Merge tag 's390-5.5-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

Pull more s390 updates from Vasily Gorbik: - Make stack unwinder reliable and suitable for livepatching. Add unwinder testing module. - Fixes for CALL_ON_STACK helper used for stack switching. - Fix unwinding from bpf code. - Fix getcpu and remove compat support in vdso code. - Fix address space control registers initialization. - Save KASLR offset for early dumps. - Handle new FILTERED_BY_HYPERVISOR reply code in crypto code. - Minor perf code cleanup and potential memory leak fix. - Add couple of error messages for corner cases during PCI device creation. * tag 's390-5.5-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (33 commits) s390: remove compat vdso code s390/livepatch: Implement reliable stack tracing for the consistency model s390/unwind: add stack pointer alignment sanity checks s390/unwind: filter out unreliable bogus %r14 s390/unwind: start unwinding from reliable state s390/test_unwind: add program check context tests s390/test_unwind: add irq context tests s390/test_unwind: print verbose unwinding results s390/test_unwind: add CALL_ON_STACK tests s390: fix register clobbering in CALL_ON_STACK s390/test_unwind: require that unwinding ended successfully s390/unwind: add a test for the internal API s390/unwind: always inline get_stack_pointer s390/pci: add error message on device number limit s390/pci: add error message for UID collision s390/cpum_sf: Check for SDBT and SDB consistency s390/cpum_sf: Use TEAR_REG macro consistantly s390/cpum_sf: Remove unnecessary check for pending SDBs s390/cpum_sf: Replace function name in debug statements s390/kaslr: store KASLR offset for early dumps ...
2026-03-06 15:25:10 -08:00 · 2019-12-03 12:50:00 -08:00
parent 4d7048f551 2115fbf721
commit 01d1dff646
35 changed files with 627 additions and 768 deletions
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -170,6 +170,7 @@ config S390
 	select HAVE_PERF_EVENTS
 	select HAVE_RCU_TABLE_FREE
 	select HAVE_REGS_AND_STACK_ACCESS_API
+	select HAVE_RELIABLE_STACKTRACE
 	select HAVE_RSEQ
 	select HAVE_SYSCALL_TRACEPOINTS
 	select HAVE_VIRT_CPU_ACCOUNTING
@@ -426,9 +427,6 @@ config COMPAT
 	  (and some other stuff like libraries and such) is needed for
 	  executing 31 bit applications.  It is safe to say "Y".

-config COMPAT_VDSO
-	def_bool COMPAT && !CC_IS_CLANG
-
 config SYSVIPC_COMPAT
 	def_bool y if COMPAT && SYSVIPC

@@ -1018,3 +1016,17 @@ config S390_GUEST
 	  the KVM hypervisor.

 endmenu
+
+menu "Selftests"
+
+config S390_UNWIND_SELFTEST
+	def_tristate n
+	prompt "Test unwind functions"
+	help
+	  This option enables s390 specific stack unwinder testing kernel
+	  module. This option is not useful for distributions or general
+	  kernels, but only for kernel developers working on architecture code.
+
+	  Say N if you are unsure.
+
+endmenu
--- a/arch/s390/Makefile
+++ b/arch/s390/Makefile
@@ -157,7 +157,6 @@ zfcpdump:

 vdso_install:
 	$(Q)$(MAKE) $(build)=arch/$(ARCH)/kernel/vdso64 $@
-	$(Q)$(MAKE) $(build)=arch/$(ARCH)/kernel/vdso32 $@

 archclean:
 	$(Q)$(MAKE) $(clean)=$(boot)
--- a/arch/s390/boot/startup.c
+++ b/arch/s390/boot/startup.c
@@ -170,6 +170,11 @@ void startup_kernel(void)
 		handle_relocs(__kaslr_offset);

 	if (__kaslr_offset) {
+		/*
+		 * Save KASLR offset for early dumps, before vmcore_info is set.
+		 * Mark as uneven to distinguish from real vmcore_info pointer.
+		 */
+		S390_lowcore.vmcore_info = __kaslr_offset | 0x1UL;
 		/* Clear non-relocated kernel */
 		if (IS_ENABLED(CONFIG_KERNEL_UNCOMPRESSED))
 			memset(img, 0, vmlinux.image_size);
--- a/arch/s390/include/asm/cpu_mf.h
+++ b/arch/s390/include/asm/cpu_mf.h
@@ -313,7 +313,7 @@ static inline unsigned long *trailer_entry_ptr(unsigned long v)
 	return (unsigned long *) ret;
 }

-/* Return if the entry in the sample data block table (sdbt)
+/* Return true if the entry in the sample data block table (sdbt)
 * is a link to the next sdbt */
 static inline int is_link_entry(unsigned long *s)
 {
--- a/arch/s390/include/asm/perf_event.h
+++ b/arch/s390/include/asm/perf_event.h
@@ -12,6 +12,7 @@

 #include <linux/perf_event.h>
 #include <linux/device.h>
+#include <asm/stacktrace.h>

 /* Per-CPU flags for PMU states */
 #define PMU_F_RESERVED			0x1000
@@ -73,4 +74,10 @@ struct perf_sf_sde_regs {
 #define SDB_FULL_BLOCKS(hwc)	(SAMPL_FLAGS(hwc) & PERF_CPUM_SF_FULL_BLOCKS)
 #define SAMPLE_FREQ_MODE(hwc)	(SAMPL_FLAGS(hwc) & PERF_CPUM_SF_FREQ_MODE)

