linux-kernel - Re: [PATCH v14 4/5] x86/boot: Parse SRAT address from RSDP and store immovable memory

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]

Date:   Mon, 17 Dec 2018 18:41:49 +0100
From:   Ingo Molnar <mingo@...nel.org>
To:     Chao Fan <fanc.fnst@...fujitsu.com>
Cc:     linux-kernel@...r.kernel.org, x86@...nel.org, bp@...en8.de,
        tglx@...utronix.de, mingo@...hat.com, hpa@...or.com,
        keescook@...omium.org, bhe@...hat.com, msys.mizuma@...il.com,
        indou.takao@...fujitsu.com, caoj.fnst@...fujitsu.com
Subject: Re: [PATCH v14 4/5] x86/boot: Parse SRAT address from RSDP and store
 immovable memory


* Chao Fan <fanc.fnst@...fujitsu.com> wrote:

> SRAT should be parsed by RSDP to fix the conflict between KASLR
> and memory-hotremove, then find the immovable memory regions and store
> them in an array called immovable_mem[]. With immovable_mem[], KASLR
> can avoid to extract kernel to specific regions.
> 
> Since 'RANDOMIZE_BASE' && 'MEMORY_HOTREMOVE' is needed, introduce
> 'CONFIG_EARLY_PARSE_RSDP' to make ifdeffery clear.
> 
> Signed-off-by: Chao Fan <fanc.fnst@...fujitsu.com>
> ---
>  arch/x86/Kconfig                  |  12 +++
>  arch/x86/boot/compressed/Makefile |   2 +
>  arch/x86/boot/compressed/acpi.c   | 128 ++++++++++++++++++++++++++++++
>  arch/x86/boot/compressed/kaslr.c  |   4 -
>  arch/x86/boot/compressed/misc.h   |  19 +++++
>  5 files changed, 161 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index ba7e3464ee92..333c383478b7 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -2149,6 +2149,18 @@ config X86_NEED_RELOCS
>  	def_bool y
>  	depends on RANDOMIZE_BASE || (X86_32 && RELOCATABLE)
>  
> +config EARLY_SRAT_PARSE
> +	bool "Early SRAT table parsing"
> +	def_bool y
> +	depends on RANDOMIZE_BASE && MEMORY_HOTREMOVE
> +	help
> +	  This option enables early SRAT parsing in compressed boot stage
> +	  so that memory hot-remove ranges do not overlap with KASLR
> +	  chosen ranges. Kernel won't be extracted in hot-removable
> +	  memory, so that make sure memory-hotremove works well with
> +	  KASLR enabled.
> +	  Say Y if you want to use both KASLR and memory-hotremove.

So why would we want to make this a config option, instead of enabling it 
unconditionally?

How reliable are the hot-removable memory markings in various firmware 
versions?

> +/* Compute SRAT table from RSDP. */
> +static struct acpi_table_header *get_acpi_srat_table(void)
> +{
> +	acpi_physical_address acpi_table;
> +	acpi_physical_address root_table;
> +	struct acpi_table_header *header;
> +	struct acpi_table_rsdp *rsdp;
> +	u32 num_entries;
> +	char arg[10];

The '10' is just a magic number attached to a meaningless local variable 
name. Please explain the limit in the code, and the role of the variable 
if it's non-obvious from the name. Or better, try to find a more obvious 
name?

Thanks,

	Ingo