linux-kernel - Re: [PATCH] x86/sev: Apply RMP table fixups for kexec.

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: <8daf448f-eb52-4b1f-9f98-099a83665321@amd.com>
Date: Tue, 2 Apr 2024 14:33:44 -0500
From: "Kalra, Ashish" <ashish.kalra@....com>
To: Borislav Petkov <bp@...en8.de>
Cc: bp@...nel.org, thomas.lendacky@....com, linux-kernel@...r.kernel.org,
 linux-tip-commits@...r.kernel.org, michael.roth@....com, x86@...nel.org
Subject: Re: [PATCH] x86/sev: Apply RMP table fixups for kexec.

On 4/2/2024 1:50 PM, Borislav Petkov wrote:
> On Tue, Apr 02, 2024 at 01:41:05PM -0500, Kalra, Ashish wrote:
>> But, that's what the e820 table fixup is doing,
> And can it be doing it in snp_probe_rmptable_info() where all the RMP
> table stuff is being done?

The e820 fixups has to be done on both e820_table_kexec and 
e820_table_firmware (to handle both kexec variants) and this has to be 
done when e820_table is used to populate both e820_table_kexec and 
e820_table_firmware and this is done in e820__memory_setup().

But, snp_probe_rmptable_info() is called as part of early_cpu_init() -> 
early_identify_cpu() which is called earlier than e820__memory_setup() 
and e820 tables are setup much later e820__memory_setup().

And we can't do this in snp_rmptable_init() as e820_table_firmware can't 
be fixed at that point and by that time this table has been mapped into 
sysfs (/sys/firmware) which is used by kexec -c variant.

The function e820__memory_setup() has an optional platform quirk 
available to override/pick up the firmware/bootloader e820 map, and 
that's why i am using this platform specific quirk to override it with 
the SNP specific RMP table e820 fixup function which internally calls 
e820__memory_setup_default() to pass the firmware/bootloader e820 map 
(set in e820_table) and then add RMP table fixups on top of it.

Thanks, Ashish