[Intel-gfx] [PATCH 02/10] drm: Add privacy-screen class (v4)
Hans de Goede
hdegoede at redhat.com
Fri Dec 10 10:46:46 UTC 2021
Hi Thorsten,
On 12/10/21 11:12, Thorsten Leemhuis wrote:
> Hi, this is your Linux kernel regression tracker speaking.
>
> Top-posting for once, to make this easy accessible to everyone.
>
> Hans, I stumbled upon your blog post about something that afaics is
> related to below patch:
> https://hansdegoede.livejournal.com/25948.html
>
> To quote:
>
>> To avoid any regressions distors should modify their initrd
>> generation tools to include privacy screen provider drivers in the
>> initrd (at least on systems with a privacy screen), before 5.17
>> kernels start showing up in their repos.
>>
>> If this change is not made, then users using a graphical bootsplash
>> (plymouth) will get an extra boot-delay of up to 8 seconds
>> (DeviceTimeout in plymouthd.defaults) before plymouth will show and
>> when using disk-encryption where the LUKS password is requested from
>> the initrd, the system will fallback to text-mode after these 8
>> seconds.
>
> Sorry for intruding, but to me as the kernel's regression tracker that
> blog post sounds a whole lot like "by kernel development standards this
> is not allowed due to the 'no regression rule', as users should never be
> required to update something in userspace when updating the kernel".
I completely understand where you are coming from here, but AFAIK
the initrd generation has always been a bit of an exception here.
For example (IIRC) over time we have seen the radeon module growing
a runtime dependency on the amdkfd module which also required updated
initrd generation tools.
Another example (which I'm sure of) is the i915 driver gaining a softdep
on kvmgt (which is now gone again) which required new enough kmod tools
to understand this as well as initrd generation tools updates to also
take softdeps into account:
https://github.com/dracutdevs/dracut/commit/4cdee66c8ed5f82bbd0638e30d867318343b0e6c
More in general if you look at e.g. dracut's git history, there are
various other cases where dracut needed to be updated to adjust
to kernel changes. For example dracut decides if a module is a block
driver and thus may be necessary to have in the initrd based on a
list of kernel-symbols the module links to and sometimes those
symbols change due to refactoring of kernel internals, see e.g. :
https://github.com/dracutdevs/dracut/commit/b292ce7295f18192124e64e5ec31161d09492160
TL;DR: initrd-generators and the kernel are simply tied together so much
that users cannot expect to be able to jump to the latest kernel without
either updating the initrd-generator, or adding some modules as modules
which must always be added to the initrd in the initrd-generator config
file (as a workaround).
Declaring kernel changes which break initrd-generation in some way as
being regressions, would mean that e.g. we cannot introduce any
kernel changes which causes some drm/block/whatever drivers to use
some new register helper functions which are not yet on the list of
symbols which dracut uses to identify drm/block/whatever drivers.
The only difference between previous initrd-generator breaking changes
and this one, is that I decided that it would be good for everyone
to be aware of this before hand; and now I get the feeling that I'm
being punished for warning people about this instead of just letting
things break silently. I know you don't intend your email this way in
any way, but still.
Also AFAIK drivers may also at some point drop support for (much) older
firmware versions requiring installing a recent linux-firmware together
with a new kernel.
In my reading of the rules the 'users should never be required to update
something in userspace when updating the kernel' rule is about keeping
people's normal programs working, IOW not breaking userspace ABI and that
is not happening here.
> But I'm not totally sure that's the case here. Could you please clarify
> what happens when a user doesn't update the initramfs. E.g. what happens
> besides the mentioned delay and the text mode (which are bad already,
> but might be a accetable compormise here -- but that's up for Linus to
> decide)? Does everything start to work normally shortly after the kernel
> mounted the rootfs and finally can load the missing module?
Yes everything will work normally once the rootfs gets mounted and the
missing module gets loaded.
Regards,
Hans
> On 05.10.21 22:23, Hans de Goede wrote:
>> On some new laptops the LCD panel has a builtin electronic privacy-screen.
>> We want to export this functionality as a property on the drm connector
>> object. But often this functionality is not exposed on the GPU but on some
>> other (ACPI) device.
>>
>> This commit adds a privacy-screen class allowing the driver for these
>> other devices to register themselves as a privacy-screen provider; and
>> allowing the drm/kms code to get a privacy-screen provider associated
>> with a specific GPU/connector combo.
