.c +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c @@ -112,6 +112,8 @@ struct seccomp_data { # define __NR_seccomp 383 # elif defined(__aarch64__) # define __NR_seccomp 277 +# elif defined(__riscv) +# define __NR_seccomp 277 # elif defined(__hppa__) # define __NR_seccomp 338 # elif defined(__powerpc__) @@ -1582,6 +1584,10 @@ TEST_F(TRACE_poke, getpid_runs_normally) # define ARCH_REGS struct user_pt_regs # define SYSCALL_NUM regs[8] # define SYSCALL_RET regs[0] +#elif defined(__riscv) && __riscv_xlen == 64 +# define ARCH_REGS struct user_regs_struct +# define SYSCALL_NUM a7 +# define SYSCALL_RET a0 #elif defined(__hppa__) # define ARCH_REGS struct user_regs_struct # define SYSCALL_NUM gr[20] @@ -1671,7 +1677,7 @@ void change_syscall(struct __test_metadata *_metadata, EXPECT_EQ(0, ret) {}
#if defined(__x86_64__) || defined(__i386__) || defined(__powerpc__) || \ - defined(__s390__) || defined(__hppa__) + defined(__s390__) || defined(__hppa__) || defined(__riscv) { regs.SYSCALL_NUM = syscall; }
On Thu, Aug 22, 2019 at 5:56 PM David Abdurachmanov david.abdurachmanov@gmail.com wrote:
This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
libseccomp (userspace) was rebased: https://github.com/seccomp/libseccomp/pull/134
Fully passes libseccomp regression testing (simulation and live).
There is one failing kernel selftest: global.user_notification_signal
v1 -> v2:
- return immediatly if secure_computing(NULL) returns -1
- fixed whitespace issues
- add missing seccomp.h
- remove patch #2 (solved now)
- add riscv to seccomp kernel selftest
Cc: keescook@chromium.org Cc: me@carlosedp.com
Signed-off-by: David Abdurachmanov david.abdurachmanov@sifive.com
arch/riscv/Kconfig | 14 ++++++++++ arch/riscv/include/asm/seccomp.h | 10 +++++++ arch/riscv/include/asm/thread_info.h | 5 +++- arch/riscv/kernel/entry.S | 27 +++++++++++++++++-- arch/riscv/kernel/ptrace.c | 10 +++++++ tools/testing/selftests/seccomp/seccomp_bpf.c | 8 +++++- 6 files changed, 70 insertions(+), 4 deletions(-) create mode 100644 arch/riscv/include/asm/seccomp.h
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig index 59a4727ecd6c..441e63ff5adc 100644 --- a/arch/riscv/Kconfig +++ b/arch/riscv/Kconfig @@ -31,6 +31,7 @@ config RISCV select GENERIC_SMP_IDLE_THREAD select GENERIC_ATOMIC64 if !64BIT select HAVE_ARCH_AUDITSYSCALL
select HAVE_ARCH_SECCOMP_FILTER select HAVE_MEMBLOCK_NODE_MAP select HAVE_DMA_CONTIGUOUS select HAVE_FUTEX_CMPXCHG if FUTEX@@ -235,6 +236,19 @@ menu "Kernel features"
source "kernel/Kconfig.hz"
+config SECCOMP
bool "Enable seccomp to safely compute untrusted bytecode"helpThis kernel feature is useful for number crunching applicationsthat may need to compute untrusted bytecode during theirexecution. By using pipes or other transports made available tothe process as file descriptors supporting the read/writesyscalls, it's possible to isolate those applications intheir own address space using seccomp. Once seccomp isenabled via prctl(PR_SET_SECCOMP), it cannot be disabledand the task is only allowed to execute a few safe syscallsdefined by each seccomp mode.endmenu
menu "Boot options" diff --git a/arch/riscv/include/asm/seccomp.h b/arch/riscv/include/asm/seccomp.h new file mode 100644 index 000000000000..bf7744ee3b3d --- /dev/null +++ b/arch/riscv/include/asm/seccomp.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_SECCOMP_H +#define _ASM_SECCOMP_H
+#include <asm/unistd.h>
+#include <asm-generic/seccomp.h>
+#endif /* _ASM_SECCOMP_H */ diff --git a/arch/riscv/include/asm/thread_info.h b/arch/riscv/include/asm/thread_info.h index 905372d7eeb8..a0b2a29a0da1 100644 --- a/arch/riscv/include/asm/thread_info.h +++ b/arch/riscv/include/asm/thread_info.h @@ -75,6 +75,7 @@ struct thread_info { #define TIF_MEMDIE 5 /* is terminating due to OOM killer */ #define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */ #define TIF_SYSCALL_AUDIT 7 /* syscall auditing */ +#define TIF_SECCOMP 8 /* syscall secure computing */
#define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -82,11 +83,13 @@ struct thread_info { #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) +#define _TIF_SECCOMP (1 << TIF_SECCOMP)
#define _TIF_WORK_MASK \ (_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | _TIF_NEED_RESCHED)
#define _TIF_SYSCALL_WORK \
(_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT)
(_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT | \_TIF_SECCOMP )#endif /* _ASM_RISCV_THREAD_INFO_H */ diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S index bc7a56e1ca6f..0bbedfa3e47d 100644 --- a/arch/riscv/kernel/entry.S +++ b/arch/riscv/kernel/entry.S @@ -203,8 +203,25 @@ check_syscall_nr: /* Check to make sure we don't jump to a bogus syscall number. */ li t0, __NR_syscalls la s0, sys_ni_syscall
/* Syscall number held in a7 */bgeu a7, t0, 1f
/** The tracer can change syscall number to valid/invalid value.* We use syscall_set_nr helper in syscall_trace_enter thus we* cannot trust the current value in a7 and have to reload from* the current task pt_regs.*/REG_L a7, PT_A7(sp)/** Syscall number held in a7.* If syscall number is above allowed value, redirect to ni_syscall.*/bge a7, t0, 1f/** Check if syscall is rejected by tracer or seccomp, i.e., a7 == -1.* If yes, we pretend it was executed.*/li t1, -1beq a7, t1, ret_from_syscall_rejected/* Call syscall */ la s0, sys_call_table slli t0, a7, RISCV_LGPTR add s0, s0, t0@@ -215,6 +232,12 @@ check_syscall_nr: ret_from_syscall: /* Set user a0 to kernel a0 */ REG_S a0, PT_A0(sp)
/** We didn't execute the actual syscall.* Seccomp already set return value for the current task pt_regs.* (If it was configured with SECCOMP_RET_ERRNO/TRACE)*/+ret_from_syscall_rejected: /* Trace syscalls, but only if requested by the user. */ REG_L t0, TASK_TI_FLAGS(tp) andi t0, t0, _TIF_SYSCALL_WORK diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c index 368751438366..63e47c9f85f0 100644 --- a/arch/riscv/kernel/ptrace.c +++ b/arch/riscv/kernel/ptrace.c @@ -154,6 +154,16 @@ void do_syscall_trace_enter(struct pt_regs *regs) if (tracehook_report_syscall_entry(regs)) syscall_set_nr(current, regs, -1);
/** Do the secure computing after ptrace; failures should be fast.* If this fails we might have return value in a0 from seccomp* (via SECCOMP_RET_ERRNO/TRACE).*/if (secure_computing(NULL) == -1) {syscall_set_nr(current, regs, -1);return;}#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS if (test_thread_flag(TIF_SYSCALL_TRACEPOINT)) trace_sys_enter(regs, syscall_get_nr(current, regs)); diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c index 6ef7f16c4cf5..492e0adad9d3 100644 --- a/tools/testing/selftests/seccomp/seccomp_bpf.c +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c @@ -112,6 +112,8 @@ struct seccomp_data { # define __NR_seccomp 383 # elif defined(__aarch64__) # define __NR_seccomp 277 +# elif defined(__riscv) +# define __NR_seccomp 277 # elif defined(__hppa__) # define __NR_seccomp 338 # elif defined(__powerpc__) @@ -1582,6 +1584,10 @@ TEST_F(TRACE_poke, getpid_runs_normally) # define ARCH_REGS struct user_pt_regs # define SYSCALL_NUM regs[8] # define SYSCALL_RET regs[0] +#elif defined(__riscv) && __riscv_xlen == 64 +# define ARCH_REGS struct user_regs_struct +# define SYSCALL_NUM a7 +# define SYSCALL_RET a0 #elif defined(__hppa__) # define ARCH_REGS struct user_regs_struct # define SYSCALL_NUM gr[20] @@ -1671,7 +1677,7 @@ void change_syscall(struct __test_metadata *_metadata, EXPECT_EQ(0, ret) {}
#if defined(__x86_64__) || defined(__i386__) || defined(__powerpc__) || \
- defined(__s390__) || defined(__hppa__)
- defined(__s390__) || defined(__hppa__) || defined(__riscv) { regs.SYSCALL_NUM = syscall; }
-- 2.21.0
Tested-by: Carlos de Paula me@carlosedp.com
On Thu, Aug 22, 2019 at 5:56 PM David Abdurachmanov david.abdurachmanov@gmail.com wrote:
This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
libseccomp (userspace) was rebased: https://github.com/seccomp/libseccomp/pull/134
Fully passes libseccomp regression testing (simulation and live).
