Take two passes to evict everything whilst searching for sufficient free
space to bind the batchbuffer. After searching for sufficient free space
using LRU eviction, evict everything that is purgeable and try again.
Only then if there is insufficient free space (or the GTT is too badly
fragmented) evict everything from the aperture and try one last time.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Accessing the uninitialised obj->pages instead of the local page lead to
an OOPs.
Reported-by: Xavier Chantry <chantry.xavier@gmail.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

My Sandybridge only reports 0 for the ring buffer registers, causing it
to hang as soon as we exhaust the available ring. As a workaround, take
advantage of our huge ring buffers and use the auto-reporting mechanism
to update the status page with the HEAD location every 64 KiB.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

After switching the MMIO registers to use pci_iomap, remember to dispose
of the mapping with pci_iounmap (for symmetry).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

It should be idle on cleanup anyway...
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

So long as we adhere to the fence registers rules for alignment and no
overlaps (including with unfenced accesses to linear memory) and account
for the tiled access in our size allocation, we do not have to allocate
the full fenced region for the object. This allows us to fight the bloat
tiling imposed on pre-i965 chipsets and frees up RAM for real use. [Inside
the GTT we still suffer the additional alignment constraints, so it doesn't
magic allow us to render larger scenes without stalls -- we need the
expanded GTT and fence pipelining to overcome those...]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Also spotted by Dan Carpenter.

obj->pin_count is unsigned so the BUG_ON(obj->pin_count<0) will never
trigger.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

The error code is only expected during the actual pruning and not during
the first measurement (nr_to_scan == 0) pass.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

In order to force a page-fault on a GTT mapping after we start using it
from the GPU and so enforce correct CPU/GPU synchronisation, we need to
invalidate the mapping.

Pointed out by Owain G. Ainsworth.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

By using read_cache_page() for individual pages during pwrite/pread we
can eliminate an unnecessary large allocation (and immediate free) of
obj->pages. Also this eliminates any potential nesting of get/put pages,
simplifying the code and preparing the path for greater things.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Since we rarely use the mmap_offset and it is easily computable from the
obj->map_list.hash, remove it.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Eliminate the racy device unload by embedding a shrinker into each
device. Smaller, simpler code.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Play safe and use the common routines which take care of the cachability
of the memory when setting up the iomapping for the PCI registers.
Whilst they should be cacheable for the current generations, actually
honouring what the device requires is a better long term strategy.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

This holds error state from the main graphics arbiter mainly involving
the DMA engine and address translation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

This is the same value as before, but it just makes the code slightly
more readable to use the local variable than converting the aperture
size into bytes every time.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

More precisely: For those that _need_ to be mappable. Also add two
BUG_ONs in fault and pin to check the consistency of the mappable
flag.

Changes in v2:
- Add tracking of gtt mappable space (to notice mappable/unmappable
  balancing issues).
- Improve the mappable working set tracking by tracking fault and pin
  separately.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

This way we can make some more educated guesses as to why exactly
we can't use 2G apertures to their full potential ;)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

On VT-d supporting platforms the GGTT is allocated in a stolen mem
section separate from graphcis stolen mem. The GMCH register contains
a bitfield specifying the size of that region. Docs suggest that this
region can only be used for GGTT and PPGTT. Hence ensure that the
PPGTT is disabled and use the complete area for the GGTT.

Unfortunately the graphics core on G33/Pineview can't cope with really
large GTTs and the BIOS usually enables the maximum of 512MB. So
don't bother with maximizing the GTT on these platforms.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

... and switch to a more classical store-reg-on-suspend, restore-on-resume
way of doing things. Obviously this is just preparation for the future,
the code is not there at all, yet.

This is needed because the next patch adjusts this register and everything
in it (not just the pagetable address) needs to be restored on resume.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

At least the part that's currently enabled by the BIOS.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

In i915_gem_object_pin obviously unbind only if mappable is true.

This is the last part to enable gtt_mappable_end != gtt_size, which
the next patch will do.

v2: Fences on g33/pineview only work in the mappable part of the
gtt.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Like before add a parameter mappable (also to gem_object_pin) and
set it depending upon the context. Only bos that are brought into
the gtt due to an execbuffer call can be put into the unmappable
part of the gtt, everything else (especially pinned objects) need
to be put into the mappable part of the gtt.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Add a mappable parameter to i915_gem_evict_something to distinguish
the two cases (non-restricted vs. mappable gtt allocations). No
functional changes because the mappable limit is set to the end of
the gtt currently.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

The render ring is not alone any more! And the other rings are just as
troublesome...
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Currently, we believe the GPU is idle if just the RENDER ring is idle.
This is obviously wrong if we only using either the BLT or the BSD
rings and so masking genuine hangs.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

In the event that we fail to dispatch the execbuffer, for example if
there is insufficient space on the ring, we were leaving the objects in
an inconsistent state. Notably they were marked as being in the GPU
write domain, but were not added to the ring or any list. This would
lead to inevitable oops:

