summaryrefslogtreecommitdiff
path: root/drivers/accel
AgeCommit message (Collapse)Author
2026-04-01accel/amdxdna: Create common PSP interfaces for AIE2 and AIE4David Zhang
The AIE2 and AIE4 use the similar interface to PSP (Platform Security Processor). Move the PSP implementation into aie_psp.c so both platforms use the same path and future AIE4 PSP work can build on it. Co-developed-by: Hayden Laccabue <Hayden.Laccabue@amd.com> Signed-off-by: Hayden Laccabue <Hayden.Laccabue@amd.com> Signed-off-by: David Zhang <yidong.zhang@amd.com> Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260330163705.3153647-4-lizhi.hou@amd.com
2026-04-01accel/amdxdna: Add basic support for AIE4 devicesDavid Zhang
Add initial support for AIE4 devices (PCI device IDs 0x17F2 and 0x1B0B), including: Device initialization Basic mailbox communication SR-IOV enablement This lays the groundwork for full AIE4 support. Co-developed-by: Hayden Laccabue <Hayden.Laccabue@amd.com> Signed-off-by: Hayden Laccabue <Hayden.Laccabue@amd.com> Signed-off-by: David Zhang <yidong.zhang@amd.com> Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260330163705.3153647-3-lizhi.hou@amd.com
2026-04-01accel/amdxdna: Create shared functions for AIE2 and AIE4Lizhi Hou
The AIE4 platform uses a mailbox management channel mechanism similar to AIE2 to communicate with the firmware. Create aie.h and aie.c and move the functions and structures that can be shared by both platforms from the AIE2-specific files into these common files. This allows AIE2 and AIE4 to reuse the same implementation and reduces code duplication. Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260330163705.3153647-2-lizhi.hou@amd.com
2026-03-31Merge drm/drm-next into drm-misc-nextThomas Zimmermann
Backmerging to get fixes and updates from v7.0-rc6. Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
2026-03-31BackMerge tag 'v7.0-rc6' into drm-nextDave Airlie
Linux 7.0-rc6 Requested by a few people on irc to resolve conflicts in other tress. Signed-off-by: Dave Airlie <airlied@redhat.com>
2026-03-30Merge drm/drm-fixes into drm-misc-next-fixesMaxime Ripard
Boris needs 7.0-rc6 for a shmem helper fix. Signed-off-by: Maxime Ripard <mripard@kernel.org>
2026-03-27accel/qaic: Retain bootlogs that overflowZack McKevitt
When a bootlog requires multiple MHI messages to transfer fully, the messages prior to the final message may have MHI overflow status set. Preserve these log messages in the accumulating bootlog. Do not treat overflow as an error. Signed-off-by: Zack McKevitt <zmckevit@qti.qualcomm.com> Signed-off-by: Troy Hanson <thanson@qti.qualcomm.com> Signed-off-by: Youssef Samir <youssef.abdulrahman@oss.qualcomm.com> Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Link: https://patch.msgid.link/20251223170511.2277302-1-youssef.abdulrahman@oss.qualcomm.com
2026-03-27accel/qaic: Simplify bootlog line handlingTroy Hanson
Instead of storing and emitting bootlogs as individual null-terminated lines, concatenate them, and later emit them to debugfs in a single call. Do not insert a null terminator on messages received from the device. Instead use the message length when subsequently storing the message. Exclude trailing nulls to normalize AIC100 and AIC200 bootlog lines. Signed-off-by: Troy Hanson <thanson@qti.qualcomm.com> Signed-off-by: Youssef Samir <youssef.abdulrahman@oss.qualcomm.com> Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Link: https://patch.msgid.link/20251223170226.2275812-1-youssef.abdulrahman@oss.qualcomm.com
2026-03-27accel/qaic: Update copyright headers to yearless formatZack McKevitt
Update copyright headers in accordance with Qualcomm's current legal policy. Signed-off-by: Zack McKevitt <zmckevit@qti.qualcomm.com> Signed-off-by: Youssef Samir <youssef.abdulrahman@oss.qualcomm.com> Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Link: https://patch.msgid.link/20251223172425.2283978-1-youssef.abdulrahman@oss.qualcomm.com
2026-03-27accel/qaic: Handle DBC deactivation if the owner went awayYoussef Samir
