Re: [PATCH v1 2/2] drm/amdgpu: adding device coredump support

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



(snip)
#ifdef CONFIG_DEV_COREDUMP
                                tmp_adev->reset_context_vram_lost = vram_lost;
tmp_adev->reset_context_task_info.pid = 0;
                                if (reset_context->job && reset_context->job->vm)
tmp_adev->reset_context_task_info =
reset_context->job->vm->task_info;
amdgpu_reset_capture_coredumpm(tmp_adev);
#endif
tmp_adev->reset_context_task_info refers to reset_context->job->vm->task_info, i am not setting PID and process name separately, instead i am having the reference to the reset_context->job->vm->task_info on valid VM.


I think you are not getting my point. struct amdgpu_task_info has 3 more parameters: process_name, task_name and tgid.

When VRAM is not lost: we are getting the whole new amdgpu_task_info structure, and all parameters are valid.

When VRAM is lost: we should reset all parameters of amdgpu_task_info structure, not just pid. Else other params may contain garbage.

So what I mean is:
instead of
tmp_adev->reset_context_task_info.pid = 0;

do something like:
memset(&tmp_adev->reset_context_task_info.pid, 0, sizeof(reset_context_task_info));

- Shashank



[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux