gmane.linux.ports.tegra

[PATCH] ARM: tegra: enable host1x support by default

Alexandre Courbot — 2013-05-18T09:17:46

Tegra DRM support used to be enabled in the default Tegra configuration,
but it now depends on CONFIG_TEGRA_HOST1X which is disabled by default.
Enable CONFIG_TEGRA_HOST1X so that DRM support is compiled in again.

Signed-off-by: Alexandre Courbot <acourbot-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>
---
 arch/arm/configs/tegra_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/configs/tegra_defconfig b/arch/arm/configs/tegra_defconfig
index a5f0485..dbf83db 100644
--- a/arch/arm/configs/tegra_defconfig
+++ b/arch/arm/configs/tegra_defconfig
< at >< at > -153,6 +153,7 < at >< at > CONFIG_MEDIA_CAMERA_SUPPORT=y
 CONFIG_MEDIA_USB_SUPPORT=y
 CONFIG_USB_VIDEO_CLASS=m
 CONFIG_DRM=y
+CONFIG_TEGRA_HOST1X=y
 CONFIG_DRM_TEGRA=y
 CONFIG_BACKLIGHT_LCD_SUPPORT=y
 # CONFIG_LCD_CLASS_DEVICE is not set

[PATCH] ARM: tegra: add various LED triggers to defconfig

Stephen Warren — 2013-05-17T23:51:30

From: Stephen Warren <swarren-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>

A couple of Tegra boards have gpio-leds present. Enable various trigger
options to make those useful. This sets all the trigger options to their
"if you don't know, say X" value.

Signed-off-by: Stephen Warren <swarren-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>
---
 arch/arm/configs/tegra_defconfig |    6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/arch/arm/configs/tegra_defconfig b/arch/arm/configs/tegra_defconfig
index 8a50fc8..ea2869c 100644
--- a/arch/arm/configs/tegra_defconfig
+++ b/arch/arm/configs/tegra_defconfig
< at >< at > -190,7 +190,13 < at >< at > CONFIG_NEW_LEDS=y
 CONFIG_LEDS_CLASS=y
 CONFIG_LEDS_GPIO=y
 CONFIG_LEDS_TRIGGERS=y
+CONFIG_LEDS_TRIGGER_TIMER=y
+CONFIG_LEDS_TRIGGER_ONESHOT=y
+CONFIG_LEDS_TRIGGER_HEARTBEAT=y
 CONFIG_LEDS_TRIGGER_GPIO=y
+CONFIG_LEDS_TRIGGER_DEFAULT_ON=y
+CONFIG_LEDS_TRIGGER_TRANSIENT=y
+CONFIG_LEDS_TRIGGER_CAMERA=y
 CONFIG_RTC_CLASS=y
 CONFIG_RTC_DRV_MAX8907=y
 CONFIG_RTC_DRV_PALMAS=y

Re: [PATCH] ARM: tegra: Add charger subnode to tps65090 node

Stephen Warren — 2013-05-17T23:57:38

What's the status of the TPS60590 bindings; are they agreed upon by
NVIDIA, TI, and SlimLogic yet? In other words, is this patch still
something I should apply for 3.11, or does it need to be reworked?

Re: [PATCH] ARM: tegra: fix memory size on Beaver

Stephen Warren — 2013-05-17T23:55:24

I have applied this to Tegra's for-3.11/dt branch.

Re: [PATCH 2/2] ARM: tegra: dts: dalmore: enable spi4

Stephen Warren — 2013-05-17T23:54:40

I have applied this to Tegra's for-3.11/dt branch.

Re: [PATCH] ARM: tegra: add various LED triggers to defconfig

Stephen Warren — 2013-05-17T23:53:45

I have applied this to (squashed it into) Tegra's for-3.11/defconfig branch.

Re: [PATCH] ARM: tegra: emc: correction of ram-code parsing from dt

Stephen Warren — 2013-05-17T23:53:19

I've applied this to Tegra's for-3.11/soc branch.


That would be a change to the DT binding. DT bindings are supposed to be
a stable ABI, so we wouldn't want to change the binding unless there was
a strong reason. I don't think there is one here.

Re: [PATCH 1/1] ARM: DT: tegra: Add gpio-leds to Tegra30 Beaver

Stephen Warren — 2013-05-17T23:52:29

I've applied this to Tegra's for-3.11/dt branch. When applying, I moved
the newly added node to maintain the node sort order.

Re: [PATCH] clk: tegra: Use common of_clk_init() function

Stephen Warren — 2013-05-17T22:54:49

Prashant, can you please resend this to Mike for inclusion in 3.11; now
that we have moved the Tegra clock driver out of arch/arm/mach-tegra, we
should send all clock driver changes through Mike if possible. Thanks.

Just one question though:



Is clk-provider.h the right header? This code isn't a clock provider?

