linux-kernel - Re: [PATCH 2/2] x86: speed cpu_up by quirking cpu_init

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:	Fri, 8 May 2015 09:51:11 +0200
From:	Ingo Molnar <mingo@...nel.org>
To:	Len Brown <lenb@...nel.org>
Cc:	x86@...nel.org, linux-pm@...r.kernel.org,
	linux-kernel@...r.kernel.org, Len Brown <len.brown@...el.com>,
	Borislav Petkov <bp@...en8.de>
Subject: Re: [PATCH 2/2] x86: speed cpu_up by quirking cpu_init_udelay


* Len Brown <lenb@...nel.org> wrote:

> From: Len Brown <len.brown@...el.com>
> 
> Modern processor familes are on a white-list to remove
> the costly cpu_init_udelay 10000.  Unknown processor families
> get the traditional 10ms delay in cpu_up().
> 
> This seemed more efficient than forcing modern processors
> to exhaustively search a black-list having all the old
> processor families that should have a 10ms delay.
> For not only are new processor familes infrequently added,
> the white list also allows a delay other than 0, if needed.

>  static unsigned int init_udelay = UDELAY_10MS_DEFAULT;
>  
> +static const struct x86_cpu_id init_udelay_ids[] = {
> +	{ X86_VENDOR_INTEL, 0x6, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{ X86_VENDOR_AMD, 0x16, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{ X86_VENDOR_AMD, 0x15, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{ X86_VENDOR_AMD, 0x14, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{ X86_VENDOR_AMD, 0x12, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{ X86_VENDOR_AMD, 0x11, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{ X86_VENDOR_AMD, 0x10, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{ X86_VENDOR_AMD, 0xF, X86_MODEL_ANY, X86_FEATURE_ANY, 0 },
> +	{}
> +};

So since especially AMD likes to iterate the family upwards, why not 
make this a simple open ended check:

	if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL &&
	    boot_cpu_data.x86 >= 6 ||
	    boot_cpu_data.x86_vendor == X86_VENDOR_AMD &&
	    boot_cpu_data.x86 >= 15) {

		... 0 delay ...
	}

... which is much smaller and more future proof?

Thanks,

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/