So far we only provide num_k8_northbridges. This is required in
different areas (e.g. L3 cache index disable, GART). But not all AMD
CPUs provide a GART. Thus it is useful to split off the GART handling
from the generic caching of AMD northbridge misc devices.
Signed-off-by: NAndreas Herrmann <andreas.herrmann3@amd.com>
LKML-Reference: <20100917160254.GC4958@loge.amd.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Stanse found pci reference leaks in uli_agp_init and nforce3_agp_init
initialization functions.

The PCI devices are bridges, so it's not critical, but still worth fixing.
Signed-off-by: NJiri Slaby <jirislaby@gmail.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NDave Airlie <airlied@redhat.com>

The current initialisation code probes 'unsupported' AGP devices
simply by calling its own probe function.  It does not lock these
devices or even check whether another driver is already bound to
them.

We must use the device core to manage this.  So if the specific
device id table didn't match anything and agp_try_unsupported=1,
switch the device id table and call driver_attach() again.
Signed-off-by: NBen Hutchings <ben@decadent.org.uk>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Convert most AGP chipset to use scratch page as default entries.
This help avoiding GPU querying 0 address and trigger computer
fault. With KMS and memory manager we bind/unbind AGP memory
constantly and it seems that some GPU are still doing AGP
traffic even after GPU report being idle with the memory segment.

Tested (radeon GPU KMS + Xorg + compiz + glxgears + quake3) on :
- SIS 1039:0001 & 1039:0003
- Intel 865 8086:2571

Compile tested for other bridges

V2 enable scratch page on uninorth
V3 fix unbound check in uninorth insert memory (Michel Dänzer)
V4 rebase on top of drm-next branch with the lastest intel AGP
   changeset (stable should use version V3 of the patch)
Signed-off-by: NJerome Glisse <jglisse@redhat.com>
Signed-off-by: NMichel Dänzer <michel@daenzer.net>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This fixes the regression introduced by commit
42590a75 ("x86/agp: Fix
agp_amd64_init and agp_amd64_cleanup").

The commit 61684cea fixed the
above regression but it's not enough. When amd64-agp is built as
a module, AGP isn't initialized, iommu is initialized, all the
aperture is owned by the iommu.
Reported-by: NMarin Mitov <mitov@issp.bas.bg>
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Tested-by: NMarin Mitov <mitov@issp.bas.bg>
LKML-Reference: <20100204090802S.fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

This fixes the regression introduced by commit
42590a75 ("x86/agp: Fix
agp_amd64_init and agp_amd64_cleanup").

The above commit changes agp_amd64_init() not to do anything if
gart_iommu_aperture is not zero.

If GART iommu calls agp_amd64_init(), we need to skip
agp_amd64_init() when it's called later via module_init.

The problem is that gart_iommu_init() calls agp_amd64_init()
with not zero gart_iommu_aperture so agp_amd64_init() is never
initialized.

When gart_iommu_init() calls agp_amd64_init(), agp should be
always initialized.
Reported-by: NMarin Mitov <mitov@issp.bas.bg>
Reported-by: NJohannes Hirte <johannes.hirte@fem.tu-ilmenau.de>
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Tested-by: NMarin Mitov <mitov@issp.bas.bg>
Tested-by: NKevin Winchester <kjwinchester@gmail.com>
Cc: davej@redhat.com
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <20100125141006O.fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

This fixes the regression introduced by the commit
f405d2c0.

The above commit fixes the following issue:

  http://marc.info/?l=linux-kernel&m=126192729110083&w=2

However, it doesn't work properly when you remove and insert the
agp_amd64 module again.

agp_amd64_init() and agp_amd64_cleanup should be called only
when gart_iommu is not called earlier (that is, the GART IOMMU
is not enabled). We need to use 'gart_iommu_aperture' to see if
GART IOMMU is enabled or not.
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: mitov@issp.bas.bg
Cc: davej@redhat.com
LKML-Reference: <20100104161603L.fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

with CONFIG_GART_IOMMU enabled drivers/char/agp/amd64-agp.c has:

 #ifndef CONFIG_GART_IOMMU
 module_init(agp_amd64_init);
 module_exit(agp_amd64_cleanup);
 #endif

agp_amd64_init() was called via gart_iommu_init with
CONFIG_GART_IOMMU=y agp_amd64_init() was called via module_init
with CONFIG_GART_IOMMU=n

The commit 75f1cdf1 changes the
x86 dma initialization routine: gart_iommu_init() is called only
when GART IOMMU is detected. So when GART IOMMU isn't detected,
agp_amd64_init isn't called.

