Re: [PATCH V6 3/8] x86/hyperv: Add new hvcall guest address host visibility support

From: Tianyu Lan <ltykernel@gmail.com>
To: Borislav Petkov <bp@alien8.de>
Cc: kys@microsoft.com, haiyangz@microsoft.com,
	sthemmin@microsoft.com, wei.liu@kernel.org, decui@microsoft.com,
	tglx@linutronix.de, mingo@redhat.com, x86@kernel.org,
	hpa@zytor.com, dave.hansen@linux.intel.com, luto@kernel.org,
	peterz@infradead.org, davem@davemloft.net, kuba@kernel.org,
	gregkh@linuxfoundation.org, arnd@arndb.de, brijesh.singh@amd.com,
	jroedel@suse.de, Tianyu.Lan@microsoft.com,
	thomas.lendacky@amd.com, pgonda@google.com,
	akpm@linux-foundation.org, rppt@kernel.org,
	kirill.shutemov@linux.intel.com, saravanand@fb.com,
	aneesh.kumar@linux.ibm.com, rientjes@google.com, tj@kernel.org,
	michael.h.kelley@microsoft.com, linux-arch@vger.kernel.org,
	linux-hyperv@vger.kernel.org, linux-kernel@vger.kernel.org,
	netdev@vger.kernel.org, vkuznets@redhat.com,
	konrad.wilk@oracle.com, hch@lst.de, robin.murphy@arm.com,
	joro@8bytes.org, parri.andrea@gmail.com, dave.hansen@intel.com
Subject: Re: [PATCH V6 3/8] x86/hyperv: Add new hvcall guest address host visibility support
Date: Fri, 1 Oct 2021 21:17:33 +0800	[thread overview]
Message-ID: <8a9a6753-d0b5-3c12-5a5c-17decae3548b@gmail.com> (raw)
In-Reply-To: <YVX7n4YM8ZirwTQu@zn.tnic>

Hi Boris:
	Thanks for your review.

On 10/1/2021 2:02 AM, Borislav Petkov wrote:
> On Thu, Sep 30, 2021 at 09:05:39AM -0400, Tianyu Lan wrote:
>> @@ -1980,15 +1982,11 @@ int set_memory_global(unsigned long addr, int numpages)
>>   				    __pgprot(_PAGE_GLOBAL), 0);
>>   }
>>   
>> -static int __set_memory_enc_dec(unsigned long addr, int numpages, bool enc)
>> +static int __set_memory_enc_pgtable(unsigned long addr, int numpages, bool enc)
> 
> What exactly is that "pgtable" at the end of the name supposed to mean?
> 

This was suggested by Dave Hansen. It gets used for the hypervisors
that get informed about "encryption" status via page tables: SEV and TDX.

https://lore.kernel.org/linux-iommu/c00e269c-da4c-c703-0182-0221c73a76cc@intel.com/

> So if you want to have different indirections here, I'd suggest you do
> this:
> 
> set_memory_encrypted/decrypted() is the external API. It calls
> 
> _set_memory_enc_dec() which does your hv_* checks. Note the single
> underscore "_" prefix.
> 
> Then, the workhorse remains __set_memory_enc_dec().
> 
> Ok?
> 
> Also, we're reworking the mem_encrypt_active() accessors:
> 
> https://lkml.kernel.org/r/20210928191009.32551-1-bp@alien8.de
> 
> so some synchronization when juggling patchsets will be needed. JFYI
> anyway.

Thanks for reminder. I know that patchset and suggested to decouple 
dependency among SEV, TDX and Hyper=v patchset.

> 
> Also 2, building your set triggers this, dunno if I'm missing some
> patches on my local branch for that.

Thanks for your test. Missing hv_set_register() when CONFIG_HYPERV is 
not selected. I will fix it in the next version.

> 
> In file included from ./arch/x86/include/asm/mshyperv.h:240,
>                   from ./include/clocksource/hyperv_timer.h:18,
>                   from ./arch/x86/include/asm/vdso/gettimeofday.h:21,
>                   from ./include/vdso/datapage.h:137,
>                   from ./arch/x86/include/asm/vgtod.h:12,
>                   from arch/x86/entry/vdso/vma.c:20:
> ./include/asm-generic/mshyperv.h: In function ‘vmbus_signal_eom’:
> ./include/asm-generic/mshyperv.h:153:3: error: implicit declaration of function ‘hv_set_register’; did you mean ‘kset_register’? [-Werror=implicit-function-declaration]
>    153 |   hv_set_register(HV_REGISTER_EOM, 0);
>        |   ^~~~~~~~~~~~~~~
>        |   kset_register
> In file included from ./arch/x86/include/asm/mshyperv.h:240,
>                   from arch/x86/mm/pat/set_memory.c:34:
> ./include/asm-generic/mshyperv.h: In function ‘vmbus_signal_eom’:
> ...
> 