Changes in drm_accel.c

45c4d994 - accel: Use XArray instead of IDR for minors

Fri, 23 Aug 2024 16:30:47 +0000

accel: Use XArray instead of IDR for minorsAccel minor management is based on DRM (and is also using structdrm_minor internally), since DRM is using XArray for minors, it makessense to also convert accel.As the two implementations are identical (only difference being theunderlying xarray), move the accel_minor_* functionality to DRM.Signed-off-by: Michał Winiarski Acked-by: James Zhu Acked-by: Christian König Link: https://patchwork.freedesktop.org/patch/msgid/20240823163048.2676257-3-michal.winiarski@intel.comSigned-off-by: Christian König List of files: /linux-6.15/drivers/accel/drm_accel.c

576d7cc5 - accel: constify the struct device_type usage

Mon, 19 Feb 2024 19:48:28 +0000

accel: constify the struct device_type usageSince commit aed65af1cc2f ("drivers: make device_type const"), the drivercore can properly handle constant struct device_type. Move theaccel_sysfs_device_minor variable to be a constant structure as well,placing it into read-only memory which can not be modified at runtime.Cc: Greg Kroah-Hartman Signed-off-by: Ricardo B. Marliere Reviewed-by: Oded Gabbay Signed-off-by: Oded Gabbay List of files: /linux-6.15/drivers/accel/drm_accel.c

64c39a93 - accel: Include

Wed, 22 Nov 2023 12:09:34 +0000

accel: Include One of the source files includes via ,which will be removed. Include drm_auth.h directly.Signed-off-by: Thomas Zimmermann Cc: Oded Gabbay Reviewed-by: David Airlie Reviewed-by: Daniel Vetter Acked-by: Alex Deucher Link: https://patchwork.freedesktop.org/patch/msgid/20231122122449.11588-6-tzimmermann@suse.de List of files: /linux-6.15/drivers/accel/drm_accel.c

e11a7d2c - accel: make accel_class a static const structure

Tue, 20 Jun 2023 18:25:29 +0000

accel: make accel_class a static const structureNow that the driver core allows for struct class to be in read-onlymemory, move the accel_class structure to be declared at build timeplacing it into read-only memory, instead of having to be dynamicallyallocated at boot time.Cc: dri-devel@lists.freedesktop.orgSuggested-by: Greg Kroah-Hartman Signed-off-by: Ivan Orlov Signed-off-by: Greg Kroah-Hartman Reviewed-by: Tomer Tayar Signed-off-by: Oded Gabbay List of files: /linux-6.15/drivers/accel/drm_accel.c

8e455145 - drm/debugfs: rework drm_debugfs_create_files implementation v2

Tue, 29 Aug 2023 11:01:15 +0000

drm/debugfs: rework drm_debugfs_create_files implementation v2Use managed memory allocation for this. That allows us to not keeptrack of all the files any more.v2: keep drm_debugfs_cleanup(), but rename to drm_debugfs_unregister(), we still need to cleanup the symlinkSigned-off-by: Christian König Link: https://patchwork.freedesktop.org/patch/msgid/20230829110115.3442-6-christian.koenig@amd.comReviewed-by: Andi Shyti List of files: /linux-6.15/drivers/accel/drm_accel.c

0b30d57a - drm/debugfs: rework debugfs directory creation v5

Tue, 29 Aug 2023 11:01:13 +0000

drm/debugfs: rework debugfs directory creation v5Instead of the per minor directories only create a single debugfsdirectory for the whole device directly when the device is initialized.For DRM devices each minor gets a symlink to the per device directoryfor now until we can be sure that this isn't useful any more in any way.Accel devices create only the per device directory and also drops the midlayer callback to create driver specific files.v2: cleanup accel component as wellv3: fix typo when debugfs is disabledv4: call drm_debugfs_dev_fini() during release as well, some kerneldoc typos fixedv5: rebased and one more kerneldoc fixSigned-off-by: Christian König Link: https://patchwork.freedesktop.org/patch/msgid/20230829110115.3442-4-christian.koenig@amd.comReviewed-by: Andi Shyti List of files: /linux-6.15/drivers/accel/drm_accel.c

e76e7ec8 - drm/debugfs: drop debugfs_init() for the render and accel node v2

Tue, 29 Aug 2023 11:01:11 +0000

drm/debugfs: drop debugfs_init() for the render and accel node v2We want to remove per minor debugfs directories. Start by stoppingdrivers from adding anything inside of those in the mid layer callback.v2: drop it for the accel node as wellSigned-off-by: Christian König Tested-by: Stanislaw Gruszka Reviewed-by: Andi Shyti Link: https://patchwork.freedesktop.org/patch/msgid/20230829110115.3442-2-christian.koenig@amd.com List of files: /linux-6.15/drivers/accel/drm_accel.c

1aaba11d - driver core: class: remove module * from class_create()

