Re: [PATCH v2 05/12] x86/sme: Replace occurrences of sme_active() with prot_guest_has()

[Borislav Petkov <bp@alien8.de>]

On Fri, Aug 13, 2021 at 11:59:24AM -0500, Tom Lendacky wrote:
> diff --git a/arch/x86/mm/mem_encrypt.c b/arch/x86/mm/mem_encrypt.c
> index edc67ddf065d..5635ca9a1fbe 100644
> --- a/arch/x86/mm/mem_encrypt.c
> +++ b/arch/x86/mm/mem_encrypt.c
> @@ -144,7 +144,7 @@ void __init sme_unmap_bootdata(char *real_mode_data)
>  	struct boot_params *boot_data;
>  	unsigned long cmdline_paddr;
>  
> -	if (!sme_active())
> +	if (!amd_prot_guest_has(PATTR_SME))
>  		return;
>  
>  	/* Get the command line address before unmapping the real_mode_data */
> @@ -164,7 +164,7 @@ void __init sme_map_bootdata(char *real_mode_data)
>  	struct boot_params *boot_data;
>  	unsigned long cmdline_paddr;
>  
> -	if (!sme_active())
> +	if (!amd_prot_guest_has(PATTR_SME))
>  		return;
>  
>  	__sme_early_map_unmap_mem(real_mode_data, sizeof(boot_params), true);
> @@ -378,7 +378,7 @@ bool sev_active(void)
>  	return sev_status & MSR_AMD64_SEV_ENABLED;
>  }
>  
> -bool sme_active(void)
> +static bool sme_active(void)

Just get rid of it altogether. Also, there's an

EXPORT_SYMBOL_GPL(sev_active);

which needs to go under the actual function. Here's a diff ontop:

---
diff --git a/arch/x86/mm/mem_encrypt.c b/arch/x86/mm/mem_encrypt.c
index 5635ca9a1fbe..a3a2396362a5 100644
--- a/arch/x86/mm/mem_encrypt.c
+++ b/arch/x86/mm/mem_encrypt.c
@@ -364,8 +364,9 @@ int __init early_set_memory_encrypted(unsigned long vaddr, unsigned long size)
 /*
  * SME and SEV are very similar but they are not the same, so there are
  * times that the kernel will need to distinguish between SME and SEV. The
- * sme_active() and sev_active() functions are used for this.  When a
- * distinction isn't needed, the mem_encrypt_active() function can be used.
+ * PATTR_HOST_MEM_ENCRYPT and PATTR_GUEST_MEM_ENCRYPT flags to
+ * amd_prot_guest_has() are used for this. When a distinction isn't needed,
+ * the mem_encrypt_active() function can be used.
  *
  * The trampoline code is a good example for this requirement.  Before
  * paging is activated, SME will access all memory as decrypted, but SEV
@@ -377,11 +378,6 @@ bool sev_active(void)
 {
 	return sev_status & MSR_AMD64_SEV_ENABLED;
 }
-
-static bool sme_active(void)
-{
-	return sme_me_mask && !sev_active();
-}
 EXPORT_SYMBOL_GPL(sev_active);
 
 /* Needs to be called from non-instrumentable code */
@@ -398,7 +394,7 @@ bool amd_prot_guest_has(unsigned int attr)
 
 	case PATTR_SME:
 	case PATTR_HOST_MEM_ENCRYPT:
-		return sme_active();
+		return sme_me_mask && !sev_active();
 
 	case PATTR_SEV:
 	case PATTR_GUEST_MEM_ENCRYPT:

>  {
>  	return sme_me_mask && !sev_active();
>  }
> @@ -428,7 +428,7 @@ bool force_dma_unencrypted(struct device *dev)
>  	 * device does not support DMA to addresses that include the
>  	 * encryption mask.
>  	 */
> -	if (sme_active()) {
> +	if (amd_prot_guest_has(PATTR_SME)) {

So I'm not sure: you add PATTR_SME which you call with
amd_prot_guest_has() and PATTR_HOST_MEM_ENCRYPT which you call with
prot_guest_has() and they both end up being the same thing on AMD.

So why even bother with PATTR_SME?

This is only going to cause confusion later and I'd say let's simply use
prot_guest_has(PATTR_HOST_MEM_ENCRYPT) everywhere...

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette