mm, vmalloc: use __GFP_HIGHMEM implicitly
Subsystems (23)
RFC PATCH 1/1 mm, vmalloc: use __GFP_HIGHMEM implicitly
From: Michal Hocko <[email protected]>
__vmalloc* allows users to provide gfp flags for the underlying allocation. This API is quite popular
$ git grep "=[[:space:]]__vmalloc\|return[[:space:]]*__vmalloc" | wc -l 77
the only problem is that many people are not aware that they really want to give __GFP_HIGHMEM along with other flags because there is really no reason to consume precious lowmemory on CONFIG_HIGHMEM systems for pages which are mapped to the kernel vmalloc space. About half of users don't use this flag, though. This signals that we make the API unnecessarily too complex.
This patch simply uses __GFP_HIGHMEM implicitly when allocating pages to be mapped to the vmalloc space. Current users which add __GFP_HIGHMEM are simplified and drop the flag.
---
Hi,
this is based on top of [1]. I believe it was Al who has brought this up quite some time ago (or maybe I just misremember). The explicit usage of __GFP_HIGHMEM in __vmalloc* seems to be too much to ask from users. I believe there is no user which doesn't want vmalloc pages be in the highmem but I might be missing something. There is vmalloc_32* API but that uses GFP_DMA* explicitly which overrides __GFP_HIGHMEM. So all current users _should_ be safe to use __GFP_HIGHMEM unconditionally. This patch should simplify things and fix many users which consume lowmem for no good reason.
I am sending this as an RFC to get some feedback, I even haven't compile tested it yet.
Any comments are welcome.
[1] http://lkml.kernel.org/r/[email protected]
arch/parisc/kernel/module.c | 2 +- arch/x86/kernel/module.c | 2 +- drivers/block/drbd/drbd_bitmap.c | 2 +- drivers/gpu/drm/etnaviv/etnaviv_dump.c | 4 ++-- drivers/md/dm-bufio.c | 2 +- fs/btrfs/free-space-tree.c | 3 +-- fs/file.c | 2 +- fs/xfs/kmem.c | 2 +- include/drm/drm_mem_util.h | 3 +-- kernel/bpf/core.c | 9 +++------ kernel/bpf/syscall.c | 3 +-- kernel/fork.c | 2 +- kernel/groups.c | 2 +- kernel/module.c | 2 +- mm/kasan/kasan.c | 2 +- mm/nommu.c | 3 +-- mm/util.c | 2 +- mm/vmalloc.c | 14 +++++++------- net/ceph/ceph_common.c | 2 +- net/netfilter/x_tables.c | 3 +-- 20 files changed, 29 insertions(+), 37 deletions(-)
arch/parisc/kernel/module.c +1 -1
@@ -218,7 +218,7 @@ void *module_alloc(unsigned long size)* easier than trying to map the text, data, init_text and * init_data correctly */ return __vmalloc_node_range (size,1 ,VMALLOC_START ,VMALLOC_END , -GFP_KERNEL | __GFP_HIGHMEM, +GFP_KERNEL ,PAGE_KERNEL_RWX ,0 ,NUMA_NO_NODE ,__builtin_return_address (0 )); }
arch/x86/kernel/module.c +1 -1
@@ -85,7 +85,7 @@ void *module_alloc(unsigned long size) p =__vmalloc_node_range (size,MODULE_ALIGN ,MODULES_VADDR +get_module_load_offset (), -MODULES_END ,GFP_KERNEL | __GFP_HIGHMEM, +MODULES_END ,GFP_KERNEL ,PAGE_KERNEL_EXEC ,0 ,NUMA_NO_NODE ,__builtin_return_address (0 ));if (p && (kasan_module_alloc (p, size) <0 )) {
drivers/block/drbd/drbd_bitmap.c +1 -1
@@ -409,7 +409,7 @@ static struct page **bm_realloc_pages(struct drbd_bitmap *b, unsigned long want) new_pages =kzalloc (bytes,GFP_NOIO | __GFP_NOWARN);if (!new_pages) { new_pages =__vmalloc (bytes, -GFP_NOIO | __GFP_HIGHMEM | __GFP_ZERO, +GFP_NOIO | __GFP_ZERO,PAGE_KERNEL );if (!new_pages)return NULL ;
drivers/gpu/drm/etnaviv/etnaviv_dump.c +2 -2
