1
0
Fork 0
alistair23-linux/include/drm/drm_connector.h

958 lines
31 KiB
C
Raw Normal View History

/*
* Copyright (c) 2016 Intel Corporation
*
* Permission to use, copy, modify, distribute, and sell this software and its
* documentation for any purpose is hereby granted without fee, provided that
* the above copyright notice appear in all copies and that both that copyright
* notice and this permission notice appear in supporting documentation, and
* that the name of the copyright holders not be used in advertising or
* publicity pertaining to distribution of the software without specific,
* written prior permission. The copyright holders make no representations
* about the suitability of this software for any purpose. It is provided "as
* is" without express or implied warranty.
*
* THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
* INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
* EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
* CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
* DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
* TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
* OF THIS SOFTWARE.
*/
#ifndef __DRM_CONNECTOR_H__
#define __DRM_CONNECTOR_H__
#include <linux/list.h>
#include <linux/ctype.h>
#include <drm/drm_mode_object.h>
#include <uapi/drm/drm_mode.h>
struct drm_device;
struct drm_connector_helper_funcs;
struct drm_device;
struct drm_crtc;
struct drm_encoder;
struct drm_property;
struct drm_property_blob;
struct drm_printer;
struct edid;
enum drm_connector_force {
DRM_FORCE_UNSPECIFIED,
DRM_FORCE_OFF,
DRM_FORCE_ON, /* force on analog part normally */
DRM_FORCE_ON_DIGITAL, /* for DVI-I use digital connector */
};
/**
* enum drm_connector_status - status for a &drm_connector
*
* This enum is used to track the connector status. There are no separate
* #defines for the uapi!
*/
enum drm_connector_status {
/**
* @connector_status_connected: The connector is definitely connected to
* a sink device, and can be enabled.
*/
connector_status_connected = 1,
/**
* @connector_status_disconnected: The connector isn't connected to a
* sink device which can be autodetect. For digital outputs like DP or
* HDMI (which can be realiable probed) this means there's really
* nothing there. It is driver-dependent whether a connector with this
* status can be lit up or not.
*/
connector_status_disconnected = 2,
/**
* @connector_status_unknown: The connector's status could not be
* reliably detected. This happens when probing would either cause
* flicker (like load-detection when the connector is in use), or when a
* hardware resource isn't available (like when load-detection needs a
* free CRTC). It should be possible to light up the connector with one
* of the listed fallback modes. For default configuration userspace
* should only try to light up connectors with unknown status when
* there's not connector with @connector_status_connected.
*/
connector_status_unknown = 3,
};
enum subpixel_order {
SubPixelUnknown = 0,
SubPixelHorizontalRGB,
SubPixelHorizontalBGR,
SubPixelVerticalRGB,
SubPixelVerticalBGR,
SubPixelNone,
};
drm: Add a new connector atomic property for link status At the time userspace does setcrtc, we've already promised the mode would work. The promise is based on the theoretical capabilities of the link, but it's possible we can't reach this in practice. The DP spec describes how the link should be reduced, but we can't reduce the link below the requirements of the mode. Black screen follows. One idea would be to have setcrtc return a failure. However, it already should not fail as the atomic checks have passed. It would also conflict with the idea of making setcrtc asynchronous in the future, returning before the actual mode setting and link training. Another idea is to train the link "upfront" at hotplug time, before pruning the mode list, so that we can do the pruning based on practical not theoretical capabilities. However, the changes for link training are pretty drastic, all for the sake of error handling and DP compliance, when the most common happy day scenario is the current approach of link training at mode setting time, using the optimal parameters for the mode. It is also not certain all hardware could do this without the pipe on; not even all our hardware can do this. Some of this can be solved, but not trivially. Both of the above ideas also fail to address link degradation *during* operation. The solution is to add a new "link-status" connector property in order to address link training failure in a way that: a) changes the current happy day scenario as little as possible, to avoid regressions, b) can be implemented the same way by all drm drivers, c) is still opt-in for the drivers and userspace, and opting out doesn't regress the user experience, d) doesn't prevent drivers from implementing better or alternate approaches, possibly without userspace involvement. And, of course, handles all the issues presented. In the usual happy day scenario, this is always "good". If something fails during or after a mode set, the kernel driver can set the link status to "bad" and issue a hotplug uevent for userspace to have it re-check the valid modes through GET_CONNECTOR IOCTL, and try modeset again. If the theoretical capabilities of the link can't be reached, the mode list is trimmed based on that. v7 by Jani: * Rebase, simplify set property while at it, checkpatch fix v6: * Fix a typo in kernel doc (Sean Paul) v5: * Clarify doc for silent rejection of atomic properties by driver (Daniel Vetter) v4: * Add comments in kernel-doc format (Daniel Vetter) * Update the kernel-doc for link-status (Sean Paul) v3: * Fixed a build error (Jani Saarinen) v2: * Removed connector->link_status (Daniel Vetter) * Set connector->state->link_status in drm_mode_connector_set_link_status_property (Daniel Vetter) * Set the connector_changed flag to true if connector->state->link_status changed. * Reset link_status to GOOD in update_output_state (Daniel Vetter) * Never allow userspace to set link status from Good To Bad (Daniel Vetter) Reviewed-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Acked-by: Tony Cheng <tony.cheng@amd.com> Acked-by: Harry Wentland <harry.wentland@amd.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Sean Paul <seanpaul@chromium.org> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Acked-by: Eric Anholt <eric@anholt.net> (for the -modesetting patch) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/0182487051aa9f1594820e35a4853de2f8747b4e.1481883920.git.jani.nikula@intel.com
2016-12-16 03:29:06 -07:00
/**
* enum drm_link_status - connector's link_status property value
*
* This enum is used as the connector's link status property value.
* It is set to the values defined in uapi.
*/
enum drm_link_status {
DRM_LINK_STATUS_GOOD = DRM_MODE_LINK_STATUS_GOOD,
DRM_LINK_STATUS_BAD = DRM_MODE_LINK_STATUS_BAD,
};
/**
* struct drm_display_info - runtime data about the connected sink
*
* Describes a given display (e.g. CRT or flat panel) and its limitations. For
* fixed display sinks like built-in panels there's not much difference between
* this and &struct drm_connector. But for sinks with a real cable this
* structure is meant to describe all the things at the other end of the cable.
*
* For sinks which provide an EDID this can be filled out by calling
* drm_add_edid_modes().
*/
struct drm_display_info {
/**
* @name: Name of the display.
*/
char name[DRM_DISPLAY_INFO_LEN];
/**
* @width_mm: Physical width in mm.
*/
unsigned int width_mm;
/**
* @height_mm: Physical height in mm.
*/
unsigned int height_mm;
/**
* @pixel_clock: Maximum pixel clock supported by the sink, in units of
* 100Hz. This mismatches the clock in &drm_display_mode (which is in
* kHZ), because that's what the EDID uses as base unit.
*/
unsigned int pixel_clock;
/**
* @bpc: Maximum bits per color channel. Used by HDMI and DP outputs.
*/
unsigned int bpc;
/**
* @subpixel_order: Subpixel order of LCD panels.
