Skip to contentSkip to conversation

fix minor error about efi memory_present

PATCH In review bibo,mao 6 participants 8 messages Oct 23, 2006 – Oct 30, 2006

Subsystems (2)
lore
Share & tools
lore

bibo,mao

author
#0 lore raw

PATCH 1/1 fix minor error about efi memory_present

Hi,
Function efi_memory_present_wrapper parameter start/end is physical address,
but function memory_present parameter is PFN, this patch converts physical address to PFN.

thanks
bibo,mao

arch/i386/kernel/setup.c +0 -0
index 519e63c..141041d 100644
@@ -846,7 +846,7 @@ efi_find_max_pfn(unsigned long start, un

static int __init
efi_memory_present_wrapper(unsigned long start, unsigned long end, void *arg) {
- memory_present(0, start, end);

+	memory_present(0, PFN_UP(start), PFN_DOWN(end));
	return 0;
}
#1 lore raw
Quoted 15 lines · On Mon, 23 Oct 2006 11:42:33 +0800 "bibo,mao" <[email protected]> wrote:

Hi,
Function efi_memory_present_wrapper parameter start/end is physical address,
but function memory_present parameter is PFN, this patch converts physical address to PFN.

thanks
bibo,mao

arch/i386/kernel/setup.c +0 -0
index 519e63c..141041d 100644

arch/i386/kernel/setup.c efi_find_max_pfn(unsigned long start, un · line 852

Quoted 8 lines
arch/i386/kernel/setup.c +1 -1
@@ -846,7 +846,7 @@ efi_find_max_pfn(unsigned long start, un
 static int __init
 efi_memory_present_wrapper(unsigned long start, unsigned long end, void *arg)
 {
-	memory_present(0, start, end);
+	memory_present(0, PFN_UP(start), PFN_DOWN(end));
 	return 0;
 }
 

It doesn't _seem_ like a "minor error". How come people's machines haven't been crashing all over the place?

#2 lore raw
Quoted 13 lines · Andrew Morton
Quoted 10 lines · On Mon, 23 Oct 2006 11:42:33 +0800 "bibo,mao" <[email protected]> wrote:

Hi,
Function efi_memory_present_wrapper parameter start/end is physical address,
but function memory_present parameter is PFN, this patch converts physical address to PFN.

thanks
bibo,mao

Quoted 16 lines

arch/i386/kernel/setup.c efi_find_max_pfn(unsigned long start, un · line 852

Quoted 13 lines
arch/i386/kernel/setup.c +1 -1
index 519e63c..141041d 100644
@@ -846,7 +846,7 @@ efi_find_max_pfn(unsigned long start, un
 static int __init
 efi_memory_present_wrapper(unsigned long start, unsigned long end, void *arg)
 {
-	memory_present(0, start, end);
+	memory_present(0, PFN_UP(start), PFN_DOWN(end));
 	return 0;
 }
 

It doesn't _seem_ like a "minor error". How come people's machines haven't been crashing all over the place?

This patch is only efi bios relative, most machines in the market are legacy bios. System only crashes when booting from EFI bios.

thanks
bibo,mao

Quoted 7 lines

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#3 lore raw

PATCH 1/1 vmalloc : optimization, cleanup, bugfixes

In reply to Andrew Morton

[PATCH] vmalloc : optimization, cleanup, bugfixes

This patch does three things

1) reorder 'struct vm_struct' to speedup lookups on CPUS with small cache lines. The fields 'next,addr,size' should be now in the same cache line, to speedup lookups.

