Re: [RFC PATCH] /dev/mem: Disable /dev/mem under TDX guest

[Nikolay Borisov <nik.borisov@suse.com>]

On 18.03.25 г. 15:27 ч., Kirill A. Shutemov wrote:
> On Tue, Mar 18, 2025 at 02:53:34PM +0200, Nikolay Borisov wrote:
>>> I think we need to think wider. What about applying a subset of LOCKDOWN_*
>>> in all coco guests by default. Many of them are relevant for the guest security.
>>
>> How do you envision this to work, by introducing another
>> CONFIG_LOCK_DOWN_KERNEL_FORCE_COCO or some such ? Will it be opt-in or
>> mandatory?
> 
> I think cc_platform_has(CC_ATTR_xxx) should enabled some subset of
> LOCKDOWN_*. No need in new config options.

Care to suggest which ones should be included? The way lockdown works at 
the moment is that it only supports 2 levels (check lock_kernel_down() 
and lockdown_is_locked_down()) at which you can lockdown - INTEGRITY_MAX 
and CONFIDENTIALITY_MAX,  where each level includes everything below it. 
So by choosing integrity max you get:

     19         LOCKDOWN_MODULE_SIGNATURE, 

     18         LOCKDOWN_DEV_MEM, 

     17         LOCKDOWN_EFI_TEST, 

     16         LOCKDOWN_KEXEC, 

     15         LOCKDOWN_HIBERNATION, 

     14         LOCKDOWN_PCI_ACCESS, 

     13         LOCKDOWN_IOPORT, 

     12         LOCKDOWN_MSR, 

     11         LOCKDOWN_ACPI_TABLES, 

     10         LOCKDOWN_DEVICE_TREE, 

      9         LOCKDOWN_PCMCIA_CIS, 

      8         LOCKDOWN_TIOCSSERIAL, 

      7         LOCKDOWN_MODULE_PARAMETERS, 

      6         LOCKDOWN_MMIOTRACE, 

      5         LOCKDOWN_DEBUGFS, 

      4         LOCKDOWN_XMON_WR, 

      3         LOCKDOWN_BPF_WRITE_USER, 

      2         LOCKDOWN_DBG_WRITE_KERNEL, 

      1         LOCKDOWN_RTAS_ERROR_INJECTION,

Given this if we for example choose to lockdown the kernel for DEV_MEM, 
we'll also get the MODULE_SIGNATURE lockdown as well. I find this 
somewhat inflexible as we might have to rejuggle the current ordering.

> 
>> Should we decide to follow the lockdown route this means the owner of the
>> coco guest will have the ability to disable it and a misbehaving userspace
>> process will still be able to induce an EPT violation.
> 
> Sure. It can shoot itself in the foot.
>