From: David Hildenbrand <david@redhat.com>
To: Vlastimil Babka <vbabka@suse.cz>,
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
Borislav Petkov <bp@alien8.de>, Andy Lutomirski <luto@kernel.org>,
Sean Christopherson <seanjc@google.com>,
Andrew Morton <akpm@linux-foundation.org>,
Joerg Roedel <jroedel@suse.de>, Ard Biesheuvel <ardb@kernel.org>
Cc: Andi Kleen <ak@linux.intel.com>,
Kuppuswamy Sathyanarayanan
<sathyanarayanan.kuppuswamy@linux.intel.com>,
David Rientjes <rientjes@google.com>,
Tom Lendacky <thomas.lendacky@amd.com>,
Thomas Gleixner <tglx@linutronix.de>,
Peter Zijlstra <peterz@infradead.org>,
Paolo Bonzini <pbonzini@redhat.com>,
Ingo Molnar <mingo@redhat.com>,
Dario Faggioli <dfaggioli@suse.com>,
Dave Hansen <dave.hansen@intel.com>,
Mike Rapoport <rppt@kernel.org>,
marcelo.cerri@canonical.com, tim.gardner@canonical.com,
khalid.elmously@canonical.com, philip.cox@canonical.com,
x86@kernel.org, linux-mm@kvack.org, linux-coco@lists.linux.dev,
linux-efi@vger.kernel.org, linux-kernel@vger.kernel.org,
Mike Rapoport <rppt@linux.ibm.com>,
Mel Gorman <mgorman@techsingularity.net>
Subject: Re: [PATCHv7 02/14] mm: Add support for unaccepted memory
Date: Fri, 5 Aug 2022 14:09:31 +0200 [thread overview]
Message-ID: <cb9d3310-3bc0-8ecf-5e71-becce980235f@redhat.com> (raw)
In-Reply-To: <8cf143e7-2b62-1a1e-de84-e3dcc6c027a4@suse.cz>
On 05.08.22 13:49, Vlastimil Babka wrote:
> On 6/14/22 14:02, Kirill A. Shutemov wrote:
>> UEFI Specification version 2.9 introduces the concept of memory
>> acceptance. Some Virtual Machine platforms, such as Intel TDX or AMD
>> SEV-SNP, require memory to be accepted before it can be used by the
>> guest. Accepting happens via a protocol specific to the Virtual Machine
>> platform.
>>
>> There are several ways kernel can deal with unaccepted memory:
>>
>> 1. Accept all the memory during the boot. It is easy to implement and
>> it doesn't have runtime cost once the system is booted. The downside
>> is very long boot time.
>>
>> Accept can be parallelized to multiple CPUs to keep it manageable
>> (i.e. via DEFERRED_STRUCT_PAGE_INIT), but it tends to saturate
>> memory bandwidth and does not scale beyond the point.
>>
>> 2. Accept a block of memory on the first use. It requires more
>> infrastructure and changes in page allocator to make it work, but
>> it provides good boot time.
>>
>> On-demand memory accept means latency spikes every time kernel steps
>> onto a new memory block. The spikes will go away once workload data
>> set size gets stabilized or all memory gets accepted.
>>
>> 3. Accept all memory in background. Introduce a thread (or multiple)
>> that gets memory accepted proactively. It will minimize time the
>> system experience latency spikes on memory allocation while keeping
>> low boot time.
>>
>> This approach cannot function on its own. It is an extension of #2:
>> background memory acceptance requires functional scheduler, but the
>> page allocator may need to tap into unaccepted memory before that.
>>
>> The downside of the approach is that these threads also steal CPU
>> cycles and memory bandwidth from the user's workload and may hurt
>> user experience.
>>
>> Implement #2 for now. It is a reasonable default. Some workloads may
>> want to use #1 or #3 and they can be implemented later based on user's
>> demands.
>>
>> Support of unaccepted memory requires a few changes in core-mm code:
>>
>> - memblock has to accept memory on allocation;
>>
>> - page allocator has to accept memory on the first allocation of the
>> page;
>>
>> Memblock change is trivial.