There is one failing kernel selftest: global.user_notification_signal
v1 -> v2:
- return immediatly if secure_computing(NULL) returns -1
- fixed whitespace issues
- add missing seccomp.h
- remove patch #2 (solved now)
- add riscv to seccomp kernel selftest
Cc: keescook@chromium.org Cc: me@carlosedp.com
Signed-off-by: David Abdurachmanov david.abdurachmanov@sifive.com Tested-by: Carlos de Paula me@carlosedp.com
arch/riscv/Kconfig | 14 ++++++++++ arch/riscv/include/asm/seccomp.h | 10 +++++++ arch/riscv/include/asm/thread_info.h | 5 +++- arch/riscv/kernel/entry.S | 27 +++++++++++++++++-- arch/riscv/kernel/ptrace.c | 10 +++++++ tools/testing/selftests/seccomp/seccomp_bpf.c | 8 +++++- 6 files changed, 70 insertions(+), 4 deletions(-) create mode 100644 arch/riscv/include/asm/seccomp.h
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig index 59a4727ecd6c..441e63ff5adc 100644 --- a/arch/riscv/Kconfig +++ b/arch/riscv/Kconfig @@ -31,6 +31,7 @@ config RISCV select GENERIC_SMP_IDLE_THREAD select GENERIC_ATOMIC64 if !64BIT select HAVE_ARCH_AUDITSYSCALL
select HAVE_ARCH_SECCOMP_FILTER select HAVE_MEMBLOCK_NODE_MAP select HAVE_DMA_CONTIGUOUS select HAVE_FUTEX_CMPXCHG if FUTEX@@ -235,6 +236,19 @@ menu "Kernel features"
source "kernel/Kconfig.hz"
+config SECCOMP
bool "Enable seccomp to safely compute untrusted bytecode"helpThis kernel feature is useful for number crunching applicationsthat may need to compute untrusted bytecode during theirexecution. By using pipes or other transports made available tothe process as file descriptors supporting the read/writesyscalls, it's possible to isolate those applications intheir own address space using seccomp. Once seccomp isenabled via prctl(PR_SET_SECCOMP), it cannot be disabledand the task is only allowed to execute a few safe syscallsdefined by each seccomp mode.endmenu
menu "Boot options" diff --git a/arch/riscv/include/asm/seccomp.h b/arch/riscv/include/asm/seccomp.h new file mode 100644 index 000000000000..bf7744ee3b3d --- /dev/null +++ b/arch/riscv/include/asm/seccomp.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_SECCOMP_H +#define _ASM_SECCOMP_H
+#include <asm/unistd.h>
+#include <asm-generic/seccomp.h>
+#endif /* _ASM_SECCOMP_H */ diff --git a/arch/riscv/include/asm/thread_info.h b/arch/riscv/include/asm/thread_info.h index 905372d7eeb8..a0b2a29a0da1 100644 --- a/arch/riscv/include/asm/thread_info.h +++ b/arch/riscv/include/asm/thread_info.h @@ -75,6 +75,7 @@ struct thread_info { #define TIF_MEMDIE 5 /* is terminating due to OOM killer */ #define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */ #define TIF_SYSCALL_AUDIT 7 /* syscall auditing */ +#define TIF_SECCOMP 8 /* syscall secure computing */
#define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -82,11 +83,13 @@ struct thread_info { #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) +#define _TIF_SECCOMP (1 << TIF_SECCOMP)
#define _TIF_WORK_MASK \ (_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | _TIF_NEED_RESCHED)
#define _TIF_SYSCALL_WORK \
(_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT)
(_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT | \_TIF_SECCOMP )#endif /* _ASM_RISCV_THREAD_INFO_H */ diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S index bc7a56e1ca6f..0bbedfa3e47d 100644 --- a/arch/riscv/kernel/entry.S +++ b/arch/riscv/kernel/entry.S @@ -203,8 +203,25 @@ check_syscall_nr: /* Check to make sure we don't jump to a bogus syscall number. */ li t0, __NR_syscalls la s0, sys_ni_syscall
/* Syscall number held in a7 */bgeu a7, t0, 1f
/** The tracer can change syscall number to valid/invalid value.* We use syscall_set_nr helper in syscall_trace_enter thus we* cannot trust the current value in a7 and have to reload from* the current task pt_regs.*/REG_L a7, PT_A7(sp)/** Syscall number held in a7.* If syscall number is above allowed value, redirect to ni_syscall.*/bge a7, t0, 1f/** Check if syscall is rejected by tracer or seccomp, i.e., a7 == -1.* If yes, we pretend it was executed.*/li t1, -1beq a7, t1, ret_from_syscall_rejected/* Call syscall */ la s0, sys_call_table slli t0, a7, RISCV_LGPTR add s0, s0, t0@@ -215,6 +232,12 @@ check_syscall_nr: ret_from_syscall: /* Set user a0 to kernel a0 */ REG_S a0, PT_A0(sp)
/** We didn't execute the actual syscall.* Seccomp already set return value for the current task pt_regs.* (If it was configured with SECCOMP_RET_ERRNO/TRACE)*/+ret_from_syscall_rejected: /* Trace syscalls, but only if requested by the user. */ REG_L t0, TASK_TI_FLAGS(tp) andi t0, t0, _TIF_SYSCALL_WORK diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c index 368751438366..63e47c9f85f0 100644 --- a/arch/riscv/kernel/ptrace.c +++ b/arch/riscv/kernel/ptrace.c @@ -154,6 +154,16 @@ void do_syscall_trace_enter(struct pt_regs *regs) if (tracehook_report_syscall_entry(regs)) syscall_set_nr(current, regs, -1);
/** Do the secure computing after ptrace; failures should be fast.* If this fails we might have return value in a0 from seccomp* (via SECCOMP_RET_ERRNO/TRACE).*/if (secure_computing(NULL) == -1) {syscall_set_nr(current, regs, -1);return;}#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS if (test_thread_flag(TIF_SYSCALL_TRACEPOINT)) trace_sys_enter(regs, syscall_get_nr(current, regs)); diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c index 6ef7f16c4cf5..492e0adad9d3 100644 --- a/tools/testing/selftests/seccomp/seccomp_bpf.c +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c @@ -112,6 +112,8 @@ struct seccomp_data { # define __NR_seccomp 383 # elif defined(__aarch64__) # define __NR_seccomp 277 +# elif defined(__riscv) +# define __NR_seccomp 277 # elif defined(__hppa__) # define __NR_seccomp 338 # elif defined(__powerpc__) @@ -1582,6 +1584,10 @@ TEST_F(TRACE_poke, getpid_runs_normally) # define ARCH_REGS struct user_pt_regs # define SYSCALL_NUM regs[8] # define SYSCALL_RET regs[0] +#elif defined(__riscv) && __riscv_xlen == 64 +# define ARCH_REGS struct user_regs_struct +# define SYSCALL_NUM a7 +# define SYSCALL_RET a0 #elif defined(__hppa__) # define ARCH_REGS struct user_regs_struct # define SYSCALL_NUM gr[20] @@ -1671,7 +1677,7 @@ void change_syscall(struct __test_metadata *_metadata, EXPECT_EQ(0, ret) {}
#if defined(__x86_64__) || defined(__i386__) || defined(__powerpc__) || \
- defined(__s390__) || defined(__hppa__)
- defined(__s390__) || defined(__hppa__) || defined(__riscv) { regs.SYSCALL_NUM = syscall; }
-- 2.21.0
Kernel selftests results:
➜ uname -a Linux fedora-unleashed 5.2.0-rc7-30159-g2d072d4-dirty #3 SMP Thu Jul 4 20:18:21 -03 2019 riscv64 riscv64 riscv64 GNU/Linux
➜ sudo ./seccomp_bpf [==========] Running 74 tests from 1 test cases. [ RUN ] global.mode_strict_support [ OK ] global.mode_strict_support [ RUN ] global.mode_strict_cannot_call_prctl [ OK ] global.mode_strict_cannot_call_prctl [ RUN ] global.no_new_privs_support [ OK ] global.no_new_privs_support [ RUN ] global.mode_filter_support [ OK ] global.mode_filter_support [ RUN ] global.mode_filter_without_nnp [ OK ] global.mode_filter_without_nnp [ RUN ] global.filter_size_limits [ OK ] global.filter_size_limits [ RUN ] global.filter_chain_limits [ OK ] global.filter_chain_limits [ RUN ] global.mode_filter_cannot_move_to_strict [ OK ] global.mode_filter_cannot_move_to_strict [ RUN ] global.mode_filter_get_seccomp [ OK ] global.mode_filter_get_seccomp [ RUN ] global.ALLOW_all [ OK ] global.ALLOW_all [ RUN ] global.empty_prog [ OK ] global.empty_prog [ RUN ] global.log_all [ OK ] global.log_all [ RUN ] global.unknown_ret_is_kill_inside [ OK ] global.unknown_ret_is_kill_inside [ RUN ] global.unknown_ret_is_kill_above_allow [ OK ] global.unknown_ret_is_kill_above_allow [ RUN ] global.KILL_all [ OK ] global.KILL_all [ RUN ] global.KILL_one [ OK ] global.KILL_one [ RUN ] global.KILL_one_arg_one [ OK ] global.KILL_one_arg_one [ RUN ] global.KILL_one_arg_six [ OK ] global.KILL_one_arg_six [ RUN ] global.KILL_thread [ OK ] global.KILL_thread [ RUN ] global.KILL_process [ OK ] global.KILL_process [ RUN ] global.arg_out_of_range [ OK ] global.arg_out_of_range [ RUN ] global.ERRNO_valid [ OK ] global.ERRNO_valid [ RUN ] global.ERRNO_zero [ OK ] global.ERRNO_zero [ RUN ] global.ERRNO_capped [ OK ] global.ERRNO_capped [ RUN ] global.ERRNO_order [ OK ] global.ERRNO_order [ RUN ] TRAP.dfl [ OK ] TRAP.dfl [ RUN ] TRAP.ign [ OK ] TRAP.ign [ RUN ] TRAP.handler [ OK ] TRAP.handler [ RUN ] precedence.allow_ok [ OK ] precedence.allow_ok [ RUN ] precedence.kill_is_highest [ OK ] precedence.kill_is_highest [ RUN ] precedence.kill_is_highest_in_any_order [ OK ] precedence.kill_is_highest_in_any_order [ RUN ] precedence.trap_is_second [ OK ] precedence.trap_is_second [ RUN ] precedence.trap_is_second_in_any_order [ OK ] precedence.trap_is_second_in_any_order [ RUN ] precedence.errno_is_third [ OK ] precedence.errno_is_third [ RUN ] precedence.errno_is_third_in_any_order [ OK ] precedence.errno_is_third_in_any_order [ RUN ] precedence.trace_is_fourth [ OK ] precedence.trace_is_fourth [ RUN ] precedence.trace_is_fourth_in_any_order [ OK ] precedence.trace_is_fourth_in_any_order [ RUN ] precedence.log_is_fifth [ OK ] precedence.log_is_fifth [ RUN ] precedence.log_is_fifth_in_any_order [ OK ] precedence.log_is_fifth_in_any_order [ RUN ] TRACE_poke.read_has_side_effects [ OK ] TRACE_poke.read_has_side_effects [ RUN ] TRACE_poke.getpid_runs_normally [ OK ] TRACE_poke.getpid_runs_normally [ RUN ] TRACE_syscall.ptrace_syscall_redirected [ OK ] TRACE_syscall.ptrace_syscall_redirected [ RUN ] TRACE_syscall.ptrace_syscall_errno [ OK ] TRACE_syscall.ptrace_syscall_errno [ RUN ] TRACE_syscall.ptrace_syscall_faked [ OK ] TRACE_syscall.ptrace_syscall_faked [ RUN ] TRACE_syscall.syscall_allowed [ OK ] TRACE_syscall.syscall_allowed [ RUN ] TRACE_syscall.syscall_redirected [ OK ] TRACE_syscall.syscall_redirected [ RUN ] TRACE_syscall.syscall_errno [ OK ] TRACE_syscall.syscall_errno [ RUN ] TRACE_syscall.syscall_faked [ OK ] TRACE_syscall.syscall_faked [ RUN ] TRACE_syscall.skip_after_RET_TRACE [ OK ] TRACE_syscall.skip_after_RET_TRACE [ RUN ] TRACE_syscall.kill_after_RET_TRACE [ OK ] TRACE_syscall.kill_after_RET_TRACE [ RUN ] TRACE_syscall.skip_after_ptrace [ OK ] TRACE_syscall.skip_after_ptrace [ RUN ] TRACE_syscall.kill_after_ptrace [ OK ] TRACE_syscall.kill_after_ptrace [ RUN ] global.seccomp_syscall [ OK ] global.seccomp_syscall [ RUN ] global.seccomp_syscall_mode_lock [ OK ] global.seccomp_syscall_mode_lock [ RUN ] global.detect_seccomp_filter_flags [ OK ] global.detect_seccomp_filter_flags [ RUN ] global.TSYNC_first [ OK ] global.TSYNC_first [ RUN ] TSYNC.siblings_fail_prctl [ OK ] TSYNC.siblings_fail_prctl [ RUN ] TSYNC.two_siblings_with_ancestor [ OK ] TSYNC.two_siblings_with_ancestor [ RUN ] TSYNC.two_sibling_want_nnp [ OK ] TSYNC.two_sibling_want_nnp [ RUN ] TSYNC.two_siblings_with_no_filter [ OK ] TSYNC.two_siblings_with_no_filter [ RUN ] TSYNC.two_siblings_with_one_divergence [ OK ] TSYNC.two_siblings_with_one_divergence [ RUN ] TSYNC.two_siblings_not_under_filter [ OK ] TSYNC.two_siblings_not_under_filter [ RUN ] global.syscall_restart [ OK ] global.syscall_restart [ RUN ] global.filter_flag_log [ OK ] global.filter_flag_log [ RUN ] global.get_action_avail [ OK ] global.get_action_avail [ RUN ] global.get_metadata [ OK ] global.get_metadata [ RUN ] global.user_notification_basic [ OK ] global.user_notification_basic [ RUN ] global.user_notification_kill_in_middle [ OK ] global.user_notification_kill_in_middle [ RUN ] global.user_notification_signal [1] 5951 alarm sudo ./seccomp_bpf
carlosedp in ~ at fedora-unleashed ➜ sudo ./seccomp_benchmark Calibrating reasonable sample size... 1564584448.964538790 - 1564584448.964529687 = 9103 1564584448.964588859 - 1564584448.964575204 = 13655 1564584448.964631342 - 1564584448.964604790 = 26552 1564584448.964710239 - 1564584448.964644997 = 65242 1564584448.964842239 - 1564584448.964726928 = 115311 1564584448.965072859 - 1564584448.964857411 = 215448 1564584448.965513618 - 1564584448.965089549 = 424069 1564584448.966417894 - 1564584448.965532584 = 885310 1564584448.968286377 - 1564584448.966443687 = 1842690 1564584448.971667549 - 1564584448.968314446 = 3353103 1564584448.978288790 - 1564584448.971694101 = 6594689 1564584448.991803618 - 1564584448.978313066 = 13490552 1564584449.017692308 - 1564584448.991836239 = 25856069 1564584449.069651756 - 1564584449.017713549 = 51938207 1564584449.173110928 - 1564584449.069673756 = 103437172 1564584449.380001204 - 1564584449.173132928 = 206868276 1564584449.793857618 - 1564584449.380041411 = 413816207 1564584450.625367342 - 1564584449.793898584 = 831468758 1564584452.299529411 - 1564584450.625426514 = 1674102897 1564584455.665938307 - 1564584452.299592376 = 3366345931 1564584462.331777479 - 1564584455.665973962 = 6665803517 Benchmarking 33554432 samples... 18.107882743 - 12.075641371 = 6032241372 getpid native: 179 ns 34.720410331 - 18.107978605 = 16612431726 getpid RET_ALLOW: 495 ns Estimated seccomp overhead per syscall: 316 n
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Is this the only failing test? Or are the rest of the selftests skipped when this test fails, and no further tests are run, as seems to be shown here:
https://lore.kernel.org/linux-riscv/CADnnUqcmDMRe1f+3jG8SPR6jRrnBsY8VVD70VbK...
For example, looking at the source, I'd naively expect to see the user_notification_closed_listener test result -- which follows right after the failing test in the selftest source. But there aren't any results?
Also - could you follow up with the author of this failing test to see if we can get some more clarity about what might be going wrong here? It appears that the failing test was added in commit 6a21cc50f0c7f ("seccomp: add a return code to trap to userspace") by Tycho Andersen tycho@tycho.ws.
- Paul
On Fri, Aug 23, 2019 at 5:30 PM Paul Walmsley paul.walmsley@sifive.com wrote:
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Is this the only failing test? Or are the rest of the selftests skipped when this test fails, and no further tests are run, as seems to be shown here:
https://lore.kernel.org/linux-riscv/CADnnUqcmDMRe1f+3jG8SPR6jRrnBsY8VVD70VbK...
Yes, it's a single test failing. After removing global.user_notification_signal test everything else pass and you get the results printed.
For example, looking at the source, I'd naively expect to see the user_notification_closed_listener test result -- which follows right after the failing test in the selftest source. But there aren't any results?
Yes, it hangs at this point. You have to manually terminate it.
Also - could you follow up with the author of this failing test to see if we can get some more clarity about what might be going wrong here? It appears that the failing test was added in commit 6a21cc50f0c7f ("seccomp: add a return code to trap to userspace") by Tycho Andersen tycho@tycho.ws.
Well the code states ".. and hope that it doesn't break when there is actually a signal :)". Maybe we are just unlucky. I don't have results from other architectures to compare.
I found that Linaro is running selftests, but SECCOMP is disabled and thus it's failing. Is there another CI which tracks selftests?
https://qa-reports.linaro.org/lkft/linux-next-oe/tests/kselftest/seccomp_sec...
- Paul
On Fri, Aug 23, 2019 at 6:04 PM David Abdurachmanov david.abdurachmanov@gmail.com wrote:
On Fri, Aug 23, 2019 at 5:30 PM Paul Walmsley paul.walmsley@sifive.com wrote:
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Is this the only failing test? Or are the rest of the selftests skipped when this test fails, and no further tests are run, as seems to be shown here:
https://lore.kernel.org/linux-riscv/CADnnUqcmDMRe1f+3jG8SPR6jRrnBsY8VVD70VbK...
Yes, it's a single test failing. After removing global.user_notification_signal test everything else pass and you get the results printed.
For example, looking at the source, I'd naively expect to see the user_notification_closed_listener test result -- which follows right after the failing test in the selftest source. But there aren't any results?
Yes, it hangs at this point. You have to manually terminate it.
Also - could you follow up with the author of this failing test to see if we can get some more clarity about what might be going wrong here? It appears that the failing test was added in commit 6a21cc50f0c7f ("seccomp: add a return code to trap to userspace") by Tycho Andersen tycho@tycho.ws.
Well the code states ".. and hope that it doesn't break when there is actually a signal :)". Maybe we are just unlucky. I don't have results from other architectures to compare.
I found that Linaro is running selftests, but SECCOMP is disabled and thus it's failing. Is there another CI which tracks selftests?
https://qa-reports.linaro.org/lkft/linux-next-oe/tests/kselftest/seccomp_sec...
Actually it seems that seccomp is enabled in kernel, but not in systemd, and somehow seccomp_bpf is missing on all arches thus causing automatic failure.
- Paul
On Fri, 23 Aug 2019, David Abdurachmanov wrote:
On Fri, Aug 23, 2019 at 5:30 PM Paul Walmsley paul.walmsley@sifive.com wrote:
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Is this the only failing test? Or are the rest of the selftests skipped when this test fails, and no further tests are run, as seems to be shown here:
https://lore.kernel.org/linux-riscv/CADnnUqcmDMRe1f+3jG8SPR6jRrnBsY8VVD70VbK...
Yes, it's a single test failing. After removing global.user_notification_signal test everything else pass and you get the results printed.
OK.
Well the code states ".. and hope that it doesn't break when there is actually a signal :)". Maybe we are just unlucky. I don't have results from other architectures to compare.
I found that Linaro is running selftests, but SECCOMP is disabled and thus it's failing. Is there another CI which tracks selftests?
0day runs the kselftests, and at least on some architectures/Kconfigs, it's succeeding:
https://lore.kernel.org/lkml/20190726083740.GG22106@shao2-debian/
https://lore.kernel.org/lkml/20190712064850.GC20848@shao2-debian/
https://lore.kernel.org/lkml/20190311074115.GC10839@shao2-debian/
etc.
- Paul
On Fri, Aug 23, 2019 at 05:30:53PM -0700, Paul Walmsley wrote:
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Is this the only failing test? Or are the rest of the selftests skipped when this test fails, and no further tests are run, as seems to be shown here:
https://lore.kernel.org/linux-riscv/CADnnUqcmDMRe1f+3jG8SPR6jRrnBsY8VVD70VbK...
For example, looking at the source, I'd naively expect to see the user_notification_closed_listener test result -- which follows right after the failing test in the selftest source. But there aren't any results?
Also - could you follow up with the author of this failing test to see if we can get some more clarity about what might be going wrong here? It appears that the failing test was added in commit 6a21cc50f0c7f ("seccomp: add a return code to trap to userspace") by Tycho Andersen tycho@tycho.ws.
So, the original email says the riscv series is tested on top of 5.2-rc7, but just for fun, can you confirm that you're building a tree that includes 9dd3fcb0ab73 ("selftests/seccomp: Handle namespace failures gracefully")? I assume it does, but I suspect something similar is happening, where the environment is slightly different than expected and the test stalls.
Does it behave the same way under emulation (i.e. can I hope to reproduce this myself?)
On Wed, Aug 28, 2019 at 10:36 AM Kees Cook keescook@chromium.org wrote:
On Fri, Aug 23, 2019 at 05:30:53PM -0700, Paul Walmsley wrote:
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Is this the only failing test? Or are the rest of the selftests skipped when this test fails, and no further tests are run, as seems to be shown here:
https://lore.kernel.org/linux-riscv/CADnnUqcmDMRe1f+3jG8SPR6jRrnBsY8VVD70VbK...
For example, looking at the source, I'd naively expect to see the user_notification_closed_listener test result -- which follows right after the failing test in the selftest source. But there aren't any results?
Also - could you follow up with the author of this failing test to see if we can get some more clarity about what might be going wrong here? It appears that the failing test was added in commit 6a21cc50f0c7f ("seccomp: add a return code to trap to userspace") by Tycho Andersen tycho@tycho.ws.
So, the original email says the riscv series is tested on top of 5.2-rc7, but just for fun, can you confirm that you're building a tree that includes 9dd3fcb0ab73 ("selftests/seccomp: Handle namespace failures gracefully")? I assume it does, but I suspect something similar is happening, where the environment is slightly different than expected and the test stalls.
Does it behave the same way under emulation (i.e. can I hope to reproduce this myself?)
This was tested in 5.2-rc7 and later in 5.3-rc with the same behavior. Also VM or physical HW doesn't matter, same result.
-- Kees Cook
Hi,
On Fri, Aug 23, 2019 at 05:30:53PM -0700, Paul Walmsley wrote:
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Also - could you follow up with the author of this failing test to see if we can get some more clarity about what might be going wrong here? It appears that the failing test was added in commit 6a21cc50f0c7f ("seccomp: add a return code to trap to userspace") by Tycho Andersen tycho@tycho.ws.
Can you post an strace and a cat of /proc/$pid/stack for both tasks where it gets stuck? I don't have any riscv hardware, and it "works for me" on x86 and arm64 with 100 tries.
Thanks,
Tycho
On Mon, Aug 26, 2019 at 7:57 AM Tycho Andersen tycho@tycho.ws wrote:
Hi,
On Fri, Aug 23, 2019 at 05:30:53PM -0700, Paul Walmsley wrote:
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
There is one failing kernel selftest: global.user_notification_signal
Also - could you follow up with the author of this failing test to see if we can get some more clarity about what might be going wrong here? It appears that the failing test was added in commit 6a21cc50f0c7f ("seccomp: add a return code to trap to userspace") by Tycho Andersen tycho@tycho.ws.
Can you post an strace and a cat of /proc/$pid/stack for both tasks where it gets stuck? I don't have any riscv hardware, and it "works for me" on x86 and arm64 with 100 tries.
I don't have the a build with SECCOMP for the board right now, so it will have to wait. I just finished a new kernel (almost rc6) for Fedora, but it will take time to assemble new repositories and a disk image.
There is older disk image available (5.2.0-rc7 kernel with v2 SECCOMP) for QEMU or libvirt/QEMU:
https://dl.fedoraproject.org/pub/alt/risc-v/disk-images/fedora/rawhide/20190... https://fedoraproject.org/wiki/Architectures/RISC-V/Installing#Boot_with_lib...
(If you are interesting trying it locally.)
IIRC I attempted to connected with strace, but it quickly returns and fails properly. Simply put strace unblocks whatever is stuck.
david
On Mon, Aug 26, 2019 at 09:39:50AM -0700, David Abdurachmanov wrote:
I don't have the a build with SECCOMP for the board right now, so it will have to wait. I just finished a new kernel (almost rc6) for Fedora,
FWIW, I don't think this should block landing the code: all the tests fail without seccomp support. ;) So this patch is an improvement!
Hi Kees,
On Mon, 26 Aug 2019, Kees Cook wrote:
On Mon, Aug 26, 2019 at 09:39:50AM -0700, David Abdurachmanov wrote:
I don't have the a build with SECCOMP for the board right now, so it will have to wait. I just finished a new kernel (almost rc6) for Fedora,
FWIW, I don't think this should block landing the code: all the tests fail without seccomp support. ;) So this patch is an improvement!
Am sympathetic to this -- we did it with the hugetlb patches for RISC-V -- but it would be good to understand a little bit more about why the test fails before we merge it.
Once we merge the patch, it will probably reduce the motivation for others to either understand and fix the underlying problem with the RISC-V code -- or, if it truly is a flaky test, to drop (or fix) the test in the seccomp_bpf kselftests.
Thanks for helping to take a closer look at this,
- Paul
On Wed, Aug 28, 2019 at 6:30 PM Paul Walmsley paul.walmsley@sifive.com wrote:
On Mon, 26 Aug 2019, Kees Cook wrote:
On Mon, Aug 26, 2019 at 09:39:50AM -0700, David Abdurachmanov wrote:
I don't have the a build with SECCOMP for the board right now, so it will have to wait. I just finished a new kernel (almost rc6) for Fedora,
FWIW, I don't think this should block landing the code: all the tests fail without seccomp support. ;) So this patch is an improvement!
Am sympathetic to this -- we did it with the hugetlb patches for RISC-V -- but it would be good to understand a little bit more about why the test fails before we merge it.
The test is almost certainly failing due to the environmental requirements (i.e. namespaces, user ids, etc). There are some corner cases in there that we've had to fix in the past. If the other tests are passing, then I would expect all the seccomp internals are fine -- it's just the case being weird. It's just a matter of figuring out what state the test environment is in so we can cover that corner case too.
Once we merge the patch, it will probably reduce the motivation for others to either understand and fix the underlying problem with the RISC-V code -- or, if it truly is a flaky test, to drop (or fix) the test in the seccomp_bpf kselftests.
Sure, I get that point -- but I don't want to block seccomp landing for riscv for that. I suggested to David offlist that the test could just be marked with a FIXME XFAIL on riscv and once someone's in a better position to reproduce it we can fix it. (I think the test bug is almost certainly not riscv specific, but just some missing requirement that we aren't handling correctly.)
How does that sound?