@@ -160,8 +160,8 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) file_size +=sizeof (*iter.hdr) * n_obj;/* Allocate the file in vmalloc memory, it's likely to be big */ - iter.start =__vmalloc (file_size,GFP_KERNEL | __GFP_HIGHMEM | - __GFP_NOWARN | __GFP_NORETRY,PAGE_KERNEL ); + iter.start =__vmalloc (file_size,GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY, +PAGE_KERNEL );if (!iter.start) {dev_warn (gpu->dev,"failed to allocate devcoredump file\n" );return ;
drivers/md/dm-bufio.c +1 -1
@@ -403,7 +403,7 @@ static void *alloc_buffer_data(struct dm_bufio_client *c, gfp_t gfp_mask,if (gfp_mask & __GFP_NORETRY) noio_flag =memalloc_noio_save (); - ptr =__vmalloc (c->block_size, gfp_mask | __GFP_HIGHMEM,PAGE_KERNEL ); + ptr =__vmalloc (c->block_size, gfp_mask,PAGE_KERNEL );if (gfp_mask & __GFP_NORETRY)memalloc_noio_restore (noio_flag);
fs/btrfs/free-space-tree.c +1 -2
@@ -167,8 +167,7 @@ static u8 *alloc_bitmap(u32 bitmap_size)if (mem)return mem; -return __vmalloc (bitmap_size,GFP_NOFS | __GFP_HIGHMEM | __GFP_ZERO, -PAGE_KERNEL ); +return __vmalloc (bitmap_size,GFP_NOFS | __GFP_ZERO,PAGE_KERNEL ); } int convert_free_space_to_bitmaps(struct btrfs_trans_handle *trans,
fs/file.c +1 -1
@@ -42,7 +42,7 @@ static void *alloc_fdmem(size_t size)if (data !=NULL )return data; } -return __vmalloc (size,GFP_KERNEL_ACCOUNT | __GFP_HIGHMEM,PAGE_KERNEL ); +return __vmalloc (size,GFP_KERNEL_ACCOUNT ,PAGE_KERNEL ); }static void __free_fdtable (struct fdtable *fdt)
fs/xfs/kmem.c +1 -1
@@ -84,7 +84,7 @@ kmem_zalloc_large(size_t size, xfs_km_flags_t flags) noio_flag =memalloc_noio_save (); lflags =kmem_flags_convert (flags); - ptr =__vmalloc (size, lflags | __GFP_HIGHMEM | __GFP_ZERO,PAGE_KERNEL ); + ptr =__vmalloc (size, lflags | __GFP_ZERO,PAGE_KERNEL );if ((current->flags &PF_FSTRANS ) || (flags &KM_NOFS ))memalloc_noio_restore (noio_flag);
include/drm/drm_mem_util.h +1 -2
@@ -37,8 +37,7 @@ static __inline__ void *drm_calloc_large(size_t nmemb, size_t size)if (size * nmemb <=PAGE_SIZE )return kcalloc (nmemb, size,GFP_KERNEL ); -return __vmalloc (size * nmemb, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO,PAGE_KERNEL ); +return __vmalloc (size * nmemb,GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL ); }/* Modeled after cairo's malloc_ab, it's like calloc but without the zeroing. */
kernel/bpf/core.c +3 -6
@@ -73,8 +73,7 @@ void *bpf_internal_load_pointer_neg_helper(const struct sk_buff *skb, int k, unsstruct bpf_prog *bpf_prog_alloc (unsigned int size,gfp_t gfp_extra_flags) { -gfp_t gfp_flags =GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO | - gfp_extra_flags; +gfp_t gfp_flags =GFP_KERNEL | __GFP_ZERO | gfp_extra_flags;struct bpf_prog_aux *aux;struct bpf_prog *fp;
@@ -102,8 +101,7 @@ EXPORT_SYMBOL_GPL(bpf_prog_alloc);struct bpf_prog *bpf_prog_realloc (struct bpf_prog *fp_old,unsigned int size,gfp_t gfp_extra_flags) { -gfp_t gfp_flags =GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO | - gfp_extra_flags; +gfp_t gfp_flags =GFP_KERNEL | __GFP_ZERO | gfp_extra_flags;struct bpf_prog *fp;u32 pages, delta;int ret;
@@ -436,8 +434,7 @@ static int bpf_jit_blind_insn(const struct bpf_insn *from,static struct bpf_prog *bpf_prog_clone_create (struct bpf_prog *fp_other,gfp_t gfp_extra_flags) { -gfp_t gfp_flags =GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO | - gfp_extra_flags; +gfp_t gfp_flags =GFP_KERNEL | __GFP_ZERO | gfp_extra_flags;struct bpf_prog *fp; fp =__vmalloc (fp_other->pages *PAGE_SIZE , gfp_flags,PAGE_KERNEL );
