[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4d051167-9419-43fe-ab80-701c3f46b19f@redhat.com>
Date: Tue, 1 Apr 2025 16:50:40 +0200
From: David Hildenbrand <david@...hat.com>
To: Gregory Price <gourry@...rry.net>
Cc: dan.j.williams@...el.com, nvdimm@...ts.linux.dev,
linux-kernel@...r.kernel.org, kernel-team@...a.com,
vishal.l.verma@...el.com, dave.jiang@...el.com, linux-cxl@...r.kernel.org
Subject: Re: [PATCH] DAX: warn when kmem regions are truncated for memory
block alignment.
On 01.04.25 16:43, Gregory Price wrote:
> On Tue, Apr 01, 2025 at 11:33:59AM +0200, David Hildenbrand wrote:
>> On 31.03.25 20:27, Gregory Price wrote:
>>> We discussed [1] how this auto-sizing can cause 1GB huge page
>>> allocation failures (assuming you online as ZONE_NORMAL). That means
>>> ACPI-informed sizing by default would potentially be harmful to existing
>>> systems and adding yet-another-boot-option just seems nasty.
>>>
>>> I've since dropped acpi-informed block size patch[2]. If there are opinions
>>> otherwise, I can continue pushing it.
>>
>> Oh, I thought we would be going forward with that. What's the reason we
>> would not want to do that?
>>
>
> It seemed like having it reduce block size by default would make 1GB huge
> pages less reliable to allocate. If you think this isn't a large concern,
> I can update and push again.
Oh, you mean with the whole memmap_on_memory thing. Even with that,
using 2GB memory blocks would only fit a single 1GB memory block ... and
it requires ZONE_NORMAL.
For ordinary boot memory, the 1GB behavior should be independent of the
memory block size (a 1GB page can span multiple blocks as long as they
are in the same zone), which is the most important thing.
So I don't think it's a concern for DAX right now. Whoever needs that,
can disable the memmap_on_memory option.
--
Cheers,
David / dhildenb
Powered by blists - more mailing lists