*/
enum subpixel_order subpixel_order;
#define DRM_COLOR_FORMAT_RGB444 (1<<0)
#define DRM_COLOR_FORMAT_YCRCB444 (1<<1)
#define DRM_COLOR_FORMAT_YCRCB422 (1<<2)
/**
* @color_formats: HDMI Color formats, selects between RGB and YCrCb
* modes. Used DRM_COLOR_FORMAT\_ defines, which are _not_ the same ones
* as used to describe the pixel format in framebuffers, and also don't
* match the formats in @bus_formats which are shared with v4l.
*/
u32 color_formats;
/**
* @bus_formats: Pixel data format on the wire, somewhat redundant with
* @color_formats. Array of size @num_bus_formats encoded using
* MEDIA_BUS_FMT\_ defines shared with v4l and media drivers.
*/
const u32 *bus_formats;
/**
* @num_bus_formats: Size of @bus_formats array.
*/
unsigned int num_bus_formats;
#define DRM_BUS_FLAG_DE_LOW (1<<0)
#define DRM_BUS_FLAG_DE_HIGH (1<<1)
/* drive data on pos. edge */
#define DRM_BUS_FLAG_PIXDATA_POSEDGE (1<<2)
/* drive data on neg. edge */
#define DRM_BUS_FLAG_PIXDATA_NEGEDGE (1<<3)
/**
* @bus_flags: Additional information (like pixel signal polarity) for
* the pixel data on the bus, using DRM_BUS_FLAGS\_ defines.
*/
u32 bus_flags;
/**
* @max_tmds_clock: Maximum TMDS clock rate supported by the
* sink in kHz. 0 means undefined.
*/
int max_tmds_clock;
/**
* @dvi_dual: Dual-link DVI sink?
*/
bool dvi_dual;
/**
* @edid_hdmi_dc_modes: Mask of supported hdmi deep color modes. Even
* more stuff redundant with @bus_formats.
*/
u8 edid_hdmi_dc_modes;
/**
* @cea_rev: CEA revision of the HDMI sink.
*/
u8 cea_rev;
};
int drm_display_info_set_bus_formats(struct drm_display_info *info,
const u32 *formats,
unsigned int num_formats);
/**
* struct drm_tv_connector_state - TV connector related states
* @subconnector: selected subconnector
* @margins: left/right/top/bottom margins
* @mode: TV mode
* @brightness: brightness in percent
* @contrast: contrast in percent
* @flicker_reduction: flicker reduction in percent
* @overscan: overscan in percent
* @saturation: saturation in percent
* @hue: hue in percent
*/
struct drm_tv_connector_state {
enum drm_mode_subconnector subconnector;
struct {
unsigned int left;
unsigned int right;
unsigned int top;
unsigned int bottom;
} margins;
unsigned int mode;
unsigned int brightness;
unsigned int contrast;
unsigned int flicker_reduction;
unsigned int overscan;
unsigned int saturation;
unsigned int hue;
};
/**
* struct drm_connector_state - mutable connector state
* @connector: backpointer to the connector
* @best_encoder: can be used by helpers and drivers to select the encoder
* @state: backpointer to global drm_atomic_state
* @tv: TV connector state
*/
struct drm_connector_state {
struct drm_connector *connector;
/**
* @crtc: CRTC to connect connector to, NULL if disabled.
*
* Do not change this directly, use drm_atomic_set_crtc_for_connector()
* instead.
*/
struct drm_crtc *crtc;
struct drm_encoder *best_encoder;
drm: Add a new connector atomic property for link status At the time userspace does setcrtc, we've already promised the mode would work. The promise is based on the theoretical capabilities of the link, but it's possible we can't reach this in practice. The DP spec describes how the link should be reduced, but we can't reduce the link below the requirements of the mode. Black screen follows. One idea would be to have setcrtc return a failure. However, it already should not fail as the atomic checks have passed. It would also conflict with the idea of making setcrtc asynchronous in the future, returning before the actual mode setting and link training. Another idea is to train the link "upfront" at hotplug time, before pruning the mode list, so that we can do the pruning based on practical not theoretical capabilities. However, the changes for link training are pretty drastic, all for the sake of error handling and DP compliance, when the most common happy day scenario is the current approach of link training at mode setting time, using the optimal parameters for the mode. It is also not certain all hardware could do this without the pipe on; not even all our hardware can do this. Some of this can be solved, but not trivially. Both of the above ideas also fail to address link degradation *during* operation. The solution is to add a new "link-status" connector property in order to address link training failure in a way that: a) changes the current happy day scenario as little as possible, to avoid regressions, b) can be implemented the same way by all drm drivers, c) is still opt-in for the drivers and userspace, and opting out doesn't regress the user experience, d) doesn't prevent drivers from implementing better or alternate approaches, possibly without userspace involvement. And, of course, handles all the issues presented. In the usual happy day scenario, this is always "good". If something fails during or after a mode set, the kernel driver can set the link status to "bad" and issue a hotplug uevent for userspace to have it re-check the valid modes through GET_CONNECTOR IOCTL, and try modeset again. If the theoretical capabilities of the link can't be reached, the mode list is trimmed based on that. v7 by Jani: * Rebase, simplify set property while at it, checkpatch fix v6: * Fix a typo in kernel doc (Sean Paul) v5: * Clarify doc for silent rejection of atomic properties by driver (Daniel Vetter) v4: * Add comments in kernel-doc format (Daniel Vetter) * Update the kernel-doc for link-status (Sean Paul) v3: * Fixed a build error (Jani Saarinen) v2: * Removed connector->link_status (Daniel Vetter) * Set connector->state->link_status in drm_mode_connector_set_link_status_property (Daniel Vetter) * Set the connector_changed flag to true if connector->state->link_status changed. * Reset link_status to GOOD in update_output_state (Daniel Vetter) * Never allow userspace to set link status from Good To Bad (Daniel Vetter) Reviewed-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Acked-by: Tony Cheng <tony.cheng@amd.com> Acked-by: Harry Wentland <harry.wentland@amd.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Sean Paul <seanpaul@chromium.org> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Acked-by: Eric Anholt <eric@anholt.net> (for the -modesetting patch) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/0182487051aa9f1594820e35a4853de2f8747b4e.1481883920.git.jani.nikula@intel.com
2016-12-16 03:29:06 -07:00
/**
* @link_status: Connector link_status to keep track of whether link is
* GOOD or BAD to notify userspace if retraining is necessary.
*/
enum drm_link_status link_status;
struct drm_atomic_state *state;
struct drm_tv_connector_state tv;
};
/**
* struct drm_connector_funcs - control connectors on a given device
*
* Each CRTC may have one or more connectors attached to it. The functions
* below allow the core DRM code to control connectors, enumerate available modes,
* etc.
*/
struct drm_connector_funcs {
/**
* @dpms:
*
* Legacy entry point to set the per-connector DPMS state. Legacy DPMS
* is exposed as a standard property on the connector, but diverted to
* this callback in the drm core. Note that atomic drivers don't
* implement the 4 level DPMS support on the connector any more, but
* instead only have an on/off "ACTIVE" property on the CRTC object.
*
* Drivers implementing atomic modeset should use
* drm_atomic_helper_connector_dpms() to implement this hook.
