Commit Graph

2900 Commits

Author SHA1 Message Date
Ben Skeggs 330c5988ee drm/nouveau: import initial work on vbios performance table parsing
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:27:00 +10:00
Ben Skeggs 4709bff02a drm/nouveau: make bios code easier to use externally
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:26:56 +10:00
Ben Skeggs 855a95e4fc drm/nouveau: make the behaviour of get_pll_limits() consistent
This replaces all the pll_types definitions for ones that match the types
used in the tables in recent VBIOS versions.

get_pll_limits() will now accept either type or register value as input
across all limits table versions, and will store the actual register ID
that a PLL type refers to in the returned structure.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:26:49 +10:00
Ben Skeggs f9aafdd30e drm/nv50: fix 100c90 write on nva3
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:26:41 +10:00
Ben Skeggs b6fd780791 drm/nouveau: zero dummy page
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:26:35 +10:00
Ben Skeggs 17b20348ea drm/nv50: mark PCIEGART pages non-present rather than using dummy page
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:26:31 +10:00
Ben Skeggs 24b102d348 drm/nouveau: we can't free ACPI EDID, so make a copy that we can
The rest of the connector code assumes we can kfree() the EDID pointer.
This causes things to blow up with the ACPI EDID pointer we get
passed.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:26:24 +10:00
Ben Skeggs e071f8cd71 drm/nouveau: handle fifo pusher errors better
The most important part of this change is that we now instruct PFIFO to
drop all pending fetches, rather than attempting to skip a single dword
and hope that things would magically sort themselves out - they usually
don't, and we end up with PFIFO being completely hung.

This commit also adds somewhat more useful logging when these exceptions
occur.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:26:02 +10:00
Ben Skeggs e1429b4c3c drm/nouveau: better handling of unmappable vram
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:55 +10:00
Francisco Jerez c16c570762 drm/nouveau: Try to fetch an EDID from OF if DDC fails.
More Apple brain damage, it fixes the modesetting failure on an eMac
G4 (fdo bug 29810).

Reported-by: Zoltan Varnagy <doi@freemail.hu>
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:48 +10:00
Francisco Jerez 9f56b1265d drm/nouveau: Simplify tile region handling.
Instead of emptying the caches to avoid a race with the PFIFO puller,
go straight ahead and try to recover from it when it happens. Also,
kill pfifo->cache_flush and tile->lock, we don't need them anymore.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:44 +10:00
Francisco Jerez 4b5c152a79 drm/nouveau: Remove implicit argument from nv_wait().
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:36 +10:00
Francisco Jerez 35fd5b42f0 drm/nouveau: PRAMIN is available from the start on pre-nv50.
This makes sure that RAMHT is cleared correctly on start up.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:28 +10:00
Francisco Jerez de5899bdac drm/nouveau: Add module parameter to override the default AGP rate.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:16 +10:00
Francisco Jerez 71d0618661 drm/nouveau: Don't enable AGP FW on nv18.
FW seems to be broken on nv18, it causes random lockups and breaks
suspend/resume even with the blob.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:10 +10:00
Francisco Jerez 3bc14b4dca drm/nouveau: Don't remove ramht entries from the neighboring channels.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:25:03 +10:00
Francisco Jerez f243423759 drm/nouveau: Break some long lines in the TV-out code.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:24:59 +10:00
Ben Skeggs 8597a1ba37 drm/nv50: fix SOR count for early chipsets
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:24:52 +10:00
Ben Skeggs 1da265662d drm/nv50: report BAR access faults
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:24:46 +10:00
Ben Skeggs d96773e7b6 drm/nv50: move vm trap to nv50_fb.c
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:24:38 +10:00
Marcin Kościelnicki e69b441882 drm/nv50: demagic grctx, and add NVAF support
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Marcin Kościelnicki <koriakin@0x04.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:24:00 +10:00
Ben Skeggs 2941482ead drm/nouveau: protect ramht_find() from oopsing if on channel without ramht
This doesn't actually happen now, but there's a test case for an earlier
kernel where a GPU error is signalled on one of nv50's fake channels, and
the ramht lookup by the IRQ handler triggered an oops.

