linux-kernel - Re: [PATCH 2/2] iommufd/selftest: Use normal IOMMU registration

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: <3a8d7ea4-c5f9-421e-84fa-2e4934cd6c92@arm.com>
Date:   Tue, 28 Nov 2023 14:53:03 +0000
From:   Robin Murphy <robin.murphy@....com>
To:     Jason Gunthorpe <jgg@...pe.ca>
Cc:     joro@...tes.org, kevin.tian@...el.com, will@...nel.org,
        iommu@...ts.linux.dev, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] iommufd/selftest: Use normal IOMMU registration

On 28/11/2023 2:35 pm, Jason Gunthorpe wrote:
> On Tue, Nov 28, 2023 at 10:42:12AM +0000, Robin Murphy wrote:
>> The IOMMU core now supports coexistence of fwspec-based drivers, which
>> the mock driver now is, so let's bring the mock bus into iommu_buses,
>> drop the special interface, and use the normal registration flow. The
>> one concession we have to make is to ensure that the mock bus is
>> registered early enough so that bus_for_each_dev() doesn't error out
>> for other IOMMU drivers registering before iommufd_test_init() runs.
> 
> This makes iommufd non-modular which becomes a total PITA for development :(

Oh fiddle, I misread the makefile, and indeed this doesn't work at all, 
sorry (turns out it fails to even build for IOMMUFD=m...) Guess I should 
have been more wary of how suspiciously easy it seemed :(

Dynamic bus registration in general would be a neat thing to explore at 
some point, since the static iommu_buses array isn't my most favourite 
part of the whole business, but I guess we leave this as-is for now.

Robin.