Resguardo inicial

linux-7.2.3/arch/s390/mm/Makefile

@@ -0,0 +1,15 @@
+# SPDX-License-Identifier: GPL-2.0
+#
+# Makefile for the linux s390-specific parts of the memory manager.
+#
+
+obj-y		:= init.o fault.o extmem.o mmap.o vmem.o maccess.o
+obj-y		+= page-states.o pageattr.o pgtable.o pgalloc.o extable.o
+
+obj-$(CONFIG_CMM)		+= cmm.o
+obj-$(CONFIG_DEBUG_VIRTUAL)	+= physaddr.o
+obj-$(CONFIG_HUGETLB_PAGE)	+= hugetlbpage.o
+obj-$(CONFIG_PTDUMP)		+= dump_pagetables.o
+obj-$(CONFIG_PFAULT)		+= pfault.o
+
+obj-$(subst m,y,$(CONFIG_KVM))	+= gmap_helpers.o

linux-7.2.3/arch/s390/mm/cmm.c

@@ -0,0 +1,447 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  Collaborative memory management interface.
+ *
+ *    Copyright IBM Corp 2003, 2010
+ *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>,
+ *
+ */
+
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/gfp.h>
+#include <linux/sched.h>
+#include <linux/string_helpers.h>
+#include <linux/sysctl.h>
+#include <linux/swap.h>
+#include <linux/kthread.h>
+#include <linux/oom.h>
+#include <linux/uaccess.h>
+
+#include <asm/diag.h>
+
+#ifdef CONFIG_CMM_IUCV
+static char *cmm_default_sender = "VMRMSVM";
+#endif
+static char *sender;
+module_param(sender, charp, 0400);
+MODULE_PARM_DESC(sender,
+		 "Guest name that may send SMSG messages (default VMRMSVM)");
+
+#include "../../../drivers/s390/net/smsgiucv.h"
+
+#define CMM_NR_PAGES ((PAGE_SIZE / sizeof(unsigned long)) - 2)
+
+struct cmm_page_array {
+	struct cmm_page_array *next;
+	unsigned long index;
+	unsigned long pages[CMM_NR_PAGES];
+};
+
+static long cmm_pages;
+static long cmm_timed_pages;
+static volatile long cmm_pages_target;
+static volatile long cmm_timed_pages_target;
+static long cmm_timeout_pages;
+static long cmm_timeout_seconds;
+
+static struct cmm_page_array *cmm_page_list;
+static struct cmm_page_array *cmm_timed_page_list;
+static DEFINE_SPINLOCK(cmm_lock);
+
+static struct task_struct *cmm_thread_ptr;
+static DECLARE_WAIT_QUEUE_HEAD(cmm_thread_wait);
+
+static void cmm_timer_fn(struct timer_list *);
+static void cmm_set_timer(void);
+static DEFINE_TIMER(cmm_timer, cmm_timer_fn);
+
+static long cmm_alloc_pages(long nr, long *counter,
+			    struct cmm_page_array **list)
+{
+	struct cmm_page_array *pa, *npa;
+	unsigned long addr;
+
+	while (nr) {
+		addr = __get_free_page(GFP_NOIO);
+		if (!addr)
+			break;
+		spin_lock(&cmm_lock);
+		pa = *list;
+		if (!pa || pa->index >= CMM_NR_PAGES) {
+			/* Need a new page for the page list. */
+			spin_unlock(&cmm_lock);
+			npa = (struct cmm_page_array *)
+				__get_free_page(GFP_NOIO);
+			if (!npa) {
+				free_page(addr);
+				break;
+			}
+			spin_lock(&cmm_lock);
+			pa = *list;
+			if (!pa || pa->index >= CMM_NR_PAGES) {
+				npa->next = pa;
+				npa->index = 0;
+				pa = npa;
+				*list = pa;
+			} else
+				free_page((unsigned long) npa);
+		}
+		diag10_range(virt_to_pfn((void *)addr), 1);
+		pa->pages[pa->index++] = addr;
+		(*counter)++;
+		spin_unlock(&cmm_lock);
+		nr--;
+		cond_resched();
+	}
+	return nr;
+}
+
+static long __cmm_free_pages(long nr, long *counter, struct cmm_page_array **list)
+{
+	struct cmm_page_array *pa;
+	unsigned long addr;
+
+	spin_lock(&cmm_lock);
+	pa = *list;
+	while (nr) {
+		if (!pa || pa->index <= 0)
+			break;
+		addr = pa->pages[--pa->index];
+		if (pa->index == 0) {
+			pa = pa->next;
+			free_page((unsigned long) *list);
+			*list = pa;
+		}
+		free_page(addr);
+		(*counter)--;
+		nr--;
+	}
+	spin_unlock(&cmm_lock);
+	return nr;
+}
+
+static long cmm_free_pages(long nr, long *counter, struct cmm_page_array **list)
+{
+	long inc = 0;
+
+	while (nr) {
+		inc = min(256L, nr);
+		nr -= inc;
+		inc = __cmm_free_pages(inc, counter, list);
+		if (inc)
+			break;
+		cond_resched();
+	}
+	return nr + inc;
+}
+
+static int cmm_oom_notify(struct notifier_block *self,
+			  unsigned long dummy, void *parm)
+{
+	unsigned long *freed = parm;
+	long nr = 256;
+
+	nr = cmm_free_pages(nr, &cmm_timed_pages, &cmm_timed_page_list);
+	if (nr > 0)
+		nr = cmm_free_pages(nr, &cmm_pages, &cmm_page_list);
+	cmm_pages_target = cmm_pages;
+	cmm_timed_pages_target = cmm_timed_pages;
+	*freed += 256 - nr;
+	return NOTIFY_OK;
+}
+
+static struct notifier_block cmm_oom_nb = {
+	.notifier_call = cmm_oom_notify,
+};
+
+static int cmm_thread(void *dummy)
+{
+	int rc;
+
+	while (1) {
+		rc = wait_event_interruptible(cmm_thread_wait,
+			cmm_pages != cmm_pages_target ||
+			cmm_timed_pages != cmm_timed_pages_target ||
+			kthread_should_stop());
+		if (kthread_should_stop() || rc == -ERESTARTSYS) {
+			cmm_pages_target = cmm_pages;
+			cmm_timed_pages_target = cmm_timed_pages;
+			break;
+		}
+		if (cmm_pages_target > cmm_pages) {
+			if (cmm_alloc_pages(1, &cmm_pages, &cmm_page_list))
+				cmm_pages_target = cmm_pages;
+		} else if (cmm_pages_target < cmm_pages) {
+			cmm_free_pages(1, &cmm_pages, &cmm_page_list);
+		}
+		if (cmm_timed_pages_target > cmm_timed_pages) {
+			if (cmm_alloc_pages(1, &cmm_timed_pages,
+					   &cmm_timed_page_list))
+				cmm_timed_pages_target = cmm_timed_pages;
+		} else if (cmm_timed_pages_target < cmm_timed_pages) {
+			cmm_free_pages(1, &cmm_timed_pages,
+				       &cmm_timed_page_list);
+		}
+		if (cmm_timed_pages > 0 && !timer_pending(&cmm_timer))
+			cmm_set_timer();
+	}
+	return 0;
+}
+
+static void cmm_kick_thread(void)
+{
+	wake_up(&cmm_thread_wait);
+}
+
+static void cmm_set_timer(void)
+{
+	if (cmm_timed_pages_target <= 0 || cmm_timeout_seconds <= 0) {
+		if (timer_pending(&cmm_timer))
+			timer_delete(&cmm_timer);
+		return;
+	}
+	mod_timer(&cmm_timer, jiffies + secs_to_jiffies(cmm_timeout_seconds));
+}
+
+static void cmm_timer_fn(struct timer_list *unused)
+{
+	long nr;
+
+	nr = cmm_timed_pages_target - cmm_timeout_pages;
+	if (nr < 0)
+		cmm_timed_pages_target = 0;
+	else
+		cmm_timed_pages_target = nr;
+	cmm_kick_thread();
+	cmm_set_timer();
+}
+
+static void cmm_set_pages(long nr)
+{
+	cmm_pages_target = nr;
+	cmm_kick_thread();
+}
+
+static long cmm_get_pages(void)
+{
+	return cmm_pages;
+}
+
+static void cmm_add_timed_pages(long nr)
+{
+	cmm_timed_pages_target += nr;
+	cmm_kick_thread();
+}
+
+static long cmm_get_timed_pages(void)
+{
+	return cmm_timed_pages;
+}
+
+static void cmm_set_timeout(long nr, long seconds)
+{
+	cmm_timeout_pages = nr;
+	cmm_timeout_seconds = seconds;
+	cmm_set_timer();
+}
+
+static int cmm_skip_blanks(char *cp, char **endp)
+{
+	char *str;
+
+	for (str = cp; *str == ' ' || *str == '\t'; str++)
+		;
+	*endp = str;
+	return str != cp;
+}
+
+static int cmm_pages_handler(const struct ctl_table *ctl, int write,
+			     void *buffer, size_t *lenp, loff_t *ppos)
+{
+	long nr = cmm_get_pages();
+	struct ctl_table ctl_entry = {
+		.procname	= ctl->procname,
+		.data		= &nr,
+		.maxlen		= sizeof(long),
+	};
+	int rc;
+
+	rc = proc_doulongvec_minmax(&ctl_entry, write, buffer, lenp, ppos);
+	if (rc < 0 || !write)
+		return rc;
+
+	cmm_set_pages(nr);
+	return 0;
+}
+
+static int cmm_timed_pages_handler(const struct ctl_table *ctl, int write,
+				   void *buffer, size_t *lenp,
+				   loff_t *ppos)
+{
+	long nr = cmm_get_timed_pages();
+	struct ctl_table ctl_entry = {
+		.procname	= ctl->procname,
+		.data		= &nr,
+		.maxlen		= sizeof(long),
+	};
+	int rc;
+
+	rc = proc_doulongvec_minmax(&ctl_entry, write, buffer, lenp, ppos);
+	if (rc < 0 || !write)
+		return rc;
+
+	cmm_add_timed_pages(nr);
+	return 0;
+}
+
+static int cmm_timeout_handler(const struct ctl_table *ctl, int write,
+			       void *buffer, size_t *lenp, loff_t *ppos)
+{
+	char buf[64], *p;
+	long nr, seconds;
+	unsigned int len;
+
+	if (!*lenp || (*ppos && !write)) {
+		*lenp = 0;
+		return 0;
+	}
+
+	if (write) {
+		len = min(*lenp, sizeof(buf));
+		memcpy(buf, buffer, len);
+		buf[len - 1] = '\0';
+		cmm_skip_blanks(buf, &p);
+		nr = simple_strtoul(p, &p, 0);
+		cmm_skip_blanks(p, &p);
+		seconds = simple_strtoul(p, &p, 0);
+		cmm_set_timeout(nr, seconds);
+		*ppos += *lenp;
+	} else {
+		len = scnprintf(buf, sizeof(buf), "%ld %ld\n",
+				cmm_timeout_pages, cmm_timeout_seconds);
+		if (len > *lenp)
+			len = *lenp;
+		memcpy(buffer, buf, len);
+		*lenp = len;
+		*ppos += len;
+	}
+	return 0;
+}
+
+static const struct ctl_table cmm_table[] = {
+	{
+		.procname	= "cmm_pages",
+		.mode		= 0644,
+		.proc_handler	= cmm_pages_handler,
+	},
+	{
+		.procname	= "cmm_timed_pages",
+		.mode		= 0644,
+		.proc_handler	= cmm_timed_pages_handler,
+	},
+	{
+		.procname	= "cmm_timeout",
+		.mode		= 0644,
+		.proc_handler	= cmm_timeout_handler,
+	},
+};
+
+#ifdef CONFIG_CMM_IUCV
+#define SMSG_PREFIX "CMM"
+static void cmm_smsg_target(const char *from, char *msg)
+{
+	long nr, seconds;
+
+	if (strlen(sender) > 0 && strcmp(from, sender) != 0)
+		return;
+	if (!cmm_skip_blanks(msg + strlen(SMSG_PREFIX), &msg))
+		return;
+	if (strncmp(msg, "SHRINK", 6) == 0) {
+		if (!cmm_skip_blanks(msg + 6, &msg))
+			return;
+		nr = simple_strtoul(msg, &msg, 0);
+		cmm_skip_blanks(msg, &msg);
+		if (*msg == '\0')
+			cmm_set_pages(nr);
+	} else if (strncmp(msg, "RELEASE", 7) == 0) {
+		if (!cmm_skip_blanks(msg + 7, &msg))
+			return;
+		nr = simple_strtoul(msg, &msg, 0);
+		cmm_skip_blanks(msg, &msg);
+		if (*msg == '\0')
+			cmm_add_timed_pages(nr);
+	} else if (strncmp(msg, "REUSE", 5) == 0) {
+		if (!cmm_skip_blanks(msg + 5, &msg))
+			return;
+		nr = simple_strtoul(msg, &msg, 0);
+		if (!cmm_skip_blanks(msg, &msg))
+			return;
+		seconds = simple_strtoul(msg, &msg, 0);
+		cmm_skip_blanks(msg, &msg);
+		if (*msg == '\0')
+			cmm_set_timeout(nr, seconds);
+	}
+}
+#endif
+
+static struct ctl_table_header *cmm_sysctl_header;
+
+static int __init cmm_init(void)
+{
+	int rc = -ENOMEM;
+
+	cmm_sysctl_header = register_sysctl("vm", cmm_table);
+	if (!cmm_sysctl_header)
+		goto out_sysctl;
+#ifdef CONFIG_CMM_IUCV
+	/* convert sender to uppercase characters */
+	if (sender)
+		string_upper(sender, sender);
+	else
+		sender = cmm_default_sender;
+
+	rc = smsg_register_callback(SMSG_PREFIX, cmm_smsg_target);
+	if (rc < 0)
+		goto out_smsg;
+#endif
+	rc = register_oom_notifier(&cmm_oom_nb);
+	if (rc < 0)
+		goto out_oom_notify;
+	cmm_thread_ptr = kthread_run(cmm_thread, NULL, "cmmthread");
+	if (!IS_ERR(cmm_thread_ptr))
+		return 0;
+
+	rc = PTR_ERR(cmm_thread_ptr);
+	unregister_oom_notifier(&cmm_oom_nb);
+out_oom_notify:
+#ifdef CONFIG_CMM_IUCV
+	smsg_unregister_callback(SMSG_PREFIX, cmm_smsg_target);
+out_smsg:
+#endif
+	unregister_sysctl_table(cmm_sysctl_header);
+out_sysctl:
+	timer_delete_sync(&cmm_timer);
+	return rc;
+}
+module_init(cmm_init);
+
+static void __exit cmm_exit(void)
+{
+	unregister_sysctl_table(cmm_sysctl_header);
+#ifdef CONFIG_CMM_IUCV
+	smsg_unregister_callback(SMSG_PREFIX, cmm_smsg_target);
+#endif
+	unregister_oom_notifier(&cmm_oom_nb);
+	kthread_stop(cmm_thread_ptr);
+	timer_delete_sync(&cmm_timer);
+	cmm_free_pages(cmm_pages, &cmm_pages, &cmm_page_list);
+	cmm_free_pages(cmm_timed_pages, &cmm_timed_pages, &cmm_timed_page_list);
+}
+module_exit(cmm_exit);
+
+MODULE_DESCRIPTION("Cooperative memory management interface");
+MODULE_LICENSE("GPL");

linux-7.2.3/arch/s390/mm/dump_pagetables.c

@@ -0,0 +1,364 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/cpufeature.h>
+#include <linux/set_memory.h>
+#include <linux/ptdump.h>
+#include <linux/seq_file.h>
+#include <linux/debugfs.h>
+#include <linux/sort.h>
+#include <linux/mm.h>
+#include <linux/kfence.h>
+#include <linux/kasan.h>
+#include <asm/kasan.h>
+#include <asm/abs_lowcore.h>
+#include <asm/nospec-branch.h>
+#include <asm/sections.h>
+#include <asm/maccess.h>
+
+static unsigned long max_addr;
+
+struct addr_marker {
+	int is_start;
+	unsigned long start_address;
+	unsigned long size;
+	const char *name;
+};
+
+static struct addr_marker *markers;
+static unsigned int markers_cnt;
+
+struct pg_state {
+	struct ptdump_state ptdump;
+	struct seq_file *seq;
+	int level;
+	unsigned int current_prot;
+	bool check_wx;
+	unsigned long wx_pages;
+	unsigned long start_address;
+	const struct addr_marker *marker;
+};
+
+#define pt_dump_seq_printf(m, fmt, args...)	\
+({						\
+	struct seq_file *__m = (m);		\
+						\
+	if (__m)				\
+		seq_printf(__m, fmt, ##args);	\
+})
+
+#define pt_dump_seq_puts(m, fmt)		\
+({						\
+	struct seq_file *__m = (m);		\
+						\
+	if (__m)				\
+		seq_puts(__m, fmt);		\
+})
+
+static void print_prot(struct seq_file *m, unsigned int pr, int level)
+{
+	static const char * const level_name[] =
+		{ "ASCE", "PGD", "PUD", "PMD", "PTE" };
+
+	pt_dump_seq_printf(m, "%s ", level_name[level]);
+	if (pr & _PAGE_INVALID) {
+		pt_dump_seq_printf(m, "I\n");
+		return;
+	}
+	pt_dump_seq_puts(m, (pr & _PAGE_PROTECT) ? "RO " : "RW ");
+	pt_dump_seq_puts(m, (pr & _PAGE_NOEXEC) ? "NX\n" : "X\n");
+}
+
+static void note_prot_wx(struct pg_state *st, unsigned long addr)
+{
+	if (!st->check_wx)
+		return;
+	if (st->current_prot & _PAGE_INVALID)
+		return;
+	if (st->current_prot & _PAGE_PROTECT)
+		return;
+	if (st->current_prot & _PAGE_NOEXEC)
+		return;
+	/*
+	 * The first lowcore page is W+X if spectre mitigations are using
+	 * trampolines or the BEAR enhancements facility is not installed,
+	 * in which case we have two lpswe instructions in lowcore that need
+	 * to be executable.
+	 */
+	if (addr == PAGE_SIZE && (nospec_uses_trampoline() || !cpu_has_bear()))
+		return;
+	WARN_ONCE(IS_ENABLED(CONFIG_DEBUG_WX),
+		  "s390/mm: Found insecure W+X mapping at address %pS\n",
+		  (void *)st->start_address);
+	st->wx_pages += (addr - st->start_address) / PAGE_SIZE;
+}
+
+static void note_page_update_state(struct pg_state *st, unsigned long addr, unsigned int prot, int level)
+{
+	struct seq_file *m = st->seq;
+
+	while (addr >= st->marker[1].start_address) {
+		st->marker++;
+		pt_dump_seq_printf(m, "---[ %s %s ]---\n", st->marker->name,
+				   st->marker->is_start ? "Start" : "End");
+	}
+	st->start_address = addr;
+	st->current_prot = prot;
+	st->level = level;
+}
+
+static void note_page(struct ptdump_state *pt_st, unsigned long addr, int level, u64 val)
+{
+	int width = sizeof(unsigned long) * 2;
+	static const char units[] = "KMGTPE";
+	const char *unit = units;
+	unsigned long delta;
+	struct pg_state *st;
+	struct seq_file *m;
+	unsigned int prot;
+
+	st = container_of(pt_st, struct pg_state, ptdump);
+	m = st->seq;
+	prot = val & (_PAGE_PROTECT | _PAGE_NOEXEC);
+	if (level == 4 && (val & _PAGE_INVALID))
+		prot = _PAGE_INVALID;
+	/* For pmd_none() & friends val gets passed as zero. */
+	if (level != 4 && !val)
+		prot = _PAGE_INVALID;
+	/* Final flush from generic code. */
+	if (level == -1)
+		addr = max_addr;
+	if (st->level == -1) {
+		pt_dump_seq_puts(m, "---[ Kernel Virtual Address Space ]---\n");
+		note_page_update_state(st, addr, prot, level);
+	} else if (prot != st->current_prot || level != st->level ||
+		   addr >= st->marker[1].start_address) {
+		note_prot_wx(st, addr);
+		pt_dump_seq_printf(m, "0x%0*lx-0x%0*lx ",
+				   width, st->start_address,
+				   width, addr);
+		delta = (addr - st->start_address) >> 10;
+		while (!(delta & 0x3ff) && unit[1]) {
+			delta >>= 10;
+			unit++;
+		}
+		pt_dump_seq_printf(m, "%9lu%c ", delta, *unit);
+		print_prot(m, st->current_prot, st->level);
+		note_page_update_state(st, addr, prot, level);
+	}
+}
+
+static void note_page_pte(struct ptdump_state *pt_st, unsigned long addr, pte_t pte)
+{
+	note_page(pt_st, addr, 4, pte_val(pte));
+}
+
+static void note_page_pmd(struct ptdump_state *pt_st, unsigned long addr, pmd_t pmd)
+{
+	note_page(pt_st, addr, 3, pmd_val(pmd));
+}
+
+static void note_page_pud(struct ptdump_state *pt_st, unsigned long addr, pud_t pud)
+{
+	note_page(pt_st, addr, 2, pud_val(pud));
+}
+
+static void note_page_p4d(struct ptdump_state *pt_st, unsigned long addr, p4d_t p4d)
+{
+	note_page(pt_st, addr, 1, p4d_val(p4d));
+}
+
+static void note_page_pgd(struct ptdump_state *pt_st, unsigned long addr, pgd_t pgd)
+{
+	note_page(pt_st, addr, 0, pgd_val(pgd));
+}
+
+static void note_page_flush(struct ptdump_state *pt_st)
+{
+	pte_t pte_zero = {0};
+
+	note_page(pt_st, 0, -1, pte_val(pte_zero));
+}
+
+bool ptdump_check_wx(void)
+{
+	struct pg_state st = {
+		.ptdump = {
+			.note_page_pte = note_page_pte,
+			.note_page_pmd = note_page_pmd,
+			.note_page_pud = note_page_pud,
+			.note_page_p4d = note_page_p4d,
+			.note_page_pgd = note_page_pgd,
+			.note_page_flush = note_page_flush,
+			.range = (struct ptdump_range[]) {
+				{.start = 0, .end = max_addr},
+				{.start = 0, .end = 0},
+			}
+		},
+		.seq = NULL,
+		.level = -1,
+		.current_prot = 0,
+		.check_wx = true,
+		.wx_pages = 0,
+		.start_address = 0,
+		.marker = (struct addr_marker[]) {
+			{ .start_address =  0, .name = NULL},
+			{ .start_address = -1, .name = NULL},
+		},
+	};
+
+	if (!cpu_has_nx())
+		return true;
+	ptdump_walk_pgd(&st.ptdump, &init_mm, NULL);
+	if (st.wx_pages) {
+		pr_warn("Checked W+X mappings: FAILED, %lu W+X pages found\n", st.wx_pages);
+
+		return false;
+	} else {
+		pr_info("Checked W+X mappings: passed, no %sW+X pages found\n",
+			(nospec_uses_trampoline() || !cpu_has_bear()) ?
+			"unexpected " : "");
+
+		return true;
+	}
+}
+
+#ifdef CONFIG_PTDUMP_DEBUGFS
+static int ptdump_show(struct seq_file *m, void *v)
+{
+	struct pg_state st = {
+		.ptdump = {
+			.note_page_pte = note_page_pte,
+			.note_page_pmd = note_page_pmd,
+			.note_page_pud = note_page_pud,
+			.note_page_p4d = note_page_p4d,
+			.note_page_pgd = note_page_pgd,
+			.note_page_flush = note_page_flush,
+			.range = (struct ptdump_range[]) {
+				{.start = 0, .end = max_addr},
+				{.start = 0, .end = 0},
+			}
+		},
+		.seq = m,
+		.level = -1,
+		.current_prot = 0,
+		.check_wx = false,
+		.wx_pages = 0,
+		.start_address = 0,
+		.marker = markers,
+	};
+
+	mutex_lock(&cpa_mutex);
+	ptdump_walk_pgd(&st.ptdump, &init_mm, NULL);
+	mutex_unlock(&cpa_mutex);
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(ptdump);
+#endif /* CONFIG_PTDUMP_DEBUGFS */
+
+static int ptdump_cmp(const void *a, const void *b)
+{
+	const struct addr_marker *ama = a;
+	const struct addr_marker *amb = b;
+
+	if (ama->start_address > amb->start_address)
+		return 1;
+	if (ama->start_address < amb->start_address)
+		return -1;
+	/*
+	 * If the start addresses of two markers are identical sort markers in an
+	 * order that considers areas contained within other areas correctly.
+	 */
+	if (ama->is_start && amb->is_start) {
+		if (ama->size > amb->size)
+			return -1;
+		if (ama->size < amb->size)
+			return 1;
+		return 0;
+	}
+	if (!ama->is_start && !amb->is_start) {
+		if (ama->size > amb->size)
+			return 1;
+		if (ama->size < amb->size)
+			return -1;
+		return 0;
+	}
+	if (ama->is_start)
+		return 1;
+	if (amb->is_start)
+		return -1;
+	return 0;
+}
+
+static int add_marker(unsigned long start, unsigned long end, const char *name)
+{
+	struct addr_marker *new;
+	size_t newsize;
+
+	newsize = (markers_cnt + 2) * sizeof(*markers);
+	new = kvrealloc(markers, newsize, GFP_KERNEL);
+	if (!new)
+		return -ENOMEM;
+	markers = new;
+	markers[markers_cnt].is_start = 1;
+	markers[markers_cnt].start_address = start;
+	markers[markers_cnt].size = end - start;
+	markers[markers_cnt].name = name;
+	markers_cnt++;
+	markers[markers_cnt].is_start = 0;
+	markers[markers_cnt].start_address = end;
+	markers[markers_cnt].size = end - start;
+	markers[markers_cnt].name = name;
+	markers_cnt++;
+	return 0;
+}
+
+static int pt_dump_init(void)
+{
+#ifdef CONFIG_KFENCE
+	unsigned long kfence_start = (unsigned long)__kfence_pool;
+#endif
+	unsigned long lowcore = (unsigned long)get_lowcore();
+	int rc;
+
+	/*
+	 * Figure out the maximum virtual address being accessible with the
+	 * kernel ASCE. We need this to keep the page table walker functions
+	 * from accessing non-existent entries.
+	 */
+	max_addr = (get_lowcore()->kernel_asce.val & _REGION_ENTRY_TYPE_MASK) >> 2;
+	max_addr = 1UL << (max_addr * 11 + 31);
+	/* start + end markers - must be added first */
+	rc = add_marker(0, -1UL, NULL);
+	rc |= add_marker((unsigned long)_stext, (unsigned long)_end, "Kernel Image");
+	rc |= add_marker(lowcore, lowcore + sizeof(struct lowcore), "Lowcore");
+	rc |= add_marker(__identity_base, __identity_base + ident_map_size, "Identity Mapping");
+	rc |= add_marker((unsigned long)__samode31, (unsigned long)__eamode31, "Amode31 Area");
+	rc |= add_marker(MODULES_VADDR, MODULES_END, "Modules Area");
+	rc |= add_marker(__abs_lowcore, __abs_lowcore + ABS_LOWCORE_MAP_SIZE, "Lowcore Area");
+	rc |= add_marker(__memcpy_real_area, __memcpy_real_area + MEMCPY_REAL_SIZE, "Real Memory Copy Area");
+	rc |= add_marker((unsigned long)vmemmap, (unsigned long)vmemmap + vmemmap_size, "vmemmap Area");
+	rc |= add_marker(VMALLOC_START, VMALLOC_END, "vmalloc Area");
+#ifdef CONFIG_KFENCE
+	rc |= add_marker(kfence_start, kfence_start + KFENCE_POOL_SIZE, "KFence Pool");
+#endif
+#ifdef CONFIG_KMSAN
+	rc |= add_marker(KMSAN_VMALLOC_SHADOW_START, KMSAN_VMALLOC_SHADOW_END, "Kmsan vmalloc Shadow");
+	rc |= add_marker(KMSAN_VMALLOC_ORIGIN_START, KMSAN_VMALLOC_ORIGIN_END, "Kmsan vmalloc Origins");
+	rc |= add_marker(KMSAN_MODULES_SHADOW_START, KMSAN_MODULES_SHADOW_END, "Kmsan Modules Shadow");
+	rc |= add_marker(KMSAN_MODULES_ORIGIN_START, KMSAN_MODULES_ORIGIN_END, "Kmsan Modules Origins");
+#endif
+#ifdef CONFIG_KASAN
+	rc |= add_marker(KASAN_SHADOW_START, KASAN_SHADOW_END, "Kasan Shadow");
+#endif
+	if (rc)
+		goto error;
+	sort(&markers[1], markers_cnt - 1, sizeof(*markers), ptdump_cmp, NULL);
+#ifdef CONFIG_PTDUMP_DEBUGFS
+	debugfs_create_file("kernel_page_tables", 0400, NULL, NULL, &ptdump_fops);
+#endif /* CONFIG_PTDUMP_DEBUGFS */
+	return 0;
+error:
+	kvfree(markers);
+	return -ENOMEM;
+}
+device_initcall(pt_dump_init);