*
* RETURNS:
*
* 0 on success or a negative error code on failure.
*/
int (*dpms)(struct drm_connector *connector, int mode);
/**
* @reset:
*
* Reset connector hardware and software state to off. This function isn't
* called by the core directly, only through drm_mode_config_reset().
* It's not a helper hook only for historical reasons.
*
* Atomic drivers can use drm_atomic_helper_connector_reset() to reset
* atomic state using this hook.
*/
void (*reset)(struct drm_connector *connector);
/**
* @detect:
*
* Check to see if anything is attached to the connector. The parameter
* force is set to false whilst polling, true when checking the
* connector due to a user request. force can be used by the driver to
* avoid expensive, destructive operations during automated probing.
*
* This callback is optional, if not implemented the connector will be
* considered as always being attached.
*
* FIXME:
*
* Note that this hook is only called by the probe helper. It's not in
* the helper library vtable purely for historical reasons. The only DRM
* core entry point to probe connector state is @fill_modes.
*
* RETURNS:
*
* drm_connector_status indicating the connector's status.
*/
enum drm_connector_status (*detect)(struct drm_connector *connector,
bool force);
/**
* @force:
*
* This function is called to update internal encoder state when the
* connector is forced to a certain state by userspace, either through
* the sysfs interfaces or on the kernel cmdline. In that case the
* @detect callback isn't called.
*
* FIXME:
*
* Note that this hook is only called by the probe helper. It's not in
* the helper library vtable purely for historical reasons. The only DRM
* core entry point to probe connector state is @fill_modes.
*/
void (*force)(struct drm_connector *connector);
/**
* @fill_modes:
*
* Entry point for output detection and basic mode validation. The
* driver should reprobe the output if needed (e.g. when hotplug
* handling is unreliable), add all detected modes to &drm_connector.modes
* and filter out any the device can't support in any configuration. It
* also needs to filter out any modes wider or higher than the
* parameters max_width and max_height indicate.
*
* The drivers must also prune any modes no longer valid from
* &drm_connector.modes. Furthermore it must update
* &drm_connector.status and &drm_connector.edid. If no EDID has been
* received for this output connector->edid must be NULL.
*
* Drivers using the probe helpers should use
* drm_helper_probe_single_connector_modes() or
* drm_helper_probe_single_connector_modes_nomerge() to implement this
* function.
*
* RETURNS:
*
* The number of modes detected and filled into &drm_connector.modes.
*/
int (*fill_modes)(struct drm_connector *connector, uint32_t max_width, uint32_t max_height);
/**
* @set_property:
*
* This is the legacy entry point to update a property attached to the
* connector.
*
* Drivers implementing atomic modeset should use
* drm_atomic_helper_connector_set_property() to implement this hook.
*
* This callback is optional if the driver does not support any legacy
* driver-private properties.
*
* RETURNS:
*
* 0 on success or a negative error code on failure.
*/
int (*set_property)(struct drm_connector *connector, struct drm_property *property,
uint64_t val);
/**
* @late_register:
*
* This optional hook can be used to register additional userspace
* interfaces attached to the connector, light backlight control, i2c,
* DP aux or similar interfaces. It is called late in the driver load
* sequence from drm_connector_register() when registering all the
* core drm connector interfaces. Everything added from this callback
* should be unregistered in the early_unregister callback.
*
* This is called while holding &drm_connector.mutex.
*
* Returns:
*
* 0 on success, or a negative error code on failure.
*/
int (*late_register)(struct drm_connector *connector);
/**
* @early_unregister:
*
* This optional hook should be used to unregister the additional
* userspace interfaces attached to the connector from
* late_register(). It is called from drm_connector_unregister(),
* early in the driver unload sequence to disable userspace access
* before data structures are torndown.
*
* This is called while holding &drm_connector.mutex.
*/
void (*early_unregister)(struct drm_connector *connector);
/**
* @destroy:
*
* Clean up connector resources. This is called at driver unload time
* through drm_mode_config_cleanup(). It can also be called at runtime
* when a connector is being hot-unplugged for drivers that support
* connector hotplugging (e.g. DisplayPort MST).
*/
void (*destroy)(struct drm_connector *connector);
/**
* @atomic_duplicate_state:
*
* Duplicate the current atomic state for this connector and return it.
* The core and helpers guarantee that any atomic state duplicated with
* this hook and still owned by the caller (i.e. not transferred to the
* driver by calling &drm_mode_config_funcs.atomic_commit) will be
* cleaned up by calling the @atomic_destroy_state hook in this
* structure.
*
* Atomic drivers which don't subclass &struct drm_connector_state should use
* drm_atomic_helper_connector_duplicate_state(). Drivers that subclass the
* state structure to extend it with driver-private state should use
* __drm_atomic_helper_connector_duplicate_state() to make sure shared state is
* duplicated in a consistent fashion across drivers.
*
* It is an error to call this hook before &drm_connector.state has been
* initialized correctly.
*
* NOTE:
*
* If the duplicate state references refcounted resources this hook must
* acquire a reference for each of them. The driver must release these
* references again in @atomic_destroy_state.
*
* RETURNS:
*
* Duplicated atomic state or NULL when the allocation failed.
*/
struct drm_connector_state *(*atomic_duplicate_state)(struct drm_connector *connector);
/**
* @atomic_destroy_state:
*
* Destroy a state duplicated with @atomic_duplicate_state and release
* or unreference all resources it references
*/
void (*atomic_destroy_state)(struct drm_connector *connector,
struct drm_connector_state *state);
/**
* @atomic_set_property:
*
* Decode a driver-private property value and store the decoded value
* into the passed-in state structure. Since the atomic core decodes all
* standardized properties (even for extensions beyond the core set of
* properties which might not be implemented by all drivers) this
* requires drivers to subclass the state structure.
*
* Such driver-private properties should really only be implemented for
* truly hardware/vendor specific state. Instead it is preferred to
* standardize atomic extension and decode the properties used to expose
* such an extension in the core.
*
* Do not call this function directly, use
* drm_atomic_connector_set_property() instead.
*
* This callback is optional if the driver does not support any
* driver-private atomic properties.
*
* NOTE:
*
* This function is called in the state assembly phase of atomic
* modesets, which can be aborted for any reason (including on
* userspace's request to just check whether a configuration would be
* possible). Drivers MUST NOT touch any persistent state (hardware or
* software) or data structures except the passed in @state parameter.
*
* Also since userspace controls in which order properties are set this
* function must not do any input validation (since the state update is
* incomplete and hence likely inconsistent). Instead any such input
* validation must be done in the various atomic_check callbacks.
*
* RETURNS:
*
* 0 if the property has been found, -EINVAL if the property isn't
* implemented by the driver (which shouldn't ever happen, the core only
* asks for properties attached to this connector). No other validation
* is allowed by the driver. The core already checks that the property
* value is within the range (integer, valid enum value, ...) the driver
* set when registering the property.
*/
int (*atomic_set_property)(struct drm_connector *connector,
struct drm_connector_state *state,
struct drm_property *property,
uint64_t val);
/**
* @atomic_get_property:
*
* Reads out the decoded driver-private property. This is used to
* implement the GETCONNECTOR IOCTL.
