Commit 2f939f96 authored by Nikita Panov's avatar Nikita Panov
Browse files

mm: introduce kernel cmdline option "kernel_replication=" 

kunpeng inclusion
category: feature
bugzilla: https://gitee.com/openeuler/kernel/issues/IBPRS0



-------------------------------------------------

The new cmdline option allows enabling/disabling
kernel replication feature.
By default, the replication feature is disabled.
This allows to set CONFIG_KERNEL_REPLICATION=y by default,
but enabling this feature is done via kernel cmdline.

Signed-off-by: default avatarNikita Panov <panov.nikita@huawei.com>
parent 2caf0ac9

Documentation/admin-guide/kernel-parameters.txt

+7 −0
Original line number Diff line number Diff line
@@ -2534,6 +2534,13 @@ 
			some extension. These two features are alternatives.

			Current only arm64 is supported.



	kernel_replication=

			[ARM64]

			Format: [on|off]

			If CONFIG_KERNEL_REPLICATION is set, it allows

			enabling/disabling the kernel replication feature

			via cmdline. Default value is off.



	kgdbdbgp=	[KGDB,HW] kgdb over EHCI usb debug port.

			Format: <Controller#>[,poll interval]

			The controller # is the number of the ehci usb debug

arch/arm64/configs/openeuler_defconfig

+1 −1
Original line number Diff line number Diff line
@@ -1212,7 +1212,7 @@ CONFIG_ARM64_HAFT=y 
CONFIG_ARCH_SUPPORTS_PER_VMA_LOCK=y

CONFIG_PER_VMA_LOCK=y

CONFIG_LOCK_MM_AND_FIND_VMA=y

# CONFIG_KERNEL_REPLICATION is not set

CONFIG_KERNEL_REPLICATION=y

CONFIG_IOMMU_MM_DATA=y

# CONFIG_ASCEND_FEATURES is not set

CONFIG_PAGE_CACHE_LIMIT=y

arch/arm64/mm/pgd.c

+42 −14
Original line number Diff line number Diff line
@@ -18,6 +18,24 @@ 


static struct kmem_cache *pgd_cache __ro_after_init;



static pgd_t *pgd_alloc_orig(struct mm_struct *mm)

{

	gfp_t gfp = GFP_PGTABLE_USER;



	if (PGD_SIZE == PAGE_SIZE)

		return (pgd_t *)__get_free_page(gfp);

	else

		return kmem_cache_alloc(pgd_cache, gfp);

}



static void pgd_free_orig(struct mm_struct *mm, pgd_t *pgd)

{

	if (PGD_SIZE == PAGE_SIZE)

		free_page((unsigned long)pgd);

	else

		kmem_cache_free(pgd_cache, pgd);

}



#ifdef CONFIG_KERNEL_REPLICATION

pgd_t *page_pgd_alloc(struct mm_struct *mm)

{
@@ -37,11 +55,11 @@ pgd_t *page_pgd_alloc(struct mm_struct *mm) 


		WARN_ON_ONCE(page_to_nid(page) != nid);



		per_node_pgd(mm, nid) = (pgd_t *)page_address(page);

		*per_node_pgd_ptr(mm, nid) = (pgd_t *)page_address(page);

	}



	for_each_online_node(nid)

		per_node_pgd(mm, nid) = per_node_pgd(mm, numa_get_memory_node(nid));

		*per_node_pgd_ptr(mm, nid) = per_node_pgd(mm, numa_get_memory_node(nid));



	mm->pgd = per_node_pgd(mm, numa_get_memory_node(0));/*!!!*/
@@ -53,7 +71,7 @@ pgd_t *page_pgd_alloc(struct mm_struct *mm) 
	return NULL;

}



pgd_t *pgd_alloc(struct mm_struct *mm)

static pgd_t *pgd_alloc_replica(struct mm_struct *mm)

{

	pgd_t **pgd_numa = (pgd_t **)kmalloc(sizeof(pgd_t *) * MAX_NUMNODES, GFP_PGTABLE_KERNEL);
@@ -81,34 +99,44 @@ static void page_pgd_free(struct mm_struct *mm, pgd_t *pgd) 
	}



	for_each_online_node(nid)

		per_node_pgd(mm, nid) = NULL;