linux-7.2.3/arch/s390/mm/extable.c

@@ -0,0 +1,129 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/bitfield.h>
+#include <linux/extable.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/panic.h>
+#include <asm/asm-extable.h>
+#include <asm/extable.h>
+#include <asm/fpu.h>
+
+const struct exception_table_entry *s390_search_extables(unsigned long addr)
+{
+	const struct exception_table_entry *fixup;
+	size_t num;
+
+	fixup = search_exception_tables(addr);
+	if (fixup)
+		return fixup;
+	num = __stop_amode31_ex_table - __start_amode31_ex_table;
+	return search_extable(__start_amode31_ex_table, num, addr);
+}
+
+static bool ex_handler_fixup(const struct exception_table_entry *ex, struct pt_regs *regs)
+{
+	regs->psw.addr = extable_fixup(ex);
+	return true;
+}
+
+static bool ex_handler_ua_fault(const struct exception_table_entry *ex, struct pt_regs *regs)
+{
+	unsigned int reg_err = FIELD_GET(EX_DATA_REG_ERR, ex->data);
+
+	regs->gprs[reg_err] = -EFAULT;
+	regs->psw.addr = extable_fixup(ex);
+	return true;
+}
+
+static bool ex_handler_ua_load_reg(const struct exception_table_entry *ex,
+				   bool pair, struct pt_regs *regs)
+{
+	unsigned int reg_zero = FIELD_GET(EX_DATA_REG_ADDR, ex->data);
+	unsigned int reg_err = FIELD_GET(EX_DATA_REG_ERR, ex->data);
+
+	regs->gprs[reg_err] = -EFAULT;
+	regs->gprs[reg_zero] = 0;
+	if (pair)
+		regs->gprs[reg_zero + 1] = 0;
+	regs->psw.addr = extable_fixup(ex);
+	return true;
+}
+
+static bool ex_handler_fpc(const struct exception_table_entry *ex, struct pt_regs *regs)
+{
+	fpu_sfpc(0);
+	regs->psw.addr = extable_fixup(ex);
+	return true;
+}
+
+struct insn_ssf {
+	u64	opc1 : 8;
+	u64	r3   : 4;
+	u64	opc2 : 4;
+	u64	b1   : 4;
+	u64	d1   : 12;
+	u64	b2   : 4;
+	u64	d2   : 12;
+} __packed;
+
+static bool ex_handler_ua_mvcos(const struct exception_table_entry *ex,
+				bool from, struct pt_regs *regs)
+{
+	unsigned long uaddr, remainder;
+	struct insn_ssf *insn;
+
+	/*
+	 * If the faulting user space access crossed a page boundary retry by
+	 * limiting the access to the first page (adjust length accordingly).
+	 * Then the mvcos instruction will either complete with condition code
+	 * zero, or generate another fault where the user space access did not
+	 * cross a page boundary.
+	 * If the faulting user space access did not cross a page boundary set
+	 * length to zero and retry. In this case no user space access will
+	 * happen, and the mvcos instruction will complete with condition code
+	 * zero.
+	 * In both cases the instruction will complete with condition code
+	 * zero (copying finished), and the register which contains the
+	 * length, indicates the number of bytes copied.
+	 */
+	regs->psw.addr = extable_fixup(ex);
+	insn = (struct insn_ssf *)regs->psw.addr;
+	if (from)
+		uaddr = regs->gprs[insn->b2] + insn->d2;
+	else
+		uaddr = regs->gprs[insn->b1] + insn->d1;
+	remainder = PAGE_SIZE - (uaddr & (PAGE_SIZE - 1));
+	if (regs->gprs[insn->r3] <= remainder)
+		remainder = 0;
+	regs->gprs[insn->r3] = remainder;
+	return true;
+}
+
+bool fixup_exception(struct pt_regs *regs)
+{
+	const struct exception_table_entry *ex;
+
+	ex = s390_search_extables(instruction_pointer(regs));
+	if (!ex)
+		return false;
+	switch (ex->type) {
+	case EX_TYPE_FIXUP:
+		return ex_handler_fixup(ex, regs);
+	case EX_TYPE_BPF:
+		return ex_handler_bpf(ex, regs);
+	case EX_TYPE_UA_FAULT:
+		return ex_handler_ua_fault(ex, regs);
+	case EX_TYPE_UA_LOAD_REG:
+		return ex_handler_ua_load_reg(ex, false, regs);
+	case EX_TYPE_UA_LOAD_REGPAIR:
+		return ex_handler_ua_load_reg(ex, true, regs);
+	case EX_TYPE_FPC:
+		return ex_handler_fpc(ex, regs);
+	case EX_TYPE_UA_MVCOS_TO:
+		return ex_handler_ua_mvcos(ex, false, regs);
+	case EX_TYPE_UA_MVCOS_FROM:
+		return ex_handler_ua_mvcos(ex, true, regs);
+	}
+	panic("invalid exception table entry");
+}

linux-7.2.3/arch/s390/mm/extmem.c

@@ -0,0 +1,678 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Author(s)......: Carsten Otte <cotte@de.ibm.com>
+ * 		    Rob M van der Heij <rvdheij@nl.ibm.com>
+ * 		    Steven Shultz <shultzss@us.ibm.com>
+ * Bugreports.to..: <Linux390@de.ibm.com>
+ * Copyright IBM Corp. 2002, 2004
+ */
+
+#define pr_fmt(fmt) "extmem: " fmt
+
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/spinlock.h>
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <linux/export.h>
+#include <linux/memblock.h>
+#include <linux/ctype.h>
+#include <linux/ioport.h>
+#include <linux/refcount.h>
+#include <linux/pgtable.h>
+#include <asm/machine.h>
+#include <asm/diag.h>
+#include <asm/page.h>
+#include <asm/ebcdic.h>
+#include <asm/errno.h>
+#include <asm/extmem.h>
+#include <asm/cpcmd.h>
+#include <asm/setup.h>
+#include <asm/asm.h>
+
+#define DCSS_PURGESEG   0x08
+#define DCSS_LOADSHRX	0x20
+#define DCSS_LOADNSRX	0x24
+#define DCSS_FINDSEGX	0x2c
+#define DCSS_SEGEXTX	0x38
+#define DCSS_FINDSEGA   0x0c
+
+struct qrange {
+	unsigned long  start; /* last byte type */
+	unsigned long  end;   /* last byte reserved */
+};
+
+struct qout64 {
+	unsigned long segstart;
+	unsigned long segend;
+	int segcnt;
+	int segrcnt;
+	struct qrange range[6];
+};
+
+struct qin64 {
+	char qopcode;
+	char rsrv1[3];
+	char qrcode;
+	char rsrv2[3];
+	char qname[8];
+	unsigned int qoutptr;
+	short int qoutlen;
+};
+
+struct dcss_segment {
+	struct list_head list;
+	char dcss_name[8];
+	char res_name[16];
+	unsigned long start_addr;
+	unsigned long end;
+	refcount_t ref_count;
+	int do_nonshared;
+	unsigned int vm_segtype;
+	struct qrange range[6];
+	int segcnt;
+	struct resource *res;
+};
+
+static DEFINE_MUTEX(dcss_lock);
+static LIST_HEAD(dcss_list);
+static char *segtype_string[] = { "SW", "EW", "SR", "ER", "SN", "EN", "SC",
+					"EW/EN-MIXED" };
+static int loadshr_scode = DCSS_LOADSHRX;
+static int loadnsr_scode = DCSS_LOADNSRX;
+static int purgeseg_scode = DCSS_PURGESEG;
+static int segext_scode = DCSS_SEGEXTX;
+
+/*
+ * Create the 8 bytes, ebcdic VM segment name from
+ * an ascii name.
+ */
+static void
+dcss_mkname(char *name, char *dcss_name)
+{
+	int i;
+
+	for (i = 0; i < 8; i++) {
+		if (name[i] == '\0')
+			break;
+		dcss_name[i] = toupper(name[i]);
+	}
+	for (; i < 8; i++)
+		dcss_name[i] = ' ';
+	ASCEBC(dcss_name, 8);
+}
+
+
+/*
+ * search all segments in dcss_list, and return the one
+ * namend *name. If not found, return NULL.
+ */
+static struct dcss_segment *
+segment_by_name (char *name)
+{
+	char dcss_name[9];
+	struct list_head *l;
+	struct dcss_segment *tmp, *retval = NULL;
+
+	BUG_ON(!mutex_is_locked(&dcss_lock));
+	dcss_mkname (name, dcss_name);
+	list_for_each (l, &dcss_list) {
+		tmp = list_entry (l, struct dcss_segment, list);
+		if (memcmp(tmp->dcss_name, dcss_name, 8) == 0) {
+			retval = tmp;
+			break;
+		}
+	}
+	return retval;
+}
+
+
+/*
+ * Perform a function on a dcss segment.
+ */
+static inline int
+dcss_diag(int *func, void *parameter,
+           unsigned long *ret1, unsigned long *ret2)
+{
+	unsigned long rx, ry;
+	int cc;
+
+	rx = virt_to_phys(parameter);
+	ry = (unsigned long) *func;
+
+	diag_stat_inc(DIAG_STAT_X064);
+	asm volatile(
+		"	diag	%[rx],%[ry],0x64\n"
+		CC_IPM(cc)
+		: CC_OUT(cc, cc), [rx] "+d" (rx), [ry] "+d" (ry)
+		:
+		: CC_CLOBBER);
+	*ret1 = rx;
+	*ret2 = ry;
+	return CC_TRANSFORM(cc);
+}
+
+static inline int
+dcss_diag_translate_rc (int vm_rc) {
+	if (vm_rc == 44)
+		return -ENOENT;
+	return -EIO;
+}
+
+
+/* do a diag to get info about a segment.
+ * fills start_address, end and vm_segtype fields
+ */
+static int
+query_segment_type (struct dcss_segment *seg)
+{
+	unsigned long dummy, vmrc;
+	int diag_cc, rc, i;
+	struct qout64 *qout;
+	struct qin64 *qin;
+
+	qin = kmalloc_obj(*qin, GFP_KERNEL | GFP_DMA);
+	qout = kmalloc_obj(*qout, GFP_KERNEL | GFP_DMA);
+	if ((qin == NULL) || (qout == NULL)) {
+		rc = -ENOMEM;
+		goto out_free;
+	}
+
+	/* initialize diag input parameters */
+	qin->qopcode = DCSS_FINDSEGA;
+	qin->qoutptr = virt_to_phys(qout);
+	qin->qoutlen = sizeof(struct qout64);
+	memcpy (qin->qname, seg->dcss_name, 8);
+
+	diag_cc = dcss_diag(&segext_scode, qin, &dummy, &vmrc);
+
+	if (diag_cc < 0) {
+		rc = diag_cc;
+		goto out_free;
+	}
+	if (diag_cc > 1) {
+		pr_warn("Querying a DCSS type failed with rc=%ld\n", vmrc);
+		rc = dcss_diag_translate_rc (vmrc);
+		goto out_free;
+	}
+
+	if (qout->segcnt > 6) {
+		rc = -EOPNOTSUPP;
+		goto out_free;
+	}
+
+	if (qout->segcnt == 1) {
+		seg->vm_segtype = qout->range[0].start & 0xff;
+	} else {
+		/* multi-part segment. only one type supported here:
+		    - all parts are contiguous
+		    - all parts are either EW or EN type
+		    - maximum 6 parts allowed */
+		unsigned long start = qout->segstart >> PAGE_SHIFT;
+		for (i=0; i<qout->segcnt; i++) {
+			if (((qout->range[i].start & 0xff) != SEG_TYPE_EW) &&
+			    ((qout->range[i].start & 0xff) != SEG_TYPE_EN)) {
+				rc = -EOPNOTSUPP;
+				goto out_free;
+			}
+			if (start != qout->range[i].start >> PAGE_SHIFT) {
+				rc = -EOPNOTSUPP;
+				goto out_free;
+			}
+			start = (qout->range[i].end >> PAGE_SHIFT) + 1;
+		}
+		seg->vm_segtype = SEG_TYPE_EWEN;
+	}
+
+	/* analyze diag output and update seg */
+	seg->start_addr = qout->segstart;
+	seg->end = qout->segend;
+
+	memcpy (seg->range, qout->range, 6*sizeof(struct qrange));
+	seg->segcnt = qout->segcnt;
+
+	rc = 0;
+
+ out_free:
+	kfree(qin);
+	kfree(qout);
+	return rc;
+}
+
+/*
+ * get info about a segment
+ * possible return values:
+ * -ENOSYS  : we are not running on VM
+ * -EIO     : could not perform query diagnose
+ * -ENOENT  : no such segment
+ * -EOPNOTSUPP: multi-part segment cannot be used with linux
+ * -ENOMEM  : out of memory
+ * 0 .. 6   : type of segment as defined in include/asm-s390/extmem.h
+ */
+int
+segment_type (char* name)
+{
+	int rc;
+	struct dcss_segment seg;
+
+	if (!machine_is_vm())
+		return -ENOSYS;
+
+	dcss_mkname(name, seg.dcss_name);
+	rc = query_segment_type (&seg);
+	if (rc < 0)
+		return rc;
+	return seg.vm_segtype;
+}
+
+/*
+ * check if segment collides with other segments that are currently loaded
+ * returns 1 if this is the case, 0 if no collision was found
+ */
+static int
+segment_overlaps_others (struct dcss_segment *seg)
+{
+	struct list_head *l;
+	struct dcss_segment *tmp;
+
+	BUG_ON(!mutex_is_locked(&dcss_lock));
+	list_for_each(l, &dcss_list) {
+		tmp = list_entry(l, struct dcss_segment, list);
+		if ((tmp->start_addr >> 20) > (seg->end >> 20))
+			continue;
+		if ((tmp->end >> 20) < (seg->start_addr >> 20))
+			continue;
+		if (seg == tmp)
+			continue;
+		return 1;
+	}
+	return 0;
+}
+
+/*
+ * real segment loading function, called from segment_load
+ * Must return either an error code < 0, or the segment type code >= 0
+ */
+static int
+__segment_load (char *name, int do_nonshared, unsigned long *addr, unsigned long *end)
+{
+	unsigned long start_addr, end_addr, dummy;
+	struct dcss_segment *seg;
+	int rc, diag_cc, segtype;
+
+	start_addr = end_addr = 0;
+	segtype = -1;
+	seg = kmalloc_obj(*seg, GFP_KERNEL | GFP_DMA);
+	if (seg == NULL) {
+		rc = -ENOMEM;
+		goto out;
+	}
+	dcss_mkname (name, seg->dcss_name);
+	rc = query_segment_type (seg);
+	if (rc < 0)
+		goto out_free;
+
+	if (segment_overlaps_others(seg)) {
+		rc = -EBUSY;
+		goto out_free;
+	}
+
+	seg->res = kzalloc_obj(struct resource);
+	if (seg->res == NULL) {
+		rc = -ENOMEM;
+		goto out_free;
+	}
+	seg->res->flags = IORESOURCE_BUSY | IORESOURCE_MEM;
+	seg->res->start = seg->start_addr;
+	seg->res->end = seg->end;
+	memcpy(&seg->res_name, seg->dcss_name, 8);
+	EBCASC(seg->res_name, 8);
+	seg->res_name[8] = '\0';
+	strlcat(seg->res_name, " (DCSS)", sizeof(seg->res_name));
+	seg->res->name = seg->res_name;
+	segtype = seg->vm_segtype;
+	if (segtype == SEG_TYPE_SC ||
+	    ((segtype == SEG_TYPE_SR || segtype == SEG_TYPE_ER) && !do_nonshared))
+		seg->res->flags |= IORESOURCE_READONLY;
+
+	/* Check for overlapping resources before adding the mapping. */
+	if (request_resource(&iomem_resource, seg->res)) {
+		rc = -EBUSY;
+		goto out_free_resource;
+	}
+
+	rc = vmem_add_mapping(seg->start_addr, seg->end - seg->start_addr + 1);
+	if (rc)
+		goto out_resource;
+
+	if (do_nonshared)
+		diag_cc = dcss_diag(&loadnsr_scode, seg->dcss_name,
+				&start_addr, &end_addr);
+	else
+		diag_cc = dcss_diag(&loadshr_scode, seg->dcss_name,
+				&start_addr, &end_addr);
+	if (diag_cc < 0) {
+		dcss_diag(&purgeseg_scode, seg->dcss_name,
+				&dummy, &dummy);
+		rc = diag_cc;
+		goto out_mapping;
+	}
+	if (diag_cc > 1) {
+		pr_warn("Loading DCSS %s failed with rc=%ld\n", name, end_addr);
+		rc = dcss_diag_translate_rc(end_addr);
+		dcss_diag(&purgeseg_scode, seg->dcss_name,
+				&dummy, &dummy);
+		goto out_mapping;
+	}
+	seg->start_addr = start_addr;
+	seg->end = end_addr;
+	seg->do_nonshared = do_nonshared;
+	refcount_set(&seg->ref_count, 1);
+	list_add(&seg->list, &dcss_list);
+	*addr = seg->start_addr;
+	*end  = seg->end;
+	if (do_nonshared)
+		pr_info("DCSS %s of range %px to %px and type %s loaded as "
+			"exclusive-writable\n", name, (void*) seg->start_addr,
+			(void*) seg->end, segtype_string[seg->vm_segtype]);
+	else {
+		pr_info("DCSS %s of range %px to %px and type %s loaded in "
+			"shared access mode\n", name, (void*) seg->start_addr,
+			(void*) seg->end, segtype_string[seg->vm_segtype]);
+	}
+	goto out;
+ out_mapping:
+	vmem_remove_mapping(seg->start_addr, seg->end - seg->start_addr + 1);
+ out_resource:
+	release_resource(seg->res);
+ out_free_resource:
+	kfree(seg->res);
+ out_free:
+	kfree(seg);
+ out:
+	return rc < 0 ? rc : segtype;
+}
+
+/*
+ * this function loads a DCSS segment
+ * name         : name of the DCSS
+ * do_nonshared : 0 indicates that the dcss should be shared with other linux images
+ *                1 indicates that the dcss should be exclusive for this linux image
+ * addr         : will be filled with start address of the segment
+ * end          : will be filled with end address of the segment
+ * return values:
+ * -ENOSYS  : we are not running on VM
+ * -EIO     : could not perform query or load diagnose
+ * -ENOENT  : no such segment
+ * -EOPNOTSUPP: multi-part segment cannot be used with linux
+ * -EBUSY   : segment cannot be used (overlaps with dcss or storage)
+ * -ERANGE  : segment cannot be used (exceeds kernel mapping range)
+ * -EPERM   : segment is currently loaded with incompatible permissions
+ * -ENOMEM  : out of memory
+ * 0 .. 6   : type of segment as defined in include/asm-s390/extmem.h
+ */
+int
+segment_load (char *name, int do_nonshared, unsigned long *addr,
+		unsigned long *end)
+{
+	struct dcss_segment *seg;
+	int rc;
+
+	if (!machine_is_vm())
+		return -ENOSYS;
+
+	mutex_lock(&dcss_lock);
+	seg = segment_by_name (name);
+	if (seg == NULL)
+		rc = __segment_load (name, do_nonshared, addr, end);
+	else {
+		if (do_nonshared == seg->do_nonshared) {
+			refcount_inc(&seg->ref_count);
+			*addr = seg->start_addr;
+			*end  = seg->end;
+			rc    = seg->vm_segtype;
+		} else {
+			*addr = *end = 0;
+			rc    = -EPERM;
+		}
+	}
+	mutex_unlock(&dcss_lock);
+	return rc;
+}
+
+/*
+ * this function modifies the shared state of a DCSS segment. note that
+ * name         : name of the DCSS
+ * do_nonshared : 0 indicates that the dcss should be shared with other linux images
+ *                1 indicates that the dcss should be exclusive for this linux image
+ * return values:
+ * -EIO     : could not perform load diagnose (segment gone!)
+ * -ENOENT  : no such segment (segment gone!)
+ * -EAGAIN  : segment is in use by other exploiters, try later
+ * -EINVAL  : no segment with the given name is currently loaded - name invalid
+ * -EBUSY   : segment can temporarily not be used (overlaps with dcss)
+ * 0	    : operation succeeded
+ */
+int
+segment_modify_shared (char *name, int do_nonshared)
+{
+	struct dcss_segment *seg;
+	unsigned long start_addr, end_addr, dummy;
+	int rc, diag_cc;
+
+	start_addr = end_addr = 0;
+	mutex_lock(&dcss_lock);
+	seg = segment_by_name (name);
+	if (seg == NULL) {
+		rc = -EINVAL;
+		goto out_unlock;
+	}
+	if (do_nonshared == seg->do_nonshared) {
+		pr_info("DCSS %s is already in the requested access "
+			"mode\n", name);
+		rc = 0;
+		goto out_unlock;
+	}
+	if (refcount_read(&seg->ref_count) != 1) {
+		pr_warn("DCSS %s is in use and cannot be reloaded\n", name);
+		rc = -EAGAIN;
+		goto out_unlock;
+	}
+	release_resource(seg->res);
+	if (do_nonshared)
+		seg->res->flags &= ~IORESOURCE_READONLY;
+	else
+		if (seg->vm_segtype == SEG_TYPE_SR ||
+		    seg->vm_segtype == SEG_TYPE_ER)
+			seg->res->flags |= IORESOURCE_READONLY;
+
+	if (request_resource(&iomem_resource, seg->res)) {
+		pr_warn("DCSS %s overlaps with used memory resources and cannot be reloaded\n",
+			name);
+		rc = -EBUSY;
+		kfree(seg->res);
+		goto out_del_mem;
+	}
+
+	dcss_diag(&purgeseg_scode, seg->dcss_name, &dummy, &dummy);
+	if (do_nonshared)
+		diag_cc = dcss_diag(&loadnsr_scode, seg->dcss_name,
+				&start_addr, &end_addr);
+	else
+		diag_cc = dcss_diag(&loadshr_scode, seg->dcss_name,
+				&start_addr, &end_addr);
+	if (diag_cc < 0) {
+		rc = diag_cc;
+		goto out_del_res;
+	}
+	if (diag_cc > 1) {
+		pr_warn("Reloading DCSS %s failed with rc=%ld\n",
+			name, end_addr);
+		rc = dcss_diag_translate_rc(end_addr);
+		goto out_del_res;
+	}
+	seg->start_addr = start_addr;
+	seg->end = end_addr;
+	seg->do_nonshared = do_nonshared;
+	rc = 0;
+	goto out_unlock;
+ out_del_res:
+	release_resource(seg->res);
+	kfree(seg->res);
+ out_del_mem:
+	vmem_remove_mapping(seg->start_addr, seg->end - seg->start_addr + 1);
+	list_del(&seg->list);
+	dcss_diag(&purgeseg_scode, seg->dcss_name, &dummy, &dummy);
+	kfree(seg);
+ out_unlock:
+	mutex_unlock(&dcss_lock);
+	return rc;
+}
+
+static void __dcss_diag_purge_on_cpu_0(void *data)
+{
+	struct dcss_segment *seg = (struct dcss_segment *)data;
+	unsigned long dummy;
+
+	dcss_diag(&purgeseg_scode, seg->dcss_name, &dummy, &dummy);
+}
+
+/*
+ * Decrease the use count of a DCSS segment and remove
+ * it from the address space if nobody is using it
+ * any longer.
+ */
+void
+segment_unload(char *name)
+{
+	struct dcss_segment *seg;
+
+	if (!machine_is_vm())
+		return;
+
+	mutex_lock(&dcss_lock);
+	seg = segment_by_name (name);
+	if (seg == NULL) {
+		pr_err("Unloading unknown DCSS %s failed\n", name);
+		goto out_unlock;
+	}
+	if (!refcount_dec_and_test(&seg->ref_count))
+		goto out_unlock;
+	release_resource(seg->res);
+	kfree(seg->res);
+	vmem_remove_mapping(seg->start_addr, seg->end - seg->start_addr + 1);
+	list_del(&seg->list);
+	/*
+	 * Workaround for z/VM issue, where calling the DCSS unload diag on
+	 * a non-IPL CPU would cause bogus sclp maximum memory detection on
+	 * next IPL.
+	 * IPL CPU 0 cannot be set offline, so the dcss_diag() call can
+	 * directly be scheduled to that CPU.
+	 */
+	smp_call_function_single(0, __dcss_diag_purge_on_cpu_0, seg, 1);
+	kfree(seg);
+out_unlock:
+	mutex_unlock(&dcss_lock);
+}
+
+/*
+ * save segment content permanently
+ */
+void
+segment_save(char *name)
+{
+	struct dcss_segment *seg;
+	char cmd1[160];
+	char cmd2[80];
+	int i, response;
+
+	if (!machine_is_vm())
+		return;
+
+	mutex_lock(&dcss_lock);
+	seg = segment_by_name (name);
+
+	if (seg == NULL) {
+		pr_err("Saving unknown DCSS %s failed\n", name);
+		goto out;
+	}
+
+	snprintf(cmd1, sizeof(cmd1), "DEFSEG %s", name);
+	for (i=0; i<seg->segcnt; i++) {
+		size_t len = strlen(cmd1);
+
+		snprintf(cmd1 + len, sizeof(cmd1) - len, " %lX-%lX %s",
+			 seg->range[i].start >> PAGE_SHIFT,
+			 seg->range[i].end >> PAGE_SHIFT,
+			 segtype_string[seg->range[i].start & 0xff]);
+	}
+	snprintf(cmd2, sizeof(cmd2), "SAVESEG %s", name);
+	response = 0;
+	cpcmd(cmd1, NULL, 0, &response);
+	if (response) {
+		pr_err("Saving a DCSS failed with DEFSEG response code "
+		       "%i\n", response);
+		goto out;
+	}
+	cpcmd(cmd2, NULL, 0, &response);
+	if (response) {
+		pr_err("Saving a DCSS failed with SAVESEG response code "
+		       "%i\n", response);
+		goto out;
+	}
+out:
+	mutex_unlock(&dcss_lock);
+}
+
+/*
+ * print appropriate error message for segment_load()/segment_type()
+ * return code
+ */
+void segment_warning(int rc, char *seg_name)
+{
+	switch (rc) {
+	case -ENOENT:
+		pr_err("DCSS %s cannot be loaded or queried\n", seg_name);
+		break;
+	case -ENOSYS:
+		pr_err("DCSS %s cannot be loaded or queried without "
+		       "z/VM\n", seg_name);
+		break;
+	case -EIO:
+		pr_err("Loading or querying DCSS %s resulted in a "
+		       "hardware error\n", seg_name);
+		break;
+	case -EOPNOTSUPP:
+		pr_err("DCSS %s has multiple page ranges and cannot be "
+		       "loaded or queried\n", seg_name);
+		break;
+	case -EBUSY:
+		pr_err("%s needs used memory resources and cannot be "
+		       "loaded or queried\n", seg_name);
+		break;
+	case -EPERM:
+		pr_err("DCSS %s is already loaded in a different access "
+		       "mode\n", seg_name);
+		break;
+	case -ENOMEM:
+		pr_err("There is not enough memory to load or query "
+		       "DCSS %s\n", seg_name);
+		break;
+	case -ERANGE: {
+		struct range mhp_range = arch_get_mappable_range();
+
+		pr_err("DCSS %s exceeds the kernel mapping range (%llu) "
+		       "and cannot be loaded\n", seg_name, mhp_range.end + 1);
+		break;
+	}
+	default:
+		break;
+	}
+}
+
+EXPORT_SYMBOL(segment_load);
+EXPORT_SYMBOL(segment_unload);
+EXPORT_SYMBOL(segment_save);
+EXPORT_SYMBOL(segment_type);
+EXPORT_SYMBOL(segment_modify_shared);
+EXPORT_SYMBOL(segment_warning);