>>
>> The page allocator is modified to accept pages on the first allocation.
>> The new page type (encoded in the _mapcount) -- PageUnaccepted() -- is
>> used to indicate that the page requires acceptance.
>>
>> Architecture has to provide two helpers if it wants to support
>> unaccepted memory:
>>
>> - accept_memory() makes a range of physical addresses accepted.
>>
>> - range_contains_unaccepted_memory() checks anything within the range
>> of physical addresses requires acceptance.
>>
>> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
>> Acked-by: Mike Rapoport <rppt@linux.ibm.com> # memblock
>> Reviewed-by: David Hildenbrand <david@redhat.com>
>
> Hmm I realize it's not ideal to raise this at v7, and maybe it was discussed
> before, but it's really not great how this affects the core page allocator
> paths. Wouldn't it be possible to only release pages to page allocator when
> accepted, and otherwise use some new per-zone variables together with the
> bitmap to track how much exactly is where to accept? Then it could be hooked
> in get_page_from_freelist() similarly to CONFIG_DEFERRED_STRUCT_PAGE_INIT -
> if we fail zone_watermark_fast() and there are unaccepted pages in the zone,
> accept them and continue. With a static key to flip in case we eventually
> accept everything. Because this is really similar scenario to the deferred
> init and that one was solved in a way that adds minimal overhead.
I kind of like just having the memory stats being correct (e.g., free
memory) and acceptance being an internal detail to be triggered when
allocating pages -- just like the arch_alloc_page() callback.
I'm sure we could optimize for the !unaccepted memory via static keys
also in this version with some checks at the right places if we find
this to hurt performance?
--
Thanks,
David / dhildenb
next prev parent reply other threads:[~2022-08-05 12:09 UTC|newest]
Thread overview: 139+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-14 12:02 [PATCHv7 00/14] mm, x86/cc: Implement support for unaccepted memory Kirill A. Shutemov
2022-06-14 12:02 ` [PATCHv7 01/14] x86/boot: Centralize __pa()/__va() definitions Kirill A. Shutemov
2022-06-23 17:37 ` Dave Hansen
2022-06-14 12:02 ` [PATCHv7 02/14] mm: Add support for unaccepted memory Kirill A. Shutemov
2022-06-14 12:57 ` Gupta, Pankaj
2022-06-17 19:28 ` Tom Lendacky
2022-06-17 20:53 ` Tom Lendacky
2022-07-21 15:14 ` Borislav Petkov
2022-07-21 15:49 ` Dave Hansen
2022-07-22 19:18 ` Borislav Petkov
2022-07-22 19:30 ` Dave Hansen
2022-07-25 12:23 ` Borislav Petkov
2022-07-25 12:38 ` David Hildenbrand
2022-07-25 12:53 ` Borislav Petkov
2022-07-26 14:30 ` David Hildenbrand
2022-07-25 13:00 ` Mike Rapoport
2022-07-25 13:05 ` Borislav Petkov
2022-08-05 11:49 ` Vlastimil Babka
2022-08-05 12:09 ` David Hildenbrand [this message]
2022-08-05 13:38 ` Vlastimil Babka
2022-08-05 14:22 ` David Hildenbrand
2022-08-05 14:53 ` Dave Hansen