*
* Do not call this function directly, use
* drm_atomic_connector_get_property() instead.
*
* This callback is optional if the driver does not support any
* driver-private atomic properties.
*
* RETURNS:
*
* 0 on success, -EINVAL if the property isn't implemented by the
* driver (which shouldn't ever happen, the core only asks for
* properties attached to this connector).
*/
int (*atomic_get_property)(struct drm_connector *connector,
const struct drm_connector_state *state,
struct drm_property *property,
uint64_t *val);
/**
* @atomic_print_state:
*
* If driver subclasses &struct drm_connector_state, it should implement
* this optional hook for printing additional driver specific state.
*
* Do not call this directly, use drm_atomic_connector_print_state()
* instead.
*/
void (*atomic_print_state)(struct drm_printer *p,
const struct drm_connector_state *state);
};
/* mode specified on the command line */
struct drm_cmdline_mode {
bool specified;
bool refresh_specified;
bool bpp_specified;
int xres, yres;
int bpp;
int refresh;
bool rb;
bool interlace;
bool cvt;
bool margins;
enum drm_connector_force force;
};
/**
* struct drm_connector - central DRM connector control structure
* @dev: parent DRM device
* @kdev: kernel device for sysfs attributes
* @attr: sysfs attributes
* @head: list management
* @base: base KMS object
* @name: human readable name, can be overwritten by the driver
* @connector_type: one of the DRM_MODE_CONNECTOR_<foo> types from drm_mode.h
* @connector_type_id: index into connector type enum
* @interlace_allowed: can this connector handle interlaced modes?
* @doublescan_allowed: can this connector handle doublescan?
* @stereo_allowed: can this connector handle stereo modes?
* @funcs: connector control functions
* @edid_blob_ptr: DRM property containing EDID if present
* @properties: property tracking for this connector
* @dpms: current dpms state
* @helper_private: mid-layer private data
* @cmdline_mode: mode line parsed from the kernel cmdline for this connector
* @force: a DRM_FORCE_<foo> state for forced mode sets
* @override_edid: has the EDID been overwritten through debugfs for testing?
* @encoder_ids: valid encoders for this connector
* @encoder: encoder driving this connector, if any
* @eld: EDID-like data, if present
* @latency_present: AV delay info from ELD, if found
* @video_latency: video latency info from ELD, if found
* @audio_latency: audio latency info from ELD, if found
* @null_edid_counter: track sinks that give us all zeros for the EDID
* @bad_edid_counter: track sinks that give us an EDID with invalid checksum
* @edid_corrupt: indicates whether the last read EDID was corrupt
* @debugfs_entry: debugfs directory for this connector
* @state: current atomic state for this connector
* @has_tile: is this connector connected to a tiled monitor
* @tile_group: tile group for the connected monitor
* @tile_is_single_monitor: whether the tile is one monitor housing
* @num_h_tile: number of horizontal tiles in the tile group
* @num_v_tile: number of vertical tiles in the tile group
* @tile_h_loc: horizontal location of this tile
* @tile_v_loc: vertical location of this tile
* @tile_h_size: horizontal size of this tile.
* @tile_v_size: vertical size of this tile.
*
* Each connector may be connected to one or more CRTCs, or may be clonable by
* another connector if they can share a CRTC. Each connector also has a specific
* position in the broader display (referred to as a 'screen' though it could
* span multiple monitors).
*/
struct drm_connector {
struct drm_device *dev;
struct device *kdev;
struct device_attribute *attr;
struct list_head head;
struct drm_mode_object base;
char *name;
/**
* @mutex: Lock for general connector state, but currently only protects
* @registered. Most of the connector state is still protected by
* &drm_mode_config.mutex.
*/
struct mutex mutex;
/**
* @index: Compacted connector index, which matches the position inside
* the mode_config.list for drivers not supporting hot-add/removing. Can
* be used as an array index. It is invariant over the lifetime of the
* connector.
*/
unsigned index;
int connector_type;
int connector_type_id;
bool interlace_allowed;
bool doublescan_allowed;
bool stereo_allowed;
/**
* @registered: Is this connector exposed (registered) with userspace?
* Protected by @mutex.
*/
bool registered;
/**
* @modes:
* Modes available on this connector (from fill_modes() + user).
* Protected by &drm_mode_config.mutex.
*/
struct list_head modes;
/**
* @status:
* One of the drm_connector_status enums (connected, not, or unknown).
* Protected by &drm_mode_config.mutex.
*/
enum drm_connector_status status;
/**
* @probed_modes:
* These are modes added by probing with DDC or the BIOS, before
* filtering is applied. Used by the probe helpers. Protected by
* &drm_mode_config.mutex.
*/
struct list_head probed_modes;
/**
* @display_info: Display information is filled from EDID information
* when a display is detected. For non hot-pluggable displays such as
* flat panels in embedded systems, the driver should initialize the
* &drm_display_info.width_mm and &drm_display_info.height_mm fields
* with the physical size of the display.
*
* Protected by &drm_mode_config.mutex.
*/
struct drm_display_info display_info;
const struct drm_connector_funcs *funcs;
struct drm_property_blob *edid_blob_ptr;
struct drm_object_properties properties;
/**
* @path_blob_ptr:
*
* DRM blob property data for the DP MST path property.
*/
struct drm_property_blob *path_blob_ptr;
/**
* @tile_blob_ptr:
*
* DRM blob property data for the tile property (used mostly by DP MST).
* This is meant for screens which are driven through separate display
* pipelines represented by &drm_crtc, which might not be running with
* genlocked clocks. For tiled panels which are genlocked, like
* dual-link LVDS or dual-link DSI, the driver should try to not expose
* the tiling and virtualize both &drm_crtc and &drm_plane if needed.
*/
struct drm_property_blob *tile_blob_ptr;
/* should we poll this connector for connects and disconnects */
/* hot plug detectable */
#define DRM_CONNECTOR_POLL_HPD (1 << 0)
/* poll for connections */
#define DRM_CONNECTOR_POLL_CONNECT (1 << 1)
/* can cleanly poll for disconnections without flickering the screen */
/* DACs should rarely do this without a lot of testing */
#define DRM_CONNECTOR_POLL_DISCONNECT (1 << 2)
/**
* @polled:
*
* Connector polling mode, a combination of
*
* DRM_CONNECTOR_POLL_HPD
* The connector generates hotplug events and doesn't need to be
* periodically polled. The CONNECT and DISCONNECT flags must not
* be set together with the HPD flag.
*
* DRM_CONNECTOR_POLL_CONNECT
* Periodically poll the connector for connection.
*
* DRM_CONNECTOR_POLL_DISCONNECT
* Periodically poll the connector for disconnection.
*
* Set to 0 for connectors that don't support connection status
* discovery.
*/
uint8_t polled;
/* requested DPMS state */
int dpms;
const struct drm_connector_helper_funcs *helper_private;
/* forced on connector */
struct drm_cmdline_mode cmdline_mode;
enum drm_connector_force force;
bool override_edid;
#define DRM_CONNECTOR_MAX_ENCODER 3
uint32_t encoder_ids[DRM_CONNECTOR_MAX_ENCODER];
struct drm_encoder *encoder; /* currently active encoder */
#define MAX_ELD_BYTES 128
/* EDID bits */
uint8_t eld[MAX_ELD_BYTES];
bool latency_present[2];
int video_latency[2]; /* [0]: progressive, [1]: interlaced */
int audio_latency[2];
int null_edid_counter; /* needed to workaround some HW bugs where we get all 0s */
unsigned bad_edid_counter;
/* Flag for raw EDID header corruption - used in Displayport
* compliance testing - * Displayport Link CTS Core 1.2 rev1.1 4.2.2.6
*/
bool edid_corrupt;
struct dentry *debugfs_entry;
struct drm_connector_state *state;
/* DisplayID bits */
bool has_tile;
struct drm_tile_group *tile_group;
bool tile_is_single_monitor;
uint8_t num_h_tile, num_v_tile;
uint8_t tile_h_loc, tile_v_loc;
uint16_t tile_h_size, tile_v_size;
};
#define obj_to_connector(x) container_of(x, struct drm_connector, base)
int drm_connector_init(struct drm_device *dev,
struct drm_connector *connector,
const struct drm_connector_funcs *funcs,
int connector_type);
int drm_connector_register(struct drm_connector *connector);
void drm_connector_unregister(struct drm_connector *connector);
int drm_mode_connector_attach_encoder(struct drm_connector *connector,
struct drm_encoder *encoder);
void drm_connector_cleanup(struct drm_connector *connector);
static inline unsigned drm_connector_index(struct drm_connector *connector)
{
return connector->index;
}
/**
* drm_connector_lookup - lookup connector object
* @dev: DRM device
* @id: connector object id
*
* This function looks up the connector object specified by id
* add takes a reference to it.
*/
static inline struct drm_connector *drm_connector_lookup(struct drm_device *dev,
uint32_t id)
{
struct drm_mode_object *mo;
mo = drm_mode_object_find(dev, id, DRM_MODE_OBJECT_CONNECTOR);
return mo ? obj_to_connector(mo) : NULL;
}
/**
* drm_connector_get - acquire a connector reference
* @connector: DRM connector
*
* This function increments the connector's refcount.
*/
static inline void drm_connector_get(struct drm_connector *connector)
{
drm_mode_object_get(&connector->base);
}
/**
* drm_connector_put - release a connector reference
* @connector: DRM connector
*
* This function decrements the connector's reference count and frees the
* object if the reference count drops to zero.
*/
static inline void drm_connector_put(struct drm_connector *connector)
{
drm_mode_object_put(&connector->base);
}
/**
* drm_connector_reference - acquire a connector reference
* @connector: DRM connector
*
* This is a compatibility alias for drm_connector_get() and should not be
* used by new code.
*/
static inline void drm_connector_reference(struct drm_connector *connector)
{
drm_connector_get(connector);
}
/**
* drm_connector_unreference - release a connector reference
* @connector: DRM connector
*
* This is a compatibility alias for drm_connector_put() and should not be
* used by new code.
*/
static inline void drm_connector_unreference(struct drm_connector *connector)
{
drm_connector_put(connector);
}
const char *drm_get_connector_status_name(enum drm_connector_status status);
const char *drm_get_subpixel_order_name(enum subpixel_order order);
const char *drm_get_dpms_name(int val);
const char *drm_get_dvi_i_subconnector_name(int val);
const char *drm_get_dvi_i_select_name(int val);
const char *drm_get_tv_subconnector_name(int val);
const char *drm_get_tv_select_name(int val);
int drm_mode_create_dvi_i_properties(struct drm_device *dev);
int drm_mode_create_tv_properties(struct drm_device *dev,
unsigned int num_modes,
const char * const modes[]);
int drm_mode_create_scaling_mode_property(struct drm_device *dev);
int drm_mode_create_aspect_ratio_property(struct drm_device *dev);
int drm_mode_create_suggested_offset_properties(struct drm_device *dev);
int drm_mode_connector_set_path_property(struct drm_connector *connector,
const char *path);
int drm_mode_connector_set_tile_property(struct drm_connector *connector);
int drm_mode_connector_update_edid_property(struct drm_connector *connector,
const struct edid *edid);
drm: Add a new connector atomic property for link status At the time userspace does setcrtc, we've already promised the mode would work. The promise is based on the theoretical capabilities of the link, but it's possible we can't reach this in practice. The DP spec describes how the link should be reduced, but we can't reduce the link below the requirements of the mode. Black screen follows. One idea would be to have setcrtc return a failure. However, it already should not fail as the atomic checks have passed. It would also conflict with the idea of making setcrtc asynchronous in the future, returning before the actual mode setting and link training. Another idea is to train the link "upfront" at hotplug time, before pruning the mode list, so that we can do the pruning based on practical not theoretical capabilities. However, the changes for link training are pretty drastic, all for the sake of error handling and DP compliance, when the most common happy day scenario is the current approach of link training at mode setting time, using the optimal parameters for the mode. It is also not certain all hardware could do this without the pipe on; not even all our hardware can do this. Some of this can be solved, but not trivially. Both of the above ideas also fail to address link degradation *during* operation. The solution is to add a new "link-status" connector property in order to address link training failure in a way that: a) changes the current happy day scenario as little as possible, to avoid regressions, b) can be implemented the same way by all drm drivers, c) is still opt-in for the drivers and userspace, and opting out doesn't regress the user experience, d) doesn't prevent drivers from implementing better or alternate approaches, possibly without userspace involvement. And, of course, handles all the issues presented. In the usual happy day scenario, this is always "good". If something fails during or after a mode set, the kernel driver can set the link status to "bad" and issue a hotplug uevent for userspace to have it re-check the valid modes through GET_CONNECTOR IOCTL, and try modeset again. If the theoretical capabilities of the link can't be reached, the mode list is trimmed based on that. v7 by Jani: * Rebase, simplify set property while at it, checkpatch fix v6: * Fix a typo in kernel doc (Sean Paul) v5: * Clarify doc for silent rejection of atomic properties by driver (Daniel Vetter) v4: * Add comments in kernel-doc format (Daniel Vetter) * Update the kernel-doc for link-status (Sean Paul) v3: * Fixed a build error (Jani Saarinen) v2: * Removed connector->link_status (Daniel Vetter) * Set connector->state->link_status in drm_mode_connector_set_link_status_property (Daniel Vetter) * Set the connector_changed flag to true if connector->state->link_status changed. * Reset link_status to GOOD in update_output_state (Daniel Vetter) * Never allow userspace to set link status from Good To Bad (Daniel Vetter) Reviewed-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Jani Nikula <jani.nikula@intel.com> Acked-by: Tony Cheng <tony.cheng@amd.com> Acked-by: Harry Wentland <harry.wentland@amd.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@intel.com> Cc: Ville Syrjala <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Sean Paul <seanpaul@chromium.org> Signed-off-by: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Acked-by: Eric Anholt <eric@anholt.net> (for the -modesetting patch) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/0182487051aa9f1594820e35a4853de2f8747b4e.1481883920.git.jani.nikula@intel.com
2016-12-16 03:29:06 -07:00
void drm_mode_connector_set_link_status_property(struct drm_connector *connector,
uint64_t link_status);
/**
* struct drm_tile_group - Tile group metadata
* @refcount: reference count
* @dev: DRM device
* @id: tile group id exposed to userspace
* @group_data: Sink-private data identifying this group
*
* @group_data corresponds to displayid vend/prod/serial for external screens
* with an EDID.