linux-7.2.3/arch/s390/mm/fault.c

@@ -0,0 +1,480 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  S390 version
+ *    Copyright IBM Corp. 1999
+ *    Author(s): Hartmut Penner (hp@de.ibm.com)
+ *		 Ulrich Weigand (uweigand@de.ibm.com)
+ *
+ *  Derived from "arch/i386/mm/fault.c"
+ *    Copyright (C) 1995  Linus Torvalds
+ */
+
+#include <linux/kernel_stat.h>
+#include <linux/mmu_context.h>
+#include <linux/cpufeature.h>
+#include <linux/perf_event.h>
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/sched/debug.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/kdebug.h>
+#include <linux/init.h>
+#include <linux/console.h>
+#include <linux/extable.h>
+#include <linux/hardirq.h>
+#include <linux/kprobes.h>
+#include <linux/uaccess.h>
+#include <linux/hugetlb.h>
+#include <linux/kfence.h>
+#include <linux/pagewalk.h>
+#include <asm/asm-extable.h>
+#include <asm/asm-offsets.h>
+#include <asm/ptrace.h>
+#include <asm/fault.h>
+#include <asm/diag.h>
+#include <asm/irq.h>
+#include <asm/facility.h>
+#include <asm/uv.h>
+#include "../kernel/entry.h"
+
+/*
+ * Find out which address space caused the exception.
+ */
+static bool is_kernel_fault(struct pt_regs *regs)
+{
+	union teid teid = { .val = regs->int_parm_long };
+
+	if (user_mode(regs))
+		return false;
+	if (teid.as == PSW_BITS_AS_SECONDARY)
+		return false;
+	return true;
+}
+
+static unsigned long get_fault_address(struct pt_regs *regs)
+{
+	union teid teid = { .val = regs->int_parm_long };
+
+	return teid.addr * PAGE_SIZE;
+}
+
+static __always_inline bool fault_is_write(struct pt_regs *regs)
+{
+	union teid teid = { .val = regs->int_parm_long };
+
+	if (test_facility(75))
+		return teid.fsi == TEID_FSI_STORE;
+	return false;
+}
+
+static void dump_pagetable(unsigned long asce, unsigned long address)
+{
+	unsigned long entry, *table = __va(asce & _ASCE_ORIGIN);
+
+	pr_alert("AS:%016lx ", asce);
+	switch (asce & _ASCE_TYPE_MASK) {
+	case _ASCE_TYPE_REGION1:
+		table += (address & _REGION1_INDEX) >> _REGION1_SHIFT;
+		if (get_kernel_nofault(entry, table))
+			goto bad;
+		pr_cont("R1:%016lx ", entry);
+		if (entry & _REGION_ENTRY_INVALID)
+			goto out;
+		table = __va(entry & _REGION_ENTRY_ORIGIN);
+		fallthrough;
+	case _ASCE_TYPE_REGION2:
+		table += (address & _REGION2_INDEX) >> _REGION2_SHIFT;
+		if (get_kernel_nofault(entry, table))
+			goto bad;
+		pr_cont("R2:%016lx ", entry);
+		if (entry & _REGION_ENTRY_INVALID)
+			goto out;
+		table = __va(entry & _REGION_ENTRY_ORIGIN);
+		fallthrough;
+	case _ASCE_TYPE_REGION3:
+		table += (address & _REGION3_INDEX) >> _REGION3_SHIFT;
+		if (get_kernel_nofault(entry, table))
+			goto bad;
+		pr_cont("R3:%016lx ", entry);
+		if (entry & (_REGION_ENTRY_INVALID | _REGION3_ENTRY_LARGE))
+			goto out;
+		table = __va(entry & _REGION_ENTRY_ORIGIN);
+		fallthrough;
+	case _ASCE_TYPE_SEGMENT:
+		table += (address & _SEGMENT_INDEX) >> _SEGMENT_SHIFT;
+		if (get_kernel_nofault(entry, table))
+			goto bad;
+		pr_cont("S:%016lx ", entry);
+		if (entry & (_SEGMENT_ENTRY_INVALID | _SEGMENT_ENTRY_LARGE))
+			goto out;
+		table = __va(entry & _SEGMENT_ENTRY_ORIGIN);
+	}
+	table += (address & _PAGE_INDEX) >> PAGE_SHIFT;
+	if (get_kernel_nofault(entry, table))
+		goto bad;
+	pr_cont("P:%016lx ", entry);
+out:
+	pr_cont("\n");
+	return;
+bad:
+	pr_cont("BAD\n");
+}
+
+static void dump_fault_info(struct pt_regs *regs)
+{
+	union teid teid = { .val = regs->int_parm_long };
+	unsigned long asce;
+
+	pr_alert("Failing address: %016lx TEID: %016lx",
+		 get_fault_address(regs), teid.val);
+	if (test_facility(131))
+		pr_cont(" ESOP-2");
+	else if (machine_has_esop())
+		pr_cont(" ESOP-1");
+	else
+		pr_cont(" SOP");
+	if (test_facility(75))
+		pr_cont(" FSI");
+	pr_cont("\n");
+	pr_alert("Fault in ");
+	switch (teid.as) {
+	case PSW_BITS_AS_HOME:
+		pr_cont("home space ");
+		break;
+	case PSW_BITS_AS_SECONDARY:
+		pr_cont("secondary space ");
+		break;
+	case PSW_BITS_AS_ACCREG:
+		pr_cont("access register ");
+		break;
+	case PSW_BITS_AS_PRIMARY:
+		pr_cont("primary space ");
+		break;
+	}
+	pr_cont("mode while using ");
+	if (is_kernel_fault(regs)) {
+		asce = get_lowcore()->kernel_asce.val;
+		pr_cont("kernel ");
+	} else {
+		asce = get_lowcore()->user_asce.val;
+		pr_cont("user ");
+	}
+	pr_cont("ASCE.\n");
+	dump_pagetable(asce, get_fault_address(regs));
+}
+
+int show_unhandled_signals = 1;
+
+static const struct ctl_table s390_fault_sysctl_table[] = {
+	{
+		.procname	= "userprocess_debug",
+		.data		= &show_unhandled_signals,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec,
+	},
+};
+
+static int __init init_s390_fault_sysctls(void)
+{
+	register_sysctl_init("kernel", s390_fault_sysctl_table);
+	return 0;
+}
+arch_initcall(init_s390_fault_sysctls);
+
+void report_user_fault(struct pt_regs *regs, long signr, int is_mm_fault)
+{
+	static DEFINE_RATELIMIT_STATE(rs, DEFAULT_RATELIMIT_INTERVAL, DEFAULT_RATELIMIT_BURST);
+
+	if ((task_pid_nr(current) > 1) && !show_unhandled_signals)
+		return;
+	if (!unhandled_signal(current, signr))
+		return;
+	if (!__ratelimit(&rs))
+		return;
+	pr_alert("User process fault: interruption code %04x ilc:%d ",
+		 regs->int_code & 0xffff, regs->int_code >> 17);
+	print_vma_addr(KERN_CONT "in ", regs->psw.addr);
+	pr_cont("\n");
+	if (is_mm_fault)
+		dump_fault_info(regs);
+	show_regs(regs);
+}
+
+static void do_sigsegv(struct pt_regs *regs, int si_code)
+{
+	report_user_fault(regs, SIGSEGV, 1);
+	force_sig_fault(SIGSEGV, si_code, (void __user *)get_fault_address(regs));
+}
+
+static void handle_fault_error_nolock(struct pt_regs *regs, int si_code)
+{
+	unsigned long address;
+	bool is_write;
+
+	if (user_mode(regs)) {
+		if (WARN_ON_ONCE(!si_code))
+			si_code = SEGV_MAPERR;
+		return do_sigsegv(regs, si_code);
+	}
+	if (fixup_exception(regs))
+		return;
+	if (is_kernel_fault(regs)) {
+		address = get_fault_address(regs);
+		is_write = fault_is_write(regs);
+		if (kfence_handle_page_fault(address, is_write, regs))
+			return;
+		pr_alert("Unable to handle kernel pointer dereference in virtual kernel address space\n");
+	} else {
+		pr_alert("Unable to handle kernel paging request in virtual user address space\n");
+	}
+	dump_fault_info(regs);
+	die(regs, "Oops");
+}
+
+static void handle_fault_error(struct pt_regs *regs, int si_code)
+{
+	struct mm_struct *mm = current->mm;
+
+	mmap_read_unlock(mm);
+	handle_fault_error_nolock(regs, si_code);
+}
+
+static void do_sigbus(struct pt_regs *regs)
+{
+	force_sig_fault(SIGBUS, BUS_ADRERR, (void __user *)get_fault_address(regs));
+}
+
+/*
+ * This routine handles page faults.  It determines the address,
+ * and the problem, and then passes it off to one of the appropriate
+ * routines.
+ *
+ * interruption code (int_code):
+ *   04       Protection	   ->  Write-Protection  (suppression)
+ *   10       Segment translation  ->  Not present	 (nullification)
+ *   11       Page translation	   ->  Not present	 (nullification)
+ *   3b       Region third trans.  ->  Not present	 (nullification)
+ */
+static void do_exception(struct pt_regs *regs, int access)
+{
+	struct vm_area_struct *vma;
+	unsigned long address;
+	struct mm_struct *mm;
+	unsigned int flags;
+	vm_fault_t fault;
+	bool is_write;
+
+	/*
+	 * The instruction that caused the program check has
+	 * been nullified. Don't signal single step via SIGTRAP.
+	 */
+	clear_thread_flag(TIF_PER_TRAP);
+	if (kprobe_page_fault(regs, 14))
+		return;
+	mm = current->mm;
+	address = get_fault_address(regs);
+	is_write = fault_is_write(regs);
+	if (is_kernel_fault(regs) || faulthandler_disabled() || !mm)
+		return handle_fault_error_nolock(regs, 0);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
+	flags = FAULT_FLAG_DEFAULT;
+	if (user_mode(regs))
+		flags |= FAULT_FLAG_USER;
+	if (is_write)
+		access = VM_WRITE;
+	if (access == VM_WRITE)
+		flags |= FAULT_FLAG_WRITE;
+	if (!(flags & FAULT_FLAG_USER))
+		goto lock_mmap;
+	vma = lock_vma_under_rcu(mm, address);
+	if (!vma)
+		goto lock_mmap;
+	if (!(vma->vm_flags & access)) {
+		vma_end_read(vma);
+		count_vm_vma_lock_event(VMA_LOCK_SUCCESS);
+		return handle_fault_error_nolock(regs, SEGV_ACCERR);
+	}
+	fault = handle_mm_fault(vma, address, flags | FAULT_FLAG_VMA_LOCK, regs);
+	if (!(fault & (VM_FAULT_RETRY | VM_FAULT_COMPLETED)))
+		vma_end_read(vma);
+	if (!(fault & VM_FAULT_RETRY)) {
+		count_vm_vma_lock_event(VMA_LOCK_SUCCESS);
+		goto done;
+	}
+	count_vm_vma_lock_event(VMA_LOCK_RETRY);
+	if (fault & VM_FAULT_MAJOR)
+		flags |= FAULT_FLAG_TRIED;
+	/* Quick path to respond to signals */
+	if (fault_signal_pending(fault, regs)) {
+		if (!user_mode(regs))
+			handle_fault_error_nolock(regs, 0);
+		return;
+	}
+lock_mmap:
+retry:
+	vma = lock_mm_and_find_vma(mm, address, regs);
+	if (!vma)
+		return handle_fault_error_nolock(regs, SEGV_MAPERR);
+	if (unlikely(!(vma->vm_flags & access)))
+		return handle_fault_error(regs, SEGV_ACCERR);
+	fault = handle_mm_fault(vma, address, flags, regs);
+	if (fault_signal_pending(fault, regs)) {
+		if (!user_mode(regs))
+			handle_fault_error_nolock(regs, 0);
+		return;
+	}
+	/* The fault is fully completed (including releasing mmap lock) */
+	if (fault & VM_FAULT_COMPLETED)
+		return;
+	if (fault & VM_FAULT_RETRY) {
+		flags |= FAULT_FLAG_TRIED;
+		goto retry;
+	}
+	mmap_read_unlock(mm);
+done:
+	if (!(fault & VM_FAULT_ERROR))
+		return;
+	if (fault & VM_FAULT_OOM) {
+		if (!user_mode(regs))
+			handle_fault_error_nolock(regs, 0);
+		else
+			pagefault_out_of_memory();
+	} else if (fault & VM_FAULT_SIGSEGV) {
+		if (!user_mode(regs))
+			handle_fault_error_nolock(regs, 0);
+		else
+			do_sigsegv(regs, SEGV_MAPERR);
+	} else if (fault & (VM_FAULT_SIGBUS | VM_FAULT_HWPOISON |
+			    VM_FAULT_HWPOISON_LARGE)) {
+		if (!user_mode(regs))
+			handle_fault_error_nolock(regs, 0);
+		else
+			do_sigbus(regs);
+	} else {
+		pr_emerg("Unexpected fault flags: %08x\n", fault);
+		BUG();
+	}
+}
+
+void do_protection_exception(struct pt_regs *regs)
+{
+	union teid teid = { .val = regs->int_parm_long };
+
+	/*
+	 * Protection exceptions are suppressing, decrement psw address.
+	 * The exception to this rule are aborted transactions, for these
+	 * the PSW already points to the correct location.
+	 */
+	if (!(regs->int_code & 0x200)) {
+		regs->psw.addr = __rewind_psw(regs->psw, regs->int_code >> 16);
+		set_pt_regs_flag(regs, PIF_PSW_ADDR_ADJUSTED);
+	}
+	/*
+	 * If bit 61 if the TEID is not set, the remainder of the
+	 * TEID is unpredictable. Special handling is required.
+	 */
+	if (unlikely(!teid.b61)) {
+		if (user_mode(regs)) {
+			dump_fault_info(regs);
+			die(regs, "Unexpected TEID");
+		}
+		/* Assume low-address protection in kernel mode. */
+		return handle_fault_error_nolock(regs, 0);
+	}
+	if (unlikely(cpu_has_nx() && teid.b56)) {
+		regs->int_parm_long = (teid.addr * PAGE_SIZE) | (regs->psw.addr & PAGE_MASK);
+		return handle_fault_error_nolock(regs, SEGV_ACCERR);
+	}
+	do_exception(regs, VM_WRITE);
+}
+NOKPROBE_SYMBOL(do_protection_exception);
+
+void do_dat_exception(struct pt_regs *regs)
+{
+	do_exception(regs, VM_ACCESS_FLAGS);
+}
+NOKPROBE_SYMBOL(do_dat_exception);
+
+#if IS_ENABLED(CONFIG_KVM)
+
+void do_secure_storage_access(struct pt_regs *regs)
+{
+	union teid teid = { .val = regs->int_parm_long };
+	unsigned long addr = get_fault_address(regs);
+	struct vm_area_struct *vma;
+	struct folio_walk fw;
+	struct mm_struct *mm;
+	struct folio *folio;
+	int rc;
+
+	/*
+	 * Bit 61 indicates if the address is valid, if it is not the
+	 * kernel should be stopped or SIGSEGV should be sent to the
+	 * process. Bit 61 is not reliable without the misc UV feature,
+	 * therefore this needs to be checked too.
+	 */
+	if (uv_has_feature(BIT_UV_FEAT_MISC) && !teid.b61) {
+		/*
+		 * When this happens, userspace did something that it
+		 * was not supposed to do, e.g. branching into secure
+		 * memory. Trigger a segmentation fault.
+		 */
+		if (user_mode(regs)) {
+			send_sig(SIGSEGV, current, 0);
+			return;
+		}
+		/*
+		 * The kernel should never run into this case and
+		 * there is no way out of this situation.
+		 */
+		panic("Unexpected PGM 0x3d with TEID bit 61=0");
+	}
+	if (is_kernel_fault(regs)) {
+		folio = virt_to_folio((void *)addr);
+		if (unlikely(!folio_try_get(folio)))
+			return;
+		rc = uv_convert_from_secure(folio_to_phys(folio));
+		if (!rc)
+			clear_bit(PG_arch_1, &folio->flags.f);
+		folio_put(folio);
+		/*
+		 * There are some valid fixup types for kernel
+		 * accesses to donated secure memory. zeropad is one
+		 * of them.
+		 */
+		if (rc)
+			return handle_fault_error_nolock(regs, 0);
+	} else {
+		if (faulthandler_disabled())
+			return handle_fault_error_nolock(regs, 0);
+		mm = current->mm;
+		mmap_read_lock(mm);
+		vma = find_vma(mm, addr);
+		if (!vma)
+			return handle_fault_error(regs, SEGV_MAPERR);
+		folio = folio_walk_start(&fw, vma, addr, 0);
+		if (!folio) {
+			mmap_read_unlock(mm);
+			return;
+		}
+		/* arch_make_folio_accessible() needs a raised refcount. */
+		folio_get(folio);
+		rc = arch_make_folio_accessible(folio);
+		folio_put(folio);
+		folio_walk_end(&fw, vma);
+		if (rc)
+			send_sig(SIGSEGV, current, 0);
+		mmap_read_unlock(mm);
+	}
+}
+NOKPROBE_SYMBOL(do_secure_storage_access);
+
+#endif /* CONFIG_KVM */

