linux-kernel - Re: [PATCH] vfio/pci: Some buggy virtual functions incorrectly report 1 for intx.

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: <20180919164741.76fbcc88@t450s.home>
Date:   Wed, 19 Sep 2018 16:47:41 -0600
From:   Alex Williamson <alex.williamson@...hat.com>
To:     "Eads, Gage" <gage.eads@...el.com>
Cc:     "Raj, Ashok" <ashok.raj@...el.com>,
        "kvm@...r.kernel.org" <kvm@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "iommu@...ts.linux-foundation.org" <iommu@...ts.linux-foundation.org>,
        Joerg Roedel <joro@...tes.org>,
        Bjorn Helgaas <helgaas@...nel.org>,
        "Alan Cox" <gnomes@...rguk.ukuu.org.uk>
Subject: Re: [PATCH] vfio/pci: Some buggy virtual functions incorrectly
 report 1 for intx.

On Wed, 19 Sep 2018 22:25:03 +0000
"Eads, Gage" <gage.eads@...el.com> wrote:

> > This looks good and also addresses Alan's concern that don't
> > silently hide under the rug for all devices. We'll also queue it
> > for testing just to confirm and keep you posted.
> > 
> > Reviewed-by: Ashok Raj <ashok.raj@...el.com>
> >   
> 
> Hi Alex,
> 
> I've confirmed that the patch works as intended for the 8086:270c
> device, and negative tested the warning by commenting out the
> device's entry in known_bogus_vf_intx_pin.
> 
> For what it's worth, my positive test case - launching a QEMU VM with
> a VFIO-owned 0x270c device - did not trigger the warning. This is
> because the change to vfio_pci_config.c caused QEMU to read a
> PCI_INTERRUPT_PIN value of 0, and so didn't execute a codepath that
> calls vfio_pci_get_irq_count(). Instead, I used a simple C program
> that calls the VFIO_DEVICE_GET_IRQ_INFO ioctl for negative testing.
> 
> Also, there's one typo in the comment: 'quite' -> 'quiet'
> 
> Tested-by: Gage Eads <gage.eads@...el.com>

Thanks for the thorough testing, Gage!  So it sounds like we're not
quite generating the desired result.  We'd really like QEMU, as the
predominant userspace driver for VFIO, to generate the warning,
otherwise there's hardly any purpose to having it.  That suggests I
should have put the warning where the config space emulation is setup
rather than the irq count path.  Thanks,

Alex