-Kees
On Fri, 27 Sep 2019, Kees Cook wrote:
On Wed, Aug 28, 2019 at 6:30 PM Paul Walmsley paul.walmsley@sifive.com wrote:
On Mon, 26 Aug 2019, Kees Cook wrote:
On Mon, Aug 26, 2019 at 09:39:50AM -0700, David Abdurachmanov wrote:
I don't have the a build with SECCOMP for the board right now, so it will have to wait. I just finished a new kernel (almost rc6) for Fedora,
FWIW, I don't think this should block landing the code: all the tests fail without seccomp support. ;) So this patch is an improvement!
Am sympathetic to this -- we did it with the hugetlb patches for RISC-V -- but it would be good to understand a little bit more about why the test fails before we merge it.
The test is almost certainly failing due to the environmental requirements (i.e. namespaces, user ids, etc). There are some corner cases in there that we've had to fix in the past. If the other tests are passing, then I would expect all the seccomp internals are fine -- it's just the case being weird. It's just a matter of figuring out what state the test environment is in so we can cover that corner case too.
Once we merge the patch, it will probably reduce the motivation for others to either understand and fix the underlying problem with the RISC-V code -- or, if it truly is a flaky test, to drop (or fix) the test in the seccomp_bpf kselftests.
Sure, I get that point -- but I don't want to block seccomp landing for riscv for that. I suggested to David offlist that the test could just be marked with a FIXME XFAIL on riscv and once someone's in a better position to reproduce it we can fix it. (I think the test bug is almost certainly not riscv specific, but just some missing requirement that we aren't handling correctly.)
OK. It might be nice to mark the seccomp_bpf.c test as flaky in the comments for the test.
How does that sound?
Let's follow your plan. Thanks for your review and feedback.
- Paul
On Thu, Aug 22, 2019 at 01:55:22PM -0700, David Abdurachmanov wrote:
This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
Oops, I see the mention of QEMU here. Where's the best place to find instructions on creating a qemu riscv image/environment?
There is one failing kernel selftest: global.user_notification_signal
This test has been fragile (and is not arch-specific), so as long as everything else is passing, I would call this patch ready to go. :)
Reviewed-by: Kees Cook keescook@chromium.org
On Aug 25, 2019, at 2:59 PM, Kees Cook keescook@chromium.org wrote:
On Thu, Aug 22, 2019 at 01:55:22PM -0700, David Abdurachmanov wrote: This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
Oops, I see the mention of QEMU here. Where's the best place to find instructions on creating a qemu riscv image/environment?
I don’t suppose one of you riscv folks would like to contribute riscv support to virtme? virtme-run —arch=riscv would be quite nice, and the total patch should be just a couple lines. Unfortunately, it helps a lot to understand the subtleties of booting the architecture to write those couple lines :)
On Wed, Aug 28, 2019 at 10:52:05AM -0700, Andy Lutomirski wrote:
On Aug 25, 2019, at 2:59 PM, Kees Cook keescook@chromium.org wrote:
On Thu, Aug 22, 2019 at 01:55:22PM -0700, David Abdurachmanov wrote: This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
Oops, I see the mention of QEMU here. Where's the best place to find instructions on creating a qemu riscv image/environment?
I don’t suppose one of you riscv folks would like to contribute riscv support to virtme? virtme-run —arch=riscv would be quite nice, and the total patch should be just a couple lines. Unfortunately, it helps a lot to understand the subtleties of booting the architecture to write those couple lines :)
As it turns out, this is where I'm stuck. All the instructions I can find are about booting a kernel off a disk image. :(
On Wed, 28 Aug 2019 10:52:05 PDT (-0700), luto@amacapital.net wrote:
On Aug 25, 2019, at 2:59 PM, Kees Cook keescook@chromium.org wrote:
On Thu, Aug 22, 2019 at 01:55:22PM -0700, David Abdurachmanov wrote: This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
Oops, I see the mention of QEMU here. Where's the best place to find instructions on creating a qemu riscv image/environment?
I don’t suppose one of you riscv folks would like to contribute riscv support to virtme? virtme-run —arch=riscv would be quite nice, and the total patch should be just a couple lines. Unfortunately, it helps a lot to understand the subtleties of booting the architecture to write those couple lines :)
What mailing list should I sent this to? You need to use the "virtme" branch of kernel.org/palmer/linux.git until I send the defconfig patches.
commit a8bd7b318691891991caea298f9a5ed0f815c322 gpg: Signature made Tue 03 Sep 2019 03:22:45 PM PDT gpg: using RSA key 00CE76D1834960DFCE886DF8EF4CA1502CCBAB41 gpg: issuer "palmer@dabbelt.com" gpg: Good signature from "Palmer Dabbelt palmer@dabbelt.com" [ultimate] gpg: aka "Palmer Dabbelt palmer@sifive.com" [ultimate] Author: Palmer Dabbelt palmer@sifive.com Date: Tue Sep 3 14:39:39 2019 -0700
Add RISC-V support
This expects a kernel with the plan 9 stuff supported (not yet in defconfig) and a new QEMU (as described in the README). I'm also not 100% sure it's working, as I'm getting
/bin/sh: exec: line 1: /run/virtme/guesttools/virtme-init: not found
Signed-off-by: Palmer Dabbelt palmer@sifive.com
diff --git a/README.md b/README.md index 51b6583..d53a456 100644 --- a/README.md +++ b/README.md @@ -112,6 +112,14 @@ PPC64
PPC64 appears to be reasonably functional.
+RISC-V +------ + +riscv64 works out of the box, but you'll neet at least QEMU-4.1.0 to be +able to run `vmlinux`-style kernels. riscv32 is not supported because +there are no existing userspace images for it. Support is provided via +QEMU's `virt` machine with OpenSBI for firmware. + Others ------
diff --git a/virtme/architectures.py b/virtme/architectures.py index 9871ea4..ee84494 100644 --- a/virtme/architectures.py +++ b/virtme/architectures.py @@ -207,6 +207,30 @@ class Arch_ppc64(Arch): # Apparently SLOF (QEMU's bundled firmware?) can't boot a zImage. return 'vmlinux'
+class Arch_riscv64(Arch): + def __init__(self, name): + Arch.__init__(self, name) + + self.defconfig_target = 'riscv64_defconfig' + self.qemuname = 'riscv64' + self.linuxname = 'riscv' + self.gccname = 'riscv64' + + def qemuargs(self, is_native): + ret = Arch.qemuargs(is_native) + + ret.extend(['-machine', 'virt']) + ret.extend(['-bios', 'default']) + + return ret + + @staticmethod + def serial_console_args(): + return ['console=ttyS0'] + + def kimg_path(self): + return 'arch/riscv/boot/Image' + class Arch_sparc64(Arch): def __init__(self, name): Arch.__init__(self, name) @@ -264,6 +288,7 @@ ARCHES = { 'arm': Arch_arm, 'aarch64': Arch_aarch64, 'ppc64': Arch_ppc64, + 'riscv64': Arch_riscv64, 'sparc64': Arch_sparc64, 's390x': Arch_s390x, }
On Tue, Sep 3, 2019 at 3:27 PM Palmer Dabbelt palmer@sifive.com wrote:
On Wed, 28 Aug 2019 10:52:05 PDT (-0700), luto@amacapital.net wrote:
On Aug 25, 2019, at 2:59 PM, Kees Cook keescook@chromium.org wrote:
On Thu, Aug 22, 2019 at 01:55:22PM -0700, David Abdurachmanov wrote: This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
Oops, I see the mention of QEMU here. Where's the best place to find instructions on creating a qemu riscv image/environment?
I don’t suppose one of you riscv folks would like to contribute riscv support to virtme? virtme-run —arch=riscv would be quite nice, and the total patch should be just a couple lines. Unfortunately, it helps a lot to understand the subtleties of booting the architecture to write those couple lines :)
What mailing list should I sent this to? You need to use the "virtme" branch of kernel.org/palmer/linux.git until I send the defconfig patches.
commit a8bd7b318691891991caea298f9a5ed0f815c322 gpg: Signature made Tue 03 Sep 2019 03:22:45 PM PDT gpg: using RSA key 00CE76D1834960DFCE886DF8EF4CA1502CCBAB41 gpg: issuer "palmer@dabbelt.com" gpg: Good signature from "Palmer Dabbelt palmer@dabbelt.com" [ultimate] gpg: aka "Palmer Dabbelt palmer@sifive.com" [ultimate] Author: Palmer Dabbelt palmer@sifive.com Date: Tue Sep 3 14:39:39 2019 -0700
Add RISC-V support
Could you rebase onto virtme master and resend in some format that isn't corrupt? git am really doesn't like your patch and, even if I fix it up manually, your gpg: lines are bogus. You could also send a PR at https://github.com/amluto/virtme
On Fri, Sep 27, 2019 at 1:58 PM Andy Lutomirski luto@kernel.org wrote:
On Tue, Sep 3, 2019 at 3:27 PM Palmer Dabbelt palmer@sifive.com wrote:
On Wed, 28 Aug 2019 10:52:05 PDT (-0700), luto@amacapital.net wrote:
On Aug 25, 2019, at 2:59 PM, Kees Cook keescook@chromium.org wrote:
On Thu, Aug 22, 2019 at 01:55:22PM -0700, David Abdurachmanov wrote: This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
Oops, I see the mention of QEMU here. Where's the best place to find instructions on creating a qemu riscv image/environment?
