All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] ARM, arm64: kvm: get rid of the bounce page
@ 2015-03-27 10:18 Ard Biesheuvel
  2015-03-27 10:18 ` [PATCH] ARM: kvm: round HYP section to page size instead of log2 upper bound Ard Biesheuvel
  0 siblings, 1 reply; 5+ messages in thread
From: Ard Biesheuvel @ 2015-03-27 10:18 UTC (permalink / raw)
  To: linux-arm-kernel

The HYP init bounce page is a runtime construct that ensures that the
HYP init code does not cross a page boundary. However, this is something
we can do perfectly well at build time, by aligning the code appropriately.

For both ARM and arm64, we just align to 4 KB, and enforce that the code
size is less than 4 KB, regardless of the chosen page size.

Note that this also fixes a benign off-by-one error in the original bounce
page code, where a bounce page would be allocated unnecessarily if the code
was exactly 1 page in size.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 arch/arm/kernel/vmlinux.lds.S   | 11 +++++++----
 arch/arm/kvm/mmu.c              | 42 +++++------------------------------------
 arch/arm64/kernel/vmlinux.lds.S | 17 +++++++++++------
 3 files changed, 23 insertions(+), 47 deletions(-)

diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
index b31aa73e8076..f2db429ea75d 100644
--- a/arch/arm/kernel/vmlinux.lds.S
+++ b/arch/arm/kernel/vmlinux.lds.S
@@ -23,7 +23,7 @@
 	VMLINUX_SYMBOL(__idmap_text_start) = .;				\
 	*(.idmap.text)							\
 	VMLINUX_SYMBOL(__idmap_text_end) = .;				\
-	. = ALIGN(32);							\
+	. = ALIGN(PAGE_SIZE);						\
 	VMLINUX_SYMBOL(__hyp_idmap_text_start) = .;			\
 	*(.hyp.idmap.text)						\
 	VMLINUX_SYMBOL(__hyp_idmap_text_end) = .;
@@ -346,8 +346,11 @@ SECTIONS
  */
 ASSERT((__proc_info_end - __proc_info_begin), "missing CPU support")
 ASSERT((__arch_info_end - __arch_info_begin), "no machine record defined")
+
 /*
- * The HYP init code can't be more than a page long.
+ * The HYP init code can't be more than a page long,
+ * and should not cross a page boundary.
  * The above comment applies as well.
  */
-ASSERT(((__hyp_idmap_text_end - __hyp_idmap_text_start) <= PAGE_SIZE), "HYP init code too big")
+ASSERT(__hyp_idmap_text_end - (__hyp_idmap_text_start & PAGE_MASK) <= PAGE_SIZE,
+	"HYP init code too big or misaligned")
diff --git a/arch/arm/kvm/mmu.c b/arch/arm/kvm/mmu.c
index 3e6859bc3e11..42a24d6b003b 100644
--- a/arch/arm/kvm/mmu.c
+++ b/arch/arm/kvm/mmu.c
@@ -37,7 +37,6 @@ static pgd_t *boot_hyp_pgd;
 static pgd_t *hyp_pgd;
 static DEFINE_MUTEX(kvm_hyp_pgd_mutex);
 
-static void *init_bounce_page;
 static unsigned long hyp_idmap_start;
 static unsigned long hyp_idmap_end;
 static phys_addr_t hyp_idmap_vector;
@@ -405,9 +404,6 @@ void free_boot_hyp_pgd(void)
 	if (hyp_pgd)
 		unmap_range(NULL, hyp_pgd, TRAMPOLINE_VA, PAGE_SIZE);
 
-	free_page((unsigned long)init_bounce_page);
-	init_bounce_page = NULL;
-
 	mutex_unlock(&kvm_hyp_pgd_mutex);
 }
 
@@ -1498,39 +1494,11 @@ int kvm_mmu_init(void)
 	hyp_idmap_end = kvm_virt_to_phys(__hyp_idmap_text_end);
 	hyp_idmap_vector = kvm_virt_to_phys(__kvm_hyp_init);
 