linux-7.2.3/arch/s390/mm/gmap_helpers.c

@@ -0,0 +1,317 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  Helper functions for KVM guest address space mapping code
+ *
+ *    Copyright IBM Corp. 2007, 2025
+ */
+
+#include <linux/export.h>
+#include <linux/mm_types.h>
+#include <linux/mmap_lock.h>
+#include <linux/mm.h>
+#include <linux/hugetlb.h>
+#include <linux/swap.h>
+#include <linux/leafops.h>
+#include <linux/pagewalk.h>
+#include <linux/ksm.h>
+#include <asm/gmap_helpers.h>
+
+/**
+ * try_get_locked_pte() - like get_locked_pte(), but atomic and with trylock
+ * @mm: the mm
+ * @vmaddr: the userspace virtual address whose pte is to be found
+ * @ptl: will be set to the pointer to the lock used to lock the pte in case
+ *       of success.
+ *
+ * This function returns the pointer to the pte corresponding to @addr in @mm,
+ * similarly to get_locked_pte(). Unlike get_locked_pte(), no attempt is made
+ * to allocate missing page tables. If a missing or large entry is found, the
+ * function will return NULL. If the ptl lock is contended, %-EAGAIN is
+ * returned.
+ *
+ * In case of success, *@ptl will point to the locked pte lock for the returned
+ * pte, like get_locked_pte() does.
+ *
+ * Context: mmap_lock or vma lock for read or for write needs to be held.
+ * Return:
+ * * %NULL if the pte cannot be reached.
+ * * %-EAGAIN if the pte can be reached, but cannot be locked.
+ * * the pointer to the pte corresponding to @addr in @mm, if it can be reached
+ *   and locked.
+ */
+pte_t *try_get_locked_pte(struct mm_struct *mm, unsigned long vmaddr, spinlock_t **ptl)
+{
+	pmd_t *pmdp, pmd, pmdval;
+	pud_t *pudp, pud;
+	p4d_t *p4dp, p4d;
+	pgd_t *pgdp, pgd;
+	pte_t *ptep;
+
+	pgdp = pgd_offset(mm, vmaddr);
+	pgd = pgdp_get(pgdp);
+	if (pgd_none(pgd) || !pgd_present(pgd))
+		return NULL;
+	p4dp = p4d_offset_lockless(pgdp, pgd, vmaddr);
+	p4d = p4dp_get(p4dp);
+	if (p4d_none(p4d) || !p4d_present(p4d))
+		return NULL;
+	pudp = pud_offset_lockless(p4dp, p4d, vmaddr);
+	pud = pudp_get(pudp);
+	if (pud_none(pud) || pud_leaf(pud) || !pud_present(pud))
+		return NULL;
+	pmdp = pmd_offset_lockless(pudp, pud, vmaddr);
+	pmd = pmdp_get_lockless(pmdp);
+	if (pmd_none(pmd) || pmd_leaf(pmd) || !pmd_present(pmd))
+		return NULL;
+	ptep = pte_offset_map_rw_nolock(mm, pmdp, vmaddr, &pmdval, ptl);
+	if (!ptep)
+		return NULL;
+
+	if (spin_trylock(*ptl)) {
+		if (unlikely(!pmd_same(pmdval, pmdp_get_lockless(pmdp)))) {
+			pte_unmap_unlock(ptep, *ptl);
+			return ERR_PTR(-EAGAIN);
+		}
+		return ptep;
+	}
+
+	pte_unmap(ptep);
+	return ERR_PTR(-EAGAIN);
+}
+EXPORT_SYMBOL_GPL(try_get_locked_pte);
+
+/**
+ * gmap_helper_zap_one_page() - discard a page if it was swapped.
+ * @mm: the mm
+ * @vmaddr: the userspace virtual address that needs to be discarded
+ *
+ * If the given address maps to a swap entry, discard it.
+ *
+ * Context: needs to be called while holding the mmap lock.
+ */
+void gmap_helper_zap_one_page(struct mm_struct *mm, unsigned long vmaddr)
+{
+	struct vm_area_struct *vma;
+	spinlock_t *ptl;	/* Lock for the host (userspace) page table */
+	softleaf_t sl;
+	pte_t *ptep;
+
+	mmap_assert_locked(mm);
+
+	/* Find the vm address for the guest address */
+	vma = vma_lookup(mm, vmaddr);
+	if (!vma || is_vm_hugetlb_page(vma))
+		return;
+
+	/* Get pointer to the page table entry */
+	ptep = try_get_locked_pte(mm, vmaddr, &ptl);
+	if (IS_ERR_OR_NULL(ptep))
+		return;
+	sl = softleaf_from_pte(*ptep);
+	if (pte_swap(*ptep) && softleaf_is_swap(sl)) {
+		dec_mm_counter(mm, MM_SWAPENTS);
+		swap_put_entries_direct(sl, 1);
+		pte_clear(mm, vmaddr, ptep);
+	}
+	pte_unmap_unlock(ptep, ptl);
+}
+EXPORT_SYMBOL_GPL(gmap_helper_zap_one_page);
+
+/**
+ * gmap_helper_discard() - discard user pages in the given range
+ * @mm: the mm
+ * @vmaddr: starting userspace address
+ * @end: end address (first address outside the range)
+ *
+ * All userpace pages in the range [@vamddr, @end) are discarded and unmapped.
+ *
+ * Context: needs to be called while holding the mmap lock.
+ */
+void gmap_helper_discard(struct mm_struct *mm, unsigned long vmaddr, unsigned long end)
+{
+	struct vm_area_struct *vma;
+
+	mmap_assert_locked(mm);
+
+	while (vmaddr < end) {
+		vma = find_vma_intersection(mm, vmaddr, end);
+		if (!vma)
+			return;
+		if (!is_vm_hugetlb_page(vma))
+			zap_vma_range(vma, vmaddr, min(end, vma->vm_end) - vmaddr);
+		vmaddr = vma->vm_end;
+	}
+}
+EXPORT_SYMBOL_GPL(gmap_helper_discard);
+
+/**
+ * gmap_helper_try_set_pte_unused() - mark a pte entry as unused
+ * @mm: the mm
+ * @vmaddr: the userspace address whose pte is to be marked
+ *
+ * Mark the pte corresponding the given address as unused. This will cause
+ * core mm code to just drop this page instead of swapping it.
+ *
+ * This function needs to be called with interrupts disabled (for example
+ * while holding a spinlock), or while holding the mmap lock. Normally this
+ * function is called as a result of an unmap operation, and thus KVM common
+ * code will already hold kvm->mmu_lock in write mode.
+ *
+ * Context: Needs to be called while holding the mmap lock or with interrupts
+ *          disabled.
+ */
+void gmap_helper_try_set_pte_unused(struct mm_struct *mm, unsigned long vmaddr)
+{
+	spinlock_t *ptl;	/* Lock for the host (userspace) page table */
+	pte_t *ptep;
+
+	/*
+	 * Several paths exists that takes the ptl lock and then call the
+	 * mmu_notifier, which takes the mmu_lock. The unmap path, instead,
+	 * takes the mmu_lock in write mode first, and then potentially
+	 * calls this function, which takes the ptl lock. This can lead to a
+	 * deadlock.
+	 * The unused page mechanism is only an optimization, if the
+	 * _PAGE_UNUSED bit is not set, the unused page is swapped as normal
+	 * instead of being discarded.
+	 * If the lock is contended the bit is not set and the deadlock is
+	 * avoided.
+	 */
+	ptep = try_get_locked_pte(mm, vmaddr, &ptl);
+	if (IS_ERR_OR_NULL(ptep))
+		return;
+
+	if (pte_present(*ptep))
+		__atomic64_or(_PAGE_UNUSED, (long *)ptep);
+	pte_unmap_unlock(ptep, ptl);
+}
+EXPORT_SYMBOL_GPL(gmap_helper_try_set_pte_unused);
+
+static int find_zeropage_pte_entry(pte_t *pte, unsigned long addr,
+				   unsigned long end, struct mm_walk *walk)
+{
+	unsigned long *found_addr = walk->private;
+
+	/* Return 1 of the page is a zeropage. */
+	if (is_zero_pfn(pte_pfn(*pte))) {
+		/*
+		 * Shared zeropage in e.g., a FS DAX mapping? We cannot do the
+		 * right thing and likely don't care: FAULT_FLAG_UNSHARE
+		 * currently only works in COW mappings, which is also where
+		 * mm_forbids_zeropage() is checked.
+		 */
+		if (!is_cow_mapping(walk->vma->vm_flags))
+			return -EFAULT;
+
+		*found_addr = addr;
+		return 1;
+	}
+	return 0;
+}
+
+static const struct mm_walk_ops find_zeropage_ops = {
+	.pte_entry      = find_zeropage_pte_entry,
+	.walk_lock      = PGWALK_WRLOCK,
+};
+
+/** __gmap_helper_unshare_zeropages() - unshare all shared zeropages
+ * @mm: the mm whose zero pages are to be unshared
+ *
+ * Unshare all shared zeropages, replacing them by anonymous pages. Note that
+ * we cannot simply zap all shared zeropages, because this could later
+ * trigger unexpected userfaultfd missing events.
+ *
+ * This must be called after mm->context.allow_cow_sharing was
+ * set to 0, to avoid future mappings of shared zeropages.
+ *
+ * mm contracts with s390, that even if mm were to remove a page table,
+ * and racing with walk_page_range_vma() calling pte_offset_map_lock()
+ * would fail, it will never insert a page table containing empty zero
+ * pages once mm_forbids_zeropage(mm) i.e.
+ * mm->context.allow_cow_sharing is set to 0.
+ */
+static int __gmap_helper_unshare_zeropages(struct mm_struct *mm)
+{
+	struct vm_area_struct *vma;
+	VMA_ITERATOR(vmi, mm, 0);
+	unsigned long addr;
+	vm_fault_t fault;
+	int rc;
+
+	for_each_vma(vmi, vma) {
+		/*
+		 * We could only look at COW mappings, but it's more future
+		 * proof to catch unexpected zeropages in other mappings and
+		 * fail.
+		 */
+		if ((vma->vm_flags & VM_PFNMAP) || is_vm_hugetlb_page(vma))
+			continue;
+		addr = vma->vm_start;
+
+retry:
+		rc = walk_page_range_vma(vma, addr, vma->vm_end,
+					 &find_zeropage_ops, &addr);
+		if (rc < 0)
+			return rc;
+		else if (!rc)
+			continue;
+
+		/* addr was updated by find_zeropage_pte_entry() */
+		fault = handle_mm_fault(vma, addr,
+					FAULT_FLAG_UNSHARE | FAULT_FLAG_REMOTE,
+					NULL);
+		if (fault & VM_FAULT_OOM)
+			return -ENOMEM;
+		/*
+		 * See break_ksm(): even after handle_mm_fault() returned 0, we
+		 * must start the lookup from the current address, because
+		 * handle_mm_fault() may back out if there's any difficulty.
+		 *
+		 * VM_FAULT_SIGBUS and VM_FAULT_SIGSEGV are unexpected but
+		 * maybe they could trigger in the future on concurrent
+		 * truncation. In that case, the shared zeropage would be gone
+		 * and we can simply retry and make progress.
+		 */
+		cond_resched();
+		goto retry;
+	}
+
+	return 0;
+}
+
+/**
+ * gmap_helper_disable_cow_sharing() - disable all COW sharing
+ *
+ * Disable most COW-sharing of memory pages for the whole process:
+ * (1) Disable KSM and unmerge/unshare any KSM pages.
+ * (2) Disallow shared zeropages and unshare any zerpages that are mapped.
+ *
+ * Not that we currently don't bother with COW-shared pages that are shared
+ * with parent/child processes due to fork().
+ */
+int gmap_helper_disable_cow_sharing(void)
+{
+	struct mm_struct *mm = current->mm;
+	int rc;
+
+	mmap_assert_write_locked(mm);
+
+	if (!mm->context.allow_cow_sharing)
+		return 0;
+
+	mm->context.allow_cow_sharing = 0;
+
+	/* Replace all shared zeropages by anonymous pages. */
+	rc = __gmap_helper_unshare_zeropages(mm);
+	/*
+	 * Make sure to disable KSM (if enabled for the whole process or
+	 * individual VMAs). Note that nothing currently hinders user space
+	 * from re-enabling it.
+	 */
+	if (!rc)
+		rc = ksm_disable(mm);
+	if (rc)
+		mm->context.allow_cow_sharing = 1;
+	return rc;
+}
+EXPORT_SYMBOL_GPL(gmap_helper_disable_cow_sharing);

linux-7.2.3/arch/s390/mm/hugetlbpage.c

@@ -0,0 +1,241 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  IBM System z Huge TLB Page Support for Kernel.
+ *
+ *    Copyright IBM Corp. 2007,2020
+ *    Author(s): Gerald Schaefer <gerald.schaefer@de.ibm.com>
+ */
+
+#define pr_fmt(fmt) "hugetlb: " fmt
+
+#include <linux/cpufeature.h>
+#include <linux/mm.h>
+#include <linux/hugetlb.h>
+#include <linux/mman.h>
+#include <linux/sched/mm.h>
+#include <linux/security.h>
+#include <asm/pgalloc.h>
+
+/*
+ * If the bit selected by single-bit bitmask "a" is set within "x", move
+ * it to the position indicated by single-bit bitmask "b".
+ */
+#define move_set_bit(x, a, b)	(((x) & (a)) >> ilog2(a) << ilog2(b))
+
+static inline unsigned long __pte_to_rste(pte_t pte)
+{
+	swp_entry_t arch_entry;
+	unsigned long rste;
+
+	/*
+	 * Convert encoding		  pte bits	pmd / pud bits
+	 *				lIR.uswrdy.p	dy..R...I...wr
+	 * empty			010.000000.0 -> 00..0...1...00
+	 * prot-none, clean, old	111.000000.1 -> 00..1...1...00
+	 * prot-none, clean, young	111.000001.1 -> 01..1...1...00
+	 * prot-none, dirty, old	111.000010.1 -> 10..1...1...00
+	 * prot-none, dirty, young	111.000011.1 -> 11..1...1...00
+	 * read-only, clean, old	111.000100.1 -> 00..1...1...01
+	 * read-only, clean, young	101.000101.1 -> 01..1...0...01
+	 * read-only, dirty, old	111.000110.1 -> 10..1...1...01
+	 * read-only, dirty, young	101.000111.1 -> 11..1...0...01
+	 * read-write, clean, old	111.001100.1 -> 00..1...1...11
+	 * read-write, clean, young	101.001101.1 -> 01..1...0...11
+	 * read-write, dirty, old	110.001110.1 -> 10..0...1...11
+	 * read-write, dirty, young	100.001111.1 -> 11..0...0...11
+	 * HW-bits: R read-only, I invalid
+	 * SW-bits: p present, y young, d dirty, r read, w write, s special,
+	 *	    u unused, l large
+	 */
+	if (pte_present(pte)) {
+		rste = pte_val(pte) & PAGE_MASK;
+		rste |= _SEGMENT_ENTRY_PRESENT;
+		rste |= move_set_bit(pte_val(pte), _PAGE_READ,
+				     _SEGMENT_ENTRY_READ);
+		rste |= move_set_bit(pte_val(pte), _PAGE_WRITE,
+				     _SEGMENT_ENTRY_WRITE);
+		rste |= move_set_bit(pte_val(pte), _PAGE_INVALID,
+				     _SEGMENT_ENTRY_INVALID);
+		rste |= move_set_bit(pte_val(pte), _PAGE_PROTECT,
+				     _SEGMENT_ENTRY_PROTECT);
+		rste |= move_set_bit(pte_val(pte), _PAGE_DIRTY,
+				     _SEGMENT_ENTRY_DIRTY);
+		rste |= move_set_bit(pte_val(pte), _PAGE_YOUNG,
+				     _SEGMENT_ENTRY_YOUNG);
+#ifdef CONFIG_MEM_SOFT_DIRTY
+		rste |= move_set_bit(pte_val(pte), _PAGE_SOFT_DIRTY,
+				     _SEGMENT_ENTRY_SOFT_DIRTY);
+#endif
+		rste |= move_set_bit(pte_val(pte), _PAGE_NOEXEC,
+				     _SEGMENT_ENTRY_NOEXEC);
+	} else if (!pte_none(pte)) {
+		/* swap pte */
+		arch_entry = __pte_to_swp_entry(pte);
+		rste = mk_swap_rste(__swp_type(arch_entry), __swp_offset(arch_entry));
+	} else
+		rste = _SEGMENT_ENTRY_EMPTY;
+	return rste;
+}
+
+static inline pte_t __rste_to_pte(unsigned long rste)
+{
+	swp_entry_t arch_entry;
+	unsigned long pteval;
+	int present, none;
+	pte_t pte;
+
+	if ((rste & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3) {
+		present = pud_present(__pud(rste));
+		none = pud_none(__pud(rste));
+	} else {
+		present = pmd_present(__pmd(rste));
+		none = pmd_none(__pmd(rste));
+	}
+
+	/*
+	 * Convert encoding		pmd / pud bits	    pte bits
+	 *				dy..R...I...wr	  lIR.uswrdy.p
+	 * empty			00..0...1...00 -> 010.000000.0
+	 * prot-none, clean, old	00..1...1...00 -> 111.000000.1
+	 * prot-none, clean, young	01..1...1...00 -> 111.000001.1
+	 * prot-none, dirty, old	10..1...1...00 -> 111.000010.1
+	 * prot-none, dirty, young	11..1...1...00 -> 111.000011.1
+	 * read-only, clean, old	00..1...1...01 -> 111.000100.1
+	 * read-only, clean, young	01..1...0...01 -> 101.000101.1
+	 * read-only, dirty, old	10..1...1...01 -> 111.000110.1
+	 * read-only, dirty, young	11..1...0...01 -> 101.000111.1
+	 * read-write, clean, old	00..1...1...11 -> 111.001100.1
+	 * read-write, clean, young	01..1...0...11 -> 101.001101.1
+	 * read-write, dirty, old	10..0...1...11 -> 110.001110.1
+	 * read-write, dirty, young	11..0...0...11 -> 100.001111.1
+	 * HW-bits: R read-only, I invalid
+	 * SW-bits: p present, y young, d dirty, r read, w write, s special,
+	 *	    u unused, l large
+	 */
+	if (present) {
+		pteval = rste & _SEGMENT_ENTRY_ORIGIN_LARGE;
+		pteval |= _PAGE_LARGE | _PAGE_PRESENT;
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_READ, _PAGE_READ);
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_WRITE, _PAGE_WRITE);
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_INVALID, _PAGE_INVALID);
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_PROTECT, _PAGE_PROTECT);
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_DIRTY, _PAGE_DIRTY);
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_YOUNG, _PAGE_YOUNG);
+#ifdef CONFIG_MEM_SOFT_DIRTY
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_SOFT_DIRTY, _PAGE_SOFT_DIRTY);
+#endif
+		pteval |= move_set_bit(rste, _SEGMENT_ENTRY_NOEXEC, _PAGE_NOEXEC);
+	} else if (!none) {
+		/* swap rste */
+		arch_entry = __rste_to_swp_entry(rste);
+		pte = mk_swap_pte(__swp_type_rste(arch_entry), __swp_offset_rste(arch_entry));
+		pteval = pte_val(pte);
+	} else
+		pteval = _PAGE_INVALID;
+	return __pte(pteval);
+}
+
+void __set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
+		     pte_t *ptep, pte_t pte)
+{
+	unsigned long rste;
+
+	rste = __pte_to_rste(pte);
+
+	/* Set correct table type for 2G hugepages */
+	if ((pte_val(ptep_get(ptep)) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3) {
+		if (likely(pte_present(pte)))
+			rste |= _REGION3_ENTRY_LARGE;
+		rste |= _REGION_ENTRY_TYPE_R3;
+	} else if (likely(pte_present(pte)))
+		rste |= _SEGMENT_ENTRY_LARGE;
+
+	set_pte(ptep, __pte(rste));
+}
+
+void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
+		     pte_t *ptep, pte_t pte, unsigned long sz)
+{
+	__set_huge_pte_at(mm, addr, ptep, pte);
+}
+
+pte_t huge_ptep_get(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+{
+	return __rste_to_pte(pte_val(ptep_get(ptep)));
+}
+
+pte_t __huge_ptep_get_and_clear(struct mm_struct *mm,
+				unsigned long addr, pte_t *ptep)
+{
+	pte_t pte = huge_ptep_get(mm, addr, ptep);
+	pmd_t *pmdp = (pmd_t *) ptep;
+	pud_t *pudp = (pud_t *) ptep;
+
+	if ((pte_val(ptep_get(ptep)) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3)
+		pudp_xchg_direct(mm, addr, pudp, __pud(_REGION3_ENTRY_EMPTY));
+	else
+		pmdp_xchg_direct(mm, addr, pmdp, __pmd(_SEGMENT_ENTRY_EMPTY));
+	return pte;
+}
+
+pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma,
+			unsigned long addr, unsigned long sz)
+{
+	pgd_t *pgdp;
+	p4d_t *p4dp;
+	pud_t *pudp;
+	pmd_t *pmdp = NULL;
+
+	pgdp = pgd_offset(mm, addr);
+	p4dp = p4d_alloc(mm, pgdp, addr);
+	if (p4dp) {
+		pudp = pud_alloc(mm, p4dp, addr);
+		if (pudp) {
+			if (sz == PUD_SIZE)
+				return (pte_t *) pudp;
+			else if (sz == PMD_SIZE)
+				pmdp = pmd_alloc(mm, pudp, addr);
+		}
+	}
+	return (pte_t *) pmdp;
+}
+
+pte_t *huge_pte_offset(struct mm_struct *mm,
+		       unsigned long addr, unsigned long sz)
+{
+	pgd_t *pgdp;
+	p4d_t *p4dp;
+	pud_t *pudp;
+	pmd_t *pmdp = NULL;
+
+	pgdp = pgd_offset(mm, addr);
+	if (pgd_present(pgdp_get(pgdp))) {
+		p4dp = p4d_offset(pgdp, addr);
+		if (p4d_present(p4dp_get(p4dp))) {
+			pudp = pud_offset(p4dp, addr);
+			if (sz == PUD_SIZE)
+				return (pte_t *)pudp;
+			if (pud_present(pudp_get(pudp)))
+				pmdp = pmd_offset(pudp, addr);
+		}
+	}
+	return (pte_t *) pmdp;
+}
+
+bool __init arch_hugetlb_valid_size(unsigned long size)
+{
+	if (cpu_has_edat1() && size == PMD_SIZE)
+		return true;
+	else if (cpu_has_edat2() && size == PUD_SIZE)
+		return true;
+	else
+		return false;
+}
+
+unsigned int __init arch_hugetlb_cma_order(void)
+{
+	if (cpu_has_edat2())
+		return PUD_SHIFT - PAGE_SHIFT;
+
+	return 0;
+}

