Just abort instead of trying to limp on if we fail to even allocate
memory during initialisation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
We need to be a notch more careful with remapping the possible clones
mask from the kernel:
- SDVO and Haswell DDI have multiple connectors mapping to the same encoder.
- The connector list can be reordered compared to the encoder list.
This is used to move eDP/LVDS to the front.
Also add even more debug output.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Start filling in the names for the parts that have been announced, the
Iris branded Haswell GT3 parts.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Perform a second sanity check that all the outputs we expect to be
active remain so following the cleanup after VT switching.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
The 'sync' parameter past along XvPutImage is not a request for us to
perform the operation synchronised to the vblank, rather a hint that the
client wishes the operation to be synchronous. In Xorg, this is
implemented as an immediate ShmCompletion event so make sure we flush
the operation to hw in this case.
Fixes regression from
commit 195a51353c [2.21.8]
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue Apr 9 19:13:46 2013 +0100
sna/video: Convert to a pure Xv backend
Reported-by: Matti Hamalainen <ccr@tnsp.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=65048
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reported-by: Jiri Slaby <jirislaby@gmail.com>
Reported-by: David Gow <david@ingeniumdigital.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=64978
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
If we undo the pending operations to a bo, then fallback we cause
corruption. For example, see wine and its 1-bit rendering.
Fixes regression from
commit 07a4400fff [2.21.7]
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri May 10 11:59:59 2013 +0100
sna: Attempt to discard overwritten operations before CopyArea
Also we need to be more careful and consider alu when marking the
operation as 'replaces'.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
At that point, the GPU bo will no longer be a clone of the source and so
we need to decouple it and make ourselves a fresh copy.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Fixes regression from
commit 16a64649e9 [2.21.7]
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri May 10 15:48:58 2013 +0100
sna: Basic copy-on-write support for cloning pixmaps
Reported-by: Axel Rohde
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Otherwise we discard a fill operation that is required to initialise the
pixmap before drawing commences.
Fixes regression from
commit ef9dc6fae5 [2.21.7]
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon May 20 11:08:51 2013 +0100
sna: Undo a few more overwritten operations upon a bo
Reported-by: Matti Hamalainen <ccr@tnsp.org>
Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=64841
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Coacher is optimistic that the issue is no longer reproducible on his
machine - and whilst I do not understand the root cause, I am confident
that the kernel code is correct as is our use.
References: https://bugs.freedesktop.org/show_bug.cgi?id=61628
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Otherwise we end up cloning the scanout only to leave it dangling if the
client copies the from the front-buffer and then writes to it.
Reported-by: Nick Bowler <nbowler@draconx.ca>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=64675
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Once cloned, we do not want to use inplace operations and instead force
a copy. However, if we do not relinquish the hints when copying across
the bo, then those hints become stale and lead to corruption.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
As Chris mentioned there is a tendency for us to find out more
PCI IDs only when users report. So let's add all new reserved Haswell IDs.
I didn't have better names for this reserved ids and didn't want to use rsvd1
and rsvd2 groups, so I decided to use "B" and "E" that stands for the last
id digit.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
When publishing first HSW ids we weren't allowed to use "GT3" codname.
But this is the correct codname and Mesa is using it already.
So to avoid people getting confused why in Mesa it is called GT3 and here
it is called GT2_PLUS let's fix this name in a standard and correct way.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Since the extended use of move_area_to_gpu for partial migration of
render sources, we exposed the lack of handling of upload caches along
that path.
Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson
Daniel Vetter
Rodrigo Vivi