>>
>> Changes in v2:
>> - Make CONFIG_DRM_PRIVACY_SCREEN a bool which controls if the drm_privacy
>> code gets built as part of the main drm module rather then making it
>> a tristate which builds its own module.
>> - Add a #if IS_ENABLED(CONFIG_DRM_PRIVACY_SCREEN) check to
>> drm_privacy_screen_consumer.h and define stubs when the check fails.
>> Together these 2 changes fix several dependency issues.
>> - Remove module related code now that this is part of the main drm.ko
>> - Use drm_class as class for the privacy-screen devices instead of
>> adding a separate class for this
>>
>> Changes in v3:
>> - Make the static inline drm_privacy_screen_get_state() stub set sw_state
>> and hw_state to PRIVACY_SCREEN_DISABLED to squelch an uninitialized
>> variable warning when CONFIG_DRM_PRIVICAY_SCREEN is not set
>>
>> Changes in v4:
>> - Make drm_privacy_screen_set_sw_state() skip calling out to the hw if
>> hw_state == new_sw_state
>>
>> Reviewed-by: Emil Velikov <emil.l.velikov at gmail.com>
>> Reviewed-by: Lyude Paul <lyude at redhat.com>
>> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
>> ---
>> Documentation/gpu/drm-kms-helpers.rst | 15 +
>> MAINTAINERS | 8 +
>> drivers/gpu/drm/Kconfig | 4 +
>> drivers/gpu/drm/Makefile | 1 +
>> drivers/gpu/drm/drm_drv.c | 4 +
>> drivers/gpu/drm/drm_privacy_screen.c | 403 ++++++++++++++++++++++
>> include/drm/drm_privacy_screen_consumer.h | 50 +++
>> include/drm/drm_privacy_screen_driver.h | 80 +++++
>> include/drm/drm_privacy_screen_machine.h | 41 +++
>> 9 files changed, 606 insertions(+)
>> create mode 100644 drivers/gpu/drm/drm_privacy_screen.c
>> create mode 100644 include/drm/drm_privacy_screen_consumer.h
>> create mode 100644 include/drm/drm_privacy_screen_driver.h
>> create mode 100644 include/drm/drm_privacy_screen_machine.h
>>
>> diff --git a/Documentation/gpu/drm-kms-helpers.rst b/Documentation/gpu/drm-kms-helpers.rst
>> index ec2f65b31930..5bb55ec1b9b5 100644
>> --- a/Documentation/gpu/drm-kms-helpers.rst
>> +++ b/Documentation/gpu/drm-kms-helpers.rst
>> @@ -435,3 +435,18 @@ Legacy CRTC/Modeset Helper Functions Reference
>>
>> .. kernel-doc:: drivers/gpu/drm/drm_crtc_helper.c
>> :export:
>> +
>> +Privacy-screen class
>> +====================
>> +
>> +.. kernel-doc:: drivers/gpu/drm/drm_privacy_screen.c
>> + :doc: overview
>> +
>> +.. kernel-doc:: include/drm/drm_privacy_screen_driver.h
>> + :internal:
>> +
>> +.. kernel-doc:: include/drm/drm_privacy_screen_machine.h
>> + :internal:
>> +
>> +.. kernel-doc:: drivers/gpu/drm/drm_privacy_screen.c
>> + :export:
>> diff --git a/MAINTAINERS b/MAINTAINERS
>> index 28e5f0ae1009..cb94bb3b8724 100644
>> --- a/MAINTAINERS
>> +++ b/MAINTAINERS
>> @@ -6423,6 +6423,14 @@ F: drivers/gpu/drm/drm_panel.c
>> F: drivers/gpu/drm/panel/
>> F: include/drm/drm_panel.h
>>
>> +DRM PRIVACY-SCREEN CLASS
>> +M: Hans de Goede <hdegoede at redhat.com>
>> +L: dri-devel at lists.freedesktop.org
>> +S: Maintained
>> +T: git git://anongit.freedesktop.org/drm/drm-misc
>> +F: drivers/gpu/drm/drm_privacy_screen*
>> +F: include/drm/drm_privacy_screen*
>> +
>> DRM TTM SUBSYSTEM
>> M: Christian Koenig <christian.koenig at amd.com>
>> M: Huang Rui <ray.huang at amd.com>
>> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
>> index 2a926d0de423..c686c08447ac 100644
>> --- a/drivers/gpu/drm/Kconfig
>> +++ b/drivers/gpu/drm/Kconfig
>> @@ -481,3 +481,7 @@ config DRM_PANEL_ORIENTATION_QUIRKS
>> config DRM_LIB_RANDOM
>> bool
>> default n
>> +
>> +config DRM_PRIVACY_SCREEN
>> + bool
>> + default n
>> diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
