From: Dave Airlie airl...@redhat.com
Looking at hibernate overwriting I though it looked like a cursor,
so I tracked down this missing piece to stop the cursor blink
timer. I've no idea if this is sufficient to fix the hibernate
problems people are seeing, but please test it.
Both radeon and
Hi Linus,
I tracked down the misc memory corruption after i915 hibernate to the
blinking fbcon cursor, and realised the i915 driver wasn't doing the fbdev
suspend/resume calls at all. nouveau and radeon have done these calls for
a long time.
This has been fairly well tested and is definitely
#part sign=pgpmime
On Thu, 29 Mar 2012 08:14:08 +0100 (IST), Dave Airlie airl...@linux.ie wrote:
Dave Airlie (1):
drm/i915: suspend fbdev device around suspend/hibernate
This has my Reviewed-by on it; Dave suggested just sending it directly
to you instead of running it through my tree
Oh YES! Finally!
Just as a side note before it went unnoticed: That is the official HDMI
register documentation from R6xx - NI hardware, not just something
reverse engineered! @Rafał and the rest of the community: We hoped that
you could stick with it and cleanup the current implementation
2012/3/28 alexdeuc...@gmail.com:
+/* digital blocks */
+#define TMDSA_CNTL 0x7880
+# define TMDSA_HDMI_EN (1 2)
+#define LVTMA_CNTL 0x7a80
+# define LVTMA_HDMI_EN (1 2)
+#define DDIA_CNTL
https://bugs.freedesktop.org/show_bug.cgi?id=26891
--- Comment #16 from James Tyrrell ja...@tyrrells.com 2012-03-29 03:03:58 PDT
---
This works for my MacBook Pro 8,3. It allows me to use my thunderbolt monitor
whilst booting via EFI, something that is not possible when using the Intel
IGD.
--
-Original Message-
From: Rafał Miłecki [mailto:zaj...@gmail.com]
Sent: Thursday, March 29, 2012 4:17 AM
To: alexdeuc...@gmail.com
Cc: airl...@gmail.com; dri-devel@lists.freedesktop.org; Deucher, Alexander
Subject: Re: [PATCH] drm/radeon/kms: add register definitions for audio
On Wed, Mar 28, 2012 at 09:55:09AM +0100, Tony Vroon wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 27/03/12 21:39, Daniel Vetter wrote:
Yet another thing to try: Can you append intel_iommu=igfx_off to
your kernel cmdline? If you can, try this both on mainline kernels
and with
From 2424354579d86a9c40eaea58b75c7d1c2e52577b Mon Sep 17 00:00:00 2001
From: majianpeng majianp...@gmail.com
Date: Thu, 29 Mar 2012 15:20:40 +0800
Subject: [PATCH] drm/i915:fix for some -Wuninitialized compiler warnings.
Signed-off-by: majianpeng majianp...@gmail.com
---
From 9f714f4192fcfccfc312466531a77f0e73162417 Mon Sep 17 00:00:00 2001
From: majianpeng majianp...@gmail.com
Date: Thu, 29 Mar 2012 15:48:02 +0800
Subject: [PATCH] drm/i915:Fix array bounds warnings.
Signed-off-by: majianpeng majianp...@gmail.com
---
drivers/gpu/drm/i915/i915_debugfs.c |1 +
On Thu, Mar 29, 2012 at 2:52 AM, Chris Wilson ch...@chris-wilson.co.uk wrote:
On Thu, 29 Mar 2012 02:26:37 +0800, Daniel Kurtz djku...@chromium.org wrote:
It is very common for an i2c device to require a small 1 or 2 byte write
followed by a read. For example, when reading from an i2c EEPROM
On Thu, Mar 29, 2012 at 2:48 AM, Chris Wilson ch...@chris-wilson.co.uk wrote:
On Thu, 29 Mar 2012 02:26:36 +0800, Daniel Kurtz djku...@chromium.org wrote:
The i915 is only able to generate a STOP cycle (i.e. finalize an i2c
transaction) during a DATA or WAIT phase. In other words, the
On Thu, Mar 29, 2012 at 2:41 AM, Chris Wilson ch...@chris-wilson.co.uk wrote:
On Thu, 29 Mar 2012 02:26:34 +0800, Daniel Kurtz djku...@chromium.org wrote:
The GMBUS controller GMBUS3 register is double-buffered. Take advantage
of this by writing two 4-byte words before the first wait for
Hello Michel Dänzer,
This is a semi-automatic email about new static checker warnings.