I don’t suppose one of you riscv folks would like to contribute riscv support to virtme? virtme-run —arch=riscv would be quite nice, and the total patch should be just a couple lines. Unfortunately, it helps a lot to understand the subtleties of booting the architecture to write those couple lines :)
FYI, it works now:
$ virtme-configkernel --arch=riscv --defconfig GEN Makefile [...] Configured. Build with 'make ARCH=riscv CROSS_COMPILE=riscv64-linux-gnu- -j4'
$ make ARCH=riscv CROSS_COMPILE=riscv64-linux-gnu- -j4 [...]
$ virtme-run --kdir=. --arch=riscv64 --mods=auto --root [path to a riscv filesystem]
This is with virtme master and a qemu-system-riscv64 from qemu git on my path. It does *not* work with Fedora 30's qemu.
So now you can all jump on the virtme bandwagon and have an easy way to test riscv kernels. :) Although, if you want to run kernel selftests, you may find the process of actually running them to be more fun if you use --rodir or --rwdir to map the kernel selftests directory into the guest.
On Wed, Aug 28, 2019 at 10:36 AM Kees Cook keescook@chromium.org wrote:
On Thu, Aug 22, 2019 at 01:55:22PM -0700, David Abdurachmanov wrote:
This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
Oops, I see the mention of QEMU here. Where's the best place to find instructions on creating a qemu riscv image/environment?
Examples from what I personally use: https://github.com/riscv/meta-riscv https://fedoraproject.org/wiki/Architectures/RISC-V/Installing#Boot_with_lib... (might be outdated)
If you are running machine with a properly working libvirt/QEMU setup:
VIRTBUILDER_IMAGE=fedora-rawhide-developer-20190703n0 FIRMWARE=fw_payload-uboot-qemu-virt-smode.elf wget https://dl.fedoraproject.org/pub/alt/risc-v/disk-images/fedora/rawhide/20190... echo riscv > /tmp/rootpw virt-builder \ --verbose \ --source https://dl.fedoraproject.org/pub/alt/risc-v/repo/virt-builder-images/images/... \ --no-check-signature \ --arch riscv64 \ --size 10G \ --format raw \ --hostname fedora-riscv \ -o disk \ --root-password file:/tmp/rootpw \ ${VIRTBUILDER_IMAGE}
sudo virt-install \ --name fedora-riscv \ --arch riscv64 \ --vcpus 4 \ --memory 3048 \ --import \ --disk path=$PWD/disk \ --boot kernel=$PWD/${FIRMWARE} \ --network network=default \ --graphics none \ --serial log.file=/tmp/fedora-riscv.serial.log \ --noautoconsole
The following does incl. SECCOMP v2 patch on top of 5.2-rc7 kernel.
There is one failing kernel selftest: global.user_notification_signal
This test has been fragile (and is not arch-specific), so as long as everything else is passing, I would call this patch ready to go. :)
Reviewed-by: Kees Cook keescook@chromium.org
-- Kees Cook
Hello Shuah,
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
libseccomp (userspace) was rebased: https://github.com/seccomp/libseccomp/pull/134
Fully passes libseccomp regression testing (simulation and live).
There is one failing kernel selftest: global.user_notification_signal
v1 -> v2:
- return immediatly if secure_computing(NULL) returns -1
- fixed whitespace issues
- add missing seccomp.h
- remove patch #2 (solved now)
- add riscv to seccomp kernel selftest
Cc: keescook@chromium.org Cc: me@carlosedp.com
Signed-off-by: David Abdurachmanov david.abdurachmanov@sifive.com
We'd like to merge this patch through the RISC-V tree. Care to ack the change to tools/testing/selftests/seccomp/seccomp_bpf.c ?
Kees has already reviewed it:
https://lore.kernel.org/linux-riscv/CAJr-aD=UnCN9E_mdVJ2H5nt=6juRSWikZnA5HxD...
- Paul
arch/riscv/Kconfig | 14 ++++++++++ arch/riscv/include/asm/seccomp.h | 10 +++++++ arch/riscv/include/asm/thread_info.h | 5 +++- arch/riscv/kernel/entry.S | 27 +++++++++++++++++-- arch/riscv/kernel/ptrace.c | 10 +++++++ tools/testing/selftests/seccomp/seccomp_bpf.c | 8 +++++- 6 files changed, 70 insertions(+), 4 deletions(-) create mode 100644 arch/riscv/include/asm/seccomp.h
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig index 59a4727ecd6c..441e63ff5adc 100644 --- a/arch/riscv/Kconfig +++ b/arch/riscv/Kconfig @@ -31,6 +31,7 @@ config RISCV select GENERIC_SMP_IDLE_THREAD select GENERIC_ATOMIC64 if !64BIT select HAVE_ARCH_AUDITSYSCALL
- select HAVE_ARCH_SECCOMP_FILTER select HAVE_MEMBLOCK_NODE_MAP select HAVE_DMA_CONTIGUOUS select HAVE_FUTEX_CMPXCHG if FUTEX
@@ -235,6 +236,19 @@ menu "Kernel features" source "kernel/Kconfig.hz" +config SECCOMP
- bool "Enable seccomp to safely compute untrusted bytecode"
- help
This kernel feature is useful for number crunching applicationsthat may need to compute untrusted bytecode during theirexecution. By using pipes or other transports made available tothe process as file descriptors supporting the read/writesyscalls, it's possible to isolate those applications intheir own address space using seccomp. Once seccomp isenabled via prctl(PR_SET_SECCOMP), it cannot be disabledand the task is only allowed to execute a few safe syscallsdefined by each seccomp mode.endmenu menu "Boot options" diff --git a/arch/riscv/include/asm/seccomp.h b/arch/riscv/include/asm/seccomp.h new file mode 100644 index 000000000000..bf7744ee3b3d --- /dev/null +++ b/arch/riscv/include/asm/seccomp.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_SECCOMP_H +#define _ASM_SECCOMP_H
+#include <asm/unistd.h>
+#include <asm-generic/seccomp.h>
+#endif /* _ASM_SECCOMP_H */ diff --git a/arch/riscv/include/asm/thread_info.h b/arch/riscv/include/asm/thread_info.h index 905372d7eeb8..a0b2a29a0da1 100644 --- a/arch/riscv/include/asm/thread_info.h +++ b/arch/riscv/include/asm/thread_info.h @@ -75,6 +75,7 @@ struct thread_info { #define TIF_MEMDIE 5 /* is terminating due to OOM killer */ #define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */ #define TIF_SYSCALL_AUDIT 7 /* syscall auditing */ +#define TIF_SECCOMP 8 /* syscall secure computing */ #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -82,11 +83,13 @@ struct thread_info { #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) +#define _TIF_SECCOMP (1 << TIF_SECCOMP) #define _TIF_WORK_MASK \ (_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | _TIF_NEED_RESCHED) #define _TIF_SYSCALL_WORK \
- (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT)
- (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT | \
_TIF_SECCOMP )#endif /* _ASM_RISCV_THREAD_INFO_H */ diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S index bc7a56e1ca6f..0bbedfa3e47d 100644 --- a/arch/riscv/kernel/entry.S +++ b/arch/riscv/kernel/entry.S @@ -203,8 +203,25 @@ check_syscall_nr: /* Check to make sure we don't jump to a bogus syscall number. */ li t0, __NR_syscalls la s0, sys_ni_syscall
- /* Syscall number held in a7 */
- bgeu a7, t0, 1f
- /*
* The tracer can change syscall number to valid/invalid value.* We use syscall_set_nr helper in syscall_trace_enter thus we* cannot trust the current value in a7 and have to reload from* the current task pt_regs.*/- REG_L a7, PT_A7(sp)
- /*
* Syscall number held in a7.* If syscall number is above allowed value, redirect to ni_syscall.*/- bge a7, t0, 1f
- /*
* Check if syscall is rejected by tracer or seccomp, i.e., a7 == -1.* If yes, we pretend it was executed.*/- li t1, -1
- beq a7, t1, ret_from_syscall_rejected
- /* Call syscall */ la s0, sys_call_table slli t0, a7, RISCV_LGPTR add s0, s0, t0
@@ -215,6 +232,12 @@ check_syscall_nr: ret_from_syscall: /* Set user a0 to kernel a0 */ REG_S a0, PT_A0(sp)
- /*
* We didn't execute the actual syscall.* Seccomp already set return value for the current task pt_regs.* (If it was configured with SECCOMP_RET_ERRNO/TRACE)*/+ret_from_syscall_rejected: /* Trace syscalls, but only if requested by the user. */ REG_L t0, TASK_TI_FLAGS(tp) andi t0, t0, _TIF_SYSCALL_WORK diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c index 368751438366..63e47c9f85f0 100644 --- a/arch/riscv/kernel/ptrace.c +++ b/arch/riscv/kernel/ptrace.c @@ -154,6 +154,16 @@ void do_syscall_trace_enter(struct pt_regs *regs) if (tracehook_report_syscall_entry(regs)) syscall_set_nr(current, regs, -1);
- /*
* Do the secure computing after ptrace; failures should be fast.* If this fails we might have return value in a0 from seccomp* (via SECCOMP_RET_ERRNO/TRACE).*/- if (secure_computing(NULL) == -1) {
syscall_set_nr(current, regs, -1);return;- }
#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS if (test_thread_flag(TIF_SYSCALL_TRACEPOINT)) trace_sys_enter(regs, syscall_get_nr(current, regs)); diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c index 6ef7f16c4cf5..492e0adad9d3 100644 --- a/tools/testing/selftests/seccomp/seccomp_bpf.c +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c @@ -112,6 +112,8 @@ struct seccomp_data { # define __NR_seccomp 383 # elif defined(__aarch64__) # define __NR_seccomp 277 +# elif defined(__riscv) +# define __NR_seccomp 277 # elif defined(__hppa__) # define __NR_seccomp 338 # elif defined(__powerpc__) @@ -1582,6 +1584,10 @@ TEST_F(TRACE_poke, getpid_runs_normally) # define ARCH_REGS struct user_pt_regs # define SYSCALL_NUM regs[8] # define SYSCALL_RET regs[0] +#elif defined(__riscv) && __riscv_xlen == 64 +# define ARCH_REGS struct user_regs_struct +# define SYSCALL_NUM a7 +# define SYSCALL_RET a0 #elif defined(__hppa__) # define ARCH_REGS struct user_regs_struct # define SYSCALL_NUM gr[20] @@ -1671,7 +1677,7 @@ void change_syscall(struct __test_metadata *_metadata, EXPECT_EQ(0, ret) {} #if defined(__x86_64__) || defined(__i386__) || defined(__powerpc__) || \
- defined(__s390__) || defined(__hppa__)
- defined(__s390__) || defined(__hppa__) || defined(__riscv) { regs.SYSCALL_NUM = syscall; }
-- 2.21.0
- Paul
Shuah,
Could you please take a quick look at this and ack it if you're OK with the tools/testing change? We'd like to get this merged soon.