>> index 0dff40bb863c..788fc37096f6 100644
>> --- a/drivers/gpu/drm/Makefile
>> +++ b/drivers/gpu/drm/Makefile
>> @@ -32,6 +32,7 @@ drm-$(CONFIG_OF) += drm_of.o
>> drm-$(CONFIG_PCI) += drm_pci.o
>> drm-$(CONFIG_DEBUG_FS) += drm_debugfs.o drm_debugfs_crc.o
>> drm-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o
>> +drm-$(CONFIG_DRM_PRIVACY_SCREEN) += drm_privacy_screen.o
>>
>> obj-$(CONFIG_DRM_DP_AUX_BUS) += drm_dp_aux_bus.o
>>
>> diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
>> index 7a5097467ba5..dc293b771c3f 100644
>> --- a/drivers/gpu/drm/drm_drv.c
>> +++ b/drivers/gpu/drm/drm_drv.c
>> @@ -43,6 +43,7 @@
>> #include <drm/drm_managed.h>
>> #include <drm/drm_mode_object.h>
>> #include <drm/drm_print.h>
>> +#include <drm/drm_privacy_screen_machine.h>
>>
>> #include "drm_crtc_internal.h"
>> #include "drm_internal.h"
>> @@ -1029,6 +1030,7 @@ static const struct file_operations drm_stub_fops = {
>>
>> static void drm_core_exit(void)
>> {
>> + drm_privacy_screen_lookup_exit();
>> unregister_chrdev(DRM_MAJOR, "drm");
>> debugfs_remove(drm_debugfs_root);
>> drm_sysfs_destroy();
>> @@ -1056,6 +1058,8 @@ static int __init drm_core_init(void)
>> if (ret < 0)
>> goto error;
>>
>> + drm_privacy_screen_lookup_init();
>> +
>> drm_core_init_complete = true;
>>
>> DRM_DEBUG("Initialized\n");
>> diff --git a/drivers/gpu/drm/drm_privacy_screen.c b/drivers/gpu/drm/drm_privacy_screen.c
>> new file mode 100644
>> index 000000000000..183a6011adf0
>> --- /dev/null
>> +++ b/drivers/gpu/drm/drm_privacy_screen.c
>> @@ -0,0 +1,403 @@
>> +// SPDX-License-Identifier: MIT
>> +/*
>> + * Copyright (C) 2020 - 2021 Red Hat, Inc.
>> + *
>> + * Authors:
>> + * Hans de Goede <hdegoede at redhat.com>
>> + */
>> +
>> +#include <linux/device.h>
>> +#include <linux/kernel.h>
>> +#include <linux/list.h>
>> +#include <linux/module.h>
>> +#include <linux/mutex.h>
>> +#include <linux/slab.h>
>> +#include <drm/drm_privacy_screen_machine.h>
>> +#include <drm/drm_privacy_screen_consumer.h>
>> +#include <drm/drm_privacy_screen_driver.h>
>> +#include "drm_internal.h"
>> +
>> +/**
>> + * DOC: overview
>> + *
>> + * This class allows non KMS drivers, from e.g. drivers/platform/x86 to
>> + * register a privacy-screen device, which the KMS drivers can then use
>> + * to implement the standard privacy-screen properties, see
>> + * :ref:`Standard Connector Properties<standard_connector_properties>`.
>> + *
>> + * KMS drivers using a privacy-screen class device are advised to use the
>> + * drm_connector_attach_privacy_screen_provider() and
>> + * drm_connector_update_privacy_screen() helpers for dealing with this.
>> + */
>> +
>> +#define to_drm_privacy_screen(dev) \
>> + container_of(dev, struct drm_privacy_screen, dev)
>> +
>> +static DEFINE_MUTEX(drm_privacy_screen_lookup_lock);
>> +static LIST_HEAD(drm_privacy_screen_lookup_list);
>> +
>> +static DEFINE_MUTEX(drm_privacy_screen_devs_lock);
>> +static LIST_HEAD(drm_privacy_screen_devs);
>> +
>> +/*** drm_privacy_screen_machine.h functions ***/
>> +
>> +/**
>> + * drm_privacy_screen_lookup_add - add an entry to the static privacy-screen
>> + * lookup list
>> + * @lookup: lookup list entry to add
>> + *
>> + * Add an entry to the static privacy-screen lookup list. Note the
>> + * &struct list_head which is part of the &struct drm_privacy_screen_lookup
>> + * gets added to a list owned by the privacy-screen core. So the passed in
>> + * &struct drm_privacy_screen_lookup must not be free-ed until it is removed
>> + * from the lookup list by calling drm_privacy_screen_lookup_remove().
