summaryrefslogtreecommitdiffstats
path: root/drivers/accel/habanalabs
Commit message (Collapse)AuthorAgeFilesLines
* accel/habanalabs: modify pci health checkOfir Bitton2024-02-261-3/+3
| | | | | | | | | | | | Today we read PCI VENDOR-ID in order to make sure PCI link is healthy. Apparently the VENDOR-ID might be stored on host and hence, when we read it we might not access the PCI bus. In order to make sure PCI health check is reliable, we will start checking the DEVICE-ID instead. Signed-off-by: Ofir Bitton <obitton@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: keep explicit size of reserved memory for FWTomer Tayar2024-02-262-2/+2
| | | | | | | | | | | | | | The reserved memory for FW is currently saved in an ASIC property in units of MB, just like the value that comes from FW. Except the fact that it is not clear from the property's name, it means also that a calculation to actual size is required everywhere that it is used. Modify the property to hold the size in bytes. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: handle reserved memory request when working with full FWTomer Tayar2024-02-261-11/+8
| | | | | | | | | | | | Currently the reserved memory request from FW is handled when running with preboot only, but this request is relevant also when running with full FW. Modify to always handle this reservation request. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs/hwmon: rate limit errors user can generateOfir Bitton2024-02-261-14/+15
| | | | | | | | | | | Fetching sensor data can fail due to various reasons. In order not to pollute the kernel log, those error prints must be rate limited. Signed-off-by: Ofir Bitton <obitton@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs/gaudi2: drain event lacks rd/wr indicationOfir Bitton2024-02-261-11/+3
| | | | | | | | | | Due to a H/W issue, AXI drain event does not include a read/write indication, hence we remove this print. Signed-off-by: Ofir Bitton <obitton@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: fix error printDani Liberman2024-02-261-2/+2
| | | | | | | | | The unmasking is for event and it can be other event than RAZWI. Signed-off-by: Dani Liberman <dliberman@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: initialize maybe-uninitialized variablesTal Risin2024-02-261-2/+2
| | | | | | | | | Prevent static analysis warning. Signed-off-by: Tal Risin <trisin@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: fix debugfs files permissionsAvri Kehat2024-02-261-7/+7
| | | | | | | | | | debugfs files are created with permissions that don't align with the access requirements. Signed-off-by: Avri Kehat <akehat@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: fix glbl error cause handlingTomer Tayar2024-02-265-18/+35
| | | | | | | | | | | The glbl error cause handling has a wrong assumption that all error bits are consecutive. Fix the handling to check all relevant error bits per ASIC. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs/gaudi2: check extended errors according to PCIe addr_dec ↵Tomer Tayar2024-02-261-10/+8
| | | | | | | | | | | | | interrupt info The FW interrupt info for a PCIe addr_dec event is set correctly, so check for either global errors or razwi according to the indications there. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: modify print for skip loading linux FW to debug logTomer Tayar2024-02-261-1/+1
| | | | | | | | | | | | | | Skip loading a linux FW image into the device with the current supported ASICs is done for test purposes only. Moreover, for future supported ASICs it is possible that there won't be a need to load such an image. The print in such a case is therefore not needed in most cases, so replace the used dev_info() with dev_dbg(). Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: remove hop size from asic propertiesFarah Kassabri2024-02-266-46/+36
| | | | | | | | | | | The hop size related properties is a MMU properties and not asic properties. As for PMMU and HMMU we could have different sizes. Signed-off-by: Farah Kassabri <fkassabri@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: use kcalloc() instead of kzalloc()Erick Archer2024-02-261-1/+1
| | | | | | | | | | | | | | | | | | | | As noted in the "Deprecated Interfaces, Language Features, Attributes, and Conventions" documentation [1], size calculations (especially multiplication) should not be performed in memory allocator (or similar) function arguments due to the risk of them overflowing. This could lead to values wrapping around and a smaller allocation being made than the caller was expecting. Using those allocations could lead to linear overflows of heap memory and other misbehaviors. So, use the purpose specific kcalloc() function instead of the argument size * count in the kzalloc() function. Link: https://www.kernel.org/doc/html/next/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments [1] Link: https://github.com/KSPP/linux/issues/162 Signed-off-by: Erick Archer <erick.archer@gmx.com> Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs/goya: remove redundant assignment to pointer 'input'Colin Ian King2024-02-261-2/+1
