[PATCH v1 0/8] arm64: MMU enabled kexec relocation

[PATCH v1 0/8] arm64: MMU enabled kexec relocation

[Pavel Tatashin]

Enable MMU during kexec relocation in order to improve reboot performance.

If kexec functionality is used for a fast system update, with a minimal
downtime, the relocation of kernel + initramfs takes a significant portion
of reboot.

The reason for slow relocation is because it is done without MMU, and thus
not benefiting from D-Cache.

Performance data
----------------
For this experiment, the size of kernel plus initramfs is small, only 25M.
If initramfs was larger, than the improvements would be greater, as time
spent in relocation is proportional to the size of relocation.

Previously:
kernel shutdown	0.022131328s
relocation	0.440510736s
kernel startup	0.294706768s

Relocation was taking: 58.2% of reboot time

Now:
kernel shutdown	0.032066576s
relocation	0.022158152s
kernel startup	0.296055880s

Now: Relocation takes 6.3% of reboot time

Total reboot is x2.16 times faster.

Previous approaches and discussions
-----------------------------------
https://lore.kernel.org/lkml/20190709182014.16052-1-pasha.tatashin@soleen.com
reserve space for kexec to avoid relocation, involves changes to generic code
to optimize a problem that exists on arm64 only:

https://lore.kernel.org/lkml/20190716165641.6990-1-pasha.tatashin@soleen.com
The first attempt to enable MMU, some bugs that prevented performance
improvement. The page tables unnecessary configured idmap for the whole
physical space.

https://lore.kernel.org/lkml/20190731153857.4045-1-pasha.tatashin@soleen.com
No linear copy, bug with EL2 reboots.

Pavel Tatashin (8):
  kexec: quiet down kexec reboot
  arm64, mm: transitional tables
  arm64: hibernate: switch to transtional page tables.
  kexec: add machine_kexec_post_load()
  arm64, kexec: move relocation function setup and clean up
  arm64, kexec: add expandable argument to relocation function
  arm64, kexec: configure transitional page table for kexec
  arm64, kexec: enable MMU during kexec relocation

 arch/arm64/Kconfig                     |   4 +
 arch/arm64/include/asm/kexec.h         |  51 ++++-
 arch/arm64/include/asm/pgtable-hwdef.h |   1 +
 arch/arm64/include/asm/trans_table.h   |  68 ++++++
 arch/arm64/kernel/asm-offsets.c        |  14 ++
 arch/arm64/kernel/cpu-reset.S          |   4 +-
 arch/arm64/kernel/cpu-reset.h          |   8 +-
 arch/arm64/kernel/hibernate.c          | 261 ++++++-----------------
 arch/arm64/kernel/machine_kexec.c      | 199 ++++++++++++++----
 arch/arm64/kernel/relocate_kernel.S    | 196 +++++++++---------
 arch/arm64/mm/Makefile                 |   1 +
 arch/arm64/mm/trans_table.c            | 273 +++++++++++++++++++++++++
 kernel/kexec.c                         |   4 +
 kernel/kexec_core.c                    |   8 +-
 kernel/kexec_file.c                    |   4 +
 kernel/kexec_internal.h                |   2 +
 16 files changed, 758 insertions(+), 340 deletions(-)
 create mode 100644 arch/arm64/include/asm/trans_table.h
 create mode 100644 arch/arm64/mm/trans_table.c

-- 
2.22.0

[PATCH v1 1/8] kexec: quiet down kexec reboot

[Pavel Tatashin]

Here is a regular kexec command sequence and output:
=====
$ kexec --reuse-cmdline -i --load Image
$ kexec -e
[  161.342002] kexec_core: Starting new kernel

Welcome to Buildroot
buildroot login:
=====

Even when "quiet" kernel parameter is specified, "kexec_core: Starting
new kernel" is printed.

This message has  KERN_EMERG level, but there is no emergency, it is a
normal kexec operation, so quiet it down to appropriate KERN_NOTICE.

Machines that have slow console baud rate benefit from less output.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Reviewed-by: Simon Horman <horms@verge.net.au>
---
 kernel/kexec_core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
index d5870723b8ad..2c5b72863b7b 100644
--- a/kernel/kexec_core.c
+++ b/kernel/kexec_core.c
@@ -1169,7 +1169,7 @@ int kernel_kexec(void)
 		 * CPU hotplug again; so re-enable it here.
 		 */
 		cpu_hotplug_enable();
-		pr_emerg("Starting new kernel\n");
+		pr_notice("Starting new kernel\n");
 		machine_shutdown();
 	}
 
-- 
2.22.0

[PATCH v1 2/8] arm64, mm: transitional tables

[Pavel Tatashin]

There are cases where normal kernel pages tables, i.e. idmap_pg_dir
and swapper_pg_dir are not sufficient because they may be overwritten.

This happens when we transition from one world to another: for example
during kexec kernel relocation transition, and also during hibernate
kernel restore transition.

In these cases, if MMU is needed, the page table memory must be allocated
from a safe place. Transitional tables is intended to allow just that.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
---
 arch/arm64/Kconfig                     |   4 +
 arch/arm64/include/asm/pgtable-hwdef.h |   1 +
 arch/arm64/include/asm/trans_table.h   |  68 ++++++
 arch/arm64/mm/Makefile                 |   1 +
 arch/arm64/mm/trans_table.c            | 273 +++++++++++++++++++++++++
 5 files changed, 347 insertions(+)
 create mode 100644 arch/arm64/include/asm/trans_table.h
 create mode 100644 arch/arm64/mm/trans_table.c

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 3adcec05b1f6..91a7416ffe4e 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -999,6 +999,10 @@ config CRASH_DUMP
 
 	  For more details see Documentation/admin-guide/kdump/kdump.rst
 
+config TRANS_TABLE
+	def_bool y
+	depends on HIBERNATION || KEXEC_CORE
+
 config XEN_DOM0
 	def_bool y
 	depends on XEN
diff --git a/arch/arm64/include/asm/pgtable-hwdef.h b/arch/arm64/include/asm/pgtable-hwdef.h
index db92950bb1a0..dcb4f13c7888 100644
--- a/arch/arm64/include/asm/pgtable-hwdef.h
+++ b/arch/arm64/include/asm/pgtable-hwdef.h
@@ -110,6 +110,7 @@
 #define PUD_TABLE_BIT		(_AT(pudval_t, 1) << 1)
 #define PUD_TYPE_MASK		(_AT(pudval_t, 3) << 0)
 #define PUD_TYPE_SECT		(_AT(pudval_t, 1) << 0)
+#define PUD_SECT_RDONLY		(_AT(pudval_t, 1) << 7)		/* AP[2] */
 
 /*
  * Level 2 descriptor (PMD).
diff --git a/arch/arm64/include/asm/trans_table.h b/arch/arm64/include/asm/trans_table.h
new file mode 100644
index 000000000000..c7aef70587a1
--- /dev/null
+++ b/arch/arm64/include/asm/trans_table.h
@@ -0,0 +1,68 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+/*
+ * Copyright (c) 2019, Microsoft Corporation.
+ * Pavel Tatashin <patatash@linux.microsoft.com>
+ */
+
+#ifndef _ASM_TRANS_TABLE_H
+#define _ASM_TRANS_TABLE_H
+
+#include <linux/bits.h>
+#include <asm/pgtable-types.h>
+
+/*
+ * trans_alloc_page
+ *	- Allocator that should return exactly one uninitilaized page, if this
+ *	 allocator fails, trans_table returns -ENOMEM error.
+ *
+ * trans_alloc_arg
+ *	- Passed to trans_alloc_page as an argument
+ *
+ * trans_flags
+ *	- bitmap with flags that control how page table is filled.
+ *	  TRANS_MKWRITE: during page table copy make PTE, PME, and PUD page
+ *			 writeable by removing RDONLY flag from PTE.
+ *	  TRANS_MKVALID: during page table copy, if PTE present, but not valid,
+ *			 make it valid.
+ *	  TRANS_CHECKPFN: During page table copy, for every PTE entry check that
+ *			  PFN that this PTE points to is valid. Otherwise return
+ *			  -ENXIO
+ *	  TRANS_FORCEMAP: During page map, if translation exists, force
+ *			  overwrite it. Otherwise -ENXIO may be returned by
+ *			  trans_table_map_* functions if conflict is detected.
+ */
+
+#define	TRANS_MKWRITE	BIT(0)
+#define	TRANS_MKVALID	BIT(1)
+#define	TRANS_CHECKPFN	BIT(2)
+#define	TRANS_FORCEMAP	BIT(3)
+
+struct trans_table_info {
+	void * (*trans_alloc_page)(void *arg);
+	void *trans_alloc_arg;
+	unsigned long trans_flags;
+};
+
+/* Create and empty trans table. */
+int trans_table_create_empty(struct trans_table_info *info,
+			     pgd_t **trans_table);
+
+/*
+ * Create trans table and copy entries from from_table to trans_table in range
+ * [start, end)
+ */
+int trans_table_create_copy(struct trans_table_info *info, pgd_t **trans_table,
+			    pgd_t *from_table, unsigned long start,
+			    unsigned long end);
+
+/*
+ * Add map entry to trans_table for a base-size page at PTE level.
+ * page:	page to be mapped.
+ * dst_addr:	new VA address for the pages
+ * pgprot:	protection for the page.
+ */
+int trans_table_map_page(struct trans_table_info *info, pgd_t *trans_table,
+			 void *page, unsigned long dst_addr, pgprot_t pgprot);
+
+#endif /* _ASM_TRANS_TABLE_H */
diff --git a/arch/arm64/mm/Makefile b/arch/arm64/mm/Makefile
index 849c1df3d214..3794fff18659 100644
--- a/arch/arm64/mm/Makefile
+++ b/arch/arm64/mm/Makefile
@@ -6,6 +6,7 @@ obj-y				:= dma-mapping.o extable.o fault.o init.o \
 obj-$(CONFIG_HUGETLB_PAGE)	+= hugetlbpage.o
 obj-$(CONFIG_ARM64_PTDUMP_CORE)	+= dump.o
 obj-$(CONFIG_ARM64_PTDUMP_DEBUGFS)	+= ptdump_debugfs.o
+obj-$(CONFIG_TRANS_TABLE)	+= trans_table.o
 obj-$(CONFIG_NUMA)		+= numa.o
 obj-$(CONFIG_DEBUG_VIRTUAL)	+= physaddr.o
 KASAN_SANITIZE_physaddr.o	+= n
diff --git a/arch/arm64/mm/trans_table.c b/arch/arm64/mm/trans_table.c
new file mode 100644
index 000000000000..e3b8d4a2fa15
--- /dev/null
+++ b/arch/arm64/mm/trans_table.c
@@ -0,0 +1,273 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Copyright (c) 2019, Microsoft Corporation.
+ * Pavel Tatashin <patatash@linux.microsoft.com>
+ */
+
+/*
+ * Transitional tables are used during system transferring from one world to
+ * another: such as during hibernate restore, and kexec reboots. During these
+ * phases one cannot rely on page table not being overwritten.
+ *
+ */
+
+#include <asm/trans_table.h>
+#include <asm/pgalloc.h>
+#include <asm/pgtable.h>
+
+static void *trans_alloc(struct trans_table_info *info)
+{
+	void *page = info->trans_alloc_page(info->trans_alloc_arg);
+
+	if (page)
+		clear_page(page);
+
+	return page;
+}
+
+static int trans_table_copy_pte(struct trans_table_info *info, pte_t *dst_ptep,
+				pte_t *src_ptep, unsigned long start,
+				unsigned long end)
+{
+	unsigned long addr = start;
+	int i = pgd_index(addr);
+
+	do {
+		pte_t src_pte = READ_ONCE(src_ptep[i]);
+
+		if (pte_none(src_pte))
+			continue;
+		if (info->trans_flags & TRANS_MKWRITE)
+			src_pte = pte_mkwrite(src_pte);
+		if (info->trans_flags & TRANS_MKVALID)
+			src_pte = pte_mkpresent(src_pte);
+		if (info->trans_flags & TRANS_CHECKPFN) {
+			if (!pfn_valid(pte_pfn(src_pte)))
+				return -ENXIO;
+		}
+		set_pte(&dst_ptep[i], src_pte);
+	} while (addr += PAGE_SIZE, i++, addr != end && i < PTRS_PER_PTE);
+
+	return 0;
+}
+
+static int trans_table_copy_pmd(struct trans_table_info *info, pmd_t *dst_pmdp,
+				pmd_t *src_pmdp, unsigned long start,
+				unsigned long end)
+{
+	unsigned long next;
+	unsigned long addr = start;
+	int i = pgd_index(addr);
+	int rc;
+
+	do {
+		pmd_t src_pmd = READ_ONCE(src_pmdp[i]);
+		pmd_t dst_pmd = READ_ONCE(dst_pmdp[i]);
+		pte_t *dst_ptep, *src_ptep;
+
+		next = pmd_addr_end(addr, end);
+		if (pmd_none(src_pmd))
+			continue;
+
+		if (!pmd_table(src_pmd)) {
+			if (info->trans_flags & TRANS_MKWRITE)
+				pmd_val(src_pmd) &= ~PMD_SECT_RDONLY;
+			set_pmd(&dst_pmdp[i], src_pmd);
+			continue;
+		}
+
+		if (pmd_none(dst_pmd)) {
+			pte_t *t = trans_alloc(info);
+
+			if (!t)
+				return -ENOMEM;
+
+			__pmd_populate(&dst_pmdp[i], __pa(t), PTE_TYPE_PAGE);
+			dst_pmd = READ_ONCE(dst_pmdp[i]);
+		}
+
+		src_ptep = __va(pmd_page_paddr(src_pmd));
+		dst_ptep = __va(pmd_page_paddr(dst_pmd));
+
+		rc = trans_table_copy_pte(info, dst_ptep, src_ptep, addr, next);
+		if (rc)
+			return rc;
+	} while (addr = next, i++, addr != end && i < PTRS_PER_PMD);
+
+	return 0;
+}
+
+static int trans_table_copy_pud(struct trans_table_info *info, pud_t *dst_pudp,
+				pud_t *src_pudp, unsigned long start,
+				unsigned long end)
+{
+	unsigned long next;
+	unsigned long addr = start;
+	int i = pgd_index(addr);
+	int rc;
+
+	do {
+		pud_t src_pud = READ_ONCE(src_pudp[i]);
+		pud_t dst_pud = READ_ONCE(dst_pudp[i]);
+		pmd_t *dst_pmdp, *src_pmdp;
+
+		next = pud_addr_end(addr, end);
+		if (pud_none(src_pud))
+			continue;
+
+		if (!pud_table(src_pud)) {
+			if (info->trans_flags & TRANS_MKWRITE)
+				pud_val(src_pud) &= ~PUD_SECT_RDONLY;
+			set_pud(&dst_pudp[i], src_pud);
+			continue;
+		}
+
+		if (pud_none(dst_pud)) {
+			pmd_t *t = trans_alloc(info);
+
+			if (!t)
+				return -ENOMEM;
+
+			__pud_populate(&dst_pudp[i], __pa(t), PMD_TYPE_TABLE);
+			dst_pud = READ_ONCE(dst_pudp[i]);
+		}
+
+		src_pmdp = __va(pud_page_paddr(src_pud));
+		dst_pmdp = __va(pud_page_paddr(dst_pud));
+
+		rc = trans_table_copy_pmd(info, dst_pmdp, src_pmdp, addr, next);
+		if (rc)
+			return rc;
+	} while (addr = next, i++, addr != end && i < PTRS_PER_PUD);
+
+	return 0;
+}
+
+static int trans_table_copy_pgd(struct trans_table_info *info, pgd_t *dst_pgdp,
+				pgd_t *src_pgdp, unsigned long start,
+				unsigned long end)
+{
+	unsigned long next;
+	unsigned long addr = start;
+	int i = pgd_index(addr);
+	int rc;
+
+	do {
+		pgd_t src_pgd;
+		pgd_t dst_pgd;
+		pud_t *dst_pudp, *src_pudp;
+
+		src_pgd = READ_ONCE(src_pgdp[i]);
+		dst_pgd = READ_ONCE(dst_pgdp[i]);
+		next = pgd_addr_end(addr, end);
+		if (pgd_none(src_pgd))
+			continue;
+
+		if (pgd_none(dst_pgd)) {
+			pud_t *t = trans_alloc(info);
+
+			if (!t)
+				return -ENOMEM;
+
+			__pgd_populate(&dst_pgdp[i], __pa(t), PUD_TYPE_TABLE);
+			dst_pgd = READ_ONCE(dst_pgdp[i]);
+		}
+
+		src_pudp = __va(pgd_page_paddr(src_pgd));
+		dst_pudp = __va(pgd_page_paddr(dst_pgd));
+
+		rc = trans_table_copy_pud(info, dst_pudp, src_pudp, addr, next);
+		if (rc)
+			return rc;
+	} while (addr = next, i++, addr != end && i < PTRS_PER_PGD);
+
+	return 0;
+}
+
+int trans_table_create_empty(struct trans_table_info *info, pgd_t **trans_table)
+{
+	pgd_t *dst_pgdp = trans_alloc(info);
+
+	if (!dst_pgdp)
+		return -ENOMEM;
+
+	*trans_table = dst_pgdp;
+
+	return 0;
+}
+
+int trans_table_create_copy(struct trans_table_info *info, pgd_t **trans_table,
+			    pgd_t *from_table, unsigned long start,
+			    unsigned long end)
+{
+	int rc;
+
+	rc = trans_table_create_empty(info, trans_table);
+	if (rc)
+		return rc;
+
+	return trans_table_copy_pgd(info, *trans_table, from_table, start, end);
+}
+
+int trans_table_map_page(struct trans_table_info *info, pgd_t *trans_table,
+			 void *page, unsigned long dst_addr, pgprot_t pgprot)
+{
+	int pgd_idx = pgd_index(dst_addr);
+	int pud_idx = pud_index(dst_addr);
+	int pmd_idx = pmd_index(dst_addr);
+	int pte_idx = pte_index(dst_addr);
+	pgd_t *pgdp = trans_table;
+	pgd_t pgd = READ_ONCE(pgdp[pgd_idx]);
+	pud_t *pudp, pud;
+	pmd_t *pmdp, pmd;
+	pte_t *ptep, pte;
+
+	if (pgd_none(pgd)) {
+		pud_t *t = trans_alloc(info);
+
+		if (!t)
+			return -ENOMEM;
+
+		__pgd_populate(&pgdp[pgd_idx], __pa(t), PUD_TYPE_TABLE);
+		pgd = READ_ONCE(pgdp[pgd_idx]);
+	}
+
+	pudp = __va(pgd_page_paddr(pgd));
+	pud = READ_ONCE(pudp[pud_idx]);
+	if (pud_sect(pud) && !(info->trans_flags & TRANS_FORCEMAP)) {
+		return -ENXIO;
+	} else if (pud_none(pud) || pud_sect(pud)) {
+		pmd_t *t = trans_alloc(info);
+
+		if (!t)
+			return -ENOMEM;
+
+		__pud_populate(&pudp[pud_idx], __pa(t), PMD_TYPE_TABLE);
+		pud = READ_ONCE(pudp[pud_idx]);
+	}
+
+	pmdp = __va(pud_page_paddr(pud));
+	pmd = READ_ONCE(pmdp[pmd_idx]);
+	if (pmd_sect(pmd) && !(info->trans_flags & TRANS_FORCEMAP)) {
+		return -ENXIO;
+	} else if (pmd_none(pmd) || pmd_sect(pmd)) {
+		pte_t *t = trans_alloc(info);
+
+		if (!t)
+			return -ENOMEM;
+
+		__pmd_populate(&pmdp[pmd_idx], __pa(t), PTE_TYPE_PAGE);
+		pmd = READ_ONCE(pmdp[pmd_idx]);
+	}
+
+	ptep = __va(pmd_page_paddr(pmd));
+	pte = READ_ONCE(ptep[pte_idx]);
+
+	if (!pte_none(pte) && !(info->trans_flags & TRANS_FORCEMAP))
+		return -ENXIO;
+
+	set_pte(&ptep[pte_idx], pfn_pte(virt_to_pfn(page), pgprot));
+
+	return 0;
+}
-- 
2.22.0