>> + */
>> +void drm_privacy_screen_lookup_add(struct drm_privacy_screen_lookup *lookup)
>> +{
>> + mutex_lock(&drm_privacy_screen_lookup_lock);
>> + list_add(&lookup->list, &drm_privacy_screen_lookup_list);
>> + mutex_unlock(&drm_privacy_screen_lookup_lock);
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_lookup_add);
>> +
>> +/**
>> + * drm_privacy_screen_lookup_remove - remove an entry to the static
>> + * privacy-screen lookup list
>> + * @lookup: lookup list entry to remove
>> + *
>> + * Remove an entry previously added with drm_privacy_screen_lookup_add()
>> + * from the static privacy-screen lookup list.
>> + */
>> +void drm_privacy_screen_lookup_remove(struct drm_privacy_screen_lookup *lookup)
>> +{
>> + mutex_lock(&drm_privacy_screen_lookup_lock);
>> + list_del(&lookup->list);
>> + mutex_unlock(&drm_privacy_screen_lookup_lock);
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_lookup_remove);
>> +
>> +/*** drm_privacy_screen_consumer.h functions ***/
>> +
>> +static struct drm_privacy_screen *drm_privacy_screen_get_by_name(
>> + const char *name)
>> +{
>> + struct drm_privacy_screen *priv;
>> + struct device *dev = NULL;
>> +
>> + mutex_lock(&drm_privacy_screen_devs_lock);
>> +
>> + list_for_each_entry(priv, &drm_privacy_screen_devs, list) {
>> + if (strcmp(dev_name(&priv->dev), name) == 0) {
>> + dev = get_device(&priv->dev);
>> + break;
>> + }
>> + }
>> +
>> + mutex_unlock(&drm_privacy_screen_devs_lock);
>> +
>> + return dev ? to_drm_privacy_screen(dev) : NULL;
>> +}
>> +
>> +/**
>> + * drm_privacy_screen_get - get a privacy-screen provider
>> + * @dev: consumer-device for which to get a privacy-screen provider
>> + * @con_id: (video)connector name for which to get a privacy-screen provider
>> + *
>> + * Get a privacy-screen provider for a privacy-screen attached to the
>> + * display described by the @dev and @con_id parameters.
>> + *
>> + * Return:
>> + * * A pointer to a &struct drm_privacy_screen on success.
>> + * * ERR_PTR(-ENODEV) if no matching privacy-screen is found
>> + * * ERR_PTR(-EPROBE_DEFER) if there is a matching privacy-screen,
>> + * but it has not been registered yet.
>> + */
>> +struct drm_privacy_screen *drm_privacy_screen_get(struct device *dev,
>> + const char *con_id)
>> +{
>> + const char *dev_id = dev ? dev_name(dev) : NULL;
>> + struct drm_privacy_screen_lookup *l;
>> + struct drm_privacy_screen *priv;
>> + const char *provider = NULL;
>> + int match, best = -1;
>> +
>> + /*
>> + * For now we only support using a static lookup table, which is
>> + * populated by the drm_privacy_screen_arch_init() call. This should
>> + * be extended with device-tree / fw_node lookup when support is added
>> + * for device-tree using hardware with a privacy-screen.
>> + *
>> + * The lookup algorithm was shamelessly taken from the clock
>> + * framework:
>> + *
>> + * We do slightly fuzzy matching here:
>> + * An entry with a NULL ID is assumed to be a wildcard.
>> + * If an entry has a device ID, it must match
>> + * If an entry has a connection ID, it must match
>> + * Then we take the most specific entry - with the following order
>> + * of precedence: dev+con > dev only > con only.