Mon, 13 Mar 2023 18:18:35 +0000

driver core: class: remove module * from class_create()The module pointer in class_create() never actually did anything, and itshouldn't have been requred to be set as a parameter even if it didsomething. So just remove it and fix up all callers of the function inthe kernel tree at the same time.Cc: "Rafael J. Wysocki" Acked-by: Benjamin Tissoires Link: https://lore.kernel.org/r/20230313181843.1207845-4-gregkh@linuxfoundation.orgSigned-off-by: Greg Kroah-Hartman List of files: /linux-6.15/drivers/accel/drm_accel.c

dcde56bb - Fix mismerge due to devnode now taking a 'const *' device

Fri, 16 Dec 2022 19:04:15 +0000

Fix mismerge due to devnode now taking a 'const *' deviceThis was a mismerge of a semantic conflict in my merge of the drivercore updates, where commit ff62b8e6588f ("driver core: make structclass.devnode() take a const *") changed the devnode function pointertype. In the meantime, the drm tree I merged earlier had introduced anew use of that in commit 8bf4889762a8 ("drivers/accel: define kconfigand register a new major").And of course this happens when I'm traveling with my laptop, and thusdidn't do a full allmodconfig build between every pull and beforepushing my work out.So I only noticed later as I was doing my full build.Brown-paper-bag-by: Linus Torvalds Fixes: 71a7507afbc3 Merge tag 'driver-core-6.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-coreSigned-off-by: Linus Torvalds List of files: /linux-6.15/drivers/accel/drm_accel.c

2c204f3d - accel: add dedicated minor for accelerator devices

Mon, 31 Oct 2022 13:33:06 +0000

accel: add dedicated minor for accelerator devicesThe accelerator devices are exposed to user-space using a dedicatedmajor. In addition, they are represented in /dev with new, dedicateddevice char names: /dev/accel/accel*. This is done to make sure anyuser-space software that tries to open a graphic card won't openthe accelerator device by mistake.The above implies that the minor numbering should be separated fromthe rest of the DRM devices. However, to avoid code duplication, wewant the drm_minor structure to be able to represent the acceleratordevice.To achieve this, we add a new drm_minor* to drm_device that representsthe accelerator device. This pointer is initialized for drivers thatdeclare they handle compute accelerator, using a new driver featureflag called DRIVER_COMPUTE_ACCEL. It is important to note that thisdriver feature is mutually exclusive with DRIVER_RENDER. Devices thatwant to expose both graphics and compute device char files should behandled by two drivers that are connected using the auxiliary busframework.In addition, we define a different IDR to handle the acceleratorsminors. This is done to make the minor's index be identical to thedevice index in /dev/. Any access to the IDR is done solelyby functions in accel_drv.c, as the IDR is define as static. TheDRM core functions call those functions in case they detect the minor'stype is DRM_MINOR_ACCEL.We define a separate accel_open function (from drm_open) that theaccel drivers should set as their open callback function. Both thesefunctions eventually call the same drm_open_helper(), which had to bechanged to be non-static so it can be called from accel_drv.c.accel_open() only partially duplicates drm_open as I removed some codefrom it that handles legacy devices.To help new drivers, I defined DEFINE_DRM_ACCEL_FOPS macro to easilyset the required function operations pointers structure.Signed-off-by: Oded Gabbay Reviewed-by: Greg Kroah-Hartman Reviewed-by: Jeffrey Hugo Reviewed-by: Dave Airlie Acked-by: Thomas Zimmermann Acked-by: Jacek Lawrynowicz Tested-by: Jacek Lawrynowicz Reviewed-by: Melissa Wen List of files: /linux-6.15/drivers/accel/drm_accel.c

8bf48897 - drivers/accel: define kconfig and register a new major

Mon, 31 Oct 2022 13:28:35 +0000

drivers/accel: define kconfig and register a new majorAdd a new Kconfig for the accel subsystem. The Kconfig currentlycontains only the basic CONFIG_DRM_ACCEL option that will be used todecide whether to compile the accel registration code. Therefore, thekconfig option is defined as bool.The accel code will be compiled as part of drm.ko and will be calleddirectly from the DRM core code. The reason we compile it as part ofdrm.ko and not as a separate module is because of cyclic dependencybetween drm.ko and the separate module (if it would have existed).This is due to the fact that DRM core code calls accel functions andvice-versa.The accelerator devices will be exposed to the user space with a new,dedicated major number - 261.The accel init function registers the new major number as a char deviceand create corresponding sysfs and debugfs root entries, similar towhat is done in DRM init function.I added a new header called drm_accel.h to include/drm/, that will holdthe prototypes of the drm_accel.c functions. In case CONFIG_DRM_ACCELis set to 'N', that header will contain empty inline implementations ofthose functions, to allow DRM core code to compile successfullywithout dependency on CONFIG_DRM_ACCEL.I Updated the MAINTAINERS file accordingly with the newly added folderand I have taken the liberty to appropriate the dri-devel mailing listand the dri-devel IRC channel for the accel subsystem.Signed-off-by: Oded Gabbay Reviewed-by: Greg Kroah-Hartman Reviewed-by: Jeffrey Hugo Reviewed-by: Dave Airlie Acked-by: Thomas Zimmermann Acked-by: Jacek Lawrynowicz Tested-by: Jacek Lawrynowicz Reviewed-by: Melissa Wen List of files: /linux-6.15/drivers/accel/drm_accel.c