-	if ((hyp_idmap_start ^ hyp_idmap_end) & PAGE_MASK) {
-		/*
-		 * Our init code is crossing a page boundary. Allocate
-		 * a bounce page, copy the code over and use that.
-		 */
-		size_t len = __hyp_idmap_text_end - __hyp_idmap_text_start;
-		phys_addr_t phys_base;
-
-		init_bounce_page = (void *)__get_free_page(GFP_KERNEL);
-		if (!init_bounce_page) {
-			kvm_err("Couldn't allocate HYP init bounce page\n");
-			err = -ENOMEM;
-			goto out;
-		}
-
-		memcpy(init_bounce_page, __hyp_idmap_text_start, len);
-		/*
-		 * Warning: the code we just copied to the bounce page
-		 * must be flushed to the point of coherency.
-		 * Otherwise, the data may be sitting in L2, and HYP
-		 * mode won't be able to observe it as it runs with
-		 * caches off at that point.
-		 */
-		kvm_flush_dcache_to_poc(init_bounce_page, len);
-
-		phys_base = kvm_virt_to_phys(init_bounce_page);
-		hyp_idmap_vector += phys_base - hyp_idmap_start;
-		hyp_idmap_start = phys_base;
-		hyp_idmap_end = phys_base + len;
-
-		kvm_info("Using HYP init bounce page @%lx\n",
-			 (unsigned long)phys_base);
-	}
+	/*
+	 * We rely on the linker script to ensure at build time that the HYP
+	 * init code does not cross a page boundary.
+	 */
+	BUG_ON((hyp_idmap_start ^ (hyp_idmap_end - 1)) & PAGE_MASK);
 
 	hyp_pgd = (pgd_t *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, hyp_pgd_order);
 	boot_hyp_pgd = (pgd_t *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, hyp_pgd_order);
diff --git a/arch/arm64/kernel/vmlinux.lds.S b/arch/arm64/kernel/vmlinux.lds.S
index 5d9d2dca530d..a2c29865c3fe 100644
--- a/arch/arm64/kernel/vmlinux.lds.S
+++ b/arch/arm64/kernel/vmlinux.lds.S
@@ -23,10 +23,14 @@ jiffies = jiffies_64;
 
 #define HYPERVISOR_TEXT					\
 	/*						\
-	 * Force the alignment to be compatible with	\
-	 * the vectors requirements			\
+	 * Align to 4 KB so that			\
+	 * a) the HYP vector table is@its minimum	\
+	 *    alignment of 2048 bytes			\
+	 * b) the HYP init code will not cross a page	\
+	 *    boundary if its size does not exceed	\
+	 *    4 KB (see related ASSERT() below)		\
 	 */						\
-	. = ALIGN(2048);				\
+	. = ALIGN(SZ_4K);				\
 	VMLINUX_SYMBOL(__hyp_idmap_text_start) = .;	\
 	*(.hyp.idmap.text)				\
 	VMLINUX_SYMBOL(__hyp_idmap_text_end) = .;	\
@@ -163,10 +167,11 @@ SECTIONS
 }
 
 /*
- * The HYP init code can't be more than a page long.
+ * The HYP init code can't be more than a page long,
+ * and should not cross a page boundary.
  */
-ASSERT(((__hyp_idmap_text_start + PAGE_SIZE) > __hyp_idmap_text_end),
-       "HYP init code too big")
+ASSERT(__hyp_idmap_text_end - (__hyp_idmap_text_start & ~(SZ_4K - 1)) <= SZ_4K,
+	"HYP init code too big or misaligned")
 
 /*
  * If padding is applied before .head.text, virt<->phys conversions will fail.
-- 
1.8.3.2

^ permalink raw reply related	[flat|nested] 5+ messages in thread

* [PATCH] ARM: kvm: round HYP section to page size instead of log2 upper bound
  2015-03-27 10:18 [PATCH] ARM, arm64: kvm: get rid of the bounce page Ard Biesheuvel
@ 2015-03-27 10:18 ` Ard Biesheuvel
  2015-03-27 11:09   ` Will Deacon
  0 siblings, 1 reply; 5+ messages in thread
From: Ard Biesheuvel @ 2015-03-27 10:18 UTC (permalink / raw)
  To: linux-arm-kernel

Older binutils do not support expressions involving the values of
external symbols so just round up the HYP region to the page size.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 arch/arm/kernel/vmlinux.lds.S | 31 +------------------------------
 arch/arm/kvm/init.S           |  3 ---
 2 files changed, 1 insertion(+), 33 deletions(-)

diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
index 808398ec024e..f2db429ea75d 100644
--- a/arch/arm/kernel/vmlinux.lds.S
+++ b/arch/arm/kernel/vmlinux.lds.S
@@ -12,26 +12,6 @@
 #include <asm/pgtable.h>
 #endif
 
