Merge branch 'for-next' of https://git.kernel.org/pub/scm/linux/kernel/git/s390/linux.git

12 files changed, 118 insertions, 27 deletions

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index 7138d923d86f..840d81f7a750 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -213,6 +213,7 @@ config S390
 	select HAVE_DYNAMIC_FTRACE_WITH_REGS
 	select HAVE_EBPF_JIT if HAVE_MARCH_Z196_FEATURES
 	select HAVE_EFFICIENT_UNALIGNED_ACCESS
+	select HAVE_FUTEX_ROBUST_UNLOCK
 	select HAVE_GENERIC_TIF_BITS
 	select HAVE_GUP_FAST
 	select HAVE_FENTRY
diff --git a/arch/s390/appldata/appldata_base.c b/arch/s390/appldata/appldata_base.c
index 836c51da18f4..f10b2f040cc2 100644
--- a/arch/s390/appldata/appldata_base.c
+++ b/arch/s390/appldata/appldata_base.c
@@ -360,21 +360,25 @@ appldata_generic_handler(const struct ctl_table *ctl, int write,
 			pr_err("Starting the data collection for %s "
 			       "failed with rc=%d\n", ops->name, rc);
 			module_put(ops->owner);
+			rc = -EIO;
 		} else
 			ops->active = 1;
 	} else if (!active && (ops->active == 1)) {
-		ops->active = 0;
 		rc = appldata_diag(ops->record_nr, APPLDATA_STOP_REC,
 				(unsigned long) ops->data, ops->size,
 				ops->mod_lvl);
-		if (rc != 0)
+		if (rc != 0) {
 			pr_err("Stopping the data collection for %s "
 			       "failed with rc=%d\n", ops->name, rc);
-		module_put(ops->owner);
+			rc = -EIO;
+		} else {
+			ops->active = 0;
+			module_put(ops->owner);
+		}
 	}
 	mutex_unlock(&appldata_ops_mutex);
 	module_put(ops->owner);
-	return 0;
+	return rc;
 }
 
 /*************************** /proc stuff <END> *******************************/
diff --git a/arch/s390/include/asm/futex_robust.h b/arch/s390/include/asm/futex_robust.h
new file mode 100644
index 000000000000..6a5fb2506192
--- /dev/null
+++ b/arch/s390/include/asm/futex_robust.h
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_S390_FUTEX_ROBUST_H
+#define _ASM_S390_FUTEX_ROBUST_H
+
+#include <asm/ptrace.h>
+
+#define VDSO_FUTEX_ROBUST_POP_GPR	4
+
+static __always_inline void __user *s390_futex_robust_unlock_get_pop(struct pt_regs *regs)
+{
+	/*
+	 * Within critical section of
+	 * __vdso_futex_robust_list64_try_unlock(__u32 *lock, __u32 tid, __u64 *pop)
+	 *
+	 * If condition code is 0, cmpxchg succeeded, and the pending
+	 * op pointer in gpr4 needs to be returned for clearing.
+	 */
+	if (psw_bits(regs->psw).cc != 0)
+		return NULL;
+	return (void __user *)regs->gprs[VDSO_FUTEX_ROBUST_POP_GPR];
+}
+
+#define arch_futex_robust_unlock_get_pop(regs)	s390_futex_robust_unlock_get_pop(regs)
+
+#endif /* _ASM_S390_FUTEX_ROBUST_H */
diff --git a/arch/s390/include/asm/ptrace.h b/arch/s390/include/asm/ptrace.h
index 3e56ef8c268b..d8f3a2dbe689 100644
--- a/arch/s390/include/asm/ptrace.h
+++ b/arch/s390/include/asm/ptrace.h
@@ -15,14 +15,14 @@
 
 #define PIF_SYSCALL			0	/* inside a system call */
 #define PIF_PSW_ADDR_ADJUSTED		1	/* psw address has been adjusted */
-#define PIF_SYSCALL_RET_SET		2	/* return value was set via ptrace */
+#define PIF_SYSCALL_ENTRY		2	/* ptrace gpr2 access redirected to syscall nr */
 #define PIF_FTRACE_FULL_REGS		3	/* all register contents valid (ftrace) */
 #define PIF_WARN_TRAP_SKIPPED		4	/* __WARN_trap() was skipped */
 #define PIF_PGM_CHECK			5	/* inside a program check */
 