| | | | | | | | | | | | | | The pointer input is assigned a value that is not read, it is being re-assigned again later with the same value. Resolve this by moving the declaration to input into the if block. Cleans up clang scan build warning: warning: Value stored to 'input' during its initialization is never read [deadcode.DeadStores] Signed-off-by: Colin Ian King <colin.i.king@intel.com> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs/gaudi2: fail memory memset when failing to copy QM packet ↵Tomer Tayar2024-02-261-2/+8
| | | | | | | | | | | | to device gaudi2_memset_memory_chunk_using_edma_qm() calls the access_dev_mem() ASIC function, but ignores its return value. Add this missing check. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: remove call to deprecated functionDani Liberman2024-02-262-3/+3
| | | | | | | | | In newer kernel versions, irq_set_affinity_hint() is deprecated. Instead, use the newer version which is irq_set_affinity_and_hint(). Signed-off-by: Dani Liberman <dliberman@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: Remove unnecessary braces from if statementMalkoot Khan2024-02-261-2/+1
| | | | | | | | | | | | The coding style in the Linux kernel prefers not to use braces for single-statement if conditions. This patch removes the unnecessary braces from an if statement in the file drivers/accel/habanalabs/common/command_submission.c, which also resolves a coding style warning. Signed-off-by: Malkoot Khan <engr.mkhan1990@gmail.com> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs/gaudi2: move HMMU page tables to device memoryFarah Kassabri2024-02-2610-382/+836
| | | | | | | | | | | | | | Currently the HMMU page tables reside in the host memory, which will cause host access from the device for every page walk. This can affect PCIe bandwidth in certain scenarios. To prevent that problem, HMMU page tables will be moved to the device memory so the miss transaction will read the hops from there instead of going to the host. Signed-off-by: Farah Kassabri <fkassabri@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: abort device reset for consecutive heartbeat failuresTomer Tayar2024-02-261-6/+8
| | | | | | | | | | | | The mechanism of aborting device reset for consecutive fatal errors is currently only for fatal errors that are reported by FW. A non-responsive FW and consecutive heartbeat failures is also considered fatal, so add them as well to this mechanism to avoid recurring device reset in such a case. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: fix DRAM BAR base address calculationTomer Tayar2024-02-261-1/+2
| | | | | | | | | | When the DRAM region size in the BAR is not a power of 2, calculating the corresponding BAR base address should be done using the offset from the DRAM start address, and not using directly the DRAM address. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: increase HL_MAX_STR to 64 bytes to avoid warningsKoby Elbaz2024-02-261-1/+1
| | | | | | | | | Fix a warning of a buffer overflow: ‘snprintf’ output between 38 and 47 bytes into a destination of size 32 Signed-off-by: Koby Elbaz <kelbaz@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs/gaudi2: add interrupt affinity for user interruptsDani Liberman2024-02-263-0/+42
| | | | | | | | | | | | | | | | User interrupts are MSIx interrupts coming from Gaudi2, that have specific range of IDs and are assigned to the sole use of the user process that opened the Gaudi2 device (reminder: there can be only a single user process running on Gaudi2 at any given time). The interrupts are allocated and managed by the driver and therefore, the user expects the driver to initialize them properly, which also includes setting the affinity to the related CPU cores of the device's NUMA node to get maximum performance. Signed-off-by: Dani Liberman <dliberman@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* Merge tag 'drm-next-2024-01-10' of git://anongit.freedesktop.org/drm/drmLinus Torvalds2024-01-1212-147/+250
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull drm updates from Dave Airlie: "This contains two major new drivers: - imagination is a first driver for Imagination Technologies devices, it only covers very specific devices, but there is hope to grow it - xe is a reboot of the i915 GPU (shares display) side using a more upstream focused development model, and trying to maximise code sharing. It's not enabled for any hw by default, and will hopefully get switched on for Intel's Lunarlake. This also drops a bunch of the old UMS ioctls. It's been dead long enough. amdgpu has a bunch of new color management code that is being used in the Steam Deck. amdgpu also has a new ACPI WBRF interaction to help avoid radio interference. Otherwise it's the usual lots of changes in lots of places. Detailed summary: new drivers: - imagination - new driver for Imagination Technologies GPU - xe - new driver for Intel GPUs using core drm concepts core: - add CLOSE_FB ioctl - remove old UMS ioctls - increase max objects to accomodate AMD color mgmt encoder: - create per-encoder debugfs directory edid: - split out drm_eld - SAD helpers - drop edid_firmware module parameter format-helper: - cache format conversion buffers sched: - move from kthread to workqueue - rename some internals - implement dynamic job-flow control gpuvm: - provide more features to handle GEM objects client: - don't acquire module reference displayport: - add mst path property documentation fdinfo: - alignment fix dma-buf: - add fence timestamp helper - add fence deadline support bridge: - transparent aux-bridge for DP/USB-C - lt8912b: add suspend/resume support and power regulator support panel: - edp: AUO B116XTN02, BOE NT116WHM-N21,836X2, NV116WHM-N49 - chromebook panel support - elida-kd35t133: rework pm - powkiddy RK2023 panel - himax-hx8394: drop prepare/unprepare and shutdown logic - BOE BP101WX1-100, Powkiddy X55, Ampire AM8001280G - Evervision VGG644804, SDC ATNA45AF01 - nv3052c: register docs, init sequence fixes, fascontek FS035VG158 - st7701: Anbernic RG-ARC support - r63353 panel controller - Ilitek ILI9805 panel controller - AUO G156HAN04.0 simplefb: - support memory regions - support power domains amdgpu: - add new 64-bit sequence number infrastructure - add AMD specific color management - ACPI WBRF support for RF interference handling - GPUVM updates - RAS updates - DCN 3.5 updates - Rework PCIe link speed handling - Document GPU reset types - DMUB fixes - eDP fixes - NBIO 7.9/7.11 updates - SubVP updates - XGMI PCIe state dumping for aqua vanjaram - GFX11 golden register updates - enable tunnelling on high pri compute amdkfd: - Migrate TLB flushing logic to amdgpu - Trap handler fixes - Fix restore workers handling on suspend/resume - Fix possible memory leak in pqm_uninit() - support import/export of dma-bufs using GEM handles radeon: - fix possible overflows in command buffer checking - check for errors in ring_lock i915: - reorg display code for reuse in xe driver - fdinfo memory stats printing - DP MST bandwidth mgmt improvements - DP panel replay enabling - MTL C20 phy state verification - MTL DP DSC fractional bpp support - Audio fastset support - use dma_fence interfaces instead of i915_sw_fence - Separate gem and display code - AUX register macro refactoring - Separate display module/device parameters - Move display capabilities debugfs under display - Makefile cleanups - Register cleanups - Move display lock inits under display/ - VLV/CHV DPIO PHY register and interface refactoring - DSI VBT sequence refactoring - C10/C20 PHY PLL hardware readout - DPLL code cleanups - Cleanup PXP plane protection checks - Improve display debug msgs - PSR selective fetch fixes/improvements - DP MST fixes - Xe2LPD FBC restrictions removed - DGFX uses direct VBT pin mapping - more MTL WAs - fix MTL eDP bug - eliminate use of kmap_atomic habanalabs: - sysfs entry to identify a device minor id with debugfs path - sysfs entry to expose device module id - add signed device info retrieval through INFO ioctl - add Gaudi2C device support - pcie reset prepare/done hooks msm: - Add support for SDM670, SM8650 - Handle the CFG interconnect to fix the obscure hangs / timeouts - Kconfig fix for QMP dependency - use managed allocators - DPU: SDM670, SM8650 support - DPU: Enable SmartDMA on SM8350 and SM8450 - DP: enable runtime PM support - GPU: add metadata UAPI - GPU: move devcoredumps to GPU device - GPU: convert to drm_exec ivpu: - update FW API - new debugfs file - a new NOP job submission test mode - improve suspend/resume - PM improvements - MMU PT optimizations - firmware profile frequency support - support for uncached buffers - switch to gem shmem helpers - replace kthread with threaded irqs rockchip: - rk3066_hdmi: convert to atomic - vop2: support nv20 and nv30 - rk3588 support mediatek: - use devm_platform_ioremap_resource - stop using iommu_present - MT8188 VDOSYS1 display support panfrost: - PM improvements - improve interrupt handling as poweroff qaic: - allow to run with single MSI - support host/device time sync - switch to persistent DRM devices exynos: - fix potential error pointer dereference - fix wrong error checking - add missing call to drm_atomic_helper_shutdown omapdrm: - dma-fence lockdep annotation fix tidss: - dma-fence lockdep annotation fix - support for AM62A7 v3d: - BCM2712 - rpi5 support - fdinfo + gputop support - uapi for CPU job handling virtio-gpu: - add context debug name" * tag 'drm-next-2024-01-10' of git://anongit.freedesktop.org/drm/drm: (2340 