2) One minor cleanup in __get_vm_area_node()

3) Bugfixes in vmalloc_user() and vmalloc_32_user()
NULL returns from __vmalloc() and __find_vm_area() were not tested.

include/linux/vmalloc.h +2 -1
@@ -23,13 +23,14 @@
 #endif
 
 struct vm_struct {
+	/* keep next,addr,size together to speedup lookups */
+	struct vm_struct	*next;
 	void			*addr;
 	unsigned long		size;
 	unsigned long		flags;
 	struct page		**pages;
 	unsigned int		nr_pages;
 	unsigned long		phys_addr;
-	struct vm_struct	*next;
 };
 
 /*
mm/vmalloc.c +17 -15
@@ -180,15 +180,13 @@
 	addr = ALIGN(start, align);
 	size = PAGE_ALIGN(size);
 
+	if (unlikely(!size))
+		return NULL;
+
 	area = kmalloc_node(sizeof(*area), GFP_KERNEL, node);
 	if (unlikely(!area))
 		return NULL;
 
-	if (unlikely(!size)) {
-		kfree (area);
-		return NULL;
-	}
-
 	/*
 	 * We always allocate a guard page.
 	 */
@@ -528,11 +526,13 @@
 	void *ret;
 
 	ret = __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, PAGE_KERNEL);
-	write_lock(&vmlist_lock);
-	area = __find_vm_area(ret);
-	area->flags |= VM_USERMAP;
-	write_unlock(&vmlist_lock);
-
+	if (ret) {
+		write_lock(&vmlist_lock);
+		area = __find_vm_area(ret);
+		if (area)
+			area->flags |= VM_USERMAP;
+		write_unlock(&vmlist_lock);
+	}
 	return ret;
 }
 EXPORT_SYMBOL(vmalloc_user);
@@ -601,11 +601,13 @@
 	void *ret;
 
 	ret = __vmalloc(size, GFP_KERNEL | __GFP_ZERO, PAGE_KERNEL);
-	write_lock(&vmlist_lock);
-	area = __find_vm_area(ret);
-	area->flags |= VM_USERMAP;
-	write_unlock(&vmlist_lock);
-
+	if (ret) {
+		write_lock(&vmlist_lock);
+		area = __find_vm_area(ret);
+		if (area)
+			area->flags |= VM_USERMAP;
+		write_unlock(&vmlist_lock);
+	}
 	return ret;
 }
 EXPORT_SYMBOL(vmalloc_32_user);
#4 lore raw

Re: [PATCH] vmalloc : optimization, cleanup, bugfixes

Quoted 10 lines · Eric Dumazet

[PATCH] vmalloc : optimization, cleanup, bugfixes

This patch does three things

1) reorder 'struct vm_struct' to speedup lookups on CPUS with small cache lines. The fields 'next,addr,size' should be now in the same cache line, to speedup lookups.

2) One minor cleanup in __get_vm_area_node()

3) Bugfixes in vmalloc_user() and vmalloc_32_user()
NULL returns from __vmalloc() and __find_vm_area() were not tested.

Hmm, so they weren't. As far as testing the return of __find_vm_area, you can just turn that into a BUG_ON(!area), because at that point, we've established that the vmalloc succeeded.

Signature · 2 lines
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com 
#5 lore raw

Re: [PATCH] vmalloc : optimization, cleanup, bugfixes

Quoted 7 lines · Nick Piggin
Quoted 2 lines · Eric Dumazet

3) Bugfixes in vmalloc_user() and vmalloc_32_user()
NULL returns from __vmalloc() and __find_vm_area() were not tested.

Hmm, so they weren't. As far as testing the return of __find_vm_area, you can just turn that into a BUG_ON(!area), because at that point, we've established that the vmalloc succeeded.

OK :)

[PATCH] vmalloc : optimization, cleanup, bugfixes

This patch does three things

1) reorder 'struct vm_struct' to speedup lookups on CPUS with small cache lines. The fields 'next,addr,size' should be now in the same cache line, to speedup lookups.