>> + */
>> + mutex_lock(&drm_privacy_screen_lookup_lock);
>> +
>> + list_for_each_entry(l, &drm_privacy_screen_lookup_list, list) {
>> + match = 0;
>> +
>> + if (l->dev_id) {
>> + if (!dev_id || strcmp(l->dev_id, dev_id))
>> + continue;
>> +
>> + match += 2;
>> + }
>> +
>> + if (l->con_id) {
>> + if (!con_id || strcmp(l->con_id, con_id))
>> + continue;
>> +
>> + match += 1;
>> + }
>> +
>> + if (match > best) {
>> + provider = l->provider;
>> + best = match;
>> + }
>> + }
>> +
>> + mutex_unlock(&drm_privacy_screen_lookup_lock);
>> +
>> + if (!provider)
>> + return ERR_PTR(-ENODEV);
>> +
>> + priv = drm_privacy_screen_get_by_name(provider);
>> + if (!priv)
>> + return ERR_PTR(-EPROBE_DEFER);
>> +
>> + return priv;
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_get);
>> +
>> +/**
>> + * drm_privacy_screen_put - release a privacy-screen reference
>> + * @priv: privacy screen reference to release
>> + *
>> + * Release a privacy-screen provider reference gotten through
>> + * drm_privacy_screen_get(). May be called with a NULL or ERR_PTR,
>> + * in which case it is a no-op.
>> + */
>> +void drm_privacy_screen_put(struct drm_privacy_screen *priv)
>> +{
>> + if (IS_ERR_OR_NULL(priv))
>> + return;
>> +
>> + put_device(&priv->dev);
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_put);
>> +
>> +/**
>> + * drm_privacy_screen_set_sw_state - set a privacy-screen's sw-state
>> + * @priv: privacy screen to set the sw-state for
>> + * @sw_state: new sw-state value to set
>> + *
>> + * Set the sw-state of a privacy screen. If the privacy-screen is not
>> + * in a locked hw-state, then the actual and hw-state of the privacy-screen
>> + * will be immediately updated to the new value. If the privacy-screen is
>> + * in a locked hw-state, then the new sw-state will be remembered as the
>> + * requested state to put the privacy-screen in when it becomes unlocked.
>> + *
>> + * Return: 0 on success, negative error code on failure.
>> + */
>> +int drm_privacy_screen_set_sw_state(struct drm_privacy_screen *priv,
>> + enum drm_privacy_screen_status sw_state)
>> +{
>> + int ret = 0;
>> +
>> + mutex_lock(&priv->lock);
>> +
>> + if (!priv->ops) {
>> + ret = -ENODEV;
>> + goto out;
>> + }
>> +
>> + /*
>> + * As per the DRM connector properties documentation, setting the
>> + * sw_state while the hw_state is locked is allowed. In this case
>> + * it is a no-op other then storing the new sw_state so that it
>> + * can be honored when the state gets unlocked.
>> + * Also skip the set if the hw already is in the desired state.
>> + */
>> + if (priv->hw_state >= PRIVACY_SCREEN_DISABLED_LOCKED ||
>> + priv->hw_state == sw_state) {
>> + priv->sw_state = sw_state;
>> + goto out;
>> + }
>> +
>> + ret = priv->ops->set_sw_state(priv, sw_state);
>> +out:
>> + mutex_unlock(&priv->lock);
>> + return ret;
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_set_sw_state);
>> +
>> +/**
>> + * drm_privacy_screen_get_state - get privacy-screen's current state
>> + * @priv: privacy screen to get the state for
>> + * @sw_state_ret: address where to store the privacy-screens current sw-state
>> + * @hw_state_ret: address where to store the privacy-screens current hw-state
>> + *
>> + * Get the current state of a privacy-screen, both the sw-state and the
>> + * hw-state.
>> + */
>> +void drm_privacy_screen_get_state(struct drm_privacy_screen *priv,
>> + enum drm_privacy_screen_status *sw_state_ret,
>> + enum drm_privacy_screen_status *hw_state_ret)
>> +{
>> + mutex_lock(&priv->lock);
>> + *sw_state_ret = priv->sw_state;
>> + *hw_state_ret = priv->hw_state;
>> + mutex_unlock(&priv->lock);
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_get_state);
>> +
>> +/*** drm_privacy_screen_driver.h functions ***/
>> +
>> +static ssize_t sw_state_show(struct device *dev,
>> + struct device_attribute *attr, char *buf)
>> +{
>> + struct drm_privacy_screen *priv = to_drm_privacy_screen(dev);
>> + const char * const sw_state_names[] = {
>> + "Disabled",
>> + "Enabled",
>> + };
>> + ssize_t ret;
>> +
>> + mutex_lock(&priv->lock);
>> +
>> + if (!priv->ops)
>> + ret = -ENODEV;
>> + else if (WARN_ON(priv->sw_state >= ARRAY_SIZE(sw_state_names)))
>> + ret = -ENXIO;
>> + else
>> + ret = sprintf(buf, "%s\n", sw_state_names[priv->sw_state]);
>> +
>> + mutex_unlock(&priv->lock);
>> + return ret;
>> +}
>> +/*
>> + * RO: Do not allow setting the sw_state through sysfs, this MUST be done
>> + * through the drm_properties on the drm_connector.
>> + */
>> +static DEVICE_ATTR_RO(sw_state);
>> +
>> +static ssize_t hw_state_show(struct device *dev,
>> + struct device_attribute *attr, char *buf)
>> +{
>> + struct drm_privacy_screen *priv = to_drm_privacy_screen(dev);
>> + const char * const hw_state_names[] = {
>> + "Disabled",
>> + "Enabled",
>> + "Disabled, locked",
>> + "Enabled, locked",
>> + };
>> + ssize_t ret;
>> +
>> + mutex_lock(&priv->lock);
>> +
>> + if (!priv->ops)
>> + ret = -ENODEV;
>> + else if (WARN_ON(priv->hw_state >= ARRAY_SIZE(hw_state_names)))
>> + ret = -ENXIO;
>> + else
>> + ret = sprintf(buf, "%s\n", hw_state_names[priv->hw_state]);
>> +
>> + mutex_unlock(&priv->lock);
>> + return ret;
>> +}
>> +static DEVICE_ATTR_RO(hw_state);
>> +
>> +static struct attribute *drm_privacy_screen_attrs[] = {
>> + &dev_attr_sw_state.attr,
>> + &dev_attr_hw_state.attr,
>> + NULL
>> +};
>> +ATTRIBUTE_GROUPS(drm_privacy_screen);
>> +
>> +static struct device_type drm_privacy_screen_type = {
>> + .name = "privacy_screen",
>> + .groups = drm_privacy_screen_groups,
>> +};
>> +
>> +static void drm_privacy_screen_device_release(struct device *dev)
>> +{
>> + struct drm_privacy_screen *priv = to_drm_privacy_screen(dev);
>> +
>> + kfree(priv);
>> +}
>> +
>> +/**
>> + * drm_privacy_screen_register - register a privacy-screen
>> + * @parent: parent-device for the privacy-screen
>> + * @ops: &struct drm_privacy_screen_ops pointer with ops for the privacy-screen
>> + *
>> + * Create and register a privacy-screen.
>> + *
>> + * Return:
>> + * * A pointer to the created privacy-screen on success.
>> + * * An ERR_PTR(errno) on failure.
>> + */
>> +struct drm_privacy_screen *drm_privacy_screen_register(
>> + struct device *parent, const struct drm_privacy_screen_ops *ops)
>> +{
>> + struct drm_privacy_screen *priv;
>> + int ret;
>> +
>> + priv = kzalloc(sizeof(*priv), GFP_KERNEL);
>> + if (!priv)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + mutex_init(&priv->lock);
>> +
>> + priv->dev.class = drm_class;
>> + priv->dev.type = &drm_privacy_screen_type;
>> + priv->dev.parent = parent;
>> + priv->dev.release = drm_privacy_screen_device_release;
>> + dev_set_name(&priv->dev, "privacy_screen-%s", dev_name(parent));
>> + priv->ops = ops;
>> +
>> + priv->ops->get_hw_state(priv);
>> +
>> + ret = device_register(&priv->dev);
>> + if (ret) {
>> + put_device(&priv->dev);
>> + return ERR_PTR(ret);
>> + }
>> +
>> + mutex_lock(&drm_privacy_screen_devs_lock);
>> + list_add(&priv->list, &drm_privacy_screen_devs);
>> + mutex_unlock(&drm_privacy_screen_devs_lock);
>> +
>> + return priv;
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_register);
>> +
>> +/**
>> + * drm_privacy_screen_unregister - unregister privacy-screen
>> + * @priv: privacy-screen to unregister
>> + *
>> + * Unregister a privacy-screen registered with drm_privacy_screen_register().
>> + * May be called with a NULL or ERR_PTR, in which case it is a no-op.
