Skip to content

Warning when accessing memory on DG2

Noticed this while working on #183 (closed)

After above issue is fixed, https://gitlab.freedesktop.org/drm/xe/kernel/uploads/8f8c3ce80e5cbfa762d36565801af186/xe_issue_res_cursor.c, triggers the below warning on test exit. If I am understanding this correctly, basically signaling is getting processed after test exit. The easiest way to fix this would be take a memory ref on pushing a job and drop the ref on cleanup but that is going to add quite a bit of overhead. @rodrigovivi - Thoughts?

  1 [  119.022986] ------------[ cut here ]------------
    2 [  119.022993] [IGT] xe_res_cursor: exiting, ret=0
    3 [  119.023117] WARNING: CPU: 4 PID: 0 at drivers/gpu/drm/xe/xe_device.h:96 xe_hw_fence_signaled+0x4a/0x60 [xe]
    4 [  119.023136] Modules linked in: xe fuse snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_hda_intel snd_intel_dspcfg x86_pkg_temp_thermal snd_hda_codec coretemp snd_hwde      p snd_hda_core snd_pcm mei_me mei e1000e crct10dif_pclmul crc32_pclmul ptp ghash_clmulni_intel pps_core i2c_i801 i2c_smbus intel_lpss_pci
    5 [  119.023158] CPU: 4 PID: 0 Comm: swapper/4 Not tainted 6.1.0-rc1-xe+ #3514
    6 [  119.023160] Hardware name: Intel Corporation CoffeeLake Client Platform/CoffeeLake S UDIMM RVP, BIOS CNLSFWR1.R00.X220.B00.2103302221 03/30/2021
    7 [  119.023161] RIP: 0010:xe_hw_fence_signaled+0x4a/0x60 [xe]
    8 [  119.023176] Code: c0 74 27 80 7a 50 00 48 8b 42 48 74 19 8b 08 8b 77 3c b8 01 00 00 00 85 f6 75 06 3b 4a 28 0f 9d c0 c3 cc cc cc cc 8b 08 eb e5 <0f> 0b eb d5 0f 0b 31 d2 eb b9 66 9      0 66 2e 0f 1f 84 00 00 00 00 00
    9 [  119.023178] RSP: 0018:ffffc900001f4f60 EFLAGS: 00010046
   10 [  119.023386] RAX: ffff888139160000 RBX: ffff8884486c8198 RCX: 0000000000000000
   11 [  119.023387] RDX: ffff8884486c8140 RSI: ffffffff82a71eec RDI: ffff8884486c8140
   12 [  119.023389] RBP: ffff888139164e98 R08: 0000000000000000 R09: 00000000fffffffe
   13 [  119.023390] R10: 0000000000000001 R11: 00000000fe71467e R12: ffffffff830fcb38
   14 [  119.023391] R13: ffff8884486c8140 R14: ffff888139164e38 R15: ffff8884486c8240
   15 [  119.023392] FS:  0000000000000000(0000) GS:ffff88845dc00000(0000) knlGS:0000000000000000
   16 [  119.023394] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
   17 [  119.023395] CR2: 00007f03c06230e0 CR3: 0000000006e12003 CR4: 00000000003706e0
   18 [  119.023396] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
   19 [  119.023397] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
   20 [  119.023399] Call Trace:
   21 [  119.023400]  <IRQ>
   22 [  119.023401]  hw_fence_irq_run_cb+0x67/0x2c0 [xe]
   23 [  119.023418]  irq_work_single+0x3c/0x90
   24 [  119.023423]  irq_work_run_list+0x28/0x40
   25 [  119.023425]  irq_work_run+0x26/0x40
   26 [  119.023428]  __sysvec_irq_work+0x3e/0x1c0
   27 [  119.023431]  sysvec_irq_work+0x85/0xb0
   28 [  119.023433]  </IRQ>
   29 [  119.023434]  <TASK>
   30 [  119.023435]  asm_sysvec_irq_work+0x16/0x20
   31 [  119.023438] RIP: 0010:cpuidle_enter_state+0x104/0x5c0