Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Cc:     brijesh.singh@amd.com, x86@kernel.org,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        Tom Lendacky <thomas.lendacky@amd.com>,
        "H. Peter Anvin" <hpa@zytor.com>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Sean Christopherson <sean.j.christopherson@intel.com>,
        =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= <rkrcmar@redhat.com>
Subject: Re: [PATCH v8 1/2] x86/mm: add .bss..decrypted section to hold shared
 variables
To:     Borislav Petkov <bp@suse.de>, Thomas Gleixner <tglx@linutronix.de>
References: <1536875471-17391-1-git-send-email-brijesh.singh@amd.com>
 <1536875471-17391-2-git-send-email-brijesh.singh@amd.com>
 <20180914071056.GA4747@zn.tnic>
 <3517a0db-2f64-6d09-7100-dced40561d08@amd.com>
 <alpine.DEB.2.21.1809141414390.10480@nanos.tec.linutronix.de>
 <20180914141216.GA29898@zn.tnic>
From:   Brijesh Singh <brijesh.singh@amd.com>
Message-ID: <43895f0e-fea7-001e-d727-d962a20f7f6c@amd.com>
Date:   Fri, 14 Sep 2018 09:27:09 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <20180914141216.GA29898@zn.tnic>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Received-SPF: None (protection.outlook.com: amd.com does not designate
 permitted sender hosts)
SpamDiagnosticOutput: 1:99
SpamDiagnosticMetadata: NSPM
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 14 Sep 2018 14:27:13.1846 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 01bec59f-ede0-4403-6d03-08d61a4e2a75
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 3dd8961f-e488-4e60-8e11-a82d994e183d
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BYAPR12MB2680
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org


