Bug 205309 - kmemleak reports various leaks in amdgpu
Summary: kmemleak reports various leaks in amdgpu
Status: RESOLVED MOVED
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel) (show other bugs)
Hardware: x86-64 Linux
: P1 normal
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2019-10-24 19:44 UTC by Erhard F.
Modified: 2020-02-18 15:48 UTC (History)
1 user (show)

See Also:
Kernel Version: 5.4-rc4
Subsystem:
Regression: No
Bisected commit-id:


Attachments
kmemleak: 34 new suspected memory leaks (31.79 KB, text/plain)
2019-10-24 19:44 UTC, Erhard F.
Details
dmesg (kernel 5.4-rc4) (67.79 KB, text/plain)
2019-10-24 19:45 UTC, Erhard F.
Details
kernel .config (kernel 5.4-rc4) (99.57 KB, text/plain)
2019-10-24 19:46 UTC, Erhard F.
Details

Description Erhard F. 2019-10-24 19:44:30 UTC
Created attachment 285635 [details]
kmemleak: 34 new suspected memory leaks

Running X I get various kmemleak dumps for amdgpu:
[...]
unreferenced object 0xffff995f56fde340 (size 224):
  comm "Xorg", pid 801, jiffies 4294890593 (age 1572.594s)
  hex dump (first 32 bytes):
    d0 e3 fd 56 5f 99 ff ff a0 61 7e c0 ff ff ff ff  ...V_....a~.....
    36 bb 4b 81 0a 00 00 00 48 fe 32 53 5f 99 ff ff  6.K.....H.2S_...
  backtrace:
    [<00000000531b3f80>] drm_sched_fence_create+0x1d/0xc0 [gpu_sched]
    [<00000000dc85bf8c>] drm_sched_job_init+0x5b/0xa0 [gpu_sched]
    [<000000005789f820>] amdgpu_job_submit+0x1e/0x90 [amdgpu]
    [<00000000f3185640>] amdgpu_copy_buffer+0x1ab/0x290 [amdgpu]
    [<0000000053449562>] amdgpu_ttm_copy_mem_to_mem+0x25d/0x520 [amdgpu]
    [<00000000810db067>] amdgpu_move_blit.constprop.25+0x85/0x1b0 [amdgpu]
    [<0000000051e06c3a>] amdgpu_bo_move+0x164/0x290 [amdgpu]
    [<00000000d0345de4>] ttm_bo_handle_move_mem+0x113/0x5b0 [ttm]
    [<000000003dbb02d9>] ttm_bo_validate+0x16e/0x1b0 [ttm]
    [<0000000096a41e6b>] amdgpu_bo_fault_reserve_notify+0xe5/0x130 [amdgpu]
    [<00000000eb688629>] ttm_bo_vm_fault+0xac/0x650 [ttm]
    [<0000000061352dd3>] __do_fault+0x31/0xf0
    [<00000000ad4b517c>] __handle_mm_fault+0xd31/0x12a0
    [<00000000037720cd>] handle_mm_fault+0x165/0x2e0
    [<00000000196e733c>] __do_page_fault+0x1a8/0x420
    [<000000000c9cfffa>] page_fault+0x31/0x40
[...]

System:    Kernel: 5.4.0-rc4 x86_64 bits: 64 Desktop: MATE 1.22.0 Distro: Gentoo Base System release 2.6 
Machine:   Type: Server Mobo: Supermicro model: H8SGL v: 1234567890 serial: <filter> BIOS: American Megatrends v: 3.5b 
           date: 03/18/2016 
CPU:       8-Core: AMD Opteron 6380 type: MT MCP speed: 1584 MHz min/max: 1400/2500 MHz 
Graphics:  Device-1: Advanced Micro Devices [AMD/ATI] Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] driver: amdgpu 
           v: kernel 
           Device-2: Matrox Systems MGA G200eW WPCM450 driver: N/A 
           Display: x11 server: X.Org 1.20.5 driver: amdgpu,ati unloaded: modesetting,radeon resolution: 1920x1080~60Hz 
           OpenGL: renderer: Radeon RX 590 Series (POLARIS10 DRM 3.35.0 5.4.0-rc4 LLVM 8.0.1) v: 4.5 Mesa 19.1.7
Comment 1 Erhard F. 2019-10-24 19:45:19 UTC
Created attachment 285637 [details]
dmesg (kernel 5.4-rc4)
Comment 2 Erhard F. 2019-10-24 19:46:13 UTC
Created attachment 285639 [details]
kernel .config (kernel 5.4-rc4)
Comment 3 Erhard F. 2019-11-16 11:03:30 UTC
Does not happen with 5.4-rc7. Seems fixed now.
Comment 4 Erhard F. 2019-11-16 11:31:08 UTC
Ok, closing this bug was a bit too early... Just got some leaks in.

Moved to https://bugs.freedesktop.org/show_bug.cgi?id=112302
(as I can select component DRM/AMDgpu here which seems more appropriate)

Note You need to log in before you can comment on or make changes to this bug.