https://bugzilla.kernel.org/show_bug.cgi?id=104551
--- Comment #6 from Ignacio ---
Those logs were taken from a Ubuntu 15.04 live USB stick
--
You are receiving this mail because:
You are watching the assignee of the bug.
attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/f2c83794/attachment.html>
At the moment it only supports ANX7814.
The ANX7814 is an ultra-low power Full-HD (1080p60) SlimPort transmitter
designed for portable devices.
This driver adds initial support and supports HDMI to DP pass-through mode.
Signed-off-by: Enric Balletbo i Serra
---
drivers/gpu/drm/bridge/Kconfig
The ANX7814 is an ultra-low power Full-HD (1080p60) SlimPort transmitter
designed for portable devices.
You can add support to your board with current binding.
Example:
anx7814: anx7814 at 38 {
compatible = "analogix,anx7814";
reg = <0x38>;
Analogix Semiconductor develops analog and mixed-signal devices for digital
media and communications interconnect applications.
Signed-off-by: Enric Balletbo i Serra
Acked-by: Rob Herring
---
Documentation/devicetree/bindings/vendor-prefixes.txt | 1 +
1 file changed, 1 insertion(+)
diff --git
Hi all,
This is the fouth version with fixes suggested by Nicolas Boichat and Dan
Carpenter.
See the changelog below for details.
The following series add initial support for the Slimport ANX7814 transmitter, a
ultra-low power Full-HD (1080p60) transmitter designed for portable device.
The dri
On Fri, Sep 25, 2015 at 04:09:46PM +0300, Tomi Valkeinen wrote:
>
>
> On 25/09/15 13:41, Kamil Lulko wrote:
> > Hi,
> >
> >> fbdev is (more or less) maintained, but it's a deprecated framework. All
> >> new Linux display drivers should be done on DRM.
> >
> > What about no-mmu platforms? DRM ha
Hi
On Thu, Sep 24, 2015 at 1:24 PM, Tomi Valkeinen
wrote:
>
> On 11/09/15 19:10, David Herrmann wrote:
>> Currently, for each open() on an fbdev device, we pin the underlying
>> fbdev device and driver module. On close(), we release both. This
>> guarantees that the fbdev object stays around unt
On 2015ë
09ì 24ì¼ 10:01, Joonyoung Shim wrote:
> Hi Inki,
>
> On 08/17/2015 06:03 PM, Inki Dae wrote:
>> On 2015ë
08ì 17ì¼ 17:17, Joonyoung Shim wrote:
>>> On 08/17/2015 04:52 PM, Inki Dae wrote:
On 2015ë
08ì 17ì¼ 14:29, Joonyoung Shim wrote:
> On 08/16/2015 02:07 PM, Inki
On Fri, Sep 25, 2015 at 04:29:44PM +0200, Thierry Reding wrote:
> On Thu, Sep 24, 2015 at 07:02:36PM +0200, Thierry Reding wrote:
> > From: Thierry Reding
> >
> > Some modules register several sub-drivers. Provide a helper that makes
> > it easy to register and unregister a list of sub-drivers, a
Hello Jyri,
Yes using or not DRM bridge we should be able to have a common
implementation
Please find,my answer belows
BR,
Arnaud
On 09/25/2015 04:11 PM, Jyri Sarha wrote:
> Despite my earlier comment this implementation and the related HW is
> quite similar in all significant aspects to the p
From: Maruthi Bayyavarapu
This adds the ACP (Audio CoProcessor) IP driver and wires
it up to the amdgpu driver. The ACP block provides the DMA
engine for i2s based ALSA driver. This is required for audio
on APUs that utilize an i2s codec.
Reviewed-by: Jammy Zhou
Reviewed-by: Maruthi Bayyavarap
From: Maruthi Srinivas Bayyavarapu
Vendor specific quirk was added to:
1. Support AMD platform which has two dwc controllers with different
base address for playback and capture. Also, I2S_COMP_PARAM_*
registers offsets differed.
2. Resume audio which was active before system suspend.
Af
From: Maruthi Srinivas Bayyavarapu
dw i2s controller can work in slave mode, codec being master.
dw i2s is made to support master/slave operation, by reading dwc
register.
Signed-off-by: Maruthi Bayyavarapu
Signed-off-by: Alex Deucher
---
v2: check register to determine master/slave mode
in
This patch set implements support for i2s audio and new AMD GPUs.
The i2s codec is fed by a DMA engine on the GPU. To handle this
we create mfd cells which we hang the i2s codec and DMA engine on.