The patch d936622c3627: drm/radeon: Only warn if the intra-domain
offset actually exceeds the limit. from Mar 28, 2012, leads to the
following Smatch complaint:
drivers/gpu/drm/radeon/radeon_object.c:244
From: Michel Dänzer michel.daen...@amd.com
Reported-by: Dan Carpenter dan.carpen...@oracle.com
Signed-off-by: Michel Dänzer michel.daen...@amd.com
---
Third time's the charm, I hope...
drivers/gpu/drm/radeon/radeon_object.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff
The following set of patches will revive the drm-render-nodes [1]
branch that has been dormant in Dave Airlie's repository for some
time. I rebased this branch to the latest drm-core-next and did some
(hopefully useful) follow-up work.
I fixed a few bugs, did a substantial cleanup, separated the
Push minor number instead. This is a preparatory
patch for introducing render nodes. It has been derived
from 7c5cc4f63556e351e9e5980ed22accad410e3fdc originally
created by Dave Airlie.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_fops.c | 19
Add support for creating multiple render nodes on the same
physical GPU.
Patch derived from 7c5cc4f63556e351e9e5980ed22accad410e3fdc
originally authored by Dave Airlie.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_crtc.c |1 +
drivers/gpu/drm/drm_stub.c
Make dev_mapping per-minor instead of per device. This is
a preparatory patch for introducing render nodes. This
will allow per-node instead of per-device mapping range,
once we introduce render nodes.
Patch derived from 7c5cc4f63556e351e9e5980ed22accad410e3fdc
originally authored by Dave Airlie.
From: Dave Airlie airl...@redhat.com
just adds some unchecked ioctls to setup the nodes.
Signed-off-by: Dave Airlie airl...@redhat.com
v2: - original ioctl numbers are now taken, use next available
- resolve some trivial conflicts due to bit-rot that
occurred since the original patch
drm_minor structure had a list_head pointer that was
used only for render nodes. control and primary nodes
do not live in the list and had this list pointer unused.
Create a separate drm_render_node structure that is used
to build the render node list and that points to the
coresponding minor.
The render-node manipulation ioctls are supposed to be
issued through control node only. Add a check and return
-EPERM to user space if access is attempted through a
node other than a control node.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_stub.c |8
Fix a few bugs in render node create and destroy ioctl
(mostly handling of error cases). Also separate some
common code into a new function for destroying
the render node.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_stub.c | 80
When a new render node is created, the number of elements
of id_list allocated in drm_mode_group_init function should
not be the sum of all CRTCs, encoders, and connectors that
the device has, but the one specified by the ioctl that
created the node.
Signed-off-by: Ilija Hadzic
Keep track of per-node open count and do not allow removal
of a render node in use.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_fops.c |2 ++
drivers/gpu/drm/drm_stub.c |2 ++
include/drm/drmP.h |1 +
3 files changed, 5 insertions(+), 0
This is the opposite function of drm_mode_group_init. It will
be needed to properly cleanup the render node after removing it.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_crtc.c | 10 ++
include/drm/drm_crtc.h |1 +
2 files changed, 11
In drm_put_dev, the whole device is being destroyed, so id_list
of the primary (legacy) node should also be cleaned up.
This plugs a memory leak that has probably existed even without
the render-node work.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_stub.c
Add fields to drm_crtc, drm_encoder and drm_connector that keep
track of which render node owns it. Assign ownership when resource
is added and revoke it when resource is destroyed. Do not allow
creation of a node that tries to claim resources that are already
in use by another node.