-/*
- * Poor man's version of LOG2CEIL(), which is
- * not available in binutils before v2.24.
- */
-#define LOG2_ROUNDUP(size) (		\
-	__LOG2_ROUNDUP(size,  2)	\
-	__LOG2_ROUNDUP(size,  3)	\
-	__LOG2_ROUNDUP(size,  4)	\
-	__LOG2_ROUNDUP(size,  5)	\
-	__LOG2_ROUNDUP(size,  6)	\
-	__LOG2_ROUNDUP(size,  7)	\
-	__LOG2_ROUNDUP(size,  8)	\
-	__LOG2_ROUNDUP(size,  9)	\
-	__LOG2_ROUNDUP(size, 10)	\
-	__LOG2_ROUNDUP(size, 11)	\
-	12)
-
-#define __LOG2_ROUNDUP(size, order)	\
-	(size) <= (1 << order) ? order :
-
 #define PROC_INFO							\
 	. = ALIGN(4);							\
 	VMLINUX_SYMBOL(__proc_info_begin) = .;				\
@@ -43,20 +23,11 @@
 	VMLINUX_SYMBOL(__idmap_text_start) = .;				\
 	*(.idmap.text)							\
 	VMLINUX_SYMBOL(__idmap_text_end) = .;				\
-	. = ALIGN(1 << LOG2_ROUNDUP(__hyp_idmap_size));			\
+	. = ALIGN(PAGE_SIZE);						\
 	VMLINUX_SYMBOL(__hyp_idmap_text_start) = .;			\
 	*(.hyp.idmap.text)						\
 	VMLINUX_SYMBOL(__hyp_idmap_text_end) = .;
 
-/*
- * If the HYP idmap .text section is populated, it needs to be positioned
- * such that it will not cross a page boundary in the final output image.
- * So align it to the section size rounded up to the next power of 2.
- * If __hyp_idmap_size is undefined, the section will be empty so define
- * it as 0 in that case.
- */
-PROVIDE(__hyp_idmap_size = 0);
-
 #ifdef CONFIG_HOTPLUG_CPU
 #define ARM_CPU_DISCARD(x)
 #define ARM_CPU_KEEP(x)		x
diff --git a/arch/arm/kvm/init.S b/arch/arm/kvm/init.S
index 11fb1d56f449..3988e72d16ff 100644
--- a/arch/arm/kvm/init.S
+++ b/arch/arm/kvm/init.S
@@ -157,6 +157,3 @@ target:	@ We're now in the trampoline code, switch page tables
 __kvm_hyp_init_end:
 
 	.popsection
-
-	.global	__hyp_idmap_size
-	.set	__hyp_idmap_size, __kvm_hyp_init_end - __kvm_hyp_init
-- 
1.8.3.2

^ permalink raw reply related	[flat|nested] 5+ messages in thread

* [PATCH] ARM: kvm: round HYP section to page size instead of log2 upper bound
  2015-03-27 10:18 ` [PATCH] ARM: kvm: round HYP section to page size instead of log2 upper bound Ard Biesheuvel
@ 2015-03-27 11:09   ` Will Deacon
  2015-03-27 11:12     ` Ard Biesheuvel
  0 siblings, 1 reply; 5+ messages in thread
From: Will Deacon @ 2015-03-27 11:09 UTC (permalink / raw)
  To: linux-arm-kernel

On Fri, Mar 27, 2015 at 10:18:01AM +0000, Ard Biesheuvel wrote:
> Older binutils do not support expressions involving the values of
> external symbols so just round up the HYP region to the page size.
> 
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

Thanks Ard. I managed to reproduce the failures in -next, and this solves
the problem for me (the only toolchain that chokes seems to be the one on
kernel.org, unfortunately).

I'll push an updated branch after a bit more testing. Let's not touch this
file ever again :)