>> + */
>> +void drm_privacy_screen_unregister(struct drm_privacy_screen *priv)
>> +{
>> + if (IS_ERR_OR_NULL(priv))
>> + return;
>> +
>> + mutex_lock(&drm_privacy_screen_devs_lock);
>> + list_del(&priv->list);
>> + mutex_unlock(&drm_privacy_screen_devs_lock);
>> +
>> + mutex_lock(&priv->lock);
>> + priv->ops = NULL;
>> + mutex_unlock(&priv->lock);
>> +
>> + device_unregister(&priv->dev);
>> +}
>> +EXPORT_SYMBOL(drm_privacy_screen_unregister);
>> diff --git a/include/drm/drm_privacy_screen_consumer.h b/include/drm/drm_privacy_screen_consumer.h
>> new file mode 100644
>> index 000000000000..0cbd23b0453d
>> --- /dev/null
>> +++ b/include/drm/drm_privacy_screen_consumer.h
>> @@ -0,0 +1,50 @@
>> +/* SPDX-License-Identifier: MIT */
>> +/*
>> + * Copyright (C) 2020 Red Hat, Inc.
>> + *
>> + * Authors:
>> + * Hans de Goede <hdegoede at redhat.com>
>> + */
>> +
>> +#ifndef __DRM_PRIVACY_SCREEN_CONSUMER_H__
>> +#define __DRM_PRIVACY_SCREEN_CONSUMER_H__
>> +
>> +#include <linux/device.h>
>> +#include <drm/drm_connector.h>
>> +
>> +struct drm_privacy_screen;
>> +
>> +#if IS_ENABLED(CONFIG_DRM_PRIVACY_SCREEN)
>> +struct drm_privacy_screen *drm_privacy_screen_get(struct device *dev,
>> + const char *con_id);
>> +void drm_privacy_screen_put(struct drm_privacy_screen *priv);
>> +
>> +int drm_privacy_screen_set_sw_state(struct drm_privacy_screen *priv,
>> + enum drm_privacy_screen_status sw_state);
>> +void drm_privacy_screen_get_state(struct drm_privacy_screen *priv,
>> + enum drm_privacy_screen_status *sw_state_ret,
>> + enum drm_privacy_screen_status *hw_state_ret);
>> +#else
>> +static inline struct drm_privacy_screen *drm_privacy_screen_get(struct device *dev,
>> + const char *con_id)
>> +{
>> + return ERR_PTR(-ENODEV);
>> +}
>> +static inline void drm_privacy_screen_put(struct drm_privacy_screen *priv)
>> +{
>> +}
>> +static inline int drm_privacy_screen_set_sw_state(struct drm_privacy_screen *priv,
>> + enum drm_privacy_screen_status sw_state)
>> +{
>> + return -ENODEV;
>> +}
>> +static inline void drm_privacy_screen_get_state(struct drm_privacy_screen *priv,
>> + enum drm_privacy_screen_status *sw_state_ret,
>> + enum drm_privacy_screen_status *hw_state_ret)
>> +{
>> + *sw_state_ret = PRIVACY_SCREEN_DISABLED;
>> + *hw_state_ret = PRIVACY_SCREEN_DISABLED;
>> +}
>> +#endif
>> +
>> +#endif
>> diff --git a/include/drm/drm_privacy_screen_driver.h b/include/drm/drm_privacy_screen_driver.h
>> new file mode 100644
>> index 000000000000..5187ae52eb03
>> --- /dev/null
>> +++ b/include/drm/drm_privacy_screen_driver.h
>> @@ -0,0 +1,80 @@
>> +/* SPDX-License-Identifier: MIT */
>> +/*
>> + * Copyright (C) 2020 Red Hat, Inc.
>> + *
>> + * Authors:
>> + * Hans de Goede <hdegoede at redhat.com>
>> + */
>> +
>> +#ifndef __DRM_PRIVACY_SCREEN_DRIVER_H__
>> +#define __DRM_PRIVACY_SCREEN_DRIVER_H__
>> +
>> +#include <linux/device.h>
>> +#include <linux/list.h>
>> +#include <linux/mutex.h>
>> +#include <drm/drm_connector.h>
>> +
>> +struct drm_privacy_screen;
>> +
>> +/**
>> + * struct drm_privacy_screen_ops - drm_privacy_screen operations
>> + *
>> + * Defines the operations which the privacy-screen class code may call.
>> + * These functions should be implemented by the privacy-screen driver.