Critical sections are parts of the code where we claim or
release resources (we don't want two render-node create or
remove ioctl called in the context of different processes
to claim part of requested resources because of the race).
Another critical section is manipulating the render node
list.
User space can send us all kinds of nonsense for num_crtc, num_encoder
and num_connector. So far, we have been checking only for presence
of at least one CRTC/encoder/connector (barring the trivial case
of a render node with no display resources, i.e., GPGPU node).
This patch makes the ioctl fail
Render node ioctl requires a list of DRM mode objects
in specific order: first all CRTCs, then all encoders,
followed by all connectors. Check that the IDs passed
from userland are in conformance with this requirement
and that they are consistent with specified num_crtc,
num_encoder and
id_list is dynamically allocated when a render node is
created. Consequently, if must be freed when the render node
is removed, otherwise we have a memory leak.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
drivers/gpu/drm/drm_stub.c |6 ++
1 files changed, 6
The following two patches add libdrm support for render node
manipulation. The first patch adds the functions for render
node creation and removal that an application can call.
The second patch add two small test utilities that allow
render node creation and removal from shell.
For example, this
Implement the user-space side of drm_render_node_create
and drm_render_node_remove ioctls. The new functions
are drmCreateRenderNode and drmRemoveRenderNode.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
include/drm/drm.h |2 +
include/drm/drm_mode.h | 16
Add two simple programs (render_node_add and render_node_rm)
that can be used to add and remove render nodes from a shell
command or a script.
Signed-off-by: Ilija Hadzic ihad...@research.bell-labs.com
---
.gitignore |2 +
configure.ac |1
On Thu, Mar 29, 2012 at 12:41:26PM -0400, Ilija Hadzic wrote:
diff --git a/include/drm/drm_mode.h b/include/drm/drm_mode.h
index 2a2acda..2a403bb 100644
--- a/include/drm/drm_mode.h
+++ b/include/drm/drm_mode.h
@@ -441,4 +441,18 @@ struct drm_mode_destroy_dumb {
uint32_t handle;
};
On Thu, Mar 29, 2012 at 12:41:22PM -0400, Ilija Hadzic wrote:
The following set of patches will revive the drm-render-nodes [1]
branch that has been dormant in Dave Airlie's repository for some
time. I rebased this branch to the latest drm-core-next and did some
(hopefully useful) follow-up
On Thu, 29 Mar 2012, Ville [iso-8859-1] Syrjälä wrote:
+/* render node create and remove functions
+ if crtc/encoders/connectors all == 0 then gpgpu node */
+struct drm_render_node_create {
+ __u32 node_minor_id;
+ __u32 num_crtc;
+ __u32 num_encoder;
+ __u32
On Thu, Mar 29, 2012 at 12:22:20PM -0500, Ilija Hadzic wrote:
On Thu, 29 Mar 2012, Ville [iso-8859-1] Syrjälä wrote:
+/* render node create and remove functions
+ if crtc/encoders/connectors all == 0 then gpgpu node */
+struct drm_render_node_create {
+ __u32 node_minor_id;
+
https://bugzilla.kernel.org/show_bug.cgi?id=29412
--- Comment #14 from Jon Dowland jon+bugzilla.kernel@alcopop.org
2012-03-29 18:08:59 ---
That second patch sorts it! Thanks! I booted to single-user; modprobe radeon;
pm-suspend; resumed: fine. Moved to multi-user; logged into GNOME;
Hi
Here is a new version of the patches.
Summary: almost all patches changed at least one line, Kernel patches
7 and 8 are RFCs.
The main difference is that now the ioctls are not for CRTC properties, but for
generic object properties. I decided to write it like this after Rob Clark's
review.
From: Paulo Zanoni paulo.r.zan...@intel.com
Move code from drm_mode_connector_property_set_ioctl to a new
function, so we can reuse this code when we add crtc properties.
v2: use bool instead of int
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
---
drivers/gpu/drm/drm_crtc.c | 41
From: Paulo Zanoni paulo.r.zan...@intel.com
For now, only connectors have it. In the future, all objects that need
properties should use it. Since the strucutre is referenced inside
struct drm_mode_object, we will be able to deal with object properties
without knowing the real type of the object.