2022-08-05 14:41 ` Dave Hansen
2022-08-05 18:17 ` Vlastimil Babka
2022-08-08 15:55 ` Dave Hansen
2022-08-10 14:19 ` Mel Gorman
2022-08-15 21:08 ` Dionna Amalie Glaze
2022-08-15 22:02 ` Tom Lendacky
2022-08-29 16:02 ` Dionna Amalie Glaze
2022-08-29 16:19 ` Dave Hansen
2022-09-06 17:50 ` Dionna Amalie Glaze
2022-09-08 12:11 ` Mike Rapoport
2022-09-08 16:23 ` Dionna Amalie Glaze
2022-09-08 19:28 ` Mike Rapoport
2022-09-22 14:31 ` Tom Lendacky
2022-09-24 1:03 ` Kirill A. Shutemov
2022-09-24 9:36 ` Mike Rapoport
2022-09-26 12:10 ` Kirill A. Shutemov
2022-09-26 13:38 ` Tom Lendacky
2022-09-26 15:42 ` Kirill A. Shutemov
2022-09-26 15:42 ` Tom Lendacky
2022-06-14 12:02 ` [PATCHv7 03/14] mm: Report unaccepted memory in meminfo Kirill A. Shutemov
2022-07-26 14:33 ` David Hildenbrand
2022-06-14 12:02 ` [PATCHv7 04/14] efi/x86: Get full memory map in allocate_e820() Kirill A. Shutemov
2022-07-25 13:02 ` Borislav Petkov
2022-06-14 12:02 ` [PATCHv7 05/14] x86/boot: Add infrastructure required for unaccepted memory support Kirill A. Shutemov
2022-06-15 10:19 ` Peter Zijlstra
2022-06-15 15:05 ` Kirill A. Shutemov
2022-07-17 17:16 ` Borislav Petkov
2022-07-25 21:33 ` Borislav Petkov
2022-06-14 12:02 ` [PATCHv7 06/14] efi/x86: Implement support for unaccepted memory Kirill A. Shutemov
2022-06-22 19:58 ` Dave Hansen
2022-07-26 8:35 ` Borislav Petkov
2022-06-14 12:02 ` [PATCHv7 07/14] x86/boot/compressed: Handle " Kirill A. Shutemov
2022-06-14 12:02 ` [PATCHv7 08/14] x86/mm: Reserve unaccepted memory bitmap Kirill A. Shutemov
2022-07-26 9:07 ` Borislav Petkov
2022-11-30 1:28 ` Kirill A. Shutemov
2022-12-01 9:37 ` Mike Rapoport
2022-12-01 13:47 ` Kirill A. Shutemov
2022-06-14 12:02 ` [PATCHv7 09/14] x86/mm: Provide helpers for unaccepted memory Kirill A. Shutemov
2022-06-14 12:02 ` [PATCHv7 10/14] x86/mm: Avoid load_unaligned_zeropad() stepping into " Kirill A. Shutemov
2022-06-23 17:19 ` Dave Hansen
2022-07-26 10:21 ` Borislav Petkov
2022-08-02 23:46 ` Dave Hansen
2022-08-03 14:02 ` Dave Hansen
2022-08-11 11:26 ` Borislav Petkov
2022-08-13 16:11 ` Andy Lutomirski
2022-08-13 21:13 ` Kirill A. Shutemov
2022-08-13 16:04 ` Andy Lutomirski
2022-08-13 20:58 ` Kirill A. Shutemov
2022-07-26 17:25 ` Borislav Petkov
2022-07-26 17:46 ` Dave Hansen
2022-07-26 20:17 ` Andy Lutomirski
2022-08-09 11:38 ` Kirill A. Shutemov
2022-08-13 16:03 ` Andy Lutomirski
2022-08-13 21:02 ` Kirill A. Shutemov
2022-06-14 12:02 ` [PATCHv7 11/14] x86: Disable kexec if system has " Kirill A. Shutemov
2022-06-23 17:23 ` Dave Hansen
2022-06-23 21:48 ` Eric W. Biederman
2022-06-24 2:00 ` Kirill A. Shutemov
2022-06-28 23:51 ` Kirill A. Shutemov
2022-06-29 0:10 ` Dave Hansen
2022-06-29 0:59 ` Kirill A. Shutemov
2022-07-04 7:18 ` Dave Young
2022-06-14 12:02 ` [PATCHv7 12/14] x86/tdx: Make _tdx_hypercall() and __tdx_module_call() available in boot stub Kirill A. Shutemov
2022-06-23 17:25 ` Dave Hansen
2022-06-14 12:02 ` [PATCHv7 13/14] x86/tdx: Refactor try_accept_one() Kirill A. Shutemov