[PATCH v1 3/8] arm64: hibernate: switch to transtional page tables.

[Pavel Tatashin]

Transitional page tables  provide the needed functionality to setup
temporary page tables needed for hibernate resume.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
---
 arch/arm64/kernel/hibernate.c | 261 ++++++++--------------------------
 1 file changed, 60 insertions(+), 201 deletions(-)

diff --git a/arch/arm64/kernel/hibernate.c b/arch/arm64/kernel/hibernate.c
index 9341fcc6e809..4120b03a02fd 100644
--- a/arch/arm64/kernel/hibernate.c
+++ b/arch/arm64/kernel/hibernate.c
@@ -16,7 +16,6 @@
 #define pr_fmt(x) "hibernate: " x
 #include <linux/cpu.h>
 #include <linux/kvm_host.h>
-#include <linux/mm.h>
 #include <linux/pm.h>
 #include <linux/sched.h>
 #include <linux/suspend.h>
@@ -31,14 +30,12 @@
 #include <asm/kexec.h>
 #include <asm/memory.h>
 #include <asm/mmu_context.h>
-#include <asm/pgalloc.h>
-#include <asm/pgtable.h>
-#include <asm/pgtable-hwdef.h>
 #include <asm/sections.h>
 #include <asm/smp.h>
 #include <asm/smp_plat.h>
 #include <asm/suspend.h>
 #include <asm/sysreg.h>
+#include <asm/trans_table.h>
 #include <asm/virt.h>
 
 /*
@@ -182,6 +179,12 @@ int arch_hibernation_header_restore(void *addr)
 }
 EXPORT_SYMBOL(arch_hibernation_header_restore);
 
+static void *
+hibernate_page_alloc(void *arg)
+{
+	return (void *)get_safe_page((gfp_t)(unsigned long)arg);
+}
+
 /*
  * Copies length bytes, starting at src_start into an new page,
  * perform cache maintentance, then maps it at the specified address low
@@ -196,57 +199,31 @@ EXPORT_SYMBOL(arch_hibernation_header_restore);
  */
 static int create_safe_exec_page(void *src_start, size_t length,
 				 unsigned long dst_addr,
-				 phys_addr_t *phys_dst_addr,
-				 void *(*allocator)(gfp_t mask),
-				 gfp_t mask)
+				 phys_addr_t *phys_dst_addr)
 {
-	int rc = 0;
-	pgd_t *pgdp;
-	pud_t *pudp;
-	pmd_t *pmdp;
-	pte_t *ptep;
-	unsigned long dst = (unsigned long)allocator(mask);
-
-	if (!dst) {
-		rc = -ENOMEM;
-		goto out;
-	}
-
-	memcpy((void *)dst, src_start, length);
-	__flush_icache_range(dst, dst + length);
+	struct trans_table_info trans_info = {
+		.trans_alloc_page	= hibernate_page_alloc,
+		.trans_alloc_arg	= (void *)GFP_ATOMIC,
+		.trans_flags		= 0,
+	};
+	void *page = (void *)get_safe_page(GFP_ATOMIC);
+	pgd_t *trans_table;
+	int rc;
+
+	if (!page)
+		return -ENOMEM;
 
-	pgdp = pgd_offset_raw(allocator(mask), dst_addr);
-	if (pgd_none(READ_ONCE(*pgdp))) {
-		pudp = allocator(mask);
-		if (!pudp) {
-			rc = -ENOMEM;
-			goto out;
-		}
-		pgd_populate(&init_mm, pgdp, pudp);
-	}
+	memcpy(page, src_start, length);
+	__flush_icache_range((unsigned long)page, (unsigned long)page + length);
 
-	pudp = pud_offset(pgdp, dst_addr);
-	if (pud_none(READ_ONCE(*pudp))) {
-		pmdp = allocator(mask);
-		if (!pmdp) {
-			rc = -ENOMEM;
-			goto out;
-		}
-		pud_populate(&init_mm, pudp, pmdp);
-	}
-
-	pmdp = pmd_offset(pudp, dst_addr);
-	if (pmd_none(READ_ONCE(*pmdp))) {
-		ptep = allocator(mask);
-		if (!ptep) {
-			rc = -ENOMEM;
-			goto out;
-		}
-		pmd_populate_kernel(&init_mm, pmdp, ptep);
-	}
+	rc = trans_table_create_empty(&trans_info, &trans_table);
+	if (rc)
+		return rc;
 
-	ptep = pte_offset_kernel(pmdp, dst_addr);
-	set_pte(ptep, pfn_pte(virt_to_pfn(dst), PAGE_KERNEL_EXEC));
+	rc = trans_table_map_page(&trans_info, trans_table, page, dst_addr,
+				  PAGE_KERNEL_EXEC);
+	if (rc)
+		return rc;
 
 	/*
 	 * Load our new page tables. A strict BBM approach requires that we
@@ -262,13 +239,12 @@ static int create_safe_exec_page(void *src_start, size_t length,
 	 */
 	cpu_set_reserved_ttbr0();
 	local_flush_tlb_all();
-	write_sysreg(phys_to_ttbr(virt_to_phys(pgdp)), ttbr0_el1);
+	write_sysreg(phys_to_ttbr(virt_to_phys(trans_table)), ttbr0_el1);
 	isb();
 
-	*phys_dst_addr = virt_to_phys((void *)dst);
+	*phys_dst_addr = virt_to_phys(page);
 
-out:
-	return rc;
+	return 0;
 }
 
 #define dcache_clean_range(start, end)	__flush_dcache_area(start, (end - start))
@@ -332,143 +308,6 @@ int swsusp_arch_suspend(void)
 	return ret;
 }
 