 #define _PIF_SYSCALL			BIT(PIF_SYSCALL)
 #define _PIF_ADDR_PSW_ADJUSTED		BIT(PIF_PSW_ADDR_ADJUSTED)
-#define _PIF_SYSCALL_RET_SET		BIT(PIF_SYSCALL_RET_SET)
+#define _PIF_SYSCALL_ENTRY		BIT(PIF_SYSCALL_ENTRY)
 #define _PIF_FTRACE_FULL_REGS		BIT(PIF_FTRACE_FULL_REGS)
 #define _PIF_WARN_TRAP_SKIPPED		BIT(PIF_WARN_TRAP_SKIPPED)
 #define _PIF_PGM_CHECK			BIT(PIF_PGM_CHECK)
diff --git a/arch/s390/include/asm/syscall.h b/arch/s390/include/asm/syscall.h
index 5f310caad1fc..47f66202977f 100644
--- a/arch/s390/include/asm/syscall.h
+++ b/arch/s390/include/asm/syscall.h
@@ -59,7 +59,6 @@ static inline void syscall_set_return_value(struct task_struct *task,
 					    struct pt_regs *regs,
 					    int error, long val)
 {
-	set_pt_regs_flag(regs, PIF_SYSCALL_RET_SET);
 	regs->gprs[2] = error ? error : val;
 }
 
diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c
index 5202f9a2d939..577fb2c3d5a8 100644
--- a/arch/s390/kernel/ptrace.c
+++ b/arch/s390/kernel/ptrace.c
@@ -204,15 +204,24 @@ static unsigned long __peek_user(struct task_struct *child, addr_t addr)
 	addr_t offset, tmp;
 
 	if (addr < offsetof(struct user, regs.acrs)) {
-		/*
-		 * psw and gprs are stored on the stack
-		 */
-		tmp = *(addr_t *)((addr_t) &task_pt_regs(child)->psw + addr);
+		struct pt_regs *regs = task_pt_regs(child);
+
+		/* psw and gprs are on the stack. */
+		tmp = *(addr_t *)((addr_t)&regs->psw + addr);
 		if (addr == offsetof(struct user, regs.psw.mask)) {
 			/* Return a clean psw mask. */
 			tmp &= PSW_MASK_USER | PSW_MASK_RI;
 			tmp |= PSW_USER_BITS;
 		}
+		/*
+		 * During system call entry gpr2 contains the return
+		 * value. Redirect reads of gpr2 to int_code so that
+		 * readers see the system call number.
+		 */
+		if (addr == offsetof(struct user, regs.gprs[2])) {
+			if (test_pt_regs_flag(regs, PIF_SYSCALL_ENTRY))
+				tmp = syscall_get_nr(child, regs);
+		}
 
 	} else if (addr < offsetof(struct user, regs.orig_gpr2)) {
 		/*
@@ -347,8 +356,17 @@ static int __poke_user(struct task_struct *child, addr_t addr, addr_t data)
 		}
 
