linux-kernel - Re: [PATCH v10 14/17] cxl/pci: Introduce CXL Endpoint protocol error handlers

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] [day] [month] [year] [list]

Message-ID: <f569fcb2-52f0-476a-946f-56db53fa7f9c@amd.com>
Date: Tue, 22 Jul 2025 13:23:13 -0500
From: "Bowman, Terry" <terry.bowman@....com>
To: Dave Jiang <dave.jiang@...el.com>, dave@...olabs.net,
 jonathan.cameron@...wei.com, alison.schofield@...el.com,
 dan.j.williams@...el.com, bhelgaas@...gle.com, shiju.jose@...wei.com,
 ming.li@...omail.com, Smita.KoralahalliChannabasappa@....com,
 rrichter@....com, dan.carpenter@...aro.org,
 PradeepVineshReddy.Kodamati@....com, lukas@...ner.de,
 Benjamin.Cheatham@....com, sathyanarayanan.kuppuswamy@...ux.intel.com,
 linux-cxl@...r.kernel.org
Cc: linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org
Subject: Re: [PATCH v10 14/17] cxl/pci: Introduce CXL Endpoint protocol error
 handlers



On 7/21/2025 5:35 PM, Dave Jiang wrote:
>
> On 6/26/25 3:42 PM, Terry Bowman wrote:
>> CXL Endpoint protocol errors are currently handled using PCI error
>> handlers. The CXL Endpoint requires CXL specific handling in the case of
>> uncorrectable error (UCE) handling not provided by the PCI handlers.
>>
>> Add CXL specific handlers for CXL Endpoints. Rename the existing
>> cxl_error_handlers to be pci_error_handlers to more correctly indicate
>> the error type and follow naming consistency.
>>
>> The PCI handlers will be called if the CXL device is not trained for
>> alternate protocol (CXL). Update the CXL Endpoint PCI handlers to call the
>> CXL UCE handlers.
> Would the CXL device still be functional if it can't train the CXL protocols? Just wondering if we still need the standard PCI handlers in that case at all. 
>
> DJ

A CXL EP failing training will not support CXL functionality. 

Once training fails the RAS registers may be unavailable. I'm concerned accesses to the 
MMIO RAS registers could possibly cause a MCE if the PCIe device doesn't respond. It will 
depend on how the training fails. This a reason to remove the PCIe handlers.

BTW, the AER status will be logged by the AER driver before a PCIe handler is called.

A while back Dan mentioned we should leave the PCIe EP handlers. He may have an opinion 
or more to add.

-Terry

[snip]