[GIT PULL] ARM: tegra: DT-related fixes needed by the USB tree

Stephen Warren — 2013-05-17T22:50:45

Felipe, (and Alan, Greg),

These are the changes from the Tegra tree that need to be merged
before the Tegra USB driver patches 4-8 from Venu can be applied. Recall
that we had previously agreed on this approach for v3.9, but Venu's
patches didn't end up being ready until now, unfortunately. I note
that Venu's patches affect both the Tegra USB and PHY drivers.

Following this, I will soon send some patches that fix up some ARM multi-
platform issues in the Tegra EHCI driver. Those won't affect the Tegra
USB PHY driver, but I do expect I'll want to rebase them on top of Venu's
series, since they will probably interact a bit with Venu's patches.

I also soon hope to see patches to:
* Convert the Tegra USB and PHY drivers from the vbus-gpios DT property
  to the new vbus-supply regulator. This will affect both drivers, and
  I really hope will happen soon.
* Add support for Tegra30 and Tegra114 SoCs. I expect this would touch
  both the USB and PHY drivers. This may or may not happen in 3.11.

As such, there wi

Re: [PATCH v4 3/8] ARM: tegra: update device trees for USB binding rework

Stephen Warren — 2013-05-17T22:06:22

I've made Whistler boot by reverting a few commits entirely unrelated to
USB or your patch series. I'll investigate those later, but for not your
patches test out fine on Whistler. So, I'll apply patches 1-3 in the
series and set up the pull request for the USB tree. Thanks.

Re: Naming convention DSI + device tree

Stephen Warren — 2013-05-17T19:39:00

...

I guess it just got overlooked. I see CONFIG_DRM is on though; perhaps
the default for the Tegra DRM driver changed in Kconfig. Feel free to
send a patch to tegra_defconfig.

[PATCH 3/6] gpu: host1x: Fix memory access in syncpt request

Arto Merilainen — 2013-05-17T11:49:45

This patch fixes a bad memory access in syncpoint request code. If
no syncpoints were available, the code accessed unreserved memory
area causing unexpected behaviour.

Signed-off-by: Arto Merilainen <amerilainen-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>
---
 drivers/gpu/host1x/syncpt.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
index 5bf5366..6b7ee88 100644
--- a/drivers/gpu/host1x/syncpt.c
+++ b/drivers/gpu/host1x/syncpt.c
< at >< at > -40,7 +40,7 < at >< at > static struct host1x_syncpt *_host1x_syncpt_alloc(struct host1x *host,
 
 for (i = 0; i < host->info->nb_pts && sp->name; i++, sp++)
 ;
-if (sp->dev)
+if (i >= host->info->nb_pts)
 return NULL;
 
 name = kasprintf(GFP_KERNEL, "%02d-%s", sp->id,

[PATCH 4/6] gpu: host1x: Fix client_managed type

Arto Merilainen — 2013-05-17T11:49:46

client_managed field in syncpoint structure was defined as an
integer. The field holds, however, only a boolean value. This patch
modifies the type to boolean.

Signed-off-by: Arto Merilainen <amerilainen-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>
---
 drivers/gpu/host1x/drm/gr2d.c |    2 +-
 drivers/gpu/host1x/syncpt.c   |    8 ++++----
 drivers/gpu/host1x/syncpt.h   |    6 +++---
 3 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/host1x/drm/gr2d.c b/drivers/gpu/host1x/drm/gr2d.c
index 6a45ae0..6c3a27b 100644
--- a/drivers/gpu/host1x/drm/gr2d.c
+++ b/drivers/gpu/host1x/drm/gr2d.c
< at >< at > -281,7 +281,7 < at >< at > static int gr2d_probe(struct platform_device *pdev)
 if (!gr2d->channel)
 return -ENOMEM;
 
-*syncpts = host1x_syncpt_request(dev, 0);
+*syncpts = host1x_syncpt_request(dev, false);
 if (!(*syncpts)) {
 host1x_channel_free(gr2d->channel);
 return -ENOMEM;
diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
index 6b7ee88..e560d67 100644
--- a/drivers/gpu/host1x/

[PATCH 1/6] gpu: host1x: Fixes to host1x firewall

Arto Merilainen — 2013-05-17T11:49:43

From: Terje Bergstrom <tbergstrom-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>

This patch adds several fixes to host1x firewall:
- Host1x firewall does not survive if it expects a reloc, but user
  space didn't pass any relocs. Also it reset the reloc table for
  each gather, whereas they should be reset only per submit. Also
  class does not need to be reset for each class - once per submit
  is enough.
- For INCR opcode the check was not working properly at all.
- The firewall verified gather buffers before copying them. This
  allowed a malicious application to rewrite the buffer content by
  timing the rewrite carefully. This patch makes the buffer
  validation occur after copying the buffers.

