Re: [PATCH Part1 RFC v2 11/20] x86/compressed: Add helper for validating pages in the decompression stage

From: Brijesh Singh <brijesh.singh@amd.com>
To: Borislav Petkov <bp@alien8.de>
Cc: brijesh.singh@amd.com, x86@kernel.org,
	linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
	tglx@linutronix.de, jroedel@suse.de, thomas.lendacky@amd.com,
	pbonzini@redhat.com, mingo@redhat.com, dave.hansen@intel.com,
	rientjes@google.com, seanjc@google.com, peterz@infradead.org,
	hpa@zytor.com, tony.luck@intel.com
Subject: Re: [PATCH Part1 RFC v2 11/20] x86/compressed: Add helper for validating pages in the decompression stage
Date: Thu, 20 May 2021 13:05:15 -0500	[thread overview]
Message-ID: <921408d1-a399-7089-8647-f9617eb12919@amd.com> (raw)
In-Reply-To: <YKah5QInPK4+7xaC@zn.tnic>

On 5/20/21 12:52 PM, Borislav Petkov wrote:
> On Fri, Apr 30, 2021 at 07:16:07AM -0500, Brijesh Singh wrote:
>> @@ -278,12 +279,28 @@ static int set_clr_page_flags(struct x86_mapping_info *info,
>>  	if ((set | clr) & _PAGE_ENC)
>>  		clflush_page(address);
>>  
>> +	/*
>> +	 * If the encryption attribute is being cleared, then change the page state to
>> +	 * shared in the RMP entry. Change of the page state must be done before the
>> +	 * PTE updates.
>> +	 */
>> +	if (clr & _PAGE_ENC)
>> +		snp_set_page_shared(pte_pfn(*ptep) << PAGE_SHIFT);
> From the last review:
>
> The statement above already looks at clr - just merge the two together.

Maybe I am missing something, the statement above was executed for
either set or clr but the page shared need to happen only for clr. So,
from code readability point I kept it outside of that if().

Otherwise we may have to do something like.

...

if ((set | clr) & _PAGE_EN) {

   if (clr)

    snp_set_page_shared(pte_pfn(*ptep) << PAGE_SHIFT);

  }

I am okay with above is the preferred approach.

>
>> @@ -136,6 +137,55 @@ static inline bool sev_snp_enabled(void)
>>  	return sev_status_val & MSR_AMD64_SEV_SNP_ENABLED ? true : false;
>>  }
>>  
>> +static void snp_page_state_change(unsigned long paddr, int op)
> From the last review:
>
> no need for too many prefixes on static functions - just call this one
> __change_page_state() or so, so that the below one can be called...

I guess I still kept the "snp" prefix because vmgexit was named that
way. Based on your feedback, I am droping the "SNP" prefix from the
VMGEXIT and will update it as well.

>> +{
>> +	u64 val;
>> +
>> +	if (!sev_snp_enabled())
>> +		return;
>> +
>> +	/*
>> +	 * If the page is getting changed from private to shard then invalidate the page
> 							shared
>
> And you can write this a lot shorter
>
> 	* If private -> shared, ...
>
>> +	 * before requesting the state change in the RMP table.
>> +	 */
>> +	if ((op == SNP_PAGE_STATE_SHARED) && pvalidate(paddr, RMP_PG_SIZE_4K, 0))
>> +		goto e_pvalidate;
>> +
>> +	/* Issue VMGEXIT to change the page state in RMP table. */
>> +	sev_es_wr_ghcb_msr(GHCB_MSR_PSC_REQ_GFN(paddr >> PAGE_SHIFT, op));
>> +	VMGEXIT();
>> +
>> +	/* Read the response of the VMGEXIT. */
>> +	val = sev_es_rd_ghcb_msr();
>> +	if ((GHCB_RESP_CODE(val) != GHCB_MSR_PSC_RESP) || GHCB_MSR_PSC_RESP_VAL(val))
>> +		goto e_psc;
> That label is used only once - just do the termination here directly and
> remove it.

Noted.

>
> Thx.
>