Because of this, this patch set covers two subsystems: drm and alsa.
The drm patches add support for
lines.
--
You are receiving this mail because:
You are the assignee for the bug.
-- next part --
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/16efc18d/attachment.html>
From: Thierry Reding
Some modules register several sub-drivers. Provide a helper that makes
it easy to register and unregister a list of sub-drivers, as well as
unwind properly on error.
Cc: Greg Kroah-Hartman
Signed-off-by: Thierry Reding
---
Changes in v2:
- properly pass around the owner mo
erry
-- next part --
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/ce5c3e41/attachment.sig>
by instrumenting platform_driver_register() to return failure
at various points in the sequence.
This works fine.
Thierry
-- next part --
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/b254adba/attachment.sig>
Despite my earlier comment this implementation and the related HW is
quite similar in all significant aspects to the patch set posted couple
of days ago [1] for Beaglebone-Black HDMI audio.
[1] http://permalink.gmane.org/gmane.linux.alsa.devel/144144
I have not yet gotten to bottom of drm-side
Oops, sorry, this email shouldn't be in the patch set, please
disconsider it.
--
Rafael
On Fri, Sep 25, 2015 at 04:54:53PM -0700, Rafael Antognolli wrote:
> Second attempt at implementing a module that allows reading/writing arbitrary
> dpcd registers. Changes to this version:
> - lseek is
This module is heavily based on i2c-dev. Once loaded, it provides one
dev node per DP AUX channel, named drm_dp_auxN, where N is an integer.
It's possible to know which connector owns this aux channel by looking
at the respective sysfs /sys/class/drm_aux_dev/drm_dp_auxN/connector, if
the connector
This is useful to determine which connector owns this AUX channel.
Signed-off-by: Rafael Antognolli
---
drivers/gpu/drm/i915/intel_dp.c | 1 +
include/drm/drm_dp_helper.h | 1 +
2 files changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
i
This series implement support to a drm_dp_aux chardev that allows reading and
writing an arbitrary amount of bytes to arbitrary dpcd register addresses using
regular read, write and lseek operations.
v2:
- lseek is used to select the register to read/write
- read/write are used instead of ioctl
Second attempt at implementing a module that allows reading/writing arbitrary
dpcd registers. Changes to this version:
- lseek is used to select the register to read/write;
- read/write are used instead of ioctl;
- no blocking_notifier is used, just a direct callback.
One t
On Tue, Sep 22, 2015 at 02:17:51PM +0200, Daniel Vetter wrote:
> On Tue, Sep 22, 2015 at 03:00:54PM +0300, Ville Syrjälä wrote:
> > On Tue, Sep 15, 2015 at 04:55:04PM -0700, Rafael Antognolli wrote:
> > > This module is heavily based on i2c-dev. Once loaded, it provides one
> > > dev node per DP
const void *data, size_t size, struct module *module);
>
> +int platform_register_drivers(struct platform_driver * const *drivers,
> + unsigned int count);
> +void platform_unregister_drivers(struct platform_driver * const *drivers,
> + unsigned int count);
> +
> /* early platform driver interface */
> struct early_platform_driver {
> const char *class_str;
> --
> 2.5.0
>
-- next part --
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/62893765/attachment.sig>
e conflicts in
> imx-drm-core.c?
Let's see what Greg thinks about patch 1. There are a bunch of other
drivers that can use these helpers, so it might be best to get the
helper merged first and hold off on the depending patches until the
next release.
I'll let you know when we
-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: OpenPGP digital signature
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/7b19a948/attachment.sig>
Am Freitag, den 25.09.2015, 15:09 +0200 schrieb Philipp Zabel:
> Am Freitag, den 25.09.2015, 14:17 +0200 schrieb Thierry Reding:
> > I think you gain much less by splitting up than you realize. Compare
> > this from before the series:
> >
> > $ du -ch drivers/gpu/drm/imx/*.ko
> > 8.0Kd
Am Freitag, den 25.09.2015, 14:17 +0200 schrieb Thierry Reding:
> I think you gain much less by splitting up than you realize. Compare
> this from before the series:
>
> $ du -ch drivers/gpu/drm/imx/*.ko
> 8.0Kdrivers/gpu/drm/imx/dw_hdmi-imx.ko
Oh right, I didn't realize that this
From: Tomasz Stanislawski
This patch fixes calling usleep_range() after taking reg_slock
using spin_lock_irqsave(). The mdelay() is used instead.