This adds a check for RAMHT's existance on a channel before looking up
an object handle.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:23:54 +10:00
Ben Skeggs 185abeccab drm/nouveau: remove nouveau_gpuobj_late_takedown
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:23:46 +10:00
Ben Skeggs e05d7eaeba drm/nouveau: protect gpuobj list + global instmem heap with spinlock
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:23:41 +10:00
Ben Skeggs eb9bcbdc45 drm/nouveau: fix gpuobj refcount to use atomics
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:23:34 +10:00
Ben Skeggs dac7900804 drm/nouveau: add spinlock around ramht modifications
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:23:30 +10:00
Ben Skeggs e05c5a317e drm/nouveau: tidy ram{ht,fc,ro} a bit
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:23:22 +10:00
Ben Skeggs fbd2895e45 drm/nouveau: rework init ordering so nv50_instmem.c can be less bad
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:22:42 +10:00
Ben Skeggs 6c3d7ef25e drm/nv50: calculate vram reordering block size
Will be used at a later point when we plug in an alternative VRAM memory
manager for GeForce 8+ boards.

Based on pscnv code to do the same.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Marcin Kościelnicki <koriakin@0x04.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:21:55 +10:00
Ben Skeggs 5125bfd886 drm/nv50: allow gpuobjs that aren't mapped into aperture
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:20:28 +10:00
Ben Skeggs 43efc9ce25 drm/nouveau: simplify fake gpu objects
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:20:23 +10:00
Ben Skeggs a8eaebc6c5 drm/nouveau: remove nouveau_gpuobj_ref completely, replace with sanity
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:20:14 +10:00
Ben Skeggs de3a6c0a3b drm/nouveau: rebase per-channel pramin heap offsets to 0
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:20:07 +10:00
Ben Skeggs b3beb167af drm/nouveau: modify object accessors, offset in bytes rather than dwords
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:20:00 +10:00
Ben Skeggs 479dcaea09 drm/nouveau: move ramht code out of nouveau_object.c, nothing to see here
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:19:55 +10:00
Ben Skeggs 2a7fdb2bc1 drm/nouveau: have nv_mask return original register value
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:19:15 +10:00
Francisco Jerez 217275d03d drm/nv40: Try to set up CRE_LCD even if it has unknown bits set.
They don't seem to do anything useful, and we really want to program
CRE_LCD if we aren't lucky enough to find the right CRTC binding
already set.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:19:09 +10:00
Francisco Jerez 7314dec95c drm/nv17-nv4x: Fix analog load detection false positive on rare occasions.
On some boards the residual current DAC outputs can draw when they're
disconnected can be high enough to give a false load detection
positive (I've only seen it in the S-video luma output of some cards,
but just to be sure). The output line capacitance is limited and
sampling twice should fix it reliably.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:19:03 +10:00
Ben Skeggs 9d59e8a1fc drm/nouveau: require explicit unmap of kmapped bos
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:18:31 +10:00
Ben Skeggs e694438d2f drm/nouveau: remove second map of notifier bo
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:18:27 +10:00
Ben Skeggs b8a6a8045d drm/nouveau: move check for no-op bo move before memcpy fallback
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:18:20 +10:00
Ben Skeggs f1ab0cc9ba drm/nv50: add new accelerated bo move funtion
Hopefully this one will be better able to cope with moving tiled buffers
around without getting them all scrambled as a result.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:18:16 +10:00
Francisco Jerez 4295f188e8 drm/nv20: Use the nv30 CRTC bandwidth calculation code.
nv2x CRTC FIFOs are as large as in nv3x (4kB it seems), and the FIFO
control registers have the same layout: we can make them share the
same implementation.

