linux-kernel - Re: ECRC configuration if AER is owned by the firmware

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

Date:   Mon, 31 Oct 2022 12:23:01 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Vidya Sagar <vidyas@...dia.com>
Cc:     "sr@...x.de" <sr@...x.de>,
        "linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: ECRC configuration if AER is owned by the firmware

On Mon, Oct 31, 2022 at 10:35:40PM +0530, Vidya Sagar wrote:
> Hi,
> With the top of the tree codebase, pcie_set_ecrc_checking() is called from
> pci_aer_init().
> This flow doesn't really take into account the ownership of the AER
> registers, and it makes sense in the case of the Linux kernel owning the
> AER.
> But, if the owner of AER is firmware and not OS, then, ideally, any
> register/fields (including ECRC bits) are not supposed to be touched by the
> OS.
> Is my understanding correct?
> If so, then, isn't the current code violating that rule?
> Could you please comment on it?

I think you are correct and the current code violates the rule.

Bjorn

P.S.  Your message was a multi-part message that will likely be
rejected by the vger mailing lists:
http://vger.kernel.org/majordomo-info.html