linux-kernel - Re: [PATCH 3/3] [RFC] x86/devmem: remove low 1MB hack for x86-64

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: <0dbb52fb-9e37-4b3e-a247-5946f08b846f@app.fastmail.com>
Date: Tue, 03 Jun 2025 21:25:55 +0200
From: "Arnd Bergmann" <arnd@...db.de>
To: "Dan Williams" <dan.j.williams@...el.com>,
 "Arnd Bergmann" <arnd@...nel.org>,
 "Dave Hansen" <dave.hansen@...ux.intel.com>
Cc: "Andy Lutomirski" <luto@...nel.org>,
 "Peter Zijlstra" <peterz@...radead.org>,
 "Thomas Gleixner" <tglx@...utronix.de>, "Ingo Molnar" <mingo@...hat.com>,
 "Borislav Petkov" <bp@...en8.de>, x86@...nel.org,
 "Greg Kroah-Hartman" <gregkh@...uxfoundation.org>,
 "Nikolay Borisov" <nik.borisov@...e.com>, linux-kernel@...r.kernel.org,
 "Kees Cook" <kees@...nel.org>, "Naveen N Rao" <naveen@...nel.org>
Subject: Re: [PATCH 3/3] [RFC] x86/devmem: remove low 1MB hack for x86-64

On Tue, Jun 3, 2025, at 20:18, Dan Williams wrote:
> [add Naveen]
>
> Arnd Bergmann wrote:
>> On Thu, May 22, 2025, at 00:14, Dan Williams wrote:
>> > Arnd Bergmann wrote:
>>
>> The third one maps the BIOS area at 0xf0000, and as far as I can tell
>> the hack explicitly allowed mapping that even though it is marked
>> busy on x86-64 since 5d94e81f69d4 ("x86: Introduce pci_map_biosrom()").
>> 
>> Is there any downside to marking this one non-busy and still allowing
>> the ROM to be mapped? Would that bring back the issue of conflicting
>> mapping flags between kernel and userspace?
>
> For the confidential VM case I expect the answer is "yes" per this patch
> attempt:
>
> http://lore.kernel.org/20250403120228.2344377-1-naveen@kernel.org

I thought the problem here was the read() on /dev/mem, not
the mmap(), are you sure it's both?

With this patch [3/3], the memremap() hack for mem_read() goes away on
64-bit, so there should be no way it gets mapped again using that,
and the generic devmem_is_allowed() just forbids it as well.

The mmap() access in turn goes through this function

pgprot_t phys_mem_access_prot(struct file *file, unsigned long pfn,
                                unsigned long size, pgprot_t vma_prot)
{
        if (!phys_mem_access_encrypted(pfn << PAGE_SHIFT, size))
                vma_prot = pgprot_decrypted(vma_prot);

        return vma_prot;
}

which I would expect to return the correct vma_prot value already.

      Arnd