2022-06-23 17:31 ` Dave Hansen
2022-07-26 10:58 ` Borislav Petkov
2022-06-14 12:02 ` [PATCHv7 14/14] x86/tdx: Add unaccepted memory support Kirill A. Shutemov
2022-06-24 16:22 ` Dave Hansen
2022-06-27 10:42 ` Kirill A. Shutemov
2022-07-26 14:51 ` Borislav Petkov
2022-08-09 11:45 ` Kirill A. Shutemov
2022-08-10 10:27 ` Borislav Petkov
2022-06-24 16:37 ` [PATCHv7 00/14] mm, x86/cc: Implement support for unaccepted memory Peter Gonda
2022-06-24 16:57 ` Dave Hansen
2022-06-24 17:06 ` Marc Orr
2022-06-24 17:09 ` Dave Hansen
2022-06-24 17:15 ` Peter Gonda
2022-06-24 17:19 ` Marc Orr
2022-06-24 17:21 ` Peter Gonda
2022-06-24 17:47 ` Dave Hansen
2022-06-24 18:10 ` Peter Gonda
2022-06-24 18:13 ` Dave Hansen
2022-06-24 17:40 ` Michael Roth
2022-06-24 17:58 ` Michael Roth
2022-06-24 18:05 ` Peter Gonda
2022-06-27 11:30 ` Kirill A. Shutemov
2022-06-27 11:54 ` Ard Biesheuvel
2022-06-27 12:22 ` Kirill A. Shutemov
2022-06-27 16:17 ` Peter Gonda
2022-06-27 16:33 ` Ard Biesheuvel
2022-06-27 22:38 ` Kirill A. Shutemov
2022-06-28 17:17 ` Ard Biesheuvel
2022-07-18 17:21 ` Kirill A. Shutemov
2022-07-18 23:32 ` Dionna Amalie Glaze
2022-07-19 0:31 ` Dionna Amalie Glaze
2022-07-19 18:29 ` Dionna Amalie Glaze
2022-07-19 19:13 ` Borislav Petkov
2022-07-19 20:45 ` Ard Biesheuvel
2022-07-19 21:23 ` Borislav Petkov
2022-07-19 21:35 ` Dave Hansen
2022-07-19 21:50 ` Borislav Petkov
2022-07-19 22:01 ` Kirill A. Shutemov
2022-07-19 22:02 ` Dave Hansen
2022-07-19 22:08 ` Tom Lendacky
2022-07-20 0:26 ` Marc Orr
2022-07-20 5:44 ` Borislav Petkov
2022-07-20 17:03 ` Marc Orr
2022-07-22 15:07 ` Borislav Petkov
2022-07-21 17:12 ` Dave Hansen
2022-07-23 11:14 ` Ard Biesheuvel
2022-07-28 22:01 ` Dionna Amalie Glaze
2022-08-09 11:14 ` Kirill A. Shutemov
2022-08-09 11:36 ` Ard Biesheuvel
2022-08-09 11:54 ` Kirill A. Shutemov
2022-08-09 21:09 ` Dionna Amalie Glaze
2022-07-19 2:48 ` Yao, Jiewen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=cb9d3310-3bc0-8ecf-5e71-becce980235f@redhat.com \
--to=david@redhat.com \
--cc=ak@linux.intel.com \
--cc=akpm@linux-foundation.org \
--cc=ardb@kernel.org \
--cc=bp@alien8.de \
--cc=dave.hansen@intel.com \
--cc=dfaggioli@suse.com \
--cc=jroedel@suse.de \
--cc=khalid.elmously@canonical.com \
--cc=kirill.shutemov@linux.intel.com \
--cc=linux-coco@lists.linux.dev \
--cc=linux-efi@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=luto@kernel.org \
--cc=marcelo.cerri@canonical.com \
--cc=mgorman@techsingularity.net \
--cc=mingo@redhat.com \
--cc=pbonzini@redhat.com \
--cc=peterz@infradead.org \
--cc=philip.cox@canonical.com \
--cc=rientjes@google.com \
--cc=rppt@kernel.org \
--cc=rppt@linux.ibm.com \
--cc=sathyanarayanan.kuppuswamy@linux.intel.com \
--cc=seanjc@google.com \
--cc=tglx@linutronix.de \
--cc=thomas.lendacky@amd.com \
--cc=tim.gardner@canonical.com \
--cc=vbabka@suse.cz \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).