When a DBC is released, the device sends a QAIC_TRANS_DEACTIVATE_FROM_DEV transaction to the host over the QAIC_CONTROL MHI channel. QAIC handles this by calling decode_deactivate() to release the resources allocated for that DBC. Since that handling is done in the qaic_manage_ioctl() context, if the user goes away before receiving and handling the deactivation, the host will be out-of-sync with the DBCs available for use, and the DBC resources will not be freed unless the device is removed. If another user loads and requests to activate a network, then the device assigns the same DBC to that network, QAIC will "indefinitely" wait for dbc->in_use = false, leading the user process to hang. As a solution to this, handle QAIC_TRANS_DEACTIVATE_FROM_DEV transactions that are received after the user has gone away. Fixes: 129776ac2e38 ("accel/qaic: Add control path") Signed-off-by: Youssef Samir <youssef.abdulrahman@oss.qualcomm.com> Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com> Link: https://patch.msgid.link/20260205123415.3870898-1-youssef.abdulrahman@oss.qualcomm.com
2026-03-27Merge tag 'drm-misc-next-2026-03-26' of ↵Dave Airlie
https://gitlab.freedesktop.org/drm/misc/kernel into drm-next drm-misc-next for v7.1: UAPI Changes: amdxdna: - support per-BO memory-usage queries docs: - Improve UAPI documentation panthor: - extend timestamp query with flags Core Changes: edid: - provide enum drm_output_color_format; mass-convert drivers gem-dma: - use drm_dev_dma_dev() for DMA mappings - set VM_DONTDUMP on mmap mipi-dbi: - drop simple-display; mass-convert drivers prime: - use drm_dev_dma_dev() for DMA mappings ttm: - improve handling of gfp_retry_mayfail Driver Changes: amdgpu: - use atomic_create_state for private_obj amdxdna: - refactor GEM implementation - fixes bridge: - provide clear-and-put helper for reliable cleanup - analogix_dp: Use DP helpers for link training - lontium-lt8713sx: Fix 64-bit division and Kconfig - samsung-dsim: Use clear-and-put imagination: - improve power-off sequence - support context-reset notification from firmware komeda: - support Arm China Linlon D6 plus DT bindings mediatek: - use drm_dev_dma_dev() for DMA mappings panel: - support Himax HX83121A plus DT bindings - support JuTouch JT070TM041 plus DT bindings - support Samsung S6E8FC0 plus DT bindings - himax-hx83102c: support Samsung S6E8FC0 plus DT bindings; support backlight - ili9806e: support Rocktech RK050HR345-CT106A plus DT bindings - simple: support Tianma TM050RDH03 plus DT bindings panthor: - support various sources for timestamp queries - fixes omapdrm: - use atomic_create_state for private_obj rcar-du: - fix suspend/resume wrt VSP interface - fix leak of device_link - clean up sun4i: - use drm_dev_dma_dev() for DMA mappings tegra: - use atomic_create_state for private_obj xe: - send 'none' recovery method for XE_WEDGED_MODE_UPON_ANY_HANG_NO_RESET Signed-off-by: Dave Airlie <airlied@redhat.com> From: Thomas Zimmermann <tzimmermann@suse.de> Link: https://patch.msgid.link/20260326151812.GA76082@linux.fritz.box
2026-03-26accel/amdxdna: Fix leak when pinning ubuf pagesMax Zhen
When pin_user_pages_fast() returns fewer pages than requested, the pages that were successfully pinned are not released, leading to a leak. Fix this by unpinning any partially pinned pages before returning failure. Fixes: bd72d4acda10 ("accel/amdxdna: Support user space allocated buffer") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Max Zhen <max.zhen@amd.com> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260326010642.2596525-1-lizhi.hou@amd.com
2026-03-26BackMerge tag 'v7.0-rc4' into drm-nextDave Airlie
Linux 7.0-rc4 Needed for rust tree. Signed-off-by: Dave Airlie <airlied@redhat.com>
2026-03-25accel/amdxdna: Add per-process BO memory usage query supportMax Zhen
Add support for querying per-process buffer object (BO) memory usage through the amdxdna GET_ARRAY UAPI. Introduce a new query type, DRM_AMDXDNA_BO_USAGE, along with struct amdxdna_drm_bo_usage to report BO memory usage statistics, including heap, total, and internal usage. Track BO memory usage on a per-client basis by maintaining counters in GEM open/close and heap allocation/free paths. This ensures the reported statistics reflect the current memory footprint of each process. Wire the new query into the GET_ARRAY implementation to expose the usage information to userspace. Link: https://github.com/amd/xdna-driver/commit/0546f2aaadbdacf1c3556410ecd71622044cd916 Signed-off-by: Max Zhen <max.zhen@amd.com> Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260324163159.2425461-1-lizhi.hou@amd.com
