[PATCH v2 23/28] um: Add stub side of SECCOMP/futex based process handling

benjamin at sipsolutions.net benjamin at sipsolutions.net
Tue Nov 22 02:07:54 PST 2022


From: Benjamin Berg <benjamin at sipsolutions.net>

This adds the stub side for the new seccomp process management code. In
this case we do register save/restore through the signal handler
mcontext. For the FS_BASE/GS_BASE register we need special handling.

Co-authored-by: Johannes Berg <johannes at sipsolutions.net>
Signed-off-by: Benjamin Berg <benjamin at sipsolutions.net>
---
 arch/um/include/shared/skas/stub-data.h | 15 +++++++
 arch/um/kernel/skas/clone.c             | 24 +++++++++++
 arch/um/kernel/skas/stub.c              | 54 +++++++++++++++++++++++++
 arch/x86/um/shared/sysdep/stub-data.h   | 12 ++++++
 arch/x86/um/shared/sysdep/stub.h        |  3 ++
 arch/x86/um/shared/sysdep/stub_32.h     |  7 ++++
 arch/x86/um/shared/sysdep/stub_64.h     | 14 +++++++
 7 files changed, 129 insertions(+)
 create mode 100644 arch/x86/um/shared/sysdep/stub-data.h

diff --git a/arch/um/include/shared/skas/stub-data.h b/arch/um/include/shared/skas/stub-data.h
index 821c1e98c051..6a6bc34f70c4 100644
--- a/arch/um/include/shared/skas/stub-data.h
+++ b/arch/um/include/shared/skas/stub-data.h
@@ -8,8 +8,13 @@
 #ifndef __STUB_DATA_H
 #define __STUB_DATA_H
 
+#include <linux/kconfig.h>
 #include <linux/compiler_types.h>
 #include <as-layout.h>
+#include <sysdep/stub-data.h>
+
+#define FUTEX_IN_CHILD 0
+#define FUTEX_IN_KERN 1
 
 #define STUB_NEXT_SYSCALL(s) \
 	((struct stub_syscall *) (((unsigned long) s) + (s)->cmd_len))
@@ -31,6 +36,16 @@ struct stub_data {
 	/* 128 leaves enough room for additional fields in the struct */
 	unsigned char syscall_data[UM_KERN_PAGE_SIZE - 128] __aligned(16);
 
+	/* data shared with signal handler (only used in seccomp mode) */
+	short restart_wait;
+	unsigned int futex;
+	int signal;
+	unsigned short si_offset;
+	unsigned short mctx_offset;
+
+	/* seccomp architecture specific state restore */
+	struct stub_data_arch arch_data;
+
 	/* Stack for our signal handlers and for calling into . */
 	unsigned char sigstack[UM_KERN_PAGE_SIZE] __aligned(UM_KERN_PAGE_SIZE);
 };
diff --git a/arch/um/kernel/skas/clone.c b/arch/um/kernel/skas/clone.c
index 8b6ea9c00133..97728b7dd54d 100644
--- a/arch/um/kernel/skas/clone.c
+++ b/arch/um/kernel/skas/clone.c
@@ -48,3 +48,27 @@ stub_clone_handler(void)
  done:
 	trap_myself();
 }
+
+#ifdef CONFIG_UML_SECCOMP
+void __attribute__ ((__section__ (".__syscall_stub")))
+stub_clone_handler_seccomp(void)
+{
+	struct stub_data *data = get_stub_page() + UM_KERN_PAGE_SIZE;
+	long err;
+
+	/* Use the syscall data as a temporary stack area (bottom half for clone). */
+	err = stub_syscall2(__NR_clone, CLONE_PARENT | CLONE_FILES | SIGCHLD,
+			    (unsigned long) data->syscall_data +
+					    sizeof(data->syscall_data) / 2 -
+					    sizeof(void *));
+	if (err) {
+		data->err = err;
+		goto done;
+	}
+
+	remap_stack_and_trap();
+
+ done:
+	trap_myself();
+}
+#endif
diff --git a/arch/um/kernel/skas/stub.c b/arch/um/kernel/skas/stub.c
index 0a13f5d21d08..4bb90395a12b 100644
--- a/arch/um/kernel/skas/stub.c
+++ b/arch/um/kernel/skas/stub.c
@@ -5,6 +5,11 @@
 
