[PATCH v2 0/22] On-demand device probing

[PATCH v2 0/22] On-demand device probing

[Tomeu Vizoso]

Hello,

I have a problem with the panel on my Tegra Chromebook taking longer
than expected to be ready during boot (Stéphane Marchesin reported what
is basically the same issue in [0]), and have looked into ordered
probing as a better way of solving this than moving nodes around in the
DT or playing with initcall levels and linking order.

While reading the thread [1] that Alexander Holler started with his
series to make probing order deterministic, it occurred to me that it
should be possible to achieve the same by probing devices as they are
referenced by other devices.

This basically reuses the information that is already implicit in the
probe() implementations, saving us from refactoring existing drivers or
adding information to DTBs.

During review of v1 of this series Linus Walleij suggested that it
should be the device driver core to make sure that dependencies are
ready before probing a device. I gave this idea a try [2] but Mark Brown
pointed out to the logic duplication between the resource acquisition
and dependency discovery code paths (though I think it's fairly minor).

To address that code duplication I experimented with Arnd's devm_probe
[3] concept of having drivers declare their dependencies instead of
acquiring them during probe, and while it worked [4], I don't think we
end up winning anything when compared to just probing devices on-demand
from resource getters.

One remaining objection is to the "sprinkling" of calls to
fwnode_ensure_device() in the resource getters of each subsystem, but I
think it's the right thing to do given that the storage of resources is
currently subsystem-specific.

We could avoid the above by moving resource storage into the core, but I
don't think there's a compelling case for that.

I have tested this on boards with Tegra, iMX.6, Exynos and OMAP SoCs,
and these patches were enough to eliminate all the deferred probes
(except one in PandaBoard because omap_dma_system doesn't have a
firmware node as of yet).

With this series I get the kernel to output to the panel in 0.5s,
instead of 2.8s.

Regards,

Tomeu

[0] http://lists.freedesktop.org/archives/dri-devel/2014-August/066527.html

[1] https://lkml.org/lkml/2014/5/12/452

[2] https://lkml.org/lkml/2015/6/17/305

[3] http://article.gmane.org/gmane.linux.ports.arm.kernel/277689

[4] https://lkml.org/lkml/2015/7/21/441

Changes in v2:
- Move delay to platform.c
- Use set_primary_fwnode()
- Use of_node_full_name()
- Move the logic for finding a platform device from its firmware node to
  of/platform.c as it's not needed for ACPI nodes.
- Add acpi_dev_get_device()
- Add fwnode_ensure_device() so the mechanism for probing devices on
  demand is independent of the firmware format.
- Acquire regulator device lock before returning from regulator_dev_lookup()

Tomeu Vizoso (22):
  platform: delay device-driver matches until late_initcall
  of/platform: Set fwnode field for new devices
  device property: add fwnode_get_name()
  of/platform: add of_platform_device_find()
  ACPI: add acpi_dev_get_device()
  device property: add fwnode_ensure_device()
  gpio: Probe GPIO drivers on demand
  gpio: Probe pinctrl devices on demand
  regulator: core: Reduce critical area in _regulator_get
  regulator: core: Probe regulators on demand
  drm: Probe panels on demand
  drm/tegra: Probe dpaux devices on demand
  i2c: core: Probe i2c master devices on demand
  pwm: Probe PWM chip devices on demand
  backlight: Probe backlight devices on demand
  usb: phy: Probe phy devices on demand
  clk: Probe clk providers on demand
  pinctrl: Probe pinctrl devices on demand
  phy: core: Probe phy providers on demand
  dma: of: Probe DMA controllers on demand
  power-supply: Probe power supplies on demand
  ASoC: core: Probe components on demand

 drivers/base/platform.c             | 28 +++++++++++
 drivers/base/property.c             | 73 +++++++++++++++++++++++++++
 drivers/clk/clk.c                   |  3 ++
 drivers/dma/of-dma.c                |  2 +
 drivers/gpio/gpiolib-of.c           |  4 ++
 drivers/gpu/drm/drm_panel.c         |  2 +
 drivers/gpu/drm/tegra/dpaux.c       |  2 +
 drivers/i2c/i2c-core.c              |  2 +
 drivers/of/platform.c               | 61 +++++++++++++++++++++++
 drivers/phy/phy-core.c              |  2 +
 drivers/pinctrl/devicetree.c        |  1 +
 drivers/power/power_supply_core.c   |  2 +
 drivers/pwm/core.c                  |  2 +
 drivers/regulator/core.c            | 99 +++++++++++++++++++++----------------
 drivers/usb/phy/phy.c               |  2 +
 drivers/video/backlight/backlight.c |  2 +
 include/linux/acpi.h                | 10 ++++
 include/linux/of_platform.h         |  1 +
 include/linux/property.h            |  4 ++
 sound/soc/soc-core.c                |  6 ++-
 20 files changed, 264 insertions(+), 44 deletions(-)

