forked from Minki/linux
e274086e47
This is to fix a memory leak issue caused by unfreed gvtg workload objects. Walk through the workload list and free all of the remained workloads before destroying kmem cache. [179.885211] INFO: Object 0xffff9cef10003b80 @offset=7040 [179.885657] kmem_cache_destroy gvt-g_vgpu_workload: Slab cache still has objects [179.886146] CPU: 2 PID: 2318 Comm: win_lucas Tainted: G B W 4.11.0+ #1 [179.887223] Call Trace: [179.887394] dump_stack+0x63/0x90 [179.887617] kmem_cache_destroy+0x1cf/0x1e0 [179.887960] intel_vgpu_clean_execlist+0x15/0x20 [i915] [179.888365] intel_gvt_destroy_vgpu+0x4c/0xd0 [i915] [179.888688] intel_vgpu_remove+0x2a/0x30 [kvmgt] [179.888988] mdev_device_remove_ops+0x23/0x50 [mdev] [179.889309] mdev_device_remove+0xe4/0x190 [mdev] [179.889615] remove_store+0x7d/0xb0 [mdev] [179.889885] dev_attr_store+0x18/0x30 [179.890129] sysfs_kf_write+0x37/0x40 [179.890371] kernfs_fop_write+0x107/0x180 [179.890632] __vfs_write+0x37/0x160 [179.890865] ? kmem_cache_alloc+0xd7/0x1b0 [179.891116] ? apparmor_file_permission+0x1a/0x20 [179.891372] ? security_file_permission+0x3b/0xc0 [179.891628] vfs_write+0xb8/0x1b0 [179.891812] SyS_write+0x55/0xc0 [179.891992] entry_SYSCALL_64_fastpath+0x1e/0xad Signed-off-by: Changbin Du <changbin.du@intel.com> Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com> |
||
|---|---|---|
| .. | ||
| aperture_gm.c | ||
| cfg_space.c | ||
| cmd_parser.c | ||
| cmd_parser.h | ||
| debug.h | ||
| display.c | ||
| display.h | ||
| edid.c | ||
| edid.h | ||
| execlist.c | ||
| execlist.h | ||
| firmware.c | ||
| gtt.c | ||
| gtt.h | ||
| gvt.c | ||
| gvt.h | ||
| handlers.c | ||
| hypercall.h | ||
| interrupt.c | ||
| interrupt.h | ||
| kvmgt.c | ||
| Makefile | ||
| mmio.c | ||
| mmio.h | ||
| mpt.h | ||
| opregion.c | ||
| reg.h | ||
| render.c | ||
| render.h | ||
| sched_policy.c | ||
| sched_policy.h | ||
| scheduler.c | ||
| scheduler.h | ||
| trace_points.c | ||
| trace.h | ||
| vgpu.c | ||