2026-03-24accel/ivpu: Add disable clock relinquish workaround for NVL-A0Karol Wachowski
Turn on disable clock relinquish workaround for Nova Lake A0. Without this workaround NPU may not power off correctly after inference, leading to unexpected system behavior. Fixes: 550f4dd2cedd ("accel/ivpu: Add support for Nova Lake's NPU") Cc: <stable@vger.kernel.org> # v6.19+ Reviewed-by: Lizhi.hou <lizhi.hou@amd.com> Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com> Link: https://patch.msgid.link/20260323095029.64613-1-karol.wachowski@linux.intel.com
2026-03-23accel/amdxdna: Return ERR_PTR on dma_alloc_noncoherent failureWendy Liang
dma_alloc_noncoherent() returns NULL on failure, but callers of aie2_alloc_msg_buffer() check for IS_ERR(). Return ERR_PTR(-ENOMEM) instead of NULL to match the amdxdna_iommu_alloc() path and the caller's error checking convention. Fixes: ece3e8980907 ("accel/amdxdna: Allow forcing IOVA-based DMA via module parameter") Signed-off-by: Wendy Liang <wendy.liang@amd.com> Reviewed-by: Karol Wachowski <karol.wachowski@linux.intel.com> Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260323173719.2311474-1-lizhi.hou@amd.com
2026-03-23accel/amdxdna: fix missing newline in pr_err messagehaoyu.lu
Add missing newline to pr_err message in amdxdna_mailbox.c. Fixes: b87f920b9344 ("accel/amdxdna: Support hardware mailbox") Signed-off-by: haoyu.lu <hechushiguitu666@gmail.com> Reviewed-by: Lizhi.hou <lizhi.hou@amd.com> Signed-off-by: Lizhi.hou <lizhi.hou@amd.com> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260323034933.216-1-hechushiguitu666@gmail.com
2026-03-20accel/amdxdna: Refactor GEM BO handling and add helper APIs for address ↵Max Zhen
retrieval Refactor amdxdna GEM buffer object (BO) handling to simplify address management and unify BO type semantics. Introduce helper APIs to retrieve commonly used BO addresses: - User virtual address (UVA) - Kernel virtual address (KVA) - Device address (IOVA/PA) These helpers centralize address lookup logic and avoid duplicating BO-specific handling across submission and execution paths. This also improves readability and reduces the risk of inconsistent address handling in future changes. As part of the refactor: - Rename SHMEM BO type to SHARE to better reflect its usage. - Merge CMD BO handling into SHARE, removing special-case logic for command buffers. - Consolidate BO type handling paths to reduce code duplication and simplify maintenance. No functional change is intended. The refactor prepares the driver for future enhancements by providing a cleaner abstraction for BO address management. Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Max Zhen <max.zhen@amd.com> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260320210615.1973016-1-lizhi.hou@amd.com
2026-03-20accel/ivpu: Perform engine reset instead of device recovery on TDRKarol Wachowski
Replace full device recovery on TDR timeout with per-context abort, allowing individual context handling instead of resetting the entire device. Extend ivpu_jsm_reset_engine() to return the list of contexts impacted by the engine reset and use that information to abort only the affected contexts. Only check for potentially faulty contexts when the engine reset was not triggered by an MMU fault or a job completion error status. This prevents misidentifying non-guilty contexts that happened to be running at the time of the fault. Trigger full device recovery if no contexts were marked by engine reset if triggered by job completion timeout, as there is no way to identify guilty one. Add engine reset counter to debugfs for engine resets bookkeeping for debugging/testing purposes. Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com> Link: https://patch.msgid.link/20260318093927.4080303-1-karol.wachowski@linux.intel.com
2026-03-17accel/amdxdna: Support retrieving hardware context debug informationLizhi Hou
The firmware implements the GET_APP_HEALTH command to collect debug information for a specific hardware context. When a command times out, the driver issues this command to collect the relevant debug information. User space tools can also retrieve this information through the hardware context query IOCTL. Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260317044906.1513133-1-lizhi.hou@amd.com