-static void _copy_pte(pte_t *dst_ptep, pte_t *src_ptep, unsigned long addr)
-{
-	pte_t pte = READ_ONCE(*src_ptep);
-
-	if (pte_valid(pte)) {
-		/*
-		 * Resume will overwrite areas that may be marked
-		 * read only (code, rodata). Clear the RDONLY bit from
-		 * the temporary mappings we use during restore.
-		 */
-		set_pte(dst_ptep, pte_mkwrite(pte));
-	} else if (debug_pagealloc_enabled() && !pte_none(pte)) {
-		/*
-		 * debug_pagealloc will removed the PTE_VALID bit if
-		 * the page isn't in use by the resume kernel. It may have
-		 * been in use by the original kernel, in which case we need
-		 * to put it back in our copy to do the restore.
-		 *
-		 * Before marking this entry valid, check the pfn should
-		 * be mapped.
-		 */
-		BUG_ON(!pfn_valid(pte_pfn(pte)));
-
-		set_pte(dst_ptep, pte_mkpresent(pte_mkwrite(pte)));
-	}
-}
-
-static int copy_pte(pmd_t *dst_pmdp, pmd_t *src_pmdp, unsigned long start,
-		    unsigned long end)
-{
-	pte_t *src_ptep;
-	pte_t *dst_ptep;
-	unsigned long addr = start;
-
-	dst_ptep = (pte_t *)get_safe_page(GFP_ATOMIC);
-	if (!dst_ptep)
-		return -ENOMEM;
-	pmd_populate_kernel(&init_mm, dst_pmdp, dst_ptep);
-	dst_ptep = pte_offset_kernel(dst_pmdp, start);
-
-	src_ptep = pte_offset_kernel(src_pmdp, start);
-	do {
-		_copy_pte(dst_ptep, src_ptep, addr);
-	} while (dst_ptep++, src_ptep++, addr += PAGE_SIZE, addr != end);
-
-	return 0;
-}
-
-static int copy_pmd(pud_t *dst_pudp, pud_t *src_pudp, unsigned long start,
-		    unsigned long end)
-{
-	pmd_t *src_pmdp;
-	pmd_t *dst_pmdp;
-	unsigned long next;
-	unsigned long addr = start;
-
-	if (pud_none(READ_ONCE(*dst_pudp))) {
-		dst_pmdp = (pmd_t *)get_safe_page(GFP_ATOMIC);
-		if (!dst_pmdp)
-			return -ENOMEM;
-		pud_populate(&init_mm, dst_pudp, dst_pmdp);
-	}
-	dst_pmdp = pmd_offset(dst_pudp, start);
-
-	src_pmdp = pmd_offset(src_pudp, start);
-	do {
-		pmd_t pmd = READ_ONCE(*src_pmdp);
-
-		next = pmd_addr_end(addr, end);
-		if (pmd_none(pmd))
-			continue;
-		if (pmd_table(pmd)) {
-			if (copy_pte(dst_pmdp, src_pmdp, addr, next))
-				return -ENOMEM;
-		} else {
-			set_pmd(dst_pmdp,
-				__pmd(pmd_val(pmd) & ~PMD_SECT_RDONLY));
-		}
-	} while (dst_pmdp++, src_pmdp++, addr = next, addr != end);
-
-	return 0;
-}
-
-static int copy_pud(pgd_t *dst_pgdp, pgd_t *src_pgdp, unsigned long start,
-		    unsigned long end)
-{
-	pud_t *dst_pudp;
-	pud_t *src_pudp;
-	unsigned long next;
-	unsigned long addr = start;
-
-	if (pgd_none(READ_ONCE(*dst_pgdp))) {
-		dst_pudp = (pud_t *)get_safe_page(GFP_ATOMIC);
-		if (!dst_pudp)
-			return -ENOMEM;
-		pgd_populate(&init_mm, dst_pgdp, dst_pudp);
-	}
-	dst_pudp = pud_offset(dst_pgdp, start);
-
-	src_pudp = pud_offset(src_pgdp, start);
-	do {
-		pud_t pud = READ_ONCE(*src_pudp);
-
-		next = pud_addr_end(addr, end);
-		if (pud_none(pud))
-			continue;
-		if (pud_table(pud)) {
-			if (copy_pmd(dst_pudp, src_pudp, addr, next))
-				return -ENOMEM;
-		} else {
-			set_pud(dst_pudp,
-				__pud(pud_val(pud) & ~PMD_SECT_RDONLY));
-		}
-	} while (dst_pudp++, src_pudp++, addr = next, addr != end);
-
-	return 0;
-}
-
-static int copy_page_tables(pgd_t *dst_pgdp, unsigned long start,
-			    unsigned long end)
-{
-	unsigned long next;
-	unsigned long addr = start;
-	pgd_t *src_pgdp = pgd_offset_k(start);
-
-	dst_pgdp = pgd_offset_raw(dst_pgdp, start);
-	do {
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(READ_ONCE(*src_pgdp)))
-			continue;
-		if (copy_pud(dst_pgdp, src_pgdp, addr, next))
-			return -ENOMEM;
-	} while (dst_pgdp++, src_pgdp++, addr = next, addr != end);
-
-	return 0;
-}
-
 /*
  * Setup then Resume from the hibernate image using swsusp_arch_suspend_exit().
  *
@@ -484,21 +323,42 @@ int swsusp_arch_resume(void)
 	phys_addr_t phys_hibernate_exit;
 	void __noreturn (*hibernate_exit)(phys_addr_t, phys_addr_t, void *,
 					  void *, phys_addr_t, phys_addr_t);
+	struct trans_table_info trans_info = {
+		.trans_alloc_page	= hibernate_page_alloc,
+		.trans_alloc_arg	= (void *)GFP_ATOMIC,
+		/*
+		 * Resume will overwrite areas that may be marked read only
+		 * (code, rodata). Clear the RDONLY bit from the temporary
+		 * mappings we use during restore.
+		 */
+		.trans_flags		= TRANS_MKWRITE,
+	};
+
+	/*
+	 * debug_pagealloc will removed the PTE_VALID bit if the page isn't in
+	 * use by the resume kernel. It may have been in use by the original
+	 * kernel, in which case we need to put it back in our copy to do the
+	 * restore.
+	 *
+	 * Before marking this entry valid, check the pfn should be mapped.
+	 */
+	if (debug_pagealloc_enabled())
+		trans_info.trans_flags |= (TRANS_MKVALID | TRANS_CHECKPFN);
 
 	/*
 	 * Restoring the memory image will overwrite the ttbr1 page tables.
 	 * Create a second copy of just the linear map, and use this when
 	 * restoring.
 	 */
-	tmp_pg_dir = (pgd_t *)get_safe_page(GFP_ATOMIC);
-	if (!tmp_pg_dir) {
-		pr_err("Failed to allocate memory for temporary page tables.\n");
-		rc = -ENOMEM;
+	rc = trans_table_create_copy(&trans_info, &tmp_pg_dir,
+				     pgd_offset_k(PAGE_OFFSET), PAGE_OFFSET, 0);
+	if (rc) {
+		if (rc == -ENOMEM)
+			pr_err("Failed to allocate memory for temporary page tables.\n");
+		else if (rc == -ENXIO)
+			pr_err("Tried to set PTE for PFN that does not exist\n");
 		goto out;
 	}
-	rc = copy_page_tables(tmp_pg_dir, PAGE_OFFSET, 0);
-	if (rc)
-		goto out;
 
 	/*
 	 * We need a zero page that is zero before & after resume in order to
@@ -523,8 +383,7 @@ int swsusp_arch_resume(void)
 	 */
 	rc = create_safe_exec_page(__hibernate_exit_text_start, exit_size,
 				   (unsigned long)hibernate_exit,
-				   &phys_hibernate_exit,
-				   (void *)get_safe_page, GFP_ATOMIC);
+				   &phys_hibernate_exit);
 	if (rc) {
 		pr_err("Failed to create safe executable page for hibernate_exit code.\n");
 		goto out;
-- 
2.22.0

[PATCH v1 4/8] kexec: add machine_kexec_post_load()

[Pavel Tatashin]

It is the same as machine_kexec_prepare(), but is called after segments are
loaded. This way, can do processing work with already loaded relocation
segments. One such example is arm64: it has to have segments loaded in
order to create a page table, but it cannot do it during kexec time,
because at that time allocations won't be possible anymore.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
---
 kernel/kexec.c          | 4 ++++
 kernel/kexec_core.c     | 6 ++++++
 kernel/kexec_file.c     | 4 ++++
 kernel/kexec_internal.h | 2 ++
 4 files changed, 16 insertions(+)

diff --git a/kernel/kexec.c b/kernel/kexec.c
index 1b018f1a6e0d..27b71dc7b35a 100644
--- a/kernel/kexec.c
+++ b/kernel/kexec.c
@@ -159,6 +159,10 @@ static int do_kexec_load(unsigned long entry, unsigned long nr_segments,
 
 	kimage_terminate(image);
 
+	ret = machine_kexec_post_load(image);
+	if (ret)
+		goto out;
+
 	/* Install the new kernel and uninstall the old */
 	image = xchg(dest_image, image);
 
diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
index 2c5b72863b7b..8360645d1bbe 100644
--- a/kernel/kexec_core.c
+++ b/kernel/kexec_core.c
@@ -587,6 +587,12 @@ static void kimage_free_extra_pages(struct kimage *image)
 	kimage_free_page_list(&image->unusable_pages);
 
 }
+
+int __weak machine_kexec_post_load(struct kimage *image)
+{
+	return 0;
+}
+
 void kimage_terminate(struct kimage *image)
 {
 	if (*image->entry != 0)
diff --git a/kernel/kexec_file.c b/kernel/kexec_file.c
index b8cc032d5620..cb531d768114 100644
--- a/kernel/kexec_file.c
+++ b/kernel/kexec_file.c
@@ -391,6 +391,10 @@ SYSCALL_DEFINE5(kexec_file_load, int, kernel_fd, int, initrd_fd,
 
 	kimage_terminate(image);
 
+	ret = machine_kexec_post_load(image);
+	if (ret)
+		goto out;
+
 	/*
 	 * Free up any temporary buffers allocated which are not needed
 	 * after image has been loaded
diff --git a/kernel/kexec_internal.h b/kernel/kexec_internal.h
index 48aaf2ac0d0d..39d30ccf8d87 100644
--- a/kernel/kexec_internal.h
+++ b/kernel/kexec_internal.h
@@ -13,6 +13,8 @@ void kimage_terminate(struct kimage *image);
 int kimage_is_destination_range(struct kimage *image,
 				unsigned long start, unsigned long end);
 
+int machine_kexec_post_load(struct kimage *image);
+
 extern struct mutex kexec_mutex;
 
 #ifdef CONFIG_KEXEC_FILE
-- 
2.22.0

[PATCH v1 5/8] arm64, kexec: move relocation function setup and clean up

[Pavel Tatashin]

Currently, kernel relocation function is configured in machine_kexec()
at the time of kexec reboot by using control_code_page.

This operation, however, is more logical to be done during kexec_load,
and thus remove from reboot time. Move, setup of this function to
newly added machine_kexec_post_load().

In addition, do some cleanup: add infor about reloction function to
kexec_image_info(), and remove extra messages from machine_kexec().

Make dtb_mem, always available, if CONFIG_KEXEC_FILE is not configured
dtb_mem is set to zero anyway.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
---
 arch/arm64/include/asm/kexec.h    |  3 +-
 arch/arm64/kernel/machine_kexec.c | 49 +++++++++++--------------------
 2 files changed, 19 insertions(+), 33 deletions(-)

diff --git a/arch/arm64/include/asm/kexec.h b/arch/arm64/include/asm/kexec.h
index 12a561a54128..d15ca1ca1e83 100644
--- a/arch/arm64/include/asm/kexec.h
+++ b/arch/arm64/include/asm/kexec.h
@@ -90,14 +90,15 @@ static inline void crash_prepare_suspend(void) {}
 static inline void crash_post_resume(void) {}
 #endif
 
-#ifdef CONFIG_KEXEC_FILE
 #define ARCH_HAS_KIMAGE_ARCH
 
 struct kimage_arch {
 	void *dtb;
 	unsigned long dtb_mem;
+	unsigned long kern_reloc;
 };
 
+#ifdef CONFIG_KEXEC_FILE
 extern const struct kexec_file_ops kexec_image_ops;
 
 struct kimage;
diff --git a/arch/arm64/kernel/machine_kexec.c b/arch/arm64/kernel/machine_kexec.c
index 0df8493624e0..9b41da50e6f7 100644
--- a/arch/arm64/kernel/machine_kexec.c
+++ b/arch/arm64/kernel/machine_kexec.c
@@ -42,6 +42,7 @@ static void _kexec_image_info(const char *func, int line,
 	pr_debug("    start:       %lx\n", kimage->start);
 	pr_debug("    head:        %lx\n", kimage->head);
 	pr_debug("    nr_segments: %lu\n", kimage->nr_segments);
+	pr_debug("    kern_reloc: %pa\n", &kimage->arch.kern_reloc);
 
 	for (i = 0; i < kimage->nr_segments; i++) {
 		pr_debug("      segment[%lu]: %016lx - %016lx, 0x%lx bytes, %lu pages\n",
@@ -58,6 +59,19 @@ void machine_kexec_cleanup(struct kimage *kimage)
 	/* Empty routine needed to avoid build errors. */
 }
 
+int machine_kexec_post_load(struct kimage *kimage)
+{
+	unsigned long kern_reloc;
+
+	kern_reloc = page_to_phys(kimage->control_code_page);
+	memcpy(__va(kern_reloc), arm64_relocate_new_kernel,
+	       arm64_relocate_new_kernel_size);
+	kimage->arch.kern_reloc = kern_reloc;
+
+	kexec_image_info(kimage);
+	return 0;
+}
+
 /**
  * machine_kexec_prepare - Prepare for a kexec reboot.
  *
@@ -67,8 +81,6 @@ void machine_kexec_cleanup(struct kimage *kimage)
  */
 int machine_kexec_prepare(struct kimage *kimage)
 {
-	kexec_image_info(kimage);
-
 	if (kimage->type != KEXEC_TYPE_CRASH && cpus_are_stuck_in_kernel()) {
 		pr_err("Can't kexec: CPUs are stuck in the kernel.\n");
 		return -EBUSY;
@@ -143,8 +155,7 @@ static void kexec_segment_flush(const struct kimage *kimage)
  */
 void machine_kexec(struct kimage *kimage)
 {
-	phys_addr_t reboot_code_buffer_phys;
-	void *reboot_code_buffer;
+	void *reboot_code_buffer = phys_to_virt(kimage->arch.kern_reloc);
 	bool in_kexec_crash = (kimage == kexec_crash_image);
 	bool stuck_cpus = cpus_are_stuck_in_kernel();
 
@@ -155,30 +166,8 @@ void machine_kexec(struct kimage *kimage)
 	WARN(in_kexec_crash && (stuck_cpus || smp_crash_stop_failed()),
 		"Some CPUs may be stale, kdump will be unreliable.\n");
 
-	reboot_code_buffer_phys = page_to_phys(kimage->control_code_page);
-	reboot_code_buffer = phys_to_virt(reboot_code_buffer_phys);
-
 	kexec_image_info(kimage);
 
-	pr_debug("%s:%d: control_code_page:        %p\n", __func__, __LINE__,
-		kimage->control_code_page);
-	pr_debug("%s:%d: reboot_code_buffer_phys:  %pa\n", __func__, __LINE__,
-		&reboot_code_buffer_phys);
-	pr_debug("%s:%d: reboot_code_buffer:       %p\n", __func__, __LINE__,
-		reboot_code_buffer);
-	pr_debug("%s:%d: relocate_new_kernel:      %p\n", __func__, __LINE__,
-		arm64_relocate_new_kernel);
-	pr_debug("%s:%d: relocate_new_kernel_size: 0x%lx(%lu) bytes\n",
-		__func__, __LINE__, arm64_relocate_new_kernel_size,
-		arm64_relocate_new_kernel_size);
-
-	/*
-	 * Copy arm64_relocate_new_kernel to the reboot_code_buffer for use
-	 * after the kernel is shut down.
-	 */
-	memcpy(reboot_code_buffer, arm64_relocate_new_kernel,
-		arm64_relocate_new_kernel_size);
-
 	/* Flush the reboot_code_buffer in preparation for its execution. */
 	__flush_dcache_area(reboot_code_buffer, arm64_relocate_new_kernel_size);
 
@@ -214,12 +203,8 @@ void machine_kexec(struct kimage *kimage)
 	 * userspace (kexec-tools).
 	 * In kexec_file case, the kernel starts directly without purgatory.
 	 */
-	cpu_soft_restart(reboot_code_buffer_phys, kimage->head, kimage->start,
-#ifdef CONFIG_KEXEC_FILE
-						kimage->arch.dtb_mem);
-#else
-						0);
-#endif
+	cpu_soft_restart(kimage->arch.kern_reloc, kimage->head, kimage->start,
+			 kimage->arch.dtb_mem);
 
 	BUG(); /* Should never get here. */
 }
-- 
2.22.0

[PATCH v1 6/8] arm64, kexec: add expandable argument to relocation function

[Pavel Tatashin]

Currently, kexec relocation function (arm64_relocate_new_kernel) accepts
the following arguments:

head:		start of array that contains relocation information.
entry:		entry point for new kernel or purgatory.
dtb_mem:	first and only argument to entry.

