From: Jason Gunthorpe <jgg@ziepe.ca>
To: "Zhijian Li (Fujitsu)" <lizhijian@fujitsu.com>
Cc: "linux-rdma@vger.kernel.org" <linux-rdma@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"zyjzyj2000@gmail.com" <zyjzyj2000@gmail.com>,
"leon@kernel.org" <leon@kernel.org>,
Yi Zhang <yi.zhang@redhat.com>,
Bob Pearson <rpearsonhpe@gmail.com>
Subject: Re: [PATCH RFC] rxe: Fix iova-to-va conversion for MR page sizes != PAGE_SIZE
Date: Mon, 5 Jan 2026 21:07:09 -0400 [thread overview]
Message-ID: <20260106010709.GO125261@ziepe.ca> (raw)
In-Reply-To: <cbbba297-7095-49f1-82e0-8f22d4f94e1a@fujitsu.com>
On Mon, Jan 05, 2026 at 06:55:22AM +0000, Zhijian Li (Fujitsu) wrote:
> After digging into the behavior during the srp/012 test again, it
> turns out this fix is incomplete. The current xarray page_list
> approach cannot correctly map memory regions composed of two or more
> scatter-gather segments.
I seem to recall there are DMA API functions that can control what
kinds of scatterlists the block stack will push down.
For real HW we already cannot support less than 4K alignment of interior SGL
segments.
Maybe rxe can tell the block stack it can only support PAGE_SIZE
alignment of interior SGL segments?
If not then this would be the reason rxe needs mr->page_size, to
support 4k.
And obviously if the mr->page size is less than PAGE_SIZE the xarray
datastructure does not work. You'd have to store physical addresses
instead..
Jason
next prev parent reply other threads:[~2026-01-06 1:07 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-26 9:52 [PATCH RFC] rxe: Fix iova-to-va conversion for MR page sizes != PAGE_SIZE Li Zhijian
2025-12-27 5:24 ` Zhu Yanjun
2025-12-30 4:02 ` Zhu Yanjun
2025-12-30 4:47 ` Zhijian Li (Fujitsu)
2025-12-30 4:57 ` Zhu Yanjun
2026-01-05 6:55 ` Zhijian Li (Fujitsu)
2026-01-05 7:39 ` Zhijian Li (Fujitsu)
2026-01-06 1:07 ` Jason Gunthorpe [this message]
2026-01-07 3:02 ` Zhijian Li (Fujitsu)
2026-01-06 1:01 ` Jason Gunthorpe
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=20260106010709.GO125261@ziepe.ca \
--to=jgg@ziepe.ca \
--cc=leon@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=lizhijian@fujitsu.com \
--cc=rpearsonhpe@gmail.com \
--cc=yi.zhang@redhat.com \
--cc=zyjzyj2000@gmail.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.