2026-03-16accel/amdxdna: Add debug prints for command submissionLizhi Hou
Add debug prints to help diagnose issues with incoming command submissions. Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260316175642.1451749-1-lizhi.hou@amd.com
2026-03-12accel/amdxdna: Allow forcing IOVA-based DMA via module parameterLizhi Hou
The amdxdna driver normally performs DMA using userspace virtual address plus PASID. For debugging and validation purposes, add a module parameter, force_iova, to force DMA to go through IOMMU IOVA mapping. When force_iova=1 is set, the driver will allocate and map DMA buffers using IOVA. Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260126193001.1400545-1-lizhi.hou@amd.com
2026-03-12Merge drm/drm-next into drm-misc-nextMaxime Ripard
Biju Das needs a patch for rz-du merged in 7.0-rc3 Signed-off-by: Maxime Ripard <mripard@kernel.org>
2026-03-11accel/amdxdna: Support sensors for column utilizationMario Limonciello (AMD)
The AMD PMF driver provides realtime column utilization (npu_busy) metrics for the NPU. Extend the DRM_IOCTL_AMDXDNA_GET_INFO sensor query to expose these metrics to userspace. Add AMDXDNA_SENSOR_TYPE_COLUMN_UTILIZATION to the sensor type enum and update aie2_get_sensors() to return both the total power and up to 8 column utilization sensors if the user buffer permits. Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> [lizhi: support legacy tool which uses small buffer. checkpatch cleanup] Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260311171842.473453-1-lizhi.hou@amd.com
2026-03-11accel/amdxdna: Add IOCTL to retrieve realtime NPU power estimateLizhi Hou
The AMD PMF driver provides an interface to obtain realtime power estimates for the NPU. Expose this information to userspace through a new DRM_IOCTL_AMDXDNA_GET_INFO parameter, allowing applications to query the current NPU power level. Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Reviewed-by: Shyam Sundar S K <Shyam-sundar.S-k@amd.com> (Update comment to indicate power and utilization) Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260228061109.361239-2-superm1@kernel.org
2026-03-11accel/amdxdna: Import AMD_PMF namespaceMario Limonciello (AMD)
The amdxdna driver uses amd_pmf_get_npu_data() which is exported in the AMD_PMF namespace. Import the AMD_PMF namespace. Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260301005028.367618-1-superm1@kernel.org
2026-03-11accel/ivpu: Apply minor code style cleanups to align with kernel styleKarol Wachowski
Replace direct import_attach test with drm_gem_is_imported() in ivpu_bo_bind(). Replace kzalloc(sizeof(*bo), GFP_KERNEL) with kzalloc_obj() in ivpu_gem_create_object(). Remove unnecessary cast to bool in ivpu_dbg_bo(). No functional changes. Reviewed-by: Maciej Falkowski <maciej.falkowski@linux.intel.com> Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com> Link: https://patch.msgid.link/20260310120736.3341679-1-karol.wachowski@linux.intel.com
2026-03-11Merge v7.0-rc3 into drm-nextSimona Vetter
Requested by Maxime Ripard for drm-misc-next because renesas people need fb797a70108f ("drm: renesas: rz-du: mipi_dsi: Set DSI divider"). Signed-off-by: Simona Vetter <simona.vetter@ffwll.ch>
2026-03-10accel/amdxdna: Fix runtime suspend deadlock when there is pending jobLizhi Hou
The runtime suspend callback drains the running job workqueue before suspending the device. If a job is still executing and calls pm_runtime_resume_and_get(), it can deadlock with the runtime suspend path. Fix this by moving pm_runtime_resume_and_get() from the job execution routine to the job submission routine, ensuring the device is resumed before the job is queued and avoiding the deadlock during runtime suspend. Fixes: 063db451832b ("accel/amdxdna: Enhance runtime power management") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260310180058.336348-1-lizhi.hou@amd.com
2026-03-10accel/ivpu: Remove boot params address setting via MMIO registerAndrzej Kacprowski