Will

> ---
>  arch/arm/kernel/vmlinux.lds.S | 31 +------------------------------
>  arch/arm/kvm/init.S           |  3 ---
>  2 files changed, 1 insertion(+), 33 deletions(-)
> 
> diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
> index 808398ec024e..f2db429ea75d 100644
> --- a/arch/arm/kernel/vmlinux.lds.S
> +++ b/arch/arm/kernel/vmlinux.lds.S
> @@ -12,26 +12,6 @@
>  #include <asm/pgtable.h>
>  #endif
>  
> -/*
> - * Poor man's version of LOG2CEIL(), which is
> - * not available in binutils before v2.24.
> - */
> -#define LOG2_ROUNDUP(size) (		\
> -	__LOG2_ROUNDUP(size,  2)	\
> -	__LOG2_ROUNDUP(size,  3)	\
> -	__LOG2_ROUNDUP(size,  4)	\
> -	__LOG2_ROUNDUP(size,  5)	\
> -	__LOG2_ROUNDUP(size,  6)	\
> -	__LOG2_ROUNDUP(size,  7)	\
> -	__LOG2_ROUNDUP(size,  8)	\
> -	__LOG2_ROUNDUP(size,  9)	\
> -	__LOG2_ROUNDUP(size, 10)	\
> -	__LOG2_ROUNDUP(size, 11)	\
> -	12)
> -
> -#define __LOG2_ROUNDUP(size, order)	\
> -	(size) <= (1 << order) ? order :
> -
>  #define PROC_INFO							\
>  	. = ALIGN(4);							\
>  	VMLINUX_SYMBOL(__proc_info_begin) = .;				\
> @@ -43,20 +23,11 @@
>  	VMLINUX_SYMBOL(__idmap_text_start) = .;				\
>  	*(.idmap.text)							\
>  	VMLINUX_SYMBOL(__idmap_text_end) = .;				\
> -	. = ALIGN(1 << LOG2_ROUNDUP(__hyp_idmap_size));			\
> +	. = ALIGN(PAGE_SIZE);						\
>  	VMLINUX_SYMBOL(__hyp_idmap_text_start) = .;			\
>  	*(.hyp.idmap.text)						\
>  	VMLINUX_SYMBOL(__hyp_idmap_text_end) = .;
>  
> -/*
> - * If the HYP idmap .text section is populated, it needs to be positioned
> - * such that it will not cross a page boundary in the final output image.
> - * So align it to the section size rounded up to the next power of 2.
> - * If __hyp_idmap_size is undefined, the section will be empty so define
> - * it as 0 in that case.
> - */
> -PROVIDE(__hyp_idmap_size = 0);
> -
>  #ifdef CONFIG_HOTPLUG_CPU
>  #define ARM_CPU_DISCARD(x)
>  #define ARM_CPU_KEEP(x)		x
> diff --git a/arch/arm/kvm/init.S b/arch/arm/kvm/init.S
> index 11fb1d56f449..3988e72d16ff 100644
> --- a/arch/arm/kvm/init.S
> +++ b/arch/arm/kvm/init.S
> @@ -157,6 +157,3 @@ target:	@ We're now in the trampoline code, switch page tables
>  __kvm_hyp_init_end:
>  
>  	.popsection
> -
> -	.global	__hyp_idmap_size
> -	.set	__hyp_idmap_size, __kvm_hyp_init_end - __kvm_hyp_init
> -- 
> 1.8.3.2
> 

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH] ARM: kvm: round HYP section to page size instead of log2 upper bound
  2015-03-27 11:09   ` Will Deacon
@ 2015-03-27 11:12     ` Ard Biesheuvel
  2015-03-27 12:13       ` Simon Horman
  0 siblings, 1 reply; 5+ messages in thread
From: Ard Biesheuvel @ 2015-03-27 11:12 UTC (permalink / raw)
  To: linux-arm-kernel

On 27 March 2015 at 12:09, Will Deacon <will.deacon@arm.com> wrote:
> On Fri, Mar 27, 2015 at 10:18:01AM +0000, Ard Biesheuvel wrote:
>> Older binutils do not support expressions involving the values of
>> external symbols so just round up the HYP region to the page size.
>>
>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>
> Thanks Ard. I managed to reproduce the failures in -next, and this solves
> the problem for me (the only toolchain that chokes seems to be the one on
> kernel.org, unfortunately).
>
> I'll push an updated branch after a bit more testing. Let's not touch this
> file ever again :)
>

Fingers crossed ...