- Paul
On Fri, 4 Oct 2019, Paul Walmsley wrote:
Hello Shuah,
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
libseccomp (userspace) was rebased: https://github.com/seccomp/libseccomp/pull/134
Fully passes libseccomp regression testing (simulation and live).
There is one failing kernel selftest: global.user_notification_signal
v1 -> v2:
- return immediatly if secure_computing(NULL) returns -1
- fixed whitespace issues
- add missing seccomp.h
- remove patch #2 (solved now)
- add riscv to seccomp kernel selftest
Cc: keescook@chromium.org Cc: me@carlosedp.com
Signed-off-by: David Abdurachmanov david.abdurachmanov@sifive.com
We'd like to merge this patch through the RISC-V tree. Care to ack the change to tools/testing/selftests/seccomp/seccomp_bpf.c ?
Kees has already reviewed it:
https://lore.kernel.org/linux-riscv/CAJr-aD=UnCN9E_mdVJ2H5nt=6juRSWikZnA5HxD...
- Paul
arch/riscv/Kconfig | 14 ++++++++++ arch/riscv/include/asm/seccomp.h | 10 +++++++ arch/riscv/include/asm/thread_info.h | 5 +++- arch/riscv/kernel/entry.S | 27 +++++++++++++++++-- arch/riscv/kernel/ptrace.c | 10 +++++++ tools/testing/selftests/seccomp/seccomp_bpf.c | 8 +++++- 6 files changed, 70 insertions(+), 4 deletions(-) create mode 100644 arch/riscv/include/asm/seccomp.h
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig index 59a4727ecd6c..441e63ff5adc 100644 --- a/arch/riscv/Kconfig +++ b/arch/riscv/Kconfig @@ -31,6 +31,7 @@ config RISCV select GENERIC_SMP_IDLE_THREAD select GENERIC_ATOMIC64 if !64BIT select HAVE_ARCH_AUDITSYSCALL
- select HAVE_ARCH_SECCOMP_FILTER select HAVE_MEMBLOCK_NODE_MAP select HAVE_DMA_CONTIGUOUS select HAVE_FUTEX_CMPXCHG if FUTEX
@@ -235,6 +236,19 @@ menu "Kernel features" source "kernel/Kconfig.hz" +config SECCOMP
- bool "Enable seccomp to safely compute untrusted bytecode"
- help
This kernel feature is useful for number crunching applicationsthat may need to compute untrusted bytecode during theirexecution. By using pipes or other transports made available tothe process as file descriptors supporting the read/writesyscalls, it's possible to isolate those applications intheir own address space using seccomp. Once seccomp isenabled via prctl(PR_SET_SECCOMP), it cannot be disabledand the task is only allowed to execute a few safe syscallsdefined by each seccomp mode.endmenu menu "Boot options" diff --git a/arch/riscv/include/asm/seccomp.h b/arch/riscv/include/asm/seccomp.h new file mode 100644 index 000000000000..bf7744ee3b3d --- /dev/null +++ b/arch/riscv/include/asm/seccomp.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_SECCOMP_H +#define _ASM_SECCOMP_H
+#include <asm/unistd.h>
+#include <asm-generic/seccomp.h>
+#endif /* _ASM_SECCOMP_H */ diff --git a/arch/riscv/include/asm/thread_info.h b/arch/riscv/include/asm/thread_info.h index 905372d7eeb8..a0b2a29a0da1 100644 --- a/arch/riscv/include/asm/thread_info.h +++ b/arch/riscv/include/asm/thread_info.h @@ -75,6 +75,7 @@ struct thread_info { #define TIF_MEMDIE 5 /* is terminating due to OOM killer */ #define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */ #define TIF_SYSCALL_AUDIT 7 /* syscall auditing */ +#define TIF_SECCOMP 8 /* syscall secure computing */ #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -82,11 +83,13 @@ struct thread_info { #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) +#define _TIF_SECCOMP (1 << TIF_SECCOMP) #define _TIF_WORK_MASK \ (_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | _TIF_NEED_RESCHED) #define _TIF_SYSCALL_WORK \
- (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT)
- (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT | \
_TIF_SECCOMP )#endif /* _ASM_RISCV_THREAD_INFO_H */ diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S index bc7a56e1ca6f..0bbedfa3e47d 100644 --- a/arch/riscv/kernel/entry.S +++ b/arch/riscv/kernel/entry.S @@ -203,8 +203,25 @@ check_syscall_nr: /* Check to make sure we don't jump to a bogus syscall number. */ li t0, __NR_syscalls la s0, sys_ni_syscall
- /* Syscall number held in a7 */
- bgeu a7, t0, 1f
- /*
* The tracer can change syscall number to valid/invalid value.* We use syscall_set_nr helper in syscall_trace_enter thus we* cannot trust the current value in a7 and have to reload from* the current task pt_regs.*/- REG_L a7, PT_A7(sp)
- /*
* Syscall number held in a7.* If syscall number is above allowed value, redirect to ni_syscall.*/- bge a7, t0, 1f
- /*
* Check if syscall is rejected by tracer or seccomp, i.e., a7 == -1.* If yes, we pretend it was executed.*/- li t1, -1
- beq a7, t1, ret_from_syscall_rejected
- /* Call syscall */ la s0, sys_call_table slli t0, a7, RISCV_LGPTR add s0, s0, t0
@@ -215,6 +232,12 @@ check_syscall_nr: ret_from_syscall: /* Set user a0 to kernel a0 */ REG_S a0, PT_A0(sp)
- /*
* We didn't execute the actual syscall.* Seccomp already set return value for the current task pt_regs.* (If it was configured with SECCOMP_RET_ERRNO/TRACE)*/+ret_from_syscall_rejected: /* Trace syscalls, but only if requested by the user. */ REG_L t0, TASK_TI_FLAGS(tp) andi t0, t0, _TIF_SYSCALL_WORK diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c index 368751438366..63e47c9f85f0 100644 --- a/arch/riscv/kernel/ptrace.c +++ b/arch/riscv/kernel/ptrace.c @@ -154,6 +154,16 @@ void do_syscall_trace_enter(struct pt_regs *regs) if (tracehook_report_syscall_entry(regs)) syscall_set_nr(current, regs, -1);
- /*
* Do the secure computing after ptrace; failures should be fast.* If this fails we might have return value in a0 from seccomp* (via SECCOMP_RET_ERRNO/TRACE).*/- if (secure_computing(NULL) == -1) {
syscall_set_nr(current, regs, -1);return;- }
#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS if (test_thread_flag(TIF_SYSCALL_TRACEPOINT)) trace_sys_enter(regs, syscall_get_nr(current, regs)); diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c index 6ef7f16c4cf5..492e0adad9d3 100644 --- a/tools/testing/selftests/seccomp/seccomp_bpf.c +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c @@ -112,6 +112,8 @@ struct seccomp_data { # define __NR_seccomp 383 # elif defined(__aarch64__) # define __NR_seccomp 277 +# elif defined(__riscv) +# define __NR_seccomp 277 # elif defined(__hppa__) # define __NR_seccomp 338 # elif defined(__powerpc__) @@ -1582,6 +1584,10 @@ TEST_F(TRACE_poke, getpid_runs_normally) # define ARCH_REGS struct user_pt_regs # define SYSCALL_NUM regs[8] # define SYSCALL_RET regs[0] +#elif defined(__riscv) && __riscv_xlen == 64 +# define ARCH_REGS struct user_regs_struct +# define SYSCALL_NUM a7 +# define SYSCALL_RET a0 #elif defined(__hppa__) # define ARCH_REGS struct user_regs_struct # define SYSCALL_NUM gr[20] @@ -1671,7 +1677,7 @@ void change_syscall(struct __test_metadata *_metadata, EXPECT_EQ(0, ret) {} #if defined(__x86_64__) || defined(__i386__) || defined(__powerpc__) || \
- defined(__s390__) || defined(__hppa__)
- defined(__s390__) || defined(__hppa__) || defined(__riscv) { regs.SYSCALL_NUM = syscall; }
-- 2.21.0
- Paul
On Mon, Oct 14, 2019 at 02:06:07PM -0700, Paul Walmsley wrote:
Shuah,
Could you please take a quick look at this and ack it if you're OK with the tools/testing change? We'd like to get this merged soon.