 		if (addr == offsetof(struct user, regs.gprs[2])) {
-			if (test_pt_regs_flag(regs, PIF_SYSCALL))
+			if (test_pt_regs_flag(regs, PIF_SYSCALL_ENTRY)) {
+				/*
+				 * System call entry path: only change system
+				 * call number in int_code, but leave gpr2
+				 * alone, which contains the return value.
+				 */
 				syscall_set_nr(child, regs, data);
+				return 0;
+			} else if (test_pt_regs_flag(regs, PIF_SYSCALL)) {
+				syscall_set_nr(child, regs, data);
+			}
 		}
 		*(addr_t *)((addr_t) &regs->psw + addr) = data;
 	} else if (addr < offsetof(struct user, regs.orig_gpr2)) {
diff --git a/arch/s390/kernel/syscall.c b/arch/s390/kernel/syscall.c
index bcc0b76bd868..a94bfcef2e8c 100644
--- a/arch/s390/kernel/syscall.c
+++ b/arch/s390/kernel/syscall.c
@@ -117,24 +117,22 @@ void noinstr __do_syscall(struct pt_regs *regs, unsigned long flags)
 		regs->int_code &= ~0xffffUL;
 		regs->int_code |= nr;
 	}
-	regs->gprs[2] = nr;
 	if (nr == __NR_restart_syscall && !(current->restart_block.arch_data & 1)) {
 		regs->psw.addr = current->restart_block.arch_data;
 		current->restart_block.arch_data = 1;
 	}
-
-	permit = syscall_enter_from_user_mode_work(regs, &nr);
-
 	/*
-	 * In the s390 ptrace ABI, both the syscall number and the return value
-	 * use gpr2. However, userspace puts the syscall number either in the
-	 * svc instruction itself, or uses gpr1. To make at least skipping syscalls
-	 * work, the ptrace code sets PIF_SYSCALL_RET_SET, which is checked here
-	 * and if set, the syscall will be skipped.
+	 * gpr2 is the return value register. Set it to -ENOSYS so that a
+	 * skipped system call gets a defined default return value.
+	 * Set PIF_SYSCALL_ENTRY so that ptrace reads and writes to gpr2
+	 * are redirected to the system call number in int_code.
 	 */
-	if (unlikely(test_and_clear_pt_regs_flag(regs, PIF_SYSCALL_RET_SET) || !permit))
+	syscall_set_return_value(current, regs, -ENOSYS, 0);
+	set_pt_regs_flag(regs, PIF_SYSCALL_ENTRY);
+	permit = syscall_enter_from_user_mode_work(regs, &nr);
+	clear_pt_regs_flag(regs, PIF_SYSCALL_ENTRY);
+	if (unlikely(!permit))
 		goto out;
-	regs->gprs[2] = -ENOSYS;
 	if (likely(nr < NR_syscalls)) {
 		nr = array_index_nospec(nr, NR_syscalls);
 		regs->gprs[2] = sys_call_table[nr](regs);
diff --git a/arch/s390/kernel/vdso.c b/arch/s390/kernel/vdso.c
index a905f989f910..efe492364a93 100644
--- a/arch/s390/kernel/vdso.c
+++ b/arch/s390/kernel/vdso.c
@@ -12,6 +12,7 @@
 #include <linux/init.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
+#include <linux/futex.h>
 #include <linux/mm.h>
 #include <linux/slab.h>
 #include <linux/smp.h>
@@ -19,14 +20,31 @@
 #include <linux/vdso_datastore.h>
 #include <vdso/datapage.h>
 #include <asm/alternative.h>
+#include <asm/vdso/vsyscall.h>
+#include <asm/vdso-symbols.h>
 #include <asm/vdso.h>
 
 extern char vdso_start[], vdso_end[];
 
+static inline void vdso_futex_robust_unlock_update_ips(void)
+{
+#ifdef CONFIG_FUTEX_ROBUST_UNLOCK
+	unsigned long vdso = current->mm->context.vdso_base;
+	struct futex_mm_data *fd = &current->mm->futex;
+
+	futex_reset_cs_ranges(fd);
+	futex_set_vdso_cs_range(fd, 0,
+				vdso + vdso_offset___futex_list64_try_unlock_cs_start,
+				vdso + vdso_offset___futex_list64_try_unlock_cs_end,
+				false);
+#endif
+}
+
 static int vdso_mremap(const struct vm_special_mapping *sm,
 		       struct vm_area_struct *vma)
 {
 	current->mm->context.vdso_base = vma->vm_start;
+	vdso_futex_robust_unlock_update_ips();
 	return 0;
 }
 
@@ -73,6 +91,7 @@ static int map_vdso(unsigned long addr, unsigned long vdso_mapping_len)
 		rc = PTR_ERR(vma);
 	} else {
 		current->mm->context.vdso_base = vdso_text_start;
+		vdso_futex_robust_unlock_update_ips();
 		rc = 0;
 	}
 out:
diff --git a/arch/s390/kernel/vdso/Makefile b/arch/s390/kernel/vdso/Makefile
index 54bcf2984ca1..35e4e3dd23c7 100644
--- a/arch/s390/kernel/vdso/Makefile
+++ b/arch/s390/kernel/vdso/Makefile
@@ -3,7 +3,8 @@
 