+#define perf_arch_fetch_caller_regs(regs, __ip) do {			\
+	(regs)->psw.addr = (__ip);					\
+	(regs)->gprs[15] = (unsigned long)__builtin_frame_address(0) -	\
+		offsetof(struct stack_frame, back_chain);		\
+} while (0)
+
 #endif /* _ASM_S390_PERF_EVENT_H */
--- a/arch/s390/include/asm/processor.h
+++ b/arch/s390/include/asm/processor.h
@@ -310,7 +310,7 @@ void enabled_wait(void);
 /*
 * Function to drop a processor into disabled wait state
 */
-static inline void __noreturn disabled_wait(void)
+static __always_inline void __noreturn disabled_wait(void)
 {
 	psw_t psw;

--- a/arch/s390/include/asm/stacktrace.h
+++ b/arch/s390/include/asm/stacktrace.h
@@ -33,12 +33,12 @@ static inline bool on_stack(struct stack_info *info,
 	return addr >= info->begin && addr + len <= info->end;
 }

-static inline unsigned long get_stack_pointer(struct task_struct *task,
-					      struct pt_regs *regs)
+static __always_inline unsigned long get_stack_pointer(struct task_struct *task,
+						       struct pt_regs *regs)
 {
 	if (regs)
 		return (unsigned long) kernel_stack_pointer(regs);
-	if (!task || task == current)
+	if (task == current)
 		return current_stack_pointer();
 	return (unsigned long) task->thread.ksp;
 }
@@ -62,6 +62,17 @@ struct stack_frame {
 };
 #endif

+/*
+ * Unlike current_stack_pointer() which simply returns current value of %r15
+ * current_frame_address() returns function stack frame address, which matches
+ * %r15 upon function invocation. It may differ from %r15 later if function
+ * allocates stack for local variables or new stack frame to call other
+ * functions.
+ */
+#define current_frame_address()						\
+	((unsigned long)__builtin_frame_address(0) -			\
+	 offsetof(struct stack_frame, back_chain))
+
 #define CALL_ARGS_0()							\
 	register unsigned long r2 asm("2")
 #define CALL_ARGS_1(arg1)						\
@@ -95,20 +106,33 @@ struct stack_frame {

 #define CALL_ON_STACK(fn, stack, nr, args...)				\
 ({									\
+	unsigned long frame = current_frame_address();			\
 	CALL_ARGS_##nr(args);						\
 	unsigned long prev;						\
 									\
 	asm volatile(							\
 		"	la	%[_prev],0(15)\n"			\
-		"	la	15,0(%[_stack])\n"			\
-		"	stg	%[_prev],%[_bc](15)\n"			\
+		"	lg	15,%[_stack]\n"				\
+		"	stg	%[_frame],%[_bc](15)\n"			\
 		"	brasl	14,%[_fn]\n"				\
 		"	la	15,0(%[_prev])\n"			\
 		: [_prev] "=&a" (prev), CALL_FMT_##nr			\
-		  [_stack] "a" (stack),					\
+		  [_stack] "R" (stack),					\
 		  [_bc] "i" (offsetof(struct stack_frame, back_chain)),	\
+		  [_frame] "d" (frame),					\
 		  [_fn] "X" (fn) : CALL_CLOBBER_##nr);			\
 	r2;								\
 })

+#define CALL_ON_STACK_NORETURN(fn, stack)				\
+({									\
+	asm volatile(							\
+		"	la	15,0(%[_stack])\n"			\
+		"	xc	%[_bc](8,15),%[_bc](15)\n"		\
+		"	brasl	14,%[_fn]\n"				\
+		::[_bc] "i" (offsetof(struct stack_frame, back_chain)),	\
+		  [_stack] "a" (stack), [_fn] "X" (fn));		\
+	BUG();								\
+})
+
 #endif /* _ASM_S390_STACKTRACE_H */
--- a/arch/s390/include/asm/unwind.h
+++ b/arch/s390/include/asm/unwind.h
@@ -35,7 +35,6 @@ struct unwind_state {
 	struct task_struct *task;
 	struct pt_regs *regs;
 	unsigned long sp, ip;
-	bool reuse_sp;
 	int graph_idx;
 	bool reliable;
 	bool error;
@@ -59,10 +58,11 @@ static inline bool unwind_error(struct unwind_state *state)
 static inline void unwind_start(struct unwind_state *state,
 				struct task_struct *task,
 				struct pt_regs *regs,
-				unsigned long sp)
+				unsigned long first_frame)
 {
-	sp = sp ? : get_stack_pointer(task, regs);
-	__unwind_start(state, task, regs, sp);
+	task = task ?: current;
+	first_frame = first_frame ?: get_stack_pointer(task, regs);
+	__unwind_start(state, task, regs, first_frame);
 }

 static inline struct pt_regs *unwind_get_entry_regs(struct unwind_state *state)
--- a/arch/s390/include/asm/vdso.h
+++ b/arch/s390/include/asm/vdso.h
@@ -41,8 +41,17 @@ struct vdso_data {
 struct vdso_per_cpu_data {
 	__u64 ectg_timer_base;
 	__u64 ectg_user_time;
-	__u32 cpu_nr;
-	__u32 node_id;
+	/*
+	 * Note: node_id and cpu_nr must be at adjacent memory locations.
+	 * VDSO userspace must read both values with a single instruction.
+	 */
+	union {
+		__u64 getcpu_val;
+		struct {
+			__u32 node_id;
+			__u32 cpu_nr;
+		};
+	};
 };

 extern struct vdso_data *vdso_data;
--- a/arch/s390/kernel/Makefile
+++ b/arch/s390/kernel/Makefile
@@ -81,4 +81,3 @@ obj-$(CONFIG_TRACEPOINTS)	+= trace.o