The NPU 60XX uses the default boot params location specified in the firmware image header, consistent with earlier generations. Remove the unnecessary MMIO register write, freeing the AON register for future use. Fixes: 44e4c88951fa ("accel/ivpu: Implement warm boot flow for NPU6 and unify boot handling") Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski@linux.intel.com> Reviewed-by: Karol Wachowski <karol.wachowski@linux.intel.com> Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com> Link: https://patch.msgid.link/20260305142226.194995-1-andrzej.kacprowski@linux.intel.com (cherry picked from commit 81e62e7bf8b9309bf0febdf00940818f98bc23d8) Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
2026-03-09accel/ivpu: Test for imported buffers with drm_gem_is_imported()Karol Wachowski
Instead of testing import_attach for imported GEM buffers, invoke drm_gem_is_imported() to do the test. The test itself does not change. Suggested-by: Thomas Zimmermann <tzimmermann@suse.de> Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de> Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com> Link: https://patch.msgid.link/20260309092755.3165130-1-karol.wachowski@linux.intel.com
2026-03-09accel/ivpu: Remove boot params address setting via MMIO registerAndrzej Kacprowski
The NPU 60XX uses the default boot params location specified in the firmware image header, consistent with earlier generations. Remove the unnecessary MMIO register write, freeing the AON register for future use. Fixes: 44e4c88951fa ("accel/ivpu: Implement warm boot flow for NPU6 and unify boot handling") Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski@linux.intel.com> Reviewed-by: Karol Wachowski <karol.wachowski@linux.intel.com> Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com> Link: https://patch.msgid.link/20260305142226.194995-1-andrzej.kacprowski@linux.intel.com
2026-03-05accel: ethosu: Handle possible underflow in IFM size calculationsRob Herring (Arm)
If the command stream has larger padding sizes than the IFM and OFM diminsions, then the calculations will underflow to a negative value. The result is a very large region bounds which is caught on submit, but it's better to catch it earlier. Current mesa ethosu driver has a signedness bug which resulted in padding of 127 (the max) and triggers this issue. Reviewed-and-Tested-by: Anders Roxell <anders.roxell@linaro.org> Link: https://patch.msgid.link/20260218-ethos-fixes-v1-3-be3fa3ea9a30@kernel.org Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
2026-03-05accel: ethosu: Fix NPU_OP_ELEMENTWISE validation with scalarRob Herring (Arm)
The NPU_OP_ELEMENTWISE instruction uses a scalar value for IFM2 if the IFM2_BROADCAST "scalar" mode is set. It is a bit (7) on the u65 and part of a field (bits 3:0) on the u85. The driver was hardcoded to the u85. Fixes: 5a5e9c0228e6 ("accel: Add Arm Ethos-U NPU driver") Reviewed-and-Tested-by: Anders Roxell <anders.roxell@linaro.org> Link: https://patch.msgid.link/20260218-ethos-fixes-v1-2-be3fa3ea9a30@kernel.org Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
2026-03-05accel: ethosu: Fix job submit error clean-up refcount underflowsRob Herring (Arm)
If the job submit fails before adding the job to the scheduler queue such as when the GEM buffer bounds checks fail, then doing a ethosu_job_put() results in a pm_runtime_put_autosuspend() without the corresponding pm_runtime_resume_and_get(). The dma_fence_put()'s are also unnecessary, but seem to be harmless. Split the ethosu_job_cleanup() function into 2 parts for the before and after the job is queued. Fixes: 5a5e9c0228e6 ("accel: Add Arm Ethos-U NPU driver") Reviewed-and-Tested-by: Anders Roxell <anders.roxell@linaro.org> Link: https://patch.msgid.link/20260218-ethos-fixes-v1-1-be3fa3ea9a30@kernel.org Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
2026-03-05accel/amdxdna: Split mailbox channel create functionLizhi Hou
The management channel used for firmware control command submission is currently created after the firmware is started. If channel creation fails (for example, due to memory allocation failure or workqueue creation interruption), the firmware remains in a pending state and is unable to receive any control commands. To avoid leaving the firmware in this inconsistent state, split xdna_mailbox_create_channel() into two separate functions so that resource allocation can be completed before interacting with the hardware. xdna_mailbox_alloc_channel() Allocates memory and initializes the workqueue. This can be called earlier, before interacting with the hardware. xdna_mailbox_start_channel() Performs the hardware interaction required to start the channel. Rename xdna_mailbox_destroy_channel() to xdna_mailbox_free_channel(). Ensure that xdna_mailbox_stop_channel() and xdna_mailbox_free_channel() properly unwind the corresponding start and allocation steps, respectively. Fixes: b87f920b9344 ("accel/amdxdna: Support hardware mailbox") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260305062041.3954024-1-lizhi.hou@amd.com
