Vignesh Raman [Fri, 22 Dec 2023 03:34:34 +0000 (09:04 +0530)]
drm/ci: uprev mesa version: fix kdl commit fetch
build-kdl.sh was doing a `clone --depth 1` of the default branch,
then checking out a commit that might not be the latest of that
branch, resulting in container build error.
https://gitlab.freedesktop.org/mesa/mesa/-/commit/5efa4d56 fixes
kdl commit fetch issue. Uprev mesa in drm-ci to fix this.
This commit updates the kernel tag and adds .never-post-merge-rules
due to the mesa uprev. It also fixes an issue where the virtio-gpu
pipeline was not getting created with the mesa uprev.
Randy Dunlap [Fri, 12 Jan 2024 05:17:31 +0000 (21:17 -0800)]
drm/doc: internals: remove section on PCI legacy support
The functions that were described in this section of
drm-internals.rst were removed in NOV-2023, along with all of the
kernel-doc comments in the source file. This now causes a
docs build warning, so remove that section of the documentation also.
drivers/gpu/drm/drm_pci.c:1: warning: no structured comments found
This happens because, although `prepare_fb` and `cleanup_fb` are
perfectly balanced, we cannot guarantee consistency in the check
plane->state->fb == state->fb. This means that sometimes we can increase
the refcount in `prepare_fb` and don't decrease it in `cleanup_fb`. The
opposite can also be true.
In fact, the struct drm_plane .state shouldn't be accessed directly
but instead, the `drm_atomic_get_new_plane_state()` helper function should
be used. So, we could stick to this check, but using
`drm_atomic_get_new_plane_state()`. But actually, this check is not really
needed. We can increase and decrease the refcount symmetrically without
problems.
This is going to make the code more simple and consistent.
Jani Nikula [Thu, 4 Jan 2024 20:16:30 +0000 (22:16 +0200)]
drm/nouveau: include drm/drm_edid.h only where needed
Including drm_edid.h from nouveau_connector.h causes the rebuild of 15
files when drm_edid.h is modified, while there are only a few files that
actually need to include drm_edid.h.
xiazhengqiao [Thu, 21 Dec 2023 09:30:57 +0000 (17:30 +0800)]
drm/bridge: Fixed a DP link training bug
To have better compatibility for DP sink, there is a retry mechanism
for the link training process to switch between different training process.
The original driver code doesn't reset the retry counter when training
state is pass. If the system triggers link training over 3 times,
there will be a chance to causes the driver to use the wrong training
method and return a training fail result.
To Fix this, we reset the retry counter when training state is pass
each time.
Dario Binacchi [Mon, 8 Jan 2024 20:15:53 +0000 (21:15 +0100)]
drm/panel: nt35510: support FRIDA FRD400B25025-A-CTK
The initialization commands are taken from the STMicroelectronics driver
found at [1].
To ensure backward compatibility, flags have been added to enable gamma
correction setting and display control. In other cases, registers have
been set to their default values according to the specifications found
in the datasheet.
Dario Binacchi [Mon, 8 Jan 2024 20:15:52 +0000 (21:15 +0100)]
drm/panel: nt35510: move hardwired parameters to configuration
This patch, preparatory for future developments, move the hardwired
parameters to configuration data to allow the addition of new
NT35510-based panels.
Hsin-Yi Wang [Wed, 20 Dec 2023 22:13:11 +0000 (14:13 -0800)]
drm/panel-edp: use put_sync in unprepare
Some edp panel requires T10 (Delay from end of valid video data transmitted
by the Source device to power-off) less than 500ms. Using autosuspend with
delay set as 1000 violates this requirement.
Use put_sync_suspend in unprepare to meet the spec. For other cases (such
as getting EDID), it still uses autosuspend.
