[PATCH] mm,ima,kexec: use memblock_free_late from ima_free_kexec_buffer

[PATCH] mm,ima,kexec: use memblock_free_late from ima_free_kexec_buffer

[Rik van Riel]

The code calling ima_free_kexec_buffer runs long after the memblock
allocator has already been torn down, potentially resulting in a use
after free in memblock_isolate_range.

With KASAN or KFENCE, this use after free will result in a BUG
from the idle task, and a subsequent kernel panic.

Switch ima_free_kexec_buffer over to memblock_free_late to avoid
that issue.

Fixes: fee3ff99bc67 ("powerpc: Move arch independent ima kexec functions to drivers/of/kexec.c")
Cc: stable@kernel.org
Signed-off-by: Rik van Riel <riel@surriel.com>
Suggested-by: Mike Rappoport <rppt@kernel.org>
---
 arch/x86/kernel/setup.c | 8 ++------
 drivers/of/kexec.c      | 3 ++-
 2 files changed, 4 insertions(+), 7 deletions(-)

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index fd975a4a5200..aa0df37c1fe7 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -359,15 +359,11 @@ static void __init add_early_ima_buffer(u64 phys_addr)
 #if defined(CONFIG_HAVE_IMA_KEXEC) && !defined(CONFIG_OF_FLATTREE)
 int __init ima_free_kexec_buffer(void)
 {
-	int rc;
-
 	if (!ima_kexec_buffer_size)
 		return -ENOENT;
 
-	rc = memblock_phys_free(ima_kexec_buffer_phys,
-				ima_kexec_buffer_size);
-	if (rc)
-		return rc;
+	memblock_free_late(ima_kexec_buffer_phys,
+			   ima_kexec_buffer_size);
 
 	ima_kexec_buffer_phys = 0;
 	ima_kexec_buffer_size = 0;
diff --git a/drivers/of/kexec.c b/drivers/of/kexec.c
index f26d2ba8a371..68278340cecf 100644
--- a/drivers/of/kexec.c
+++ b/drivers/of/kexec.c
@@ -184,7 +184,8 @@ int __init ima_free_kexec_buffer(void)
 	if (ret)
 		return ret;
 
-	return memblock_phys_free(addr, size);
+	memblock_free_late(addr, size);
+	return 0;
 }
 #endif
 
-- 
2.41.0

Re: [PATCH] mm,ima,kexec: use memblock_free_late from ima_free_kexec_buffer

[Rob Herring]

On Thu, Jul 20, 2023 at 10:14:31AM -0400, Rik van Riel wrote:
> The code calling ima_free_kexec_buffer runs long after the memblock
> allocator has already been torn down, potentially resulting in a use
> after free in memblock_isolate_range.
> 
> With KASAN or KFENCE, this use after free will result in a BUG
> from the idle task, and a subsequent kernel panic.
> 
> Switch ima_free_kexec_buffer over to memblock_free_late to avoid
> that issue.
> 
> Fixes: fee3ff99bc67 ("powerpc: Move arch independent ima kexec functions to drivers/of/kexec.c")

Fixes: b69a2afd5afc ("x86/kexec: Carry forward IMA measurement log on kexec")

Acked-by: Rob Herring <robh@kernel.org>

(I'm assuming someone else is taking this)