-- 
2.4.3

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 13/22] i2c: core: Probe i2c master devices on demand

[Tomeu Vizoso]

When looking up an i2c master through its firmware node, probe it if it
hasn't already.

The goal is to reduce deferred probes to a minimum, as it makes it very
cumbersome to find out why a device failed to probe, and can introduce
very big delays in when a critical device is probed.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
---

Changes in v2: None

 drivers/i2c/i2c-core.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c
index e6d4935161e4..5520b413e3db 100644
--- a/drivers/i2c/i2c-core.c
+++ b/drivers/i2c/i2c-core.c
@@ -1353,6 +1353,8 @@ struct i2c_adapter *of_find_i2c_adapter_by_node(struct device_node *node)
 {
 	struct device *dev;
 
+	fwnode_ensure_device(&node->fwnode);
+
 	dev = bus_find_device(&i2c_bus_type, NULL, node,
 					 of_dev_node_match);
 	if (!dev)
-- 
2.4.3

Re: [PATCH v2 0/22] On-demand device probing

[Rafael J. Wysocki]

On Tuesday, July 28, 2015 03:19:31 PM Tomeu Vizoso wrote:
> Hello,
> 
> I have a problem with the panel on my Tegra Chromebook taking longer
> than expected to be ready during boot (Stéphane Marchesin reported what
> is basically the same issue in [0]), and have looked into ordered
> probing as a better way of solving this than moving nodes around in the
> DT or playing with initcall levels and linking order.
> 
> While reading the thread [1] that Alexander Holler started with his
> series to make probing order deterministic, it occurred to me that it
> should be possible to achieve the same by probing devices as they are
> referenced by other devices.
> 
> This basically reuses the information that is already implicit in the
> probe() implementations, saving us from refactoring existing drivers or
> adding information to DTBs.
> 
> During review of v1 of this series Linus Walleij suggested that it
> should be the device driver core to make sure that dependencies are
> ready before probing a device. I gave this idea a try [2] but Mark Brown
> pointed out to the logic duplication between the resource acquisition
> and dependency discovery code paths (though I think it's fairly minor).
> 
> To address that code duplication I experimented with Arnd's devm_probe
> [3] concept of having drivers declare their dependencies instead of
> acquiring them during probe, and while it worked [4], I don't think we
> end up winning anything when compared to just probing devices on-demand
> from resource getters.
> 
> One remaining objection is to the "sprinkling" of calls to
> fwnode_ensure_device() in the resource getters of each subsystem, but I
> think it's the right thing to do given that the storage of resources is
> currently subsystem-specific.
> 
> We could avoid the above by moving resource storage into the core, but I
> don't think there's a compelling case for that.
> 
> I have tested this on boards with Tegra, iMX.6, Exynos and OMAP SoCs,
> and these patches were enough to eliminate all the deferred probes
> (except one in PandaBoard because omap_dma_system doesn't have a
> firmware node as of yet).
> 
> With this series I get the kernel to output to the panel in 0.5s,
> instead of 2.8s.

Can you trim your CC list somewhat, please?

I'm definitely going to look at this, but not before then next week.
Sorry about that.