 # vdso
 obj-y				+= vdso64/
-obj-$(CONFIG_COMPAT_VDSO)	+= vdso32/
--- a/arch/s390/kernel/asm-offsets.c
+++ b/arch/s390/kernel/asm-offsets.c
@@ -78,8 +78,7 @@ int main(void)
 	OFFSET(__VDSO_TS_END, vdso_data, ts_end);
 	OFFSET(__VDSO_ECTG_BASE, vdso_per_cpu_data, ectg_timer_base);
 	OFFSET(__VDSO_ECTG_USER, vdso_per_cpu_data, ectg_user_time);
-	OFFSET(__VDSO_CPU_NR, vdso_per_cpu_data, cpu_nr);
-	OFFSET(__VDSO_NODE_ID, vdso_per_cpu_data, node_id);
+	OFFSET(__VDSO_GETCPU_VAL, vdso_per_cpu_data, getcpu_val);
 	BLANK();
 	/* constants used by the vdso */
 	DEFINE(__CLOCK_REALTIME, CLOCK_REALTIME);
--- a/arch/s390/kernel/dumpstack.c
+++ b/arch/s390/kernel/dumpstack.c
@@ -38,6 +38,7 @@ const char *stack_type_name(enum stack_type type)
 		return "unknown";
 	}
 }
+EXPORT_SYMBOL_GPL(stack_type_name);

 static inline bool in_stack(unsigned long sp, struct stack_info *info,
 			    enum stack_type type, unsigned long low,
@@ -93,7 +94,9 @@ int get_stack_info(unsigned long sp, struct task_struct *task,
 	if (!sp)
 		goto unknown;

-	task = task ? : current;
+	/* Sanity check: ABI requires SP to be aligned 8 bytes. */
+	if (sp & 0x7)
+		goto unknown;

 	/* Check per-task stack */
 	if (in_task_stack(sp, task, info))
@@ -128,8 +131,6 @@ void show_stack(struct task_struct *task, unsigned long *stack)
 	struct unwind_state state;

 	printk("Call Trace:\n");
-	if (!task)
-		task = current;
 	unwind_for_each_frame(&state, task, NULL, (unsigned long) stack)
 		printk(state.reliable ? " [<%016lx>] %pSR \n" :
 					"([<%016lx>] %pSR)\n",
--- a/arch/s390/kernel/head64.S
+++ b/arch/s390/kernel/head64.S
@@ -31,7 +31,7 @@ ENTRY(startup_continue)
 #
 	larl	%r14,init_task
 	stg	%r14,__LC_CURRENT
-	larl	%r15,init_thread_union+THREAD_SIZE-STACK_FRAME_OVERHEAD
+	larl	%r15,init_thread_union+THREAD_SIZE-STACK_FRAME_OVERHEAD-__PT_SIZE
 #ifdef CONFIG_KASAN
 	brasl	%r14,kasan_early_init
 #endif
--- a/arch/s390/kernel/machine_kexec.c
+++ b/arch/s390/kernel/machine_kexec.c
@@ -164,7 +164,9 @@ static bool kdump_csum_valid(struct kimage *image)
 #ifdef CONFIG_CRASH_DUMP
 	int rc;

+	preempt_disable();
 	rc = CALL_ON_STACK(do_start_kdump, S390_lowcore.nodat_stack, 1, image);
+	preempt_enable();
 	return rc == 0;
 #else
 	return false;
@@ -254,10 +256,10 @@ void arch_crash_save_vmcoreinfo(void)
 	VMCOREINFO_SYMBOL(lowcore_ptr);
 	VMCOREINFO_SYMBOL(high_memory);
 	VMCOREINFO_LENGTH(lowcore_ptr, NR_CPUS);
-	mem_assign_absolute(S390_lowcore.vmcore_info, paddr_vmcoreinfo_note());
 	vmcoreinfo_append_str("SDMA=%lx\n", __sdma);
 	vmcoreinfo_append_str("EDMA=%lx\n", __edma);
 	vmcoreinfo_append_str("KERNELOFFSET=%lx\n", kaslr_offset());
+	mem_assign_absolute(S390_lowcore.vmcore_info, paddr_vmcoreinfo_note());
 }

 void machine_shutdown(void)
--- a/arch/s390/kernel/perf_cpum_sf.c
+++ b/arch/s390/kernel/perf_cpum_sf.c
@@ -156,8 +156,8 @@ static void free_sampling_buffer(struct sf_buffer *sfb)
 		}
 	}

-	debug_sprintf_event(sfdbg, 5, "%s freed sdbt %p\n", __func__,
-			    sfb->sdbt);
+	debug_sprintf_event(sfdbg, 5, "%s: freed sdbt %#lx\n", __func__,
+			    (unsigned long)sfb->sdbt);
 	memset(sfb, 0, sizeof(*sfb));
 }

@@ -193,7 +193,7 @@ static int realloc_sampling_buffer(struct sf_buffer *sfb,
 				   unsigned long num_sdb, gfp_t gfp_flags)
 {
 	int i, rc;
-	unsigned long *new, *tail;
+	unsigned long *new, *tail, *tail_prev = NULL;

 	if (!sfb->sdbt || !sfb->tail)
 		return -EINVAL;
@@ -213,9 +213,10 @@ static int realloc_sampling_buffer(struct sf_buffer *sfb,
 	 */
 	if (sfb->sdbt != get_next_sdbt(tail)) {
 		debug_sprintf_event(sfdbg, 3, "%s: "
-				    "sampling buffer is not linked: origin %p"
-				    " tail %p\n", __func__,
-				    (void *)sfb->sdbt, (void *)tail);
+				    "sampling buffer is not linked: origin %#lx"
+				    " tail %#lx\n", __func__,
+				    (unsigned long)sfb->sdbt,
+				    (unsigned long)tail);
 		return -EINVAL;
 	}

@@ -232,6 +233,7 @@ static int realloc_sampling_buffer(struct sf_buffer *sfb,
 			sfb->num_sdbt++;
 			/* Link current page to tail of chain */
 			*tail = (unsigned long)(void *) new + 1;
+			tail_prev = tail;
 			tail = new;
 		}