kernel/bpf/syscall.c +1 -2
@@ -67,8 +67,7 @@ void *bpf_map_area_alloc(size_t size)return area; } -return __vmalloc (size,GFP_KERNEL | __GFP_HIGHMEM | flags, -PAGE_KERNEL ); +return __vmalloc (size,GFP_KERNEL | flags,PAGE_KERNEL ); }void bpf_map_area_free (void *area)
kernel/fork.c +1 -1
@@ -192,7 +192,7 @@ static unsigned long *alloc_thread_stack_node(struct task_struct *tsk, int node) stack =__vmalloc_node_range (THREAD_SIZE ,THREAD_SIZE ,VMALLOC_START ,VMALLOC_END , -THREADINFO_GFP | __GFP_HIGHMEM, +THREADINFO_GFP ,PAGE_KERNEL ,0 , node,__builtin_return_address (0 ));
kernel/groups.c +1 -1
@@ -18,7 +18,7 @@ struct group_info *groups_alloc(int gidsetsize) len =sizeof (struct group_info ) +sizeof (kgid_t) * gidsetsize; gi =kmalloc (len,GFP_KERNEL_ACCOUNT |__GFP_NOWARN|__GFP_NORETRY);if (!gi) - gi =__vmalloc (len,GFP_KERNEL_ACCOUNT |__GFP_HIGHMEM,PAGE_KERNEL ); + gi =__vmalloc (len,GFP_KERNEL_ACCOUNT ,PAGE_KERNEL );if (!gi)return NULL ;
kernel/module.c +1 -1
@@ -2848,7 +2848,7 @@ static int copy_module_from_user(const void __user *umod, unsigned long len,/* Suck in entire file: we'll want most of it. */ info->hdr =__vmalloc (info->len, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_NOWARN,PAGE_KERNEL ); +GFP_KERNEL | __GFP_NOWARN,PAGE_KERNEL );if (!info->hdr)return -ENOMEM ;
mm/kasan/kasan.c +1 -1
@@ -679,7 +679,7 @@ int kasan_module_alloc(void *addr, size_t size) ret =__vmalloc_node_range (shadow_size,1 , shadow_start, shadow_start + shadow_size, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, +GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL ,VM_NO_GUARD ,NUMA_NO_NODE ,__builtin_return_address (0 ));
mm/nommu.c +1 -2
@@ -245,8 +245,7 @@ void *vmalloc_user(unsigned long size) {void *ret; - ret =__vmalloc (size,GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, -PAGE_KERNEL ); + ret =__vmalloc (size,GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL );if (ret) {struct vm_area_struct *vma;
mm/util.c +1 -1
@@ -371,7 +371,7 @@ void *kvmalloc_node(size_t size, gfp_t flags, int node)if (ret || size <=PAGE_SIZE )return ret; -return __vmalloc_node_flags (size, node, flags | __GFP_HIGHMEM); +return __vmalloc_node_flags (size, node, flags); }EXPORT_SYMBOL (kvmalloc_node);
mm/vmalloc.c +7 -7
@@ -1619,7 +1619,7 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,struct page **pages;unsigned int nr_pages, array_size, i;const gfp_t nested_gfp = (gfp_mask &GFP_RECLAIM_MASK ) | __GFP_ZERO; -const gfp_t alloc_mask = gfp_mask | __GFP_NOWARN; +const gfp_t alloc_mask = gfp_mask | __GFP_HIGHMEM | __GFP_NOWARN; nr_pages =get_vm_area_size (area) >>PAGE_SHIFT ; array_size = (nr_pages *sizeof (struct page *));
@@ -1783,7 +1783,7 @@ void *__vmalloc_node_flags(unsigned long size,void *vmalloc (unsigned long size) {return __vmalloc_node_flags (size,NUMA_NO_NODE , -GFP_KERNEL | __GFP_HIGHMEM); +GFP_KERNEL ); }EXPORT_SYMBOL (vmalloc);
@@ -1800,7 +1800,7 @@ EXPORT_SYMBOL(vmalloc);void *vzalloc (unsigned long size) {return __vmalloc_node_flags (size,NUMA_NO_NODE , -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO); +GFP_KERNEL | __GFP_ZERO); }EXPORT_SYMBOL (vzalloc);