On 09/14/2018 09:12 AM, Borislav Petkov wrote:
> On Fri, Sep 14, 2018 at 02:17:05PM +0200, Thomas Gleixner wrote:
>>> The sme_encrypt_kernel() does not have access to pmd (after pointer
>>> fixup is applied). You can extend the sme_encrypt_kernel() to pass an
>>> additional arguments but then we start getting in include hell. The pmd
>>> is defined as "pmdval_t". If we extend the sme_encrypt_kernel() then
>>> asm/mem_encrypt.h need to include the header file which defines
>>> "pmdval_t". Adding the 'asm/pgtable_type.h' was causing all kind of
>>> compilation errors. I didn't spend much time on it. IMO, we really don't
>>> need to go in this path unless we see some value from doing this.
>>
>> Keep it here then.
> 
> *For what is worth*, a simple forward declaration works. I've taken the
> 64-bit forward declaration of pmdval_t as SME is 64-bit only anyway.
> 
> The below diff ontop passes the mandatory all*config smoke builds:
> 
> ---
>   arch/x86/include/asm/mem_encrypt.h |  6 ++++--
>   arch/x86/kernel/head64.c           | 18 +-----------------
>   arch/x86/mm/mem_encrypt_identity.c | 18 +++++++++++++++++-
>   3 files changed, 22 insertions(+), 20 deletions(-)
> 
> diff --git a/arch/x86/include/asm/mem_encrypt.h b/arch/x86/include/asm/mem_encrypt.h
> index 616f8e637bc3..67c0e6cfdfb3 100644
> --- a/arch/x86/include/asm/mem_encrypt.h
> +++ b/arch/x86/include/asm/mem_encrypt.h
> @@ -19,6 +19,8 @@
>   
>   #include <asm/bootparam.h>
>   
> +typedef unsigned long pmdval_t;
> +
>   #ifdef CONFIG_AMD_MEM_ENCRYPT
>   
>   extern u64 sme_me_mask;
> @@ -40,7 +42,7 @@ void __init sme_unmap_bootdata(char *real_mode_data);
>   
>   void __init sme_early_init(void);
>   
> -void __init sme_encrypt_kernel(struct boot_params *bp);
> +void __init sme_encrypt_kernel(struct boot_params *bp, pmdval_t *pmd);
>   void __init sme_enable(struct boot_params *bp);
>   
>   int __init early_set_memory_decrypted(unsigned long vaddr, unsigned long size);
> @@ -69,7 +71,7 @@ static inline void __init sme_unmap_bootdata(char *real_mode_data) { }
>   
>   static inline void __init sme_early_init(void) { }
>   
> -static inline void __init sme_encrypt_kernel(struct boot_params *bp) { }
> +static inline void __init sme_encrypt_kernel(struct boot_params *bp, pmdval_t *pmd) { }
>   static inline void __init sme_enable(struct boot_params *bp) { }
>   
>   static inline bool sme_active(void) { return false; }
> diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
> index c16af27eb23f..6f8e9b534e80 100644
> --- a/arch/x86/kernel/head64.c
> +++ b/arch/x86/kernel/head64.c
> @@ -112,7 +112,6 @@ static bool __head check_la57_support(unsigned long physaddr)
>   unsigned long __head __startup_64(unsigned long physaddr,
>   				  struct boot_params *bp)
>   {
> -	unsigned long vaddr, vaddr_end;
>   	unsigned long load_delta, *p;
>   	unsigned long pgtable_flags;
>   	pgdval_t *pgd;
> @@ -233,22 +232,7 @@ unsigned long __head __startup_64(unsigned long physaddr,
>   	*fixup_long(&phys_base, physaddr) += load_delta - sme_get_me_mask();
>   
>   	/* Encrypt the kernel and related (if SME is active) */
> -	sme_encrypt_kernel(bp);
> -
> -	/*
> -	 * Clear the memory encryption mask from the .bss..decrypted section.
> -	 * The bss section will be memset to zero later in the initialization so
> -	 * there is no need to zero it after changing the memory encryption
> -	 * attribute.
> -	 */
> -	if (mem_encrypt_active()) {
> -		vaddr = (unsigned long)__start_bss_decrypted;
> -		vaddr_end = (unsigned long)__end_bss_decrypted;
> -		for (; vaddr < vaddr_end; vaddr += PMD_SIZE) {
> -			i = pmd_index(vaddr);
> -			pmd[i] -= sme_get_me_mask();
> -		}
> -	}
> +	sme_encrypt_kernel(bp, pmd);
>   
>   	/*
>   	 * Return the SME encryption mask (if SME is active) to be used as a
> diff --git a/arch/x86/mm/mem_encrypt_identity.c b/arch/x86/mm/mem_encrypt_identity.c
> index a19ef1a416ff..9dbc145d10f8 100644
> --- a/arch/x86/mm/mem_encrypt_identity.c
> +++ b/arch/x86/mm/mem_encrypt_identity.c
> @@ -267,15 +267,17 @@ static unsigned long __init sme_pgtable_calc(unsigned long len)
>   	return entries + tables;
>   }
>   
> -void __init sme_encrypt_kernel(struct boot_params *bp)
> +void __init sme_encrypt_kernel(struct boot_params *bp, pmdval_t *pmd)
>   {
>   	unsigned long workarea_start, workarea_end, workarea_len;
>   	unsigned long execute_start, execute_end, execute_len;
>   	unsigned long kernel_start, kernel_end, kernel_len;
>   	unsigned long initrd_start, initrd_end, initrd_len;
>   	struct sme_populate_pgd_data ppd;
> +	unsigned long vaddr, vaddr_end;
>   	unsigned long pgtable_area_len;
>   	unsigned long decrypted_base;
> +	int i;
>   
>   	if (!sme_active())
>   		return;
> @@ -467,6 +469,20 @@ void __init sme_encrypt_kernel(struct boot_params *bp)
>   
>   	/* Flush the TLB - no globals so cr3 is enough */
>   	native_write_cr3(__native_read_cr3());
> +
> +	/*
> +	 * Clear the memory encryption mask from the .bss..decrypted section.
> +	 * The bss section will be memset to zero later in the initialization so
> +	 * there is no need to zero it after changing the memory encryption
> +	 * attribute.
> +	 */
> +	vaddr = (unsigned long)__start_bss_decrypted;
> +	vaddr_end = (unsigned long)__end_bss_decrypted;
> +
> +	for (; vaddr < vaddr_end; vaddr += PMD_SIZE) {
> +		i = pmd_index(vaddr);
> +		pmd[i] -= sme_get_me_mask();
> +	}
>   }


The above code will never get executed for the SEV case.

See if (!sme_active()) check in the start of function.

If we decide to go on this patch, then we have to do something like
this:

sme_encrypt_kernel(...)
{
	if (!mem_encrypt_active())
		return;

	if (sev_active())
		goto out;

	/* Do kernel and initrd in-place encrypts for SME only case */
	.....
	.....

out:
	/* Clear the C-bit from .bss..decrypted section */
	...
	...
}


>   
>   void __init sme_enable(struct boot_params *bp)
>