Vegard Nossum [Mon, 25 Dec 2023 06:51:45 +0000 (07:51 +0100)]
drm/nouveau: uapi: fix kerneldoc warnings
As of commit b77fdd6a48e6 ("scripts/kernel-doc: restore warning for
Excess struct/union"), we see the following warnings when running 'make
htmldocs':
./include/uapi/drm/nouveau_drm.h:292: warning: Excess struct member 'DRM_NOUVEAU_VM_BIND_OP_MAP' description in 'drm_nouveau_vm_bind_op'
./include/uapi/drm/nouveau_drm.h:292: warning: Excess struct member 'DRM_NOUVEAU_VM_BIND_OP_UNMAP' description in 'drm_nouveau_vm_bind_op'
./include/uapi/drm/nouveau_drm.h:292: warning: Excess struct member 'DRM_NOUVEAU_VM_BIND_SPARSE' description in 'drm_nouveau_vm_bind_op'
./include/uapi/drm/nouveau_drm.h:336: warning: Excess struct member 'DRM_NOUVEAU_VM_BIND_RUN_ASYNC' description in 'drm_nouveau_vm_bind'
The problem is that these values are #define constants, but had kerneldoc
comments attached to them as if they were actual struct members.
There are a number of ways we could fix this, but I chose to draw
inspiration from include/uapi/drm/i915_drm.h, which pulls them into the
corresponding kerneldoc comment for the struct member that they are
intended to be used with.
To keep the diff readable, there are a number of things I _didn't_ do in
this patch, but which we should also consider:
- This is pretty good documentation, but it ends up in gpu/driver-uapi,
which is part of subsystem-apis/ when it really ought to display under
userspace-api/ (the "Linux kernel user-space API guide" book of the
documentation).
- More generally, we might want a warning if include/uapi/ files are
kerneldoc'd outside userspace-api/.
- I'd consider it cleaner if the #defines appeared between the kerneldoc
for the member and the member itself (which is something other DRM-
related UAPI docs do).
- The %IDENTIFIER kerneldoc syntax is intended for "constants", and is
more appropriate in this context than ``IDENTIFIER`` or &IDENTIFIER.
The DRM docs aren't very consistent on this.
Change kernel-doc "/**" comments to common "/*" comments to prevent
kernel-doc warnings:
gf100.c:1044: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
* Wait until GR goes idle. GR is considered idle if it is disabled by the
gf100.c:1044: warning: missing initial short description on line:
* Wait until GR goes idle. GR is considered idle if it is disabled by the
Randy Dunlap [Sun, 31 Dec 2023 23:36:31 +0000 (15:36 -0800)]
drm/nouveau: don't misuse kernel-doc comments
Change kernel-doc "/**" comments to common "/*" comments to prevent
kernel-doc warnings:
nouveau_ioc32.c:2: warning: Cannot understand * \file mga_ioc32.c
on line 2 - I thought it was a doc line
nouveau_ioc32.c:52: warning: Function parameter or member 'filp' not described in 'nouveau_compat_ioctl'
nouveau_ioc32.c:52: warning: Function parameter or member 'cmd' not described in 'nouveau_compat_ioctl'
nouveau_ioc32.c:52: warning: Function parameter or member 'arg' not described in 'nouveau_compat_ioctl'
nouveau_ioc32.c:52: warning: expecting prototype for Called whenever a 32-bit process running under a 64(). Prototype was for nouveau_compat_ioctl() instead
Change kernel-doc "/**" comments to common "/*" comments to prevent
kernel-doc warnings:
crtc.c:453: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
* Sets up registers for the given mode/adjusted_mode pair.
crtc.c:453: warning: missing initial short description on line:
* Sets up registers for the given mode/adjusted_mode pair.
crtc.c:629: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
* Sets up registers for the given mode/adjusted_mode pair.
crtc.c:629: warning: missing initial short description on line:
* Sets up registers for the given mode/adjusted_mode pair.
Randy Dunlap [Sat, 16 Dec 2023 20:11:52 +0000 (12:11 -0800)]
drm/nouveau/bios/init: drop kernel-doc notation
The "/**" comments in this file are not kernel-doc comments. They are
used on static functions which can have kernel-doc comments, but that
is not the primary focus of kernel-doc comments.
Since these comments are incomplete for kernel-doc notation, remove
the kernel-doc "/**" markers and make them common comments.