The number of arguments cannot be easily expended, because this
function is also called from HVC_SOFT_RESTART, which preserves only
three arguments. And, also arm64_relocate_new_kernel is written in
assembly but called without stack, thus no place to move extra
arguments to free registers.

Soon, we will need to pass more arguments: once we enable MMU we
will need to pass information about page tables.

Another benefit of allowing this function to accept more arguments, is that
kernel can actually accept up to 4 arguments (x0-x3), however currently
only one is used, but if in the future we will need for more (for example,
pass information about when previous kernel exited to have a precise
measurement in time spent in purgatory), we won't be easilty do that
if arm64_relocate_new_kernel can't accept more arguments.

So, add a new struct: kern_reloc_arg, and place it in kexec safe page (i.e
memory that is not overwritten during relocation).
Thus, make arm64_relocate_new_kernel to only take one argument, that
contains all the needed information.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
---
 arch/arm64/include/asm/kexec.h      | 18 ++++++
 arch/arm64/kernel/asm-offsets.c     |  9 +++
 arch/arm64/kernel/cpu-reset.S       |  4 +-
 arch/arm64/kernel/cpu-reset.h       |  8 +--
 arch/arm64/kernel/machine_kexec.c   | 28 ++++++++-
 arch/arm64/kernel/relocate_kernel.S | 88 ++++++++++-------------------
 6 files changed, 86 insertions(+), 69 deletions(-)

diff --git a/arch/arm64/include/asm/kexec.h b/arch/arm64/include/asm/kexec.h
index d15ca1ca1e83..d5b79d4c7fae 100644
--- a/arch/arm64/include/asm/kexec.h
+++ b/arch/arm64/include/asm/kexec.h
@@ -90,12 +90,30 @@ static inline void crash_prepare_suspend(void) {}
 static inline void crash_post_resume(void) {}
 #endif
 
+/*
+ * kern_reloc_arg is passed to kernel relocation function as an argument.
+ * head		kimage->head, allows to traverse through relocation segments.
+ * entry_addr	kimage->start, where to jump from relocation function (new
+ *		kernel, or purgatory entry address).
+ * kern_arg0	first argument to kernel is its dtb address. The other
+ *		arguments are currently unused, and must be set to 0
+ */
+struct kern_reloc_arg {
+	unsigned long	head;
+	unsigned long	entry_addr;
+	unsigned long	kern_arg0;
+	unsigned long	kern_arg1;
+	unsigned long	kern_arg2;
+	unsigned long	kern_arg3;
+};
+
 #define ARCH_HAS_KIMAGE_ARCH
 
 struct kimage_arch {
 	void *dtb;
 	unsigned long dtb_mem;
 	unsigned long kern_reloc;
+	unsigned long kern_reloc_arg;
 };
 
 #ifdef CONFIG_KEXEC_FILE
diff --git a/arch/arm64/kernel/asm-offsets.c b/arch/arm64/kernel/asm-offsets.c
index 214685760e1c..900394907fd8 100644
--- a/arch/arm64/kernel/asm-offsets.c
+++ b/arch/arm64/kernel/asm-offsets.c
@@ -23,6 +23,7 @@
 #include <asm/suspend.h>
 #include <linux/kbuild.h>
 #include <linux/arm-smccc.h>
+#include <linux/kexec.h>
 
 int main(void)
 {
@@ -126,6 +127,14 @@ int main(void)
 #ifdef CONFIG_ARM_SDE_INTERFACE
   DEFINE(SDEI_EVENT_INTREGS,	offsetof(struct sdei_registered_event, interrupted_regs));
   DEFINE(SDEI_EVENT_PRIORITY,	offsetof(struct sdei_registered_event, priority));
+#endif
+#ifdef CONFIG_KEXEC_CORE
+  DEFINE(KRELOC_HEAD,		offsetof(struct kern_reloc_arg, head));
+  DEFINE(KRELOC_ENTRY_ADDR,	offsetof(struct kern_reloc_arg, entry_addr));
+  DEFINE(KRELOC_KERN_ARG0,	offsetof(struct kern_reloc_arg, kern_arg0));
+  DEFINE(KRELOC_KERN_ARG1,	offsetof(struct kern_reloc_arg, kern_arg1));
+  DEFINE(KRELOC_KERN_ARG2,	offsetof(struct kern_reloc_arg, kern_arg2));
+  DEFINE(KRELOC_KERN_ARG3,	offsetof(struct kern_reloc_arg, kern_arg3));
 #endif
   return 0;
 }
diff --git a/arch/arm64/kernel/cpu-reset.S b/arch/arm64/kernel/cpu-reset.S
index 6ea337d464c4..64c78a42919f 100644
--- a/arch/arm64/kernel/cpu-reset.S
+++ b/arch/arm64/kernel/cpu-reset.S
@@ -43,9 +43,7 @@ ENTRY(__cpu_soft_restart)
 	hvc	#0				// no return
 
 1:	mov	x18, x1				// entry
-	mov	x0, x2				// arg0
-	mov	x1, x3				// arg1
-	mov	x2, x4				// arg2
+	mov	x0, x2				// arg
 	br	x18
 ENDPROC(__cpu_soft_restart)
 
diff --git a/arch/arm64/kernel/cpu-reset.h b/arch/arm64/kernel/cpu-reset.h
index ed50e9587ad8..7a8720ff186f 100644
--- a/arch/arm64/kernel/cpu-reset.h
+++ b/arch/arm64/kernel/cpu-reset.h
@@ -11,12 +11,10 @@
 #include <asm/virt.h>
 
 void __cpu_soft_restart(unsigned long el2_switch, unsigned long entry,
-	unsigned long arg0, unsigned long arg1, unsigned long arg2);
+			unsigned long arg);
 
 static inline void __noreturn cpu_soft_restart(unsigned long entry,
-					       unsigned long arg0,
-					       unsigned long arg1,
-					       unsigned long arg2)
+					       unsigned long arg)
 {
 	typeof(__cpu_soft_restart) *restart;
 
@@ -25,7 +23,7 @@ static inline void __noreturn cpu_soft_restart(unsigned long entry,
 	restart = (void *)__pa_symbol(__cpu_soft_restart);
 
 	cpu_install_idmap();
-	restart(el2_switch, entry, arg0, arg1, arg2);
+	restart(el2_switch, entry, arg);
 	unreachable();
 }
 
diff --git a/arch/arm64/kernel/machine_kexec.c b/arch/arm64/kernel/machine_kexec.c
index 9b41da50e6f7..d745ea2051df 100644
--- a/arch/arm64/kernel/machine_kexec.c
+++ b/arch/arm64/kernel/machine_kexec.c
@@ -43,6 +43,7 @@ static void _kexec_image_info(const char *func, int line,
 	pr_debug("    head:        %lx\n", kimage->head);
 	pr_debug("    nr_segments: %lu\n", kimage->nr_segments);
 	pr_debug("    kern_reloc: %pa\n", &kimage->arch.kern_reloc);
+	pr_debug("    kern_reloc_arg: %pa\n", &kimage->arch.kern_reloc_arg);
 
 	for (i = 0; i < kimage->nr_segments; i++) {
 		pr_debug("      segment[%lu]: %016lx - %016lx, 0x%lx bytes, %lu pages\n",
@@ -59,14 +60,38 @@ void machine_kexec_cleanup(struct kimage *kimage)
 	/* Empty routine needed to avoid build errors. */
 }
 
+/* Allocates pages for kexec page table */
+static void *kexec_page_alloc(void *arg)
+{
+	struct kimage *kimage = (struct kimage *)arg;
+	struct page *page = kimage_alloc_control_pages(kimage, 0);
+
+	if (!page)
+		return NULL;
+
+	return page_address(page);
+}
+
 int machine_kexec_post_load(struct kimage *kimage)
 {
 	unsigned long kern_reloc;
+	struct kern_reloc_arg *kern_reloc_arg;
 
 	kern_reloc = page_to_phys(kimage->control_code_page);
 	memcpy(__va(kern_reloc), arm64_relocate_new_kernel,
 	       arm64_relocate_new_kernel_size);
+
+	kern_reloc_arg = kexec_page_alloc(kimage);
+	if (!kern_reloc_arg)
+		return -ENOMEM;
+	memset(kern_reloc_arg, 0, sizeof(struct kern_reloc_arg));
+
 	kimage->arch.kern_reloc = kern_reloc;
+	kimage->arch.kern_reloc_arg = __pa(kern_reloc_arg);
+
+	kern_reloc_arg->head = kimage->head;
+	kern_reloc_arg->entry_addr = kimage->start;
+	kern_reloc_arg->kern_arg0 = kimage->arch.dtb_mem;
 
 	kexec_image_info(kimage);
 	return 0;
@@ -203,8 +228,7 @@ void machine_kexec(struct kimage *kimage)
 	 * userspace (kexec-tools).
 	 * In kexec_file case, the kernel starts directly without purgatory.
 	 */
-	cpu_soft_restart(kimage->arch.kern_reloc, kimage->head, kimage->start,
-			 kimage->arch.dtb_mem);
+	cpu_soft_restart(kimage->arch.kern_reloc, kimage->arch.kern_reloc_arg);
 
 	BUG(); /* Should never get here. */
 }
diff --git a/arch/arm64/kernel/relocate_kernel.S b/arch/arm64/kernel/relocate_kernel.S
index c1d7db71a726..d352faf7cbe6 100644
--- a/arch/arm64/kernel/relocate_kernel.S
+++ b/arch/arm64/kernel/relocate_kernel.S
@@ -8,7 +8,7 @@
 
 #include <linux/kexec.h>
 #include <linux/linkage.h>
-
+#include <asm/asm-offsets.h>
 #include <asm/assembler.h>
 #include <asm/kexec.h>
 #include <asm/page.h>
@@ -17,86 +17,58 @@
 /*
  * arm64_relocate_new_kernel - Put a 2nd stage image in place and boot it.
  *
- * The memory that the old kernel occupies may be overwritten when coping the
+ * The memory that the old kernel occupies may be overwritten when copying the
  * new image to its final location.  To assure that the
  * arm64_relocate_new_kernel routine which does that copy is not overwritten,
  * all code and data needed by arm64_relocate_new_kernel must be between the
  * symbols arm64_relocate_new_kernel and arm64_relocate_new_kernel_end.  The
  * machine_kexec() routine will copy arm64_relocate_new_kernel to the kexec
- * control_code_page, a special page which has been set up to be preserved
- * during the copy operation.
+ * safe memory that has been set up to be preserved during the copy operation.
  */
 ENTRY(arm64_relocate_new_kernel)
-
-	/* Setup the list loop variables. */
-	mov	x18, x2				/* x18 = dtb address */
-	mov	x17, x1				/* x17 = kimage_start */
-	mov	x16, x0				/* x16 = kimage_head */
-	raw_dcache_line_size x15, x0		/* x15 = dcache line size */
-	mov	x14, xzr			/* x14 = entry ptr */
-	mov	x13, xzr			/* x13 = copy dest */
-
 	/* Clear the sctlr_el2 flags. */
-	mrs	x0, CurrentEL
-	cmp	x0, #CurrentEL_EL2
+	mrs	x2, CurrentEL
+	cmp	x2, #CurrentEL_EL2
 	b.ne	1f
-	mrs	x0, sctlr_el2
+	mrs	x2, sctlr_el2
 	ldr	x1, =SCTLR_ELx_FLAGS
-	bic	x0, x0, x1
+	bic	x2, x2, x1
 	pre_disable_mmu_workaround
-	msr	sctlr_el2, x0
+	msr	sctlr_el2, x2
 	isb
-1:
-
-	/* Check if the new image needs relocation. */
+1:	/* Check if the new image needs relocation. */
+	ldr	x16, [x0, #KRELOC_HEAD]		/* x16 = kimage_head */
 	tbnz	x16, IND_DONE_BIT, .Ldone
-
+	raw_dcache_line_size x15, x1		/* x15 = dcache line size */
 .Lloop:
 	and	x12, x16, PAGE_MASK		/* x12 = addr */
-
 	/* Test the entry flags. */
 .Ltest_source:
 	tbz	x16, IND_SOURCE_BIT, .Ltest_indirection
 
 	/* Invalidate dest page to PoC. */
-	mov     x0, x13
-	add     x20, x0, #PAGE_SIZE
+	mov     x2, x13
+	add     x20, x2, #PAGE_SIZE
 	sub     x1, x15, #1
-	bic     x0, x0, x1
-2:	dc      ivac, x0
-	add     x0, x0, x15
-	cmp     x0, x20
+	bic     x2, x2, x1
+2:	dc      ivac, x2
+	add     x2, x2, x15
+	cmp     x2, x20
 	b.lo    2b
 	dsb     sy
 
-	mov x20, x13
-	mov x21, x12
-	copy_page x20, x21, x0, x1, x2, x3, x4, x5, x6, x7
-
-	/* dest += PAGE_SIZE */
-	add	x13, x13, PAGE_SIZE
+	copy_page x13, x12, x1, x2, x3, x4, x5, x6, x7, x8
 	b	.Lnext
-
 .Ltest_indirection:
 	tbz	x16, IND_INDIRECTION_BIT, .Ltest_destination
-
-	/* ptr = addr */
-	mov	x14, x12
+	mov	x14, x12			/* ptr = addr */
 	b	.Lnext
-
 .Ltest_destination:
 	tbz	x16, IND_DESTINATION_BIT, .Lnext
-
-	/* dest = addr */
-	mov	x13, x12
-
+	mov	x13, x12			/* dest = addr */
 .Lnext:
-	/* entry = *ptr++ */
-	ldr	x16, [x14], #8
-
-	/* while (!(entry & DONE)) */
-	tbz	x16, IND_DONE_BIT, .Lloop
-
+	ldr	x16, [x14], #8			/* entry = *ptr++ */
+	tbz	x16, IND_DONE_BIT, .Lloop	/* while (!(entry & DONE)) */
 .Ldone:
 	/* wait for writes from copy_page to finish */
 	dsb	nsh
@@ -105,18 +77,16 @@ ENTRY(arm64_relocate_new_kernel)
 	isb
 
 	/* Start new image. */
-	mov	x0, x18
-	mov	x1, xzr
-	mov	x2, xzr
-	mov	x3, xzr
-	br	x17
-
-ENDPROC(arm64_relocate_new_kernel)
+	ldr	x4, [x0, #KRELOC_ENTRY_ADDR]	/* x4 = kimage_start */
+	ldr	x3, [x0, #KRELOC_KERN_ARG3]
+	ldr	x2, [x0, #KRELOC_KERN_ARG2]
+	ldr	x1, [x0, #KRELOC_KERN_ARG1]
+	ldr	x0, [x0, #KRELOC_KERN_ARG0]	/* x0 = dtb address */
+	br	x4
+END(arm64_relocate_new_kernel)
 
 .ltorg
-
 .align 3	/* To keep the 64-bit values below naturally aligned. */
-
 .Lcopy_end:
 .org	KEXEC_CONTROL_PAGE_SIZE
 
-- 
2.22.0

[PATCH v1 7/8] arm64, kexec: configure transitional page table for kexec

[Pavel Tatashin]

Configure a page table located in kexec-safe memory that has
the following mappings:

1. identity mapping for text of relocation function with executable permission.
2. identity mapping for argument for relocation function.
3. linear mappings for all source ranges
4. linear mappings for all destination ranges.