2026-03-04accel/amdxdna: Fix major version check on NPU1 platformLizhi Hou
Add the missing major number in npu1_fw_feature_table. Without the major version specified, the firmware feature check fails, preventing new firmware commands from being enabled on the NPU1 platform. With the correct major version populated, the driver properly detects firmware support and enables the new command. Fixes: f1eac46fe5f7 ("accel/amdxdna: Update firmware version check for latest firmware") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260304195012.3616908-1-lizhi.hou@amd.com
2026-03-03accel/ivpu: Limit number of maximum contexts and doorbells per userKarol Wachowski
Implement per-user resource limits to prevent resource exhaustion. Root users can allocate all available contexts (128) and doorbells (255), while non-root users are limited to half of the available resources (64 contexts and 127 doorbells respectively). This prevents scenarios where a single user could monopolize NPU resources and starve other users on multi-user systems. Change doorbell ID and command queue ID allocation errors to debug messages as those are user triggered. Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com> Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> Signed-off-by: Maciej Falkowski <maciej.falkowski@linux.intel.com> Link: https://patch.msgid.link/20260302202207.469442-1-maciej.falkowski@linux.intel.com
2026-03-03Merge drm/drm-next into drm-misc-nextThomas Zimmermann
Backmerge fixes from v7.0-rc2 into drm-misc-next. Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
2026-03-02accel/amdxdna: Fix NULL pointer dereference of mgmt_channLizhi Hou
mgmt_chann may be set to NULL if the firmware returns an unexpected error in aie2_send_mgmt_msg_wait(). This can later lead to a NULL pointer dereference in aie2_hw_stop(). Fix this by introducing a dedicated helper to destroy mgmt_chann and by adding proper NULL checks before accessing it. Fixes: b87f920b9344 ("accel/amdxdna: Support hardware mailbox") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260226213857.3068474-1-lizhi.hou@amd.com
2026-02-27accel/amdxdna: Fill invalid payload for failed commandLizhi Hou
Newer userspace applications may read the payload of a failed command to obtain detailed error information. However, the driver and old firmware versions may not support returning advanced error information. In this case, initialize the command payload with an invalid value so userspace can detect that no detailed error information is available. Fixes: aac243092b70 ("accel/amdxdna: Add command execution") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260227004841.3080241-1-lizhi.hou@amd.com
2026-02-26accel/ivpu: Update FW Boot API to version 3.29.4Maciej Falkowski
Update firmware boot API to the version 3.29.4. Remove unused boot parameters from the vpu_firmware_header structure. Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> Signed-off-by: Maciej Falkowski <maciej.falkowski@linux.intel.com> Link: https://patch.msgid.link/20260220160116.220367-1-maciej.falkowski@linux.intel.com
2026-02-25accel/amdxdna: Use a different name for latest firmwareLizhi Hou
Using legacy driver with latest firmware causes a power off issue. Fix this by assigning a different filename (npu_7.sbin) to the latest firmware. The driver attempts to load the latest firmware first and falls back to the previous firmware version if loading fails. Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/5009 Fixes: f1eac46fe5f7 ("accel/amdxdna: Update firmware version check for latest firmware") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260225204752.2711734-1-lizhi.hou@amd.com
2026-02-23accel/amdxdna: Validate command buffer payload countLizhi Hou
The count field in the command header is used to determine the valid payload size. Verify that the valid payload does not exceed the remaining buffer space. Fixes: aac243092b70 ("accel/amdxdna: Add command execution") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260219211946.1920485-1-lizhi.hou@amd.com