This prevents scripts/kernel-doc from issuing 68 warnings:
init.c:584: warning: Function parameter or member 'init' not described in 'init_reserved'
and 67 warnings like this one:
init.c:611: warning: expecting prototype for INIT_DONE(). Prototype was for init_done() instead
Remove the unused drm_connector_helper_get_modes_from_ddc()
function. Most drivers should probably have this functionality split to
detect and get modes parts, so the helper is not the best abstraction.
drm/mgag200: Fix caching setup for remapped video memory
I/O video memory for the framebuffer supports write-combine caching
mode. Simplify the driver's code that sets up the caching mode.
* Map video memory with ioremap_wc(), which automatically sets up
the PAT entry with write-combine caching.
* Remove the now obsolete call to devm_arch_io_reserve_memtype_wc().
It is only required to mmap the video memory to user space, which the
driver doesn't do.
* According to the PAT documentation, arch_phys_wc_add() is best
called after remapping I/O memory, so move it after ioremap.
Markus Elfring [Tue, 26 Dec 2023 15:30:25 +0000 (16:30 +0100)]
drm/sched: One function call less in drm_sched_init() after error detection
The kfree() function was called in one case by the
drm_sched_init() function during error handling
even if the passed data structure member contained a null pointer.
This issue was detected by using the Coccinelle software.
drm: Move drm_set_preferred_mode() helper from drm_edid to drm_modes
The helper is generic, it doesn't use the opaque EDID type struct drm_edid
and is also used by drivers that only support non-probeable displays such
as fixed panels.
These drivers add a list of modes using drm_mode_probed_add() and then set
a preferred mode using the drm_set_preferred_mode() helper.
It seems more logical to have the helper definition in drm_modes.o instead
of drm_edid.o, since the former contains modes helper while the latter has
helpers to manage the EDID information.
Since both drm_edid.o and drm_modes.o object files are built-in the drm.o
object, there are no functional changes. But besides being a more logical
place for this helper, it could also allow to eventually make drm_edid.o
optional and not included in drm.o if only fixed panels must be supported
in a given system.
Quentin Schulz [Thu, 4 Jan 2024 12:41:57 +0000 (13:41 +0100)]
drm/panel: ltk050h3146w: only print message when GPIO getting is not EPROBE_DEFER
devm_gpiod_get_optional may return EPROBE_DEFER in case the GPIO
controller isn't yet probed when the panel driver is being probed.
In that case, a spurious and confusing error message about not being
able to get the reset GPIO is printed even though later on the device
actually manages to get probed.
Use dev_err_probe instead so that the message is only printed when it
truly matters.
Quentin Schulz [Mon, 20 Nov 2023 12:29:49 +0000 (13:29 +0100)]
drm/rockchip: lvds: do not print scary message when probing defer
This scary message can misled the user into thinking something bad has
happened and needs to be fixed, however it could simply be part of a
normal boot process where EPROBE_DEFER is taken into account. Therefore,
let's use dev_err_probe so that this message doesn't get shown (by
default) when the return code is EPROBE_DEFER.
Quentin Schulz [Mon, 20 Nov 2023 12:29:48 +0000 (13:29 +0100)]
drm/rockchip: lvds: do not overwrite error code
ret variable stores the return value of drm_of_find_panel_or_bridge
which can return error codes different from EPROBE_DEFER. Therefore,
let's just return that error code instead of forcing it to EPROBE_DEFER.
Alex Bee [Fri, 22 Dec 2023 17:42:17 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Drop custom fill_modes hook
Now that we have proper pixelclock-based mode validation we can drop the
custom fill_modes hook.
CRTC size validation for the display controller has been added with
Commit 8e140cb60270 ("drm/rockchip: vop: limit maximum resolution to
hardware capabilities")
As per TRM this controller supports pixelclocks starting from 25 MHz. The
maximum supported pixelclocks are defined by the phy configurations we
have. Also it can't support modes that require doubled clocks. If the
variant has a phy reference clock we can additionally validate against VESA
DMT'srecommendations.
Alex Bee [Fri, 22 Dec 2023 17:42:15 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Add RK3128 support
This variant requires the phy reference clock to be enabled before the DDC
block can work and the (initial) DDC bus frequency is calculated based on
the rate of this clock. Besides the only difference is phy configuration
required to make the driver working for this variant as well.
Alex Bee [Fri, 22 Dec 2023 17:42:14 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Add variant support
In preparation to support RK3128's integration of the controller, this
patch adds a simple variant implementation. They mainly differ in the phy
configuration required, so those are part of the match_data. The values
have been taken from downstream. The pixelclocks in there are meant to be
max-inclusive.
Zack Rusin [Sun, 24 Dec 2023 05:25:40 +0000 (00:25 -0500)]
drm/vmwgfx: Unmap the surface before resetting it on a plane state
Switch to a new plane state requires unreferencing of all held surfaces.
In the work required for mob cursors the mapped surfaces started being
cached but the variable indicating whether the surface is currently
mapped was not being reset. This leads to crashes as the duplicated
state, incorrectly, indicates the that surface is mapped even when
no surface is present. That's because after unreferencing the surface
it's perfectly possible for the plane to be backed by a bo instead of a
surface.
Reset the surface mapped flag when unreferencing the plane state surface
to fix null derefs in cleanup. Fixes crashes in KDE KWin 6.0 on Wayland:
Randy Dunlap [Fri, 15 Dec 2023 23:56:38 +0000 (15:56 -0800)]
drm/vmwgfx: fix kernel-doc Excess struct member 'base'
Fix a new kernel-doc warning reported by kernel test robot:
vmwgfx_surface.c:55: warning: Excess struct member 'base' description in 'vmw_user_surface'
The other warning is not correct: it is confused by "__counted_by".
Kees has made a separate patch for that.
In -Wall mode, kernel-doc still reports 20 warnings of this nature:
vmwgfx_surface.c:198: warning: No description found for return value of 'vmw_surface_dma_size'
but I am not addressing those.
Randy Dunlap [Fri, 15 Dec 2023 23:41:02 +0000 (15:41 -0800)]
drm/vmwgfx: fix all kernel-doc warnings in stdu
kernel test robot reports one kernel-doc warning in stdu, but
running scripts/kernel-doc in -Wall mode reports several more,
so fix all of them at one time:
vmwgfx_stdu.c:76: warning: Excess struct member 'transfer' description in 'vmw_stdu_dirty'
vmwgfx_stdu.c:103: warning: missing initial short description on line:
* struct vmw_screen_target_display_unit
vmwgfx_stdu.c:215: warning: No description found for return value of 'vmw_stdu_bind_st'
vmwgfx_stdu.c:320: warning: No description found for return value of 'vmw_stdu_destroy_st'
vmwgfx_stdu.c:551: warning: No description found for return value of 'vmw_kms_stdu_readback'
vmwgfx_stdu.c:719: warning: No description found for return value of 'vmw_kms_stdu_surface_dirty'
vmwgfx_stdu.c:895: warning: No description found for return value of 'vmw_stdu_primary_plane_prepare_fb'
vmwgfx_stdu.c:1470: warning: No description found for return value of 'vmw_stdu_init'
Andy Shevchenko [Tue, 19 Dec 2023 15:19:55 +0000 (17:19 +0200)]
drm/virtio: Spelling fixes
While making a spelling mistake myself for `git grep kvalloc`
I found that the only file has such a typo. Fix it and update
to the standard de facto of how we refer to the functions.
Also spell usr-out as user-out, it seems this driver uses its
own terminology nobody else can decypher, make it more readable.
Harry Wentland [Wed, 8 Nov 2023 16:36:24 +0000 (11:36 -0500)]
drm/vkms: Avoid reading beyond LUT array
When the floor LUT index (drm_fixp2int(lut_index) is the last
index of the array the ceil LUT index will point to an entry
beyond the array. Make sure we guard against it and use the
value of the floor LUT index.
v3:
- Drop bits from commit description that didn't contribute
anything of value
Harry Wentland [Wed, 8 Nov 2023 16:36:20 +0000 (11:36 -0500)]
drm: Don't treat 0 as -1 in drm_fixp2int_ceil
Unit testing this in VKMS shows that passing 0 into
this function returns -1, which is highly counter-
intuitive. Fix it by checking whether the input is
>= 0 instead of > 0.
dt-bindings: display: ssd132x: Add vendor prefix to width and height
Commit 2d23e7d6bacb ("dt-bindings: display: Add SSD132x OLED controllers")
used the wrong properties for width and height, instead of the correct
"solomon,width" and "solomon,height" properties.