From: Paulo Zanoni paulo.r.zan...@intel.com
Useless for connector properties (since they already have their own
ioctls), but useful when we add properties to CRTCs, planes and other
objects.
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
---
drivers/gpu/drm/drm_crtc.c | 180
From: Paulo Zanoni paulo.r.zan...@intel.com
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
---
drivers/gpu/drm/drm_crtc.c | 96 +--
1 files changed, 12 insertions(+), 84 deletions(-)
diff --git a/drivers/gpu/drm/drm_crtc.c
From: Paulo Zanoni paulo.r.zan...@intel.com
The i915 driver needs this for the rotation and overscan compensation
properties. Other drivers might need this too.
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
---
drivers/gpu/drm/drm_crtc.c | 20
From: Paulo Zanoni paulo.r.zan...@intel.com
This property is needed so we can inform the KVMr feature about our
current rotation: whenever we change the rotation, we should change
that property so that the KVMr knows that the screen is rotated.
How to reproduce the problem:
- on an AMT machine,
From: Paulo Zanoni paulo.r.zan...@intel.com
They're named underscan hborder and underscan vborder. The
properties accept values from 0 to 100, where 0 is don't compensate
and 100 is shrink the screen as much as possible (not necessarily
100 pixels).
V2: Rename to underscan hborder and underscan
From: Paulo Zanoni paulo.r.zan...@intel.com
Use unsigned int instead of int:
- modetest.c:89:1: warning: comparison between signed and unsigned integer
expressions [-Wsign-compare]
- modetest.c:97:1: warning: comparison between signed and unsigned integer
expressions [-Wsign-compare]
-
From: Paulo Zanoni paulo.r.zan...@intel.com
Don't continue without freeing the connector.
192 bytes in 6 blocks are indirectly lost in loss record 6 of 12
at 0x4C2779D: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x4E30DD8: drmMalloc (xf86drm.c:147)
by 0x4E35024:
From: Paulo Zanoni paulo.r.zan...@intel.com
New library calls:
- drmModeObjectGetProperties
- drmModeFreeObjectProperties
- drmModeObjectSetProperties
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
---
include/drm/drm.h |2 +
include/drm/drm_mode.h | 24 ++
From: Paulo Zanoni paulo.r.zan...@intel.com
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
---
tests/modetest/modetest.c | 15 +++
1 files changed, 15 insertions(+), 0 deletions(-)
diff --git a/tests/modetest/modetest.c b/tests/modetest/modetest.c
index 02ea579..a6a2a8c
From: Paulo Zanoni paulo.r.zan...@intel.com
Don't worry if that fails: only the KVMr feature will be affected.
We still need to change the sna/ code.
We also need to add the dependency on libdrm.
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
---
src/intel.h |3 ++
From: Paulo Zanoni paulo.r.zan...@intel.com
In the Kernel side, these are crtc properties (since in the hardware,
underscan use the panel fitters, which are attached to the pipes).
Ideally we should make these as crtc properties too, but since xrandr
doesn't have support for them, we expose the
On Thu, 29 Mar 2012 18:27:20 -0300, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
Also return void instead of int. We have more than 100 callers and
no one checks for the return value.
If this function fails the property won't be exposed by the get/set
On Thu, 29 Mar 2012 18:30:19 -0300, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
Don't worry if that fails: only the KVMr feature will be affected.
We still need to change the sna/ code.
We also need to add the dependency on libdrm.
Signed-off-by:
On Thu, 29 Mar 2012 18:30:20 -0300, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
In the Kernel side, these are crtc properties (since in the hardware,
underscan use the panel fitters, which are attached to the pipes).
