mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Bo Liu (刘波)-浪潮信息" <liubo03@inspur.com>
To: "kristen.c.accardi@intel.com" <kristen.c.accardi@intel.com>,
	"vinicius.gomes@intel.com" <vinicius.gomes@intel.com>
Cc: "linux-crypto@vger.kernel.org" <linux-crypto@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"Dou zhi chong Dou (窦志冲)" <douzhichong@inspur.com>
Subject: Intel IAA issue consultation
Date: Fri, 28 Mar 2025 03:26:27 +0000	[thread overview]
Message-ID: <7f9bf073f04d4ba6b0d98f6189d951bc@inspur.com> (raw)

[-- Attachment #1: Type: text/plain, Size: 1008 bytes --]

Hi Kristen:

We recently encountered issues when using Intel IAA (In-Memory Analytics Accelerator) for data decompression in the Linux kernel. Specifically, while using the iaa_crypto driver, decompression fails when processing large datasets composed of non-contiguous memory pages. Initially, we attempted to address this by using scatterlists (SGLs), but decompression still failed. After analyzing the iaa_crypto driver code, we discovered that ​non-contiguous pages require enabling IOMMU to map them into a single virtual contiguous memory block for the device. However, while this approach worked for small numbers of non-contiguous pages, it failed for larger quantities due to IOMMU limitations in mapping fragmented pages into a unified virtual address space.

We would like to consult with you

Is it possible to use iaa_crypto for decompression with ​numerous non-contiguous pages?
If yes, what modifications are required to resolve the IOMMU mapping constraints?

Thanks,
Bo Liu.

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 3777 bytes --]

                 reply	other threads:[~2025-03-28  3:26 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=7f9bf073f04d4ba6b0d98f6189d951bc@inspur.com \
    --to=liubo03@inspur.com \
    --cc=douzhichong@inspur.com \
    --cc=kristen.c.accardi@intel.com \
    --cc=linux-crypto@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=vinicius.gomes@intel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®