>> ---
>>  arch/arm/kernel/vmlinux.lds.S | 31 +------------------------------
>>  arch/arm/kvm/init.S           |  3 ---
>>  2 files changed, 1 insertion(+), 33 deletions(-)
>>
>> diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
>> index 808398ec024e..f2db429ea75d 100644
>> --- a/arch/arm/kernel/vmlinux.lds.S
>> +++ b/arch/arm/kernel/vmlinux.lds.S
>> @@ -12,26 +12,6 @@
>>  #include <asm/pgtable.h>
>>  #endif
>>
>> -/*
>> - * Poor man's version of LOG2CEIL(), which is
>> - * not available in binutils before v2.24.
>> - */
>> -#define LOG2_ROUNDUP(size) (         \
>> -     __LOG2_ROUNDUP(size,  2)        \
>> -     __LOG2_ROUNDUP(size,  3)        \
>> -     __LOG2_ROUNDUP(size,  4)        \
>> -     __LOG2_ROUNDUP(size,  5)        \
>> -     __LOG2_ROUNDUP(size,  6)        \
>> -     __LOG2_ROUNDUP(size,  7)        \
>> -     __LOG2_ROUNDUP(size,  8)        \
>> -     __LOG2_ROUNDUP(size,  9)        \
>> -     __LOG2_ROUNDUP(size, 10)        \
>> -     __LOG2_ROUNDUP(size, 11)        \
>> -     12)
>> -
>> -#define __LOG2_ROUNDUP(size, order)  \
>> -     (size) <= (1 << order) ? order :
>> -
>>  #define PROC_INFO                                                    \
>>       . = ALIGN(4);                                                   \
>>       VMLINUX_SYMBOL(__proc_info_begin) = .;                          \
>> @@ -43,20 +23,11 @@
>>       VMLINUX_SYMBOL(__idmap_text_start) = .;                         \
>>       *(.idmap.text)                                                  \
>>       VMLINUX_SYMBOL(__idmap_text_end) = .;                           \
>> -     . = ALIGN(1 << LOG2_ROUNDUP(__hyp_idmap_size));                 \
>> +     . = ALIGN(PAGE_SIZE);                                           \
>>       VMLINUX_SYMBOL(__hyp_idmap_text_start) = .;                     \
>>       *(.hyp.idmap.text)                                              \
>>       VMLINUX_SYMBOL(__hyp_idmap_text_end) = .;
>>
>> -/*
>> - * If the HYP idmap .text section is populated, it needs to be positioned
>> - * such that it will not cross a page boundary in the final output image.
>> - * So align it to the section size rounded up to the next power of 2.
>> - * If __hyp_idmap_size is undefined, the section will be empty so define
>> - * it as 0 in that case.
>> - */
>> -PROVIDE(__hyp_idmap_size = 0);
>> -
>>  #ifdef CONFIG_HOTPLUG_CPU
>>  #define ARM_CPU_DISCARD(x)
>>  #define ARM_CPU_KEEP(x)              x
>> diff --git a/arch/arm/kvm/init.S b/arch/arm/kvm/init.S
>> index 11fb1d56f449..3988e72d16ff 100644
>> --- a/arch/arm/kvm/init.S
>> +++ b/arch/arm/kvm/init.S
>> @@ -157,6 +157,3 @@ target:   @ We're now in the trampoline code, switch page tables
>>  __kvm_hyp_init_end:
>>
>>       .popsection
>> -
>> -     .global __hyp_idmap_size
>> -     .set    __hyp_idmap_size, __kvm_hyp_init_end - __kvm_hyp_init
>> --
>> 1.8.3.2
>>

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH] ARM: kvm: round HYP section to page size instead of log2 upper bound
  2015-03-27 11:12     ` Ard Biesheuvel
@ 2015-03-27 12:13       ` Simon Horman
  0 siblings, 0 replies; 5+ messages in thread
From: Simon Horman @ 2015-03-27 12:13 UTC (permalink / raw)
  To: linux-arm-kernel

Hi Ard, Hi Will,

On Fri, Mar 27, 2015 at 12:12:00PM +0100, Ard Biesheuvel wrote:
> On 27 March 2015 at 12:09, Will Deacon <will.deacon@arm.com> wrote:
> > On Fri, Mar 27, 2015 at 10:18:01AM +0000, Ard Biesheuvel wrote:
> >> Older binutils do not support expressions involving the values of
> >> external symbols so just round up the HYP region to the page size.
> >>
> >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> >
> > Thanks Ard. I managed to reproduce the failures in -next, and this solves
> > the problem for me (the only toolchain that chokes seems to be the one on
> > kernel.org, unfortunately).

For the record that is the one I am using.

> > I'll push an updated branch after a bit more testing. Let's not touch this
> > file ever again :)

:)

> Fingers crossed ...

I have tested this patch applied on top of next-20150326 and the build
problem I observed is gone. I have also tested that the resulting kernel
boots on a range of Renesas boards.

Tested-by: Simon Horman <horms+renesas@verge.net.au>

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2015-03-27 12:13 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-03-27 10:18 [PATCH] ARM, arm64: kvm: get rid of the bounce page Ard Biesheuvel
2015-03-27 10:18 ` [PATCH] ARM: kvm: round HYP section to page size instead of log2 upper bound Ard Biesheuvel
2015-03-27 11:09   ` Will Deacon
2015-03-27 11:12     ` Ard Biesheuvel
2015-03-27 12:13       ` Simon Horman

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.