linux-kernel - Re: [PATCH] iommu/amd: Fix IOMMU page flush when detach all devices from a domain

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]

Message-ID: <20190116170852.GD4681@8bytes.org>
Date:   Wed, 16 Jan 2019 18:08:52 +0100
From:   "joro@...tes.org" <joro@...tes.org>
To:     "Suthikulpanit, Suravee" <Suravee.Suthikulpanit@....com>
Cc:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "iommu@...ts.linux-foundation.org" <iommu@...ts.linux-foundation.org>,
        Boris Ostrovsky <boris.ostrovsky@...cle.com>,
        "Singh, Brijesh" <brijesh.singh@....com>
Subject: Re: [PATCH] iommu/amd: Fix IOMMU page flush when detach all devices
 from a domain

On Wed, Jan 16, 2019 at 02:08:55PM +0000, Suthikulpanit, Suravee wrote:
> Actually, I am not sure how we would be missing the flush on the last device.
> In my test, I am seeing the flush command being issued correctly during
> vfio_unmap_unpin(), which is after all devices are detached.
> Although, I might be missing your point here. Could you please elaborate?

Okay, you are right, the patch effectivly adds an unconditional flush of
the domain on all iommus when the last device is detached. So it is
correct in that regard. But that code path is also used in the
iommu_unmap() path.

The problem now is, that with your change we send flush commands to all
IOMMUs in the unmap path when no device is attached to the domain.
This will hurt performance there, no?

Regards,

	Joerg