kernel.git/drivers/gpu/drm/imagination, branch linux-6.15.y

drm/imagination: Clear runtime PM errors while resetting the GPU

2025-08-20T16:36:08Z

[ Upstream commit 551507e0d0bf32ce1d7d27533c4b98307380804c ] The runtime PM might be left in error state if one of the callbacks returned an error, e.g. if the (auto)suspend callback failed following a firmware crash. When that happens, any further attempt to acquire or release a power reference will then also fail, making it impossible to do anything else with the GPU. The driver logic will eventually reach the reset code. In pvr_power_reset(), replace pvr_power_get() with a new API pvr_power_get_clear() which also attempts to clear any runtime PM error state if acquiring a power reference is not possible. Signed-off-by: Alessio Belle Reviewed-by: Matt Coster Link: https://lore.kernel.org/r/20250624-clear-rpm-errors-gpu-reset-v1-1-b8ff2ae55aac@imgtec.com Signed-off-by: Matt Coster Signed-off-by: Sasha Levin

drm/imagination: Fix kernel crash when hard resetting the GPU

2025-07-17T16:43:48Z

commit d38376b3ee48d073c64e75e150510d7e6b4b04f7 upstream. The GPU hard reset sequence calls pm_runtime_force_suspend() and pm_runtime_force_resume(), which according to their documentation should only be used during system-wide PM transitions to sleep states. The main issue though is that depending on some internal runtime PM state as seen by pm_runtime_force_suspend() (whether the usage count is <= 1), pm_runtime_force_resume() might not resume the device unless needed. If that happens, the runtime PM resume callback pvr_power_device_resume() is not called, the GPU clocks are not re-enabled, and the kernel crashes on the next attempt to access GPU registers as part of the power-on sequence. Replace calls to pm_runtime_force_suspend() and pm_runtime_force_resume() with direct calls to the driver's runtime PM callbacks, pvr_power_device_suspend() and pvr_power_device_resume(), to ensure clocks are re-enabled and avoid the kernel crash. Fixes: cc1aeedb98ad ("drm/imagination: Implement firmware infrastructure and META FW support") Signed-off-by: Alessio Belle Reviewed-by: Matt Coster Link: https://lore.kernel.org/r/20250624-fix-kernel-crash-gpu-hard-reset-v1-1-6d24810d72a6@imgtec.com Cc: stable@vger.kernel.org Signed-off-by: Matt Coster Signed-off-by: Greg Kroah-Hartman

Merge drm/drm-fixes into drm-misc-fixes

2025-04-08T08:15:47Z

Backmerging to get updates from v6.15-rc1. Signed-off-by: Thomas Zimmermann

drm/imagination: fix firmware memory leaks

2025-03-25T17:18:15Z

Free the memory used to hold the results of firmware image processing when the module is unloaded. Fix the related issue of the same memory being leaked if processing of the firmware image fails during module load. Ensure all firmware GEM objects are destroyed if firmware image processing fails. Fixes memory leaks on powervr module unload detected by Kmemleak: unreferenced object 0xffff000042e20000 (size 94208): comm "modprobe", pid 470, jiffies 4295277154 hex dump (first 32 bytes): 02 ae 7f ed bf 45 84 00 3c 5b 1f ed 9f 45 45 05 .....E..<[...EE. d5 4f 5d 14 6c 00 3d 23 30 d0 3a 4a 66 0e 48 c8 .O].l.=#0.:Jf.H. backtrace (crc dd329dec): kmemleak_alloc+0x30/0x40 ___kmalloc_large_node+0x140/0x188 __kmalloc_large_node_noprof+0x2c/0x13c __kmalloc_noprof+0x48/0x4c0 pvr_fw_init+0xaa4/0x1f50 [powervr] unreferenced object 0xffff000042d20000 (size 20480): comm "modprobe", pid 470, jiffies 4295277154 hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 09 00 00 00 0b 00 00 00 ................ 00 00 00 00 00 00 00 00 07 00 00 00 08 00 00 00 ................ backtrace (crc 395b02e3): kmemleak_alloc+0x30/0x40 ___kmalloc_large_node+0x140/0x188 __kmalloc_large_node_noprof+0x2c/0x13c __kmalloc_noprof+0x48/0x4c0 pvr_fw_init+0xb0c/0x1f50 [powervr] Cc: stable@vger.kernel.org Fixes: cc1aeedb98ad ("drm/imagination: Implement firmware infrastructure and META FW support") Signed-off-by: Brendan King Reviewed-by: Matt Coster Link: https://lore.kernel.org/r/20250318-ddkopsrc-1339-firmware-related-memory-leak-on-module-unload-v1-1-155337c57bb4@imgtec.com Signed-off-by: Matt Coster

drm/imagination: take paired job reference

2025-03-25T17:17:37Z

For paired jobs, have the fragment job take a reference on the geometry job, so that the geometry job cannot be freed until the fragment job has finished with it. The geometry job structure is accessed when the fragment job is being prepared by the GPU scheduler. Taking the reference prevents the geometry job being freed until the fragment job no longer requires it. Fixes a use after free bug detected by KASAN: [ 124.256386] BUG: KASAN: slab-use-after-free in pvr_queue_prepare_job+0x108/0x868 [powervr] [ 124.264893] Read of size 1 at addr ffff0000084cb960 by task kworker/u16:4/63 Cc: stable@vger.kernel.org Fixes: eaf01ee5ba28 ("drm/imagination: Implement job submission and scheduling") Signed-off-by: Brendan King Reviewed-by: Matt Coster Link: https://lore.kernel.org/r/20250318-ddkopsrc-1337-use-after-free-in-pvr_queue_prepare_job-v1-1-80fb30d044a6@imgtec.com Signed-off-by: Matt Coster