linux-7.2.3/arch/s390/mm/init.c

@@ -0,0 +1,319 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  S390 version
+ *    Copyright IBM Corp. 1999
+ *    Author(s): Hartmut Penner (hp@de.ibm.com)
+ *
+ *  Derived from "arch/i386/mm/init.c"
+ *    Copyright (C) 1995  Linus Torvalds
+ */
+
+#include <linux/cpufeature.h>
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/swiotlb.h>
+#include <linux/smp.h>
+#include <linux/init.h>
+#include <linux/pagemap.h>
+#include <linux/memblock.h>
+#include <linux/memory.h>
+#include <linux/pfn.h>
+#include <linux/poison.h>
+#include <linux/initrd.h>
+#include <linux/export.h>
+#include <linux/cma.h>
+#include <linux/gfp.h>
+#include <linux/dma-direct.h>
+#include <linux/percpu.h>
+#include <asm/processor.h>
+#include <linux/uaccess.h>
+#include <asm/pgalloc.h>
+#include <asm/ctlreg.h>
+#include <asm/kfence.h>
+#include <asm/dma.h>
+#include <asm/abs_lowcore.h>
+#include <asm/tlbflush.h>
+#include <asm/sections.h>
+#include <asm/sclp.h>
+#include <asm/set_memory.h>
+#include <asm/kasan.h>
+#include <asm/dma-mapping.h>
+#include <asm/uv.h>
+#include <linux/virtio_anchor.h>
+#include <linux/virtio_config.h>
+#include <linux/execmem.h>
+
+pgd_t swapper_pg_dir[PTRS_PER_PGD] __section(".bss..swapper_pg_dir");
+pgd_t invalid_pg_dir[PTRS_PER_PGD] __section(".bss..invalid_pg_dir");
+
+struct ctlreg __bootdata_preserved(s390_invalid_asce);
+
+unsigned long __bootdata_preserved(page_noexec_mask);
+EXPORT_SYMBOL(page_noexec_mask);
+
+unsigned long __bootdata_preserved(segment_noexec_mask);
+EXPORT_SYMBOL(segment_noexec_mask);
+
+unsigned long __bootdata_preserved(region_noexec_mask);
+EXPORT_SYMBOL(region_noexec_mask);
+
+unsigned long empty_zero_page, zero_page_mask;
+EXPORT_SYMBOL(empty_zero_page);
+EXPORT_SYMBOL(zero_page_mask);
+
+void __init arch_setup_zero_pages(void)
+{
+	unsigned long total_pages = memblock_estimated_nr_free_pages();
+	unsigned int order;
+
+	/* Latest machines require a mapping granularity of 512KB */
+	order = 7;
+
+	/* Limit number of empty zero pages for small memory sizes */
+	while (order > 2 && (total_pages >> 10) < (1UL << order))
+		order--;
+
+	empty_zero_page = (unsigned long)memblock_alloc_or_panic(PAGE_SIZE << order, PAGE_SIZE);
+
+	zero_page_mask = ((PAGE_SIZE << order) - 1) & PAGE_MASK;
+
+	set_memory_ro(empty_zero_page, 1UL << order);
+}
+
+void __init arch_zone_limits_init(unsigned long *max_zone_pfns)
+{
+	max_zone_pfns[ZONE_DMA] = virt_to_pfn(MAX_DMA_ADDRESS);
+	max_zone_pfns[ZONE_NORMAL] = max_low_pfn;
+}
+
+/*
+ * paging_init() sets up the page tables
+ */
+void __init paging_init(void)
+{
+	vmem_map_init();
+	zone_dma_limit = DMA_BIT_MASK(31);
+}
+
+void mark_rodata_ro(void)
+{
+	unsigned long size = __end_ro_after_init - __start_ro_after_init;
+
+	if (cpu_has_nx())
+		system_ctl_set_bit(0, CR0_INSTRUCTION_EXEC_PROTECTION_BIT);
+	__set_memory_ro(__start_ro_after_init, __end_ro_after_init);
+	pr_info("Write protected read-only-after-init data: %luk\n", size >> 10);
+}
+
+int set_memory_encrypted(unsigned long vaddr, int numpages)
+{
+	int i;
+
+	/* make specified pages unshared, (swiotlb, dma_free) */
+	for (i = 0; i < numpages; ++i) {
+		uv_remove_shared(virt_to_phys((void *)vaddr));
+		vaddr += PAGE_SIZE;
+	}
+	return 0;
+}
+
+int set_memory_decrypted(unsigned long vaddr, int numpages)
+{
+	int i;
+	/* make specified pages shared (swiotlb, dma_alloca) */
+	for (i = 0; i < numpages; ++i) {
+		uv_set_shared(virt_to_phys((void *)vaddr));
+		vaddr += PAGE_SIZE;
+	}
+	return 0;
+}
+
+/* are we a protected virtualization guest? */
+bool force_dma_unencrypted(struct device *dev)
+{
+	return is_prot_virt_guest();
+}
+
+/* protected virtualization */
+static void __init pv_init(void)
+{
+	if (!is_prot_virt_guest())
+		return;
+
+	virtio_set_mem_acc_cb(virtio_require_restricted_mem_acc);
+
+	/* make sure bounce buffers are shared */
+	swiotlb_init(true, SWIOTLB_FORCE | SWIOTLB_VERBOSE);
+	swiotlb_update_mem_attributes();
+}
+
+void __init arch_mm_preinit(void)
+{
+	cpumask_set_cpu(0, &init_mm.context.cpu_attach_mask);
+	cpumask_set_cpu(0, mm_cpumask(&init_mm));
+
+	pv_init();
+}
+
+unsigned long memory_block_size_bytes(void)
+{
+	/*
+	 * Make sure the memory block size is always greater
+	 * or equal than the memory increment size.
+	 */
+	return max_t(unsigned long, MIN_MEMORY_BLOCK_SIZE, sclp.rzm);
+}
+
+unsigned long __per_cpu_offset[NR_CPUS] __read_mostly;
+EXPORT_SYMBOL(__per_cpu_offset);
+
+static int __init pcpu_cpu_distance(unsigned int from, unsigned int to)
+{
+	return LOCAL_DISTANCE;
+}
+
+static int __init pcpu_cpu_to_node(int cpu)
+{
+	return 0;
+}
+
+void __init setup_per_cpu_areas(void)
+{
+	unsigned long delta;
+	unsigned int cpu;
+	int rc;
+
+	/*
+	 * Always reserve area for module percpu variables.  That's
+	 * what the legacy allocator did.
+	 */
+	rc = pcpu_embed_first_chunk(PERCPU_MODULE_RESERVE,
+				    PERCPU_DYNAMIC_RESERVE, PAGE_SIZE,
+				    pcpu_cpu_distance,
+				    pcpu_cpu_to_node);
+	if (rc < 0)
+		panic("Failed to initialize percpu areas.");
+
+	delta = (unsigned long)pcpu_base_addr - (unsigned long)__per_cpu_start;
+	for_each_possible_cpu(cpu)
+		__per_cpu_offset[cpu] = delta + pcpu_unit_offsets[cpu];
+}
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+
+#ifdef CONFIG_CMA
+
+/* Prevent memory blocks which contain cma regions from going offline */
+
+struct s390_cma_mem_data {
+	unsigned long start;
+	unsigned long end;
+};
+
+static int s390_cma_check_range(struct cma *cma, void *data)
+{
+	struct s390_cma_mem_data *mem_data;
+
+	mem_data = data;
+
+	if (cma_intersects(cma, mem_data->start, mem_data->end))
+		return -EBUSY;
+
+	return 0;
+}
+
+static int s390_cma_mem_notifier(struct notifier_block *nb,
+				 unsigned long action, void *data)
+{
+	struct s390_cma_mem_data mem_data;
+	struct memory_notify *arg;
+	int rc = 0;
+
+	arg = data;
+	mem_data.start = arg->start_pfn << PAGE_SHIFT;
+	mem_data.end = mem_data.start + (arg->nr_pages << PAGE_SHIFT);
+	if (action == MEM_GOING_OFFLINE)
+		rc = cma_for_each_area(s390_cma_check_range, &mem_data);
+	return notifier_from_errno(rc);
+}
+
+static struct notifier_block s390_cma_mem_nb = {
+	.notifier_call = s390_cma_mem_notifier,
+};
+
+static int __init s390_cma_mem_init(void)
+{
+	return register_memory_notifier(&s390_cma_mem_nb);
+}
+device_initcall(s390_cma_mem_init);
+
+#endif /* CONFIG_CMA */
+
+int arch_add_memory(int nid, u64 start, u64 size,
+		    struct mhp_params *params)
+{
+	unsigned long start_pfn = PFN_DOWN(start);
+	unsigned long size_pages = PFN_DOWN(size);
+	int rc;
+
+	if (WARN_ON_ONCE(pgprot_val(params->pgprot) != pgprot_val(PAGE_KERNEL)))
+		return -EINVAL;
+
+	VM_BUG_ON(!mhp_range_allowed(start, size, true));
+	rc = vmem_add_mapping(start, size);
+	if (rc)
+		return rc;
+
+	rc = __add_pages(nid, start_pfn, size_pages, params);
+	if (rc)
+		vmem_remove_mapping(start, size);
+	return rc;
+}
+
+void arch_remove_memory(u64 start, u64 size, struct vmem_altmap *altmap,
+			struct dev_pagemap *pgmap)
+{
+	unsigned long start_pfn = start >> PAGE_SHIFT;
+	unsigned long nr_pages = size >> PAGE_SHIFT;
+
+	__remove_pages(start_pfn, nr_pages, altmap, pgmap);
+	vmem_remove_mapping(start, size);
+}
+#endif /* CONFIG_MEMORY_HOTPLUG */
+
+#ifdef CONFIG_EXECMEM
+static struct execmem_info execmem_info __ro_after_init;
+
+struct execmem_info __init *execmem_arch_setup(void)
+{
+	unsigned long module_load_offset = 0;
+	unsigned long start;
+
+	if (kaslr_enabled())
+		module_load_offset = get_random_u32_inclusive(1, 1024) * PAGE_SIZE;
+
+	start = MODULES_VADDR + module_load_offset;
+
+	execmem_info = (struct execmem_info){
+		.ranges = {
+			[EXECMEM_DEFAULT] = {
+				.flags	= EXECMEM_KASAN_SHADOW,
+				.start	= start,
+				.end	= MODULES_END,
+				.pgprot	= PAGE_KERNEL,
+				.alignment = MODULE_ALIGN,
+			},
+		},
+	};
+
+	return &execmem_info;
+}
+#endif /* CONFIG_EXECMEM */

linux-7.2.3/arch/s390/mm/maccess.c

@@ -0,0 +1,194 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Access kernel memory without faulting -- s390 specific implementation.
+ *
+ * Copyright IBM Corp. 2009, 2015
+ *
+ */
+
+#include <linux/uaccess.h>
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/gfp.h>
+#include <linux/cpu.h>
+#include <linux/uio.h>
+#include <linux/io.h>
+#include <asm/asm-extable.h>
+#include <asm/abs_lowcore.h>
+#include <asm/stacktrace.h>
+#include <asm/sections.h>
+#include <asm/maccess.h>
+#include <asm/ctlreg.h>
+
+unsigned long __bootdata_preserved(__memcpy_real_area);
+pte_t *__bootdata_preserved(memcpy_real_ptep);
+static DEFINE_MUTEX(memcpy_real_mutex);
+
+static notrace long s390_kernel_write_odd(void *dst, const void *src, size_t size)
+{
+	unsigned long aligned, offset, count;
+	char tmp[8];
+
+	aligned = (unsigned long) dst & ~7UL;
+	offset = (unsigned long) dst & 7UL;
+	size = min(8UL - offset, size);
+	count = size - 1;
+	asm volatile(
+		"	bras	1,0f\n"
+		"	mvc	0(1,%4),0(%5)\n"
+		"0:	mvc	0(8,%3),0(%0)\n"
+		"	ex	%1,0(1)\n"
+		"	lg	%1,0(%3)\n"
+		"	lra	%0,0(%0)\n"
+		"	sturg	%1,%0"
+		: "+&a" (aligned), "+&a" (count), "=m" (tmp)
+		: "a" (&tmp), "a" (&tmp[offset]), "a" (src)
+		: "cc", "memory", "1");
+	return size;
+}
+
+/*
+ * __s390_kernel_write - write to kernel memory bypassing DAT
+ * @dst: destination address
+ * @src: source address
+ * @size: number of bytes to copy
+ *
+ * This function writes to kernel memory bypassing DAT and possible page table
+ * write protection. It writes to the destination using the sturg instruction.
+ * Therefore we have a read-modify-write sequence: the function reads eight
+ * bytes from destination at an eight byte boundary, modifies the bytes
+ * requested and writes the result back in a loop.
+ */
+static DEFINE_SPINLOCK(s390_kernel_write_lock);
+
+notrace void *__s390_kernel_write(void *dst, const void *src, size_t size)
+{
+	void *tmp = dst;
+	unsigned long flags;
+	long copied;
+
+	spin_lock_irqsave(&s390_kernel_write_lock, flags);
+	while (size) {
+		copied = s390_kernel_write_odd(tmp, src, size);
+		tmp += copied;
+		src += copied;
+		size -= copied;
+	}
+	spin_unlock_irqrestore(&s390_kernel_write_lock, flags);
+
+	return dst;
+}
+
+size_t memcpy_real_iter(struct iov_iter *iter, unsigned long src, size_t count)
+{
+	size_t len, copied, res = 0;
+	unsigned long phys, offset;
+	void *chunk;
+	pte_t pte;
+
+	BUILD_BUG_ON(MEMCPY_REAL_SIZE != PAGE_SIZE);
+	while (count) {
+		phys = src & MEMCPY_REAL_MASK;
+		offset = src & ~MEMCPY_REAL_MASK;
+		chunk = (void *)(__memcpy_real_area + offset);
+		len = min(count, MEMCPY_REAL_SIZE - offset);
+		pte = mk_pte_phys(phys, PAGE_KERNEL_RO);
+
+		mutex_lock(&memcpy_real_mutex);
+		if (pte_val(pte) != pte_val(*memcpy_real_ptep)) {
+			__ptep_ipte(__memcpy_real_area, memcpy_real_ptep, 0, 0, IPTE_GLOBAL);
+			set_pte(memcpy_real_ptep, pte);
+		}
+		copied = copy_to_iter(chunk, len, iter);
+		mutex_unlock(&memcpy_real_mutex);
+
+		count -= copied;
+		src += copied;
+		res += copied;
+		if (copied < len)
+			break;
+	}
+	return res;
+}
+
+int memcpy_real(void *dest, unsigned long src, size_t count)
+{
+	struct iov_iter iter;
+	struct kvec kvec;
+
+	kvec.iov_base = dest;
+	kvec.iov_len = count;
+	iov_iter_kvec(&iter, ITER_DEST, &kvec, 1, count);
+	if (memcpy_real_iter(&iter, src, count) < count)
+		return -EFAULT;
+	return 0;
+}
+
+/*
+ * Find CPU that owns swapped prefix page
+ */
+static int get_swapped_owner(phys_addr_t addr)
+{
+	phys_addr_t lc;
+	int cpu;
+
+	for_each_online_cpu(cpu) {
+		lc = virt_to_phys(lowcore_ptr[cpu]);
+		if (addr > lc + sizeof(struct lowcore) - 1 || addr < lc)
+			continue;
+		return cpu;
+	}
+	return -1;
+}
+
+/*
+ * Convert a physical pointer for /dev/mem access
+ *
+ * For swapped prefix pages a new buffer is returned that contains a copy of
+ * the absolute memory. The buffer size is maximum one page large.
+ */
+void *xlate_dev_mem_ptr(phys_addr_t addr)
+{
+	void *ptr = phys_to_virt(addr);
+	void *bounce = ptr;
+	struct lowcore *abs_lc;
+	unsigned long size;
+	int this_cpu, cpu;
+
+	cpus_read_lock();
+	this_cpu = get_cpu();
+	if (addr >= sizeof(struct lowcore)) {
+		cpu = get_swapped_owner(addr);
+		if (cpu < 0)
+			goto out;
+	}
+	bounce = (void *)__get_free_page(GFP_ATOMIC);
+	if (!bounce)
+		goto out;
+	size = PAGE_SIZE - (addr & ~PAGE_MASK);
+	if (addr < sizeof(struct lowcore)) {
+		abs_lc = get_abs_lowcore();
+		ptr = (void *)abs_lc + addr;
+		memcpy(bounce, ptr, size);
+		put_abs_lowcore(abs_lc);
+	} else if (cpu == this_cpu) {
+		ptr = (void *)(addr - virt_to_phys(lowcore_ptr[cpu]));
+		memcpy(bounce, ptr, size);
+	} else {
+		memcpy(bounce, ptr, size);
+	}
+out:
+	put_cpu();
+	cpus_read_unlock();
+	return bounce;
+}
+
+/*
+ * Free converted buffer for /dev/mem access (if necessary)
+ */
+void unxlate_dev_mem_ptr(phys_addr_t addr, void *ptr)
+{
+	if (addr != virt_to_phys(ptr))
+		free_page((unsigned long)ptr);
+}

linux-7.2.3/arch/s390/mm/mmap.c

@@ -0,0 +1,215 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ *  flexible mmap layout support
+ *
+ * Copyright 2003-2004 Red Hat Inc., Durham, North Carolina.
+ * All Rights Reserved.
+ *
+ * Started by Ingo Molnar <mingo@elte.hu>
+ */
+
+#include <linux/elf-randomize.h>
+#include <linux/personality.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/sched/signal.h>
+#include <linux/sched/mm.h>
+#include <linux/random.h>
+#include <linux/security.h>
+#include <linux/hugetlb.h>
+#include <asm/elf.h>
+
+static unsigned long stack_maxrandom_size(void)
+{
+	if (!(current->flags & PF_RANDOMIZE))
+		return 0;
+	return STACK_RND_MASK << PAGE_SHIFT;
+}
+
+static inline int mmap_is_legacy(const struct rlimit *rlim_stack)
+{
+	if (current->personality & ADDR_COMPAT_LAYOUT)
+		return 1;
+	if (rlim_stack->rlim_cur == RLIM_INFINITY)
+		return 1;
+	return sysctl_legacy_va_layout;
+}
+
+unsigned long arch_mmap_rnd(void)
+{
+	return (get_random_u32() & MMAP_RND_MASK) << PAGE_SHIFT;
+}
+
+static unsigned long mmap_base_legacy(unsigned long rnd)
+{
+	return TASK_UNMAPPED_BASE + rnd;
+}
+
+static inline unsigned long mmap_base(unsigned long rnd,
+				      const struct rlimit *rlim_stack)
+{
+	unsigned long gap = rlim_stack->rlim_cur;
+	unsigned long pad = stack_maxrandom_size() + stack_guard_gap;
+
+	/* Values close to RLIM_INFINITY can overflow. */
+	if (gap + pad > gap)
+		gap += pad;
+
+	/*
+	 * Top of mmap area (just below the process stack).
+	 * Leave at least a ~128 MB hole.
+	 */
+	gap = clamp(gap, SZ_128M, (STACK_TOP / 6) * 5);
+
+	return PAGE_ALIGN(STACK_TOP - gap - rnd);
+}
+
+static unsigned long get_align_mask(struct file *filp, unsigned long flags)
+{
+	if (filp && is_file_hugepages(filp))
+		return huge_page_mask_align(filp);
+	if (!(current->flags & PF_RANDOMIZE))
+		return 0;
+	if (filp || (flags & MAP_SHARED))
+		return MMAP_ALIGN_MASK << PAGE_SHIFT;
+	return 0;
+}
+
+unsigned long arch_get_unmapped_area(struct file *filp, unsigned long addr,
+				     unsigned long len, unsigned long pgoff,
+				     unsigned long flags, vm_flags_t vm_flags)
+{
+	struct mm_struct *mm = current->mm;
+	struct vm_area_struct *vma;
+	struct vm_unmapped_area_info info = {};
+
+	if (len > TASK_SIZE - mmap_min_addr)
+		return -ENOMEM;
+
+	if (flags & MAP_FIXED)
+		goto check_asce_limit;
+
+	if (addr) {
+		addr = PAGE_ALIGN(addr);
+		vma = find_vma(mm, addr);
+		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
+		    (!vma || addr + len <= vm_start_gap(vma)))
+			goto check_asce_limit;
+	}
+
+	info.length = len;
+	info.low_limit = mm->mmap_base;
+	info.high_limit = TASK_SIZE;
+	info.align_mask = get_align_mask(filp, flags);
+	if (!(filp && is_file_hugepages(filp)))
+		info.align_offset = pgoff << PAGE_SHIFT;
+	addr = vm_unmapped_area(&info);
+	if (offset_in_page(addr))
+		return addr;
+
+check_asce_limit:
+	return check_asce_limit(mm, addr, len);
+}
+
+unsigned long arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
+					     unsigned long len, unsigned long pgoff,
+					     unsigned long flags, vm_flags_t vm_flags)
+{
+	struct vm_area_struct *vma;
+	struct mm_struct *mm = current->mm;
+	struct vm_unmapped_area_info info = {};
+
+	/* requested length too big for entire address space */
+	if (len > TASK_SIZE - mmap_min_addr)
+		return -ENOMEM;
+
+	if (flags & MAP_FIXED)
+		goto check_asce_limit;
+
+	/* requesting a specific address */
+	if (addr) {
+		addr = PAGE_ALIGN(addr);
+		vma = find_vma(mm, addr);
+		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
+				(!vma || addr + len <= vm_start_gap(vma)))
+			goto check_asce_limit;
+	}
+
+	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
+	info.length = len;
+	info.low_limit = PAGE_SIZE;
+	info.high_limit = mm->mmap_base;
+	info.align_mask = get_align_mask(filp, flags);
+	if (!(filp && is_file_hugepages(filp)))
+		info.align_offset = pgoff << PAGE_SHIFT;
+	addr = vm_unmapped_area(&info);
+
+	/*
+	 * A failed mmap() very likely causes application failure,
+	 * so fall back to the bottom-up function here. This scenario
+	 * can happen with large stack limits and large mmap()
+	 * allocations.
+	 */
+	if (offset_in_page(addr)) {
+		VM_BUG_ON(addr != -ENOMEM);
+		info.flags = 0;
+		info.low_limit = TASK_UNMAPPED_BASE;
+		info.high_limit = TASK_SIZE;
+		addr = vm_unmapped_area(&info);
+		if (offset_in_page(addr))
+			return addr;
+	}
+
+check_asce_limit:
+	return check_asce_limit(mm, addr, len);
+}
+
+/*
+ * This function, called very early during the creation of a new
+ * process VM image, sets up which VM layout function to use:
+ */
+void arch_pick_mmap_layout(struct mm_struct *mm, const struct rlimit *rlim_stack)
+{
+	unsigned long random_factor = 0UL;
+
+	if (current->flags & PF_RANDOMIZE)
+		random_factor = arch_mmap_rnd();
+
+	/*
+	 * Fall back to the standard layout if the personality
+	 * bit is set, or if the expected stack growth is unlimited:
+	 */
+	if (mmap_is_legacy(rlim_stack)) {
+		mm->mmap_base = mmap_base_legacy(random_factor);
+		mm_flags_clear(MMF_TOPDOWN, mm);
+	} else {
+		mm->mmap_base = mmap_base(random_factor, rlim_stack);
+		mm_flags_set(MMF_TOPDOWN, mm);
+	}
+}
+
+static pgprot_t protection_map[16] __ro_after_init;
+
+void __init setup_protection_map(void)
+{
+	pgprot_t *pm = protection_map;
+
+	pm[VM_NONE]					= PAGE_NONE;
+	pm[VM_READ]					= PAGE_RO;
+	pm[VM_WRITE]					= PAGE_RO;
+	pm[VM_WRITE | VM_READ]				= PAGE_RO;
+	pm[VM_EXEC]					= PAGE_RX;
+	pm[VM_EXEC | VM_READ]				= PAGE_RX;
+	pm[VM_EXEC | VM_WRITE]				= PAGE_RX;
+	pm[VM_EXEC | VM_WRITE | VM_READ]		= PAGE_RX;
+	pm[VM_SHARED]					= PAGE_NONE;
+	pm[VM_SHARED | VM_READ]				= PAGE_RO;
+	pm[VM_SHARED | VM_WRITE]			= PAGE_RW;
+	pm[VM_SHARED | VM_WRITE | VM_READ]		= PAGE_RW;
+	pm[VM_SHARED | VM_EXEC]				= PAGE_RX;
+	pm[VM_SHARED | VM_EXEC | VM_READ]		= PAGE_RX;
+	pm[VM_SHARED | VM_EXEC | VM_WRITE]		= PAGE_RWX;
+	pm[VM_SHARED | VM_EXEC | VM_WRITE | VM_READ]	= PAGE_RWX;
+}
+
+DECLARE_VM_GET_PAGE_PROT

linux-7.2.3/arch/s390/mm/page-states.c

@@ -0,0 +1,33 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright IBM Corp. 2008
+ *
+ * Guest page hinting for unused pages.
+ *
+ * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/mm.h>
+#include <asm/page-states.h>
+#include <asm/sections.h>
+#include <asm/page.h>
+
+int __bootdata_preserved(cmma_flag);
+EXPORT_SYMBOL(cmma_flag);
+
+void arch_free_page(struct page *page, int order)
+{
+	if (!cmma_flag)
+		return;
+	__set_page_unused(page_to_virt(page), 1UL << order);
+}
+
+void arch_alloc_page(struct page *page, int order)
+{
+	if (!cmma_flag)
+		return;
+	if (cmma_flag < 2)
+		__set_page_stable_dat(page_to_virt(page), 1UL << order);
+	else
+		__set_page_stable_nodat(page_to_virt(page), 1UL << order);
+}