Ideally we should make these as crtc
https://bugzilla.kernel.org/show_bug.cgi?id=42678
--- Comment #10 from Jérôme Glisse gli...@freedesktop.org 2012-03-29
22:57:06 ---
Adrien question is: Is Xorg stuck inside the kernel ? Fixing root cause of GPU
lockup is a different matter (basicly you have to go though several G of datas
https://bugzilla.kernel.org/show_bug.cgi?id=29412
--- Comment #15 from Alex Deucher alexdeuc...@gmail.com 2012-03-29 23:01:52
---
Perfect. No need for the vbios. I'll send the patch to Dave.
--
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
--- You are
From: Alex Deucher alexander.deuc...@amd.com
On pre-R600 asics, the SpeedFanControl table is not
executed as part of ASIC_Init as it is on newer asics.
Fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=29412
Signed-off-by: Alex Deucher alexander.deuc...@amd.com
Cc: sta...@vger.kernel.org
---
From: Rob Clark r...@ti.com
A bitmask property is similar to an enum. The enum value is a bit
position (0-63), and valid property values consist of a mask of
zero or more of (1 enum_val[n]).
TODO: word commit msg better
TODO: maybe flags would be a better name for the property type?
---
See
From: Rob Clark r...@ti.com
The omapdrm driver uses this for setting per-overlay rotation. It
is likely also useful for setting YUV-RGB colorspace conversion
matrix, etc.
---
drivers/gpu/drm/drm_crtc.c | 19 +++
include/drm/drm_crtc.h |5 +
2 files changed, 24
On Thu, Mar 29, 2012 at 8:02 PM, Rob Clark rob.cl...@linaro.org wrote:
From: Rob Clark r...@ti.com
A bitmask property is similar to an enum. The enum value is a bit
position (0-63), and valid property values consist of a mask of
zero or more of (1 enum_val[n]).
TODO: word commit msg
Hi everyone,
is there anything else I can provide?
On Fr, 09 Mär 2012, Norbert Preining wrote:
Hi everyone,
currently 3.3.0-rc6+ I just was hit by that after wake up from ram:
On Di, 28 Feb 2012, Daniel Vetter wrote:
Feb 28 11:42:47 mithrandir kernel: [15627.756071]
On Thu, Mar 29, 2012 at 18:28, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
Use unsigned int instead of int:
- modetest.c:89:1: warning: comparison between signed and unsigned integer
expressions [-Wsign-compare]
- modetest.c:97:1: warning: comparison
On Thu, Mar 29, 2012 at 18:28, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
In the future we'll have more than just connector properties, so create
a dump_prop function that can handle any property (instead of the
current dump_props function that only
On Thu, Mar 29, 2012 at 18:28, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
New library calls:
- drmModeObjectGetProperties
- drmModeFreeObjectProperties
- drmModeObjectSetProperties
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
Reviewed-by:
On Thu, Mar 29, 2012 at 18:28, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
Reviewed-by: Eugeni Dodonov eugeni.dodo...@intel.com
--
Eugeni Dodonov
http://eugeni.dodonov.net/
From: Rob Clark r...@ti.com
In syncing with the corresponding kernel header, the wrong license
header was inadvertantly copied over. The intention was for the
userspace headers to have a MIT license following the convention
of the rest of libdrm, xorg, etc.
Signed-off-by: Rob Clark r...@ti.com
On Thu, Mar 29, 2012 at 18:27, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
For now, only connectors have it. In the future, all objects that need
properties should use it. Since the strucutre is referenced inside
*structure
struct drm_mode_object,
On Thu, Mar 29, 2012 at 18:27, Paulo Zanoni przan...@gmail.com wrote:
+ switch (arg_obj-type) {
+ case DRM_MODE_OBJECT_CONNECTOR:
+ ret = drm_mode_connector_set_obj_prop(arg_obj, property,
+ arg-value);
+
On Thu, Mar 29, 2012 at 18:27, Paulo Zanoni przan...@gmail.com wrote:
From: Paulo Zanoni paulo.r.zan...@intel.com
Signed-off-by: Paulo Zanoni paulo.r.zan...@intel.com
Reviewed-by: Eugeni Dodonov eugeni.dodo...@intel.com
--
Eugeni Dodonov
http://eugeni.dodonov.net/
Multi buffer plane pixel formats are added as like kernel header.