@@ -1817,7 +1817,7 @@ void *vmalloc_user(unsigned long size)void *ret; ret =__vmalloc_node (size,SHMLBA , -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, +GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL ,NUMA_NO_NODE ,__builtin_return_address (0 ));if (ret) {
@@ -1841,7 +1841,7 @@ EXPORT_SYMBOL(vmalloc_user);*/ void *vmalloc_node (unsigned long size,int node) { -return __vmalloc_node (size,1 ,GFP_KERNEL | __GFP_HIGHMEM,PAGE_KERNEL , +return __vmalloc_node (size,1 ,GFP_KERNEL ,PAGE_KERNEL , node,__builtin_return_address (0 )); }EXPORT_SYMBOL (vmalloc_node);
@@ -1861,7 +1861,7 @@ EXPORT_SYMBOL(vmalloc_node);void *vzalloc_node (unsigned long size,int node) {return __vmalloc_node_flags (size, node, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO); +GFP_KERNEL | __GFP_ZERO); }EXPORT_SYMBOL (vzalloc_node);
@@ -1883,7 +1883,7 @@ EXPORT_SYMBOL(vzalloc_node);void *vmalloc_exec (unsigned long size) { -return __vmalloc_node (size,1 ,GFP_KERNEL | __GFP_HIGHMEM,PAGE_KERNEL_EXEC , +return __vmalloc_node (size,1 ,GFP_KERNEL ,PAGE_KERNEL_EXEC ,NUMA_NO_NODE ,__builtin_return_address (0 )); }
net/ceph/ceph_common.c +1 -1
@@ -187,7 +187,7 @@ void *ceph_kvmalloc(size_t size, gfp_t flags)return ptr; } -return __vmalloc (size, flags | __GFP_HIGHMEM,PAGE_KERNEL ); +return __vmalloc (size, flags,PAGE_KERNEL ); }
net/netfilter/x_tables.c +1 -2
@@ -998,8 +998,7 @@ struct xt_table_info *xt_alloc_table_info(unsigned int size)if (sz <= (PAGE_SIZE <<PAGE_ALLOC_COSTLY_ORDER )) info =kmalloc (sz,GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY);if (!info) { - info =__vmalloc (sz,GFP_KERNEL | __GFP_NOWARN | - __GFP_NORETRY | __GFP_HIGHMEM, + info =__vmalloc (sz,GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY,PAGE_KERNEL );if (!info)return NULL ;
Signature · 7 lines
2.11.0 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Quoted 31 lines · Michal Hocko
From: Michal Hocko <[email protected]>
__vmalloc* allows users to provide gfp flags for the underlying allocation. This API is quite popular
$ git grep "=[[:space:]]__vmalloc\|return[[:space:]]*__vmalloc" | wc -l 77the only problem is that many people are not aware that they really want to give
__GFP_HIGHMEMalong with other flags because there is really no reason to consume precious lowmemory onCONFIG_HIGHMEMsystems for pages which are mapped to the kernel vmalloc space. About half of users don't use this flag, though. This signals that we make the API unnecessarily too complex.This patch simply uses
__GFP_HIGHMEMimplicitly when allocating pages to be mapped to the vmalloc space. Current users which add__GFP_HIGHMEMare simplified and drop the flag.Signed-off-by Michal Hocko <[email protected]>---
Hi,
this is based on top of [1]. I believe it was Al who has brought this up quite some time ago (or maybe I just misremember). The explicit usage of__GFP_HIGHMEMin __vmalloc* seems to be too much to ask from users. I believe there is no user which doesn't want vmalloc pages be in the highmem but I might be missing something. There is vmalloc_32* API but that uses GFP_DMA* explicitly which overrides__GFP_HIGHMEM. So all current users _should_ be safe to use__GFP_HIGHMEMunconditionally. This patch should simplify things and fix many users which consume lowmem for no good reason.
Quoted 2 lines
I am sending this as an RFC to get some feedback, I even haven't compile tested it yet.
Any thoughts, objections?
Quoted 371 lines
Any comments are welcome.
[1] http://lkml.kernel.org/r/[email protected]
arch/parisc/kernel/module.c | 2 +- arch/x86/kernel/module.c | 2 +- drivers/block/drbd/drbd_bitmap.c | 2 +- drivers/gpu/drm/etnaviv/etnaviv_dump.c | 4 ++-- drivers/md/dm-bufio.c | 2 +- fs/btrfs/free-space-tree.c | 3 +-- fs/file.c | 2 +- fs/xfs/kmem.c | 2 +- include/drm/drm_mem_util.h | 3 +-- kernel/bpf/core.c | 9 +++------ kernel/bpf/syscall.c | 3 +-- kernel/fork.c | 2 +- kernel/groups.c | 2 +- kernel/module.c | 2 +- mm/kasan/kasan.c | 2 +- mm/nommu.c | 3 +-- mm/util.c | 2 +- mm/vmalloc.c | 14 +++++++------- net/ceph/ceph_common.c | 2 +- net/netfilter/x_tables.c | 3 +-- 20 files changed, 29 insertions(+), 37 deletions(-)arch/parisc/kernel/module.c +1
-1@@ -218,7 +218,7 @@ void *module_alloc(unsigned long size)* easier than trying to map the text, data, init_text and * init_data correctly */ return __vmalloc_node_range (size,1 ,VMALLOC_START ,VMALLOC_END , -GFP_KERNEL | __GFP_HIGHMEM, +GFP_KERNEL ,PAGE_KERNEL_RWX ,0 ,NUMA_NO_NODE ,__builtin_return_address (0 )); }arch/x86/kernel/module.c +1
-1@@ -85,7 +85,7 @@ void *module_alloc(unsigned long size) p =__vmalloc_node_range (size,MODULE_ALIGN ,MODULES_VADDR +get_module_load_offset (), -MODULES_END ,GFP_KERNEL | __GFP_HIGHMEM, +MODULES_END ,GFP_KERNEL ,PAGE_KERNEL_EXEC ,0 ,NUMA_NO_NODE ,__builtin_return_address (0 ));if (p && (kasan_module_alloc (p, size) <0 )) {drivers/block/drbd/drbd_bitmap.c +1
-1@@ -409,7 +409,7 @@ static struct page **bm_realloc_pages(struct drbd_bitmap *b, unsigned long want) new_pages =kzalloc (bytes,GFP_NOIO | __GFP_NOWARN);if (!new_pages) { new_pages =__vmalloc (bytes, -GFP_NOIO | __GFP_HIGHMEM | __GFP_ZERO, +GFP_NOIO | __GFP_ZERO,PAGE_KERNEL );if (!new_pages)return NULL ;drivers/gpu/drm/etnaviv/etnaviv_dump.c +2
-2@@ -160,8 +160,8 @@ void etnaviv_core_dump(struct etnaviv_gpu *gpu) file_size +=sizeof (*iter.hdr) * n_obj;/* Allocate the file in vmalloc memory, it's likely to be big */ - iter.start =__vmalloc (file_size,GFP_KERNEL | __GFP_HIGHMEM | - __GFP_NOWARN | __GFP_NORETRY,PAGE_KERNEL ); + iter.start =__vmalloc (file_size,GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY, +PAGE_KERNEL );if (!iter.start) {dev_warn (gpu->dev,"failed to allocate devcoredump file\n" );return ;drivers/md/dm-bufio.c +1
-1@@ -403,7 +403,7 @@ static void *alloc_buffer_data(struct dm_bufio_client *c, gfp_t gfp_mask,if (gfp_mask & __GFP_NORETRY) noio_flag =memalloc_noio_save (); - ptr =__vmalloc (c->block_size, gfp_mask | __GFP_HIGHMEM,PAGE_KERNEL ); + ptr =__vmalloc (c->block_size, gfp_mask,PAGE_KERNEL );if (gfp_mask & __GFP_NORETRY)memalloc_noio_restore (noio_flag);fs/btrfs/free-space-tree.c +1
-2@@ -167,8 +167,7 @@ static u8 *alloc_bitmap(u32 bitmap_size)if (mem)return mem; -return __vmalloc (bitmap_size,GFP_NOFS | __GFP_HIGHMEM | __GFP_ZERO, -PAGE_KERNEL ); +return __vmalloc (bitmap_size,GFP_NOFS | __GFP_ZERO,PAGE_KERNEL ); } int convert_free_space_to_bitmaps(struct btrfs_trans_handle *trans,fs/file.c +1
-1@@ -42,7 +42,7 @@ static void *alloc_fdmem(size_t size)if (data !=NULL )return data; } -return __vmalloc (size,GFP_KERNEL_ACCOUNT | __GFP_HIGHMEM,PAGE_KERNEL ); +return __vmalloc (size,GFP_KERNEL_ACCOUNT ,PAGE_KERNEL ); }static void __free_fdtable (struct fdtable *fdt)fs/xfs/kmem.c +1
-1@@ -84,7 +84,7 @@ kmem_zalloc_large(size_t size, xfs_km_flags_t flags) noio_flag =memalloc_noio_save (); lflags =kmem_flags_convert (flags); - ptr =__vmalloc (size, lflags | __GFP_HIGHMEM | __GFP_ZERO,PAGE_KERNEL ); + ptr =__vmalloc (size, lflags | __GFP_ZERO,PAGE_KERNEL );if ((current->flags &PF_FSTRANS ) || (flags &KM_NOFS ))memalloc_noio_restore (noio_flag);include/drm/drm_mem_util.h +1
-2@@ -37,8 +37,7 @@ static __inline__ void *drm_calloc_large(size_t nmemb, size_t size)if (size * nmemb <=PAGE_SIZE )return kcalloc (nmemb, size,GFP_KERNEL ); -return __vmalloc (size * nmemb, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO,PAGE_KERNEL ); +return __vmalloc (size * nmemb,GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL ); }/* Modeled after cairo's malloc_ab, it's like calloc but without the zeroing. */ kernel/bpf/core.c +3
-6@@ -73,8 +73,7 @@ void *bpf_internal_load_pointer_neg_helper(const struct sk_buff *skb, int k, unsstruct bpf_prog *bpf_prog_alloc (unsigned int size,gfp_t gfp_extra_flags) { -gfp_t gfp_flags =GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO | - gfp_extra_flags; +gfp_t gfp_flags =GFP_KERNEL | __GFP_ZERO | gfp_extra_flags;struct bpf_prog_aux *aux;struct bpf_prog *fp;@@ -102,8 +101,7 @@ EXPORT_SYMBOL_GPL(bpf_prog_alloc);struct bpf_prog *bpf_prog_realloc (struct bpf_prog *fp_old,unsigned int size,gfp_t gfp_extra_flags) { -gfp_t gfp_flags =GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO | - gfp_extra_flags; +gfp_t gfp_flags =GFP_KERNEL | __GFP_ZERO | gfp_extra_flags;struct bpf_prog *fp;u32 pages, delta;int ret;@@ -436,8 +434,7 @@ static int bpf_jit_blind_insn(const struct bpf_insn *from,static struct bpf_prog *bpf_prog_clone_create (struct bpf_prog *fp_other,gfp_t gfp_extra_flags) { -gfp_t gfp_flags =GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO | - gfp_extra_flags; +gfp_t gfp_flags =GFP_KERNEL | __GFP_ZERO | gfp_extra_flags;struct bpf_prog *fp; fp =__vmalloc (fp_other->pages *PAGE_SIZE , gfp_flags,PAGE_KERNEL );kernel/bpf/syscall.c +1
-2@@ -67,8 +67,7 @@ void *bpf_map_area_alloc(size_t size)return area; } -return __vmalloc (size,GFP_KERNEL | __GFP_HIGHMEM | flags, -PAGE_KERNEL ); +return __vmalloc (size,GFP_KERNEL | flags,PAGE_KERNEL ); }void bpf_map_area_free (void *area)kernel/fork.c +1
-1@@ -192,7 +192,7 @@ static unsigned long *alloc_thread_stack_node(struct task_struct *tsk, int node) stack =__vmalloc_node_range (THREAD_SIZE ,THREAD_SIZE ,VMALLOC_START ,VMALLOC_END , -THREADINFO_GFP | __GFP_HIGHMEM, +THREADINFO_GFP ,PAGE_KERNEL ,0 , node,__builtin_return_address (0 ));kernel/groups.c +1
-1@@ -18,7 +18,7 @@ struct group_info *groups_alloc(int gidsetsize) len =sizeof (struct group_info ) +sizeof (kgid_t) * gidsetsize; gi =kmalloc (len,GFP_KERNEL_ACCOUNT |__GFP_NOWARN|__GFP_NORETRY);if (!gi) - gi =__vmalloc (len,GFP_KERNEL_ACCOUNT |__GFP_HIGHMEM,PAGE_KERNEL ); + gi =__vmalloc (len,GFP_KERNEL_ACCOUNT ,PAGE_KERNEL );if (!gi)return NULL ;kernel/module.c +1
-1@@ -2848,7 +2848,7 @@ static int copy_module_from_user(const void __user *umod, unsigned long len,/* Suck in entire file: we'll want most of it. */ info->hdr =__vmalloc (info->len, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_NOWARN,PAGE_KERNEL ); +GFP_KERNEL | __GFP_NOWARN,PAGE_KERNEL );if (!info->hdr)return -ENOMEM ;mm/kasan/kasan.c +1
-1@@ -679,7 +679,7 @@ int kasan_module_alloc(void *addr, size_t size) ret =__vmalloc_node_range (shadow_size,1 , shadow_start, shadow_start + shadow_size, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, +GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL ,VM_NO_GUARD ,NUMA_NO_NODE ,__builtin_return_address (0 ));mm/nommu.c +1
-2@@ -245,8 +245,7 @@ void *vmalloc_user(unsigned long size) {void *ret; - ret =__vmalloc (size,GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, -PAGE_KERNEL ); + ret =__vmalloc (size,GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL );if (ret) {struct vm_area_struct *vma;mm/util.c +1
