From: "Mike Rapoport (Microsoft)" Instead of checking if kho is finalized in each caller of __kho_preserve_order(), do it in the core function itself. Signed-off-by: Mike Rapoport (Microsoft) --- kernel/kexec_handover.c | 55 +++++++++++++++++++---------------------- 1 file changed, 26 insertions(+), 29 deletions(-) diff --git a/kernel/kexec_handover.c b/kernel/kexec_handover.c index 8079fc4b9189..f421acc58c1f 100644 --- a/kernel/kexec_handover.c +++ b/kernel/kexec_handover.c @@ -91,6 +91,29 @@ struct kho_serialization { struct khoser_mem_chunk *preserved_mem_map; }; +struct kho_out { + struct blocking_notifier_head chain_head; + + struct dentry *dir; + + struct mutex lock; /* protects KHO FDT finalization */ + + struct kho_serialization ser; + bool finalized; +}; + +static struct kho_out kho_out = { + .chain_head = BLOCKING_NOTIFIER_INIT(kho_out.chain_head), + .lock = __MUTEX_INITIALIZER(kho_out.lock), + .ser = { + .fdt_list = LIST_HEAD_INIT(kho_out.ser.fdt_list), + .track = { + .orders = XARRAY_INIT(kho_out.ser.track.orders, 0), + }, + }, + .finalized = false, +}; + static void *xa_load_or_alloc(struct xarray *xa, unsigned long index, size_t sz) { void *elm, *res; @@ -149,6 +172,9 @@ static int __kho_preserve_order(struct kho_mem_track *track, unsigned long pfn, might_sleep(); + if (kho_out.finalized) + return -EBUSY; + physxa = xa_load(&track->orders, order); if (!physxa) { int err; @@ -640,29 +666,6 @@ int kho_add_subtree(struct kho_serialization *ser, const char *name, void *fdt) } EXPORT_SYMBOL_GPL(kho_add_subtree); -struct kho_out { - struct blocking_notifier_head chain_head; - - struct dentry *dir; - - struct mutex lock; /* protects KHO FDT finalization */ - - struct kho_serialization ser; - bool finalized; -}; - -static struct kho_out kho_out = { - .chain_head = BLOCKING_NOTIFIER_INIT(kho_out.chain_head), - .lock = __MUTEX_INITIALIZER(kho_out.lock), - .ser = { - .fdt_list = LIST_HEAD_INIT(kho_out.ser.fdt_list), - .track = { - .orders = XARRAY_INIT(kho_out.ser.track.orders, 0), - }, - }, - .finalized = false, -}; - int register_kho_notifier(struct notifier_block *nb) { return blocking_notifier_chain_register(&kho_out.chain_head, nb); @@ -690,9 +693,6 @@ int kho_preserve_folio(struct folio *folio) const unsigned int order = folio_order(folio); struct kho_mem_track *track = &kho_out.ser.track; - if (kho_out.finalized) - return -EBUSY; - return __kho_preserve_order(track, pfn, order); } EXPORT_SYMBOL_GPL(kho_preserve_folio); @@ -716,9 +716,6 @@ int kho_preserve_phys(phys_addr_t phys, size_t size) int err = 0; struct kho_mem_track *track = &kho_out.ser.track; - if (kho_out.finalized) - return -EBUSY; - if (!PAGE_ALIGNED(phys) || !PAGE_ALIGNED(size)) return -EINVAL; -- 2.50.1 From: "Mike Rapoport (Microsoft)" to make it clear that KHO operates on pages rather than on a random physical address. The kho_preserve_pages() will be also used in upcoming support for vmalloc preservation. Signed-off-by: Mike Rapoport (Microsoft) --- include/linux/kexec_handover.h | 5 +++-- kernel/kexec_handover.c | 25 +++++++++++-------------- mm/memblock.c | 4 +++- 3 files changed, 17 insertions(+), 17 deletions(-) diff --git a/include/linux/kexec_handover.h b/include/linux/kexec_handover.h index 348844cffb13..cc5c49b0612b 100644 --- a/include/linux/kexec_handover.h +++ b/include/linux/kexec_handover.h @@ -18,6 +18,7 @@ enum kho_event { struct folio; struct notifier_block; +struct page; #define DECLARE_KHOSER_PTR(name, type) \ union { \ @@ -42,7 +43,7 @@ struct kho_serialization; bool kho_is_enabled(void); int kho_preserve_folio(struct folio *folio); -int kho_preserve_phys(phys_addr_t phys, size_t size); +int kho_preserve_pages(struct page *page, unsigned int nr_pages); struct folio *kho_restore_folio(phys_addr_t phys); int kho_add_subtree(struct kho_serialization *ser, const char *name, void *fdt); int kho_retrieve_subtree(const char *name, phys_addr_t *phys); @@ -65,7 +66,7 @@ static inline int kho_preserve_folio(struct folio *folio) return -EOPNOTSUPP; } -static inline int kho_preserve_phys(phys_addr_t phys, size_t size) +static inline int kho_preserve_pages(struct page *page, unsigned int nr_pages) { return -EOPNOTSUPP; } diff --git a/kernel/kexec_handover.c b/kernel/kexec_handover.c index f421acc58c1f..3ad59c5f9eaa 100644 --- a/kernel/kexec_handover.c +++ b/kernel/kexec_handover.c @@ -698,26 +698,23 @@ int kho_preserve_folio(struct folio *folio) EXPORT_SYMBOL_GPL(kho_preserve_folio); /** - * kho_preserve_phys - preserve a physically contiguous range across kexec. - * @phys: physical address of the range. - * @size: size of the range. + * kho_preserve_pages - preserve contiguous pages across kexec + * @page: first page in the list. + * @nr_pages: number of pages. * - * Instructs KHO to preserve the memory range from @phys to @phys + @size - * across kexec. + * Preserve a contiguous list of order 0 pages. Must be restored using + * kho_restore_page() on each order 0 page. * * Return: 0 on success, error code on failure */ -int kho_preserve_phys(phys_addr_t phys, size_t size) +int kho_preserve_pages(struct page *page, unsigned int nr_pages) { - unsigned long pfn = PHYS_PFN(phys); + struct kho_mem_track *track = &kho_out.ser.track; + const unsigned long start_pfn = page_to_pfn(page); + const unsigned long end_pfn = start_pfn + nr_pages; + unsigned long pfn = start_pfn; unsigned long failed_pfn = 0; - const unsigned long start_pfn = pfn; - const unsigned long end_pfn = PHYS_PFN(phys + size); int err = 0; - struct kho_mem_track *track = &kho_out.ser.track; - - if (!PAGE_ALIGNED(phys) || !PAGE_ALIGNED(size)) - return -EINVAL; while (pfn < end_pfn) { const unsigned int order = @@ -737,7 +734,7 @@ int kho_preserve_phys(phys_addr_t phys, size_t size) return err; } -EXPORT_SYMBOL_GPL(kho_preserve_phys); +EXPORT_SYMBOL_GPL(kho_preserve_pages); /* Handling for debug/kho/out */ diff --git a/mm/memblock.c b/mm/memblock.c index 117d963e677c..6ec3eaa4e8d1 100644 --- a/mm/memblock.c +++ b/mm/memblock.c @@ -2516,8 +2516,10 @@ static int reserve_mem_kho_finalize(struct kho_serialization *ser) for (i = 0; i < reserved_mem_count; i++) { struct reserve_mem_table *map = &reserved_mem_table[i]; + struct page *page = phys_to_page(map->start); + unsigned int nr_pages = map->size >> PAGE_SHIFT; - err |= kho_preserve_phys(map->start, map->size); + err |= kho_preserve_pages(page, nr_pages); } err |= kho_preserve_folio(page_folio(kho_fdt)); -- 2.50.1 From: "Mike Rapoport (Microsoft)" A vmalloc allocation is preserved using binary structure similar to global KHO memory tracker. It's a linked list of pages where each page is an array of physical address of pages in vmalloc area. kho_preserve_vmalloc() hands out the physical address of the head page to the caller. This address is used as the argument to kho_vmalloc_restore() to restore the mapping in the vmalloc address space and populate it with the preserved pages. Signed-off-by: Mike Rapoport (Microsoft) --- include/linux/kexec_handover.h | 21 +++ kernel/kexec_handover.c | 244 +++++++++++++++++++++++++++++++++ 2 files changed, 265 insertions(+) diff --git a/include/linux/kexec_handover.h b/include/linux/kexec_handover.h index cc5c49b0612b..1cb515d8257a 100644 --- a/include/linux/kexec_handover.h +++ b/include/linux/kexec_handover.h @@ -39,12 +39,22 @@ struct page; struct kho_serialization; +struct kho_vmalloc_chunk; +struct kho_vmalloc { + DECLARE_KHOSER_PTR(first, struct kho_vmalloc_chunk *); + unsigned int total_pages; + unsigned short flags; + unsigned short order; +}; + #ifdef CONFIG_KEXEC_HANDOVER bool kho_is_enabled(void); int kho_preserve_folio(struct folio *folio); int kho_preserve_pages(struct page *page, unsigned int nr_pages); +int kho_preserve_vmalloc(void *ptr, struct kho_vmalloc *preservation); struct folio *kho_restore_folio(phys_addr_t phys); +void *kho_restore_vmalloc(const struct kho_vmalloc *preservation); int kho_add_subtree(struct kho_serialization *ser, const char *name, void *fdt); int kho_retrieve_subtree(const char *name, phys_addr_t *phys); @@ -71,11 +81,22 @@ static inline int kho_preserve_pages(struct page *page, unsigned int nr_pages) return -EOPNOTSUPP; } +static inline int kho_preserve_vmalloc(void *ptr, + struct kho_vmalloc *preservation) +{ + return -EOPNOTSUPP; +} + static inline struct folio *kho_restore_folio(phys_addr_t phys) { return NULL; } +static inline void *kho_restore_vmalloc(const struct kho_vmalloc *preservation) +{ + return NULL; +} + static inline int kho_add_subtree(struct kho_serialization *ser, const char *name, void *fdt) { diff --git a/kernel/kexec_handover.c b/kernel/kexec_handover.c index 3ad59c5f9eaa..d670caf6d07f 100644 --- a/kernel/kexec_handover.c +++ b/kernel/kexec_handover.c @@ -18,6 +18,7 @@ #include #include #include +#include #include @@ -736,6 +737,249 @@ int kho_preserve_pages(struct page *page, unsigned int nr_pages) } EXPORT_SYMBOL_GPL(kho_preserve_pages); +struct kho_vmalloc_hdr { + DECLARE_KHOSER_PTR(next, struct kho_vmalloc_chunk *); +}; + +#define KHO_VMALLOC_SIZE \ + ((PAGE_SIZE - sizeof(struct kho_vmalloc_hdr)) / \ + sizeof(phys_addr_t)) + +struct kho_vmalloc_chunk { + struct kho_vmalloc_hdr hdr; + phys_addr_t phys[KHO_VMALLOC_SIZE]; +}; + +static_assert(sizeof(struct kho_vmalloc_chunk) == PAGE_SIZE); + +/* vmalloc flags KHO supports */ +#define KHO_VMALLOC_SUPPORTED_FLAGS (VM_ALLOC | VM_ALLOW_HUGE_VMAP) + +/* KHO internal flags for vmalloc preservations */ +#define KHO_VMALLOC_ALLOC 0x0001 +#define KHO_VMALLOC_HUGE_VMAP 0x0002 + +static unsigned short vmalloc_flags_to_kho(unsigned int vm_flags) +{ + unsigned short kho_flags = 0; + + if (vm_flags & VM_ALLOC) + kho_flags |= KHO_VMALLOC_ALLOC; + if (vm_flags & VM_ALLOW_HUGE_VMAP) + kho_flags |= KHO_VMALLOC_HUGE_VMAP; + + return kho_flags; +} + +static unsigned int kho_flags_to_vmalloc(unsigned short kho_flags) +{ + unsigned int vm_flags = 0; + + if (kho_flags & KHO_VMALLOC_ALLOC) + vm_flags |= VM_ALLOC; + if (kho_flags & KHO_VMALLOC_HUGE_VMAP) + vm_flags |= VM_ALLOW_HUGE_VMAP; + + return vm_flags; +} + +static struct kho_vmalloc_chunk *new_vmalloc_chunk(struct kho_vmalloc_chunk *cur) +{ + struct kho_vmalloc_chunk *chunk; + int err; + + chunk = (struct kho_vmalloc_chunk *)get_zeroed_page(GFP_KERNEL); + if (!chunk) + return NULL; + + err = kho_preserve_pages(virt_to_page(chunk), 1); + if (err) + goto err_free; + if (cur) + KHOSER_STORE_PTR(cur->hdr.next, chunk); + return chunk; + +err_free: + free_page((unsigned long)chunk); + return NULL; +} + +static void kho_vmalloc_unpreserve_chunk(struct kho_vmalloc_chunk *chunk) +{ + struct kho_mem_track *track = &kho_out.ser.track; + unsigned long pfn = PHYS_PFN(virt_to_phys(chunk)); + + __kho_unpreserve(track, pfn, pfn + 1); + + for (int i = 0; chunk->phys[i]; i++) { + pfn = PHYS_PFN(chunk->phys[i]); + __kho_unpreserve(track, pfn, pfn + 1); + } +} + +static void kho_vmalloc_free_chunks(struct kho_vmalloc *kho_vmalloc) +{ + struct kho_vmalloc_chunk *chunk = KHOSER_LOAD_PTR(kho_vmalloc->first); + + while (chunk) { + struct kho_vmalloc_chunk *tmp = chunk; + + kho_vmalloc_unpreserve_chunk(chunk); + + chunk = KHOSER_LOAD_PTR(chunk->hdr.next); + kfree(tmp); + } +} + +/** + * kho_preserve_vmalloc - preserve memory allocated with vmalloc() across kexec + * @ptr: pointer to the area in vmalloc address space + * @preservation: placeholder for preservation metadata + * + * Instructs KHO to preserve the area in vmalloc address space at @ptr. The + * physical pages mapped at @ptr will be preserved and on successful return + * @preservation will hold the physical address of a structure that describes + * the preservation. + * + * NOTE: The memory allocated with vmalloc_node() variants cannot be reliably + * restored on the same node + * + * Return: 0 on success, error code on failure + */ +int kho_preserve_vmalloc(void *ptr, struct kho_vmalloc *preservation) +{ + struct kho_vmalloc_chunk *chunk; + struct vm_struct *vm = find_vm_area(ptr); + unsigned int order, flags, nr_contig_pages; + unsigned int idx = 0; + int err; + + if (!vm) + return -EINVAL; + + if (vm->flags & ~KHO_VMALLOC_SUPPORTED_FLAGS) + return -EOPNOTSUPP; + + flags = vmalloc_flags_to_kho(vm->flags); + order = get_vm_area_page_order(vm); + + chunk = new_vmalloc_chunk(NULL); + if (!chunk) + return -ENOMEM; + KHOSER_STORE_PTR(preservation->first, chunk); + + nr_contig_pages = (1 << order); + for (int i = 0; i < vm->nr_pages; i += nr_contig_pages) { + phys_addr_t phys = page_to_phys(vm->pages[i]); + + err = kho_preserve_pages(vm->pages[i], nr_contig_pages); + if (err) + goto err_free; + + chunk->phys[idx++] = phys; + if (idx == ARRAY_SIZE(chunk->phys)) { + chunk = new_vmalloc_chunk(chunk); + if (!chunk) + goto err_free; + idx = 0; + } + } + + preservation->total_pages = vm->nr_pages; + preservation->flags = flags; + preservation->order = order; + + return 0; + +err_free: + kho_vmalloc_free_chunks(preservation); + return err; +} +EXPORT_SYMBOL_GPL(kho_preserve_vmalloc); + +/** + * kho_restore_vmalloc - recreates and populates an area in vmalloc address + * space from the preserved memory. + * @preservation: preservation metadata. + * + * Recreates an area in vmalloc address space and populates it with memory that + * was preserved using kho_preserve_vmalloc(). + * + * Return: pointer to the area in the vmalloc address space, NULL on failure. + */ +void *kho_restore_vmalloc(const struct kho_vmalloc *preservation) +{ + struct kho_vmalloc_chunk *chunk = KHOSER_LOAD_PTR(preservation->first); + unsigned int align, order, shift, vm_flags; + unsigned int idx = 0, nr; + unsigned long addr, size; + struct vm_struct *area; + struct page **pages; + int err; + + vm_flags = kho_flags_to_vmalloc(preservation->flags); + if (vm_flags & ~KHO_VMALLOC_SUPPORTED_FLAGS) + return NULL; + + nr = preservation->total_pages; + pages = kvmalloc_array(nr, sizeof(*pages), GFP_KERNEL); + if (!pages) + return NULL; + order = preservation->order; + shift = PAGE_SHIFT + order; + align = 1 << shift; + + while (chunk) { + struct page *page; + + for (int i = 0; chunk->phys[i]; i++) { + phys_addr_t phys = chunk->phys[i]; + + if (idx + (1 << order) > nr) + goto err_free_pages_array; + + for (int j = 0; j < (1 << order); j++) { + page = phys_to_page(phys); + kho_restore_page(page, 0); + pages[idx++] = page; + phys += PAGE_SIZE; + } + } + + page = virt_to_page(chunk); + chunk = KHOSER_LOAD_PTR(chunk->hdr.next); + kho_restore_page(page, 0); + __free_page(page); + } + + if (idx != nr) + goto err_free_pages_array; + + area = __get_vm_area_node(nr * PAGE_SIZE, align, shift, vm_flags, + VMALLOC_START, VMALLOC_END, NUMA_NO_NODE, + GFP_KERNEL, __builtin_return_address(0)); + if (!area) + goto err_free_pages_array; + + addr = (unsigned long)area->addr; + size = get_vm_area_size(area); + err = vmap_pages_range(addr, addr + size, PAGE_KERNEL, pages, shift); + if (err) + goto err_free_vm_area; + + area->pages = pages; + area->nr_pages = nr; + + return area->addr; + +err_free_vm_area: + free_vm_area(area); +err_free_pages_array: + kvfree(pages); + return NULL; +} +EXPORT_SYMBOL_GPL(kho_restore_vmalloc); + /* Handling for debug/kho/out */ static struct dentry *debugfs_root; -- 2.50.1 From: "Mike Rapoport (Microsoft)" KHO test stores physical addresses of the preserved folios directly in fdt. Use kho_preserve_vmalloc() instead of it and kho_restore_vmalloc() to retrieve the addresses after kexec. This makes the test more scalable from one side and adds tests coverage for kho_preserve_vmalloc() from the other. Signed-off-by: Mike Rapoport (Microsoft) --- lib/test_kho.c | 41 +++++++++++++++++++++++++++++------------ 1 file changed, 29 insertions(+), 12 deletions(-) diff --git a/lib/test_kho.c b/lib/test_kho.c index fe8504e3407b..60cd899ea745 100644 --- a/lib/test_kho.c +++ b/lib/test_kho.c @@ -32,6 +32,7 @@ module_param(max_mem, long, 0644); struct kho_test_state { unsigned int nr_folios; struct folio **folios; + phys_addr_t *folios_info; struct folio *fdt; __wsum csum; }; @@ -67,18 +68,15 @@ static struct notifier_block kho_test_nb = { static int kho_test_save_data(struct kho_test_state *state, void *fdt) { - phys_addr_t *folios_info; + phys_addr_t *folios_info __free(kvfree) = NULL; + struct kho_vmalloc folios_info_phys; int err = 0; - err |= fdt_begin_node(fdt, "data"); - err |= fdt_property(fdt, "nr_folios", &state->nr_folios, - sizeof(state->nr_folios)); - err |= fdt_property_placeholder(fdt, "folios_info", - state->nr_folios * sizeof(*folios_info), - (void **)&folios_info); - err |= fdt_property(fdt, "csum", &state->csum, sizeof(state->csum)); - err |= fdt_end_node(fdt); + folios_info = vmalloc_array(state->nr_folios, sizeof(*folios_info)); + if (!folios_info) + return -ENOMEM; + err = kho_preserve_vmalloc(folios_info, &folios_info_phys); if (err) return err; @@ -93,6 +91,17 @@ static int kho_test_save_data(struct kho_test_state *state, void *fdt) break; } + err |= fdt_begin_node(fdt, "data"); + err |= fdt_property(fdt, "nr_folios", &state->nr_folios, + sizeof(state->nr_folios)); + err |= fdt_property(fdt, "folios_info", &folios_info_phys, + sizeof(folios_info_phys)); + err |= fdt_property(fdt, "csum", &state->csum, sizeof(state->csum)); + err |= fdt_end_node(fdt); + + if (!err) + state->folios_info = no_free_ptr(folios_info); + return err; } @@ -209,8 +218,9 @@ static int kho_test_save(void) static int kho_test_restore_data(const void *fdt, int node) { + const struct kho_vmalloc *folios_info_phys; const unsigned int *nr_folios; - const phys_addr_t *folios_info; + phys_addr_t *folios_info; const __wsum *old_csum; __wsum csum = 0; int len; @@ -225,8 +235,12 @@ static int kho_test_restore_data(const void *fdt, int node) if (!old_csum || len != sizeof(*old_csum)) return -EINVAL; - folios_info = fdt_getprop(fdt, node, "folios_info", &len); - if (!folios_info || len != sizeof(*folios_info) * *nr_folios) + folios_info_phys = fdt_getprop(fdt, node, "folios_info", &len); + if (!folios_info_phys || len != sizeof(*folios_info_phys)) + return -EINVAL; + + folios_info = kho_restore_vmalloc(folios_info_phys); + if (!folios_info) return -EINVAL; for (int i = 0; i < *nr_folios; i++) { @@ -246,6 +260,8 @@ static int kho_test_restore_data(const void *fdt, int node) folio_put(folio); } + vfree(folios_info); + if (csum != *old_csum) return -EINVAL; @@ -304,6 +320,7 @@ static void kho_test_cleanup(void) folio_put(kho_test_state.folios[i]); kvfree(kho_test_state.folios); + vfree(kho_test_state.folios_info); folio_put(kho_test_state.fdt); } -- 2.50.1