Also, configure el2_vector, that is used to jump to new kernel from EL2 on
non-VHE kernels.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
---
 arch/arm64/include/asm/kexec.h      |  32 +++++++
 arch/arm64/kernel/asm-offsets.c     |   6 ++
 arch/arm64/kernel/machine_kexec.c   | 129 ++++++++++++++++++++++++++--
 arch/arm64/kernel/relocate_kernel.S |  16 +++-
 4 files changed, 174 insertions(+), 9 deletions(-)

diff --git a/arch/arm64/include/asm/kexec.h b/arch/arm64/include/asm/kexec.h
index d5b79d4c7fae..450d8440f597 100644
--- a/arch/arm64/include/asm/kexec.h
+++ b/arch/arm64/include/asm/kexec.h
@@ -90,6 +90,23 @@ static inline void crash_prepare_suspend(void) {}
 static inline void crash_post_resume(void) {}
 #endif
 
+#if defined(CONFIG_KEXEC_CORE)
+/* Global variables for the arm64_relocate_new_kernel routine. */
+extern const unsigned char arm64_relocate_new_kernel[];
+extern const unsigned long arm64_relocate_new_kernel_size;
+
+/* Body of the vector for escalating to EL2 from relocation routine */
+extern const unsigned char kexec_el1_sync[];
+extern const unsigned long kexec_el1_sync_size;
+
+#define KEXEC_EL2_VECTOR_TABLE_SIZE	2048
+#define KEXEC_EL2_SYNC_OFFSET		(KEXEC_EL2_VECTOR_TABLE_SIZE / 2)
+
+#endif
+
+#define KEXEC_SRC_START	PAGE_OFFSET
+#define KEXEC_DST_START	(PAGE_OFFSET + \
+			((UL(0xffffffffffffffff) - PAGE_OFFSET) >> 1) + 1)
 /*
  * kern_reloc_arg is passed to kernel relocation function as an argument.
  * head		kimage->head, allows to traverse through relocation segments.
@@ -97,6 +114,15 @@ static inline void crash_post_resume(void) {}
  *		kernel, or purgatory entry address).
  * kern_arg0	first argument to kernel is its dtb address. The other
  *		arguments are currently unused, and must be set to 0
+ * trans_ttbr0	idmap for relocation function and its argument
+ * trans_ttbr1	linear map for source/destination addresses.
+ * el2_vector	If present means that relocation routine will go to EL1
+ *		from EL2 to do the copy, and then back to EL2 to do the jump
+ *		to new world. This vector contains only the final jump
+ *		instruction at KEXEC_EL2_SYNC_OFFSET.
+ * src_addr	linear map for source pages.
+ * dst_addr	linear map for destination pages.
+ * copy_len	Number of bytes that need to be copied
  */
 struct kern_reloc_arg {
 	unsigned long	head;
@@ -105,6 +131,12 @@ struct kern_reloc_arg {
 	unsigned long	kern_arg1;
 	unsigned long	kern_arg2;
 	unsigned long	kern_arg3;
+	unsigned long	trans_ttbr0;
+	unsigned long	trans_ttbr1;
+	unsigned long	el2_vector;
+	unsigned long	src_addr;
+	unsigned long	dst_addr;
+	unsigned long	copy_len;
 };
 
 #define ARCH_HAS_KIMAGE_ARCH
diff --git a/arch/arm64/kernel/asm-offsets.c b/arch/arm64/kernel/asm-offsets.c
index 900394907fd8..7c2ba09a8ceb 100644
--- a/arch/arm64/kernel/asm-offsets.c
+++ b/arch/arm64/kernel/asm-offsets.c
@@ -135,6 +135,12 @@ int main(void)
   DEFINE(KRELOC_KERN_ARG1,	offsetof(struct kern_reloc_arg, kern_arg1));
   DEFINE(KRELOC_KERN_ARG2,	offsetof(struct kern_reloc_arg, kern_arg2));
   DEFINE(KRELOC_KERN_ARG3,	offsetof(struct kern_reloc_arg, kern_arg3));
+  DEFINE(KRELOC_TRANS_TTBR0,	offsetof(struct kern_reloc_arg, trans_ttbr0));
+  DEFINE(KRELOC_TRANS_TTBR1,	offsetof(struct kern_reloc_arg, trans_ttbr1));
+  DEFINE(KRELOC_EL2_VECTOR,	offsetof(struct kern_reloc_arg, el2_vector));
+  DEFINE(KRELOC_SRC_ADDR,	offsetof(struct kern_reloc_arg, src_addr));
+  DEFINE(KRELOC_DST_ADDR,	offsetof(struct kern_reloc_arg, dst_addr));
+  DEFINE(KRELOC_COPY_LEN,	offsetof(struct kern_reloc_arg, copy_len));
 #endif
   return 0;
 }
diff --git a/arch/arm64/kernel/machine_kexec.c b/arch/arm64/kernel/machine_kexec.c
index d745ea2051df..16f761fc50c8 100644
--- a/arch/arm64/kernel/machine_kexec.c
+++ b/arch/arm64/kernel/machine_kexec.c
@@ -20,13 +20,10 @@
 #include <asm/mmu.h>
 #include <asm/mmu_context.h>
 #include <asm/page.h>
+#include <asm/trans_table.h>
 
 #include "cpu-reset.h"
 
-/* Global variables for the arm64_relocate_new_kernel routine. */
-extern const unsigned char arm64_relocate_new_kernel[];
-extern const unsigned long arm64_relocate_new_kernel_size;
-
 /**
  * kexec_image_info - For debugging output.
  */
@@ -72,15 +69,128 @@ static void *kexec_page_alloc(void *arg)
 	return page_address(page);
 }
 
+/*
+ * Map source segments starting from KEXEC_SRC_START, and map destination
+ * segments starting from KEXEC_DST_START, and return size of copy in
+ * *copy_len argument.
+ * Relocation function essentially needs to do:
+ * memcpy(KEXEC_DST_START, KEXEC_SRC_START, copy_len);
+ */
+static int map_segments(struct kimage *kimage, pgd_t *pgdp,
+			struct trans_table_info *info,
+			unsigned long *copy_len)
+{
+	unsigned long *ptr = 0;
+	unsigned long dest = 0;
+	unsigned long src_va = KEXEC_SRC_START;
+	unsigned long dst_va = KEXEC_DST_START;
+	unsigned long len = 0;
+	unsigned long entry, addr;
+	int rc;
+
+	for (entry = kimage->head; !(entry & IND_DONE); entry = *ptr++) {
+		addr = entry & PAGE_MASK;
+
+		switch (entry & IND_FLAGS) {
+		case IND_DESTINATION:
+			dest = addr;
+			break;
+		case IND_INDIRECTION:
+			ptr = __va(addr);
+			if (rc)
+				return rc;
+			break;
+		case IND_SOURCE:
+			rc = trans_table_map_page(info, pgdp, __va(addr),
+						  src_va, PAGE_KERNEL);
+			if (rc)
+				return rc;
+			rc = trans_table_map_page(info, pgdp, __va(dest),
+						  dst_va, PAGE_KERNEL);
+			if (rc)
+				return rc;
+			dest += PAGE_SIZE;
+			src_va += PAGE_SIZE;
+			dst_va += PAGE_SIZE;
+			len += PAGE_SIZE;
+		}
+	}
+	*copy_len = len;
+
+	return 0;
+}
+
+static int mmu_relocate_setup(struct kimage *kimage, unsigned long kern_reloc,
+			      struct kern_reloc_arg *kern_reloc_arg)
+{
+	struct trans_table_info info = {
+		.trans_alloc_page	= kexec_page_alloc,
+		.trans_alloc_arg	= kimage,
+		.trans_flags		= 0,
+	};
+	pgd_t *trans_ttbr0, *trans_ttbr1;
+	int rc;
+
+	rc = trans_table_create_empty(&info, &trans_ttbr0);
+	if (rc)
+		return rc;
+
+	rc = trans_table_create_empty(&info, &trans_ttbr1);
+	if (rc)
+		return rc;
+
+	rc = map_segments(kimage, trans_ttbr1, &info,
+			  &kern_reloc_arg->copy_len);
+	if (rc)
+		return rc;
+
+	/* Map relocation function va == pa */
+	rc = trans_table_map_page(&info, trans_ttbr0,  __va(kern_reloc),
+				  kern_reloc, PAGE_KERNEL_EXEC);
+	if (rc)
+		return rc;
+
+	/* Map relocation function argument va == pa */
+	rc = trans_table_map_page(&info, trans_ttbr0, kern_reloc_arg,
+				  __pa(kern_reloc_arg), PAGE_KERNEL);
+	if (rc)
+		return rc;
+
+	kern_reloc_arg->trans_ttbr0 = phys_to_ttbr(__pa(trans_ttbr0));
+	kern_reloc_arg->trans_ttbr1 = phys_to_ttbr(__pa(trans_ttbr1));
+	kern_reloc_arg->src_addr = KEXEC_SRC_START;
+	kern_reloc_arg->dst_addr = KEXEC_DST_START;
+
+	return 0;
+}
+
 int machine_kexec_post_load(struct kimage *kimage)
 {
+	unsigned long el2_vector = 0;
 	unsigned long kern_reloc;
 	struct kern_reloc_arg *kern_reloc_arg;
+	int rc = 0;
+
+	/*
+	 * Sanity check that relocation function + el2_vector fit into one
+	 * page.
+	 */
+	if (arm64_relocate_new_kernel_size > KEXEC_EL2_VECTOR_TABLE_SIZE) {
+		pr_err("can't fit relocation function and el2_vector in one page");
+		return -ENOMEM;
+	}
 
 	kern_reloc = page_to_phys(kimage->control_code_page);
 	memcpy(__va(kern_reloc), arm64_relocate_new_kernel,
 	       arm64_relocate_new_kernel_size);
 
+	/* Setup vector table only when EL2 is available, but no VHE */
+	if (is_hyp_mode_available() && !is_kernel_in_hyp_mode()) {
+		el2_vector = kern_reloc + KEXEC_EL2_VECTOR_TABLE_SIZE;
+		memcpy(__va(el2_vector + KEXEC_EL2_SYNC_OFFSET), kexec_el1_sync,
+		       kexec_el1_sync_size);
+	}
+
 	kern_reloc_arg = kexec_page_alloc(kimage);
 	if (!kern_reloc_arg)
 		return -ENOMEM;
@@ -91,10 +201,19 @@ int machine_kexec_post_load(struct kimage *kimage)
 
 	kern_reloc_arg->head = kimage->head;
 	kern_reloc_arg->entry_addr = kimage->start;
+	kern_reloc_arg->el2_vector = el2_vector;
 	kern_reloc_arg->kern_arg0 = kimage->arch.dtb_mem;
 
+	/*
+	 * If relocation is not needed, we do not need to enable MMU in
+	 * relocation routine, therefore do not create page tables for
+	 * scenarios such as crash kernel
+	 */
+	if (!(kimage->head & IND_DONE))
+		rc = mmu_relocate_setup(kimage, kern_reloc, kern_reloc_arg);
+
 	kexec_image_info(kimage);
-	return 0;
+	return rc;
 }
 
 /**
diff --git a/arch/arm64/kernel/relocate_kernel.S b/arch/arm64/kernel/relocate_kernel.S
index d352faf7cbe6..14243a678277 100644
--- a/arch/arm64/kernel/relocate_kernel.S
+++ b/arch/arm64/kernel/relocate_kernel.S
@@ -83,17 +83,25 @@ ENTRY(arm64_relocate_new_kernel)
 	ldr	x1, [x0, #KRELOC_KERN_ARG1]
 	ldr	x0, [x0, #KRELOC_KERN_ARG0]	/* x0 = dtb address */
 	br	x4
+.ltorg
+.Larm64_relocate_new_kernel_end:
 END(arm64_relocate_new_kernel)
 
-.ltorg
+ENTRY(kexec_el1_sync)
+	br	x4				/* Jump to new world from el2 */
+.Lkexec_el1_sync_end:
+END(kexec_el1_sync)
+
 .align 3	/* To keep the 64-bit values below naturally aligned. */
-.Lcopy_end:
 .org	KEXEC_CONTROL_PAGE_SIZE
-
 /*
  * arm64_relocate_new_kernel_size - Number of bytes to copy to the
  * control_code_page.
  */
 .globl arm64_relocate_new_kernel_size
 arm64_relocate_new_kernel_size:
-	.quad	.Lcopy_end - arm64_relocate_new_kernel
+	.quad	.Larm64_relocate_new_kernel_end - arm64_relocate_new_kernel
+
+.globl kexec_el1_sync_size
+kexec_el1_sync_size:
+	.quad	.Lkexec_el1_sync_end - kexec_el1_sync
-- 
2.22.0

[PATCH v1 8/8] arm64, kexec: enable MMU during kexec relocation

[Pavel Tatashin]

Now, that we have transitional page tables configured, temporarily enable
MMU to allow faster relocation of segments to final destination.