-1@@ -371,7 +371,7 @@ void *kvmalloc_node(size_t size, gfp_t flags, int node)if (ret || size <=PAGE_SIZE )return ret; -return __vmalloc_node_flags (size, node, flags | __GFP_HIGHMEM); +return __vmalloc_node_flags (size, node, flags); }EXPORT_SYMBOL (kvmalloc_node);mm/vmalloc.c +7
-7@@ -1619,7 +1619,7 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,struct page **pages;unsigned int nr_pages, array_size, i;const gfp_t nested_gfp = (gfp_mask &GFP_RECLAIM_MASK ) | __GFP_ZERO; -const gfp_t alloc_mask = gfp_mask | __GFP_NOWARN; +const gfp_t alloc_mask = gfp_mask | __GFP_HIGHMEM | __GFP_NOWARN; nr_pages =get_vm_area_size (area) >>PAGE_SHIFT ; array_size = (nr_pages *sizeof (struct page *));@@ -1783,7 +1783,7 @@ void *__vmalloc_node_flags(unsigned long size,void *vmalloc (unsigned long size) {return __vmalloc_node_flags (size,NUMA_NO_NODE , -GFP_KERNEL | __GFP_HIGHMEM); +GFP_KERNEL ); }EXPORT_SYMBOL (vmalloc);@@ -1800,7 +1800,7 @@ EXPORT_SYMBOL(vmalloc);void *vzalloc (unsigned long size) {return __vmalloc_node_flags (size,NUMA_NO_NODE , -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO); +GFP_KERNEL | __GFP_ZERO); }EXPORT_SYMBOL (vzalloc);@@ -1817,7 +1817,7 @@ void *vmalloc_user(unsigned long size)void *ret; ret =__vmalloc_node (size,SHMLBA , -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, +GFP_KERNEL | __GFP_ZERO,PAGE_KERNEL ,NUMA_NO_NODE ,__builtin_return_address (0 ));if (ret) {@@ -1841,7 +1841,7 @@ EXPORT_SYMBOL(vmalloc_user);*/ void *vmalloc_node (unsigned long size,int node) { -return __vmalloc_node (size,1 ,GFP_KERNEL | __GFP_HIGHMEM,PAGE_KERNEL , +return __vmalloc_node (size,1 ,GFP_KERNEL ,PAGE_KERNEL , node,__builtin_return_address (0 )); }EXPORT_SYMBOL (vmalloc_node);@@ -1861,7 +1861,7 @@ EXPORT_SYMBOL(vmalloc_node);void *vzalloc_node (unsigned long size,int node) {return __vmalloc_node_flags (size, node, -GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO); +GFP_KERNEL | __GFP_ZERO); }EXPORT_SYMBOL (vzalloc_node);@@ -1883,7 +1883,7 @@ EXPORT_SYMBOL(vzalloc_node);void *vmalloc_exec (unsigned long size) { -return __vmalloc_node (size,1 ,GFP_KERNEL | __GFP_HIGHMEM,PAGE_KERNEL_EXEC , +return __vmalloc_node (size,1 ,GFP_KERNEL ,PAGE_KERNEL_EXEC ,NUMA_NO_NODE ,__builtin_return_address (0 )); }net/ceph/ceph_common.c +1
-1@@ -187,7 +187,7 @@ void *ceph_kvmalloc(size_t size, gfp_t flags)return ptr; } -return __vmalloc (size, flags | __GFP_HIGHMEM,PAGE_KERNEL ); +return __vmalloc (size, flags,PAGE_KERNEL ); }net/netfilter/x_tables.c +1
-2@@ -998,8 +998,7 @@ struct xt_table_info *xt_alloc_table_info(unsigned int size)if (sz <= (PAGE_SIZE <<PAGE_ALLOC_COSTLY_ORDER )) info =kmalloc (sz,GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY);if (!info) { - info =__vmalloc (sz,GFP_KERNEL | __GFP_NOWARN | - __GFP_NORETRY | __GFP_HIGHMEM, + info =__vmalloc (sz,GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY,PAGE_KERNEL );if (!info)return NULL ;Signature · 1 line
2.11.0
Signature · 8 lines
Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Quoted 31 lines · Michal Hocko
From: Michal Hocko <[email protected]>
__vmalloc* allows users to provide gfp flags for the underlying allocation. This API is quite popular
$ git grep "=[[:space:]]__vmalloc\|return[[:space:]]*__vmalloc" | wc -l 77the only problem is that many people are not aware that they really want to give