		*per_node_pgd_ptr(mm, nid) = NULL;



}



void pgd_free(struct mm_struct *mm, pgd_t *pgd)

static void pgd_free_replica(struct mm_struct *mm, pgd_t *pgd)

{

	page_pgd_free(mm, pgd);



	kfree(mm->pgd_numa);

}



#else /* !CONFIG_KERNEL_REPLICATION */

pgd_t *pgd_alloc(struct mm_struct *mm)

{

	gfp_t gfp = GFP_PGTABLE_USER;



	if (PGD_SIZE == PAGE_SIZE)

		return (pgd_t *)__get_free_page(gfp);

	mm->pgd_numa = NULL;

	if (is_text_replicated())

		return pgd_alloc_replica(mm);

	else

		return kmem_cache_alloc(pgd_cache, gfp);

		return pgd_alloc_orig(mm);



}



void pgd_free(struct mm_struct *mm, pgd_t *pgd)

{

	if (PGD_SIZE == PAGE_SIZE)

		free_page((unsigned long)pgd);

	if (is_text_replicated())

		pgd_free_replica(mm, pgd);

	else

		kmem_cache_free(pgd_cache, pgd);

		pgd_free_orig(mm, pgd);

}



#else /* !CONFIG_KERNEL_REPLICATION */

pgd_t *pgd_alloc(struct mm_struct *mm)

{

	return pgd_alloc_orig(mm);

}



void pgd_free(struct mm_struct *mm, pgd_t *pgd)

{

	pgd_free_orig(mm, pgd);

}

#endif /* CONFIG_KERNEL_REPLICATION */

include/linux/numa_kernel_replication.h

+25 −3
Original line number Diff line number Diff line
@@ -42,8 +42,31 @@ extern nodemask_t replica_nodes; 
	     nid != MAX_NUMNODES;			\

	     nid = next_node(nid, replica_nodes))



#define this_node_pgd(mm) ((mm)->pgd_numa[numa_node_id()])

#define per_node_pgd(mm, nid) ((mm)->pgd_numa[nid])

bool is_text_replicated(void);



static inline pgd_t *this_node_pgd(struct mm_struct *mm)

{

	if (is_text_replicated())

		return mm->pgd_numa[numa_node_id()];

	else

		return mm->pgd;

}



static inline pgd_t *per_node_pgd(struct mm_struct *mm, int nid)

{

	if (is_text_replicated())

		return mm->pgd_numa[nid];

	else

		return mm->pgd;

}



static inline pgd_t **per_node_pgd_ptr(struct mm_struct *mm, int nid)

{

	if (is_text_replicated())

		return &mm->pgd_numa[nid];

	else

		return &mm->pgd;

}



static inline bool numa_addr_has_replica(const void *addr)

{
@@ -56,7 +79,6 @@ void __init numa_replicate_kernel_text(void); 
void numa_replicate_kernel_rodata(void);

void numa_replication_fini(void);



bool is_text_replicated(void);

propagation_level_t get_propagation_level(void);

void numa_setup_pgd(void);

void __init_or_module *numa_get_replica(void *vaddr, int nid);

kernel/module/main.c

+8 −3
Original line number Diff line number Diff line
@@ -1217,6 +1217,9 @@ static void module_replicate_sections(struct module *mod) 
{

	int i;



	if (!is_text_replicated())

		return;



	for (i = 0; i < ARRAY_SIZE(sections_to_replicate); i++)

		module_replicate(mod->mem[sections_to_replicate[i]].base);

}
@@ -1228,10 +1231,12 @@ static void *module_memory_alloc(unsigned int size, enum mod_mem_type type) 
	if (mod_mem_use_vmalloc(type))

		return vzalloc(size);



	if (is_text_replicated()) {

		for (i = 0; i < ARRAY_SIZE(sections_to_replicate); i++) {

			if (type == sections_to_replicate[i])

				return module_alloc_replica(size);

		}

	}

	return module_alloc(size);

}