*/
struct drm_tile_group {
struct kref refcount;
struct drm_device *dev;
int id;
u8 group_data[8];
};
struct drm_tile_group *drm_mode_create_tile_group(struct drm_device *dev,
char topology[8]);
struct drm_tile_group *drm_mode_get_tile_group(struct drm_device *dev,
char topology[8]);
void drm_mode_put_tile_group(struct drm_device *dev,
struct drm_tile_group *tg);
/**
* drm_for_each_connector - iterate over all connectors
* @connector: the loop cursor
* @dev: the DRM device
*
* Iterate over all connectors of @dev.
drm: locking&new iterators for connector_list The requirements for connector_list locking are a bit tricky: - We need to be able to jump over zombie conectors (i.e. with refcount == 0, but not yet removed from the list). If instead we require that there's no zombies on the list then the final kref_put must happen under the list protection lock, which means that locking context leaks all over the place. Not pretty - better to deal with zombies and wrap the locking just around the list_del in the destructor. - When we walk the list we must _not_ hold the connector list lock. We walk the connector list at an absolutely massive amounts of places, if all those places can't ever call drm_connector_unreference the code would get unecessarily complicated. - connector_list needs it own lock, again too many places that walk it that we could reuse e.g. mode_config.mutex without resulting in inversions. - Lots of code uses these loops to look-up a connector, i.e. they want to be able to call drm_connector_reference. But on the other hand we want connectors to stay on that list until they're dead (i.e. connector_list can't hold a full reference), which means despite the "can't hold lock for the loop body" rule we need to make sure a connector doesn't suddenly become a zombie. At first Dave&I discussed various horror-show approaches using srcu, but turns out it's fairly easy: - For the loop body we always hold an additional reference to the current connector. That means it can't zombify, and it also means it'll stay on the list, which means we can use it as our iterator to find the next connector. - When we try to find the next connector we only have to jump over zombies. To make sure we don't chase bad pointers that entire loop is protected with the new connect_list_lock spinlock. And because we know that we're starting out with a non-zombie (need to drop our reference for the old connector only after we have our new one), we're guranteed to still be on the connector_list and either find the next non-zombie or complete the iteration. - Only downside is that we need to make sure that the temporary reference for the loop body doesn't leak. iter_get/put() functions + lockdep make sure that's the case. - To avoid a flag day the new iterator macro has an _iter postfix. We can rename it back once all the users of the unsafe version are gone (there's about 100 list walkers for the connector_list). For now this patch only converts all the list walking in the core, leaving helpers and drivers for later patches. The nice thing is that we can now finally remove 2 FIXME comments from the register/unregister functions. v2: - use irqsafe spinlocks, so that we can use this in drm_state_dump too. - nuke drm_modeset_lock_all from drm_connector_init, now entirely cargo-culted nonsense. v3: - do {} while (!kref_get_unless_zero), makes for a tidier loop (Dave). - pretty kerneldoc - add EXPORT_SYMBOL, helpers&drivers are supposed to use this. v4: Change lockdep annotations to only check whether we release the iter fake lock again (i.e. make sure that iter_put is called), but not check any locking dependecies itself. That seams to require a recursive read lock in trylock mode. Cc: Dave Airlie <airlied@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161213230814.19598-6-daniel.vetter@ffwll.ch
2016-12-13 16:08:06 -07:00
*
* WARNING:
*
* This iterator is not safe against hotadd/removal of connectors and is
* deprecated. Use drm_for_each_connector_iter() instead.
*/
#define drm_for_each_connector(connector, dev) \
list_for_each_entry(connector, &(dev)->mode_config.connector_list, head)
drm: locking&new iterators for connector_list The requirements for connector_list locking are a bit tricky: - We need to be able to jump over zombie conectors (i.e. with refcount == 0, but not yet removed from the list). If instead we require that there's no zombies on the list then the final kref_put must happen under the list protection lock, which means that locking context leaks all over the place. Not pretty - better to deal with zombies and wrap the locking just around the list_del in the destructor. - When we walk the list we must _not_ hold the connector list lock. We walk the connector list at an absolutely massive amounts of places, if all those places can't ever call drm_connector_unreference the code would get unecessarily complicated. - connector_list needs it own lock, again too many places that walk it that we could reuse e.g. mode_config.mutex without resulting in inversions. - Lots of code uses these loops to look-up a connector, i.e. they want to be able to call drm_connector_reference. But on the other hand we want connectors to stay on that list until they're dead (i.e. connector_list can't hold a full reference), which means despite the "can't hold lock for the loop body" rule we need to make sure a connector doesn't suddenly become a zombie. At first Dave&I discussed various horror-show approaches using srcu, but turns out it's fairly easy: - For the loop body we always hold an additional reference to the current connector. That means it can't zombify, and it also means it'll stay on the list, which means we can use it as our iterator to find the next connector. - When we try to find the next connector we only have to jump over zombies. To make sure we don't chase bad pointers that entire loop is protected with the new connect_list_lock spinlock. And because we know that we're starting out with a non-zombie (need to drop our reference for the old connector only after we have our new one), we're guranteed to still be on the connector_list and either find the next non-zombie or complete the iteration. - Only downside is that we need to make sure that the temporary reference for the loop body doesn't leak. iter_get/put() functions + lockdep make sure that's the case. - To avoid a flag day the new iterator macro has an _iter postfix. We can rename it back once all the users of the unsafe version are gone (there's about 100 list walkers for the connector_list). For now this patch only converts all the list walking in the core, leaving helpers and drivers for later patches. The nice thing is that we can now finally remove 2 FIXME comments from the register/unregister functions. v2: - use irqsafe spinlocks, so that we can use this in drm_state_dump too. - nuke drm_modeset_lock_all from drm_connector_init, now entirely cargo-culted nonsense. v3: - do {} while (!kref_get_unless_zero), makes for a tidier loop (Dave). - pretty kerneldoc - add EXPORT_SYMBOL, helpers&drivers are supposed to use this. v4: Change lockdep annotations to only check whether we release the iter fake lock again (i.e. make sure that iter_put is called), but not check any locking dependecies itself. That seams to require a recursive read lock in trylock mode. Cc: Dave Airlie <airlied@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161213230814.19598-6-daniel.vetter@ffwll.ch
2016-12-13 16:08:06 -07:00
/**
* struct drm_connector_list_iter - connector_list iterator
*
* This iterator tracks state needed to be able to walk the connector_list
* within struct drm_mode_config. Only use together with
* drm_connector_list_iter_begin(), drm_connector_list_iter_end() and