Thanks,
Rafael

Re: [PATCH v2 0/22] On-demand device probing

[Rob Herring]

On Tue, Jul 28, 2015 at 8:19 AM, Tomeu Vizoso
<tomeu.vizoso@collabora.com> wrote:
> Hello,
>
> I have a problem with the panel on my Tegra Chromebook taking longer
> than expected to be ready during boot (Stéphane Marchesin reported what
> is basically the same issue in [0]), and have looked into ordered
> probing as a better way of solving this than moving nodes around in the
> DT or playing with initcall levels and linking order.
>
> While reading the thread [1] that Alexander Holler started with his
> series to make probing order deterministic, it occurred to me that it
> should be possible to achieve the same by probing devices as they are
> referenced by other devices.
>
> This basically reuses the information that is already implicit in the
> probe() implementations, saving us from refactoring existing drivers or
> adding information to DTBs.
>
> During review of v1 of this series Linus Walleij suggested that it
> should be the device driver core to make sure that dependencies are
> ready before probing a device. I gave this idea a try [2] but Mark Brown
> pointed out to the logic duplication between the resource acquisition
> and dependency discovery code paths (though I think it's fairly minor).
>
> To address that code duplication I experimented with Arnd's devm_probe
> [3] concept of having drivers declare their dependencies instead of
> acquiring them during probe, and while it worked [4], I don't think we
> end up winning anything when compared to just probing devices on-demand
> from resource getters.
>
> One remaining objection is to the "sprinkling" of calls to
> fwnode_ensure_device() in the resource getters of each subsystem, but I
> think it's the right thing to do given that the storage of resources is
> currently subsystem-specific.

Seems like a minor change to me.

> We could avoid the above by moving resource storage into the core, but I
> don't think there's a compelling case for that.
>
> I have tested this on boards with Tegra, iMX.6, Exynos and OMAP SoCs,
> and these patches were enough to eliminate all the deferred probes
> (except one in PandaBoard because omap_dma_system doesn't have a
> firmware node as of yet).
>
> With this series I get the kernel to output to the panel in 0.5s,
> instead of 2.8s.

Generally, I think this looks pretty good. It is simple and the error
path is simply falling back to deferred probe.

One overall comment is I'm not so sure if fwnode_ensure_device
shouldn't just be of_ensure_device. At least currently, it looks like
all the calling locations are DT specific functions anyway. There's
very little logic within the function to really benefit sharing with
ACPI. It is basically just a call to of_platform_device_find and then
bus_probe_device. I expect the get functions will always call into
DT/ACPI specific functions which can then call the firmware specific
device find function.

Rob

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v2 0/22] On-demand device probing

[Tomeu Vizoso]

On 30 July 2015 at 05:06, Rob Herring <robherring2@gmail.com> wrote:
> On Tue, Jul 28, 2015 at 8:19 AM, Tomeu Vizoso
> <tomeu.vizoso@collabora.com> wrote:
>> Hello,
>>
>> I have a problem with the panel on my Tegra Chromebook taking longer
>> than expected to be ready during boot (Stéphane Marchesin reported what
>> is basically the same issue in [0]), and have looked into ordered
>> probing as a better way of solving this than moving nodes around in the
>> DT or playing with initcall levels and linking order.
>>
>> While reading the thread [1] that Alexander Holler started with his
>> series to make probing order deterministic, it occurred to me that it
>> should be possible to achieve the same by probing devices as they are
>> referenced by other devices.
>>
>> This basically reuses the information that is already implicit in the
>> probe() implementations, saving us from refactoring existing drivers or
>> adding information to DTBs.
>>
>> During review of v1 of this series Linus Walleij suggested that it
>> should be the device driver core to make sure that dependencies are
>> ready before probing a device. I gave this idea a try [2] but Mark Brown
>> pointed out to the logic duplication between the resource acquisition
>> and dependency discovery code paths (though I think it's fairly minor).
>>
>> To address that code duplication I experimented with Arnd's devm_probe
>> [3] concept of having drivers declare their dependencies instead of
>> acquiring them during probe, and while it worked [4], I don't think we
>> end up winning anything when compared to just probing devices on-demand
>> from resource getters.
>>
>> One remaining objection is to the "sprinkling" of calls to
>> fwnode_ensure_device() in the resource getters of each subsystem, but I
>> think it's the right thing to do given that the storage of resources is
>> currently subsystem-specific.
>
> Seems like a minor change to me.
>
>> We could avoid the above by moving resource storage into the core, but I
>> don't think there's a compelling case for that.
>>
>> I have tested this on boards with Tegra, iMX.6, Exynos and OMAP SoCs,
>> and these patches were enough to eliminate all the deferred probes
>> (except one in PandaBoard because omap_dma_system doesn't have a
>> firmware node as of yet).
>>
>> With this series I get the kernel to output to the panel in 0.5s,
>> instead of 2.8s.
>
> Generally, I think this looks pretty good. It is simple and the error
> path is simply falling back to deferred probe.
>
> One overall comment is I'm not so sure if fwnode_ensure_device
> shouldn't just be of_ensure_device. At least currently, it looks like
> all the calling locations are DT specific functions anyway. There's
> very little logic within the function to really benefit sharing with
> ACPI. It is basically just a call to of_platform_device_find and then
> bus_probe_device. I expect the get functions will always call into
> DT/ACPI specific functions which can then call the firmware specific
> device find function.