> Cc: stable@kernel.org
> Signed-off-by: Rik van Riel <riel@surriel.com>
> Suggested-by: Mike Rappoport <rppt@kernel.org>
> ---
>  arch/x86/kernel/setup.c | 8 ++------
>  drivers/of/kexec.c      | 3 ++-
>  2 files changed, 4 insertions(+), 7 deletions(-)
> 
> diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> index fd975a4a5200..aa0df37c1fe7 100644
> --- a/arch/x86/kernel/setup.c
> +++ b/arch/x86/kernel/setup.c
> @@ -359,15 +359,11 @@ static void __init add_early_ima_buffer(u64 phys_addr)
>  #if defined(CONFIG_HAVE_IMA_KEXEC) && !defined(CONFIG_OF_FLATTREE)
>  int __init ima_free_kexec_buffer(void)
>  {
> -	int rc;
> -
>  	if (!ima_kexec_buffer_size)
>  		return -ENOENT;
>  
> -	rc = memblock_phys_free(ima_kexec_buffer_phys,
> -				ima_kexec_buffer_size);
> -	if (rc)
> -		return rc;
> +	memblock_free_late(ima_kexec_buffer_phys,
> +			   ima_kexec_buffer_size);
>  
>  	ima_kexec_buffer_phys = 0;
>  	ima_kexec_buffer_size = 0;
> diff --git a/drivers/of/kexec.c b/drivers/of/kexec.c
> index f26d2ba8a371..68278340cecf 100644
> --- a/drivers/of/kexec.c
> +++ b/drivers/of/kexec.c
> @@ -184,7 +184,8 @@ int __init ima_free_kexec_buffer(void)
>  	if (ret)
>  		return ret;
>  
> -	return memblock_phys_free(addr, size);
> +	memblock_free_late(addr, size);
> +	return 0;
>  }
>  #endif
>  
> -- 
> 2.41.0
> 

Re: [PATCH] mm,ima,kexec: use memblock_free_late from ima_free_kexec_buffer

[Rik van Riel]

[-- Attachment #1: Type: text/plain, Size: 970 bytes --]

On Fri, 2023-07-21 at 13:38 -0600, Rob Herring wrote:
> On Thu, Jul 20, 2023 at 10:14:31AM -0400, Rik van Riel wrote:
> > The code calling ima_free_kexec_buffer runs long after the memblock
> > allocator has already been torn down, potentially resulting in a
> > use
> > after free in memblock_isolate_range.
> > 
> > With KASAN or KFENCE, this use after free will result in a BUG
> > from the idle task, and a subsequent kernel panic.
> > 
> > Switch ima_free_kexec_buffer over to memblock_free_late to avoid
> > that issue.
> > 
> > Fixes: fee3ff99bc67 ("powerpc: Move arch independent ima kexec
> > functions to drivers/of/kexec.c")
> 
> Fixes: b69a2afd5afc ("x86/kexec: Carry forward IMA measurement log on
> kexec")
> 
Thank you for digging further back in the history of that code.

> Acked-by: Rob Herring <robh@kernel.org>
> 
> (I'm assuming someone else is taking this)

I hope so, but I don't know who...

-- 
All Rights Reversed.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH] mm,ima,kexec: use memblock_free_late from ima_free_kexec_buffer

[Rob Herring]

On Fri, Jul 21, 2023 at 08:40:41PM -0400, Rik van Riel wrote:
> On Fri, 2023-07-21 at 13:38 -0600, Rob Herring wrote:
> > On Thu, Jul 20, 2023 at 10:14:31AM -0400, Rik van Riel wrote:
> > > The code calling ima_free_kexec_buffer runs long after the memblock
> > > allocator has already been torn down, potentially resulting in a
> > > use
> > > after free in memblock_isolate_range.
> > > 
> > > With KASAN or KFENCE, this use after free will result in a BUG
> > > from the idle task, and a subsequent kernel panic.
> > > 
> > > Switch ima_free_kexec_buffer over to memblock_free_late to avoid
> > > that issue.
> > > 
> > > Fixes: fee3ff99bc67 ("powerpc: Move arch independent ima kexec
> > > functions to drivers/of/kexec.c")
> > 
> > Fixes: b69a2afd5afc ("x86/kexec: Carry forward IMA measurement log on
> > kexec")
> > 
> Thank you for digging further back in the history of that code.
> 
> > Acked-by: Rob Herring <robh@kernel.org>
> > 
> > (I'm assuming someone else is taking this)
> 
> I hope so, but I don't know who...

You could split this into 2 patches, then there's no doubt. And each 
Fixes is clear.

Rob