2017-06-12 01:52:35

by Hoeun Ryu

[permalink] [raw]
Subject: [PATCH] arm:lpae: build TTB control register value from scratch in v7_ttb_setup

Reading TTBCR in early boot stage might return the value of the previous
kernel's configuration, especially in case of kexec. For example, if
normal kernel (first kernel) had run on a configuration of PHYS_OFFSET <=
PAGE_OFFSET and crash kernel (second kernel) is running on a configuration
PHYS_OFFSET > PAGE_OFFSET, which can happen because it depends on the
reserved area for crash kernel, reading TTBCR and using the value to OR
other bit fields might be risky because it doesn't have a reset value for
TTBCR.

Acked-by: Russell King <[email protected]>
Suggested-by: Robin Murphy <[email protected]>
Signed-off-by: Hoeun Ryu <[email protected]>

---

* add Acked-by: Russell King <[email protected]>
* v1: amended based on
- "[PATCHv2] arm: LPAE: kexec: clear TTBCR.T1SZ explicitly when
PHYS_OFFSET > PAGE_OFFSET"
- https://lkml.org/lkml/2017/6/5/239

arch/arm/mm/proc-v7-3level.S | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/arch/arm/mm/proc-v7-3level.S b/arch/arm/mm/proc-v7-3level.S
index 5e5720e..7d16bbc 100644
--- a/arch/arm/mm/proc-v7-3level.S
+++ b/arch/arm/mm/proc-v7-3level.S
@@ -129,8 +129,7 @@ ENDPROC(cpu_v7_set_pte_ext)
.macro v7_ttb_setup, zero, ttbr0l, ttbr0h, ttbr1, tmp
ldr \tmp, =swapper_pg_dir @ swapper_pg_dir virtual address
cmp \ttbr1, \tmp, lsr #12 @ PHYS_OFFSET > PAGE_OFFSET?
- mrc p15, 0, \tmp, c2, c0, 2 @ TTB control egister
- orr \tmp, \tmp, #TTB_EAE
+ mov \tmp, #TTB_EAE @ for TTB control egister
ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP)
ALT_UP(orr \tmp, \tmp, #TTB_FLAGS_UP)
ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP << 16)
--
2.7.4


2017-07-03 08:22:43

by Hoeun Ryu

[permalink] [raw]
Subject: Re: [PATCH] arm:lpae: build TTB control register value from scratch in v7_ttb_setup

Hello, Russell King.

Do you have a plan to include this patch in your tree ?

Thank you.

On Mon, 2017-06-12 at 10:47 +0900, Hoeun Ryu wrote:
>  Reading TTBCR in early boot stage might return the value of the
> previous
> kernel's configuration, especially in case of kexec. For example, if
> normal kernel (first kernel) had run on a configuration of
> PHYS_OFFSET <=
> PAGE_OFFSET and crash kernel (second kernel) is running on a
> configuration
> PHYS_OFFSET > PAGE_OFFSET, which can happen because it depends on the
> reserved area for crash kernel, reading TTBCR and using the value to
> OR
> other bit fields might be risky because it doesn't have a reset value
> for
> TTBCR.
>
> Acked-by: Russell King <[email protected]>
> Suggested-by: Robin Murphy <[email protected]>
> Signed-off-by: Hoeun Ryu <[email protected]>
>
> ---
>
>  * add Acked-by: Russell King <[email protected]>
>  * v1: amended based on
>      - "[PATCHv2] arm: LPAE: kexec: clear TTBCR.T1SZ explicitly when
>         PHYS_OFFSET > PAGE_OFFSET"
>      - https://lkml.org/lkml/2017/6/5/239
>
>  arch/arm/mm/proc-v7-3level.S | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/arch/arm/mm/proc-v7-3level.S b/arch/arm/mm/proc-v7-
> 3level.S
> index 5e5720e..7d16bbc 100644
> --- a/arch/arm/mm/proc-v7-3level.S
> +++ b/arch/arm/mm/proc-v7-3level.S
> @@ -129,8 +129,7 @@ ENDPROC(cpu_v7_set_pte_ext)
>   .macro v7_ttb_setup, zero, ttbr0l, ttbr0h, ttbr1, tmp
>   ldr \tmp, =swapper_pg_dir @
> swapper_pg_dir virtual address
>   cmp \ttbr1, \tmp, lsr #12 @
> PHYS_OFFSET > PAGE_OFFSET?
> - mrc p15, 0, \tmp, c2, c0, 2 @ TTB
> control egister
> - orr \tmp, \tmp, #TTB_EAE
> + mov \tmp, #TTB_EAE @ for TTB
> control egister
>   ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP)
>   ALT_UP(orr \tmp, \tmp, #TTB_FLAGS_UP)
>   ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP << 16)

2017-08-04 06:12:23

by Hoeun Ryu

[permalink] [raw]
Subject: Re: [PATCH] arm:lpae: build TTB control register value from scratch in v7_ttb_setup

Hello, Russell King.