@@ -241,18 +243,30 @@ static int realloc_sampling_buffer(struct sf_buffer *sfb,
 		 * issue, a new realloc call (if required) might succeed.
 		 */
 		rc = alloc_sample_data_block(tail, gfp_flags);
-		if (rc)
+		if (rc) {
+			/* Undo last SDBT. An SDBT with no SDB at its first
+			 * entry but with an SDBT entry instead can not be
+			 * handled by the interrupt handler code.
+			 * Avoid this situation.
+			 */
+			if (tail_prev) {
+				sfb->num_sdbt--;
+				free_page((unsigned long) new);
+				tail = tail_prev;
+			}
 			break;
+		}
 		sfb->num_sdb++;
 		tail++;
+		tail_prev = new = NULL;	/* Allocated at least one SBD */
 	}

 	/* Link sampling buffer to its origin */
 	*tail = (unsigned long) sfb->sdbt + 1;
 	sfb->tail = tail;

-	debug_sprintf_event(sfdbg, 4, "realloc_sampling_buffer: new buffer"
-			    " settings: sdbt %lu sdb %lu\n",
+	debug_sprintf_event(sfdbg, 4, "%s: new buffer"
+			    " settings: sdbt %lu sdb %lu\n", __func__,
 			    sfb->num_sdbt, sfb->num_sdb);
 	return rc;
 }
@@ -292,12 +306,13 @@ static int alloc_sampling_buffer(struct sf_buffer *sfb, unsigned long num_sdb)
 	rc = realloc_sampling_buffer(sfb, num_sdb, GFP_KERNEL);
 	if (rc) {
 		free_sampling_buffer(sfb);
-		debug_sprintf_event(sfdbg, 4, "alloc_sampling_buffer: "
-			"realloc_sampling_buffer failed with rc %i\n", rc);
+		debug_sprintf_event(sfdbg, 4, "%s: "
+			"realloc_sampling_buffer failed with rc %i\n",
+			__func__, rc);
 	} else
 		debug_sprintf_event(sfdbg, 4,
-			"alloc_sampling_buffer: tear %p dear %p\n",
-			sfb->sdbt, (void *)*sfb->sdbt);
+			"%s: tear %#lx dear %#lx\n", __func__,
+			(unsigned long)sfb->sdbt, (unsigned long)*sfb->sdbt);
 	return rc;
 }

@@ -465,8 +480,8 @@ static void sfb_account_overflows(struct cpu_hw_sf *cpuhw,
 	if (num)
 		sfb_account_allocs(num, hwc);

-	debug_sprintf_event(sfdbg, 5, "sfb: overflow: overflow %llu ratio %lu"
-			    " num %lu\n", OVERFLOW_REG(hwc), ratio, num);
+	debug_sprintf_event(sfdbg, 5, "%s: overflow %llu ratio %lu num %lu\n",
+			    __func__, OVERFLOW_REG(hwc), ratio, num);
 	OVERFLOW_REG(hwc) = 0;
 }

@@ -504,13 +519,13 @@ static void extend_sampling_buffer(struct sf_buffer *sfb,
 	 */
 	rc = realloc_sampling_buffer(sfb, num, GFP_ATOMIC);
 	if (rc)
-		debug_sprintf_event(sfdbg, 5, "sfb: extend: realloc "
-				    "failed with rc %i\n", rc);
+		debug_sprintf_event(sfdbg, 5, "%s: realloc failed with rc %i\n",
+				    __func__, rc);

 	if (sfb_has_pending_allocs(sfb, hwc))
-		debug_sprintf_event(sfdbg, 5, "sfb: extend: "
+		debug_sprintf_event(sfdbg, 5, "%s: "
 				    "req %lu alloc %lu remaining %lu\n",
-				    num, sfb->num_sdb - num_old,
+				    __func__, num, sfb->num_sdb - num_old,
 				    sfb_pending_allocs(sfb, hwc));
 }

@@ -600,13 +615,6 @@ static void hw_init_period(struct hw_perf_event *hwc, u64 period)
 	local64_set(&hwc->period_left, hwc->sample_period);
 }

-static void hw_reset_registers(struct hw_perf_event *hwc,
-			       unsigned long *sdbt_origin)
-{
-	/* (Re)set to first sample-data-block-table */
-	TEAR_REG(hwc) = (unsigned long) sdbt_origin;
-}
-
 static unsigned long hw_limit_rate(const struct hws_qsi_info_block *si,
 				   unsigned long rate)
 {
@@ -698,9 +706,9 @@ static unsigned long getrate(bool freq, unsigned long sample,
 		 */
 		if (sample_rate_to_freq(si, rate) >
 		    sysctl_perf_event_sample_rate) {
-			debug_sprintf_event(sfdbg, 1,
+			debug_sprintf_event(sfdbg, 1, "%s: "
 					    "Sampling rate exceeds maximum "
-					    "perf sample rate\n");
+					    "perf sample rate\n", __func__);
 			rate = 0;
 		}
 	}
@@ -745,10 +753,9 @@ static int __hw_perf_event_init_rate(struct perf_event *event,
 	attr->sample_period = rate;
 	SAMPL_RATE(hwc) = rate;
 	hw_init_period(hwc, SAMPL_RATE(hwc));
-	debug_sprintf_event(sfdbg, 4, "__hw_perf_event_init_rate:"
-			    "cpu:%d period:%#llx freq:%d,%#lx\n", event->cpu,
-			    event->attr.sample_period, event->attr.freq,
-			    SAMPLE_FREQ_MODE(hwc));
+	debug_sprintf_event(sfdbg, 4, "%s: cpu %d period %#llx freq %d,%#lx\n",
+			    __func__, event->cpu, event->attr.sample_period,
+			    event->attr.freq, SAMPLE_FREQ_MODE(hwc));
 	return 0;
 }