That's fine with me. I just went that way because I assumed the plan
was for subsystems to move to consume fw data through fwnode and drop
as much fw-specific code as possible.

But I have just looked at fwnode_get_named_gpiod and the OF and ACPI
code paths are so dissimilar that I guess that's not so and would be
better to do as you say.

Thanks,

Tomeu

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v2 13/22] i2c: core: Probe i2c master devices on demand

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 1160 bytes --]

On Tue, Jul 28, 2015 at 03:19:44PM +0200, Tomeu Vizoso wrote:
> When looking up an i2c master through its firmware node, probe it if it
> hasn't already.
> 
> The goal is to reduce deferred probes to a minimum, as it makes it very
> cumbersome to find out why a device failed to probe, and can introduce
> very big delays in when a critical device is probed.
> 
> Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>

What is the status of this series? The boot time reduction sounds great.

> ---
> 
> Changes in v2: None
> 
>  drivers/i2c/i2c-core.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c
> index e6d4935161e4..5520b413e3db 100644
> --- a/drivers/i2c/i2c-core.c
> +++ b/drivers/i2c/i2c-core.c
> @@ -1353,6 +1353,8 @@ struct i2c_adapter *of_find_i2c_adapter_by_node(struct device_node *node)
>  {
>  	struct device *dev;
>  
> +	fwnode_ensure_device(&node->fwnode);

TBH, the function name doesn't tell me a lot. It ensures what?

> +
>  	dev = bus_find_device(&i2c_bus_type, NULL, node,
>  					 of_dev_node_match);
>  	if (!dev)
> -- 
> 2.4.3
> 

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [PATCH v2 13/22] i2c: core: Probe i2c master devices on demand

[Tomeu Vizoso]

On 9 August 2015 at 14:34, Wolfram Sang <wsa-z923LK4zBo2bacvFa/9K2g@public.gmane.org> wrote:
> On Tue, Jul 28, 2015 at 03:19:44PM +0200, Tomeu Vizoso wrote:
>> When looking up an i2c master through its firmware node, probe it if it
>> hasn't already.
>>
>> The goal is to reduce deferred probes to a minimum, as it makes it very
>> cumbersome to find out why a device failed to probe, and can introduce
>> very big delays in when a critical device is probed.
>>
>> Signed-off-by: Tomeu Vizoso <tomeu.vizoso-ZGY8ohtN/8qB+jHODAdFcQ@public.gmane.org>
>
> What is the status of this series?

See here for a summary:

http://lkml.kernel.org/g/1438870315-18689-1-git-send-email-tomeu.vizoso-ZGY8ohtN/8qB+jHODAdFcQ@public.gmane.org

> The boot time reduction sounds great.

Note that this should not reduce total boot time substantially, as the
time wasted due to superfluous probe deferrals is really small when
compared to the time spent in delays due to hardware constraints. If
you want to reduce total boot time, async probing may be helpful in
some scenarios.

Regards,

Tomeu

>> ---
>>
>> Changes in v2: None
>>
>>  drivers/i2c/i2c-core.c | 2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c
>> index e6d4935161e4..5520b413e3db 100644
>> --- a/drivers/i2c/i2c-core.c
>> +++ b/drivers/i2c/i2c-core.c
>> @@ -1353,6 +1353,8 @@ struct i2c_adapter *of_find_i2c_adapter_by_node(struct device_node *node)
>>  {
>>       struct device *dev;
>>
>> +     fwnode_ensure_device(&node->fwnode);
>
> TBH, the function name doesn't tell me a lot. It ensures what?
>
>> +
>>       dev = bus_find_device(&i2c_bus_type, NULL, node,
>>                                        of_dev_node_match);
>>       if (!dev)
>> --
>> 2.4.3
>>
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html