The following patch has not merged yet.
Do you have a plan to accept and merge this patch ?

Thank you.

On Mon, 2017-06-12 at 10:47 +0900, Hoeun Ryu wrote:
>  Reading TTBCR in early boot stage might return the value of the previous
> kernel's configuration, especially in case of kexec. For example, if
> normal kernel (first kernel) had run on a configuration of PHYS_OFFSET <=
> PAGE_OFFSET and crash kernel (second kernel) is running on a configuration
> PHYS_OFFSET > PAGE_OFFSET, which can happen because it depends on the
> reserved area for crash kernel, reading TTBCR and using the value to OR
> other bit fields might be risky because it doesn't have a reset value for
> TTBCR.
>
> Acked-by: Russell King <[email protected]>
> Suggested-by: Robin Murphy <[email protected]>
> Signed-off-by: Hoeun Ryu <[email protected]>
>
> ---
>
>  * add Acked-by: Russell King <[email protected]>
>  * v1: amended based on
>      - "[PATCHv2] arm: LPAE: kexec: clear TTBCR.T1SZ explicitly when
>         PHYS_OFFSET > PAGE_OFFSET"
>      - https://lkml.org/lkml/2017/6/5/239
>
>  arch/arm/mm/proc-v7-3level.S | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/arch/arm/mm/proc-v7-3level.S b/arch/arm/mm/proc-v7-3level.S
> index 5e5720e..7d16bbc 100644
> --- a/arch/arm/mm/proc-v7-3level.S
> +++ b/arch/arm/mm/proc-v7-3level.S
> @@ -129,8 +129,7 @@ ENDPROC(cpu_v7_set_pte_ext)
>   .macro v7_ttb_setup, zero, ttbr0l, ttbr0h, ttbr1, tmp
>   ldr \tmp, =swapper_pg_dir @ swapper_pg_dir virtual address
>   cmp \ttbr1, \tmp, lsr #12 @ PHYS_OFFSET > PAGE_OFFSET?
> - mrc p15, 0, \tmp, c2, c0, 2 @ TTB control egister
> - orr \tmp, \tmp, #TTB_EAE
> + mov \tmp, #TTB_EAE @ for TTB control egister
>   ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP)
>   ALT_UP(orr \tmp, \tmp, #TTB_FLAGS_UP)
>   ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP << 16)

2017-08-04 10:04:59

by Robin Murphy

[permalink] [raw]
Subject: Re: [PATCH] arm:lpae: build TTB control register value from scratch in v7_ttb_setup

On 04/08/17 07:07, Hoeun Ryu wrote:
> Hello, Russell King.
>
> The following patch has not merged yet.
> Do you have a plan to accept and merge this patch ?

This should probably go through the ARM tree, so please submit it to
Russell's patch-tracking system here:

http://www.armlinux.org.uk/developer/patches/

Robin.

>
> Thank you.
>
> On Mon, 2017-06-12 at 10:47 +0900, Hoeun Ryu wrote:
>> Reading TTBCR in early boot stage might return the value of the previous
>> kernel's configuration, especially in case of kexec. For example, if
>> normal kernel (first kernel) had run on a configuration of PHYS_OFFSET <=
>> PAGE_OFFSET and crash kernel (second kernel) is running on a configuration
>> PHYS_OFFSET > PAGE_OFFSET, which can happen because it depends on the
>> reserved area for crash kernel, reading TTBCR and using the value to OR
>> other bit fields might be risky because it doesn't have a reset value for
>> TTBCR.
>>
>> Acked-by: Russell King <[email protected]>
>> Suggested-by: Robin Murphy <[email protected]>
>> Signed-off-by: Hoeun Ryu <[email protected]>
>>
>> ---
>>
>> * add Acked-by: Russell King <[email protected]>
>> * v1: amended based on
>> - "[PATCHv2] arm: LPAE: kexec: clear TTBCR.T1SZ explicitly when
>> PHYS_OFFSET > PAGE_OFFSET"
>> - https://lkml.org/lkml/2017/6/5/239
>>
>> arch/arm/mm/proc-v7-3level.S | 3 +--
>> 1 file changed, 1 insertion(+), 2 deletions(-)
>>
>> diff --git a/arch/arm/mm/proc-v7-3level.S b/arch/arm/mm/proc-v7-3level.S
>> index 5e5720e..7d16bbc 100644
>> --- a/arch/arm/mm/proc-v7-3level.S
>> +++ b/arch/arm/mm/proc-v7-3level.S
>> @@ -129,8 +129,7 @@ ENDPROC(cpu_v7_set_pte_ext)
>> .macro v7_ttb_setup, zero, ttbr0l, ttbr0h, ttbr1, tmp
>> ldr \tmp, =swapper_pg_dir @ swapper_pg_dir virtual address
>> cmp \ttbr1, \tmp, lsr #12 @ PHYS_OFFSET > PAGE_OFFSET?
>> - mrc p15, 0, \tmp, c2, c0, 2 @ TTB control egister
>> - orr \tmp, \tmp, #TTB_EAE
>> + mov \tmp, #TTB_EAE @ for TTB control egister
>> ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP)
>> ALT_UP(orr \tmp, \tmp, #TTB_FLAGS_UP)
>> ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP << 16)