FWIW, I regularly carry these kinds of selftest changes via my seccomp tree, so if Shuah is busy, I think it'll be fine to take this in riscv. If not, I'll take responsibility of apologizing to Shuah! :) :)
-Kees
- Paul
On Fri, 4 Oct 2019, Paul Walmsley wrote:
Hello Shuah,
On Thu, 22 Aug 2019, David Abdurachmanov wrote:
This patch was extensively tested on Fedora/RISCV (applied by default on top of 5.2-rc7 kernel for <2 months). The patch was also tested with 5.3-rc on QEMU and SiFive Unleashed board.
libseccomp (userspace) was rebased: https://github.com/seccomp/libseccomp/pull/134
Fully passes libseccomp regression testing (simulation and live).
There is one failing kernel selftest: global.user_notification_signal
v1 -> v2:
- return immediatly if secure_computing(NULL) returns -1
- fixed whitespace issues
- add missing seccomp.h
- remove patch #2 (solved now)
- add riscv to seccomp kernel selftest
Cc: keescook@chromium.org Cc: me@carlosedp.com
Signed-off-by: David Abdurachmanov david.abdurachmanov@sifive.com
We'd like to merge this patch through the RISC-V tree. Care to ack the change to tools/testing/selftests/seccomp/seccomp_bpf.c ?
Kees has already reviewed it:
https://lore.kernel.org/linux-riscv/CAJr-aD=UnCN9E_mdVJ2H5nt=6juRSWikZnA5HxD...
- Paul
arch/riscv/Kconfig | 14 ++++++++++ arch/riscv/include/asm/seccomp.h | 10 +++++++ arch/riscv/include/asm/thread_info.h | 5 +++- arch/riscv/kernel/entry.S | 27 +++++++++++++++++-- arch/riscv/kernel/ptrace.c | 10 +++++++ tools/testing/selftests/seccomp/seccomp_bpf.c | 8 +++++- 6 files changed, 70 insertions(+), 4 deletions(-) create mode 100644 arch/riscv/include/asm/seccomp.h
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig index 59a4727ecd6c..441e63ff5adc 100644 --- a/arch/riscv/Kconfig +++ b/arch/riscv/Kconfig @@ -31,6 +31,7 @@ config RISCV select GENERIC_SMP_IDLE_THREAD select GENERIC_ATOMIC64 if !64BIT select HAVE_ARCH_AUDITSYSCALL
- select HAVE_ARCH_SECCOMP_FILTER select HAVE_MEMBLOCK_NODE_MAP select HAVE_DMA_CONTIGUOUS select HAVE_FUTEX_CMPXCHG if FUTEX
@@ -235,6 +236,19 @@ menu "Kernel features" source "kernel/Kconfig.hz" +config SECCOMP
- bool "Enable seccomp to safely compute untrusted bytecode"
- help
This kernel feature is useful for number crunching applicationsthat may need to compute untrusted bytecode during theirexecution. By using pipes or other transports made available tothe process as file descriptors supporting the read/writesyscalls, it's possible to isolate those applications intheir own address space using seccomp. Once seccomp isenabled via prctl(PR_SET_SECCOMP), it cannot be disabledand the task is only allowed to execute a few safe syscallsdefined by each seccomp mode.endmenu menu "Boot options" diff --git a/arch/riscv/include/asm/seccomp.h b/arch/riscv/include/asm/seccomp.h new file mode 100644 index 000000000000..bf7744ee3b3d --- /dev/null +++ b/arch/riscv/include/asm/seccomp.h @@ -0,0 +1,10 @@ +/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_SECCOMP_H +#define _ASM_SECCOMP_H
+#include <asm/unistd.h>
+#include <asm-generic/seccomp.h>
+#endif /* _ASM_SECCOMP_H */ diff --git a/arch/riscv/include/asm/thread_info.h b/arch/riscv/include/asm/thread_info.h index 905372d7eeb8..a0b2a29a0da1 100644 --- a/arch/riscv/include/asm/thread_info.h +++ b/arch/riscv/include/asm/thread_info.h @@ -75,6 +75,7 @@ struct thread_info { #define TIF_MEMDIE 5 /* is terminating due to OOM killer */ #define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */ #define TIF_SYSCALL_AUDIT 7 /* syscall auditing */ +#define TIF_SECCOMP 8 /* syscall secure computing */ #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -82,11 +83,13 @@ struct thread_info { #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT) +#define _TIF_SECCOMP (1 << TIF_SECCOMP) #define _TIF_WORK_MASK \ (_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | _TIF_NEED_RESCHED) #define _TIF_SYSCALL_WORK \
- (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT)
- (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_TRACEPOINT | _TIF_SYSCALL_AUDIT | \
_TIF_SECCOMP )#endif /* _ASM_RISCV_THREAD_INFO_H */ diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S index bc7a56e1ca6f..0bbedfa3e47d 100644 --- a/arch/riscv/kernel/entry.S +++ b/arch/riscv/kernel/entry.S @@ -203,8 +203,25 @@ check_syscall_nr: /* Check to make sure we don't jump to a bogus syscall number. */ li t0, __NR_syscalls la s0, sys_ni_syscall
- /* Syscall number held in a7 */
- bgeu a7, t0, 1f
- /*
* The tracer can change syscall number to valid/invalid value.* We use syscall_set_nr helper in syscall_trace_enter thus we* cannot trust the current value in a7 and have to reload from* the current task pt_regs.*/- REG_L a7, PT_A7(sp)
- /*
* Syscall number held in a7.* If syscall number is above allowed value, redirect to ni_syscall.*/- bge a7, t0, 1f
- /*
* Check if syscall is rejected by tracer or seccomp, i.e., a7 == -1.* If yes, we pretend it was executed.*/- li t1, -1
- beq a7, t1, ret_from_syscall_rejected
- /* Call syscall */ la s0, sys_call_table slli t0, a7, RISCV_LGPTR add s0, s0, t0
@@ -215,6 +232,12 @@ check_syscall_nr: ret_from_syscall: /* Set user a0 to kernel a0 */ REG_S a0, PT_A0(sp)
- /*
* We didn't execute the actual syscall.* Seccomp already set return value for the current task pt_regs.* (If it was configured with SECCOMP_RET_ERRNO/TRACE)*/+ret_from_syscall_rejected: /* Trace syscalls, but only if requested by the user. */ REG_L t0, TASK_TI_FLAGS(tp) andi t0, t0, _TIF_SYSCALL_WORK diff --git a/arch/riscv/kernel/ptrace.c b/arch/riscv/kernel/ptrace.c index 368751438366..63e47c9f85f0 100644 --- a/arch/riscv/kernel/ptrace.c +++ b/arch/riscv/kernel/ptrace.c @@ -154,6 +154,16 @@ void do_syscall_trace_enter(struct pt_regs *regs) if (tracehook_report_syscall_entry(regs)) syscall_set_nr(current, regs, -1);
- /*
* Do the secure computing after ptrace; failures should be fast.* If this fails we might have return value in a0 from seccomp* (via SECCOMP_RET_ERRNO/TRACE).*/- if (secure_computing(NULL) == -1) {
syscall_set_nr(current, regs, -1);return;- }
#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS if (test_thread_flag(TIF_SYSCALL_TRACEPOINT)) trace_sys_enter(regs, syscall_get_nr(current, regs)); diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c index 6ef7f16c4cf5..492e0adad9d3 100644 --- a/tools/testing/selftests/seccomp/seccomp_bpf.c +++ b/tools/testing/selftests/seccomp/seccomp_bpf.c @@ -112,6 +112,8 @@ struct seccomp_data { # define __NR_seccomp 383 # elif defined(__aarch64__) # define __NR_seccomp 277 +# elif defined(__riscv) +# define __NR_seccomp 277 # elif defined(__hppa__) # define __NR_seccomp 338 # elif defined(__powerpc__) @@ -1582,6 +1584,10 @@ TEST_F(TRACE_poke, getpid_runs_normally) # define ARCH_REGS struct user_pt_regs # define SYSCALL_NUM regs[8] # define SYSCALL_RET regs[0] +#elif defined(__riscv) && __riscv_xlen == 64 +# define ARCH_REGS struct user_regs_struct +# define SYSCALL_NUM a7 +# define SYSCALL_RET a0 #elif defined(__hppa__) # define ARCH_REGS struct user_regs_struct # define SYSCALL_NUM gr[20] @@ -1671,7 +1677,7 @@ void change_syscall(struct __test_metadata *_metadata, EXPECT_EQ(0, ret) {} #if defined(__x86_64__) || defined(__i386__) || defined(__powerpc__) || \
- defined(__s390__) || defined(__hppa__)
- defined(__s390__) || defined(__hppa__) || defined(__riscv) { regs.SYSCALL_NUM = syscall; }
-- 2.21.0
- Paul
linux-kselftest-mirror@lists.linaro.org
-
Andy Lutomirski -
Andy Lutomirski -
Carlos Eduardo de Paula -
David Abdurachmanov -
Kees Cook -
Palmer Dabbelt -
Paul Walmsley -
Tycho Andersen