linux-kernel - Re: [RESEND PATCH 2/3] PCI: aardvark: Fix checking for PIO status

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: <20210720183042.GA108003@bjorn-Precision-5520>
Date:   Tue, 20 Jul 2021 13:30:42 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Pali Rohár <pali@...nel.org>
Cc:     Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
        Marek Behún <kabel@...nel.org>,
        linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RESEND PATCH 2/3] PCI: aardvark: Fix checking for PIO status

On Tue, Jul 20, 2021 at 08:11:55PM +0200, Pali Rohár wrote:
> On Tuesday 20 July 2021 11:34:51 Bjorn Helgaas wrote:
> > On Tue, Jul 20, 2021 at 04:49:55PM +0200, Pali Rohár wrote:

> > > So what should pci-aardvark driver in this case do? Return ~0 or re-send
> > > this config read request (and how many times)?
> > 
> > That's a good question.  I don't know what other hardware
> > implementations do, but I doubt they retry forever.  Since the spec
> > doesn't specify a number of retries, I think you can choose to do
> > none and immediately return ~0.
> 
> Ok, so I will change implementation to return ~0 without any retry. This
> is simple and easy implementation. Anyway, it would be nice to know what
> other real-hardware implementations of PCIe controllers are doing.
> 
> PCI_EXP_RTCTL_CRSSVE bit is only part of kernel emulated PCIe Bridge and
> has no real register in aardvark hw.
> 
> So for me it looks like that aardvark's pci_ops.read() should set read
> value to ~0 or 0xffff0001 based on what is stored in kernel emulated
> PCI_EXP_RTCTL_CRSSVE bit. Right?

Yes, that would be my advice.  Of course it's only if we're reading
the Vendor ID *and* PCI_EXP_RTCTL_CRSSVE is set.