Previously we were using the nv1x code, but the calculated FIFO
watermarks are usually too low for nv2x and they cause horrible
scanout artifacts. They've gone unnoticed until now because we've been
leaving one of the bandwidth regs uninitialized (CRE 47, which
contains the most significant bits of FFLWM), so everything seemed to
work fine except in some cases after a cold boot, depending on the
memory bandwidth and pixel clocks used.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:17:59 +10:00
Francisco Jerez 316f60a120 drm/nouveau: Fix suspend on some nv4x AGP cards.
On some nv4x cards (specifically, the ones that use an internal
PCIE->AGP bridge) the AGP controller state isn't preserved after a
suspend/resume cycle, and the AGP control registers have moved from
0x18xx to 0x100xx, so the FW check in nouveau_mem_reset_agp() doesn't
quite work. Check "dev->agp->mode" instead.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
2010-09-24 16:17:52 +10:00
Dave Airlie 4a445f291a Merge remote branch 'origin/master' of /home/airlied/kernel//linux-2.6 into drm-core-next 2010-09-24 15:37:33 +10:00
Chris Wilson db8c076b92 Merge remote branch 'linus' into drm-intel-fixes 2010-09-21 09:14:55 +01:00
Chris Wilson af62610313 drm/i915: Hold a reference to the object whilst unbinding the eviction list
During heavy aperture thrashing we may be forced to wait upon several active
objects during eviction. The active list may be the last reference to
these objects and so the action of waiting upon one of them may cause
another to be freed (and itself unbound). To prevent the object
disappearing underneath us, we need to acquire and hold a reference
whilst unbinding.

This should fix the reported page refcount OOPS:

kernel BUG at drivers/gpu/drm/i915/i915_gem.c:1444!
...
RIP: 0010:[<ffffffffa0093026>]  [<ffffffffa0093026>] i915_gem_object_put_pages+0x25/0xf5 [i915]
Call Trace:
 [<ffffffffa009481d>] i915_gem_object_unbind+0xc5/0x1a7 [i915]
 [<ffffffffa0098ab2>] i915_gem_evict_something+0x3bd/0x409 [i915]
 [<ffffffffa0027923>] ? drm_gem_object_lookup+0x27/0x57 [drm]
 [<ffffffffa0093bc3>] i915_gem_object_bind_to_gtt+0x1d3/0x279 [i915]
 [<ffffffffa0095b30>] i915_gem_object_pin+0xa3/0x146 [i915]
 [<ffffffffa0027948>] ? drm_gem_object_lookup+0x4c/0x57 [drm]
 [<ffffffffa00961bc>] i915_gem_do_execbuffer+0x50d/0xe32 [i915]

Reported-by: Shawn Starr <shawn.starr@rogers.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=18902
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2010-09-20 20:50:00 +01:00
Chris Wilson 41a5142891 drm/i915,agp/intel: Add second set of PCI-IDs for B43
There is a second revision of B43 (a desktop gen4 part) floating around,
functionally equivalent to the original B43, so simply add the new
PCI-IDs.

Bugzilla: https://bugs.freedesktop.org/show_bugs.cgi?id=30221
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
2010-09-17 08:22:30 +01:00
Chris Wilson e259befd90 drm/i915: Fix Sandybridge fence registers
With 5 places to update when adding handling for fence registers, it is
easy to overlook one or two. Correct that oversight, but fence
management should be improved before a new set of registers is added.

Bugzilla: https://bugs.freedesktop.org/show_bug?id=30199
Original patch by: Yuanhan Liu <yuanhan.liu@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
2010-09-17 08:18:30 +01:00
Chris Wilson 79077319d7 drm/i915/crt: Downgrade warnings for hotplug failures
These are not fatal errors, so do not alarm the user by filling the
logs with *** ERROR ***. Especially as we know that g4x CRT detection
is a little sticky.

On the one hand the errors are valid since they are warning us of a
stall -- we poll the register whilst holding the mode lock so not even
the mouse will update. On the other hand, those stalls were already present
yet nobody complained.

Reported-by: Andi Kleen <andi@firstfloor.org>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=18332
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2010-09-17 08:05:28 +01:00