@@ -951,8 +958,7 @@ static void cpumsf_pmu_enable(struct pmu *pmu)
 			 * buffer extents
 			 */
 			sfb_account_overflows(cpuhw, hwc);
-			if (sfb_has_pending_allocs(&cpuhw->sfb, hwc))
-				extend_sampling_buffer(&cpuhw->sfb, hwc);
+			extend_sampling_buffer(&cpuhw->sfb, hwc);
 		}
 		/* Rate may be adjusted with ioctl() */
 		cpuhw->lsctl.interval = SAMPL_RATE(&cpuhw->event->hw);
@@ -973,12 +979,11 @@ static void cpumsf_pmu_enable(struct pmu *pmu)
 	/* Load current program parameter */
 	lpp(&S390_lowcore.lpp);

-	debug_sprintf_event(sfdbg, 6, "pmu_enable: es %i cs %i ed %i cd %i "
-			    "interval %#lx tear %p dear %p\n",
+	debug_sprintf_event(sfdbg, 6, "%s: es %i cs %i ed %i cd %i "
+			    "interval %#lx tear %#lx dear %#lx\n", __func__,
 			    cpuhw->lsctl.es, cpuhw->lsctl.cs, cpuhw->lsctl.ed,
 			    cpuhw->lsctl.cd, cpuhw->lsctl.interval,
-			    (void *) cpuhw->lsctl.tear,
-			    (void *) cpuhw->lsctl.dear);
+			    cpuhw->lsctl.tear, cpuhw->lsctl.dear);
 }

 static void cpumsf_pmu_disable(struct pmu *pmu)
@@ -1019,8 +1024,8 @@ static void cpumsf_pmu_disable(struct pmu *pmu)
 			cpuhw->lsctl.dear = si.dear;
 		}
 	} else
-		debug_sprintf_event(sfdbg, 3, "cpumsf_pmu_disable: "
-				    "qsi() failed with err %i\n", err);
+		debug_sprintf_event(sfdbg, 3, "%s: qsi() failed with err %i\n",
+				    __func__, err);

 	cpuhw->flags &= ~PMU_F_ENABLED;
 }
@@ -1265,9 +1270,9 @@ static void hw_perf_event_update(struct perf_event *event, int flush_all)
 			sampl_overflow += te->overflow;

 		/* Timestamps are valid for full sample-data-blocks only */
-		debug_sprintf_event(sfdbg, 6, "%s: sdbt %p "
+		debug_sprintf_event(sfdbg, 6, "%s: sdbt %#lx "
 				    "overflow %llu timestamp %#llx\n",
-				    __func__, sdbt, te->overflow,
+				    __func__, (unsigned long)sdbt, te->overflow,
 				    (te->f) ? trailer_timestamp(te) : 0ULL);

 		/* Collect all samples from a single sample-data-block and
@@ -1312,8 +1317,10 @@ static void hw_perf_event_update(struct perf_event *event, int flush_all)
 						 sampl_overflow, 1 + num_sdb);
 	if (sampl_overflow || event_overflow)
 		debug_sprintf_event(sfdbg, 4, "%s: "
-				    "overflow stats: sample %llu event %llu\n",
-				    __func__, sampl_overflow, event_overflow);
+				    "overflows: sample %llu event %llu"
+				    " total %llu num_sdb %llu\n",
+				    __func__, sampl_overflow, event_overflow,
+				    OVERFLOW_REG(hwc), num_sdb);
 }

 #define AUX_SDB_INDEX(aux, i) ((i) % aux->sfb.num_sdb)
@@ -1424,10 +1431,10 @@ static int aux_output_begin(struct perf_output_handle *handle,
 	cpuhw->lsctl.tear = base + offset * sizeof(unsigned long);
 	cpuhw->lsctl.dear = aux->sdb_index[head];

-	debug_sprintf_event(sfdbg, 6, "aux_output_begin: "
+	debug_sprintf_event(sfdbg, 6, "%s: "
 			    "head->alert_mark->empty_mark (num_alert, range)"
 			    "[%#lx -> %#lx -> %#lx] (%#lx, %#lx) "
-			    "tear index %#lx, tear %#lx dear %#lx\n",
+			    "tear index %#lx, tear %#lx dear %#lx\n", __func__,
 			    aux->head, aux->alert_mark, aux->empty_mark,
 			    AUX_SDB_NUM_ALERT(aux), range,
 			    head / CPUM_SF_SDB_PER_TABLE,
@@ -1571,7 +1578,9 @@ static void hw_collect_aux(struct cpu_hw_sf *cpuhw)
 			pr_err("The AUX buffer with %lu pages for the "
 			       "diagnostic-sampling mode is full\n",
 				num_sdb);
-			debug_sprintf_event(sfdbg, 1, "AUX buffer used up\n");
+			debug_sprintf_event(sfdbg, 1,
+					    "%s: AUX buffer used up\n",
+					    __func__);
 			break;
 		}
 		if (WARN_ON_ONCE(!aux))
@@ -1594,23 +1603,25 @@ static void hw_collect_aux(struct cpu_hw_sf *cpuhw)
 			perf_aux_output_end(&cpuhw->handle, size);
 			pr_err("Sample data caused the AUX buffer with %lu "
 			       "pages to overflow\n", num_sdb);
-			debug_sprintf_event(sfdbg, 1, "head %#lx range %#lx "
-					    "overflow %#llx\n",
+			debug_sprintf_event(sfdbg, 1, "%s: head %#lx range %#lx "
+					    "overflow %#llx\n", __func__,
 					    aux->head, range, overflow);
 		} else {
 			size = AUX_SDB_NUM_ALERT(aux) << PAGE_SHIFT;
 			perf_aux_output_end(&cpuhw->handle, size);
-			debug_sprintf_event(sfdbg, 6, "head %#lx alert %#lx "
+			debug_sprintf_event(sfdbg, 6, "%s: head %#lx alert %#lx "
 					    "already full, try another\n",
+					    __func__,
 					    aux->head, aux->alert_mark);
 		}
 	}

 	if (done)
-		debug_sprintf_event(sfdbg, 6, "aux_reset_buffer: "
+		debug_sprintf_event(sfdbg, 6, "%s: aux_reset_buffer "
 				    "[%#lx -> %#lx -> %#lx] (%#lx, %#lx)\n",
-				    aux->head, aux->alert_mark, aux->empty_mark,
-				    AUX_SDB_NUM_ALERT(aux), range);
+				    __func__, aux->head, aux->alert_mark,
+				    aux->empty_mark, AUX_SDB_NUM_ALERT(aux),
+				    range);
 }