Backmerge tag 'v6.14-rc6' into drm-next

2025-03-11T23:43:12Z

This is a backmerge from Linux 6.14-rc6, needed for the nova PR. Signed-off-by: Dave Airlie

drm/imagination: Fix timestamps in firmware traces

2025-03-03T12:15:55Z

When firmware traces are enabled, the firmware dumps 48-bit timestamps for each trace as two 32-bit values, highest 32 bits (of which only 16 useful) first. The driver was reassembling them the other way round i.e. interpreting the first value in memory as the lowest 32 bits, and the second value as the highest 32 bits (then truncated to 16 bits). Due to this, firmware trace dumps showed very large timestamps even for traces recorded shortly after GPU boot. The timestamps in these dumps would also sometimes jump backwards because of the truncation. Example trace dumped after loading the powervr module and enabling firmware traces, where each line is commented with the timestamp value in hexadecimal to better show both issues: [93540092739584] : Host Sync Partition marker: 1 // 0x551300000000 [28419798597632] : GPU units deinit // 0x19d900000000 [28548647616512] : GPU deinit // 0x19f700000000 Update logic to reassemble the timestamps halves in the correct order. Fixes: cb56cd610866 ("drm/imagination: Add firmware trace to debugfs") Signed-off-by: Alessio Belle Reviewed-by: Matt Coster Link: https://patchwork.freedesktop.org/patch/msgid/20250221-fix-fw-trace-timestamps-v1-1-dba4aeb030ca@imgtec.com Signed-off-by: Matt Coster

drm/imagination: only init job done fences once

2025-03-03T12:10:13Z

Ensure job done fences are only initialised once. This fixes a memory manager not clean warning from drm_mm_takedown on module unload. Cc: stable@vger.kernel.org Fixes: eaf01ee5ba28 ("drm/imagination: Implement job submission and scheduling") Signed-off-by: Brendan King Reviewed-by: Matt Coster Link: https://patchwork.freedesktop.org/patch/msgid/20250226-init-done-fences-once-v2-1-c1b2f556b329@imgtec.com Signed-off-by: Matt Coster

drm/imagination: Hold drm_gem_gpuva lock for unmap

2025-03-03T12:09:00Z

Avoid a warning from drm_gem_gpuva_assert_lock_held in drm_gpuva_unlink. The Imagination driver uses the GEM object reservation lock to protect the gpuva list, but the GEM object was not always known in the code paths that ended up calling drm_gpuva_unlink. When the GEM object isn't known, it is found by calling drm_gpuva_find to lookup the object associated with a given virtual address range, or by calling drm_gpuva_find_first when removing all mappings. Cc: stable@vger.kernel.org Fixes: 4bc736f890ce ("drm/imagination: vm: make use of GPUVM's drm_exec helper") Signed-off-by: Brendan King Reviewed-by: Matt Coster Link: https://patchwork.freedesktop.org/patch/msgid/20250226-hold-drm_gem_gpuva-lock-for-unmap-v2-1-3fdacded227f@imgtec.com Signed-off-by: Matt Coster

drm/imagination: avoid deadlock on fence release

2025-03-03T12:07:56Z

Do scheduler queue fence release processing on a workqueue, rather than in the release function itself. Fixes deadlock issues such as the following: [ 607.400437] ============================================ [ 607.405755] WARNING: possible recursive locking detected [ 607.415500] -------------------------------------------- [ 607.420817] weston:zfq0/24149 is trying to acquire lock: [ 607.426131] ffff000017d041a0 (reservation_ww_class_mutex){+.+.}-{3:3}, at: pvr_gem_object_vunmap+0x40/0xc0 [powervr] [ 607.436728] but task is already holding lock: [ 607.442554] ffff000017d105a0 (reservation_ww_class_mutex){+.+.}-{3:3}, at: dma_buf_ioctl+0x250/0x554 [ 607.451727] other info that might help us debug this: [ 607.458245] Possible unsafe locking scenario: [ 607.464155] CPU0 [ 607.466601] ---- [ 607.469044] lock(reservation_ww_class_mutex); [ 607.473584] lock(reservation_ww_class_mutex); [ 607.478114] *** DEADLOCK *** Cc: stable@vger.kernel.org Fixes: eaf01ee5ba28 ("drm/imagination: Implement job submission and scheduling") Signed-off-by: Brendan King Reviewed-by: Matt Coster Link: https://patchwork.freedesktop.org/patch/msgid/20250226-fence-release-deadlock-v2-1-6fed2fc1fe88@imgtec.com Signed-off-by: Matt Coster