2) One minor cleanup in __get_vm_area_node()

3) Bugfixes in vmalloc_user() and vmalloc_32_user()
NULL returns from __vmalloc() and __find_vm_area() were not tested.

include/linux/vmalloc.h +2 -1
@@ -23,13 +23,14 @@
 #endif
 
 struct vm_struct {
+	/* keep next,addr,size together to speedup lookups */
+	struct vm_struct	*next;
 	void			*addr;
 	unsigned long		size;
 	unsigned long		flags;
 	struct page		**pages;
 	unsigned int		nr_pages;
 	unsigned long		phys_addr;
-	struct vm_struct	*next;
 };
 
 /*
mm/vmalloc.c +17 -15
@@ -180,15 +180,13 @@
 	addr = ALIGN(start, align);
 	size = PAGE_ALIGN(size);
 
+	if (unlikely(!size))
+		return NULL;
+
 	area = kmalloc_node(sizeof(*area), GFP_KERNEL, node);
 	if (unlikely(!area))
 		return NULL;
 
-	if (unlikely(!size)) {
-		kfree (area);
-		return NULL;
-	}
-
 	/*
 	 * We always allocate a guard page.
 	 */
@@ -528,11 +526,13 @@
 	void *ret;
 
 	ret = __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO, PAGE_KERNEL);
-	write_lock(&vmlist_lock);
-	area = __find_vm_area(ret);
-	area->flags |= VM_USERMAP;
-	write_unlock(&vmlist_lock);
-
+	if (ret) {
+		write_lock(&vmlist_lock);
+		area = __find_vm_area(ret);
+		BUG_ON(!area);
+		area->flags |= VM_USERMAP;
+		write_unlock(&vmlist_lock);
+	}
 	return ret;
 }
 EXPORT_SYMBOL(vmalloc_user);
@@ -601,11 +601,13 @@
 	void *ret;
 
 	ret = __vmalloc(size, GFP_KERNEL | __GFP_ZERO, PAGE_KERNEL);
-	write_lock(&vmlist_lock);
-	area = __find_vm_area(ret);
-	area->flags |= VM_USERMAP;
-	write_unlock(&vmlist_lock);
-
+	if (ret) {
+		write_lock(&vmlist_lock);
+		area = __find_vm_area(ret);
+		BUG_ON(!area);
+		area->flags |= VM_USERMAP;
+		write_unlock(&vmlist_lock);
+	}
 	return ret;
 }
 EXPORT_SYMBOL(vmalloc_32_user);
#6 lore raw

Re: [PATCH] vmalloc : optimization, cleanup, bugfixes

In reply to Nick Piggin
Quoted 9 lines · Nick Piggin
Quoted 8 lines · Eric Dumazet

[PATCH] vmalloc : optimization, cleanup, bugfixes

This patch does three things

1) reorder 'struct vm_struct' to speedup lookups on CPUS with small cache lines. The fields 'next,addr,size' should be now in the same cache line, to speedup lookups.

Quoted 8 lines
Quoted 4 lines

2) One minor cleanup in __get_vm_area_node()

3) Bugfixes in vmalloc_user() and vmalloc_32_user()
NULL returns from __vmalloc() and __find_vm_area() were not tested.

Hmm, so they weren't. As far as testing the return of __find_vm_area, you can just turn that into a BUG_ON(!area), because at that point, we've established that the vmalloc succeeded.

No need for a BUG_ON it'll simply be a NULL dereference, at which point we're back to the original code.

Zwane

#7 lore raw

Re: [PATCH] vmalloc : optimization, cleanup, bugfixes

Quoted 13 lines · Zwane Mwaikambo
Quoted 11 lines · Nick Piggin
Quoted 10 lines · Eric Dumazet

[PATCH] vmalloc : optimization, cleanup, bugfixes

This patch does three things

1) reorder 'struct vm_struct' to speedup lookups on CPUS with small cache lines. The fields 'next,addr,size' should be now in the same cache line, to speedup lookups.

2) One minor cleanup in __get_vm_area_node()

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

3) Bugfixes in vmalloc_user() and vmalloc_32_user()
NULL returns from __vmalloc() and __find_vm_area() were not tested.

Hmm, so they weren't. As far as testing the return of __find_vm_area, you can just turn that into a BUG_ON(!area), because at that point, we've established that the vmalloc succeeded.

No need for a BUG_ON it'll simply be a NULL dereference, at which point we're back to the original code.

Indeed

This is what Andrew said one week ago :)