linux-7.2.3/arch/s390/mm/pageattr.c

@@ -0,0 +1,471 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright IBM Corp. 2011
+ * Author(s): Jan Glauber <jang@linux.vnet.ibm.com>
+ */
+#include <linux/cpufeature.h>
+#include <linux/hugetlb.h>
+#include <linux/proc_fs.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <asm/cacheflush.h>
+#include <asm/facility.h>
+#include <asm/pgalloc.h>
+#include <asm/kfence.h>
+#include <asm/page.h>
+#include <asm/asm.h>
+#include <asm/set_memory.h>
+
+void __storage_key_init_range(unsigned long start, unsigned long end)
+{
+	unsigned long boundary, size;
+
+	while (start < end) {
+		if (cpu_has_edat1()) {
+			/* set storage keys for a 1MB frame */
+			size = 1UL << 20;
+			boundary = (start + size) & ~(size - 1);
+			if (boundary <= end) {
+				do {
+					start = sske_frame(start, PAGE_DEFAULT_KEY);
+				} while (start < boundary);
+				continue;
+			}
+		}
+		page_set_storage_key(start, PAGE_DEFAULT_KEY, 1);
+		start += PAGE_SIZE;
+	}
+}
+
+#ifdef CONFIG_PROC_FS
+atomic_long_t __bootdata_preserved(direct_pages_count[PG_DIRECT_MAP_MAX]);
+
+void arch_report_meminfo(struct seq_file *m)
+{
+	seq_printf(m, "DirectMap4k:    %8lu kB\n",
+		   atomic_long_read(&direct_pages_count[PG_DIRECT_MAP_4K]) << 2);
+	seq_printf(m, "DirectMap1M:    %8lu kB\n",
+		   atomic_long_read(&direct_pages_count[PG_DIRECT_MAP_1M]) << 10);
+	seq_printf(m, "DirectMap2G:    %8lu kB\n",
+		   atomic_long_read(&direct_pages_count[PG_DIRECT_MAP_2G]) << 21);
+}
+#endif /* CONFIG_PROC_FS */
+
+static void pgt_set(unsigned long *old, unsigned long new, unsigned long addr,
+		    unsigned long dtt)
+{
+	unsigned long *table, mask;
+
+	mask = 0;
+	if (cpu_has_edat2()) {
+		switch (dtt) {
+		case CRDTE_DTT_REGION3:
+			mask = ~(PTRS_PER_PUD * sizeof(pud_t) - 1);
+			break;
+		case CRDTE_DTT_SEGMENT:
+			mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1);
+			break;
+		case CRDTE_DTT_PAGE:
+			mask = ~(PTRS_PER_PTE * sizeof(pte_t) - 1);
+			break;
+		}
+		table = (unsigned long *)((unsigned long)old & mask);
+		crdte(*old, new, table, dtt, addr, get_lowcore()->kernel_asce.val);
+	} else {
+		cspg(old, *old, new);
+	}
+}
+
+static int walk_pte_level(pmd_t *pmdp, unsigned long addr, unsigned long end,
+			  unsigned long flags)
+{
+	pte_t *ptep, new;
+
+	if (flags == SET_MEMORY_4K)
+		return 0;
+	ptep = pte_offset_kernel(pmdp, addr);
+	do {
+		new = ptep_get(ptep);
+		if (pte_none(new))
+			return -EINVAL;
+		if (flags & SET_MEMORY_RO)
+			new = pte_wrprotect(new);
+		else if (flags & SET_MEMORY_RW)
+			new = pte_mkwrite_novma(pte_mkdirty(new));
+		if (flags & SET_MEMORY_NX)
+			new = set_pte_bit(new, __pgprot(_PAGE_NOEXEC));
+		else if (flags & SET_MEMORY_X)
+			new = clear_pte_bit(new, __pgprot(_PAGE_NOEXEC));
+		if (flags & SET_MEMORY_INV) {
+			new = set_pte_bit(new, __pgprot(_PAGE_INVALID));
+		} else if (flags & SET_MEMORY_DEF) {
+			new = __pte(pte_val(new) & PAGE_MASK);
+			new = set_pte_bit(new, PAGE_KERNEL);
+		}
+		pgt_set((unsigned long *)ptep, pte_val(new), addr, CRDTE_DTT_PAGE);
+		ptep++;
+		addr += PAGE_SIZE;
+		cond_resched();
+	} while (addr < end);
+	return 0;
+}
+
+static int split_pmd_page(pmd_t *pmdp, unsigned long addr)
+{
+	unsigned long pte_addr, prot;
+	pte_t *pt_dir, *ptep;
+	pmd_t new, pmd;
+	int i, ro, nx;
+
+	pt_dir = vmem_pte_alloc();
+	if (!pt_dir)
+		return -ENOMEM;
+	pmd = pmdp_get(pmdp);
+	pte_addr = pmd_pfn(pmd) << PAGE_SHIFT;
+	ro = !!(pmd_val(pmd) & _SEGMENT_ENTRY_PROTECT);
+	nx = !!(pmd_val(pmd) & _SEGMENT_ENTRY_NOEXEC);
+	prot = pgprot_val(ro ? PAGE_KERNEL_RO : PAGE_KERNEL);
+	if (!nx)
+		prot &= ~_PAGE_NOEXEC;
+	ptep = pt_dir;
+	for (i = 0; i < PTRS_PER_PTE; i++) {
+		set_pte(ptep, __pte(pte_addr | prot));
+		pte_addr += PAGE_SIZE;
+		ptep++;
+	}
+	new = __pmd(__pa(pt_dir) | _SEGMENT_ENTRY);
+	pgt_set((unsigned long *)pmdp, pmd_val(new), addr, CRDTE_DTT_SEGMENT);
+	update_page_count(PG_DIRECT_MAP_4K, PTRS_PER_PTE);
+	update_page_count(PG_DIRECT_MAP_1M, -1);
+	return 0;
+}
+
+static void modify_pmd_page(pmd_t *pmdp, unsigned long addr,
+			    unsigned long flags)
+{
+	pmd_t new = pmdp_get(pmdp);
+
+	if (flags & SET_MEMORY_RO)
+		new = pmd_wrprotect(new);
+	else if (flags & SET_MEMORY_RW)
+		new = pmd_mkwrite_novma(pmd_mkdirty(new));
+	if (flags & SET_MEMORY_NX)
+		new = set_pmd_bit(new, __pgprot(_SEGMENT_ENTRY_NOEXEC));
+	else if (flags & SET_MEMORY_X)
+		new = clear_pmd_bit(new, __pgprot(_SEGMENT_ENTRY_NOEXEC));
+	if (flags & SET_MEMORY_INV) {
+		new = set_pmd_bit(new, __pgprot(_SEGMENT_ENTRY_INVALID));
+	} else if (flags & SET_MEMORY_DEF) {
+		new = __pmd(pmd_val(new) & PMD_MASK);
+		new = set_pmd_bit(new, SEGMENT_KERNEL);
+	}
+	pgt_set((unsigned long *)pmdp, pmd_val(new), addr, CRDTE_DTT_SEGMENT);
+}
+
+static int walk_pmd_level(pud_t *pudp, unsigned long addr, unsigned long end,
+			  unsigned long flags)
+{
+	unsigned long next;
+	pmd_t *pmdp, pmd;
+	int need_split;
+	int rc = 0;
+
+	pmdp = pmd_offset(pudp, addr);
+	do {
+		pmd = pmdp_get(pmdp);
+		if (pmd_none(pmd))
+			return -EINVAL;
+		next = pmd_addr_end(addr, end);
+		if (pmd_leaf(pmd)) {
+			need_split  = !!(flags & SET_MEMORY_4K);
+			need_split |= !!(addr & ~PMD_MASK);
+			need_split |= !!(addr + PMD_SIZE > next);
+			if (need_split) {
+				rc = split_pmd_page(pmdp, addr);
+				if (rc)
+					return rc;
+				continue;
+			}
+			modify_pmd_page(pmdp, addr, flags);
+		} else {
+			rc = walk_pte_level(pmdp, addr, next, flags);
+			if (rc)
+				return rc;
+		}
+		pmdp++;
+		addr = next;
+		cond_resched();
+	} while (addr < end);
+	return rc;
+}
+
+int split_pud_page(pud_t *pudp, unsigned long addr)
+{
+	unsigned long pmd_addr, prot;
+	pmd_t *pm_dir, *pmdp;
+	pud_t new, pud;
+	int i, ro, nx;
+
+	pm_dir = vmem_crst_alloc(_SEGMENT_ENTRY_EMPTY);
+	if (!pm_dir)
+		return -ENOMEM;
+	pud = pudp_get(pudp);
+	pmd_addr = pud_pfn(pud) << PAGE_SHIFT;
+	ro = !!(pud_val(pud) & _REGION_ENTRY_PROTECT);
+	nx = !!(pud_val(pud) & _REGION_ENTRY_NOEXEC);
+	prot = pgprot_val(ro ? SEGMENT_KERNEL_RO : SEGMENT_KERNEL);
+	if (!nx)
+		prot &= ~_SEGMENT_ENTRY_NOEXEC;
+	pmdp = pm_dir;
+	for (i = 0; i < PTRS_PER_PMD; i++) {
+		set_pmd(pmdp, __pmd(pmd_addr | prot));
+		pmd_addr += PMD_SIZE;
+		pmdp++;
+	}
+	new = __pud(__pa(pm_dir) | _REGION3_ENTRY);
+	pgt_set((unsigned long *)pudp, pud_val(new), addr, CRDTE_DTT_REGION3);
+	update_page_count(PG_DIRECT_MAP_1M, PTRS_PER_PMD);
+	update_page_count(PG_DIRECT_MAP_2G, -1);
+	return 0;
+}
+
+static void modify_pud_page(pud_t *pudp, unsigned long addr,
+			    unsigned long flags)
+{
+	pud_t new = pudp_get(pudp);
+
+	if (flags & SET_MEMORY_RO)
+		new = pud_wrprotect(new);
+	else if (flags & SET_MEMORY_RW)
+		new = pud_mkwrite(pud_mkdirty(new));
+	if (flags & SET_MEMORY_NX)
+		new = set_pud_bit(new, __pgprot(_REGION_ENTRY_NOEXEC));
+	else if (flags & SET_MEMORY_X)
+		new = clear_pud_bit(new, __pgprot(_REGION_ENTRY_NOEXEC));
+	if (flags & SET_MEMORY_INV) {
+		new = set_pud_bit(new, __pgprot(_REGION_ENTRY_INVALID));
+	} else if (flags & SET_MEMORY_DEF) {
+		new = __pud(pud_val(new) & PUD_MASK);
+		new = set_pud_bit(new, REGION3_KERNEL);
+	}
+	pgt_set((unsigned long *)pudp, pud_val(new), addr, CRDTE_DTT_REGION3);
+}
+
+static int walk_pud_level(p4d_t *p4d, unsigned long addr, unsigned long end,
+			  unsigned long flags)
+{
+	unsigned long next;
+	pud_t *pudp, pud;
+	int need_split;
+	int rc = 0;
+
+	pudp = pud_offset(p4d, addr);
+	do {
+		pud = pudp_get(pudp);
+		if (pud_none(pud))
+			return -EINVAL;
+		next = pud_addr_end(addr, end);
+		if (pud_leaf(pud)) {
+			need_split  = !!(flags & SET_MEMORY_4K);
+			need_split |= !!(addr & ~PUD_MASK);
+			need_split |= !!(addr + PUD_SIZE > next);
+			if (need_split) {
+				rc = split_pud_page(pudp, addr);
+				if (rc)
+					break;
+				continue;
+			}
+			modify_pud_page(pudp, addr, flags);
+		} else {
+			rc = walk_pmd_level(pudp, addr, next, flags);
+		}
+		pudp++;
+		addr = next;
+		cond_resched();
+	} while (addr < end && !rc);
+	return rc;
+}
+
+static int walk_p4d_level(pgd_t *pgd, unsigned long addr, unsigned long end,
+			  unsigned long flags)
+{
+	unsigned long next;
+	p4d_t *p4dp;
+	int rc = 0;
+
+	p4dp = p4d_offset(pgd, addr);
+	do {
+		if (p4d_none(p4dp_get(p4dp)))
+			return -EINVAL;
+		next = p4d_addr_end(addr, end);
+		rc = walk_pud_level(p4dp, addr, next, flags);
+		p4dp++;
+		addr = next;
+		cond_resched();
+	} while (addr < end && !rc);
+	return rc;
+}
+
+DEFINE_MUTEX(cpa_mutex);
+
+static int change_page_attr(unsigned long addr, unsigned long end,
+			    unsigned long flags)
+{
+	unsigned long next;
+	int rc = -EINVAL;
+	pgd_t *pgdp;
+
+	pgdp = pgd_offset_k(addr);
+	do {
+		if (pgd_none(pgdp_get(pgdp)))
+			break;
+		next = pgd_addr_end(addr, end);
+		rc = walk_p4d_level(pgdp, addr, next, flags);
+		if (rc)
+			break;
+		cond_resched();
+	} while (pgdp++, addr = next, addr < end && !rc);
+	return rc;
+}
+
+static int change_page_attr_alias(unsigned long addr, unsigned long end,
+				  unsigned long flags)
+{
+	unsigned long alias, offset, va_start, va_end;
+	struct vm_struct *area;
+	int rc = 0;
+
+	/*
+	 * Changes to read-only permissions on kernel VA mappings are also
+	 * applied to the kernel direct mapping. Execute permissions are
+	 * intentionally not transferred to keep all allocated pages within
+	 * the direct mapping non-executable.
+	 */
+	flags &= SET_MEMORY_RO | SET_MEMORY_RW;
+	if (!flags)
+		return 0;
+	area = NULL;
+	while (addr < end) {
+		if (!area)
+			area = find_vm_area((void *)addr);
+		if (!area || !(area->flags & VM_ALLOC))
+			return 0;
+		va_start = (unsigned long)area->addr;
+		va_end = va_start + area->nr_pages * PAGE_SIZE;
+		offset = (addr - va_start) >> PAGE_SHIFT;
+		alias = (unsigned long)page_address(area->pages[offset]);
+		rc = change_page_attr(alias, alias + PAGE_SIZE, flags);
+		if (rc)
+			break;
+		addr += PAGE_SIZE;
+		if (addr >= va_end)
+			area = NULL;
+	}
+	return rc;
+}
+
+int __set_memory(unsigned long addr, unsigned long numpages, unsigned long flags)
+{
+	unsigned long end;
+	int rc;
+
+	if (!cpu_has_nx())
+		flags &= ~(SET_MEMORY_NX | SET_MEMORY_X);
+	if (!flags)
+		return 0;
+	if (!numpages)
+		return 0;
+	addr &= PAGE_MASK;
+	end = addr + numpages * PAGE_SIZE;
+	mutex_lock(&cpa_mutex);
+	rc = change_page_attr(addr, end, flags);
+	if (rc)
+		goto out;
+	rc = change_page_attr_alias(addr, end, flags);
+out:
+	mutex_unlock(&cpa_mutex);
+	return rc;
+}
+
+int set_direct_map_invalid_noflush(struct page *page)
+{
+	return __set_memory((unsigned long)page_to_virt(page), 1, SET_MEMORY_INV);
+}
+
+int set_direct_map_default_noflush(struct page *page)
+{
+	return __set_memory((unsigned long)page_to_virt(page), 1, SET_MEMORY_DEF);
+}
+
+int set_direct_map_valid_noflush(struct page *page, unsigned nr, bool valid)
+{
+	unsigned long flags;
+
+	if (valid)
+		flags = SET_MEMORY_DEF;
+	else
+		flags = SET_MEMORY_INV;
+
+	return __set_memory((unsigned long)page_to_virt(page), nr, flags);
+}
+
+bool kernel_page_present(struct page *page)
+{
+	unsigned long addr;
+	unsigned int cc;
+
+	addr = (unsigned long)page_address(page);
+	asm volatile(
+		"	lra	%[addr],0(%[addr])\n"
+		CC_IPM(cc)
+		: CC_OUT(cc, cc), [addr] "+a" (addr)
+		:
+		: CC_CLOBBER);
+	return CC_TRANSFORM(cc) == 0;
+}
+
+#if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KFENCE)
+
+static void ipte_range(pte_t *pte, unsigned long address, int nr)
+{
+	int i;
+
+	if (test_facility(13)) {
+		__ptep_ipte_range(address, nr - 1, pte, IPTE_GLOBAL);
+		return;
+	}
+	for (i = 0; i < nr; i++) {
+		__ptep_ipte(address, pte, 0, 0, IPTE_GLOBAL);
+		address += PAGE_SIZE;
+		pte++;
+	}
+}
+
+void __kernel_map_pages(struct page *page, int numpages, int enable)
+{
+	unsigned long address;
+	pte_t *ptep, pte;
+	int nr, i, j;
+
+	for (i = 0; i < numpages;) {
+		address = (unsigned long)page_to_virt(page + i);
+		ptep = virt_to_kpte(address);
+		nr = (unsigned long)ptep >> ilog2(sizeof(long));
+		nr = PTRS_PER_PTE - (nr & (PTRS_PER_PTE - 1));
+		nr = min(numpages - i, nr);
+		if (enable) {
+			for (j = 0; j < nr; j++) {
+				pte = ptep_get(ptep);
+				pte = clear_pte_bit(pte, __pgprot(_PAGE_INVALID));
+				set_pte(ptep, pte);
+				address += PAGE_SIZE;
+				ptep++;
+			}
+		} else {
+			ipte_range(ptep, address, nr);
+		}
+		i += nr;
+	}
+}
+
+#endif /* CONFIG_DEBUG_PAGEALLOC */

linux-7.2.3/arch/s390/mm/pfault.c

@@ -0,0 +1,248 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright IBM Corp. 1999, 2023
+ */
+
+#include <linux/cpuhotplug.h>
+#include <linux/sched/task.h>
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/irq.h>
+#include <asm/asm-extable.h>
+#include <asm/asm-offsets.h>
+#include <asm/pfault.h>
+#include <asm/diag.h>
+
+#define __SUBCODE_MASK 0x0600
+#define __PF_RES_FIELD 0x8000000000000000UL
+
+/*
+ * 'pfault' pseudo page faults routines.
+ */
+static int pfault_disable;
+
+static int __init nopfault(char *str)
+{
+	pfault_disable = 1;
+	return 1;
+}
+early_param("nopfault", nopfault);
+
+struct pfault_refbk {
+	u16 refdiagc;
+	u16 reffcode;
+	u16 refdwlen;
+	u16 refversn;
+	u64 refgaddr;
+	u64 refselmk;
+	u64 refcmpmk;
+	u64 reserved;
+};
+
+static struct pfault_refbk pfault_init_refbk = {
+	.refdiagc = 0x258,
+	.reffcode = 0,
+	.refdwlen = 5,
+	.refversn = 2,
+	.refgaddr = __LC_LPP,
+	.refselmk = 1UL << 48,
+	.refcmpmk = 1UL << 48,
+	.reserved = __PF_RES_FIELD
+};
+
+int __pfault_init(void)
+{
+	int rc = -EOPNOTSUPP;
+
+	if (pfault_disable)
+		return rc;
+	diag_stat_inc(DIAG_STAT_X258);
+	asm_inline volatile(
+		"	diag	%[refbk],%[rc],0x258\n"
+		"0:	nopr	%%r7\n"
+		EX_TABLE(0b, 0b)
+		: [rc] "+d" (rc)
+		: [refbk] "a" (virt_to_phys(&pfault_init_refbk)), "m" (pfault_init_refbk)
+		: "cc");
+	return rc;
+}
+
+static struct pfault_refbk pfault_fini_refbk = {
+	.refdiagc = 0x258,
+	.reffcode = 1,
+	.refdwlen = 5,
+	.refversn = 2,
+};
+
+void __pfault_fini(void)
+{
+	if (pfault_disable)
+		return;
+	diag_stat_inc(DIAG_STAT_X258);
+	asm_inline volatile(
+		"	diag	%[refbk],0,0x258\n"
+		"0:	nopr	%%r7\n"
+		EX_TABLE(0b, 0b)
+		:
+		: [refbk] "a" (virt_to_phys(&pfault_fini_refbk)), "m" (pfault_fini_refbk)
+		: "cc");
+}
+
+static DEFINE_SPINLOCK(pfault_lock);
+static LIST_HEAD(pfault_list);
+
+#define PF_COMPLETE	0x0080
+
+/*
+ * The mechanism of our pfault code: if Linux is running as guest, runs a user
+ * space process and the user space process accesses a page that the host has
+ * paged out we get a pfault interrupt.
+ *
+ * This allows us, within the guest, to schedule a different process. Without
+ * this mechanism the host would have to suspend the whole virtual cpu until
+ * the page has been paged in.
+ *
+ * So when we get such an interrupt then we set the state of the current task
+ * to uninterruptible and also set the need_resched flag. Both happens within
+ * interrupt context(!). If we later on want to return to user space we
+ * recognize the need_resched flag and then call schedule().  It's not very
+ * obvious how this works...
+ *
+ * Of course we have a lot of additional fun with the completion interrupt (->
+ * host signals that a page of a process has been paged in and the process can
+ * continue to run). This interrupt can arrive on any cpu and, since we have
+ * virtual cpus, actually appear before the interrupt that signals that a page
+ * is missing.
+ */
+static void pfault_interrupt(struct ext_code ext_code,
+			     unsigned int param32, unsigned long param64)
+{
+	struct task_struct *tsk;
+	__u16 subcode;
+	pid_t pid;
+
+	/*
+	 * Get the external interruption subcode & pfault initial/completion
+	 * signal bit. VM stores this in the 'cpu address' field associated
+	 * with the external interrupt.
+	 */
+	subcode = ext_code.subcode;
+	if ((subcode & 0xff00) != __SUBCODE_MASK)
+		return;
+	inc_irq_stat(IRQEXT_PFL);
+	/* Get the token (= pid of the affected task). */
+	pid = param64 & LPP_PID_MASK;
+	rcu_read_lock();
+	tsk = find_task_by_pid_ns(pid, &init_pid_ns);
+	if (tsk)
+		get_task_struct(tsk);
+	rcu_read_unlock();
+	if (!tsk)
+		return;
+	spin_lock(&pfault_lock);
+	if (subcode & PF_COMPLETE) {
+		/* signal bit is set -> a page has been swapped in by VM */
+		if (tsk->thread.pfault_wait == 1) {
+			/*
+			 * Initial interrupt was faster than the completion
+			 * interrupt. pfault_wait is valid. Set pfault_wait
+			 * back to zero and wake up the process. This can
+			 * safely be done because the task is still sleeping
+			 * and can't produce new pfaults.
+			 */
+			tsk->thread.pfault_wait = 0;
+			list_del(&tsk->thread.list);
+			wake_up_process(tsk);
+			put_task_struct(tsk);
+		} else {
+			/*
+			 * Completion interrupt was faster than initial
+			 * interrupt. Set pfault_wait to -1 so the initial
+			 * interrupt doesn't put the task to sleep.
+			 * If the task is not running, ignore the completion
+			 * interrupt since it must be a leftover of a PFAULT
+			 * CANCEL operation which didn't remove all pending
+			 * completion interrupts.
+			 */
+			if (task_is_running(tsk))
+				tsk->thread.pfault_wait = -1;
+		}
+	} else {
+		/* signal bit not set -> a real page is missing. */
+		if (WARN_ON_ONCE(tsk != current))
+			goto out;
+		if (tsk->thread.pfault_wait == 1) {
+			/* Already on the list with a reference: put to sleep */
+			goto block;
+		} else if (tsk->thread.pfault_wait == -1) {
+			/*
+			 * Completion interrupt was faster than the initial
+			 * interrupt (pfault_wait == -1). Set pfault_wait
+			 * back to zero and exit.
+			 */
+			tsk->thread.pfault_wait = 0;
+		} else {
+			/*
+			 * Initial interrupt arrived before completion
+			 * interrupt. Let the task sleep.
+			 * An extra task reference is needed since a different
+			 * cpu may set the task state to TASK_RUNNING again
+			 * before the scheduler is reached.
+			 */
+			get_task_struct(tsk);
+			tsk->thread.pfault_wait = 1;
+			list_add(&tsk->thread.list, &pfault_list);
+block:
+			/*
+			 * Since this must be a userspace fault, there
+			 * is no kernel task state to trample. Rely on the
+			 * return to userspace schedule() to block.
+			 */
+			__set_current_state(TASK_UNINTERRUPTIBLE);
+			set_need_resched_current();
+		}
+	}
+out:
+	spin_unlock(&pfault_lock);
+	put_task_struct(tsk);
+}
+
+static int pfault_cpu_dead(unsigned int cpu)
+{
+	struct thread_struct *thread, *next;
+	struct task_struct *tsk;
+
+	spin_lock_irq(&pfault_lock);
+	list_for_each_entry_safe(thread, next, &pfault_list, list) {
+		thread->pfault_wait = 0;
+		list_del(&thread->list);
+		tsk = container_of(thread, struct task_struct, thread);
+		wake_up_process(tsk);
+		put_task_struct(tsk);
+	}
+	spin_unlock_irq(&pfault_lock);
+	return 0;
+}
+
+static int __init pfault_irq_init(void)
+{
+	int rc;
+
+	rc = register_external_irq(EXT_IRQ_CP_SERVICE, pfault_interrupt);
+	if (rc)
+		goto out_extint;
+	rc = pfault_init() == 0 ? 0 : -EOPNOTSUPP;
+	if (rc)
+		goto out_pfault;
+	irq_subclass_register(IRQ_SUBCLASS_SERVICE_SIGNAL);
+	cpuhp_setup_state_nocalls(CPUHP_S390_PFAULT_DEAD, "s390/pfault:dead",
+				  NULL, pfault_cpu_dead);
+	return 0;
+
+out_pfault:
+	unregister_external_irq(EXT_IRQ_CP_SERVICE, pfault_interrupt);
+out_extint:
+	pfault_disable = 1;
+	return rc;
+}
+early_initcall(pfault_irq_init);

