Currently, the __is_lm_address() check just masks out the top 12 bits
of the address, but if they are 0, it still yields a true result.
This has as a side effect that virt_addr_valid() returns true even for
invalid virtual addresses (e.g. 0x0).
Fix the detection checking that it's actually a kernel address starting
at PAGE_OFFSET.
Fixes: f4693c2716b35 ("arm64: mm: extend linear region for 52-bit VA configurations")
Cc: <[email protected]> # 5.4.x
Cc: Catalin Marinas <[email protected]>
Cc: Will Deacon <[email protected]>
Suggested-by: Catalin Marinas <[email protected]>
Reviewed-by: Catalin Marinas <[email protected]>
Acked-by: Mark Rutland <[email protected]>
Signed-off-by: Vincenzo Frascino <[email protected]>
---
arch/arm64/include/asm/memory.h | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h
index 18fce223b67b..99d7e1494aaa 100644
--- a/arch/arm64/include/asm/memory.h
+++ b/arch/arm64/include/asm/memory.h
@@ -247,9 +247,11 @@ static inline const void *__tag_set(const void *addr, u8 tag)
/*
- * The linear kernel range starts at the bottom of the virtual address space.
+ * Check whether an arbitrary address is within the linear map, which
+ * lives in the [PAGE_OFFSET, PAGE_END) interval at the bottom of the
+ * kernel's TTBR1 address range.
*/
-#define __is_lm_address(addr) (((u64)(addr) & ~PAGE_OFFSET) < (PAGE_END - PAGE_OFFSET))
+#define __is_lm_address(addr) (((u64)(addr) ^ PAGE_OFFSET) < (PAGE_END - PAGE_OFFSET))
#define __lm_to_phys(addr) (((addr) & ~PAGE_OFFSET) + PHYS_OFFSET)
#define __kimg_to_phys(addr) ((addr) - kimage_voffset)
--
2.30.0
On Tue, Jan 26, 2021 at 01:40:56PM +0000, Vincenzo Frascino wrote:
> Currently, the __is_lm_address() check just masks out the top 12 bits
> of the address, but if they are 0, it still yields a true result.
> This has as a side effect that virt_addr_valid() returns true even for
> invalid virtual addresses (e.g. 0x0).
>
> Fix the detection checking that it's actually a kernel address starting
> at PAGE_OFFSET.
>
> Fixes: f4693c2716b35 ("arm64: mm: extend linear region for 52-bit VA configurations")
> Cc: <[email protected]> # 5.4.x
Not sure what happened with the Fixes tag but that's definitely not what
it fixes. The above is a 5.11 commit that preserves the semantics of an
older commit. So it should be:
Fixes: 68dd8ef32162 ("arm64: memory: Fix virt_addr_valid() using __is_lm_address()")
The above also had a fix for another commit but no need to add two
entries, we just fix the original fix: 14c127c957c1 ("arm64: mm: Flip
kernel VA space").
Anyway, no need to repost, I can update the fixes tag myself.
In terms of stable backports, it may be cleaner to backport 7bc1a0f9e176
("arm64: mm: use single quantity to represent the PA to VA translation")
which has a Fixes tag already but never made it to -stable. On top of
this, we can backport Ard's latest f4693c2716b35 ("arm64: mm: extend
linear region for 52-bit VA configurations"). I just tried these locally
and the conflicts were fairly trivial.
--
Catalin
On 1/26/21 4:36 PM, Catalin Marinas wrote:
> On Tue, Jan 26, 2021 at 01:40:56PM +0000, Vincenzo Frascino wrote:
>> Currently, the __is_lm_address() check just masks out the top 12 bits
>> of the address, but if they are 0, it still yields a true result.
>> This has as a side effect that virt_addr_valid() returns true even for
>> invalid virtual addresses (e.g. 0x0).
>>
>> Fix the detection checking that it's actually a kernel address starting
>> at PAGE_OFFSET.
>>
>> Fixes: f4693c2716b35 ("arm64: mm: extend linear region for 52-bit VA configurations")
>> Cc: <[email protected]> # 5.4.x
>
> Not sure what happened with the Fixes tag but that's definitely not what
> it fixes. The above is a 5.11 commit that preserves the semantics of an
> older commit. So it should be:
>
> Fixes: 68dd8ef32162 ("arm64: memory: Fix virt_addr_valid() using __is_lm_address()")
>
Yes that is correct. I moved the release to which applies backword but I forgot
to update the fixes tag I suppose.
...
>
> Anyway, no need to repost, I can update the fixes tag myself.
>
Thank you for this.
> In terms of stable backports, it may be cleaner to backport 7bc1a0f9e176
> ("arm64: mm: use single quantity to represent the PA to VA translation")
> which has a Fixes tag already but never made it to -stable. On top of
> this, we can backport Ard's latest f4693c2716b35 ("arm64: mm: extend
> linear region for 52-bit VA configurations"). I just tried these locally
> and the conflicts were fairly trivial.
>
Ok, thank you for digging it. I will give it a try tomorrow.
--
Regards,
Vincenzo
On Tue, 26 Jan 2021 13:40:56 +0000, Vincenzo Frascino wrote:
> Currently, the __is_lm_address() check just masks out the top 12 bits
> of the address, but if they are 0, it still yields a true result.
> This has as a side effect that virt_addr_valid() returns true even for
> invalid virtual addresses (e.g. 0x0).
>
> Fix the detection checking that it's actually a kernel address starting
> at PAGE_OFFSET.
Applied to arm64 (for-next/fixes), thanks!
[1/1] arm64: Fix kernel address detection of __is_lm_address()
https://git.kernel.org/arm64/c/519ea6f1c82f
--
Catalin