linux-kernel - Re: [PATCH] x86/tdx: Fix crash on kexec with CONFIG

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: <fe39a04d-2b3a-476f-9505-b874e3ad93b4@amd.com>
Date: Fri, 30 Aug 2024 15:47:50 -0500
From: "Kalra, Ashish" <ashish.kalra@....com>
To: "Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>
Cc: Tom Lendacky <thomas.lendacky@....com>,
 "Maciej W. Rozycki" <macro@...am.me.uk>, Thomas Gleixner
 <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>,
 Borislav Petkov <bp@...en8.de>, Dave Hansen <dave.hansen@...ux.intel.com>,
 x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>,
 Kevin Loughlin <kevinloughlin@...gle.com>, linux-kernel@...r.kernel.org,
 Rick Edgecombe <rick.p.edgecombe@...el.com>,
 Isaku Yamahata <isaku.yamahata@...el.com>
Subject: Re: [PATCH] x86/tdx: Fix crash on kexec with CONFIG_EISA

Hello Kirill,

On 8/29/2024 7:53 AM, Kirill A. Shutemov wrote:
> On Wed, Aug 28, 2024 at 02:43:23PM -0500, Kalra, Ashish wrote:
>> Hello Kirill,
>>
>> On 8/28/2024 1:21 AM, Kirill A. Shutemov wrote:
>>>>>> memremap()
>>>>>>   arch_memremap_wb()
>>>>>>     ioremap_cache()
>>>>>>       __ioremap_caller(.encrytped = false)
>>>>>>
>>>>>> I think arch_memremap_wb() should be mapped ioremap_encrypted() in x86
>>>>>> case. See the patch below.
>>>>>>
>>>>>> It seems to be working fine on TDX, but I am not sure about SEV.
>>>>>>
>>>>>> Tom, any comments?
>>>>> I haven't dug through the code that thoroughly, but I don't think making
>>>>> arch_memremap_wb() be ioremap_encrypted() will work for SME, where some
>>>>> data, e.g. setup data, is unencrypted and needs to be mapped shared.
>>>>>
>>>>> Let me add @Ashish to the thread and have him investigate this since he
>>>>> has been working on the kexec support under SNP. Can someone provide the
>>>>> specific kernel options that need to be in place?
>>>> As Tom asked for, please provide the specific kernel options to test
>>>> with this configuration.
>>> It is not about testing a specific configuration. The question is if it
>>> safe for memremap() to map all WB memory as encrypted by default.
>>>
>>> Looks like it is safe for TDX, but I am not sure about SME/SEV.
>> For SEV it may make sense, but for SME we don't want memremap() to map
>> all WB memory as encrypted by default.
> Could you elaborate why?
>
> I mean if it is specific memory ranges that can be identified as such we
> could make exception for them.
Looks like that the exception is already made for some of these memory ranges with MEMREMAP_DEC flag set along with MEMREMAP_WB.
>>> Maybe we want a specific flag to make memremap() map WB memory as
>>> decrypted/shared. Make everything encrypted by default seems like a sane
>>> default.
>> What are MEMREMAP_ENC, MEMREMAP_DEC flags being used for currently ?
> Good question.
>
> I see zero use of MEMREMAP_ENC in the kernel. And MEMREMAP_DEC only used
> for setup data which I believe AMD thing.
>
> If it is the only memory that must be decrypted, I guess we can make it
> work with encrypted by default.

Yes, and this looks pretty much covered with:

arch_memremap_can_ram_remap(..)

-> if (CC_ATTR_HOST_MEM_REMAP)

--> memremap_is_setup_data()

---> memremap(.., MEMREMAP_WB | MEMREMAP_DEC);

It does look like that it can work with encrypted by default and any memory ranges which need special handling and need to be setup as decrypted can use MEMREMAP_WB | MEMREMAP_DEC flag.

Thanks, Ashish