 /*
@@ -1633,8 +1644,8 @@ static void aux_buffer_free(void *data)
 	kfree(aux->sdb_index);
 	kfree(aux);

-	debug_sprintf_event(sfdbg, 4, "aux_buffer_free: free "
-			    "%lu SDBTs\n", num_sdbt);
+	debug_sprintf_event(sfdbg, 4, "%s: free "
+			    "%lu SDBTs\n", __func__, num_sdbt);
 }

 static void aux_sdb_init(unsigned long sdb)
@@ -1742,9 +1753,8 @@ static void *aux_buffer_setup(struct perf_event *event, void **pages,
 	 */
 	aux->empty_mark = sfb->num_sdb - 1;

-	debug_sprintf_event(sfdbg, 4, "aux_buffer_setup: setup %lu SDBTs"
-			    " and %lu SDBs\n",
-			    sfb->num_sdbt, sfb->num_sdb);
+	debug_sprintf_event(sfdbg, 4, "%s: setup %lu SDBTs and %lu SDBs\n",
+			    __func__, sfb->num_sdbt, sfb->num_sdb);

 	return aux;

@@ -1797,9 +1807,9 @@ static int cpumsf_pmu_check_period(struct perf_event *event, u64 value)
 	event->attr.sample_period = rate;
 	SAMPL_RATE(&event->hw) = rate;
 	hw_init_period(&event->hw, SAMPL_RATE(&event->hw));
-	debug_sprintf_event(sfdbg, 4, "cpumsf_pmu_check_period:"
-			    "cpu:%d value:%#llx period:%#llx freq:%d\n",
-			    event->cpu, value,
+	debug_sprintf_event(sfdbg, 4, "%s:"
+			    " cpu %d value %#llx period %#llx freq %d\n",
+			    __func__, event->cpu, value,
 			    event->attr.sample_period, do_freq);
 	return 0;
 }
@@ -1875,7 +1885,7 @@ static int cpumsf_pmu_add(struct perf_event *event, int flags)
 	if (!SAMPL_DIAG_MODE(&event->hw)) {
 		cpuhw->lsctl.tear = (unsigned long) cpuhw->sfb.sdbt;
 		cpuhw->lsctl.dear = *(unsigned long *) cpuhw->sfb.sdbt;
-		hw_reset_registers(&event->hw, cpuhw->sfb.sdbt);
+		TEAR_REG(&event->hw) = (unsigned long) cpuhw->sfb.sdbt;
 	}

 	/* Ensure sampling functions are in the disabled state.  If disabled,
@@ -2030,7 +2040,7 @@ static void cpumf_measurement_alert(struct ext_code ext_code,

 	/* Report measurement alerts only for non-PRA codes */
 	if (alert != CPU_MF_INT_SF_PRA)
-		debug_sprintf_event(sfdbg, 6, "measurement alert: %#x\n",
+		debug_sprintf_event(sfdbg, 6, "%s: alert %#x\n", __func__,
 				    alert);

 	/* Sampling authorization change request */
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -355,7 +355,6 @@ early_initcall(async_stack_realloc);

 void __init arch_call_rest_init(void)
 {
-	struct stack_frame *frame;
 	unsigned long stack;

 	stack = stack_alloc();
@@ -368,13 +367,7 @@ void __init arch_call_rest_init(void)
 	set_task_stack_end_magic(current);
 	stack += STACK_INIT_OFFSET;
 	S390_lowcore.kernel_stack = stack;
-	frame = (struct stack_frame *) stack;
-	memset(frame, 0, sizeof(*frame));
-	/* Branch to rest_init on the new stack, never returns */
-	asm volatile(
-		"	la	15,0(%[_frame])\n"
-		"	jg	rest_init\n"
-		: : [_frame] "a" (frame));
+	CALL_ON_STACK_NORETURN(rest_init, stack);
 }

 static void __init setup_lowcore_dat_off(void)
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -262,10 +262,13 @@ static void pcpu_prepare_secondary(struct pcpu *pcpu, int cpu)
 	lc->spinlock_index = 0;
 	lc->percpu_offset = __per_cpu_offset[cpu];
 	lc->kernel_asce = S390_lowcore.kernel_asce;
+	lc->user_asce = S390_lowcore.kernel_asce;
 	lc->machine_flags = S390_lowcore.machine_flags;
 	lc->user_timer = lc->system_timer =
 		lc->steal_timer = lc->avg_steal_timer = 0;
 	__ctl_store(lc->cregs_save_area, 0, 15);
+	lc->cregs_save_area[1] = lc->kernel_asce;
+	lc->cregs_save_area[7] = lc->vdso_asce;
 	save_access_regs((unsigned int *) lc->access_regs_save_area);
 	memcpy(lc->stfle_fac_list, S390_lowcore.stfle_fac_list,
 	       sizeof(lc->stfle_fac_list));
@@ -844,6 +847,8 @@ static void smp_init_secondary(void)