Marin Mitov reported this issue:

 http://marc.info/?l=linux-kernel&m=126192729110083&w=2

With this patch, agp_amd64_init() is always called via
module_init (the above ifndef is removed). If agp_amd64_init()
is called via gart_iommu_init() earlier, agp_amd64_init()
finishes without doing anything (when it is called via
module_init).
Reported-by: NMarin Mitov <mitov@issp.bas.bg>
Tested-by: NMarin Mitov <mitov@issp.bas.bg>
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: davej@redhat.com
LKML-Reference: <20091228181118C.fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

There seems to be no reason for these -- they're a 1:1 mapping on all
platforms.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

In commit 07613ba2 ("agp: switch AGP to use page array instead of
unsigned long array") we switched the mask_memory() method to take a
'struct page *' instead of an address. This is painful, because in some
cases it has to be an IOMMU-mapped virtual bus address (in fact,
shouldn't it _always_ be a dma_addr_t returned from pci_map_xxx(), and
we just happen to get lucky most of the time?)
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>

This switches AGP to use an array of pages for tracking the
pages allocated to the GART. This should enable GEM on PAE to work
a lot better as we can pass highmem pages to the PAT code and it will
do the right thing with them.
Signed-off-by: NDave Airlie <airlied@redhat.com>

Impact: fix bug to make agp work with dri

Jeffrey reported that dri does work with 64bit, but doesn't work with
32bit it turns out NB aperture is 32M, aperture on agp is 128M

64bit is using 64M for vaidation for 64 iommu/gart 32bit is only using
32M..., and will not update the nb aperture.

So try to compare nb apterture and agp apterture before leaving not
touch nb aperture.
Reported-by: NJeffrey Trull <jetrull@sbcglobal.net>
Tested-by: NJeffrey Trull <jetrull@sbcglobal.net>
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Acked-by: NDave Airlie <airlied@linux.ie>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Update assorted email addresses and related info to point
to a single current, valid address.

additionally
- trivial CREDITS entry updates. (Not that this file means much any more)
- remove arjans dead redhat.com address from powernow driver
Signed-off-by: NDave Jones <davej@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The pageattr-array patch that you currently have in tip/master only
enables it for intel-agp, not the others. The attached enables it for
all drivers currently directly using agp_generic_alloc_page() and
agp_generic_destroy_page() (ocal driver is amd-k7-agp).

The new agp_generic_alloc_pages() interface uses the also new
pageattr array interface API. This makes all AGP drivers that
up to now used generic_{alloc,destroy}_page() use it.
Signed-off-by: NRene Herman <rene.herman@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Convert printks to use dev_printk().
Signed-off-by: NBjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NDave Airlie <airlied@redhat.com>

I think the intent was that if no bridges matched agp_amd64_pci_table[],
we would fall back to checking for any bridge with the AGP capability.
But in the current code, we execute the fallback path only when
pci_register_driver() itself fails, which is unrelated to whether any
matching devices were found.

This patch counts the AGP bridges found in the probe() method and executes
the fallback path when none is found.
Signed-off-by: NBjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Use boolean in AGP instead of having own TRUE/FALSE

--
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Factor-out common aperture_valid code.
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Cleanup gart handling on amd64 a bit: move common code into
enable_gart_translation , and use symbolic register names where
appropriate.
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

some systems are using 32M for gart and agp when memory is less than 4G.
Kernel will reject and try to allcate another 64M that is not needed,
and we will waste 64M of perfectly good RAM.

this patch adds a workaround by checking aper_base/order between NB and
agp bridge. If they are the same, and memory size is less than 4G, it
will allow it.
Signed-off-by: NYinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

while looking at Rafael J. Wysocki's system boot log,

I found a funny printout:

	Node 0: aperture @ de000000 size 32 MB
	Aperture too small (32 MB)
	AGP bridge at 00:04:00
	Aperture from AGP @ de000000 size 4096 MB (APSIZE 0)
	Aperture too small (0 MB)
	Your BIOS doesn't leave a aperture memory hole
	Please enable the IOMMU option in the BIOS setup
	This costs you 64 MB of RAM
	Mapping aperture over 65536 KB of RAM @ 4000000

	...

	agpgart: Detected AGP bridge 20
	agpgart: Aperture pointing to RAM
	agpgart: Aperture from AGP @ de000000 size 4096 MB
	agpgart: Aperture too small (0 MB)
	agpgart: No usable aperture found.
	agpgart: Consider rebooting with iommu=memaper=2 to get a good aperture.

it means BIOS allocated the correct gart on the NB and AGP bridge, but
because a bug in the silicon (the agp bridge reports the wrong order,
it wants 4G instead) the kernel will reject that allocation.

Also, because the size is only 32MB, and we try to get another 64M for gart,
late fix_northbridge can not revert that change because it still reads
the wrong size from agp bridge.

So try to double check the order value from the agp bridge, before calling
aperture_valid().

[ mingo@elte.hu: 32-bit fix. ]
Signed-off-by: NYinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Move symbolic constants into gart.h, and use them instead of hardcoded
constant.
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

This patch renames the IOMMU config option to GART_IOMMU because in fact it
means the GART and not general support for an IOMMU on x86.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Acked-by: NMuli Ben-Yehuda <muli@il.ibm.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>

Instead of all drivers reading pci config space to get the revision
ID, they can now use the pci_device->revision member.

This exposes some issues where drivers where reading a word or a dword
for the revision number, and adding useless error-handling around the
read. Some drivers even just read it for no purpose of all.