Fix this by adding the vendor prefix to the width and height properties.
dt-bindings: display: ssd1307fb: Add vendor prefix to width and height
The commit 591825fba8a2 ("dt-bindings: display: ssd1307fb: Remove default
width and height values") used the wrong properties for width and height,
instead of the correct "solomon,width" and "solomon,height" properties.
Fix this by adding the vendor prefix to the width and height properties.
Alex Bee [Fri, 22 Dec 2023 17:42:12 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Don't power up the phy after resetting
inno_hdmi_reset is only ever called when initializing the controller. At
this point it’s completely uneccessary to power up the PHY, since all
what has to work at this point is the DDC bus. The phy will be powered up
correctly when a mode is set in inno_hdmi_encoder_enable and disabled in
inno_hdmi_encoder_disable.
Set it to LOWER_PWR after resetting the controller.
Alex Bee [Fri, 22 Dec 2023 17:42:11 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Correctly setup HDMI quantization range
The display controller will always give full range RGB regardless of the
mode set, but HDMI requires certain modes to be transmitted in limited
range RGB. This is especially required for HDMI sinks which do not support
non-standard quantization ranges.
This enables color space conversion for those modes and sets the
quantization range accordingly in the AVI infoframe.
Alex Bee [Fri, 22 Dec 2023 17:42:10 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Subclass connector state
The data which is currently hold in hdmi_data should not be part of device
itself but of the connector state.
Introduce a connector state subclass and move the data from hdmi_data in
there.
Maxime Ripard [Fri, 22 Dec 2023 17:42:05 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Move infoframe disable to separate function
The code to upload infoframes to the controller uses a weird construct
which, based on the previous function call return code, will either
disable or enable that infoframe.
In order to get rid of that argument, let's split the function to
disable the infoframe into a separate function and make it obvious what
we are doing in the error path.
Maxime Ripard [Fri, 22 Dec 2023 17:42:04 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Drop HDMI Vendor Infoframe support
The HDMI vendor infoframe is only meant to be sent with 4k60 modes and
higher, but the controller doesn't support them. Let's drop them from
the kernel.
Maxime Ripard [Fri, 22 Dec 2023 17:42:03 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Remove tmds rate from structure
The tmds_rate field in the inno_hdmi structure is used mostly to
configure the internal i2c controller divider through a call to the
inno_hdmi_i2c_init() function.
We can simply make that rate an argument to that function, which also
removes a workaround to initialize the divider at probe time when we
don't have a mode yet.
Maxime Ripard [Fri, 22 Dec 2023 17:42:01 +0000 (18:42 +0100)]
drm/rockchip: inno_hdmi: Remove useless input format
The driver has a lot of logic to deal with multiple input formats, but
hardcodes it to RGB. This means that most of that code has been dead
code, so let's get rid of it.
Alex Bee [Fri, 22 Dec 2023 17:41:54 +0000 (18:41 +0100)]
drm/rockchip: inno_hdmi: Fix video timing
The controller wants the difference between *total and *sync_start in the
HDMI_VIDEO_EXT_*DELAY registers. Otherwise the signal is very unstable for
certain non-VIC modes. See downstream commit [0].
Alex Bee [Fri, 22 Dec 2023 17:41:53 +0000 (18:41 +0100)]
drm/rockchip: vop: Add output selection registers for RK312x
In contrast to RK3036, RK312x SoCs have multiple output channels such as
RGB (i.e. LVDS TTL), LVDS, DSI and HDMI.
In order to support that, this splits output from RK3036 and defines an
separate one for RK3126 with the registers required to enable the
appropriate output and setup the correct polarity.
Philipp Stanner [Fri, 22 Dec 2023 11:52:17 +0000 (12:52 +0100)]
drm/tilcdc: request and mapp iomem with devres
tilcdc currently just ioremaps its iomem, without doing the (a bit more
robust) request on the memory first. The devm_ functions provide a handy
way to both request and ioremap the memory with automatic cleanup.