Signed-off-by: Seung-Woo Kim sw0312@samsung.com
---
include/drm/drm_fourcc.h |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/include/drm/drm_fourcc.h b/include/drm/drm_fourcc.h
index
https://bugs.freedesktop.org/show_bug.cgi?id=47955
--- Comment #15 from Andrew Randrianasulu rand...@mail.ru 2012-03-29 20:54:13
PDT ---
.. and spriteblast mesa demo surely don't work correctly here. (I see something
remotely like _giant_ sprites flashing on screen)
--
Configure bugmail:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 29/03/12 13:39, Daniel Vetter wrote:
Is the for-chainsaw branch plus intel_iommu=igfx_off now stable for
you?
It appears to crash when I shut my machine down, but has otherwise
behaved quite well.
Regards,
- --
Tony Vroon
UNIX systems
On Thu, Mar 29, 2012 at 9:34 AM, Keith Packard kei...@keithp.com wrote:
#part sign=pgpmime
On Thu, 29 Mar 2012 08:14:08 +0100 (IST), Dave Airlie airl...@linux.ie
wrote:
Dave Airlie (1):
drm/i915: suspend fbdev device around suspend/hibernate
I've reported the issue one year ago and
The GMBUS controller can report a NAK condition while a transaction is
still active. If the driver is fast enough, and the bus is slow enough,
the driver may clear the NAK condition while the controller is still
busy, resulting in a confused GMBUS controller. This will leave the
controller in a
A common method of probing an i2c bus is trying to do a zero-length write.
Handle this case by checking the length first before decrementing it.
This is actually important, since attempting a zero-length write is one
of the ways that i2cdetect and i2c_new_probed_device detect whether
there is
It is very common for an i2c device to require a small 1 or 2 byte write
followed by a read. For example, when reading from an i2c EEPROM it is
common to write and address, offset or index followed by a reading some
values.
The i915 gmbus controller provides a special "INDEX" cycle for
This patchset addresses a couple of issues with the i915 gmbus
implementation.
v6 is a rebased and trimmed version of the patchset,
since the first few patches have already been merged onto drm-intel-next-queued.
* Cleans up INDEX cycle implementation
* Only remove POSTING_READ() when it
Save the GMBUS2 value read while polling for state changes, and then
reuse this value when determining for which reason the loops were exited.
This is a small optimization which saves a couple of bus accesses for
memory mapped IO registers.
To avoid "assigning in if clause" checkpatch errors",
The GMBUS controller GMBUS3 register is double-buffered. Take advantage
of this by writing two 4-byte words before the first wait for HW_RDY.
This helps keep the GMBUS controller from becoming idle during long writes.
Signed-off-by: Daniel Kurtz
---
drivers/gpu/drm/i915/intel_i2c.c | 14
The POSTING_READ() calls were originally added to make sure the writes
were flushed before any timing delays and across loops.
Now that the code has settled a bit, let's remove them.
Signed-off-by: Daniel Kurtz
---
drivers/gpu/drm/i915/intel_i2c.c |3 ---
1 files changed, 0 insertions(+), 3
The i915 is only able to generate a STOP cycle (i.e. finalize an i2c
transaction) during a DATA or WAIT phase. In other words, the
controller rejects a STOP requested as part of the first transaction in a
sequence.
Thus, for the first transaction we must always use a WAIT cycle, detect
when the
From: Dave Airlie
Looking at hibernate overwriting I though it looked like a cursor,
so I tracked down this missing piece to stop the cursor blink
timer. I've no idea if this is sufficient to fix the hibernate
problems people are seeing, but please test it.
Both radeon and
Hi Linus,
I tracked down the misc memory corruption after i915 hibernate to the
blinking fbcon cursor, and realised the i915 driver wasn't doing the fbdev
suspend/resume calls at all. nouveau and radeon have done these calls for
a long time.