 #include <sysdep/stub.h>
 
+#ifdef CONFIG_UML_SECCOMP
+#include <linux/futex.h>
+#include <errno.h>
+#endif
+
 static __always_inline int syscall_handler(struct stub_data *d)
 {
 	struct stub_syscall *sc;
@@ -45,3 +50,52 @@ stub_syscall_handler(void)
 
 	trap_myself();
 }
+
+#ifdef CONFIG_UML_SECCOMP
+void __attribute__ ((__section__ (".__syscall_stub")))
+stub_signal_interrupt(int sig, siginfo_t *info, void *p)
+{
+	struct stub_data *d = get_stub_page();
+	ucontext_t *uc = p;
+	long res;
+
+	d->signal = sig;
+	d->si_offset = (unsigned long)info - (unsigned long)&d->sigstack[0];
+	d->mctx_offset = (unsigned long)&uc->uc_mcontext - (unsigned long)&d->sigstack[0];
+
+restart_wait:
+	d->futex = FUTEX_IN_KERN;
+	do {
+		res = stub_syscall3(__NR_futex, (unsigned long)&d->futex,
+				    FUTEX_WAKE, 1);
+	} while (res == -EINTR);
+	do {
+		res = stub_syscall4(__NR_futex, (unsigned long)&d->futex,
+				    FUTEX_WAIT, FUTEX_IN_KERN, 0);
+	} while (res == -EINTR || d->futex == FUTEX_IN_KERN);
+
+	if (res < 0 && res != -EAGAIN)
+		stub_syscall2(__NR_kill, 0, SIGKILL);
+
+	/* Try running queued syscalls. */
+	if (syscall_handler(d) < 0 || d->restart_wait) {
+		/* Report SIGTRAP if we restart. */
+		d->signal = SIGTRAP;
+		d->restart_wait = 0;
+		goto restart_wait;
+	}
+
+	/* Return so that the host modified mcontext is restored. */
+}
+
+void __attribute__ ((__section__ (".__syscall_stub")))
+stub_signal_restorer(void)
+{
+	struct stub_data *d = get_stub_page();
+
+	/* Restore arch dependent state that is not part of the mcontext */
+	stub_seccomp_restore_state(&d->arch_data);
+
+	stub_syscall0(__NR_rt_sigreturn);
+}
+#endif
diff --git a/arch/x86/um/shared/sysdep/stub-data.h b/arch/x86/um/shared/sysdep/stub-data.h
new file mode 100644
index 000000000000..b0b6f2a95f0e
--- /dev/null
+++ b/arch/x86/um/shared/sysdep/stub-data.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifdef __i386__
+struct stub_data_arch { };
+#else
+#define STUB_SYNC_FS_BASE (1 << 0)
+#define STUB_SYNC_GS_BASE (1 << 1)
+struct stub_data_arch {
+	int sync;
+	unsigned long fs_base;
+	unsigned long gs_base;
+};
+#endif
diff --git a/arch/x86/um/shared/sysdep/stub.h b/arch/x86/um/shared/sysdep/stub.h
index 579681d12158..eb2e3a24d40b 100644
--- a/arch/x86/um/shared/sysdep/stub.h
+++ b/arch/x86/um/shared/sysdep/stub.h
@@ -14,3 +14,6 @@
 extern void stub_segv_handler(int, siginfo_t *, void *);
 extern void stub_syscall_handler(void);
 extern void stub_clone_handler(void);
+extern void stub_signal_interrupt(int, siginfo_t *, void *);
+extern void stub_signal_restorer(void);
+extern void stub_clone_handler_seccomp(void);
diff --git a/arch/x86/um/shared/sysdep/stub_32.h b/arch/x86/um/shared/sysdep/stub_32.h
index 9531e52b0a3e..c8328a1949ee 100644
--- a/arch/x86/um/shared/sysdep/stub_32.h
+++ b/arch/x86/um/shared/sysdep/stub_32.h
@@ -144,4 +144,11 @@ static __always_inline void *get_stub_page(void)
 
 	return (void *)ret;
 }
+
+static __always_inline void
+stub_seccomp_restore_state(struct stub_data_arch *arch)
+{
+	/* No extra arch specific restore */
+}
+
 #endif
diff --git a/arch/x86/um/shared/sysdep/stub_64.h b/arch/x86/um/shared/sysdep/stub_64.h
index bd2f62530bcc..99c5e54654a7 100644
--- a/arch/x86/um/shared/sysdep/stub_64.h
+++ b/arch/x86/um/shared/sysdep/stub_64.h
@@ -10,6 +10,7 @@
 #include <sysdep/ptrace_user.h>
 #include <generated/asm-offsets.h>
 #include <linux/stddef.h>
+#include <asm/prctl.h>
 
 #define STUB_MMAP_NR __NR_mmap
 #define MMAP_OFFSET(o) (o)
@@ -140,4 +141,17 @@ static __always_inline void *get_stub_page(void)
 
 	return (void *)ret - UM_KERN_PAGE_SIZE;
 }
+
+static __always_inline void
+stub_seccomp_restore_state(struct stub_data_arch *arch)
+{
+	/* TODO: Use _writefsbase_u64/_writegsbase_u64 when possible */
+	if (arch->sync & STUB_SYNC_FS_BASE)
+		stub_syscall2(__NR_arch_prctl, ARCH_SET_FS, arch->fs_base);
+	if (arch->sync & STUB_SYNC_GS_BASE)
+		stub_syscall2(__NR_arch_prctl, ARCH_SET_GS, arch->gs_base);
+
+	arch->sync = 0;
+}
+
 #endif
-- 
2.38.1




More information about the linux-um mailing list