linux-7.2.3/arch/s390/mm/pgalloc.c

@@ -0,0 +1,457 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  Page table allocation functions
+ *
+ *    Copyright IBM Corp. 2016
+ *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/sysctl.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <asm/mmu_context.h>
+#include <asm/page-states.h>
+#include <asm/pgalloc.h>
+#include <asm/tlbflush.h>
+
+unsigned long *crst_table_alloc_noprof(struct mm_struct *mm)
+{
+	gfp_t gfp = GFP_KERNEL_ACCOUNT;
+	struct ptdesc *ptdesc;
+	unsigned long *table;
+
+	if (mm == &init_mm)
+		gfp &= ~__GFP_ACCOUNT;
+	ptdesc = pagetable_alloc_noprof(gfp, CRST_ALLOC_ORDER);
+	if (!ptdesc)
+		return NULL;
+	table = ptdesc_address(ptdesc);
+	__arch_set_page_dat(table, 1UL << CRST_ALLOC_ORDER);
+	return table;
+}
+
+void crst_table_free(struct mm_struct *mm, unsigned long *table)
+{
+	if (!table)
+		return;
+	pagetable_free(virt_to_ptdesc(table));
+}
+
+static void __crst_table_upgrade(void *arg)
+{
+	struct mm_struct *mm = arg;
+	struct ctlreg asce;
+
+	/* change all active ASCEs to avoid the creation of new TLBs */
+	if (current->active_mm == mm) {
+		asce.val = mm->context.asce;
+		get_lowcore()->user_asce = asce;
+		local_ctl_load(7, &asce);
+		if (!test_thread_flag(TIF_ASCE_PRIMARY))
+			local_ctl_load(1, &asce);
+	}
+	__tlb_flush_local();
+}
+
+int crst_table_upgrade(struct mm_struct *mm, unsigned long end)
+{
+	unsigned long *pgd = NULL, *p4d = NULL, *__pgd;
+	unsigned long asce_limit = mm->context.asce_limit;
+
+	mmap_assert_write_locked(mm);
+
+	/* upgrade should only happen from 3 to 4, 3 to 5, or 4 to 5 levels */
+	VM_BUG_ON(asce_limit < _REGION2_SIZE);
+
+	if (end <= asce_limit)
+		return 0;
+
+	if (asce_limit == _REGION2_SIZE) {
+		p4d = crst_table_alloc(mm);
+		if (unlikely(!p4d))
+			goto err_p4d;
+		crst_table_init(p4d, _REGION2_ENTRY_EMPTY);
+		pagetable_p4d_ctor(virt_to_ptdesc(p4d));
+	}
+	if (end > _REGION1_SIZE) {
+		pgd = crst_table_alloc(mm);
+		if (unlikely(!pgd))
+			goto err_pgd;
+		crst_table_init(pgd, _REGION1_ENTRY_EMPTY);
+		pagetable_pgd_ctor(virt_to_ptdesc(pgd));
+	}
+
+	spin_lock_bh(&mm->page_table_lock);
+
+	if (p4d) {
+		__pgd = (unsigned long *) mm->pgd;
+		p4d_populate(mm, (p4d_t *) p4d, (pud_t *) __pgd);
+		mm->pgd = (pgd_t *) p4d;
+		mm->context.asce_limit = _REGION1_SIZE;
+		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
+			_ASCE_USER_BITS | _ASCE_TYPE_REGION2;
+		mm_inc_nr_puds(mm);
+	}
+	if (pgd) {
+		__pgd = (unsigned long *) mm->pgd;
+		pgd_populate(mm, (pgd_t *) pgd, (p4d_t *) __pgd);
+		mm->pgd = (pgd_t *) pgd;
+		mm->context.asce_limit = TASK_SIZE_MAX;
+		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
+			_ASCE_USER_BITS | _ASCE_TYPE_REGION1;
+	}
+
+	spin_unlock_bh(&mm->page_table_lock);
+
+	on_each_cpu(__crst_table_upgrade, mm, 0);
+
+	return 0;
+
+err_pgd:
+	pagetable_dtor(virt_to_ptdesc(p4d));
+	crst_table_free(mm, p4d);
+err_p4d:
+	return -ENOMEM;
+}
+
+unsigned long *page_table_alloc_noprof(struct mm_struct *mm)
+{
+	gfp_t gfp = GFP_KERNEL_ACCOUNT;
+	struct ptdesc *ptdesc;
+	unsigned long *table;
+
+	if (mm == &init_mm)
+		gfp &= ~__GFP_ACCOUNT;
+	ptdesc = pagetable_alloc_noprof(gfp, 0);
+	if (!ptdesc)
+		return NULL;
+	if (!pagetable_pte_ctor(mm, ptdesc)) {
+		pagetable_free(ptdesc);
+		return NULL;
+	}
+	table = ptdesc_address(ptdesc);
+	__arch_set_page_dat(table, 1);
+	memset64((u64 *)table, _PAGE_INVALID, PTRS_PER_PTE);
+	memset64((u64 *)table + PTRS_PER_PTE, 0, PTRS_PER_PTE);
+	return table;
+}
+
+void page_table_free(struct mm_struct *mm, unsigned long *table)
+{
+	struct ptdesc *ptdesc = virt_to_ptdesc(table);
+
+	if (pagetable_is_reserved(ptdesc))
+		return free_reserved_ptdesc(ptdesc);
+	pagetable_dtor_free(ptdesc);
+}
+
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+static void pte_free_now(struct rcu_head *head)
+{
+	struct ptdesc *ptdesc = container_of(head, struct ptdesc, pt_rcu_head);
+
+	pagetable_dtor_free(ptdesc);
+}
+
+void pte_free_defer(struct mm_struct *mm, pgtable_t pgtable)
+{
+	struct ptdesc *ptdesc = virt_to_ptdesc(pgtable);
+
+	call_rcu(&ptdesc->pt_rcu_head, pte_free_now);
+}
+#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+
+/*
+ * Base infrastructure required to generate basic asces, region, segment,
+ * and page tables that do not make use of enhanced features like EDAT1.
+ */
+
+static struct kmem_cache *base_pgt_cache;
+
+static unsigned long *base_pgt_alloc(void)
+{
+	unsigned long *table;
+
+	table = kmem_cache_alloc(base_pgt_cache, GFP_KERNEL);
+	if (table)
+		memset64((u64 *)table, _PAGE_INVALID, PTRS_PER_PTE);
+	return table;
+}
+
+static void base_pgt_free(unsigned long *table)
+{
+	kmem_cache_free(base_pgt_cache, table);
+}
+
+static unsigned long *base_crst_alloc(unsigned long val)
+{
+	unsigned long *table;
+	struct ptdesc *ptdesc;
+
+	ptdesc = pagetable_alloc(GFP_KERNEL, CRST_ALLOC_ORDER);
+	if (!ptdesc)
+		return NULL;
+	table = ptdesc_address(ptdesc);
+	crst_table_init(table, val);
+	return table;
+}
+
+static void base_crst_free(unsigned long *table)
+{
+	if (!table)
+		return;
+	pagetable_free(virt_to_ptdesc(table));
+}
+
+#define BASE_ADDR_END_FUNC(NAME, SIZE)					\
+static inline unsigned long base_##NAME##_addr_end(unsigned long addr,	\
+						   unsigned long end)	\
+{									\
+	unsigned long next = (addr + (SIZE)) & ~((SIZE) - 1);		\
+									\
+	return (next - 1) < (end - 1) ? next : end;			\
+}
+
+BASE_ADDR_END_FUNC(page,    PAGE_SIZE)
+BASE_ADDR_END_FUNC(segment, _SEGMENT_SIZE)
+BASE_ADDR_END_FUNC(region3, _REGION3_SIZE)
+BASE_ADDR_END_FUNC(region2, _REGION2_SIZE)
+BASE_ADDR_END_FUNC(region1, _REGION1_SIZE)
+
+static inline unsigned long base_lra(unsigned long address)
+{
+	unsigned long real;
+
+	asm volatile(
+		"	lra	%0,0(%1)"
+		: "=d" (real) : "a" (address) : "cc");
+	return real;
+}
+
+static int base_page_walk(unsigned long *origin, unsigned long addr,
+			  unsigned long end, int alloc)
+{
+	unsigned long *pte, next;
+
+	if (!alloc)
+		return 0;
+	pte = origin;
+	pte += (addr & _PAGE_INDEX) >> PAGE_SHIFT;
+	do {
+		next = base_page_addr_end(addr, end);
+		*pte = base_lra(addr);
+	} while (pte++, addr = next, addr < end);
+	return 0;
+}
+
+static int base_segment_walk(unsigned long *origin, unsigned long addr,
+			     unsigned long end, int alloc)
+{
+	unsigned long *ste, next, *table;
+	int rc;
+
+	ste = origin;
+	ste += (addr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT;
+	do {
+		next = base_segment_addr_end(addr, end);
+		if (*ste & _SEGMENT_ENTRY_INVALID) {
+			if (!alloc)
+				continue;
+			table = base_pgt_alloc();
+			if (!table)
+				return -ENOMEM;
+			*ste = __pa(table) | _SEGMENT_ENTRY;
+		}
+		table = __va(*ste & _SEGMENT_ENTRY_ORIGIN);
+		rc = base_page_walk(table, addr, next, alloc);
+		if (rc)
+			return rc;
+		if (!alloc)
+			base_pgt_free(table);
+		cond_resched();
+	} while (ste++, addr = next, addr < end);
+	return 0;
+}
+
+static int base_region3_walk(unsigned long *origin, unsigned long addr,
+			     unsigned long end, int alloc)
+{
+	unsigned long *rtte, next, *table;
+	int rc;
+
+	rtte = origin;
+	rtte += (addr & _REGION3_INDEX) >> _REGION3_SHIFT;
+	do {
+		next = base_region3_addr_end(addr, end);
+		if (*rtte & _REGION_ENTRY_INVALID) {
+			if (!alloc)
+				continue;
+			table = base_crst_alloc(_SEGMENT_ENTRY_EMPTY);
+			if (!table)
+				return -ENOMEM;
+			*rtte = __pa(table) | _REGION3_ENTRY;
+		}
+		table = __va(*rtte & _REGION_ENTRY_ORIGIN);
+		rc = base_segment_walk(table, addr, next, alloc);
+		if (rc)
+			return rc;
+		if (!alloc)
+			base_crst_free(table);
+	} while (rtte++, addr = next, addr < end);
+	return 0;
+}
+
+static int base_region2_walk(unsigned long *origin, unsigned long addr,
+			     unsigned long end, int alloc)
+{
+	unsigned long *rste, next, *table;
+	int rc;
+
+	rste = origin;
+	rste += (addr & _REGION2_INDEX) >> _REGION2_SHIFT;
+	do {
+		next = base_region2_addr_end(addr, end);
+		if (*rste & _REGION_ENTRY_INVALID) {
+			if (!alloc)
+				continue;
+			table = base_crst_alloc(_REGION3_ENTRY_EMPTY);
+			if (!table)
+				return -ENOMEM;
+			*rste = __pa(table) | _REGION2_ENTRY;
+		}
+		table = __va(*rste & _REGION_ENTRY_ORIGIN);
+		rc = base_region3_walk(table, addr, next, alloc);
+		if (rc)
+			return rc;
+		if (!alloc)
+			base_crst_free(table);
+	} while (rste++, addr = next, addr < end);
+	return 0;
+}
+
+static int base_region1_walk(unsigned long *origin, unsigned long addr,
+			     unsigned long end, int alloc)
+{
+	unsigned long *rfte, next, *table;
+	int rc;
+
+	rfte = origin;
+	rfte += (addr & _REGION1_INDEX) >> _REGION1_SHIFT;
+	do {
+		next = base_region1_addr_end(addr, end);
+		if (*rfte & _REGION_ENTRY_INVALID) {
+			if (!alloc)
+				continue;
+			table = base_crst_alloc(_REGION2_ENTRY_EMPTY);
+			if (!table)
+				return -ENOMEM;
+			*rfte = __pa(table) | _REGION1_ENTRY;
+		}
+		table = __va(*rfte & _REGION_ENTRY_ORIGIN);
+		rc = base_region2_walk(table, addr, next, alloc);
+		if (rc)
+			return rc;
+		if (!alloc)
+			base_crst_free(table);
+	} while (rfte++, addr = next, addr < end);
+	return 0;
+}
+
+/**
+ * base_asce_free - free asce and tables returned from base_asce_alloc()
+ * @asce: asce to be freed
+ *
+ * Frees all region, segment, and page tables that were allocated with a
+ * corresponding base_asce_alloc() call.
+ */
+void base_asce_free(unsigned long asce)
+{
+	unsigned long *table = __va(asce & _ASCE_ORIGIN);
+
+	if (!asce)
+		return;
+	switch (asce & _ASCE_TYPE_MASK) {
+	case _ASCE_TYPE_SEGMENT:
+		base_segment_walk(table, 0, _REGION3_SIZE, 0);
+		break;
+	case _ASCE_TYPE_REGION3:
+		base_region3_walk(table, 0, _REGION2_SIZE, 0);
+		break;
+	case _ASCE_TYPE_REGION2:
+		base_region2_walk(table, 0, _REGION1_SIZE, 0);
+		break;
+	case _ASCE_TYPE_REGION1:
+		base_region1_walk(table, 0, TASK_SIZE_MAX, 0);
+		break;
+	}
+	base_crst_free(table);
+}
+
+static int base_pgt_cache_init(void)
+{
+	static DEFINE_MUTEX(base_pgt_cache_mutex);
+	unsigned long sz = _PAGE_TABLE_SIZE;
+
+	if (base_pgt_cache)
+		return 0;
+	mutex_lock(&base_pgt_cache_mutex);
+	if (!base_pgt_cache)
+		base_pgt_cache = kmem_cache_create("base_pgt", sz, sz, 0, NULL);
+	mutex_unlock(&base_pgt_cache_mutex);
+	return base_pgt_cache ? 0 : -ENOMEM;
+}
+
+/**
+ * base_asce_alloc - create kernel mapping without enhanced DAT features
+ * @addr: virtual start address of kernel mapping
+ * @num_pages: number of consecutive pages
+ *
+ * Generate an asce, including all required region, segment and page tables,
+ * that can be used to access the virtual kernel mapping. The difference is
+ * that the returned asce does not make use of any enhanced DAT features like
+ * e.g. large pages. This is required for some I/O functions that pass an
+ * asce, like e.g. some service call requests.
+ *
+ * Note: the returned asce may NEVER be attached to any cpu. It may only be
+ *	 used for I/O requests. tlb entries that might result because the
+ *	 asce was attached to a cpu won't be cleared.
+ */
+unsigned long base_asce_alloc(unsigned long addr, unsigned long num_pages)
+{
+	unsigned long asce, *table, end;
+	int rc;
+
+	if (base_pgt_cache_init())
+		return 0;
+	end = addr + num_pages * PAGE_SIZE;
+	if (end <= _REGION3_SIZE) {
+		table = base_crst_alloc(_SEGMENT_ENTRY_EMPTY);
+		if (!table)
+			return 0;
+		rc = base_segment_walk(table, addr, end, 1);
+		asce = __pa(table) | _ASCE_TYPE_SEGMENT | _ASCE_TABLE_LENGTH;
+	} else if (end <= _REGION2_SIZE) {
+		table = base_crst_alloc(_REGION3_ENTRY_EMPTY);
+		if (!table)
+			return 0;
+		rc = base_region3_walk(table, addr, end, 1);
+		asce = __pa(table) | _ASCE_TYPE_REGION3 | _ASCE_TABLE_LENGTH;
+	} else if (end <= _REGION1_SIZE) {
+		table = base_crst_alloc(_REGION2_ENTRY_EMPTY);
+		if (!table)
+			return 0;
+		rc = base_region2_walk(table, addr, end, 1);
+		asce = __pa(table) | _ASCE_TYPE_REGION2 | _ASCE_TABLE_LENGTH;
+	} else {
+		table = base_crst_alloc(_REGION1_ENTRY_EMPTY);
+		if (!table)
+			return 0;
+		rc = base_region1_walk(table, addr, end, 1);
+		asce = __pa(table) | _ASCE_TYPE_REGION1 | _ASCE_TABLE_LENGTH;
+	}
+	if (rc) {
+		base_asce_free(asce);
+		asce = 0;
+	}
+	return asce;
+}

linux-7.2.3/arch/s390/mm/pgtable.c

@@ -0,0 +1,355 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *    Copyright IBM Corp. 2007, 2011
+ *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/cpufeature.h>
+#include <linux/export.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/gfp.h>
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/smp.h>
+#include <linux/spinlock.h>
+#include <linux/rcupdate.h>
+#include <linux/slab.h>
+#include <linux/leafops.h>
+#include <linux/sysctl.h>
+#include <linux/ksm.h>
+#include <linux/mman.h>
+
+#include <asm/tlbflush.h>
+#include <asm/mmu_context.h>
+#include <asm/page-states.h>
+#include <asm/machine.h>
+
+pgprot_t pgprot_writecombine(pgprot_t prot)
+{
+	/*
+	 * mio_wb_bit_mask may be set on a different CPU, but it is only set
+	 * once at init and only read afterwards.
+	 */
+	return __pgprot(pgprot_val(prot) | mio_wb_bit_mask);
+}
+EXPORT_SYMBOL_GPL(pgprot_writecombine);
+
+static inline void ptep_ipte_local(struct mm_struct *mm, unsigned long addr,
+				   pte_t *ptep, int nodat)
+{
+	unsigned long opt, asce;
+
+	if (machine_has_tlb_guest()) {
+		opt = 0;
+		asce = READ_ONCE(mm->context.gmap_asce);
+		if (asce == 0UL || nodat)
+			opt |= IPTE_NODAT;
+		if (asce != -1UL) {
+			asce = asce ? : mm->context.asce;
+			opt |= IPTE_GUEST_ASCE;
+		}
+		__ptep_ipte(addr, ptep, opt, asce, IPTE_LOCAL);
+	} else {
+		__ptep_ipte(addr, ptep, 0, 0, IPTE_LOCAL);
+	}
+}
+
+static inline void ptep_ipte_global(struct mm_struct *mm, unsigned long addr,
+				    pte_t *ptep, int nodat)
+{
+	unsigned long opt, asce;
+
+	if (machine_has_tlb_guest()) {
+		opt = 0;
+		asce = READ_ONCE(mm->context.gmap_asce);
+		if (asce == 0UL || nodat)
+			opt |= IPTE_NODAT;
+		if (asce != -1UL) {
+			asce = asce ? : mm->context.asce;
+			opt |= IPTE_GUEST_ASCE;
+		}
+		__ptep_ipte(addr, ptep, opt, asce, IPTE_GLOBAL);
+	} else {
+		__ptep_ipte(addr, ptep, 0, 0, IPTE_GLOBAL);
+	}
+}
+
+static inline pte_t ptep_flush_direct(struct mm_struct *mm,
+				      unsigned long addr, pte_t *ptep,
+				      int nodat)
+{
+	pte_t old;
+
+	old = *ptep;
+	if (unlikely(pte_val(old) & _PAGE_INVALID))
+		return old;
+	atomic_inc(&mm->context.flush_count);
+	if (cpu_has_tlb_lc() &&
+	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id())))
+		ptep_ipte_local(mm, addr, ptep, nodat);
+	else
+		ptep_ipte_global(mm, addr, ptep, nodat);
+	atomic_dec(&mm->context.flush_count);
+	return old;
+}
+
+static inline pte_t ptep_flush_lazy(struct mm_struct *mm,
+				    unsigned long addr, pte_t *ptep,
+				    int nodat)
+{
+	pte_t old;
+
+	old = *ptep;
+	if (unlikely(pte_val(old) & _PAGE_INVALID))
+		return old;
+	atomic_inc(&mm->context.flush_count);
+	if (cpumask_equal(&mm->context.cpu_attach_mask,
+			  cpumask_of(smp_processor_id()))) {
+		set_pte(ptep, set_pte_bit(*ptep, __pgprot(_PAGE_INVALID)));
+		mm->context.flush_mm = 1;
+	} else
+		ptep_ipte_global(mm, addr, ptep, nodat);
+	atomic_dec(&mm->context.flush_count);
+	return old;
+}
+
+pte_t ptep_xchg_direct(struct mm_struct *mm, unsigned long addr,
+		       pte_t *ptep, pte_t new)
+{
+	pte_t old;
+
+	preempt_disable();
+	old = ptep_flush_direct(mm, addr, ptep, 1);
+	set_pte(ptep, new);
+	preempt_enable();
+	return old;
+}
+EXPORT_SYMBOL(ptep_xchg_direct);
+
+/*
+ * Caller must check that new PTE only differs in _PAGE_PROTECT HW bit, so that
+ * RDP can be used instead of IPTE. See also comments at pte_allow_rdp().
+ */
+void ptep_reset_dat_prot(struct mm_struct *mm, unsigned long addr, pte_t *ptep,
+			 pte_t new)
+{
+	preempt_disable();
+	atomic_inc(&mm->context.flush_count);
+	if (cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id())))
+		__ptep_rdp(addr, ptep, 1);
+	else
+		__ptep_rdp(addr, ptep, 0);
+	/*
+	 * PTE is not invalidated by RDP, only _PAGE_PROTECT is cleared. That
+	 * means it is still valid and active, and must not be changed according
+	 * to the architecture. But writing a new value that only differs in SW
+	 * bits is allowed.
+	 */
+	set_pte(ptep, new);
+	atomic_dec(&mm->context.flush_count);
+	preempt_enable();
+}
+EXPORT_SYMBOL(ptep_reset_dat_prot);
+
+pte_t ptep_xchg_lazy(struct mm_struct *mm, unsigned long addr,
+		     pte_t *ptep, pte_t new)
+{
+	pte_t old;
+
+	preempt_disable();
+	old = ptep_flush_lazy(mm, addr, ptep, 1);
+	set_pte(ptep, new);
+	preempt_enable();
+	return old;
+}
+EXPORT_SYMBOL(ptep_xchg_lazy);
+
+pte_t ptep_modify_prot_start(struct vm_area_struct *vma, unsigned long addr,
+			     pte_t *ptep)
+{
+	return ptep_flush_lazy(vma->vm_mm, addr, ptep, 1);
+}
+
+void ptep_modify_prot_commit(struct vm_area_struct *vma, unsigned long addr,
+			     pte_t *ptep, pte_t old_pte, pte_t pte)
+{
+	set_pte(ptep, pte);
+}
+
+static inline void pmdp_idte_local(struct mm_struct *mm,
+				   unsigned long addr, pmd_t *pmdp)
+{
+	if (machine_has_tlb_guest())
+		__pmdp_idte(addr, pmdp, IDTE_NODAT | IDTE_GUEST_ASCE, mm->context.asce, IDTE_LOCAL);
+	else
+		__pmdp_idte(addr, pmdp, 0, 0, IDTE_LOCAL);
+}
+
+static inline void pmdp_idte_global(struct mm_struct *mm,
+				    unsigned long addr, pmd_t *pmdp)
+{
+	if (machine_has_tlb_guest()) {
+		__pmdp_idte(addr, pmdp, IDTE_NODAT | IDTE_GUEST_ASCE,
+			    mm->context.asce, IDTE_GLOBAL);
+	} else {
+		__pmdp_idte(addr, pmdp, 0, 0, IDTE_GLOBAL);
+	}
+}
+
+static inline pmd_t pmdp_flush_direct(struct mm_struct *mm,
+				      unsigned long addr, pmd_t *pmdp)
+{
+	pmd_t old;
+
+	old = *pmdp;
+	if (pmd_val(old) & _SEGMENT_ENTRY_INVALID)
+		return old;
+	atomic_inc(&mm->context.flush_count);
+	if (cpu_has_tlb_lc() &&
+	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id())))
+		pmdp_idte_local(mm, addr, pmdp);
+	else
+		pmdp_idte_global(mm, addr, pmdp);
+	atomic_dec(&mm->context.flush_count);
+	return old;
+}
+
+static inline pmd_t pmdp_flush_lazy(struct mm_struct *mm,
+				    unsigned long addr, pmd_t *pmdp)
+{
+	pmd_t old;
+
+	old = *pmdp;
+	if (pmd_val(old) & _SEGMENT_ENTRY_INVALID)
+		return old;
+	atomic_inc(&mm->context.flush_count);
+	if (cpumask_equal(&mm->context.cpu_attach_mask,
+			  cpumask_of(smp_processor_id()))) {
+		set_pmd(pmdp, set_pmd_bit(*pmdp, __pgprot(_SEGMENT_ENTRY_INVALID)));
+		mm->context.flush_mm = 1;
+	} else {
+		pmdp_idte_global(mm, addr, pmdp);
+	}
+	atomic_dec(&mm->context.flush_count);
+	return old;
+}
+
+pmd_t pmdp_xchg_direct(struct mm_struct *mm, unsigned long addr,
+		       pmd_t *pmdp, pmd_t new)
+{
+	pmd_t old;
+
+	preempt_disable();
+	old = pmdp_flush_direct(mm, addr, pmdp);
+	set_pmd(pmdp, new);
+	preempt_enable();
+	return old;
+}
+EXPORT_SYMBOL(pmdp_xchg_direct);
+
+pmd_t pmdp_xchg_lazy(struct mm_struct *mm, unsigned long addr,
+		     pmd_t *pmdp, pmd_t new)
+{
+	pmd_t old;
+
+	preempt_disable();
+	old = pmdp_flush_lazy(mm, addr, pmdp);
+	set_pmd(pmdp, new);
+	preempt_enable();
+	return old;
+}
+EXPORT_SYMBOL(pmdp_xchg_lazy);
+
+static inline void pudp_idte_local(struct mm_struct *mm,
+				   unsigned long addr, pud_t *pudp)
+{
+	if (machine_has_tlb_guest())
+		__pudp_idte(addr, pudp, IDTE_NODAT | IDTE_GUEST_ASCE,
+			    mm->context.asce, IDTE_LOCAL);
+	else
+		__pudp_idte(addr, pudp, 0, 0, IDTE_LOCAL);
+}
+
+static inline void pudp_idte_global(struct mm_struct *mm,
+				    unsigned long addr, pud_t *pudp)
+{
+	if (machine_has_tlb_guest())
+		__pudp_idte(addr, pudp, IDTE_NODAT | IDTE_GUEST_ASCE,
+			    mm->context.asce, IDTE_GLOBAL);
+	else
+		__pudp_idte(addr, pudp, 0, 0, IDTE_GLOBAL);
+}
+
+static inline pud_t pudp_flush_direct(struct mm_struct *mm,
+				      unsigned long addr, pud_t *pudp)
+{
+	pud_t old;
+
+	old = *pudp;
+	if (pud_val(old) & _REGION_ENTRY_INVALID)
+		return old;
+	atomic_inc(&mm->context.flush_count);
+	if (cpu_has_tlb_lc() &&
+	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id())))
+		pudp_idte_local(mm, addr, pudp);
+	else
+		pudp_idte_global(mm, addr, pudp);
+	atomic_dec(&mm->context.flush_count);
+	return old;
+}
+
+pud_t pudp_xchg_direct(struct mm_struct *mm, unsigned long addr,
+		       pud_t *pudp, pud_t new)
+{
+	pud_t old;
+
+	preempt_disable();
+	old = pudp_flush_direct(mm, addr, pudp);
+	set_pud(pudp, new);
+	preempt_enable();
+	return old;
+}
+EXPORT_SYMBOL(pudp_xchg_direct);
+
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+void pgtable_trans_huge_deposit(struct mm_struct *mm, pmd_t *pmdp,
+				pgtable_t pgtable)
+{
+	struct list_head *lh = (struct list_head *) pgtable;
+
+	assert_spin_locked(pmd_lockptr(mm, pmdp));
+
+	/* FIFO */
+	if (!pmd_huge_pte(mm, pmdp))
+		INIT_LIST_HEAD(lh);
+	else
+		list_add(lh, (struct list_head *) pmd_huge_pte(mm, pmdp));
+	pmd_huge_pte(mm, pmdp) = pgtable;
+}
+
+pgtable_t pgtable_trans_huge_withdraw(struct mm_struct *mm, pmd_t *pmdp)
+{
+	struct list_head *lh;
+	pgtable_t pgtable;
+	pte_t *ptep;
+
+	assert_spin_locked(pmd_lockptr(mm, pmdp));
+
+	/* FIFO */
+	pgtable = pmd_huge_pte(mm, pmdp);
+	lh = (struct list_head *) pgtable;
+	if (list_empty(lh))
+		pmd_huge_pte(mm, pmdp) = NULL;
+	else {
+		pmd_huge_pte(mm, pmdp) = (pgtable_t) lh->next;
+		list_del(lh);
+	}
+	ptep = (pte_t *) pgtable;
+	set_pte(ptep, __pte(_PAGE_INVALID));
+	ptep++;
+	set_pte(ptep, __pte(_PAGE_INVALID));
+	return pgtable;
+}
+#endif /* CONFIG_TRANSPARENT_HUGEPAGE */