 # Include the generic Makefile to check the built vdso.
 include $(srctree)/lib/vdso/Makefile.include
-obj-vdso = vdso_user_wrapper.o note.o vgetrandom-chacha.o
+obj-vdso-y = vdso_user_wrapper.o note.o vgetrandom-chacha.o
+obj-vdso-$(CONFIG_FUTEX_ROBUST_UNLOCK) += vfutex.o
 obj-cvdso = vdso_generic.o getcpu.o vgetrandom.o
 VDSO_CFLAGS_REMOVE := -pg $(CC_FLAGS_FTRACE) $(CC_FLAGS_EXPOLINE)
 CFLAGS_REMOVE_getcpu.o = $(VDSO_CFLAGS_REMOVE)
@@ -16,8 +17,8 @@ endif
 
 # Build rules
 
-targets := $(obj-vdso) $(obj-cvdso) vdso.so vdso.so.dbg
-obj-vdso := $(addprefix $(obj)/, $(obj-vdso))
+targets := $(obj-vdso-y) $(obj-cvdso) vdso.so vdso.so.dbg
+obj-vdso := $(addprefix $(obj)/, $(obj-vdso-y))
 obj-cvdso := $(addprefix $(obj)/, $(obj-cvdso))
 
 KBUILD_AFLAGS_VDSO := $(KBUILD_AFLAGS) -DBUILD_VDSO
diff --git a/arch/s390/kernel/vdso/gen_vdso_offsets.sh b/arch/s390/kernel/vdso/gen_vdso_offsets.sh
index 359982fb002d..4039e6cfca51 100755
--- a/arch/s390/kernel/vdso/gen_vdso_offsets.sh
+++ b/arch/s390/kernel/vdso/gen_vdso_offsets.sh
@@ -12,4 +12,5 @@
 #
 
 LC_ALL=C
-sed -n 's/\([0-9a-f]*\) . __kernel_\(.*\)/\#define vdso_offset_\2\t0x\1/p'
+sed -n -e 's/\([0-9a-f]*\) . __kernel_\(.*\)/\#define vdso_offset_\2\t0x\1/p' \
+       -e 's/\([0-9a-f]*\) . \(__futex_.*\)/\#define vdso_offset_\2\t0x\1/p'
diff --git a/arch/s390/kernel/vdso/vdso.lds.S b/arch/s390/kernel/vdso/vdso.lds.S
index 6bcdb64315f2..1172f05f609d 100644
--- a/arch/s390/kernel/vdso/vdso.lds.S
+++ b/arch/s390/kernel/vdso/vdso.lds.S
@@ -110,6 +110,9 @@ VERSION
 		__kernel_rt_sigreturn;
 		__kernel_sigreturn;
 		__kernel_getrandom;
+#ifdef CONFIG_FUTEX_ROBUST_UNLOCK
+		__vdso_futex_robust_list64_try_unlock;
+#endif
 	local: *;
 	};
 }
diff --git a/arch/s390/kernel/vdso/vfutex.S b/arch/s390/kernel/vdso/vfutex.S
new file mode 100644
index 000000000000..8a25bd4ecea4
--- /dev/null
+++ b/arch/s390/kernel/vdso/vfutex.S
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+#include <linux/linkage.h>
+#include <asm/dwarf.h>
+
+/*
+ * __u32 __vdso_futex_robust_list64_try_unlock(__u32 *lock, __u32 tid, __u64 *pop)
+ *
+ * Returns tid on success, or lock value on failure.
+ */
+SYM_FUNC_START(__vdso_futex_robust_list64_try_unlock)
+	CFI_STARTPROC
+	lghi	%r1,0
+	cs	%r3,%r1,0(%r2)
+__futex_list64_try_unlock_cs_start:
+	jnz	__futex_list64_try_unlock_cs_end
+__futex_list64_try_unlock_cs_success:
+	mvghi	0(%r4),0
+__futex_list64_try_unlock_cs_end:
+	llgfr	%r2,%r3
+	br	%r14
+	CFI_ENDPROC
+SYM_FUNC_END(__vdso_futex_robust_list64_try_unlock)