crash_dump: Don't log dm-crypt key bytes in read_key_from_user_keying

[PATCH] crash_dump: Don't log dm-crypt key bytes in read_key_from_user_keying

Posted by Thorsten Blum 1 month, 2 weeks ago

When debug logging is enabled, read_key_from_user_keying() logs the
first 8 bytes of the key payload and partially exposes the dm-crypt key.
Stop logging any key bytes.

Fixes: 479e58549b0f ("crash_dump: store dm crypt keys in kdump reserved memory")
Cc: stable@vger.kernel.org
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
---
 kernel/crash_dump_dm_crypt.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/crash_dump_dm_crypt.c b/kernel/crash_dump_dm_crypt.c
index 27a144920562..5ce958d069dd 100644
--- a/kernel/crash_dump_dm_crypt.c
+++ b/kernel/crash_dump_dm_crypt.c
@@ -168,8 +168,8 @@ static int read_key_from_user_keying(struct dm_crypt_key *dm_key)
 
 	memcpy(dm_key->data, ukp->data, ukp->datalen);
 	dm_key->key_size = ukp->datalen;
-	kexec_dprintk("Get dm crypt key (size=%u) %s: %8ph\n", dm_key->key_size,
-		      dm_key->key_desc, dm_key->data);
+	kexec_dprintk("Get dm crypt key (size=%u) %s\n", dm_key->key_size,
+		      dm_key->key_desc);
 
 out:
 	up_read(&key->sem);
-- 
Thorsten Blum <thorsten.blum@linux.dev>
GPG: 1D60 735E 8AEF 3BE4 73B6  9D84 7336 78FD 8DFE EAD4

Re: [PATCH] crash_dump: Don't log dm-crypt key bytes in read_key_from_user_keying

Posted by Baoquan He 1 month, 2 weeks ago

On 02/28/26 at 12:00am, Thorsten Blum wrote:
> When debug logging is enabled, read_key_from_user_keying() logs the
> first 8 bytes of the key payload and partially exposes the dm-crypt key.
> Stop logging any key bytes.
> 
> Fixes: 479e58549b0f ("crash_dump: store dm crypt keys in kdump reserved memory")
> Cc: stable@vger.kernel.org
> Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
> ---
>  kernel/crash_dump_dm_crypt.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/crash_dump_dm_crypt.c b/kernel/crash_dump_dm_crypt.c
> index 27a144920562..5ce958d069dd 100644
> --- a/kernel/crash_dump_dm_crypt.c
> +++ b/kernel/crash_dump_dm_crypt.c
> @@ -168,8 +168,8 @@ static int read_key_from_user_keying(struct dm_crypt_key *dm_key)
>  
>  	memcpy(dm_key->data, ukp->data, ukp->datalen);
>  	dm_key->key_size = ukp->datalen;
> -	kexec_dprintk("Get dm crypt key (size=%u) %s: %8ph\n", dm_key->key_size,
> -		      dm_key->key_desc, dm_key->data);
> +	kexec_dprintk("Get dm crypt key (size=%u) %s\n", dm_key->key_size,

Make sense to me.

The kexec_dprintk() is only for debug printing. We can remove above line
or change it to pr_debug() if security is worried.

Coiby, what do you think?

Thanks
Baoquan

Re: [PATCH] crash_dump: Don't log dm-crypt key bytes in read_key_from_user_keying

Posted by Coiby Xu 1 month, 1 week ago

On Mon, Mar 02, 2026 at 11:48:22AM +0800, Baoquan He wrote:
>On 02/28/26 at 12:00am, Thorsten Blum wrote:
>> When debug logging is enabled, read_key_from_user_keying() logs the
>> first 8 bytes of the key payload and partially exposes the dm-crypt key.
>> Stop logging any key bytes.
>>
>> Fixes: 479e58549b0f ("crash_dump: store dm crypt keys in kdump reserved memory")
>> Cc: stable@vger.kernel.org
>> Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
>> ---
>>  kernel/crash_dump_dm_crypt.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/crash_dump_dm_crypt.c b/kernel/crash_dump_dm_crypt.c
>> index 27a144920562..5ce958d069dd 100644
>> --- a/kernel/crash_dump_dm_crypt.c
>> +++ b/kernel/crash_dump_dm_crypt.c
>> @@ -168,8 +168,8 @@ static int read_key_from_user_keying(struct dm_crypt_key *dm_key)
>>
>>  	memcpy(dm_key->data, ukp->data, ukp->datalen);
>>  	dm_key->key_size = ukp->datalen;
>> -	kexec_dprintk("Get dm crypt key (size=%u) %s: %8ph\n", dm_key->key_size,
>> -		      dm_key->key_desc, dm_key->data);
>> +	kexec_dprintk("Get dm crypt key (size=%u) %s\n", dm_key->key_size,
>
>Make sense to me.
>
>The kexec_dprintk() is only for debug printing. We can remove above line
>or change it to pr_debug() if security is worried.
>
>Coiby, what do you think?

I think we can assume a key can be reliably read thus no need to print
the first 8 bytes to check it. So this patch looks good to me. And
thanks to Thorsten for suggesting multiple improvements on
kernel/crash_dump_dm_crypt.c!

>
>Thanks
>Baoquan
>

-- 
Best regards,
Coiby