In devices where the revision ID is being copied over and used in what
appears to be the equivalent of hotpath, I have left the copy code
and the cached copy as not to influence the driver's performance.

Compile tested with make all{yes,mod}config on x86_64 and i386.
Signed-off-by: NAuke Kok <auke-jan.h.kok@intel.com>
Acked-by: NDave Jones <davej@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

I'm using a custom BIOS to configure the northbridge GART at address
0x80000000, size 2G.  Linux complains:

"Aperture from northbridge cpu 0 beyond 4GB. Ignoring."

I think there's an off-by-two error in arch/x86_64/kernel/aperture.c:

AK: use correct types for i386
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Under CONFIG_DISCONTIGMEM, assuming that a !pfn_valid() implies all
subsequent pfn-s are also invalid is wrong. Thus replace this by
explicitly checking against the E820 map.

AK: make e820 on x86-64 not initdata
Signed-off-by: NJan Beulich <jbeulich@novell.com>
Signed-off-by: NAndi Kleen <ak@suse.de>
Acked-by: NMark Langsdorf <mark.langsdorf@amd.com>

Make agp_bridge_driver->aperture_sizes and ->masks const.
Also agp_bridge_data->driver
Signed-off-by: NDave Jones <davej@redhat.com>

This patch allows drm to populate an agpgart structure with pages of its own.
It's needed for the new drm memory manager which dynamically flips pages in and out of AGP.

The patch modifies the generic functions as well as the intel agp driver. The intel drm driver is
currently the only one supporting the new memory manager.

Other agp drivers may need some minor fixing up once they have a corresponding memory manager enabled drm driver.

AGP memory types >= AGP_USER_TYPES are not populated by the agpgart driver, but the drm is expected
to do that, as well as taking care of cache- and tlb flushing when needed.

It's not possible to request these types from user space using agpgart ioctls.

The Intel driver also gets a new memory type for pages that can be bound cached to the intel GTT.
Signed-off-by: NThomas Hellstrom <thomas@tungstengraphics.com>
Signed-off-by: NDave Jones <davej@redhat.com>

Culled from the VIA codedrop.
Also fixes up one ID used in amd64-agp to use the
VIA part number instead of the board name in its ID.
Signed-off-by: NDave Jones <davej@redhat.com>

Signed-off-by: NDave Jones <davej@redhat.com>

When CONFIG_HOTPLUG=n, agp_amd64_resume() calls nforce3_agp_init(),
which is __devinit == __init, so has been discarded and is not
usable for resume.

WARNING: drivers/char/agp/amd64-agp.o - Section mismatch: reference to .init.text: from .text between 'agp_amd64_resume' (at offset 0x249) and 'amd64_tlbflush'
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NDave Jones <davej@redhat.com>

When CONFIG_HOTPLUG=n, agp_amd64_resume() calls nforce3_agp_init(), which is
__devinit == __init, so has been discarded and is not usable for resume.

WARNING: drivers/char/agp/amd64-agp.o - Section mismatch: reference to .init.text: from .text between 'agp_amd64_resume' (at offset 0x249) and 'amd64_tlbflush'
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NDave Jones <davej@redhat.com>

Signed-off-by: NJörn Engel <joern@wohnheim.fh-wedel.de>
Signed-off-by: NAdrian Bunk <bunk@stusta.de>

 - Rename the GART_IOMMU option to IOMMU to make clear it's not
   just for AMD
 - Rewrite the help text to better emphatise this fact
 - Make it an embedded option because too many people get it wrong.

To my astonishment I discovered the aacraid driver tests this
symbol directly. This looks quite broken to me - it's an internal
implementation detail of the PCI DMA API. Can the maintainer
please clarify what this test was intended to do?

Cc: linux-scsi@vger.kernel.org
Cc: alan@redhat.com
Cc: markh@osdl.org
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

 - Factor out the duplicated access/cache code into a single file
   * Shared between i386/x86-64.
 - Share flush code between AGP and IOMMU
   * Fix a bug: AGP didn't wait for end of flush before
 - Drop 8 northbridges limit and allocate dynamically
 - Add lock to serialize AGP and IOMMU GART flushes
 - Add PCI ID for next AMD northbridge
 - Random related cleanups

The old K8 NUMA discovery code is unchanged. New systems
should all use SRAT for this.

Cc: "Navin Boppuri" <navin.boppuri@newisys.com>
Cc: Dave Jones <davej@redhat.com>
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

pci_register_driver() never returns a positive number.
Signed-off-by: NDave Jones <davej@redhat.com>

kernel.org bugzilla #6206

Based on patch from Serge Belyshev <belyshev@depni.sinp.msu.ru>
Signed-off-by: NDave Jones <davej@redhat.com>

[description by AK]
Made a cut'n'paste error when adding the entry for the ALI M1695
AGP bridge and added a second entry for the 1689

Cc: Dave Jones <davej@redhat.com>
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

[description by AK]
Made a cut'n'paste error when adding the entry for the ALI M1695
AGP bridge and added a second entry for the 1689
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NDave Jones <davej@redhat.com>