linux-kernel - Re: [RFC PATCH v2 15/20] iommu/amd: AMD IOMMU support for memory encryption

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <27bc5c87-3a74-a1ee-55b1-7f19ec9cd6cc@amd.com>
Date:   Wed, 14 Sep 2016 08:45:44 -0500
From:   Tom Lendacky <thomas.lendacky@....com>
To:     Borislav Petkov <bp@...en8.de>
CC:     <linux-arch@...r.kernel.org>, <linux-efi@...r.kernel.org>,
        <kvm@...r.kernel.org>, <linux-doc@...r.kernel.org>,
        <x86@...nel.org>, <linux-kernel@...r.kernel.org>,
        <kasan-dev@...glegroups.com>, <linux-mm@...ck.org>,
        <iommu@...ts.linux-foundation.org>,
        Radim Krčmář <rkrcmar@...hat.com>,
        Arnd Bergmann <arnd@...db.de>,
        Jonathan Corbet <corbet@....net>,
        Matt Fleming <matt@...eblueprint.co.uk>,
        Joerg Roedel <joro@...tes.org>,
        Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>,
        Andrey Ryabinin <aryabinin@...tuozzo.com>,
        Ingo Molnar <mingo@...hat.com>,
        Andy Lutomirski <luto@...nel.org>,
        "H. Peter Anvin" <hpa@...or.com>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Alexander Potapenko <glider@...gle.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Dmitry Vyukov <dvyukov@...gle.com>
Subject: Re: [RFC PATCH v2 15/20] iommu/amd: AMD IOMMU support for memory
 encryption

On 09/12/2016 06:45 AM, Borislav Petkov wrote:
> On Mon, Aug 22, 2016 at 05:38:20PM -0500, Tom Lendacky wrote:
>> Add support to the AMD IOMMU driver to set the memory encryption mask if
>> memory encryption is enabled.
>>
>> Signed-off-by: Tom Lendacky <thomas.lendacky@....com>
>> ---
>>  arch/x86/include/asm/mem_encrypt.h |    2 ++
>>  arch/x86/mm/mem_encrypt.c          |    5 +++++
>>  drivers/iommu/amd_iommu.c          |   10 ++++++++++
>>  3 files changed, 17 insertions(+)
>>
>> diff --git a/arch/x86/include/asm/mem_encrypt.h b/arch/x86/include/asm/mem_encrypt.h
>> index 384fdfb..e395729 100644
>> --- a/arch/x86/include/asm/mem_encrypt.h
>> +++ b/arch/x86/include/asm/mem_encrypt.h
>> @@ -36,6 +36,8 @@ void __init sme_early_init(void);
>>  /* Architecture __weak replacement functions */
>>  void __init mem_encrypt_init(void);
>>  
>> +unsigned long amd_iommu_get_me_mask(void);
>> +
>>  unsigned long swiotlb_get_me_mask(void);
>>  void swiotlb_set_mem_dec(void *vaddr, unsigned long size);
>>  
>> diff --git a/arch/x86/mm/mem_encrypt.c b/arch/x86/mm/mem_encrypt.c
>> index 6b2e8bf..2f28d87 100644
>> --- a/arch/x86/mm/mem_encrypt.c
>> +++ b/arch/x86/mm/mem_encrypt.c
>> @@ -185,6 +185,11 @@ void __init mem_encrypt_init(void)
>>  	swiotlb_clear_encryption();
>>  }
>>  
>> +unsigned long amd_iommu_get_me_mask(void)
>> +{
>> +	return sme_me_mask;
>> +}
>> +
>>  unsigned long swiotlb_get_me_mask(void)
>>  {
>>  	return sme_me_mask;
>> diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
>> index 96de97a..63995e3 100644
>> --- a/drivers/iommu/amd_iommu.c
>> +++ b/drivers/iommu/amd_iommu.c
>> @@ -166,6 +166,15 @@ struct dma_ops_domain {
>>  static struct iova_domain reserved_iova_ranges;
>>  static struct lock_class_key reserved_rbtree_key;
>>  
>> +/*
>> + * Support for memory encryption. If memory encryption is supported, then an
>> + * override to this function will be provided.
>> + */
>> +unsigned long __weak amd_iommu_get_me_mask(void)
>> +{
>> +	return 0;
>> +}
> 
> So instead of adding a function each time which returns sme_me_mask
> for each user it has, why don't you add a single function which
> returns sme_me_mask in mem_encrypt.c and add an inline in the header
> mem_encrypt.h which returns 0 for the !CONFIG_AMD_MEM_ENCRYPT case.

Currently, mem_encrypt.h only lives in the arch/x86 directory so it
wouldn't be able to be included here without breaking other archs.

> 
> This all is still funny because we access sme_me_mask directly for the
> different KERNEL_* masks but then you're adding an accessor function.

Because this lives outside of the arch/x86 I need to use the weak
function.

> 
> So what you should do instead, IMHO, is either hide sme_me_mask
> altogether and use the accessor functions only (not sure if that would
> work in all cases) or expose sme_me_mask unconditionally and have it be
> 0 if CONFIG_AMD_MEM_ENCRYPT is not enabled so that it just works.
> 
> Or is there a third, more graceful variant?

Is there a better way to do this given the support is only in x86?

Thanks,
Tom

>