Waiting in atomic context is not the best idea in general.
Hopefully, waiting occurs only when Video Processor fails
to reset correctly.
Signed-off-by
The patch removes unused hdmi_context field.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c
b/drivers/gpu/drm/exynos/exynos_hdmi.c
index a4ec8b9..b0f5ff4 100644
--- a/drivers/gpu/drm/e
Simple formula can be used to calculate CTS and N coefficients.
Additionaly ACR registers have different offsets for different versions
of IP.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 71 +++-
drivers/gpu/drm/exynos/regs-hdmi.h | 2
Inline function is safer than macro, also the name has been changed to
be consistent with other inline function encoder_to_hdmi.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 15 +--
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/d
hdmi_resources structure was filled by old platform data code and is not
necessary anymore. The patch removes it at groups together resource related
fields in hdmi_context.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 105 +++
1 file cha
The patch converts API to gpiod and moves initialization code
to hdmi_resources_init.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 37 ++--
1 file changed, 14 insertions(+), 23 deletions(-)
diff --git a/drivers/gpu/drm/exynos/exynos_hdm
Driver tries to disable sclk_hdmi during re-parenting, to avoid possible
glitches. It is ineffective as the clock is used also by other devices (mixer).
Anyway driver works without disabling sclk_hdmi.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 4
1 file changed
Driver always uses the same number of regulators, so there is no point in
dynamic allocation.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 25 +
1 file changed, 9 insertions(+), 16 deletions(-)
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c
hdmi-en is an optional regulator so it should be better handled by
devm_regulator_get_optional call.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 24 +++-
1 file changed, 11 insertions(+), 13 deletions(-)
diff --git a/drivers/gpu/drm/exynos/exynos_
DRM device should be destroyed before releasing resources.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c
b/drivers/gpu/drm/exynos/exynos_hdmi.c
index 50ced709
Registers are 32-bit, even if only lower 8-bits are used.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c
b/drivers/gpu/drm/exynos/exynos_hdmi.c
index 814dd8c..50
Currently driver tries to set specific HDMI-PHY registers in three situations:
- before reset,
- before power off,
- after applying HDMI-PHY configuration.
First two cases seems to be unnecessary - register contents will be lost
anyway. The third case can be merged with HDMI-PHY configuration by f
The patch moves PLL stabilization check to separate function, adjust timeout
parameters and de-duplicates code common for both HW variants.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 68 ++--
1 file changed, 26 insertions(+), 42 deleti
Some registers resides at different offsets depending on device version.
This patch adds infrastructure for mapping such registers to proper address
based on hdmi_type. It adds also mappings to some registers.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 56 +++
samsung,exynos5-hdmi compatible was marked as deprecated in Jun 2013.
It was never used since then.
Signed-off-by: Andrzej Hajda
---
Documentation/devicetree/bindings/video/exynos_hdmi.txt | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/Documentation/devicetree/binding
This compatible was marked as deprecated in Jun 2013 and it is not used since
then. Additionally its driver data points to wrong pll settings, so it
cannot work anyway.
Signed-off-by: Andrzej Hajda
---
drivers/gpu/drm/exynos/exynos_hdmi.c | 10 --
1 file changed, 10 deletions(-)
diff --
Hi,
This is another set of cleanup/improvement patches for HDMI.
The patchset is based on exynos-drm-next.
It was tested on Universal and Odroid U3.
Regards
Andrzej
Andrzej Hajda (15):
drm/exynos/hdmi: remove support for deprecated compatible
dt-bindings: remove deprecated compatible strin
On Thu, Sep 24, 2015 at 06:51:34PM +0200, Thierry Reding wrote:
> From: Thierry Reding
>
> SCDC is a mechanism defined in the HDMI 2.0 specification that allows
> the source and sink devices to communicate.
>
> This commit introduces helpers to access the SCDC and provides the
> symbolic names f
On Thu, Sep 24, 2015 at 06:51:35PM +0200, Thierry Reding wrote:
> From: Thierry Reding
>
> Sinks compliant with the HDMI 2.0 specification may support SCDC, a
> mechanism for the source and the sink to communicate. Sinks advertise
> support for this feature in the HDMI Forum Vendor Specific Data
The "i" variable should be signed or it leads to a crash in the error
handling code.