 	S390_lowcore.last_update_clock = get_tod_clock();
 	restore_access_regs(S390_lowcore.access_regs_save_area);
+	set_cpu_flag(CIF_ASCE_PRIMARY);
+	set_cpu_flag(CIF_ASCE_SECONDARY);
 	cpu_init();
 	preempt_disable();
 	init_cpu_timer();
@@ -871,7 +876,7 @@ static void __no_sanitize_address smp_start_secondary(void *cpuvoid)
 	S390_lowcore.restart_source = -1UL;
 	__ctl_load(S390_lowcore.cregs_save_area, 0, 15);
 	__load_psw_mask(PSW_KERNEL_BITS | PSW_MASK_DAT);
-	CALL_ON_STACK(smp_init_secondary, S390_lowcore.kernel_stack, 0);
+	CALL_ON_STACK_NORETURN(smp_init_secondary, S390_lowcore.kernel_stack);
 }

 /* Upping and downing of CPUs */
--- a/arch/s390/kernel/stacktrace.c
+++ b/arch/s390/kernel/stacktrace.c
@@ -9,6 +9,7 @@
 #include <linux/stacktrace.h>
 #include <asm/stacktrace.h>
 #include <asm/unwind.h>
+#include <asm/kprobes.h>

 void arch_stack_walk(stack_trace_consume_fn consume_entry, void *cookie,
 		     struct task_struct *task, struct pt_regs *regs)
@@ -22,3 +23,45 @@ void arch_stack_walk(stack_trace_consume_fn consume_entry, void *cookie,
 			break;
 	}
 }
+
+/*
+ * This function returns an error if it detects any unreliable features of the
+ * stack.  Otherwise it guarantees that the stack trace is reliable.
+ *
+ * If the task is not 'current', the caller *must* ensure the task is inactive.
+ */
+int arch_stack_walk_reliable(stack_trace_consume_fn consume_entry,
+			     void *cookie, struct task_struct *task)
+{
+	struct unwind_state state;
+	unsigned long addr;
+
+	unwind_for_each_frame(&state, task, NULL, 0) {
+		if (state.stack_info.type != STACK_TYPE_TASK)
+			return -EINVAL;
+
+		if (state.regs)
+			return -EINVAL;
+
+		addr = unwind_get_return_address(&state);
+		if (!addr)
+			return -EINVAL;
+
+#ifdef CONFIG_KPROBES
+		/*
+		 * Mark stacktraces with kretprobed functions on them
+		 * as unreliable.
+		 */
+		if (state.ip == (unsigned long)kretprobe_trampoline)
+			return -EINVAL;
+#endif
+
+		if (!consume_entry(cookie, addr, false))
+			return -EINVAL;
+	}
+
+	/* Check for stack corruption */
+	if (unwind_error(&state))
+		return -EINVAL;
+	return 0;
+}
--- a/arch/s390/kernel/unwind_bc.c
+++ b/arch/s390/kernel/unwind_bc.c
@@ -36,6 +36,12 @@ static bool update_stack_info(struct unwind_state *state, unsigned long sp)
 	return true;
 }

+static inline bool is_task_pt_regs(struct unwind_state *state,
+				   struct pt_regs *regs)
+{
+	return task_pt_regs(state->task) == regs;
+}
+
 bool unwind_next_frame(struct unwind_state *state)
 {
 	struct stack_info *info = &state->stack_info;
@@ -46,20 +52,16 @@ bool unwind_next_frame(struct unwind_state *state)

 	regs = state->regs;
 	if (unlikely(regs)) {
-		if (state->reuse_sp) {
-			sp = state->sp;
-			state->reuse_sp = false;
-		} else {
-			sp = READ_ONCE_NOCHECK(regs->gprs[15]);
-			if (unlikely(outside_of_stack(state, sp))) {
-				if (!update_stack_info(state, sp))
-					goto out_err;
-			}
-		}
+		sp = state->sp;
 		sf = (struct stack_frame *) sp;
 		ip = READ_ONCE_NOCHECK(sf->gprs[8]);
 		reliable = false;
 		regs = NULL;
+		if (!__kernel_text_address(ip)) {
+			/* skip bogus %r14 */
+			state->regs = NULL;
+			return unwind_next_frame(state);
+		}
 	} else {
 		sf = (struct stack_frame *) state->sp;
 		sp = READ_ONCE_NOCHECK(sf->back_chain);
@@ -76,15 +78,24 @@ bool unwind_next_frame(struct unwind_state *state)
 			/* No back-chain, look for a pt_regs structure */
 			sp = state->sp + STACK_FRAME_OVERHEAD;
 			if (!on_stack(info, sp, sizeof(struct pt_regs)))
-				goto out_stop;
+				goto out_err;
 			regs = (struct pt_regs *) sp;
-			if (READ_ONCE_NOCHECK(regs->psw.mask) & PSW_MASK_PSTATE)
+			if (is_task_pt_regs(state, regs))
 				goto out_stop;
 			ip = READ_ONCE_NOCHECK(regs->psw.addr);
+			sp = READ_ONCE_NOCHECK(regs->gprs[15]);
+			if (unlikely(outside_of_stack(state, sp))) {
+				if (!update_stack_info(state, sp))
+					goto out_err;
+			}
 			reliable = true;
 		}
 	}

+	/* Sanity check: ABI requires SP to be aligned 8 bytes. */
+	if (sp & 0x7)
+		goto out_err;
+
 	ip = ftrace_graph_ret_addr(state->task, &state->graph_idx, ip, (void *) sp);

 	/* Update unwind state */
@@ -103,13 +114,11 @@ out_stop:
 EXPORT_SYMBOL_GPL(unwind_next_frame);

 void __unwind_start(struct unwind_state *state, struct task_struct *task,
-		    struct pt_regs *regs, unsigned long sp)
+		    struct pt_regs *regs, unsigned long first_frame)
 {
 	struct stack_info *info = &state->stack_info;
-	unsigned long *mask = &state->stack_mask;
-	bool reliable, reuse_sp;
 	struct stack_frame *sf;
-	unsigned long ip;
+	unsigned long ip, sp;

 	memset(state, 0, sizeof(*state));
 	state->task = task;
@@ -121,25 +130,28 @@ void __unwind_start(struct unwind_state *state, struct task_struct *task,
 		return;
 	}