drm: locking&new iterators for connector_list The requirements for connector_list locking are a bit tricky: - We need to be able to jump over zombie conectors (i.e. with refcount == 0, but not yet removed from the list). If instead we require that there's no zombies on the list then the final kref_put must happen under the list protection lock, which means that locking context leaks all over the place. Not pretty - better to deal with zombies and wrap the locking just around the list_del in the destructor. - When we walk the list we must _not_ hold the connector list lock. We walk the connector list at an absolutely massive amounts of places, if all those places can't ever call drm_connector_unreference the code would get unecessarily complicated. - connector_list needs it own lock, again too many places that walk it that we could reuse e.g. mode_config.mutex without resulting in inversions. - Lots of code uses these loops to look-up a connector, i.e. they want to be able to call drm_connector_reference. But on the other hand we want connectors to stay on that list until they're dead (i.e. connector_list can't hold a full reference), which means despite the "can't hold lock for the loop body" rule we need to make sure a connector doesn't suddenly become a zombie. At first Dave&I discussed various horror-show approaches using srcu, but turns out it's fairly easy: - For the loop body we always hold an additional reference to the current connector. That means it can't zombify, and it also means it'll stay on the list, which means we can use it as our iterator to find the next connector. - When we try to find the next connector we only have to jump over zombies. To make sure we don't chase bad pointers that entire loop is protected with the new connect_list_lock spinlock. And because we know that we're starting out with a non-zombie (need to drop our reference for the old connector only after we have our new one), we're guranteed to still be on the connector_list and either find the next non-zombie or complete the iteration. - Only downside is that we need to make sure that the temporary reference for the loop body doesn't leak. iter_get/put() functions + lockdep make sure that's the case. - To avoid a flag day the new iterator macro has an _iter postfix. We can rename it back once all the users of the unsafe version are gone (there's about 100 list walkers for the connector_list). For now this patch only converts all the list walking in the core, leaving helpers and drivers for later patches. The nice thing is that we can now finally remove 2 FIXME comments from the register/unregister functions. v2: - use irqsafe spinlocks, so that we can use this in drm_state_dump too. - nuke drm_modeset_lock_all from drm_connector_init, now entirely cargo-culted nonsense. v3: - do {} while (!kref_get_unless_zero), makes for a tidier loop (Dave). - pretty kerneldoc - add EXPORT_SYMBOL, helpers&drivers are supposed to use this. v4: Change lockdep annotations to only check whether we release the iter fake lock again (i.e. make sure that iter_put is called), but not check any locking dependecies itself. That seams to require a recursive read lock in trylock mode. Cc: Dave Airlie <airlied@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161213230814.19598-6-daniel.vetter@ffwll.ch
2016-12-13 16:08:06 -07:00
* drm_connector_list_iter_next() respectively the convenience macro
* drm_for_each_connector_iter().
*/
struct drm_connector_list_iter {
/* private: */
struct drm_device *dev;
struct drm_connector *conn;
};
void drm_connector_list_iter_begin(struct drm_device *dev,
struct drm_connector_list_iter *iter);
drm: locking&new iterators for connector_list The requirements for connector_list locking are a bit tricky: - We need to be able to jump over zombie conectors (i.e. with refcount == 0, but not yet removed from the list). If instead we require that there's no zombies on the list then the final kref_put must happen under the list protection lock, which means that locking context leaks all over the place. Not pretty - better to deal with zombies and wrap the locking just around the list_del in the destructor. - When we walk the list we must _not_ hold the connector list lock. We walk the connector list at an absolutely massive amounts of places, if all those places can't ever call drm_connector_unreference the code would get unecessarily complicated. - connector_list needs it own lock, again too many places that walk it that we could reuse e.g. mode_config.mutex without resulting in inversions. - Lots of code uses these loops to look-up a connector, i.e. they want to be able to call drm_connector_reference. But on the other hand we want connectors to stay on that list until they're dead (i.e. connector_list can't hold a full reference), which means despite the "can't hold lock for the loop body" rule we need to make sure a connector doesn't suddenly become a zombie. At first Dave&I discussed various horror-show approaches using srcu, but turns out it's fairly easy: - For the loop body we always hold an additional reference to the current connector. That means it can't zombify, and it also means it'll stay on the list, which means we can use it as our iterator to find the next connector. - When we try to find the next connector we only have to jump over zombies. To make sure we don't chase bad pointers that entire loop is protected with the new connect_list_lock spinlock. And because we know that we're starting out with a non-zombie (need to drop our reference for the old connector only after we have our new one), we're guranteed to still be on the connector_list and either find the next non-zombie or complete the iteration. - Only downside is that we need to make sure that the temporary reference for the loop body doesn't leak. iter_get/put() functions + lockdep make sure that's the case. - To avoid a flag day the new iterator macro has an _iter postfix. We can rename it back once all the users of the unsafe version are gone (there's about 100 list walkers for the connector_list). For now this patch only converts all the list walking in the core, leaving helpers and drivers for later patches. The nice thing is that we can now finally remove 2 FIXME comments from the register/unregister functions. v2: - use irqsafe spinlocks, so that we can use this in drm_state_dump too. - nuke drm_modeset_lock_all from drm_connector_init, now entirely cargo-culted nonsense. v3: - do {} while (!kref_get_unless_zero), makes for a tidier loop (Dave). - pretty kerneldoc - add EXPORT_SYMBOL, helpers&drivers are supposed to use this. v4: Change lockdep annotations to only check whether we release the iter fake lock again (i.e. make sure that iter_put is called), but not check any locking dependecies itself. That seams to require a recursive read lock in trylock mode. Cc: Dave Airlie <airlied@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161213230814.19598-6-daniel.vetter@ffwll.ch
2016-12-13 16:08:06 -07:00
struct drm_connector *
drm_connector_list_iter_next(struct drm_connector_list_iter *iter);
void drm_connector_list_iter_end(struct drm_connector_list_iter *iter);
drm: locking&new iterators for connector_list The requirements for connector_list locking are a bit tricky: - We need to be able to jump over zombie conectors (i.e. with refcount == 0, but not yet removed from the list). If instead we require that there's no zombies on the list then the final kref_put must happen under the list protection lock, which means that locking context leaks all over the place. Not pretty - better to deal with zombies and wrap the locking just around the list_del in the destructor. - When we walk the list we must _not_ hold the connector list lock. We walk the connector list at an absolutely massive amounts of places, if all those places can't ever call drm_connector_unreference the code would get unecessarily complicated. - connector_list needs it own lock, again too many places that walk it that we could reuse e.g. mode_config.mutex without resulting in inversions. - Lots of code uses these loops to look-up a connector, i.e. they want to be able to call drm_connector_reference. But on the other hand we want connectors to stay on that list until they're dead (i.e. connector_list can't hold a full reference), which means despite the "can't hold lock for the loop body" rule we need to make sure a connector doesn't suddenly become a zombie. At first Dave&I discussed various horror-show approaches using srcu, but turns out it's fairly easy: - For the loop body we always hold an additional reference to the current connector. That means it can't zombify, and it also means it'll stay on the list, which means we can use it as our iterator to find the next connector. - When we try to find the next connector we only have to jump over zombies. To make sure we don't chase bad pointers that entire loop is protected with the new connect_list_lock spinlock. And because we know that we're starting out with a non-zombie (need to drop our reference for the old connector only after we have our new one), we're guranteed to still be on the connector_list and either find the next non-zombie or complete the iteration. - Only downside is that we need to make sure that the temporary reference for the loop body doesn't leak. iter_get/put() functions + lockdep make sure that's the case. - To avoid a flag day the new iterator macro has an _iter postfix. We can rename it back once all the users of the unsafe version are gone (there's about 100 list walkers for the connector_list). For now this patch only converts all the list walking in the core, leaving helpers and drivers for later patches. The nice thing is that we can now finally remove 2 FIXME comments from the register/unregister functions. v2: - use irqsafe spinlocks, so that we can use this in drm_state_dump too. - nuke drm_modeset_lock_all from drm_connector_init, now entirely cargo-culted nonsense. v3: - do {} while (!kref_get_unless_zero), makes for a tidier loop (Dave). - pretty kerneldoc - add EXPORT_SYMBOL, helpers&drivers are supposed to use this. v4: Change lockdep annotations to only check whether we release the iter fake lock again (i.e. make sure that iter_put is called), but not check any locking dependecies itself. That seams to require a recursive read lock in trylock mode. Cc: Dave Airlie <airlied@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161213230814.19598-6-daniel.vetter@ffwll.ch