Fixes: 1d263474c441 ('drm/amdgpu: unwind properly in amdgpu_cs_parser_init()')
Signed-off-by: Dan Carpenter
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 749
On Fri, Sep 25, 2015 at 02:16:07PM +0300, Jani Nikula wrote:
> On Fri, 25 Sep 2015, Geliang Tang wrote:
> > Leak a task reference in i915_ppgtt_info(), add put_task_struct()
> > to fix it.
>
> Introduced by
>
> commit 1c60fef535d143860d5bf6593e24ab6417f5227c
> Author: Ben Widawsky
> Date: Fri
because, at least in most of the cases I've seen, it refers to the pipe
rather than the CRTC object ID.
That said this is all only temporary, though it may take us a while to
get rid of it. The ultimate goal is to convert all drivers over to using
struct drm_crtc * everywhere, at which point we don't have to bother
with what to call the CRTC index. As a preliminary step, having drivers
use consistent naming is hopefully going to make it easier to replace.
Thierry
-- next part --
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/e6feb675/attachment.sig>
Another regression for Jairo to track.
-Daniel
On Thu, Sep 24, 2015 at 05:22:09PM -0400, Nick Bowler wrote:
> Hi,
>
> Testing out 4.3-rc2, first thing I notice is that the VGA output is
> not working. Specifically, the display is continuously powering on
> and off -- at no point is any image vis
Hi Dave,
Another attempt at drm-misc for 4.4 ...
- better atomic helpers for runtime pm drivers
- atomic fbdev
- dp aux i2c STATUS_UPDATE handling (for short i2c replies from the sink)
- bunch of constify patches
- inital kerneldoc for vga switcheroo
- some vblank code cleanups from Ville and Thie
ion/pgp-signature
Size: 819 bytes
Desc: not available
URL:
<http://lists.freedesktop.org/archives/dri-devel/attachments/20150925/c0712def/attachment.sig>
On Fri, 25 Sep 2015, Geliang Tang wrote:
> Leak a task reference in i915_ppgtt_info(), add put_task_struct()
> to fix it.
Introduced by
commit 1c60fef535d143860d5bf6593e24ab6417f5227c
Author: Ben Widawsky
Date: Fri Dec 6 14:11:30 2013 -0800
drm/i915: Dump all ppgtt
>
> Signed-off-by: G
Neither userspace nor the kernel internal functions use the CMDSTREAM
GEM type anymore. Remove it from the public API and clean up all related
functions.
Signed-off-by: Lucas Stach
---
drivers/staging/etnaviv/etnaviv_gem.c | 127 +++---
include/uapi/drm/etnaviv_drm.h
Instead of using a GEM buffer for the kernel internal ring buffer
use the newly introduced cmdbuf object. This removes the last remaining
user of the CMDSTREAM GEM flag.
Signed-off-by: Lucas Stach
---
drivers/staging/etnaviv/etnaviv_buffer.c | 90 ++--
drivers/staging
This rewrites the submit interface to copy the command stream from user
memory. This mitigates a potential attack vector of the old interface
where userspace could submit a command buffer that would be validated by
the kernel, but is still mapped into userspace. This could be exploited
by changing
There is no point in having a context restore buffer, as the need for tracking
GPU hardware state in userspace and the fact that we need to submit all states
that reference memory buffers anyway to ensure proper patching of the
relocations, eat away from the potential benefit.
Signed-off-by: Lucas
This makes sure that we are satifying the cache handling rules outlined
in the previous commit. Cached buffers are pulled into the CPU domain
before access and pushed to the GPU again when the CPU is done.
Signed-off-by: Lucas Stach
---
drivers/staging/etnaviv/etnaviv_gem.c | 41
This redefines how we do cache handling in the following way:
All buffer are pushed into the GPU domain initially, this will only
be done when populating the buffers backing store, but for simplicity
userspace can assume that the buffer is owned by the GPU as soon as it
constructs a GEM handle. Th
Not bullet proof yet, as this possibly shoots down more submits than
necessary. However it allows for fairly rapid turnarounds during
userspace development.
Signed-off-by: Lucas Stach
---
drivers/staging/etnaviv/etnaviv_gpu.c | 30 +-
1 file changed, 29 insertions(+),
From: Christian Gmeiner
etnaviv_buffer_init() creates a very simple command buffer
to be able to start the FE. FE fetches 'prefetch' number of 64 bit
words via DMA and starts to execute the read buffer.
This is a very simple code cleanup and pushes the whole buffer
logic (alignment, cmd buffer c
From: Christian Gmeiner
This is very useful for debugging issues regarding command
buffer processing.