linux-7.2.3/arch/s390/mm/physaddr.c

@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/mmdebug.h>
+#include <linux/export.h>
+#include <linux/mm.h>
+#include <asm/page.h>
+
+unsigned long __phys_addr(unsigned long x, bool is_31bit)
+{
+	VIRTUAL_BUG_ON(is_vmalloc_or_module_addr((void *)(x)));
+	x = __pa_nodebug(x);
+	if (is_31bit)
+		VIRTUAL_BUG_ON(x >> 31);
+	return x;
+}
+EXPORT_SYMBOL(__phys_addr);

linux-7.2.3/arch/s390/mm/vmem.c

@@ -0,0 +1,690 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *    Copyright IBM Corp. 2006
+ */
+
+#include <linux/memory_hotplug.h>
+#include <linux/cpufeature.h>
+#include <linux/memblock.h>
+#include <linux/pfn.h>
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <linux/hugetlb.h>
+#include <linux/slab.h>
+#include <linux/sort.h>
+#include <asm/page-states.h>
+#include <asm/abs_lowcore.h>
+#include <asm/cacheflush.h>
+#include <asm/maccess.h>
+#include <asm/nospec-branch.h>
+#include <asm/ctlreg.h>
+#include <asm/pgalloc.h>
+#include <asm/setup.h>
+#include <asm/tlbflush.h>
+#include <asm/sections.h>
+#include <asm/set_memory.h>
+#include <asm/physmem_info.h>
+
+static DEFINE_MUTEX(vmem_mutex);
+
+static void __ref *vmem_alloc_pages(unsigned int order)
+{
+	unsigned long size = PAGE_SIZE << order;
+
+	if (slab_is_available())
+		return (void *)__get_free_pages(GFP_KERNEL, order);
+	return memblock_alloc(size, size);
+}
+
+static void vmem_free_pages(unsigned long addr, int order, struct vmem_altmap *altmap)
+{
+	unsigned int nr_pages = 1 << order;
+	struct page *page;
+
+	if (altmap) {
+		vmem_altmap_free(altmap, 1 << order);
+		return;
+	}
+	page = virt_to_page((void *)addr);
+	if (PageReserved(page)) {
+		/* allocated from memblock */
+		while (nr_pages--)
+			free_reserved_page(page++);
+	} else {
+		free_pages(addr, order);
+	}
+}
+
+void *vmem_crst_alloc(unsigned long val)
+{
+	unsigned long *table;
+
+	table = vmem_alloc_pages(CRST_ALLOC_ORDER);
+	if (!table)
+		return NULL;
+	crst_table_init(table, val);
+	__arch_set_page_dat(table, 1UL << CRST_ALLOC_ORDER);
+	return table;
+}
+
+pte_t __ref *vmem_pte_alloc(void)
+{
+	pte_t *pte;
+
+	if (slab_is_available())
+		pte = (pte_t *)page_table_alloc(&init_mm);
+	else
+		pte = (pte_t *)memblock_alloc(PAGE_SIZE, PAGE_SIZE);
+	if (!pte)
+		return NULL;
+	memset64((u64 *)pte, _PAGE_INVALID, PTRS_PER_PTE);
+	__arch_set_page_dat(pte, 1);
+	return pte;
+}
+
+static void vmem_pte_free(unsigned long *table)
+{
+	page_table_free(&init_mm, table);
+}
+
+#define PAGE_UNUSED 0xFD
+
+/*
+ * The unused vmemmap range, which was not yet memset(PAGE_UNUSED) ranges
+ * from unused_sub_pmd_start to next PMD_SIZE boundary.
+ */
+static unsigned long unused_sub_pmd_start;
+
+static void vmemmap_flush_unused_sub_pmd(void)
+{
+	if (!unused_sub_pmd_start)
+		return;
+	memset((void *)unused_sub_pmd_start, PAGE_UNUSED,
+	       ALIGN(unused_sub_pmd_start, PMD_SIZE) - unused_sub_pmd_start);
+	unused_sub_pmd_start = 0;
+}
+
+static void vmemmap_mark_sub_pmd_used(unsigned long start, unsigned long end)
+{
+	/*
+	 * As we expect to add in the same granularity as we remove, it's
+	 * sufficient to mark only some piece used to block the memmap page from
+	 * getting removed (just in case the memmap never gets initialized,
+	 * e.g., because the memory block never gets onlined).
+	 */
+	memset((void *)start, 0, sizeof(struct page));
+}
+
+static void vmemmap_use_sub_pmd(unsigned long start, unsigned long end)
+{
+	/*
+	 * We only optimize if the new used range directly follows the
+	 * previously unused range (esp., when populating consecutive sections).
+	 */
+	if (unused_sub_pmd_start == start) {
+		unused_sub_pmd_start = end;
+		if (likely(IS_ALIGNED(unused_sub_pmd_start, PMD_SIZE)))
+			unused_sub_pmd_start = 0;
+		return;
+	}
+	vmemmap_flush_unused_sub_pmd();
+	vmemmap_mark_sub_pmd_used(start, end);
+}
+
+static void vmemmap_use_new_sub_pmd(unsigned long start, unsigned long end)
+{
+	unsigned long page = ALIGN_DOWN(start, PMD_SIZE);
+
+	vmemmap_flush_unused_sub_pmd();
+
+	/* Could be our memmap page is filled with PAGE_UNUSED already ... */
+	vmemmap_mark_sub_pmd_used(start, end);
+
+	/* Mark the unused parts of the new memmap page PAGE_UNUSED. */
+	if (!IS_ALIGNED(start, PMD_SIZE))
+		memset((void *)page, PAGE_UNUSED, start - page);
+	/*
+	 * We want to avoid memset(PAGE_UNUSED) when populating the vmemmap of
+	 * consecutive sections. Remember for the last added PMD the last
+	 * unused range in the populated PMD.
+	 */
+	if (!IS_ALIGNED(end, PMD_SIZE))
+		unused_sub_pmd_start = end;
+}
+
+/* Returns true if the PMD is completely unused and can be freed. */
+static bool vmemmap_unuse_sub_pmd(unsigned long start, unsigned long end)
+{
+	unsigned long page = ALIGN_DOWN(start, PMD_SIZE);
+
+	vmemmap_flush_unused_sub_pmd();
+	memset((void *)start, PAGE_UNUSED, end - start);
+	return !memchr_inv((void *)page, PAGE_UNUSED, PMD_SIZE);
+}
+
+/* __ref: we'll only call vmemmap_alloc_block() via vmemmap_populate() */
+static int __ref modify_pte_table(pmd_t *pmd, unsigned long addr,
+				  unsigned long end, bool add, bool direct,
+				  struct vmem_altmap *altmap)
+{
+	unsigned long prot, pages = 0;
+	int ret = -ENOMEM;
+	pte_t *pte, entry;
+
+	prot = pgprot_val(PAGE_KERNEL);
+	pte = pte_offset_kernel(pmd, addr);
+	for (; addr < end; addr += PAGE_SIZE, pte++) {
+		entry = ptep_get(pte);
+		if (!add) {
+			if (pte_none(entry))
+				continue;
+			if (!direct)
+				vmem_free_pages((unsigned long)pfn_to_virt(pte_pfn(entry)), get_order(PAGE_SIZE), altmap);
+			pte_clear(&init_mm, addr, pte);
+		} else if (pte_none(entry)) {
+			if (!direct) {
+				void *new_page = vmemmap_alloc_block_buf(PAGE_SIZE, NUMA_NO_NODE, altmap);
+
+				if (!new_page)
+					goto out;
+				set_pte(pte, __pte(__pa(new_page) | prot));
+			} else {
+				set_pte(pte, __pte(__pa(addr) | prot));
+			}
+		} else {
+			continue;
+		}
+		pages++;
+	}
+	ret = 0;
+out:
+	if (direct)
+		update_page_count(PG_DIRECT_MAP_4K, add ? pages : -pages);
+	return ret;
+}
+
+static void try_free_pte_table(pmd_t *pmd, unsigned long start)
+{
+	pte_t *pte;
+	int i;
+
+	/* We can safely assume this is fully in 1:1 mapping & vmemmap area */
+	pte = pte_offset_kernel(pmd, start);
+	for (i = 0; i < PTRS_PER_PTE; i++, pte++) {
+		if (!pte_none(ptep_get(pte)))
+			return;
+	}
+	vmem_pte_free((unsigned long *)pmd_deref(pmdp_get(pmd)));
+	pmd_clear(pmd);
+}
+
+/* __ref: we'll only call vmemmap_alloc_block() via vmemmap_populate() */
+static int __ref modify_pmd_table(pud_t *pud, unsigned long addr,
+				  unsigned long end, bool add, bool direct,
+				  struct vmem_altmap *altmap)
+{
+	unsigned long next, prot, pages = 0;
+	int ret = -ENOMEM;
+	pmd_t entry;
+	pmd_t *pmd;
+	pte_t *pte;
+
+	prot = pgprot_val(SEGMENT_KERNEL);
+	pmd = pmd_offset(pud, addr);
+	for (; addr < end; addr = next, pmd++) {
+		next = pmd_addr_end(addr, end);
+		entry = pmdp_get(pmd);
+		if (!add) {
+			if (pmd_none(entry))
+				continue;
+			if (pmd_leaf(entry)) {
+				if (IS_ALIGNED(addr, PMD_SIZE) &&
+				    IS_ALIGNED(next, PMD_SIZE)) {
+					if (!direct)
+						vmem_free_pages(pmd_deref(entry), get_order(PMD_SIZE), altmap);
+					pmd_clear(pmd);
+					pages++;
+				} else if (!direct && vmemmap_unuse_sub_pmd(addr, next)) {
+					vmem_free_pages(pmd_deref(entry), get_order(PMD_SIZE), altmap);
+					pmd_clear(pmd);
+				}
+				continue;
+			}
+		} else if (pmd_none(entry)) {
+			if (IS_ALIGNED(addr, PMD_SIZE) &&
+			    IS_ALIGNED(next, PMD_SIZE) &&
+			    cpu_has_edat1() && direct &&
+			    !debug_pagealloc_enabled()) {
+				set_pmd(pmd, __pmd(__pa(addr) | prot));
+				pages++;
+				continue;
+			} else if (!direct && cpu_has_edat1()) {
+				void *new_page;
+
+				/*
+				 * Use 1MB frames for vmemmap if available. We
+				 * always use large frames even if they are only
+				 * partially used. Otherwise we would have also
+				 * page tables since vmemmap_populate gets
+				 * called for each section separately.
+				 */
+				new_page = vmemmap_alloc_block_buf(PMD_SIZE, NUMA_NO_NODE, altmap);
+				if (new_page) {
+					set_pmd(pmd, __pmd(__pa(new_page) | prot));
+					if (!IS_ALIGNED(addr, PMD_SIZE) ||
+					    !IS_ALIGNED(next, PMD_SIZE)) {
+						vmemmap_use_new_sub_pmd(addr, next);
+					}
+					continue;
+				}
+			}
+			pte = vmem_pte_alloc();
+			if (!pte)
+				goto out;
+			pmd_populate(&init_mm, pmd, pte);
+		} else if (pmd_leaf(entry)) {
+			if (!direct)
+				vmemmap_use_sub_pmd(addr, next);
+			continue;
+		}
+		ret = modify_pte_table(pmd, addr, next, add, direct, altmap);
+		if (ret)
+			goto out;
+		if (!add)
+			try_free_pte_table(pmd, addr & PMD_MASK);
+	}
+	ret = 0;
+out:
+	if (direct)
+		update_page_count(PG_DIRECT_MAP_1M, add ? pages : -pages);
+	return ret;
+}
+
+static void try_free_pmd_table(pud_t *pud, unsigned long start)
+{
+	pmd_t *pmd;
+	int i;
+
+	pmd = pmd_offset(pud, start);
+	for (i = 0; i < PTRS_PER_PMD; i++, pmd++)
+		if (!pmd_none(pmdp_get(pmd)))
+			return;
+	vmem_free_pages(pud_deref(pudp_get(pud)), CRST_ALLOC_ORDER, NULL);
+	pud_clear(pud);
+}
+
+static int modify_pud_table(p4d_t *p4d, unsigned long addr, unsigned long end,
+			    bool add, bool direct, struct vmem_altmap *altmap)
+{
+	unsigned long next, prot, pages = 0;
+	int ret = -ENOMEM;
+	pud_t *pud, entry;
+	pmd_t *pmd;
+
+	prot = pgprot_val(REGION3_KERNEL);
+	pud = pud_offset(p4d, addr);
+	for (; addr < end; addr = next, pud++) {
+		next = pud_addr_end(addr, end);
+		entry = pudp_get(pud);
+		if (!add) {
+			if (pud_none(entry))
+				continue;
+			if (pud_leaf(entry)) {
+				if (IS_ALIGNED(addr, PUD_SIZE) &&
+				    IS_ALIGNED(next, PUD_SIZE)) {
+					if (!direct)
+						vmem_free_pages(pud_deref(entry), get_order(PUD_SIZE), altmap);
+					pud_clear(pud);
+					pages++;
+					continue;
+				} else {
+					split_pud_page(pud, addr & PUD_MASK);
+				}
+			}
+		} else if (pud_none(entry)) {
+			if (IS_ALIGNED(addr, PUD_SIZE) &&
+			    IS_ALIGNED(next, PUD_SIZE) &&
+			    cpu_has_edat2() && direct &&
+			    !debug_pagealloc_enabled()) {
+				set_pud(pud, __pud(__pa(addr) | prot));
+				pages++;
+				continue;
+			}
+			pmd = vmem_crst_alloc(_SEGMENT_ENTRY_EMPTY);
+			if (!pmd)
+				goto out;
+			pud_populate(&init_mm, pud, pmd);
+		} else if (pud_leaf(entry)) {
+			continue;
+		}
+		ret = modify_pmd_table(pud, addr, next, add, direct, altmap);
+		if (ret)
+			goto out;
+		if (!add)
+			try_free_pmd_table(pud, addr & PUD_MASK);
+	}
+	ret = 0;
+out:
+	if (direct)
+		update_page_count(PG_DIRECT_MAP_2G, add ? pages : -pages);
+	return ret;
+}
+
+static void try_free_pud_table(p4d_t *p4d, unsigned long start)
+{
+	pud_t *pud;
+	int i;
+
+	pud = pud_offset(p4d, start);
+	for (i = 0; i < PTRS_PER_PUD; i++, pud++) {
+		if (!pud_none(pudp_get(pud)))
+			return;
+	}
+	vmem_free_pages(p4d_deref(p4dp_get(p4d)), CRST_ALLOC_ORDER, NULL);
+	p4d_clear(p4d);
+}
+
+static int modify_p4d_table(pgd_t *pgd, unsigned long addr, unsigned long end,
+			    bool add, bool direct, struct vmem_altmap *altmap)
+{
+	unsigned long next;
+	int ret = -ENOMEM;
+	p4d_t *p4d, entry;
+	pud_t *pud;
+
+	p4d = p4d_offset(pgd, addr);
+	for (; addr < end; addr = next, p4d++) {
+		next = p4d_addr_end(addr, end);
+		entry = p4dp_get(p4d);
+		if (!add) {
+			if (p4d_none(entry))
+				continue;
+		} else if (p4d_none(entry)) {
+			pud = vmem_crst_alloc(_REGION3_ENTRY_EMPTY);
+			if (!pud)
+				goto out;
+			p4d_populate(&init_mm, p4d, pud);
+		}
+		ret = modify_pud_table(p4d, addr, next, add, direct, altmap);
+		if (ret)
+			goto out;
+		if (!add)
+			try_free_pud_table(p4d, addr & P4D_MASK);
+	}
+	ret = 0;
+out:
+	return ret;
+}
+
+static void try_free_p4d_table(pgd_t *pgd, unsigned long start)
+{
+	p4d_t *p4d;
+	int i;
+
+	p4d = p4d_offset(pgd, start);
+	for (i = 0; i < PTRS_PER_P4D; i++, p4d++) {
+		if (!p4d_none(p4dp_get(p4d)))
+			return;
+	}
+	vmem_free_pages(pgd_deref(pgdp_get(pgd)), CRST_ALLOC_ORDER, NULL);
+	pgd_clear(pgd);
+}
+
+static int modify_pagetable(unsigned long start, unsigned long end, bool add,
+			    bool direct, struct vmem_altmap *altmap)
+{
+	unsigned long addr, next;
+	int ret = -ENOMEM;
+	pgd_t *pgd, entry;
+	p4d_t *p4d;
+
+	if (WARN_ON_ONCE(!PAGE_ALIGNED(start | end)))
+		return -EINVAL;
+	/* Don't mess with any tables not fully in 1:1 mapping, vmemmap & kasan area */
+#ifdef CONFIG_KASAN
+	if (WARN_ON_ONCE(!(start >= KASAN_SHADOW_START && end <= KASAN_SHADOW_END) &&
+			 end > __abs_lowcore))
+		return -EINVAL;
+#else
+	if (WARN_ON_ONCE(end > __abs_lowcore))
+		return -EINVAL;
+#endif
+	for (addr = start; addr < end; addr = next) {
+		next = pgd_addr_end(addr, end);
+		pgd = pgd_offset_k(addr);
+		entry = pgdp_get(pgd);
+
+		if (!add) {
+			if (pgd_none(entry))
+				continue;
+		} else if (pgd_none(entry)) {
+			p4d = vmem_crst_alloc(_REGION2_ENTRY_EMPTY);
+			if (!p4d)
+				goto out;
+			pgd_populate(&init_mm, pgd, p4d);
+		}
+		ret = modify_p4d_table(pgd, addr, next, add, direct, altmap);
+		if (ret)
+			goto out;
+		if (!add)
+			try_free_p4d_table(pgd, addr & PGDIR_MASK);
+	}
+	ret = 0;
+out:
+	if (!add)
+		flush_tlb_kernel_range(start, end);
+	return ret;
+}
+
+static int add_pagetable(unsigned long start, unsigned long end, bool direct,
+			 struct vmem_altmap *altmap)
+{
+	return modify_pagetable(start, end, true, direct, altmap);
+}
+
+static int remove_pagetable(unsigned long start, unsigned long end, bool direct,
+			    struct vmem_altmap *altmap)
+{
+	return modify_pagetable(start, end, false, direct, altmap);
+}
+
+/*
+ * Add a physical memory range to the 1:1 mapping.
+ */
+static int vmem_add_range(unsigned long start, unsigned long size)
+{
+	start = (unsigned long)__va(start);
+	return add_pagetable(start, start + size, true, NULL);
+}
+
+/*
+ * Remove a physical memory range from the 1:1 mapping.
+ */
+static void vmem_remove_range(unsigned long start, unsigned long size)
+{
+	start = (unsigned long)__va(start);
+	remove_pagetable(start, start + size, true, NULL);
+}
+
+/*
+ * Add a backed mem_map array to the virtual mem_map array.
+ */
+int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node,
+			       struct vmem_altmap *altmap)
+{
+	int ret;
+
+	mutex_lock(&vmem_mutex);
+	/* We don't care about the node, just use NUMA_NO_NODE on allocations */
+	ret = add_pagetable(start, end, false, altmap);
+	if (ret)
+		remove_pagetable(start, end, false, altmap);
+	mutex_unlock(&vmem_mutex);
+	return ret;
+}
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+
+void vmemmap_free(unsigned long start, unsigned long end,
+		  struct vmem_altmap *altmap)
+{
+	mutex_lock(&vmem_mutex);
+	remove_pagetable(start, end, false, altmap);
+	mutex_unlock(&vmem_mutex);
+}
+
+#endif
+
+void vmem_remove_mapping(unsigned long start, unsigned long size)
+{
+	mutex_lock(&vmem_mutex);
+	vmem_remove_range(start, size);
+	mutex_unlock(&vmem_mutex);
+}
+
+struct range arch_get_mappable_range(void)
+{
+	struct range mhp_range;
+
+	mhp_range.start = 0;
+	mhp_range.end = max_mappable - 1;
+	return mhp_range;
+}
+
+int vmem_add_mapping(unsigned long start, unsigned long size)
+{
+	struct range range = arch_get_mappable_range();
+	int ret;
+
+	if (start < range.start ||
+	    start + size > range.end + 1 ||
+	    start + size < start)
+		return -ERANGE;
+
+	mutex_lock(&vmem_mutex);
+	ret = vmem_add_range(start, size);
+	if (ret)
+		vmem_remove_range(start, size);
+	mutex_unlock(&vmem_mutex);
+	return ret;
+}
+
+/*
+ * Allocate new or return existing page-table entry, but do not map it
+ * to any physical address. If missing, allocate segment- and region-
+ * table entries along. Meeting a large segment- or region-table entry
+ * while traversing is an error, since the function is expected to be
+ * called against virtual regions reserved for 4KB mappings only.
+ */
+pte_t *vmem_get_alloc_pte(unsigned long addr, bool alloc)
+{
+	pte_t *ptep = NULL;
+	pud_t pud_entry;
+	pmd_t pmd_entry;
+	pgd_t *pgd;
+	p4d_t *p4d;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+
+	pgd = pgd_offset_k(addr);
+	if (pgd_none(pgdp_get(pgd))) {
+		if (!alloc)
+			goto out;
+		p4d = vmem_crst_alloc(_REGION2_ENTRY_EMPTY);
+		if (!p4d)
+			goto out;
+		pgd_populate(&init_mm, pgd, p4d);
+	}
+	p4d = p4d_offset(pgd, addr);
+	if (p4d_none(p4dp_get(p4d))) {
+		if (!alloc)
+			goto out;
+		pud = vmem_crst_alloc(_REGION3_ENTRY_EMPTY);
+		if (!pud)
+			goto out;
+		p4d_populate(&init_mm, p4d, pud);
+	}
+	pud = pud_offset(p4d, addr);
+	pud_entry = pudp_get(pud);
+	if (pud_none(pud_entry)) {
+		if (!alloc)
+			goto out;
+		pmd = vmem_crst_alloc(_SEGMENT_ENTRY_EMPTY);
+		if (!pmd)
+			goto out;
+		pud_populate(&init_mm, pud, pmd);
+	} else if (WARN_ON_ONCE(pud_leaf(pud_entry))) {
+		goto out;
+	}
+	pmd = pmd_offset(pud, addr);
+	pmd_entry = pmdp_get(pmd);
+	if (pmd_none(pmd_entry)) {
+		if (!alloc)
+			goto out;
+		pte = vmem_pte_alloc();
+		if (!pte)
+			goto out;
+		pmd_populate(&init_mm, pmd, pte);
+	} else if (WARN_ON_ONCE(pmd_leaf(pmd_entry))) {
+		goto out;
+	}
+	ptep = pte_offset_kernel(pmd, addr);
+out:
+	return ptep;
+}
+
+int __vmem_map_4k_page(unsigned long addr, unsigned long phys, pgprot_t prot, bool alloc)
+{
+	pte_t *ptep, pte;
+
+	if (!IS_ALIGNED(addr, PAGE_SIZE))
+		return -EINVAL;
+	ptep = vmem_get_alloc_pte(addr, alloc);
+	if (!ptep)
+		return -ENOMEM;
+	__ptep_ipte(addr, ptep, 0, 0, IPTE_GLOBAL);
+	pte = mk_pte_phys(phys, prot);
+	set_pte(ptep, pte);
+	return 0;
+}
+
+int vmem_map_4k_page(unsigned long addr, unsigned long phys, pgprot_t prot)
+{
+	int rc;
+
+	mutex_lock(&vmem_mutex);
+	rc = __vmem_map_4k_page(addr, phys, prot, true);
+	mutex_unlock(&vmem_mutex);
+	return rc;
+}
+
+void vmem_unmap_4k_page(unsigned long addr)
+{
+	pte_t *ptep;
+
+	mutex_lock(&vmem_mutex);
+	ptep = virt_to_kpte(addr);
+	__ptep_ipte(addr, ptep, 0, 0, IPTE_GLOBAL);
+	pte_clear(&init_mm, addr, ptep);
+	mutex_unlock(&vmem_mutex);
+}
+
+void __init vmem_map_init(void)
+{
+	__set_memory_rox(_stext, _etext);
+	__set_memory_ro(_etext, __end_rodata);
+	__set_memory_rox(__stext_amode31, __etext_amode31);
+	/*
+	 * If the BEAR-enhancement facility is not installed the first
+	 * prefix page is used to return to the previous context with
+	 * an LPSWE instruction and therefore must be executable.
+	 */
+	if (!cpu_has_bear())
+		set_memory_x(0, 1);
+	if (debug_pagealloc_enabled())
+		__set_memory_4k(__va(0), absolute_pointer(__va(0)) + ident_map_size);
+	pr_info("Write protected kernel read-only data: %luk\n",
+		(unsigned long)(__end_rodata - _stext) >> 10);
+}