linux-kernel - Re: [PATCH v2] x86/mm: use max memory block size on bare metal

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:   Thu, 11 Jun 2020 10:05:38 -0700
From:   Dave Hansen <dave.hansen@...el.com>
To:     Daniel Jordan <daniel.m.jordan@...cle.com>
Cc:     linux-mm@...ck.org, linux-kernel@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Andy Lutomirski <luto@...nel.org>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        David Hildenbrand <david@...hat.com>,
        Michal Hocko <mhocko@...nel.org>,
        Pavel Tatashin <pasha.tatashin@...een.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Steven Sistare <steven.sistare@...cle.com>
Subject: Re: [PATCH v2] x86/mm: use max memory block size on bare metal

On 6/11/20 9:59 AM, Daniel Jordan wrote:
> On Thu, Jun 11, 2020 at 07:16:02AM -0700, Dave Hansen wrote:
>> On 6/9/20 3:54 PM, Daniel Jordan wrote:
>>> +	/*
>>> +	 * Use max block size to minimize overhead on bare metal, where
>>> +	 * alignment for memory hotplug isn't a concern.
>>> +	 */
>>> +	if (hypervisor_is_type(X86_HYPER_NATIVE)) {
>>> +		bz = MAX_BLOCK_SIZE;
>>> +		goto done;
>>> +	}
>> What ends up being the worst case scenario?  Booting a really small
>> bare-metal x86 system, say with 64MB or 128MB of RAM?  What's the
>> overhead there?
> Might not be following you, so bear with me, but we only get to this check on a
> system with a physical address end of at least MEM_SIZE_FOR_LARGE_BLOCK (64G),
> and this would still (ever so slightly...) reduce overhead of memory block init
> at boot in that case.

Ahh, I see now.  That is just above the hunk you added, but just wasn't
in the diff context or mentioned in the changelog.

One other nit for this.  We *do* have actual hardware hotplug, and I'm
pretty sure the alignment guarantees for hardware hotplug are pretty
weak.  For instance, the alignment guarantees for persistent memory are
still only 64MB even on modern platforms.

Let's say we're on bare metal and we see an SRAT table that has some
areas that show that hotplug might happen there.  Is this patch still
ideal there?