commits) drm/amd/display: Allow z8/z10 from driver drm/amd/display: fix bandwidth validation failure on DCN 2.1 drm/amdgpu: apply the RV2 system aperture fix to RN/CZN as well drm/amd/display: Move fixpt_from_s3132 to amdgpu_dm drm/amd/display: Fix recent checkpatch errors in amdgpu_dm Revert "drm/amdkfd: Relocate TBA/TMA to opposite side of VM hole" drm/amd/display: avoid stringop-overflow warnings for dp_decide_lane_settings() drm/amd/display: Fix power_helpers.c codestyle drm/amd/display: Fix hdcp_log.h codestyle drm/amd/display: Fix hdcp2_execution.c codestyle drm/amd/display: Fix hdcp_psp.h codestyle drm/amd/display: Fix freesync.c codestyle drm/amd/display: Fix hdcp_psp.c codestyle drm/amd/display: Fix hdcp1_execution.c codestyle drm/amd/pm/smu7: fix a memleak in smu7_hwmgr_backend_init drm/amdkfd: Fix iterator used outside loop in 'kfd_add_peer_prop()' drm/amdgpu: Drop 'fence' check in 'to_amdgpu_amdkfd_fence()' drm/amdkfd: Confirm list is non-empty before utilizing list_first_entry in kfd_topology.c drm/amdgpu: Fix '*fw' from request_firmware() not released in 'amdgpu_ucode_request()' drm/amdgpu: Fix variable 'mca_funcs' dereferenced before NULL check in 'amdgpu_mca_smu_get_mca_entry()' ...
| * accel/habanalabs: fix information leak in sec_attest_info()Xingyuan Mo2023-12-191-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | This function may copy the pad0 field of struct hl_info_sec_attest to user mode which has not been initialized, resulting in leakage of kernel heap data to user mode. To prevent this, use kzalloc() to allocate and zero out the buffer, which can also eliminate other uninitialized holes, if any. Fixes: 0c88760f8f5e ("habanalabs/gaudi2: add secured attestation info uapi") Signed-off-by: Xingyuan Mo <hdthky0@gmail.com> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs/gaudi2: avoid overriding existing undefined opcode dataTomer Tayar2023-12-191-21/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Part of the undefined opcode data is updated in gaudi2_handle_qman_err_generic() and some in handle_lower_qman_data_on_err(). However, the 'write_enable' flag is checked only in gaudi2_handle_qman_err_generic(), and information of more than a single error can be mixed there. Moreover, handle_lower_qman_data_on_err() is called only for the lower QMAN, so for an error in the upper QMAN there is only a partial info. Move all the data update to be done in a single place, protected by the 'write_enable' flag. As mainly the lower QMAN's info is interesting, avoid saving the partial info for the upper QMAN. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: add parent_device sysfs attributeTomer Tayar2023-12-192-0/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The device debugfs directory was modified to be named as the device-name. This name is the parent device name, i.e. either the PCI address in case of an ASIC, or the simulator device name in case of a simulator. This change makes it more difficult for a user to access the debugfs directory for a specific accel device, because he can't just use the accel minor id, but he needs to do more device-dependent operations to get the device name. To make it easier to get this name, add a 'parent_device' sysfs attribute that the user can read using the minor id before accessing debugfs. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs/gaudi2: add zero padding when printing QM CP instructionTomer Tayar2023-12-191-1/+1
| | | | | | | | | | | | | | | | | | | | | | QM instructions are in multiples of 64 bits and the command type is in the upper bits of first QWORD. To make it clearer that an undefined command is due to a type of 0x0, always print all 64 bits and add a zero padding if needed. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: report 3 instances of Infineon second stageAriel Suller2023-12-191-2/+18
| | | | | | | | | | | | | | | | | | Infineon controller second stage has 3 instances that their version need to be reported by driver. Signed-off-by: Ariel Suller <asuller@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs/gaudi2: add signed dev info uAPIMoti Haimovski2023-12-193-0/+63
| | | | | | | | | | | | | | | | | | User will provide a nonce via the INFO ioctl, and will retrieve the signed device info generated using given nonce. Signed-off-by: Moti Haimovski <mhaimovski@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs/gaudi2: use correct registers to dump QM CQ infoTomer Tayar2023-12-192-12/+12