2017-08-04 13:26:36

by Hoeun Ryu

[permalink] [raw]
Subject: Re: [PATCH] arm:lpae: build TTB control register value from scratch in v7_ttb_setup



2017. 8. 4. ???? 7:04 Robin Murphy <[email protected]> ?ۼ?:

>> On 04/08/17 07:07, Hoeun Ryu wrote:
>> Hello, Russell King.
>>
>> The following patch has not merged yet.
>> Do you have a plan to accept and merge this patch ?
>
> This should probably go through the ARM tree, so please submit it to
> Russell's patch-tracking system here:
>
> http://www.armlinux.org.uk/developer/patches/

Thank you for the reply, I'll try it.

>
> Robin.
>
>>
>> Thank you.
>>
>>> On Mon, 2017-06-12 at 10:47 +0900, Hoeun Ryu wrote:
>>> Reading TTBCR in early boot stage might return the value of the previous
>>> kernel's configuration, especially in case of kexec. For example, if
>>> normal kernel (first kernel) had run on a configuration of PHYS_OFFSET <=
>>> PAGE_OFFSET and crash kernel (second kernel) is running on a configuration
>>> PHYS_OFFSET > PAGE_OFFSET, which can happen because it depends on the
>>> reserved area for crash kernel, reading TTBCR and using the value to OR
>>> other bit fields might be risky because it doesn't have a reset value for
>>> TTBCR.
>>>
>>> Acked-by: Russell King <[email protected]>
>>> Suggested-by: Robin Murphy <[email protected]>
>>> Signed-off-by: Hoeun Ryu <[email protected]>
>>>
>>> ---
>>>
>>> * add Acked-by: Russell King <[email protected]>
>>> * v1: amended based on
>>> - "[PATCHv2] arm: LPAE: kexec: clear TTBCR.T1SZ explicitly when
>>> PHYS_OFFSET > PAGE_OFFSET"
>>> - https://lkml.org/lkml/2017/6/5/239
>>>
>>> arch/arm/mm/proc-v7-3level.S | 3 +--
>>> 1 file changed, 1 insertion(+), 2 deletions(-)
>>>
>>> diff --git a/arch/arm/mm/proc-v7-3level.S b/arch/arm/mm/proc-v7-3level.S
>>> index 5e5720e..7d16bbc 100644
>>> --- a/arch/arm/mm/proc-v7-3level.S
>>> +++ b/arch/arm/mm/proc-v7-3level.S
>>> @@ -129,8 +129,7 @@ ENDPROC(cpu_v7_set_pte_ext)
>>> .macro v7_ttb_setup, zero, ttbr0l, ttbr0h, ttbr1, tmp
>>> ldr \tmp, =swapper_pg_dir @ swapper_pg_dir virtual address
>>> cmp \ttbr1, \tmp, lsr #12 @ PHYS_OFFSET > PAGE_OFFSET?
>>> - mrc p15, 0, \tmp, c2, c0, 2 @ TTB control egister
>>> - orr \tmp, \tmp, #TTB_EAE
>>> + mov \tmp, #TTB_EAE @ for TTB control egister
>>> ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP)
>>> ALT_UP(orr \tmp, \tmp, #TTB_FLAGS_UP)
>>> ALT_SMP(orr \tmp, \tmp, #TTB_FLAGS_SMP << 16)
>