Signed-off-by: Christian Gmeiner
---
drivers/staging/etnaviv/etnaviv_drv.c | 39 +++
1 file changed, 39 insertions(+)
diff --git a/drivers/staging/etnaviv/etna
From: Christian Gmeiner
Signed-off-by: Christian Gmeiner
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gpu.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/staging/etnaviv/etnaviv_gpu.c
b/drivers/staging/etnaviv/etnaviv_gpu.c
index f89fa869885e..948c85a8a591 1
From: Christian Gmeiner
Signed-off-by: Christian Gmeiner
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gem.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/staging/etnaviv/etnaviv_gem.c
b/drivers/staging/etnaviv/etnaviv_gem.c
index 5c2462d6130d..
From: Russell King
The CMA memory area is shared between CMA and other users. In order to
allow CMA to work, non-CMA allocations in this area must be for movable
pages, so that a CMA allocation can reclaim its memory from other users
when required.
Page cache backed allocations, such as pages a
From: Russell King
Add support for the new GEM_WAIT ioctl, which waits for the kernel to
finish with the GEM object (all render complete, and has been retired).
This is different from the WAIT_FENCE ioctl, which just waits for
rendering to complete.
This is an important distinction, as when a ge
From: Russell King
Move the mapping teardown directly into etnaviv_gem_free_object().
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gem.c | 17 -
1 file changed, 4 insertions(+), 13 deletions(-)
diff --git a/drivers/staging/etnaviv/etnaviv_gem.c
b/drivers/st
From: Russell King
We've already validated the command buffer offset, there's no need to
repeat it elsewhere.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gem_submit.c | 5 -
1 file changed, 5 deletions(-)
diff --git a/drivers/staging/etnaviv/etnaviv_gem_submit.c
b/dri
From: Russell King
On GPU submission, copy the submit command and bo structures in one go
and outside of dev->struct_mutex. There are two reasons:
- This avoids the overhead from calling and checking copy_from_user()
multiple times.
- Holding dev->struct_mutex over a page fault should be avoi
From: Russell King
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gem.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/etnaviv/etnaviv_gem.c
b/drivers/staging/etnaviv/etnaviv_gem.c
index 3a68295eaee4..02c5e46dca75 100644
--- a/drivers/s
From: Russell King
Performance analysis indicates that the command parser is a hot spot
in the code. We can improve the efficiency of the command parser by
storing the command length in a table, rather than decoding the
command manually via a switch statement.
Signed-off-by: Russell King
---
From: Russell King
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_buffer.c | 1 -
drivers/staging/etnaviv/etnaviv_gpu.c| 4 +---
drivers/staging/etnaviv/etnaviv_gpu.h| 2 --
3 files changed, 1 insertion(+), 6 deletions(-)
diff --git a/drivers/staging/etnaviv/etnaviv_bu
From: Russell King
When resuming a core, especially if it has been powered down by a PM
domain, we need to re-initialise the clock control register, and
perform a reload of the context.
This is important where we have a GPU with multiple execution states,
where we must select the appropriate exe
From: Russell King
Rename the GPU clock functions for what they actually are, splitting
the clock disable out from etnaviv_gpu_suspend(). This allows us to
fix etnaviv_gpu_rpm_resume(), which needed only to disable the clocks
on failure, rather than trying to stop an uninitialised GPU.
Signed-o
From: Russell King
The GPU requires a double-write to set the clock divisor. Rather than
open-coding this knowledge in a couple of places, provide a helper to
do this instead. This avoids spreading this knowledge around the
driver.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etna
From: Russell King
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gpu.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/staging/etnaviv/etnaviv_gpu.c
b/drivers/staging/etnaviv/etnaviv_gpu.c
index 233b2197eb37..b12d46075732 100644
--- a/drivers/stagi
From: Russell King
Running the cairo trace "firefox-36-20090609" results in the Xorg server
trying to queue up 319M of buffer objects in a single commit (due to the
number of drawing operations that will fit into a 32K command buffer.)
This causes the commit ioctl to return -ENOSPC as the MMU is
From: Russell King
The mapping structure contains everything we need; avoid passing extra
data in rather than using the data we have stored in the structure.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gem.c | 2 +-
drivers/staging/etnaviv/etnaviv_mmu.c | 31 ++
From: Russell King
The drm_mm scanning for free MMU blocks walks the list of inactive
objects, looking up their vram node. Unfortunately, the code assumes
that if it has a vram node, it is registered into drm_mm.