Signed-off-by: Terje Bergstrom <tbergstrom-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>
Signed-off-by: Arto Merilainen <amerilainen-DDmLM1+adcrQT0dZR+AlfA< at >public.gmane.org>
---
 drivers/gpu/host1x/job.c |  120 ++++++++++++++++++++--------------------------
 1 file changed, 53 insertions(+), 67 deletions(-)

diff --git a/

[PATCH 0/6] Miscellaneous fixes to host1x

Arto Merilainen — 2013-05-17T11:49:42

This patch series fixes two issues in the host1x driver: First, the
command buffer validation routine had vulnerabilities that were not
detected in earlier testing. Second, the return codes of some
functions were misleading or completely missing. This caused the
driver to give wrong return codes also to the userspace.

The series is based on top of 3.10rc1. I have tested the patch series
on cardhu by running host1x and gr2d test cases (available at [0]).
I would appreciate any help in testing/reviewing these patches.

[0] https://gitorious.org/linux-host1x/libdrm-host1x

Arto Merilainen (5):
  gpu: host1x: Fix syncpoint wait return value
  gpu: host1x: Fix memory access in syncpt request
  gpu: host1x: Fix client_managed type
  gpu: host1x: Rework CPU syncpoint increment
  drm/tegra: Fix syncpoint increment return code

Terje Bergstrom (1):
  gpu: host1x: Fixes to host1x firewall

 drivers/gpu/host1x/dev.h          |    8 +--
 drivers/gpu/host1x/drm/drm.c      |    3 +-
 drivers/gpu/host1x/drm/gr2d.c     |

Re: [PATCH 0/6] ARM: tegra114: add CPU hotplug support

Russell King - ARM Linux — 2013-05-17T10:27:00

No.  This is the code I removed:

-int tegra_cpu_disable(unsigned int cpu)
-{
-       /*
-        * we don't allow CPU 0 to be shutdown (it is still too special
-        * e.g. clock tick interrupts)
-        */
-       return cpu == 0 ? -EPERM : 0;
-}

This gets called from here:

static int platform_cpu_disable(unsigned int cpu)
{
        if (smp_ops.cpu_disable)
                return smp_ops.cpu_disable(cpu);

        /*
         * By default, allow disabling all CPUs except the first one,
         * since this is special on a lot of platforms, e.g. because
         * of clock tick interrupts.
         */
        return cpu == 0 ? -EPERM : 0;
}

Notice the useless utterly nature of tegra_cpu_disable() given that it's
doing _precisely_ the same thing as the common code.  It might as well
not even exist - and it doesn't for that very reason.  Never provide
implementations for methods which are copies of the default action if
no method is provided.

Re: [PATCH 0/6] ARM: tegra114: add CPU hotplug support

Russell King - ARM Linux — 2013-05-17T10:23:45

It got removed for the simple reason that the implementation of that
function was exactly the same as the default path in cpu_disable().
If you have a need to do something else other than the default, then
provide the function (and *only* provide it if you need to do something
different.)

However, if the default suffices, do not provide extra code even if you
later intend to change the function.  Only add the function when you
need to do something else.

I see too much of this "I'll add this function just becase, and maybe
I'll change it later" crap getting into the kernel; this needs to stop.
Things like providing an empty function for every single method in a
struct even through the behaviour would be to do nothing had that
function not been provided.  So whenever I see this, I remove the
offending functions if I'm touching that code.

Re: [PATCH 0/6] ARM: tegra114: add CPU hotplug support

Joseph Lo — 2013-05-17T10:15:59

I think the CPU0 is the boot CPU for most of the case. We always need at
least one CPU online. That assumes to be the boot CPU.


OK. Thanks.

Re: Naming convention DSI + device tree

Alex Courbot — 2013-05-17T05:36:34

Current version is RFC v2 - I know Laurent is working on a v3, but it is 
taking some time. It will probably differ considerably from v2, so be 
warned if you use it.

Still, if you want to play around with the CDF, you can have a look at 
the tegra-cdf branch of my github repo: 
https://github.com/Gnurou/linux/commits/tegra-cdf

It was kind of old so I rebased it on top of 3.10rc1. It includes the 
CDF v2 patch, adds CDF support to the tegra-drm driver, and also 
implements a driver for the Tegra 2-based Ventana panel. It's pretty 
straightforward to read so I guess you can use that as a guide.

I checked whether the rebase worked and it does - you just need to 
enable host1x/tegradrm support on top of the default settings (Stephen, 
why aren't these enabled by default?)

So long story short, CDF works well and is small enough to be easy to 
maintain out-of-tree. A proper panel framework is desperately needed in 
the kernel, so I hope it will move forward soon.

Hope this helps,
Alex.

Re: [PATCH v4 7/8] usb: phy: tegra: Add error handling & clean up.

Stephen Warren — 2013-05-16T21:08:29

I'm being really anal here, but note this wasn't a change to a comment,
but to the error message printed by the code.