Re: [Patch v3 05/14] x86/mm: Handle decryption/re-encryption of bss_decrypted consistently

[Tom Lendacky <thomas.lendacky@amd.com>]

On 11/16/22 14:35, Tom Lendacky wrote:
> On 11/16/22 12:41, Michael Kelley wrote:
>> Current code in sme_postprocess_startup() decrypts the bss_decrypted
>> section when sme_me_mask is non-zero.  But code in
>> mem_encrypt_free_decrytped_mem() re-encrypts the unused portion based
>> on CC_ATTR_MEM_ENCRYPT.  In a Hyper-V guest VM using vTOM, these
>> conditions are not equivalent as sme_me_mask is always zero when
>> using vTOM.  Consequently, mem_encrypt_free_decrypted_mem() attempts
>> to re-encrypt memory that was never decrypted.
>>
>> Fix this in mem_encrypt_free_decrypted_mem() by conditioning the
>> re-encryption on the same test for non-zero sme_me_mask.  Hyper-V
>> guests using vTOM don't need the bss_decrypted section to be
>> decrypted, so skipping the decryption/re-encryption doesn't cause
>> a problem.
>>
>> Signed-off-by: Michael Kelley <mikelley@microsoft.com>

Meant to add this in the previous reply...

With the change to use sme_me_mask directly

Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com>

>> ---
>>   arch/x86/mm/mem_encrypt_amd.c | 10 +++++++---
>>   1 file changed, 7 insertions(+), 3 deletions(-)
>>
>> diff --git a/arch/x86/mm/mem_encrypt_amd.c b/arch/x86/mm/mem_encrypt_amd.c
>> index 9c4d8db..5a51343 100644
>> --- a/arch/x86/mm/mem_encrypt_amd.c
>> +++ b/arch/x86/mm/mem_encrypt_amd.c
>> @@ -513,10 +513,14 @@ void __init mem_encrypt_free_decrypted_mem(void)
>>       npages = (vaddr_end - vaddr) >> PAGE_SHIFT;
>>       /*
>> -     * The unused memory range was mapped decrypted, change the encryption
>> -     * attribute from decrypted to encrypted before freeing it.
>> +     * If the unused memory range was mapped decrypted, change the 
>> encryption
>> +     * attribute from decrypted to encrypted before freeing it. Base the
>> +     * re-encryption on the same condition used for the decryption in
>> +     * sme_postprocess_startup(). Higher level abstractions, such as
>> +     * CC_ATTR_MEM_ENCRYPT, aren't necessarily equivalent in a Hyper-V VM
>> +     * using vTOM, where sme_me_mask is always zero.
>>        */
>> -    if (cc_platform_has(CC_ATTR_MEM_ENCRYPT)) {
>> +    if (sme_get_me_mask()) {
> 
> To be consistent within this file, you should use sme_me_mask directly.
> 
> Thanks,
> Tom
> 
>>           r = set_memory_encrypted(vaddr, npages);
>>           if (r) {
>>               pr_warn("failed to free unused decrypted pages\n");