linux-kernel - Re: [PATCH 04/15] PCI: Add pci_find_vsec_capability() to find a specific VSEC

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: <CAPcyv4ia_0Sn8paGi7y7JGNXQrbCoFhT7st2VOD=L_LKNEMOEg@mail.gmail.com>
Date:   Mon, 1 Feb 2021 15:24:45 -0800
From:   Dan Williams <dan.j.williams@...el.com>
To:     Bjorn Helgaas <helgaas@...nel.org>
Cc:     Gustavo Pimentel <Gustavo.Pimentel@...opsys.com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Linux PCI <linux-pci@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Vinod Koul <vkoul@...nel.org>, dmaengine@...r.kernel.org,
        Ben Widawsky <ben.widawsky@...el.com>
Subject: Re: [PATCH 04/15] PCI: Add pci_find_vsec_capability() to find a
 specific VSEC

[ add Ben ]

On Mon, Feb 1, 2021 at 2:39 PM Bjorn Helgaas <helgaas@...nel.org> wrote:
>
> [+cc Vinod, Dan, dmaengine]
>
> On Tue, Dec 15, 2020 at 06:30:13PM +0100, Gustavo Pimentel wrote:
> > Add pci_find_vsec_capability() that crawls through the device config
> > space searching in all Vendor-Specific Extended Capabilities for a
> > particular capability ID.
> >
> > Vendor-Specific Extended Capability (VSEC) is a PCIe capability (acts
> > like a wrapper) specified by PCI-SIG that allows the vendor to create
> > their own and specific capability in the device config space.
> >
> > Signed-off-by: Gustavo Pimentel <gustavo.pimentel@...opsys.com>
>
> If you fix the below, feel free to add my
>
> Acked-by: Bjorn Helgaas <bhelgaas@...gle.com>
>
> Otherwise, I can take it myself.  But that will be an ordering issue
> in the merge window if you merge the rest of the series via another
> tree.

I wonder if this warrants and if you'd be willing to stand up a stable
branch for just this commit for concerned parties to integrate,
because CXL development should adopt it as well.

>
> > ---
> >  drivers/pci/pci.c             | 29 +++++++++++++++++++++++++++++
> >  include/linux/pci.h           |  1 +
> >  include/uapi/linux/pci_regs.h |  5 +++++
> >  3 files changed, 35 insertions(+)
> >
> > diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> > index 6d4d5a2..235d0b2 100644
> > --- a/drivers/pci/pci.c
> > +++ b/drivers/pci/pci.c
> > @@ -623,6 +623,35 @@ u64 pci_get_dsn(struct pci_dev *dev)
> >  }
> >  EXPORT_SYMBOL_GPL(pci_get_dsn);
> >
> > +/**
> > + * pci_find_vsec_capability - Find a vendor-specific extended capability
> > + * @dev: PCI device to query
> > + * @cap: vendor-specific capability id code
>
> s/id/ID/
>
> > + *
> > + * Returns the address of the vendor-specific structure that matches the
> > + * requested capability id code within the device's PCI configuration space
>
> s/id/ID/
>
> > + * or 0 if it does not find a match.
> > + */
> > +int pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id)
> > +{
> > +     u32 header;
> > +     int vsec;
>
>   int vsec;
>   u32 header;
>
> since that's the order they're used.
>
> > +
> > +     vsec = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_VNDR);
> > +     while (vsec) {
> > +             if (pci_read_config_dword(dev, vsec + 0x4,
>
> s/0x4/PCI_VSEC_HDR/
>
> > +                                       &header) == PCIBIOS_SUCCESSFUL &&
> > +                 PCI_VSEC_CAP_ID(header) == vsec_cap_id)
> > +                     break;
>
>   return vsec;
>
> > +
> > +             vsec = pci_find_next_ext_capability(dev, vsec,
> > +                                                 PCI_EXT_CAP_ID_VNDR);
> > +     }
> > +
> > +     return vsec;
>
>   return 0;
>
> > +}
> > +EXPORT_SYMBOL_GPL(pci_find_vsec_capability);
> > +
> >  static int __pci_find_next_ht_cap(struct pci_dev *dev, int pos, int ht_cap)
> >  {
> >       int rc, ttl = PCI_FIND_CAP_TTL;
> > diff --git a/include/linux/pci.h b/include/linux/pci.h
> > index 22207a7..effecb0 100644
> > --- a/include/linux/pci.h
> > +++ b/include/linux/pci.h
> > @@ -1067,6 +1067,7 @@ int pci_find_capability(struct pci_dev *dev, int cap);
> >  int pci_find_next_capability(struct pci_dev *dev, u8 pos, int cap);
> >  int pci_find_ext_capability(struct pci_dev *dev, int cap);
> >  int pci_find_next_ext_capability(struct pci_dev *dev, int pos, int cap);
> > +int pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id);
> >  int pci_find_ht_capability(struct pci_dev *dev, int ht_cap);
> >  int pci_find_next_ht_capability(struct pci_dev *dev, int pos, int ht_cap);
> >  struct pci_bus *pci_find_next_bus(const struct pci_bus *from);
> > diff --git a/include/uapi/linux/pci_regs.h b/include/uapi/linux/pci_regs.h
> > index a95d55f..f5d17be 100644
> > --- a/include/uapi/linux/pci_regs.h
> > +++ b/include/uapi/linux/pci_regs.h
> > @@ -730,6 +730,11 @@
> >  #define PCI_EXT_CAP_DSN_SIZEOF       12
> >  #define PCI_EXT_CAP_MCAST_ENDPOINT_SIZEOF 40
> >
> > +/* Vendor-Specific Extended Capabilities */
> > +#define PCI_VSEC_CAP_ID(header)              (header & 0x0000ffff)
> > +#define PCI_VSEC_CAP_REV(header)     ((header >> 16) & 0xf)
> > +#define PCI_VSEC_CAP_LEN(header)     ((header >> 20) & 0xffc)
>
> Please put these next to the existing PCI_VSEC_HDR.
>
> Why does PCI_VSEC_CAP_LEN mask with 0xffc instead of 0xfff?  I don't
> see anything in the spec about VSEC Length having to be a multiple of
> 4 (PCIe r5.0, sec 7.9.5.2).
>
> But you don't use this anyway, so I'd just drop it (and
> PCI_VSEC_CAP_REV) altogether.
>
> >  /* Advanced Error Reporting */
> >  #define PCI_ERR_UNCOR_STATUS 4       /* Uncorrectable Error Status */
> >  #define  PCI_ERR_UNC_UND     0x00000001      /* Undefined */
> > --
> > 2.7.4
> >