The performance data: for a moderate size kernel + initramfs: 25M the
relocation was taking 0.382s, with enabled MMU it now takes
0.019s only or x20 improvement.

The time is proportional to the size of relocation, therefore if initramfs
is larger, 100M it could take over a second.

Also, remove reloc_arg->head, as it is not needed anymore once MMU is
enabled.

Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
---
 arch/arm64/include/asm/kexec.h      |   2 -
 arch/arm64/kernel/asm-offsets.c     |   1 -
 arch/arm64/kernel/machine_kexec.c   |   1 -
 arch/arm64/kernel/relocate_kernel.S | 136 +++++++++++++++++-----------
 4 files changed, 84 insertions(+), 56 deletions(-)

diff --git a/arch/arm64/include/asm/kexec.h b/arch/arm64/include/asm/kexec.h
index 450d8440f597..ad81ed3e5751 100644
--- a/arch/arm64/include/asm/kexec.h
+++ b/arch/arm64/include/asm/kexec.h
@@ -109,7 +109,6 @@ extern const unsigned long kexec_el1_sync_size;
 			((UL(0xffffffffffffffff) - PAGE_OFFSET) >> 1) + 1)
 /*
  * kern_reloc_arg is passed to kernel relocation function as an argument.
- * head		kimage->head, allows to traverse through relocation segments.
  * entry_addr	kimage->start, where to jump from relocation function (new
  *		kernel, or purgatory entry address).
  * kern_arg0	first argument to kernel is its dtb address. The other
@@ -125,7 +124,6 @@ extern const unsigned long kexec_el1_sync_size;
  * copy_len	Number of bytes that need to be copied
  */
 struct kern_reloc_arg {
-	unsigned long	head;
 	unsigned long	entry_addr;
 	unsigned long	kern_arg0;
 	unsigned long	kern_arg1;
diff --git a/arch/arm64/kernel/asm-offsets.c b/arch/arm64/kernel/asm-offsets.c
index 7c2ba09a8ceb..13ad00b1b90f 100644
--- a/arch/arm64/kernel/asm-offsets.c
+++ b/arch/arm64/kernel/asm-offsets.c
@@ -129,7 +129,6 @@ int main(void)
   DEFINE(SDEI_EVENT_PRIORITY,	offsetof(struct sdei_registered_event, priority));
 #endif
 #ifdef CONFIG_KEXEC_CORE
-  DEFINE(KRELOC_HEAD,		offsetof(struct kern_reloc_arg, head));
   DEFINE(KRELOC_ENTRY_ADDR,	offsetof(struct kern_reloc_arg, entry_addr));
   DEFINE(KRELOC_KERN_ARG0,	offsetof(struct kern_reloc_arg, kern_arg0));
   DEFINE(KRELOC_KERN_ARG1,	offsetof(struct kern_reloc_arg, kern_arg1));
diff --git a/arch/arm64/kernel/machine_kexec.c b/arch/arm64/kernel/machine_kexec.c
index 16f761fc50c8..b5ff5fdb4777 100644
--- a/arch/arm64/kernel/machine_kexec.c
+++ b/arch/arm64/kernel/machine_kexec.c
@@ -199,7 +199,6 @@ int machine_kexec_post_load(struct kimage *kimage)
 	kimage->arch.kern_reloc = kern_reloc;
 	kimage->arch.kern_reloc_arg = __pa(kern_reloc_arg);
 
-	kern_reloc_arg->head = kimage->head;
 	kern_reloc_arg->entry_addr = kimage->start;
 	kern_reloc_arg->el2_vector = el2_vector;
 	kern_reloc_arg->kern_arg0 = kimage->arch.dtb_mem;
diff --git a/arch/arm64/kernel/relocate_kernel.S b/arch/arm64/kernel/relocate_kernel.S
index 14243a678277..96ff6760bd9c 100644
--- a/arch/arm64/kernel/relocate_kernel.S
+++ b/arch/arm64/kernel/relocate_kernel.S
@@ -4,6 +4,8 @@
  *
  * Copyright (C) Linaro.
  * Copyright (C) Huawei Futurewei Technologies.
+ * Copyright (c) 2019, Microsoft Corporation.
+ * Pavel Tatashin <patatash@linux.microsoft.com>
  */
 
 #include <linux/kexec.h>
@@ -14,6 +16,49 @@
 #include <asm/page.h>
 #include <asm/sysreg.h>
 
+/* Invalidae TLB */
+.macro tlb_invalidate
+	dsb	sy
+	dsb	ish
+	tlbi	vmalle1
+	dsb	ish
+	isb
+.endm
+
+/* Turn-off mmu at level specified by sctlr */
+.macro turn_off_mmu sctlr, tmp1, tmp2
+	mrs	\tmp1, \sctlr
+	ldr	\tmp2, =SCTLR_ELx_FLAGS
+	bic	\tmp1, \tmp1, \tmp2
+	pre_disable_mmu_workaround
+	msr	\sctlr, \tmp1
+	isb
+.endm
+
+/* Turn-on mmu at level specified by sctlr */
+.macro turn_on_mmu sctlr, tmp1, tmp2
+	mrs	\tmp1, \sctlr
+	ldr	\tmp2, =SCTLR_ELx_FLAGS
+	orr	\tmp1, \tmp1, \tmp2
+	msr	\sctlr, \tmp1
+	ic	iallu
+	dsb	nsh
+	isb
+.endm
+
+/*
+ * Set ttbr0 and ttbr1, called while MMU is disabled, so no need to temporarily
+ * set zero_page table. Invalidate TLB after new tables are set.
+ */
+.macro set_ttbr arg, tmp
+	ldr	\tmp, [\arg, #KRELOC_TRANS_TTBR0]
+	msr	ttbr0_el1, \tmp
+	ldr	\tmp, [\arg, #KRELOC_TRANS_TTBR1]
+	offset_ttbr1 \tmp
+	msr	ttbr1_el1, \tmp
+	isb
+.endm
+
 /*
  * arm64_relocate_new_kernel - Put a 2nd stage image in place and boot it.
  *
@@ -24,65 +69,52 @@
  * symbols arm64_relocate_new_kernel and arm64_relocate_new_kernel_end.  The
  * machine_kexec() routine will copy arm64_relocate_new_kernel to the kexec
  * safe memory that has been set up to be preserved during the copy operation.
+ *
+ * This function temporarily enables MMU if kernel relocation is needed.
+ * Also, if we enter this function at EL2 on non-VHE kernel, we temporarily go
+ * to EL1 to enable MMU, and escalate back to EL2 at the end to do the jump to
+ * the new kernel. This is determined by presence of el2_vector.
  */
 ENTRY(arm64_relocate_new_kernel)
-	/* Clear the sctlr_el2 flags. */
-	mrs	x2, CurrentEL
-	cmp	x2, #CurrentEL_EL2
+	mrs	x1, CurrentEL
+	cmp	x1, #CurrentEL_EL2
 	b.ne	1f
-	mrs	x2, sctlr_el2
-	ldr	x1, =SCTLR_ELx_FLAGS
-	bic	x2, x2, x1
-	pre_disable_mmu_workaround
-	msr	sctlr_el2, x2
-	isb
-1:	/* Check if the new image needs relocation. */
-	ldr	x16, [x0, #KRELOC_HEAD]		/* x16 = kimage_head */
-	tbnz	x16, IND_DONE_BIT, .Ldone
-	raw_dcache_line_size x15, x1		/* x15 = dcache line size */
-.Lloop:
-	and	x12, x16, PAGE_MASK		/* x12 = addr */
-	/* Test the entry flags. */
-.Ltest_source:
-	tbz	x16, IND_SOURCE_BIT, .Ltest_indirection
-
-	/* Invalidate dest page to PoC. */
-	mov     x2, x13
-	add     x20, x2, #PAGE_SIZE
-	sub     x1, x15, #1
-	bic     x2, x2, x1
-2:	dc      ivac, x2
-	add     x2, x2, x15
-	cmp     x2, x20
-	b.lo    2b
-	dsb     sy
-
-	copy_page x13, x12, x1, x2, x3, x4, x5, x6, x7, x8
-	b	.Lnext
-.Ltest_indirection:
-	tbz	x16, IND_INDIRECTION_BIT, .Ltest_destination
-	mov	x14, x12			/* ptr = addr */
-	b	.Lnext
-.Ltest_destination:
-	tbz	x16, IND_DESTINATION_BIT, .Lnext
-	mov	x13, x12			/* dest = addr */
-.Lnext:
-	ldr	x16, [x14], #8			/* entry = *ptr++ */
-	tbz	x16, IND_DONE_BIT, .Lloop	/* while (!(entry & DONE)) */
-.Ldone:
-	/* wait for writes from copy_page to finish */
-	dsb	nsh
-	ic	iallu
-	dsb	nsh
-	isb
-
-	/* Start new image. */
-	ldr	x4, [x0, #KRELOC_ENTRY_ADDR]	/* x4 = kimage_start */
+	turn_off_mmu sctlr_el2, x1, x2		/* Turn off MMU at EL2 */
+1:	mov	x20, xzr			/* x20 will hold vector value */
+	ldr	x11, [x0, #KRELOC_COPY_LEN]
+	cbz	x11, 5f				/* Check if need to relocate */
+	ldr	x20, [x0, #KRELOC_EL2_VECTOR]
+	cbz	x20, 2f				/* need to reduce to EL1? */
+	msr	vbar_el2, x20			/* el2_vector present, means */
+	adr	x1, 2f				/* we will do copy in el1 but */
+	msr	elr_el2, x1			/* do final jump from el2 */
+	eret					/* Reduce to EL1 */
+2:	set_ttbr x0, x1				/* Set our page tables */
+	tlb_invalidate
+	turn_on_mmu sctlr_el1, x1, x2		/* Turn MMU back on */
+	ldr	x1, [x0, #KRELOC_DST_ADDR];
+	ldr	x2, [x0, #KRELOC_SRC_ADDR];
+	mov	x12, x1				/* x12 dst backup */
+3:	copy_page x1, x2, x3, x4, x5, x6, x7, x8, x9, x10
+	sub	x11, x11, #PAGE_SIZE
+	cbnz	x11, 3b				/* page copy loop */
+	raw_dcache_line_size x2, x3		/* x2 = dcache line size */
+	sub	x3, x2, #1			/* x3 = dcache_size - 1 */
+	bic	x12, x12, x3
+4:	dc	cvau, x12			/* Flush D-cache */
+	add	x12, x12, x2
+	cmp	x12, x1				/* Compare to dst + len */
+	b.ne	4b				/* D-cache flush loop */
+	turn_off_mmu sctlr_el1, x1, x2		/* Turn off MMU */
+	tlb_invalidate				/* Invalidate TLB */
+5:	ldr	x4, [x0, #KRELOC_ENTRY_ADDR]	/* x4 = kimage_start */
 	ldr	x3, [x0, #KRELOC_KERN_ARG3]
 	ldr	x2, [x0, #KRELOC_KERN_ARG2]
 	ldr	x1, [x0, #KRELOC_KERN_ARG1]
 	ldr	x0, [x0, #KRELOC_KERN_ARG0]	/* x0 = dtb address */
-	br	x4
+	cbnz	x20, 6f				/* need to escalate to el2? */
+	br	x4				/* Jump to new world */
+6:	hvc	#0				/* enters kexec_el1_sync */
 .ltorg
 .Larm64_relocate_new_kernel_end:
 END(arm64_relocate_new_kernel)
-- 
2.22.0

Re: [PATCH v1 0/8] arm64: MMU enabled kexec relocation

[Pavel Tatashin]

Just a friendly reminder, please send your comments on this series.
It's been a week since I sent out these patches, and no feedback yet.
Also, I'd appreciate if anyone could test this series on vhe hardware
with vhe kernel, it does not look like QEMU can emulate it yet

Thank you,
Pasha

On Thu, Aug 1, 2019 at 11:24 AM Pavel Tatashin
<pasha.tatashin@soleen.com> wrote:
>
> Enable MMU during kexec relocation in order to improve reboot performance.
>
> If kexec functionality is used for a fast system update, with a minimal
> downtime, the relocation of kernel + initramfs takes a significant portion
> of reboot.
>
> The reason for slow relocation is because it is done without MMU, and thus
> not benefiting from D-Cache.
>
> Performance data
> ----------------
> For this experiment, the size of kernel plus initramfs is small, only 25M.
> If initramfs was larger, than the improvements would be greater, as time
> spent in relocation is proportional to the size of relocation.
>
> Previously:
> kernel shutdown 0.022131328s
> relocation      0.440510736s
> kernel startup  0.294706768s
>
> Relocation was taking: 58.2% of reboot time
>
> Now:
> kernel shutdown 0.032066576s
> relocation      0.022158152s
> kernel startup  0.296055880s
>
> Now: Relocation takes 6.3% of reboot time
>
> Total reboot is x2.16 times faster.
>
> Previous approaches and discussions
> -----------------------------------
> https://lore.kernel.org/lkml/20190709182014.16052-1-pasha.tatashin@soleen.com
> reserve space for kexec to avoid relocation, involves changes to generic code
> to optimize a problem that exists on arm64 only:
>
> https://lore.kernel.org/lkml/20190716165641.6990-1-pasha.tatashin@soleen.com
> The first attempt to enable MMU, some bugs that prevented performance
> improvement. The page tables unnecessary configured idmap for the whole
> physical space.
>
> https://lore.kernel.org/lkml/20190731153857.4045-1-pasha.tatashin@soleen.com
> No linear copy, bug with EL2 reboots.
>
> Pavel Tatashin (8):
>   kexec: quiet down kexec reboot
>   arm64, mm: transitional tables
>   arm64: hibernate: switch to transtional page tables.
>   kexec: add machine_kexec_post_load()
>   arm64, kexec: move relocation function setup and clean up
>   arm64, kexec: add expandable argument to relocation function
>   arm64, kexec: configure transitional page table for kexec
>   arm64, kexec: enable MMU during kexec relocation
>
>  arch/arm64/Kconfig                     |   4 +
>  arch/arm64/include/asm/kexec.h         |  51 ++++-
>  arch/arm64/include/asm/pgtable-hwdef.h |   1 +
>  arch/arm64/include/asm/trans_table.h   |  68 ++++++
>  arch/arm64/kernel/asm-offsets.c        |  14 ++
>  arch/arm64/kernel/cpu-reset.S          |   4 +-
>  arch/arm64/kernel/cpu-reset.h          |   8 +-
>  arch/arm64/kernel/hibernate.c          | 261 ++++++-----------------
>  arch/arm64/kernel/machine_kexec.c      | 199 ++++++++++++++----
>  arch/arm64/kernel/relocate_kernel.S    | 196 +++++++++---------
>  arch/arm64/mm/Makefile                 |   1 +
>  arch/arm64/mm/trans_table.c            | 273 +++++++++++++++++++++++++
>  kernel/kexec.c                         |   4 +
>  kernel/kexec_core.c                    |   8 +-
>  kernel/kexec_file.c                    |   4 +
>  kernel/kexec_internal.h                |   2 +
>  16 files changed, 758 insertions(+), 340 deletions(-)
>  create mode 100644 arch/arm64/include/asm/trans_table.h
>  create mode 100644 arch/arm64/mm/trans_table.c
>
> --
> 2.22.0
>

Re: [PATCH v1 0/8] arm64: MMU enabled kexec relocation

[Pavel Tatashin]

Hi,

It is been two weeks, and no review activity yet. Please help with
reviewing this work.