Replace the manual ioremap with the devm_ version.
Jeffrey Hugo [Fri, 8 Dec 2023 16:34:57 +0000 (09:34 -0700)]
accel/qaic: Order pci_remove() operations in reverse of probe()
In probe() we create the drm_device, and then register the MHI controller.
In remove(), we should unregister the controller first, then remove the
drm_device. Update the remove() operations to match.
accel/qaic: Leverage DRM managed APIs to release resources
Offload the balancing of init and destroy calls to DRM managed APIs.
mutex destroy for ->cntl_mutex is not called during device release and
destroy workqueue is not called in error path of create_qdev(). So, use
DRM managed APIs to manage the release of resources and avoid such
problems.
Pin-yen Lin [Thu, 14 Dec 2023 15:27:52 +0000 (23:27 +0800)]
drm/panel-edp: Add some panels with conservative timings
These panels are used by Mediatek MT8173 Chromebooks, and they used to
work with the downstream v4.19 kernel without any specified delay.
Back in the v4.19 kernel, they used the "little white lie" approach,
which is making the devicetree claim a specific panel's compatible
string for many different panels. That was a common solution before the
generic edp-panel driver.
After we uprevved the device to a newer kernel and used the edp-panel
driver, we saw multiple devices reporting warnings of using an unknown
panel and falling back to the conservative timings, which means that
they turn on/off much more slowly than they should. We tried to fill in
the timings for those panels, but we failed to find all the data sheets
for them.
Therefore, instead of having them use the default conservative timings,
update them with less-conservative timings from other panels of the same
vendor. The panels should still work under those timings, and we can
save some delays and suppress the warnings.
Ville Syrjälä [Fri, 15 Dec 2023 11:11:29 +0000 (13:11 +0200)]
drm/mm: Allow CONFIG_DRM_MM_DEBUG with DRM=m
The original rationale for
commit cd456f8d06d2 ("drm: Restrict stackdepot usage to builtin drm.ko")
was that depot_save_stack() (which is what we used back then)
wasn't exported. stack_depot_save() (which is what we use now) is
exported however, so relax the dependency allow CONFIG_DRM_MM_DEBUG
with DRM=m.
Maxime Ripard [Thu, 14 Dec 2023 10:09:15 +0000 (11:09 +0100)]
drm/atomic: Make the drm_atomic_state documentation less ambiguous
The current documentation of drm_atomic_state says that it's the "global
state object". This is confusing since, while it does contain all the
objects affected by an update and their respective states, if an object
isn't affected by this update it won't be part of it.
Thus, it's not truly a "global state", unlike object state structures
that do contain the entire state of a given object.
Maxime Ripard [Thu, 14 Dec 2023 10:09:14 +0000 (11:09 +0100)]
drm/atomic: Rework the object doc a bit
Commits 63e83c1dba54 ("drm: Consolidate connector arrays in
drm_atomic_state"), b8b5342b699b ("drm: Consolidate plane arrays in
drm_atomic_state") and 5d943aa6c0d4 ("drm: Consolidate crtc arrays in
drm_atomic_state") moved the object pointer and their state pointer to
an intermediate structure storing both.
The CRTC commit didn't update the doc of the crtcs field to reflect
that, and the doc for the planes and connectors fields mention that they
are pointers to an array of structures with per-$OBJECT data.
The private_objs field was added later on by commit b430c27a7de3 ("drm:
Add driver-private objects to atomic state") reusing the same sentence
than the crtcs field, probably due to copy and paste.
While these fields are indeed pointers to an array, each item of that
array contain a pointer to the object structure affected by the update,
and its old and new state. There's no per-object data there, and there's
more than just a pointer to the objects.
Let's rephrase those fields a bit to better match the current situation.
Maxime Ripard [Thu, 14 Dec 2023 10:09:13 +0000 (11:09 +0100)]
drm/atomic: Remove inexistent reference
Commit 63e83c1dba54 ("drm: Consolidate connector arrays in
drm_atomic_state") removed the connector_states field but didn't remove
its mention in the num_connectors documentation.
projects / linux.git / log