__GFP_HIGHMEMalong with other flags because there is really no reason to consume precious lowmemory onCONFIG_HIGHMEMsystems for pages which are mapped to the kernel vmalloc space. About half of users don't use this flag, though. This signals that we make the API unnecessarily too complex.This patch simply uses
__GFP_HIGHMEMimplicitly when allocating pages to be mapped to the vmalloc space. Current users which add__GFP_HIGHMEMare simplified and drop the flag.Signed-off-by Michal Hocko <[email protected]>---
Hi,
this is based on top of [1]. I believe it was Al who has brought this up quite some time ago (or maybe I just misremember). The explicit usage of__GFP_HIGHMEMin __vmalloc* seems to be too much to ask from users. I believe there is no user which doesn't want vmalloc pages be in the highmem but I might be missing something. There is vmalloc_32* API but that uses GFP_DMA* explicitly which overrides__GFP_HIGHMEM. So all current users _should_ be safe to use__GFP_HIGHMEMunconditionally. This patch should simplify things and fix many users which consume lowmem for no good reason.
Quoted 4 lines
I am sending this as an RFC to get some feedback, I even haven't compile tested it yet.
Any comments are welcome.
The idea sounds good. What are the potential dangers? That somebody of the current callers without __GFP_HIGHMEM would take a physical address of the page and then tried to access it via direct mapping?
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Quoted 32 lines · Vlastimil Babka
Quoted 31 lines · Michal Hocko
From: Michal Hocko <[email protected]>
__vmalloc* allows users to provide gfp flags for the underlying allocation. This API is quite popular
$ git grep "=[[:space:]]__vmalloc\|return[[:space:]]*__vmalloc" | wc -l 77the only problem is that many people are not aware that they really want to give
__GFP_HIGHMEMalong with other flags because there is really no reason to consume precious lowmemory onCONFIG_HIGHMEMsystems for pages which are mapped to the kernel vmalloc space. About half of users don't use this flag, though. This signals that we make the API unnecessarily too complex.This patch simply uses
__GFP_HIGHMEMimplicitly when allocating pages to be mapped to the vmalloc space. Current users which add__GFP_HIGHMEMare simplified and drop the flag.Signed-off-by Michal Hocko <[email protected]>---
Hi,
this is based on top of [1]. I believe it was Al who has brought this up quite some time ago (or maybe I just misremember). The explicit usage of__GFP_HIGHMEMin __vmalloc* seems to be too much to ask from users. I believe there is no user which doesn't want vmalloc pages be in the highmem but I might be missing something. There is vmalloc_32* API but that uses GFP_DMA* explicitly which overrides__GFP_HIGHMEM. So all current users _should_ be safe to use__GFP_HIGHMEMunconditionally. This patch should simplify things and fix many users which consume lowmem for no good reason.
Quoted 8 lines
Quoted 4 lines
I am sending this as an RFC to get some feedback, I even haven't compile tested it yet.
Any comments are welcome.
The idea sounds good. What are the potential dangers? That somebody of the current callers without
__GFP_HIGHMEMwould take a physical address of the page and then tried to access it via direct mapping?
Yes, that wouldn't work but I do not think anybody would want to do something like that. Another risk would be that somebody really wanted to use vmalloc_32* but didn't use the proper API. The physically allocated page would then be used for a device which wouldn't be able to access it because it would be out of its addressable space.
Signature · 8 lines
Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>