However, there are two cases when a vram node is created - whenever a
MMU entry ha
From: Russell King
There's no need to initialise a list_head which is only going to be
added with list_add() to an existing list. Remove this redundant code.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_mmu.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/stagin
From: Russell King
The Linux kernel dislikes the use of stdint.h types. In order to
prepare this code for merging, we need to convert to the preferred
types, which are u8/u16/u32/u64 for unsigned ints of the specified
size, and s8/s16/s32/s64 for signed ints of the same.
Signed-off-by: Russell
From: Russell King
Provide a helper to queue a work_struct onto our private workqueue.
This avoids repetitions of:
struct drm_device *dev = ...
struct etnaviv_private *priv = dev->dev_private;
queue_work(priv->wq, w);
Signed-off-by: Russell King
---
drivers/staging/et
From: Russell King
Having a global fence completion implementation in the presence of
mutliple GPU cores which can complete out of order is buggy. Consider
the case of two processes submitting work for two GPU cores:
GPU A GPU B
priv->completed_fence =
From: Russell King
Tracking the current context on a system wide level is incorrect: if the
context for one GPU core changes, the context does not change on other
GPU cores. Make the context tracking a per-GPU thing.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_drv.c | 9 ++
From: Russell King
Move the locking out of etnaviv_gem_mmap_offset() into its solitary
caller, and take the mutex using mutex_lock_interruptible(). This
allows etnaviv_ioctl_gem_info() to handle signals while trying to
obtain this lock.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/
From: Russell King
etnaviv_gem_mmap_offset() returned zero if drm_gem_create_mmap_offset()
failed. This is incorrect behaviour as etnaviv_ioctl_gem_info() does
not detect this condition, so it returns success.
Fix this by re-architecting etnaviv_gem_mmap_offset(). Merge
mmap_offset() into this
From: Russell King
Reduce the exposure of struct_mutex to an absolute minimum. We only
need to take this lock over the call to etnaviv_gem_get_pages() as
vm_insert_page() already does a properly synchronised update of the
PTE (using the pte table lock.)
Reducing the code covered by this lock al
From: Russell King
We're only ever going to insert using a pointer to a struct page. Use
vm_insert_page() rather than converting back and forth between a
struct page and a PFN.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gem.c | 9 -
1 file changed, 4 insertions(+)
From: Russell King
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_drv.c | 43 ---
drivers/staging/etnaviv/etnaviv_gpu.c | 21 ++---
drivers/staging/etnaviv/etnaviv_gpu.h | 2 +-
3 files changed, 44 insertions(+), 22 deletions(-)
dif
From: Russell King
Avoid the below circular dependency warning by avoiding calling
pm_runtime_get_sync() at a point where struct_mutex is held, and we
may well provoke the power domain to be resumed.
We avoid this simply by referencing the runtime PM in the
etnaviv_gem_submit() function just bef
From: Russell King
The aperture end address is inclusive, not exclusive.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_iommu.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/staging/etnaviv/etnaviv_iommu.c
b/drivers/staging/etnaviv/etnaviv_iommu.
From: Russell King
The only failure which etnaviv_load() implemented was for kzalloc().
Other failures (alloc_ordered_workqueue(), component_bind_all()) were
not implemented. Add these.
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_drv.c | 14 +-
1 file changed,
From: Russell King
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_drv.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/staging/etnaviv/etnaviv_drv.c
b/drivers/staging/etnaviv/etnaviv_drv.c
index 5a53c6969fcd..b2ce4a6467ed 100644
--- a/drivers/staging/etnaviv/etnav
Signed-off-by: Lucas Stach
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_buffer.c | 2 +-
drivers/staging/etnaviv/etnaviv_cmd_parser.c | 16
drivers/staging/etnaviv/etnaviv_drv.c| 3 +--
drivers/staging/etnaviv/etnaviv_drv.h| 3 +--
drive
From: Christian Gmeiner
There is no need to spam the kernel logs with the GPU specs and features
at startup. If someone wants to know about this stuff debugfs should be
the right place to look at.
Also use better format specifiers to make it easier for humans to read.
Signed-off-by: Christian G
Signed-off-by: Lucas Stach
Signed-off-by: Russell King
---
drivers/staging/etnaviv/etnaviv_gem_submit.c | 2 +-
drivers/staging/etnaviv/etnaviv_gpu.c| 2 +-
drivers/staging/etnaviv/etnaviv_gpu.h| 4 ++--
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/stagi
1 - 100 of 127 matches
Mail list logo