Thank you,
Pasha

On Thu, Aug 8, 2019 at 2:44 PM Pavel Tatashin <pasha.tatashin@soleen.com> wrote:
>
> Just a friendly reminder, please send your comments on this series.
> It's been a week since I sent out these patches, and no feedback yet.
> Also, I'd appreciate if anyone could test this series on vhe hardware
> with vhe kernel, it does not look like QEMU can emulate it yet
>
> Thank you,
> Pasha
>
> On Thu, Aug 1, 2019 at 11:24 AM Pavel Tatashin
> <pasha.tatashin@soleen.com> wrote:
> >
> > Enable MMU during kexec relocation in order to improve reboot performance.
> >
> > If kexec functionality is used for a fast system update, with a minimal
> > downtime, the relocation of kernel + initramfs takes a significant portion
> > of reboot.
> >
> > The reason for slow relocation is because it is done without MMU, and thus
> > not benefiting from D-Cache.
> >
> > Performance data
> > ----------------
> > For this experiment, the size of kernel plus initramfs is small, only 25M.
> > If initramfs was larger, than the improvements would be greater, as time
> > spent in relocation is proportional to the size of relocation.
> >
> > Previously:
> > kernel shutdown 0.022131328s
> > relocation      0.440510736s
> > kernel startup  0.294706768s
> >
> > Relocation was taking: 58.2% of reboot time
> >
> > Now:
> > kernel shutdown 0.032066576s
> > relocation      0.022158152s
> > kernel startup  0.296055880s
> >
> > Now: Relocation takes 6.3% of reboot time
> >
> > Total reboot is x2.16 times faster.
> >
> > Previous approaches and discussions
> > -----------------------------------
> > https://lore.kernel.org/lkml/20190709182014.16052-1-pasha.tatashin@soleen.com
> > reserve space for kexec to avoid relocation, involves changes to generic code
> > to optimize a problem that exists on arm64 only:
> >
> > https://lore.kernel.org/lkml/20190716165641.6990-1-pasha.tatashin@soleen.com
> > The first attempt to enable MMU, some bugs that prevented performance
> > improvement. The page tables unnecessary configured idmap for the whole
> > physical space.
> >
> > https://lore.kernel.org/lkml/20190731153857.4045-1-pasha.tatashin@soleen.com
> > No linear copy, bug with EL2 reboots.
> >
> > Pavel Tatashin (8):
> >   kexec: quiet down kexec reboot
> >   arm64, mm: transitional tables
> >   arm64: hibernate: switch to transtional page tables.
> >   kexec: add machine_kexec_post_load()
> >   arm64, kexec: move relocation function setup and clean up
> >   arm64, kexec: add expandable argument to relocation function
> >   arm64, kexec: configure transitional page table for kexec
> >   arm64, kexec: enable MMU during kexec relocation
> >
> >  arch/arm64/Kconfig                     |   4 +
> >  arch/arm64/include/asm/kexec.h         |  51 ++++-
> >  arch/arm64/include/asm/pgtable-hwdef.h |   1 +
> >  arch/arm64/include/asm/trans_table.h   |  68 ++++++
> >  arch/arm64/kernel/asm-offsets.c        |  14 ++
> >  arch/arm64/kernel/cpu-reset.S          |   4 +-
> >  arch/arm64/kernel/cpu-reset.h          |   8 +-
> >  arch/arm64/kernel/hibernate.c          | 261 ++++++-----------------
> >  arch/arm64/kernel/machine_kexec.c      | 199 ++++++++++++++----
> >  arch/arm64/kernel/relocate_kernel.S    | 196 +++++++++---------
> >  arch/arm64/mm/Makefile                 |   1 +
> >  arch/arm64/mm/trans_table.c            | 273 +++++++++++++++++++++++++
> >  kernel/kexec.c                         |   4 +
> >  kernel/kexec_core.c                    |   8 +-
> >  kernel/kexec_file.c                    |   4 +
> >  kernel/kexec_internal.h                |   2 +
> >  16 files changed, 758 insertions(+), 340 deletions(-)
> >  create mode 100644 arch/arm64/include/asm/trans_table.h
> >  create mode 100644 arch/arm64/mm/trans_table.c
> >
> > --
> > 2.22.0
> >

Re: [PATCH v1 0/8] arm64: MMU enabled kexec relocation

[James Morse]

Hi Pavel,

On 08/08/2019 19:44, Pavel Tatashin wrote:
> Just a friendly reminder, please send your comments on this series.

(Please don't top-post)


> It's been a week since I sent out these patches, and no feedback yet.

A week is not a lot of time, people are busy, go to conferences, some even dare to take
holiday!


> Also, I'd appreciate if anyone could test this series on vhe hardware
> with vhe kernel, it does not look like QEMU can emulate it yet

This locks up during resume from hibernate on my AMD Seattle, a regular v8.0 machine.


Please try and build the series to reduce review time. What you have here is an all-new
page-table generation API, which you switch hibernate and kexec too. This is effectively a
new implementation of hibernate and kexec. There are three things here that need review.

You have a regression in your all-new implementation of hibernate. It took six months (and
lots of review) to get the existing code right, please don't rip it out if there is
nothing wrong with it.


Instead, please just move the hibernate copy_page_tables() code, and then wire kexec up.
You shouldn't need to change anything in the copy_page_tables() code as the linear map is
the same in both cases.


It looks like you are creating the page tables just after the kexec:segments have been
loaded. This will go horribly wrong if anything changes between then and kexec time. (e.g.
memory you've got mapped gets hot-removed).
This needs to be done as late as possible, so we don't waste memory, and the world can't
change around us. Reboot notifiers run before kexec, can't we do the memory-allocation there?


> On Thu, Aug 1, 2019 at 11:24 AM Pavel Tatashin
> <pasha.tatashin@soleen.com> wrote:
>>
>> Enable MMU during kexec relocation in order to improve reboot performance.
>>
>> If kexec functionality is used for a fast system update, with a minimal
>> downtime, the relocation of kernel + initramfs takes a significant portion
>> of reboot.
>>
>> The reason for slow relocation is because it is done without MMU, and thus
>> not benefiting from D-Cache.
>>
>> Performance data
>> ----------------
>> For this experiment, the size of kernel plus initramfs is small, only 25M.
>> If initramfs was larger, than the improvements would be greater, as time
>> spent in relocation is proportional to the size of relocation.
>>
>> Previously:
>> kernel shutdown 0.022131328s
>> relocation      0.440510736s
>> kernel startup  0.294706768s
>>
>> Relocation was taking: 58.2% of reboot time
>>
>> Now:
>> kernel shutdown 0.032066576s
>> relocation      0.022158152s
>> kernel startup  0.296055880s
>>
>> Now: Relocation takes 6.3% of reboot time
>>
>> Total reboot is x2.16 times faster.

When I first saw these numbers they were ~'0.29s', which I wrongly assumed was 29 seconds.
Savings in milliseconds, for _reboot_ is a hard sell. I'm hoping that on the machines that
take minutes to kexec we'll get numbers that make this change more convincing.


Thanks,

James

Re: [PATCH v1 2/8] arm64, mm: transitional tables

[James Morse]

Hi Pavel,

On 01/08/2019 16:24, Pavel Tatashin wrote:
> There are cases where normal kernel pages tables, i.e. idmap_pg_dir
> and swapper_pg_dir are not sufficient because they may be overwritten.
> 
> This happens when we transition from one world to another: for example
> during kexec kernel relocation transition, and also during hibernate
> kernel restore transition.
> 
> In these cases, if MMU is needed, the page table memory must be allocated
> from a safe place. Transitional tables is intended to allow just that.

> diff --git a/arch/arm64/include/asm/pgtable-hwdef.h b/arch/arm64/include/asm/pgtable-hwdef.h
> index db92950bb1a0..dcb4f13c7888 100644
> --- a/arch/arm64/include/asm/pgtable-hwdef.h
> +++ b/arch/arm64/include/asm/pgtable-hwdef.h
> @@ -110,6 +110,7 @@
>  #define PUD_TABLE_BIT		(_AT(pudval_t, 1) << 1)
>  #define PUD_TYPE_MASK		(_AT(pudval_t, 3) << 0)
>  #define PUD_TYPE_SECT		(_AT(pudval_t, 1) << 0)
> +#define PUD_SECT_RDONLY		(_AT(pudval_t, 1) << 7)		/* AP[2] */

This shouldn't be needed. As far as I'm aware, we only get read-only pages in the linear
map from debug-pagealloc, and the module aliases. Both of which require the linear map to
be made of page-size mappings.

Where are you seeing these?


> diff --git a/arch/arm64/include/asm/trans_table.h b/arch/arm64/include/asm/trans_table.h
> new file mode 100644
> index 000000000000..c7aef70587a1
> --- /dev/null
> +++ b/arch/arm64/include/asm/trans_table.h
> @@ -0,0 +1,68 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +
> +/*
> + * Copyright (c) 2019, Microsoft Corporation.
> + * Pavel Tatashin <patatash@linux.microsoft.com>
> + */
> +
> +#ifndef _ASM_TRANS_TABLE_H
> +#define _ASM_TRANS_TABLE_H
> +
> +#include <linux/bits.h>
> +#include <asm/pgtable-types.h>
> +
> +/*
> + * trans_alloc_page
> + *	- Allocator that should return exactly one uninitilaized page, if this
> + *	 allocator fails, trans_table returns -ENOMEM error.
> + *
> + * trans_alloc_arg
> + *	- Passed to trans_alloc_page as an argument
> + *
> + * trans_flags
> + *	- bitmap with flags that control how page table is filled.
> + *	  TRANS_MKWRITE: during page table copy make PTE, PME, and PUD page
> + *			 writeable by removing RDONLY flag from PTE.
> + *	  TRANS_MKVALID: during page table copy, if PTE present, but not valid,
> + *			 make it valid.
> + *	  TRANS_CHECKPFN: During page table copy, for every PTE entry check that
> + *			  PFN that this PTE points to is valid. Otherwise return
> + *			  -ENXIO

Adding top-level global knobs to manipulate the copied linear map is going to lead to
bugs. The existing code will only change the PTE in specific circumstances, that it tests
for, that only happen at the PTE level.


> + *	  TRANS_FORCEMAP: During page map, if translation exists, force
> + *			  overwrite it. Otherwise -ENXIO may be returned by
> + *			  trans_table_map_* functions if conflict is detected.

This one, sounds like a very bad idea.


Thanks,

James

Re: [PATCH v1 0/8] arm64: MMU enabled kexec relocation

[Pavel Tatashin]

Hi James,

Thank you for your feedback. My replies below:

> > Also, I'd appreciate if anyone could test this series on vhe hardware
> > with vhe kernel, it does not look like QEMU can emulate it yet
>
> This locks up during resume from hibernate on my AMD Seattle, a regular v8.0 machine.

Thanks for reporting a bug I will root cause and fix it.

> Please try and build the series to reduce review time. What you have here is an all-new
> page-table generation API, which you switch hibernate and kexec too. This is effectively a
> new implementation of hibernate and kexec. There are three things here that need review.
>
> You have a regression in your all-new implementation of hibernate. It took six months (and
> lots of review) to get the existing code right, please don't rip it out if there is
> nothing wrong with it.

> Instead, please just move the hibernate copy_page_tables() code, and then wire kexec up.
> You shouldn't need to change anything in the copy_page_tables() code as the linear map is
> the same in both cases.

It is not really an all-new implementation of hibernate (for kexec it
is true though). I used the current implementation of hibernate as
bases, and simply generalized the functions by providing a flexible
interface. So what you are asking is actually exactly what I am doing.
I realize, that I introduced a bug that I will fix.

> It looks like you are creating the page tables just after the kexec:segments have been
> loaded. This will go horribly wrong if anything changes between then and kexec time. (e.g.
> memory you've got mapped gets hot-removed).
> This needs to be done as late as possible, so we don't waste memory, and the world can't
> change around us. Reboot notifiers run before kexec, can't we do the memory-allocation there?

Kexec by design does not allow allocate during kexec time. This is
because we cannot fail during kexec syscall. All allocations must be
done during kexec load time. Kernel memory cannot be hot-removed, as
it is not part of ZONE_MOVABLE, and cannot be migrated.

The current implementation relies on this assumption as well: during
load time the (struct kimage) -> head contains the physical addresses
of sources and destinations. If sources can be moved, this array will
be broken.