2026-02-23accel/amdxdna: Prevent ubuf size overflowLizhi Hou
The ubuf size calculation may overflow, resulting in an undersized allocation and possible memory corruption. Use check_add_overflow() helpers to validate the size calculation before allocation. Fixes: bd72d4acda10 ("accel/amdxdna: Support user space allocated buffer") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260217192815.1784689-1-lizhi.hou@amd.com
2026-02-23accel/amdxdna: Fix out-of-bounds memset in command slot handlingLizhi Hou
The remaining space in a command slot may be smaller than the size of the command header. Clearing the command header with memset() before verifying the available slot space can result in an out-of-bounds write and memory corruption. Fix this by moving the memset() call after the size validation. Fixes: 3d32eb7a5ecf ("accel/amdxdna: Fix cu_idx being cleared by memset() during command setup") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260217185415.1781908-1-lizhi.hou@amd.com
2026-02-23accel/amdxdna: Fix command hang on suspended hardware contextLizhi Hou
When a hardware context is suspended, the job scheduler is stopped. If a command is submitted while the context is suspended, the job is queued in the scheduler but aie2_sched_job_run() is never invoked to restart the hardware context. As a result, the command hangs. Fix this by modifying the hardware context suspend routine to keep the job scheduler running so that queued jobs can trigger context restart properly. Fixes: aac243092b70 ("accel/amdxdna: Add command execution") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260211205341.722982-1-lizhi.hou@amd.com
2026-02-23accel/amdxdna: Fix suspend failure after enabling turbo modeLizhi Hou
Enabling turbo mode disables hardware clock gating. Suspend requires hardware clock gating to be re-enabled, otherwise suspend will fail. Fix this by calling aie2_runtime_cfg() from aie2_hw_stop() to re-enable clock gating during suspend. Also ensure that firmware is initialized in aie2_hw_start() before modifying clock-gating settings during resume. Fixes: f4d7b8a6bc8c ("accel/amdxdna: Enhance power management settings") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260211204716.722788-1-lizhi.hou@amd.com
2026-02-23accel/amdxdna: Fix dead lock for suspend and resumeLizhi Hou
When an application issues a query IOCTL while auto suspend is running, a deadlock can occur. The query path holds dev_lock and then calls pm_runtime_resume_and_get(), which waits for the ongoing suspend to complete. Meanwhile, the suspend callback attempts to acquire dev_lock and blocks, resulting in a deadlock. Fix this by releasing dev_lock before calling pm_runtime_resume_and_get() and reacquiring it after the call completes. Also acquire dev_lock in the resume callback to keep the locking consistent. Fixes: 063db451832b ("accel/amdxdna: Enhance runtime power management") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260211204644.722758-1-lizhi.hou@amd.com
2026-02-23accel/amdxdna: Reduce log noise during process terminationMario Limonciello
During process termination, several error messages are logged that are not actual errors but expected conditions when a process is killed or interrupted. This creates unnecessary noise in the kernel log. The specific scenarios are: 1. HMM invalidation returns -ERESTARTSYS when the wait is interrupted by a signal during process cleanup. This is expected when a process is being terminated and should not be logged as an error. 2. Context destruction returns -ENODEV when the firmware or device has already stopped, which commonly occurs during cleanup if the device was already torn down. This is also an expected condition during orderly shutdown. Downgrade these expected error conditions from error level to debug level to reduce log noise while still keeping genuine errors visible. Fixes: 97f27573837e ("accel/amdxdna: Fix potential NULL pointer dereference in context cleanup") Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com> Link: https://patch.msgid.link/20260210164521.1094274-3-mario.limonciello@amd.com
generated by cgit v1.2.3 (git 2.0.4) at 2026-07-23 04:06:35 +0000