linux-kernel - Re: [PATCH] iommu: fix crash in report_iommu

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20250410130006.GD1727154@ziepe.ca>
Date: Thu, 10 Apr 2025 10:00:06 -0300
From: Jason Gunthorpe <jgg@...pe.ca>
To: "Tian, Kevin" <kevin.tian@...el.com>
Cc: Robin Murphy <robin.murphy@....com>,
	Fedor Pchelkin <pchelkin@...ras.ru>, Joerg Roedel <joro@...tes.org>,
	Will Deacon <will@...nel.org>, Nicolin Chen <nicolinc@...dia.com>,
	"iommu@...ts.linux.dev" <iommu@...ts.linux.dev>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"lvc-project@...uxtesting.org" <lvc-project@...uxtesting.org>
Subject: Re: [PATCH] iommu: fix crash in report_iommu_fault()

On Thu, Apr 10, 2025 at 05:29:51AM +0000, Tian, Kevin wrote:

> It sounds a general requirement to all IOMMU drivers, but in reality
> only a subset of iommu drivers call it (e.g. intel/smmuv3 don't). So
> there seems to be an implicit assumption from drivers on whether
> the underlying IOMMU provides such facility...

Yeah, it's a bit wonky. And a different intersection calls the new
fault handling API instead :\

There are only 3 things using iommu_set_fault_handler():
 - drivers/gpu/drm/msm
   Seems to actually do restartable page faulting?? Maybe it needs to
   move to the PRI API..

 - drivers/infiniband/hw/usnic
   This just prints a log, we should remove it

 - drivers/remoteproc
   This prints a log and does some crash_handler

I just quickly typed in a small series to improve on this..

Jason