linux-kernel - Re: [PATCH v1 1/2] PCI/AER: Decode Error Source Requester ID

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: <CACK8Z6H+sZztSjKbHFqK1efCwruxVSdDs+5kt96_-NPdM2eGmw@mail.gmail.com>
Date:   Thu, 31 May 2018 10:35:51 -0700
From:   Rajat Jain <rajatja@...gle.com>
To:     Bjorn Helgaas <helgaas@...nel.org>
Cc:     linux-pci <linux-pci@...r.kernel.org>,
        Oza Pawandeep <poza@...eaurora.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 1/2] PCI/AER: Decode Error Source Requester ID

On Wed, May 30, 2018 at 9:42 PM Bjorn Helgaas <helgaas@...nel.org> wrote:
>
> On Wed, May 30, 2018 at 11:41:23AM -0700, Rajat Jain wrote:
> > On Wed, May 30, 2018 at 10:54 AM Bjorn Helgaas <helgaas@...nel.org> wrote:
> >
> > > From: Bjorn Helgaas <bhelgaas@...gle.com>
> >
> > > Decode the Requester ID from the AER Error Source Register into domain/
> > > bus/device/function format to match other logging.  In cases where the ID
> > > matches the device used for pci_err(), drop the extra ID completely so we
> > > don't print it twice.
> >
> > > Signed-off-by: Bjorn Helgaas <bhelgaas@...gle.com>
> > > ---
> > >   drivers/pci/pcie/aer/aerdrv_errprint.c |   18 +++++++++++-------
> > >   1 file changed, 11 insertions(+), 7 deletions(-)
> >
> > > diff --git a/drivers/pci/pcie/aer/aerdrv_errprint.c
> > b/drivers/pci/pcie/aer/aerdrv_errprint.c
> > > index 21ca5e1b0ded..d7fde8368d81 100644
> > > --- a/drivers/pci/pcie/aer/aerdrv_errprint.c
> > > +++ b/drivers/pci/pcie/aer/aerdrv_errprint.c
> > > @@ -163,17 +163,17 @@ void aer_print_error(struct pci_dev *dev, struct
> > aer_err_info *info)
> > >          int id = ((dev->bus->number << 8) | dev->devfn);
>
> > >          if (!info->status) {
> > > -               pci_err(dev, "PCIe Bus Error: severity=%s,
> > type=Unaccessible, id=%04x(Unregistered Agent ID)\n",
> > > -                       aer_error_severity_string[info->severity], id);
> > > +               pci_err(dev, "PCIe Bus Error: severity=%s,
> > type=Inaccessible, (Unregistered Agent ID)\n",
> > > +                       aer_error_severity_string[info->severity]);
> >
> > Does this code path indicate that a requester id was decoded to a device
> > that is not registered with the kernel? If so, shouldn't we log the bad
> > requester ID for better debugging, specifically since there is not going to
> > be any subsequent print about this ID (since we return from this function
> > in this case)?
>
> Previously we printed "id", which was constructed above from "dev":
>
>   id = ((dev->bus->number << 8) | dev->devfn);
>
> so even if we print "id=%04x", it contains exactly the same
> information as the bus/device/function printed using "dev".

Sorry, my bad, I missed it, despite it being right there in my face :-).

>
> So no, I don't think "Unregistered Agent ID" means a device not registered
> with the kernel.  At any rate, we do have a pci_dev for it.
>
> I *think* "info->status == 0" means PCI_ERR_COR_STATUS (or
> PCI_ERR_UNCOR_STATUS) was zero, i.e., we didn't find any error status
> bits set for this device.  I don't think "Unregistered Agent ID" is a
> very good description of this situation.

Agree, may be something along the lines of "Unknown Error Status"
might be better.

Thanks,

Rajat

>
> Bjorn