>> + */
>> +struct drm_privacy_screen_ops {
>> + /**
>> + * @set_sw_state: Called to request a change of the privacy-screen
>> + * state. The privacy-screen class code contains a check to avoid this
>> + * getting called when the hw_state reports the state is locked.
>> + * It is the driver's responsibility to update sw_state and hw_state.
>> + * This is always called with the drm_privacy_screen's lock held.
>> + */
>> + int (*set_sw_state)(struct drm_privacy_screen *priv,
>> + enum drm_privacy_screen_status sw_state);
>> + /**
>> + * @get_hw_state: Called to request that the driver gets the current
>> + * privacy-screen state from the hardware and then updates sw_state and
>> + * hw_state accordingly. This will be called by the core just before
>> + * the privacy-screen is registered in sysfs.
>> + */
>> + void (*get_hw_state)(struct drm_privacy_screen *priv);
>> +};
>> +
>> +/**
>> + * struct drm_privacy_screen - central privacy-screen structure
>> + *
>> + * Central privacy-screen structure, this contains the struct device used
>> + * to register the screen in sysfs, the screen's state, ops, etc.
>> + */
>> +struct drm_privacy_screen {
>> + /** @dev: device used to register the privacy-screen in sysfs. */
>> + struct device dev;
>> + /** @lock: mutex protection all fields in this struct. */
>> + struct mutex lock;
>> + /** @list: privacy-screen devices list list-entry. */
>> + struct list_head list;
>> + /**
>> + * @ops: &struct drm_privacy_screen_ops for this privacy-screen.
>> + * This is NULL if the driver has unregistered the privacy-screen.
>> + */
>> + const struct drm_privacy_screen_ops *ops;
>> + /**
>> + * @sw_state: The privacy-screen's software state, see
>> + * :ref:`Standard Connector Properties<standard_connector_properties>`
>> + * for more info.
>> + */
>> + enum drm_privacy_screen_status sw_state;
>> + /**
>> + * @hw_state: The privacy-screen's hardware state, see
>> + * :ref:`Standard Connector Properties<standard_connector_properties>`
>> + * for more info.
>> + */
>> + enum drm_privacy_screen_status hw_state;
>> +};
>> +
>> +struct drm_privacy_screen *drm_privacy_screen_register(
>> + struct device *parent, const struct drm_privacy_screen_ops *ops);
>> +void drm_privacy_screen_unregister(struct drm_privacy_screen *priv);
>> +
>> +#endif
>> diff --git a/include/drm/drm_privacy_screen_machine.h b/include/drm/drm_privacy_screen_machine.h
>> new file mode 100644
>> index 000000000000..aaa0d38cce92
>> --- /dev/null
>> +++ b/include/drm/drm_privacy_screen_machine.h
>> @@ -0,0 +1,41 @@
>> +/* SPDX-License-Identifier: MIT */
>> +/*
>> + * Copyright (C) 2020 Red Hat, Inc.
>> + *
>> + * Authors:
>> + * Hans de Goede <hdegoede at redhat.com>
>> + */
>> +
>> +#ifndef __DRM_PRIVACY_SCREEN_MACHINE_H__
>> +#define __DRM_PRIVACY_SCREEN_MACHINE_H__
>> +
>> +#include <linux/list.h>
>> +
>> +/**
>> + * struct drm_privacy_screen_lookup - static privacy-screen lookup list entry
>> + *
>> + * Used for the static lookup-list for mapping privacy-screen consumer
>> + * dev-connector pairs to a privacy-screen provider.
>> + */
>> +struct drm_privacy_screen_lookup {
>> + /** @list: Lookup list list-entry. */
>> + struct list_head list;
>> + /** @dev_id: Consumer device name or NULL to match all devices. */
>> + const char *dev_id;
>> + /** @con_id: Consumer connector name or NULL to match all connectors. */
>> + const char *con_id;
>> + /** @provider: dev_name() of the privacy_screen provider. */
>> + const char *provider;
>> +};
>> +
>> +void drm_privacy_screen_lookup_add(struct drm_privacy_screen_lookup *lookup);
>> +void drm_privacy_screen_lookup_remove(struct drm_privacy_screen_lookup *lookup);
>> +
>> +static inline void drm_privacy_screen_lookup_init(void)
>> +{
>> +}
>> +static inline void drm_privacy_screen_lookup_exit(void)
>> +{
>> +}
>> +
>> +#endif
>>
>
More information about the Intel-gfx
mailing list