> >> Previously:
> >> kernel shutdown 0.022131328s
> >> relocation      0.440510736s
> >> kernel startup  0.294706768s
> >>
> >> Relocation was taking: 58.2% of reboot time
> >>
> >> Now:
> >> kernel shutdown 0.032066576s
> >> relocation      0.022158152s
> >> kernel startup  0.296055880s
> >>
> >> Now: Relocation takes 6.3% of reboot time
> >>
> >> Total reboot is x2.16 times faster.
>
> When I first saw these numbers they were ~'0.29s', which I wrongly assumed was 29 seconds.
> Savings in milliseconds, for _reboot_ is a hard sell. I'm hoping that on the machines that
> take minutes to kexec we'll get numbers that make this change more convincing.

Sure, this userland is very small kernel+userland is only 47M. Here is
another data point: fitImage: 380M, it contains a larger userland.
The numbers for kernel shutdown and startup are the same as this is
the same kernel, but relocation takes: 3.58s

shutdown: 0.02s
relocation: 3.58s
startup:  0.30s

Relocation take 88% of reboot time. And, we must have it under one second.

Thank you,
Pasha

Re: [PATCH v1 2/8] arm64, mm: transitional tables

[Pavel Tatashin]

On Thu, Aug 15, 2019 at 2:11 PM James Morse <james.morse@arm.com> wrote:
>
> Hi Pavel,
>
> On 01/08/2019 16:24, Pavel Tatashin wrote:
> > There are cases where normal kernel pages tables, i.e. idmap_pg_dir
> > and swapper_pg_dir are not sufficient because they may be overwritten.
> >
> > This happens when we transition from one world to another: for example
> > during kexec kernel relocation transition, and also during hibernate
> > kernel restore transition.
> >
> > In these cases, if MMU is needed, the page table memory must be allocated
> > from a safe place. Transitional tables is intended to allow just that.
>
> > diff --git a/arch/arm64/include/asm/pgtable-hwdef.h b/arch/arm64/include/asm/pgtable-hwdef.h
> > index db92950bb1a0..dcb4f13c7888 100644
> > --- a/arch/arm64/include/asm/pgtable-hwdef.h
> > +++ b/arch/arm64/include/asm/pgtable-hwdef.h
> > @@ -110,6 +110,7 @@
> >  #define PUD_TABLE_BIT                (_AT(pudval_t, 1) << 1)
> >  #define PUD_TYPE_MASK                (_AT(pudval_t, 3) << 0)
> >  #define PUD_TYPE_SECT                (_AT(pudval_t, 1) << 0)
> > +#define PUD_SECT_RDONLY              (_AT(pudval_t, 1) << 7)         /* AP[2] */
>
> This shouldn't be needed. As far as I'm aware, we only get read-only pages in the linear
> map from debug-pagealloc, and the module aliases. Both of which require the linear map to
> be made of page-size mappings.
>
> Where are you seeing these?

This was done simply for generalization.

In old copy_pud:

445   set_pud(dst_pudp,
446   __pud(pud_val(pud) & ~PMD_SECT_RDONLY));

In new trans_table_copy_pud():
119                 if (!pud_table(src_pud)) {
120                         if (info->trans_flags & TRANS_MKWRITE)
121                                 pud_val(src_pud) &= ~PUD_SECT_RDONLY;

If you want, I can replace it with PMD_SECT_RDONLY

> > + * trans_flags
> > + *   - bitmap with flags that control how page table is filled.
> > + *     TRANS_MKWRITE: during page table copy make PTE, PME, and PUD page
> > + *                    writeable by removing RDONLY flag from PTE.
> > + *     TRANS_MKVALID: during page table copy, if PTE present, but not valid,
> > + *                    make it valid.
> > + *     TRANS_CHECKPFN: During page table copy, for every PTE entry check that
> > + *                     PFN that this PTE points to is valid. Otherwise return
> > + *                     -ENXIO
>
> Adding top-level global knobs to manipulate the copied linear map is going to lead to
> bugs. The existing code will only change the PTE in specific circumstances, that it tests
> for, that only happen at the PTE level.

I am simply doing the same what the old code is doing:

hibernate sets this flag if: debug_pagealloc_enabled() and trans_table
in this case check if pfn_valid() or not. This is part of
generalization, without which it is not possible to re-use the code
between kexec and hibernate.

>
>
> > + *     TRANS_FORCEMAP: During page map, if translation exists, force
> > + *                     overwrite it. Otherwise -ENXIO may be returned by
> > + *                     trans_table_map_* functions if conflict is detected.
>

I will remove it.

Thank you,
Pasha

Re: [PATCH v1 0/8] arm64: MMU enabled kexec relocation

[James Morse]

Hi Pavel,

On 15/08/2019 21:09, Pavel Tatashin wrote:
>>> Also, I'd appreciate if anyone could test this series on vhe hardware
>>> with vhe kernel, it does not look like QEMU can emulate it yet
>>
>> This locks up during resume from hibernate on my AMD Seattle, a regular v8.0 machine.
> 
> Thanks for reporting a bug I will root cause and fix it.

>> Please try and build the series to reduce review time. What you have here is an all-new
>> page-table generation API, which you switch hibernate and kexec too. This is effectively a
>> new implementation of hibernate and kexec. There are three things here that need review.
>>
>> You have a regression in your all-new implementation of hibernate. It took six months (and
>> lots of review) to get the existing code right, please don't rip it out if there is
>> nothing wrong with it.
> 
>> Instead, please just move the hibernate copy_page_tables() code, and then wire kexec up.
>> You shouldn't need to change anything in the copy_page_tables() code as the linear map is
>> the same in both cases.

> It is not really an all-new implementation of hibernate (for kexec it
> is true though). I used the current implementation of hibernate as
> bases, and simply generalized the functions by providing a flexible
> interface. So what you are asking is actually exactly what I am doing.

I disagree. The resume page-table code is the bulk of the complexity in hibernate.c. Your
first patch dumps ~200 lines of differently-complex code, and your second switches
hibernate over to it.

Instead, please move that code, keeping it as it is. git will spot the move, and the
generated diffstat should only reflect the build-system changes. You don't need to 'switch
hibernate to transitional page tables.'

Adding kexec will then show-up what needs changing, each change comes with a commit
message explaining why. Having these as 'generalisations' in the first patch is a mess.

There is existing code that we don't want to break. Any changes need to be done as a
sequence of small incremental changes. It can't be reviewed any other way.


> I realize, that I introduced a bug that I will fix.

Done as a sequence of small incremental changes, I could bisect it to the patch that
introduces the bug, and probably fix it from the description in the commit message.


>> It looks like you are creating the page tables just after the kexec:segments have been
>> loaded. This will go horribly wrong if anything changes between then and kexec time. (e.g.
>> memory you've got mapped gets hot-removed).
>> This needs to be done as late as possible, so we don't waste memory, and the world can't
>> change around us. Reboot notifiers run before kexec, can't we do the memory-allocation there?

> Kexec by design does not allow allocate during kexec time. This is
> because we cannot fail during kexec syscall.

This problem needs solving.

| Reboot notifiers run before kexec, can't we do the memory-allocation there?


> All allocations must be done during kexec load time.

This increases the memory footprint. I don't think we should waste ~2MB per GB of kernel
memory on this feature. (Assuming 4K pages and rodata_full)

Another option is to allocate this memory at load time, but then free it so it can be used
in the meantime. You can keep the list of allocated pfn, as we know they aren't in use by
the running kernel, kexec metadata, loaded images etc.

Memory hotplug would need handling carefully, as would anything that 'donates' memory to
another agent. (I suspect the TEE stuff does this, I don't know how it interacts with kexec)


> Kernel memory cannot be hot-removed, as
> it is not part of ZONE_MOVABLE, and cannot be migrated.

Today, yes. Tomorrow?, "arm64/mm: Enable memory hot remove":
https://lore.kernel.org/r/1563171470-3117-1-git-send-email-anshuman.khandual@arm.com


>>>> Previously:
>>>> kernel shutdown 0.022131328s
>>>> relocation      0.440510736s
>>>> kernel startup  0.294706768s
>>>>
>>>> Relocation was taking: 58.2% of reboot time
>>>>
>>>> Now:
>>>> kernel shutdown 0.032066576s
>>>> relocation      0.022158152s
>>>> kernel startup  0.296055880s
>>>>
>>>> Now: Relocation takes 6.3% of reboot time
>>>>
>>>> Total reboot is x2.16 times faster.
>>
>> When I first saw these numbers they were ~'0.29s', which I wrongly assumed was 29 seconds.
>> Savings in milliseconds, for _reboot_ is a hard sell. I'm hoping that on the machines that
>> take minutes to kexec we'll get numbers that make this change more convincing.

> Sure, this userland is very small kernel+userland is only 47M. Here is
> another data point: fitImage: 380M, it contains a larger userland.
> The numbers for kernel shutdown and startup are the same as this is
> the same kernel, but relocation takes: 3.58s
> shutdown: 0.02s
> relocation: 3.58s
> startup:  0.30s
> 
> Relocation take 88% of reboot time. And, we must have it under one second.

Where does this one second number come from? (was it ever a reasonable starting point?)


Thanks,

James

Re: [PATCH v1 0/8] arm64: MMU enabled kexec relocation

[Pavel Tatashin]

Hi James,

Thank you for your feedback, my replies below:

> > It is not really an all-new implementation of hibernate (for kexec it
> > is true though). I used the current implementation of hibernate as
> > bases, and simply generalized the functions by providing a flexible
> > interface. So what you are asking is actually exactly what I am doing.
>
> I disagree. The resume page-table code is the bulk of the complexity in hibernate.c. Your
> first patch dumps ~200 lines of differently-complex code, and your second switches
> hibernate over to it.

OK, I will make the change incremental.

>
> Instead, please move that code, keeping it as it is. git will spot the move, and the
> generated diffstat should only reflect the build-system changes. You don't need to 'switch
> hibernate to transitional page tables.'
>
> Adding kexec will then show-up what needs changing, each change comes with a commit
> message explaining why. Having these as 'generalisations' in the first patch is a mess.

Makes sense, I will fix it.

>
> There is existing code that we don't want to break. Any changes need to be done as a
> sequence of small incremental changes. It can't be reviewed any other way.
>
>
> > I realize, that I introduced a bug that I will fix.
>
> Done as a sequence of small incremental changes, I could bisect it to the patch that
> introduces the bug, and probably fix it from the description in the commit message.

BTW, I root caused it, there were two trivial errors:
1. In "arm64, mm: transitional tables"
int i = pgd_index(addr);
In trans_table_copy_*:
should be: pte_index(), pmd_index(), pud_index(), accordingly.
2. In trans_table_create_copy()
pgd_offset_k(PAGE_OFFSET) should be: mm_init.pgd

> >> It looks like you are creating the page tables just after the kexec:segments have been
> >> loaded. This will go horribly wrong if anything changes between then and kexec time. (e.g.
> >> memory you've got mapped gets hot-removed).
> >> This needs to be done as late as possible, so we don't waste memory, and the world can't
> >> change around us. Reboot notifiers run before kexec, can't we do the memory-allocation there?
>
> > Kexec by design does not allow allocate during kexec time. This is
> > because we cannot fail during kexec syscall.
>
> This problem needs solving.
>
> | Reboot notifiers run before kexec, can't we do the memory-allocation there?
>
>
> > All allocations must be done during kexec load time.
>
> This increases the memory footprint. I don't think we should waste ~2MB per GB of kernel
> memory on this feature. (Assuming 4K pages and rodata_full)
>
> Another option is to allocate this memory at load time, but then free it so it can be used
> in the meantime. You can keep the list of allocated pfn, as we know they aren't in use by
> the running kernel, kexec metadata, loaded images etc.

This is until a new kernel module is loaded, I do not think this is safe to do.

In my opinion 2M per 1 GB is a fair trade off for a faster kexec
performance. Unlike with crash kexec for which we do not add any
memory useage, the kernel does not have to be all the time in memory,
but can be loaded by user before reboot. If machine is so scare on
memory resources that 2M per 1G matters, user simply won't keep new
kernel in memory until it is actually needed.

>
> Memory hotplug would need handling carefully, as would anything that 'donates' memory to
> another agent. (I suspect the TEE stuff does this, I don't know how it interacts with kexec)
>
>
> > Kernel memory cannot be hot-removed, as
> > it is not part of ZONE_MOVABLE, and cannot be migrated.
>
> Today, yes. Tomorrow?, "arm64/mm: Enable memory hot remove":
> https://lore.kernel.org/r/1563171470-3117-1-git-send-email-anshuman.khandual@arm.com

I understand that ARM64 is about to get hot-remove feature, but what I
am saying is that my feature does not introduce new problem because
the current kexec code assumes that kernel memory is not movable
(array of sparse physical source dest addresses in kimage->head). It
is possible to offline and hot-remove only memory that can be freed by
page migration, the pages that were allocated for kexec kernel are not
one of them.

> >>>> Previously:
> >>>> kernel shutdown 0.022131328s
> >>>> relocation      0.440510736s
> >>>> kernel startup  0.294706768s
> >>>>
> >>>> Relocation was taking: 58.2% of reboot time
> >>>>
> >>>> Now:
> >>>> kernel shutdown 0.032066576s
> >>>> relocation      0.022158152s
> >>>> kernel startup  0.296055880s
> >>>>
> >>>> Now: Relocation takes 6.3% of reboot time
> >>>>
> >>>> Total reboot is x2.16 times faster.
> >>
> >> When I first saw these numbers they were ~'0.29s', which I wrongly assumed was 29 seconds.
> >> Savings in milliseconds, for _reboot_ is a hard sell. I'm hoping that on the machines that
> >> take minutes to kexec we'll get numbers that make this change more convincing.
>
> > Sure, this userland is very small kernel+userland is only 47M. Here is
> > another data point: fitImage: 380M, it contains a larger userland.
> > The numbers for kernel shutdown and startup are the same as this is
> > the same kernel, but relocation takes: 3.58s
> > shutdown: 0.02s
> > relocation: 3.58s
> > startup:  0.30s
> >
> > Relocation take 88% of reboot time. And, we must have it under one second.
>
> Where does this one second number come from? (was it ever a reasonable starting point?)

Currently we have two fitImages for this system in development: one
that has a bare minimal userland, only ~40 packages, and another has a
more complete userland. So, my first experiment shows the data from
this first bare minimum ftImage, the second experiment from the second
more complete fitImage. As I stated in cover letter, kexec time is
proportional to the size of the image and this series fixes this
scalability issue by making relocation  ~20 times faster.

Pasha