2016-12-13 16:08:06 -07:00
/**
* drm_for_each_connector_iter - connector_list iterator macro
* @connector: &struct drm_connector pointer used as cursor
* @iter: &struct drm_connector_list_iter
drm: locking&new iterators for connector_list The requirements for connector_list locking are a bit tricky: - We need to be able to jump over zombie conectors (i.e. with refcount == 0, but not yet removed from the list). If instead we require that there's no zombies on the list then the final kref_put must happen under the list protection lock, which means that locking context leaks all over the place. Not pretty - better to deal with zombies and wrap the locking just around the list_del in the destructor. - When we walk the list we must _not_ hold the connector list lock. We walk the connector list at an absolutely massive amounts of places, if all those places can't ever call drm_connector_unreference the code would get unecessarily complicated. - connector_list needs it own lock, again too many places that walk it that we could reuse e.g. mode_config.mutex without resulting in inversions. - Lots of code uses these loops to look-up a connector, i.e. they want to be able to call drm_connector_reference. But on the other hand we want connectors to stay on that list until they're dead (i.e. connector_list can't hold a full reference), which means despite the "can't hold lock for the loop body" rule we need to make sure a connector doesn't suddenly become a zombie. At first Dave&I discussed various horror-show approaches using srcu, but turns out it's fairly easy: - For the loop body we always hold an additional reference to the current connector. That means it can't zombify, and it also means it'll stay on the list, which means we can use it as our iterator to find the next connector. - When we try to find the next connector we only have to jump over zombies. To make sure we don't chase bad pointers that entire loop is protected with the new connect_list_lock spinlock. And because we know that we're starting out with a non-zombie (need to drop our reference for the old connector only after we have our new one), we're guranteed to still be on the connector_list and either find the next non-zombie or complete the iteration. - Only downside is that we need to make sure that the temporary reference for the loop body doesn't leak. iter_get/put() functions + lockdep make sure that's the case. - To avoid a flag day the new iterator macro has an _iter postfix. We can rename it back once all the users of the unsafe version are gone (there's about 100 list walkers for the connector_list). For now this patch only converts all the list walking in the core, leaving helpers and drivers for later patches. The nice thing is that we can now finally remove 2 FIXME comments from the register/unregister functions. v2: - use irqsafe spinlocks, so that we can use this in drm_state_dump too. - nuke drm_modeset_lock_all from drm_connector_init, now entirely cargo-culted nonsense. v3: - do {} while (!kref_get_unless_zero), makes for a tidier loop (Dave). - pretty kerneldoc - add EXPORT_SYMBOL, helpers&drivers are supposed to use this. v4: Change lockdep annotations to only check whether we release the iter fake lock again (i.e. make sure that iter_put is called), but not check any locking dependecies itself. That seams to require a recursive read lock in trylock mode. Cc: Dave Airlie <airlied@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161213230814.19598-6-daniel.vetter@ffwll.ch
2016-12-13 16:08:06 -07:00
*
* Note that @connector is only valid within the list body, if you want to use
* @connector after calling drm_connector_list_iter_end() then you need to grab
* your own reference first using drm_connector_begin().
drm: locking&new iterators for connector_list The requirements for connector_list locking are a bit tricky: - We need to be able to jump over zombie conectors (i.e. with refcount == 0, but not yet removed from the list). If instead we require that there's no zombies on the list then the final kref_put must happen under the list protection lock, which means that locking context leaks all over the place. Not pretty - better to deal with zombies and wrap the locking just around the list_del in the destructor. - When we walk the list we must _not_ hold the connector list lock. We walk the connector list at an absolutely massive amounts of places, if all those places can't ever call drm_connector_unreference the code would get unecessarily complicated. - connector_list needs it own lock, again too many places that walk it that we could reuse e.g. mode_config.mutex without resulting in inversions. - Lots of code uses these loops to look-up a connector, i.e. they want to be able to call drm_connector_reference. But on the other hand we want connectors to stay on that list until they're dead (i.e. connector_list can't hold a full reference), which means despite the "can't hold lock for the loop body" rule we need to make sure a connector doesn't suddenly become a zombie. At first Dave&I discussed various horror-show approaches using srcu, but turns out it's fairly easy: - For the loop body we always hold an additional reference to the current connector. That means it can't zombify, and it also means it'll stay on the list, which means we can use it as our iterator to find the next connector. - When we try to find the next connector we only have to jump over zombies. To make sure we don't chase bad pointers that entire loop is protected with the new connect_list_lock spinlock. And because we know that we're starting out with a non-zombie (need to drop our reference for the old connector only after we have our new one), we're guranteed to still be on the connector_list and either find the next non-zombie or complete the iteration. - Only downside is that we need to make sure that the temporary reference for the loop body doesn't leak. iter_get/put() functions + lockdep make sure that's the case. - To avoid a flag day the new iterator macro has an _iter postfix. We can rename it back once all the users of the unsafe version are gone (there's about 100 list walkers for the connector_list). For now this patch only converts all the list walking in the core, leaving helpers and drivers for later patches. The nice thing is that we can now finally remove 2 FIXME comments from the register/unregister functions. v2: - use irqsafe spinlocks, so that we can use this in drm_state_dump too. - nuke drm_modeset_lock_all from drm_connector_init, now entirely cargo-culted nonsense. v3: - do {} while (!kref_get_unless_zero), makes for a tidier loop (Dave). - pretty kerneldoc - add EXPORT_SYMBOL, helpers&drivers are supposed to use this. v4: Change lockdep annotations to only check whether we release the iter fake lock again (i.e. make sure that iter_put is called), but not check any locking dependecies itself. That seams to require a recursive read lock in trylock mode. Cc: Dave Airlie <airlied@gmail.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161213230814.19598-6-daniel.vetter@ffwll.ch
2016-12-13 16:08:06 -07:00
*/
#define drm_for_each_connector_iter(connector, iter) \
while ((connector = drm_connector_list_iter_next(iter)))
#endif