| | | | | | | | | | | | | | | | | | | | | | The QM CQ PTR_LO/PTR_HI/TSIZE registers are for pushing a CQ entry, and although they are updated by HW even when descriptors are fetched by PQ and CB addresses are fed into CQ, the correct registers to use when dumping the CQ info are the ones with the _STS suffix. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: expose module id through sysfsDani Liberman2023-12-191-0/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Module ID exposes the physical location of the device in the server, from the pov of the devices in regard to how they are connected by internal fabric. This information is already exposed in our INFO ioctl, but there are utilities and scripts running in data-center which are already accessing sysfs for topology information and it is easier for them to continue getting that information from sysfs instead of opening a file descriptor. Signed-off-by: Dani Liberman <dliberman@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: print error code when mapping failsDani Liberman2023-12-191-3/+4
| | | | | | | | | | | | | | | | | | Failure to map is considered a non-trivial error and we need to notify the user about it. Signed-off-by: Dani Liberman <dliberman@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs/gaudi2: get the correct QM CQ info upon an errorTomer Tayar2023-12-192-22/+23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Upon a QM error, the address/size from both the CQ and the ARC_CQ are printed, although the instruction that led to the error was received from only one of them. Moreover, in case of a QM undefined opcode, only one of these address/size sets will be captured based on the value of ARC_CQ_PTR. However, this value can be non-zero even if currently the CQ is used, in case the CQ/ARC_CQ are alternately used. Under the assumption of having a stop-on-error configuration, modify to use CP_STS.CUR_CQ field to get the relevant CQ for the QM error. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: set hard reset flag if graceful reset is skippedTomer Tayar2023-12-191-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | hl_device_cond_reset() might be called with the hard reset flag unset, because a compute reset upon device release as part of a graceful reset is valid. If the conditions for graceful reset are not met, hl_device_reset() will be called for an immediate reset. In this case a compute reset is not valid, so it will be replaced with a hard reset together with a debug message about it. This message might be confusing, as it implies that a compute reset was requested when it shouldn't. To prevent this confusion, set the hard reset flag in hl_device_cond_reset() if going to an immediate reset. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: remove 'get temperature' debug printOfir Bitton2023-12-191-4/+0
| | | | | | | | | | | | | | | | | | The print was added long back for a specific debug and can now be removed. Signed-off-by: Ofir Bitton <obitton@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs/gaudi2: fix undef opcode reportingDafna Hirschfeld2023-12-191-8/+6
| | | | | | | | | | | | | | | | | | | | currently the undefined opcode event bit in set only for lower cp and only if 'write_enable' is true. It should be set anyway and for all streams in order to report that event to userspace. Signed-off-by: Dafna Hirschfeld <dhirschfeld@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: fix EQ heartbeat mechanismFarah Kassabri2023-12-191-7/+7
| | | | | | | | | | | | | | | | | | Stop rescheduling another heartbeat check when EQ heartbeat check fails as it generates confusing logs in dmesg that the heartbeat fails. Signed-off-by: Farah Kassabri <fkassabri@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: add support for Gaudi2C deviceOded Gabbay2023-12-196-0/+13
| | | | | | | | | | | | | | Gaudi2 with PCI revision ID with the value of '3' represents Gaudi2C device and should be detected and initialized as Gaudi2. Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: add log when eq event is not receivedFarah Kassabri2023-12-191-2/+4
| | | | | | | | | | | | | | | | | | | | | | | | Add error log when no eq event is received from FW, to cover a scenario when FW is stuck for some reason. In such case driver will not receive neither the eq error interrupt or the eq heartbeat event, and will just initiate a reset without indication in the dmesg about the reason. Signed-off-by: Farah Kassabri <fkassabri@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs/gaudi2: assume hard-reset by FW upon PCIe AXI drainTomer Tayar2023-12-192-0/+10
| | | | | | | | | | | | | | | | | | | | | | | | When a PCIe AXI drain event happens, it is possible that the driver cannot access the device through PCIe, and therefore cannot send a hard-reset request to FW. Starting from FW version 1.13, FW will initiate a hard-reset in such a case without waiting for a reset request from the driver. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: update device boot error checkFarah Kassabri2023-12-191-83/+32
| | | | | | | | | | | | | | | | | | | | Use a predefined mask which set the device critical boot errors. Driver will fail and stop its loading, only upon detecting at least one of those errors defined in this mask. Signed-off-by: Farah Kassabri <fkassabri@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
