Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1217579 > unrolled thread

[PATCH v2] x86: fix small LDT allocation for Xen

Back to article view | Back to linux.kernel

Contents

  [PATCH v2] x86: fix small LDT allocation for Xen "Jan Beulich" <JBeulich@suse.com> - 2015-09-02 14:50 +0200
    Re: [PATCH v2] x86: fix small LDT allocation for Xen Andy Lutomirski <luto@amacapital.net> - 2015-09-02 16:10 +0200
      Re: [PATCH v2] x86: fix small LDT allocation for Xen "Jan Beulich" <JBeulich@suse.com> - 2015-09-02 17:50 +0200
        Re: [PATCH v2] x86: fix small LDT allocation for Xen Ingo Molnar <mingo@kernel.org> - 2015-09-02 19:00 +0200

#1217579 — [PATCH v2] x86: fix small LDT allocation for Xen

While commit 37868fe113 ("x86/ldt: Make modify_ldt synchronous") added
a nice comment explaining that Xen needs page-aligned whole page chunks
for guest descriptor tables, it then nevertheless used kzalloc() on the
small size path. As I'm unaware of guarantees for kmalloc(PAGE_SIZE, )
to return page-aligned memory blocks, I believe this needs to be
switched back to __get_free_page().

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
---
v2: Also adjust the freeing side.
---
 arch/x86/kernel/ldt.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

--- 4.2/arch/x86/kernel/ldt.c
+++ 4.2-x86-LDT-alloc/arch/x86/kernel/ldt.c
@@ -58,7 +58,7 @@ static struct ldt_struct *alloc_ldt_stru
 	if (alloc_size > PAGE_SIZE)
 		new_ldt->entries = vzalloc(alloc_size);
 	else
-		new_ldt->entries = kzalloc(PAGE_SIZE, GFP_KERNEL);
+		new_ldt->entries = (void *)__get_free_page(GFP_KERNEL|__GFP_ZERO);
 
 	if (!new_ldt->entries) {
 		kfree(new_ldt);
@@ -95,7 +95,7 @@ static void free_ldt_struct(struct ldt_s
 	if (ldt->size * LDT_ENTRY_SIZE > PAGE_SIZE)
 		vfree(ldt->entries);
 	else
-		kfree(ldt->entries);
+		put_page(virt_to_page(ldt->entries));
 	kfree(ldt);
 }
 



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

[toc] | [next] | [standalone]

#1217622

On Sep 2, 2015 5:46 AM, "Jan Beulich" <JBeulich@suse.com> wrote:
>
> While commit 37868fe113 ("x86/ldt: Make modify_ldt synchronous") added
> a nice comment explaining that Xen needs page-aligned whole page chunks
> for guest descriptor tables, it then nevertheless used kzalloc() on the
> small size path. As I'm unaware of guarantees for kmalloc(PAGE_SIZE, )
> to return page-aligned memory blocks, I believe this needs to be
> switched back to __get_free_page().
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> Cc: Andy Lutomirski <luto@kernel.org>
> Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
> Cc: David Vrabel <david.vrabel@citrix.com>
> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
> ---
> v2: Also adjust the freeing side.
> ---
>  arch/x86/kernel/ldt.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> --- 4.2/arch/x86/kernel/ldt.c
> +++ 4.2-x86-LDT-alloc/arch/x86/kernel/ldt.c
> @@ -58,7 +58,7 @@ static struct ldt_struct *alloc_ldt_stru
>         if (alloc_size > PAGE_SIZE)
>                 new_ldt->entries = vzalloc(alloc_size);
>         else
> -               new_ldt->entries = kzalloc(PAGE_SIZE, GFP_KERNEL);
> +               new_ldt->entries = (void *)__get_free_page(GFP_KERNEL|__GFP_ZERO);
>
>         if (!new_ldt->entries) {
>                 kfree(new_ldt);
> @@ -95,7 +95,7 @@ static void free_ldt_struct(struct ldt_s
>         if (ldt->size * LDT_ENTRY_SIZE > PAGE_SIZE)
>                 vfree(ldt->entries);
>         else
> -               kfree(ldt->entries);
> +               put_page(virt_to_page(ldt->entries));

FWIW, I'm not convinced this is or was correct.  Using free_page looks
a bit safer, and free_page does more than just put_page.

Does anyone here know the right way to do this?

--Andy

>         kfree(ldt);
>  }
>
>
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1217709

>>> On 02.09.15 at 16:08, <luto@amacapital.net> wrote:
> On Sep 2, 2015 5:46 AM, "Jan Beulich" <JBeulich@suse.com> wrote:
>>
>> While commit 37868fe113 ("x86/ldt: Make modify_ldt synchronous") added
>> a nice comment explaining that Xen needs page-aligned whole page chunks
>> for guest descriptor tables, it then nevertheless used kzalloc() on the
>> small size path. As I'm unaware of guarantees for kmalloc(PAGE_SIZE, )
>> to return page-aligned memory blocks, I believe this needs to be
>> switched back to __get_free_page().
>>
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>> Cc: Andy Lutomirski <luto@kernel.org>
>> Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
>> Cc: David Vrabel <david.vrabel@citrix.com>
>> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
>> ---
>> v2: Also adjust the freeing side.
>> ---
>>  arch/x86/kernel/ldt.c |    4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> --- 4.2/arch/x86/kernel/ldt.c
>> +++ 4.2-x86-LDT-alloc/arch/x86/kernel/ldt.c
>> @@ -58,7 +58,7 @@ static struct ldt_struct *alloc_ldt_stru
>>         if (alloc_size > PAGE_SIZE)
>>                 new_ldt->entries = vzalloc(alloc_size);
>>         else
>> -               new_ldt->entries = kzalloc(PAGE_SIZE, GFP_KERNEL);
>> +               new_ldt->entries = (void 
> *)__get_free_page(GFP_KERNEL|__GFP_ZERO);
>>
>>         if (!new_ldt->entries) {
>>                 kfree(new_ldt);
>> @@ -95,7 +95,7 @@ static void free_ldt_struct(struct ldt_s
>>         if (ldt->size * LDT_ENTRY_SIZE > PAGE_SIZE)
>>                 vfree(ldt->entries);
>>         else
>> -               kfree(ldt->entries);
>> +               put_page(virt_to_page(ldt->entries));
> 
> FWIW, I'm not convinced this is or was correct.  Using free_page looks
> a bit safer, and free_page does more than just put_page.

Actually I agree. put_page() is meant to be paired with get_page();
__get_free_pages() is just misleading (i.e. doesn't imply a get_page())
and instead is to be paired with free_pages(). Will do a v3 then.

Jan

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

[toc] | [prev] | [next] | [standalone]

#1217754

* Jan Beulich <JBeulich@suse.com> wrote:

> >> @@ -95,7 +95,7 @@ static void free_ldt_struct(struct ldt_s
> >>         if (ldt->size * LDT_ENTRY_SIZE > PAGE_SIZE)
> >>                 vfree(ldt->entries);
> >>         else
> >> -               kfree(ldt->entries);
> >> +               put_page(virt_to_page(ldt->entries));
> > 
> > FWIW, I'm not convinced this is or was correct.  Using free_page looks
> > a bit safer, and free_page does more than just put_page.
> 
> Actually I agree. put_page() is meant to be paired with get_page();
> __get_free_pages() is just misleading (i.e. doesn't imply a get_page())
> and instead is to be paired with free_pages(). Will do a v3 then.

So put_page() is a page cache primitive and should not be used outside of it.

As it happens, it is safe to put_page() a regularly allocated page as well, but 
it's a little bit slower than doing a free_page().

Thanks,

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

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web