[ 1010.522940] [drm:i915_gem_do_execbuffer] *ERROR* dispatch failed -16
[ 1010.523055] BUG: unable to handle kernel NULL pointer dereference at
0000000000000088
[ 1010.523097] IP: [<ffffffff8122d006>] i915_gem_flush_ring+0x26/0x140
[ 1010.523120] PGD 14cf2f067 PUD 14ce04067 PMD 0
[ 1010.523140] Oops: 0000 [#1] SMP
[ 1010.523154] last sysfs file: /sys/devices/virtual/vc/vcsa2/uevent
[ 1010.523173] CPU 0
[ 1010.523183] Pid: 716, comm: X Not tainted 2.6.36+ #34 LosLunas
CRB/SandyBridge Platform
[ 1010.523206] RIP: 0010:[<ffffffff8122d006>]  [<ffffffff8122d006>]
i915_gem_flush_ring+0x26/0x140
[ 1010.523233] RSP: 0018:ffff88014bf97cd8  EFLAGS: 00010296
[ 1010.523249] RAX: ffff88014e2d1808 RBX: 0000000000000000 RCX: 0000000000000000
[ 1010.523270] RDX: 0000000000000002 RSI: 0000000000000000 RDI: 0000000000000000
[ 1010.523290] RBP: ffff88014e2d1000 R08: 0000000000000002 R09: 00000000400c645f
[ 1010.523311] R10: 0000000000000001 R11: 0000000000000246 R12: 0000000000000002
[ 1010.523331] R13: ffff88014e29a000 R14: 00000000000000c8 R15: ffffffff8162eb28
[ 1010.523352] FS:  00007fc62379d700(0000) GS:ffff88001fc00000(0000) knlGS:0000000000000000
[ 1010.523375] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1010.523392] CR2: 0000000000000088 CR3: 000000014bf87000 CR4: 00000000000406f0
[ 1010.523412] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 1010.523433] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[ 1010.523454] Process X (pid: 716, threadinfo ffff88014bf96000, task ffff88014cc1ee40)
[ 1010.523475] Stack:
[ 1010.523483]  ffff88014d5199c0 0000000000000200 0000000000000000 ffff88014bcc6400
[ 1010.523509] <0> 0000000000000000 0000000000000001 ffff88014e29a000 ffff88014bcc6400
[ 1010.523537] <0> ffffffff8162eb28 ffffffff8122faa8 ffff88014e29a000 ffff88014bcc6400
[ 1010.523568] Call Trace:
[ 1010.523578]  [<ffffffff8122faa8>] ?  i915_gem_object_flush_gpu_write_domain+0x48/0x80
[ 1010.523601]  [<ffffffff8122fb8e>] ?  i915_gem_object_set_to_gtt_domain+0x2e/0xb0
[ 1010.523623]  [<ffffffff8123113b>] ?  i915_gem_set_domain_ioctl+0xdb/0x1f0
[ 1010.523644]  [<ffffffff8120a3f1>] ? drm_ioctl+0x3d1/0x460
[ 1010.523660]  [<ffffffff81231060>] ?  i915_gem_set_domain_ioctl+0x0/0x1f0
[ 1010.523682]  [<ffffffff81092618>] ?  vma_prio_tree_insert+0x28/0x120
[ 1010.523701]  [<ffffffff8109f379>] ? vma_link+0x99/0xf0
[ 1010.523717]  [<ffffffff810a111d>] ? mmap_region+0x1ed/0x4f0
[ 1010.523734]  [<ffffffff810c306f>] ? do_vfs_ioctl+0x9f/0x580
[ 1010.523750]  [<ffffffff810c3599>] ? sys_ioctl+0x49/0x80
[ 1010.523767]  [<ffffffff810022eb>] ?  system_call_fastpath+0x16/0x1b
[ 1010.523785] Code: 00 00 00 00 00 41 57 89 ce 41 56 41 55 41 54 45 89 c4 55 48 89 fd 53 48 89 d3 44 89 c2 48 89 df 4c 8d b3 c8 00 00 00 48 83 ec 18 <ff> 93 88 00 00 00 48 8b 83 c8 00 00 00 4c 8b bd 30 03 00 00 48
[ 1010.523946] RIP  [<ffffffff8122d006>] i915_gem_flush_ring+0x26/0x140
[ 1010.523966]  RSP <ffff88014bf97cd8>
[ 1010.523977] CR2: 0000000000000088
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Preparing the ringbuffer for adding new commands can fail (a timeout
whilst waiting for the GPU to catch up and free some space). So check
for any potential error before overwriting HEAD with new commands, and
propagate that error back to the user where possible.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

The ringbuffer keeps a pointer to the parent device, so we can use that
instead of passing around the pointer on the stack.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>