| * accel/habanalabs: add pcie reset prepare/done hooksfarah kassabri2023-12-191-0/+34
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When working on a bare-metal system, if FLR will happen the firmware will handle it and driver will have no knowledge of it, and this will cause two issues: 1.The driver will be in operational state while it should be in reset. This will cause the heartbeat mechanism to keep sending messages to FW while pci device is in reset. Eventually heartbeat will fail and the device will end up in non-operational state. 2. After FW handles the FLR, and due to the reset it'll go back to preboot stage, and driver need to perform hard reset in order to load the boot fit binary. This patch will add reset_prepare hook that will set the device to be in disabled state, so it'll be not operational, and also reset_done hook which will be called after the actual FLR handling, then it will perform hard reset. Signed-off-by: farah kassabri <fkassabri@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* | eventfd: simplify eventfd_signal()Christian Brauner2023-11-281-1/+1
|/ | | | | | | | | | | | | | | Ever since the eventfd type was introduced back in 2007 in commit e1ad7468c77d ("signal/timer/event: eventfd core") the eventfd_signal() function only ever passed 1 as a value for @n. There's no point in keeping that additional argument. Link: https://lore.kernel.org/r/20231122-vfs-eventfd-signal-v2-2-bd549b14ce0c@kernel.org Acked-by: Xu Yilun <yilun.xu@intel.com> Acked-by: Andrew Donnellan <ajd@linux.ibm.com> # ocxl Acked-by: Eric Farman <farman@linux.ibm.com> # s390 Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Christian Brauner <brauner@kernel.org>
* accel/habanalabs/gaudi2: fix spmu mask creationOded Gabbay2023-10-091-1/+2
| | | | | | | | | | | | event_types_num received from the user can be 0. In that case, the event_mask should be 0. In addition, to create a correct mask we need to match the number of event types to the bit location such that bit 0 represents a single event type, bit 1 represents 2 types and so on. Signed-off-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Ofir Bitton <obitton@habana.ai>
* accel/habanalabs/gaudi2: perform hard-reset upon PCIe AXI drain eventTomer Tayar2023-10-091-1/+1
| | | | | | | | | | | | Non-completed transactions from PCIe towards the device are handled by the AXI drain mechanism. This handling is in the PCIe level, but the transactions are still there in the device consuming some queues entries, and therefore the device must be reset. Modify to perform hard-reset upon PCIe AXI drain events. Signed-off-by: Tomer Tayar <ttayar@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: fix bug in decoder wait for cs completionfarah kassabri2023-10-091-7/+7
| | | | | | | | | | | | The decoder interrupts are handled in the interrupt context same as all user interrupts. In such case, the wait list should be protected by spin_lock_irqsave in order to avoid deadlock that might happen with the user submission flow. Signed-off-by: farah kassabri <fkassabri@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: remove wrong doc for init_phys_pg_pack_from_userptrDafna Hirschfeld2023-10-091-1/+0
| | | | | | | | The function does not pin the pages so remove that from the inline doc. Signed-off-by: Dafna Hirschfeld <dhirschfeld@habana.ai> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: add missing debugfs function stubsArnd Bergmann2023-10-091-0/+9
| | | | | | | | | | | | | | | | Two function stubs were removed in an earlier commit but are now needed again: drivers/accel/habanalabs/common/device.c: In function 'hl_device_init': drivers/accel/habanalabs/common/device.c:2231:14: error: implicit declaration of function 'hl_debugfs_device_init'; did you mean 'drm_debugfs_dev_init'? [-Werror=implicit-function-declaration] 2231 | rc = hl_debugfs_device_init(hdev); drivers/accel/habanalabs/common/device.c:2367:9: error: implicit declaration of function 'hl_debugfs_device_fini'; did you mean 'hl_debugfs_remove_file'? [-Werror=implicit-function-declaration] 2367 | hl_debugfs_device_fini(hdev); | ^~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Oded Gabbay <ogabbay@kernel.org> Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
* accel/habanalabs: minor cosmetic update to habanalabs.hOded Gabbay2023-10-091-16/+15
| | | | | | | | | - Update copyright years - Align fields in struct hl_userptr - Fix comments Signed-off-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Ofir Bitton <obitton@habana.ai>
* accel/habanalabs/gaudi: remove define used for simulatorOded Gabbay2023-10-092-4/+1
| | | | | | | We don't support simulator in upstream. Signed-off-by: Oded Gabbay <ogabbay@kernel.org> Reviewed-by: Ofir Bitton <obitton@habana.ai>
generated by cgit v1.2.3 (git 2.25.1) at 2024-08-15 07:38:13 +0000