+	/* Get the instruction pointer from pt_regs or the stack frame */
+	if (regs) {
+		ip = regs->psw.addr;
+		sp = regs->gprs[15];
+	} else if (task == current) {
+		sp = current_frame_address();
+	} else {
+		sp = task->thread.ksp;
+	}
+
 	/* Get current stack pointer and initialize stack info */
-	if (get_stack_info(sp, task, info, mask) != 0 ||
-	    !on_stack(info, sp, sizeof(struct stack_frame))) {
+	if (!update_stack_info(state, sp)) {
 		/* Something is wrong with the stack pointer */
 		info->type = STACK_TYPE_UNKNOWN;
 		state->error = true;
 		return;
 	}

-	/* Get the instruction pointer from pt_regs or the stack frame */
-	if (regs) {
-		ip = READ_ONCE_NOCHECK(regs->psw.addr);
-		reliable = true;
-		reuse_sp = true;
-	} else {
-		sf = (struct stack_frame *) sp;
+	if (!regs) {
+		/* Stack frame is within valid stack */
+		sf = (struct stack_frame *)sp;
 		ip = READ_ONCE_NOCHECK(sf->gprs[8]);
-		reliable = false;
-		reuse_sp = false;
 	}

 	ip = ftrace_graph_ret_addr(state->task, &state->graph_idx, ip, NULL);
@@ -147,7 +159,17 @@ void __unwind_start(struct unwind_state *state, struct task_struct *task,
 	/* Update unwind state */
 	state->sp = sp;
 	state->ip = ip;
-	state->reliable = reliable;
-	state->reuse_sp = reuse_sp;
+	state->reliable = true;
+
+	if (!first_frame)
+		return;
+	/* Skip through the call chain to the specified starting frame */
+	while (!unwind_done(state)) {
+		if (on_stack(&state->stack_info, first_frame, sizeof(struct stack_frame))) {
+			if (state->sp >= first_frame)
+				break;
+		}
+		unwind_next_frame(state);
+	}
 }
 EXPORT_SYMBOL_GPL(__unwind_start);
--- a/arch/s390/kernel/vdso.c
+++ b/arch/s390/kernel/vdso.c
@@ -29,13 +29,6 @@
 #include <asm/vdso.h>
 #include <asm/facility.h>

-#ifdef CONFIG_COMPAT_VDSO
-extern char vdso32_start, vdso32_end;
-static void *vdso32_kbase = &vdso32_start;
-static unsigned int vdso32_pages;
-static struct page **vdso32_pagelist;
-#endif
-
 extern char vdso64_start, vdso64_end;
 static void *vdso64_kbase = &vdso64_start;
 static unsigned int vdso64_pages;
@@ -55,12 +48,6 @@ static vm_fault_t vdso_fault(const struct vm_special_mapping *sm,

 	vdso_pagelist = vdso64_pagelist;
 	vdso_pages = vdso64_pages;
-#ifdef CONFIG_COMPAT_VDSO
-	if (vma->vm_mm->context.compat_mm) {
-		vdso_pagelist = vdso32_pagelist;
-		vdso_pages = vdso32_pages;
-	}
-#endif

 	if (vmf->pgoff >= vdso_pages)
 		return VM_FAULT_SIGBUS;
@@ -76,10 +63,6 @@ static int vdso_mremap(const struct vm_special_mapping *sm,
 	unsigned long vdso_pages;

 	vdso_pages = vdso64_pages;
-#ifdef CONFIG_COMPAT_VDSO
-	if (vma->vm_mm->context.compat_mm)
-		vdso_pages = vdso32_pages;
-#endif

 	if ((vdso_pages << PAGE_SHIFT) != vma->vm_end - vma->vm_start)
 		return -EINVAL;
@@ -209,12 +192,10 @@ int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
 	if (!vdso_enabled)
 		return 0;

+	if (is_compat_task())
+		return 0;
+
 	vdso_pages = vdso64_pages;
-#ifdef CONFIG_COMPAT_VDSO
-	mm->context.compat_mm = is_compat_task();
-	if (mm->context.compat_mm)
-		vdso_pages = vdso32_pages;
-#endif
 	/*
 	 * vDSO has a problem and was disabled, just don't "enable" it for
 	 * the process
@@ -267,23 +248,6 @@ static int __init vdso_init(void)
 	int i;

 	vdso_init_data(vdso_data);
-#ifdef CONFIG_COMPAT_VDSO
-	/* Calculate the size of the 32 bit vDSO */
-	vdso32_pages = ((&vdso32_end - &vdso32_start
-			 + PAGE_SIZE - 1) >> PAGE_SHIFT) + 1;
-
-	/* Make sure pages are in the correct state */
-	vdso32_pagelist = kcalloc(vdso32_pages + 1, sizeof(struct page *),
-				  GFP_KERNEL);
-	BUG_ON(vdso32_pagelist == NULL);
-	for (i = 0; i < vdso32_pages - 1; i++) {
-		struct page *pg = virt_to_page(vdso32_kbase + i*PAGE_SIZE);
-		get_page(pg);
-		vdso32_pagelist[i] = pg;
-	}
-	vdso32_pagelist[vdso32_pages - 1] = virt_to_page(vdso_data);
-	vdso32_pagelist[vdso32_pages] = NULL;
-#endif

 	/* Calculate the size of the 64 bit vDSO */
 	vdso64_pages = ((&vdso64_end - &vdso64_start
--- a/Show More
+++ b/Show More