This has been fairly well tested and is definitely
<#part sign=pgpmime>
On Thu, 29 Mar 2012 08:14:08 +0100 (IST), Dave Airlie
wrote:
> Dave Airlie (1):
> drm/i915: suspend fbdev device around suspend/hibernate
This has my Reviewed-by on it; Dave suggested just sending it directly
to you instead of running it through my tree and then back
Oh YES! Finally!
Just as a side note before it went unnoticed: That is the official HDMI
register documentation from R6xx - NI hardware, not just something
reverse engineered! @Rafa? and the rest of the community: We hoped that
you could stick with it and cleanup the current implementation
2012/3/28 :
> +/* digital blocks */
> +#define TMDSA_CNTL ? ? ? ? ? ? ? ? ? ? ? 0x7880
> +# ? ? ? define TMDSA_HDMI_EN ? ? ? ? ? ? (1 << 2)
> +#define LVTMA_CNTL ? ? ? ? ? ? ? ? ? ? ? 0x7a80
> +# ? ? ? define LVTMA_HDMI_EN ? ? ? ? ? ? (1 << 2)
> +#define DDIA_CNTL ? ? ? ? ? ? ? ? ? ? ? ?0x7200
>
On Thu, Mar 29, 2012 at 04:46:39PM +0800, Daniel Kurtz wrote:
> On Thu, Mar 29, 2012 at 2:41 AM, Chris Wilson
> wrote:
> > On Thu, 29 Mar 2012 02:26:34 +0800, Daniel Kurtz
> > wrote:
> >> The GMBUS controller GMBUS3 register is double-buffered. ?Take advantage
> >> of this ?by writing two
On Thu, Mar 29, 2012 at 04:37:18PM +0800, Daniel Kurtz wrote:
> On Thu, Mar 29, 2012 at 2:52 AM, Chris Wilson
> wrote:
> > On Thu, 29 Mar 2012 02:26:37 +0800, Daniel Kurtz
> > wrote:
> >> It is very common for an i2c device to require a small 1 or 2 byte write
> >> followed by a read. ?For
https://bugs.freedesktop.org/show_bug.cgi?id=26891
--- Comment #16 from James Tyrrell 2012-03-29 03:03:58
PDT ---
This works for my MacBook Pro 8,3. It allows me to use my thunderbolt monitor
whilst booting via EFI, something that is not possible when using the Intel
IGD.
--
Configure
> -Original Message-
> From: Rafa? Mi?ecki [mailto:zajec5 at gmail.com]
> Sent: Thursday, March 29, 2012 4:17 AM
> To: alexdeucher at gmail.com
> Cc: airlied at gmail.com; dri-devel at lists.freedesktop.org; Deucher,
> Alexander
> Subject: Re: [PATCH] drm/radeon/kms: add register
On Wed, Mar 28, 2012 at 09:55:09AM +0100, Tony Vroon wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> On 27/03/12 21:39, Daniel Vetter wrote:
> > Yet another thing to try: Can you append intel_iommu=igfx_off to
> > your kernel cmdline? If you can, try this both on mainline kernels
> >
Signed-off-by: majianpeng
---
drivers/gpu/drm/i915/i915_gem.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 1f441f5..c071b66 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++
Signed-off-by: majianpeng
---
drivers/gpu/drm/i915/i915_debugfs.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c
b/drivers/gpu/drm/i915/i915_debugfs.c
index fdb7cce..8ad59dd 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++
On Thu, Mar 29, 2012 at 2:52 AM, Chris Wilson
wrote:
> On Thu, 29 Mar 2012 02:26:37 +0800, Daniel Kurtz
> wrote:
>> It is very common for an i2c device to require a small 1 or 2 byte write
>> followed by a read. ?For example, when reading from an i2c EEPROM it is
>> common to write and
On Thu, Mar 29, 2012 at 2:48 AM, Chris Wilson
wrote:
> On Thu, 29 Mar 2012 02:26:36 +0800, Daniel Kurtz
> wrote:
>> The i915 is only able to generate a STOP cycle (i.e. finalize an i2c
>> transaction) during a DATA or WAIT phase. ?In other words, the
>> controller rejects a STOP requested as
1 - 100 of 154 matches
Mail list logo