lkml.org 
[lkml]   [2015]   [Apr]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [PATCH] x86, aperture: Check for GART before accessing GART registers

    * Borislav Petkov <bp@suse.de> wrote:

    > On Thu, Apr 02, 2015 at 12:04:21PM -0500, Aravind Gopalakrishnan wrote:
    > > >No need - I can amend the local copy I have here.
    >
    > Here's what I did:
    >
    > ---
    > From: Aravind Gopalakrishnan <Aravind.Gopalakrishnan@amd.com>
    > Date: Wed, 1 Apr 2015 09:32:08 -0500
    > Subject: [PATCH] x86/gart: Check for GART support before accessing GART
    > registers
    > MIME-Version: 1.0
    > Content-Type: text/plain; charset=UTF-8
    > Content-Transfer-Encoding: 8bit
    >
    > GART registers are not present in newer AMD processors (Fam15h,
    > Model 10h and later). So, avoid accesses to GART registers in PCI
    > config space by returning early in early_gart_iommu_check() and
    > gart_iommu_hole_init() if GART is not available.
    >
    > Current code doesn't break on existing processors but there are some
    > side effects:
    >
    > We get bogus AGP aperture messages which are simply noise on
    > GART-less processors:
    >
    > AGP: Node 0: aperture [bus addr 0x00000000-0x01ffffff] (32MB)
    > AGP: Your BIOS doesn't leave a aperture memory hole

    hah, someone should fix the typo in that message:

    s/a a/a

    > AGP: Please enable the IOMMU option in the BIOS setup
    > AGP: This costs you 64MB of RAM
    > AGP: Mapping aperture over RAM [mem 0xd4000000-0xd7ffffff]
    >
    > We can avoid calling allocate_aperture() and would not have to
    > wastefully reserve 64MB of RAM with memblock_reserve(). Also, we can
    > avoid having to loop through all PCI buses and devices twice, searching
    > for a non-existent AGP bridge if we bail out early.
    >
    > Refactor the family check used in amd_nb into an inline function so we

    s/amd_nb/amd_nb.c

    >
    > +static inline bool amd_gart_present(void)
    > +{
    > + /* GART present only on Fam15h upto model 0fh */

    s/h u/h, u

    > + if (amd_gart_present())
    > amd_northbridges.flags |= AMD_NB_GART;
    >
    > /*
    > diff --git a/arch/x86/kernel/aperture_64.c b/arch/x86/kernel/aperture_64.c
    > index 76164e173a24..1cb170b06853 100644
    > --- a/arch/x86/kernel/aperture_64.c
    > +++ b/arch/x86/kernel/aperture_64.c
    > @@ -262,7 +262,7 @@ void __init early_gart_iommu_check(void)
    > u64 aper_base = 0, last_aper_base = 0;
    > int aper_enabled = 0, last_aper_enabled = 0, last_valid = 0;
    >
    > - if (!early_pci_allowed())
    > + if (!early_pci_allowed() || !amd_gart_present())
    > return;
    >
    > /* This is mostly duplicate of iommu_hole_init */
    > @@ -356,7 +356,7 @@ int __init gart_iommu_hole_init(void)
    > int i, node;
    >
    > if (gart_iommu_aperture_disabled || !fix_aperture ||
    > - !early_pci_allowed())
    > + !early_pci_allowed() || !amd_gart_present())
    > return -ENODEV;

    So what happens if !early_pci_allowed() but the GART is present? We'll
    set amd_northbridges.flags |= AMD_NB_GART, but won't run any of the
    setup code in aperture_64.c, right? Is that a valid setup?

    Thanks,

    Ingo


    \
     
     \ /
      Last update: 2015-04-02 21:01    [W:2.692 / U:0.464 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site