linux-stable.git/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c, branch linux-5.9.y

drm/amdkfd: fix a memory leak issue

2020-09-15T22:16:04+00:00

In the resume stage of GPU recovery, start_cpsch will call pm_init
which set pm->allocated as false, cause the next pm_release_ib has
no chance to release ib memory.

Add pm_release_ib in stop_cpsch which will be called in the suspend
stage of GPU recovery.

Reviewed-by: Felix Kuehling 
Signed-off-by: Dennis Li 
Signed-off-by: Alex Deucher

drm/kfd: fix a system crash issue during GPU recovery

2020-09-15T22:15:28+00:00

The crash log as the below:

[Thu Aug 20 23:18:14 2020] general protection fault: 0000 [#1] SMP NOPTI
[Thu Aug 20 23:18:14 2020] CPU: 152 PID: 1837 Comm: kworker/152:1 Tainted: G           OE     5.4.0-42-generic #46~18.04.1-Ubuntu
[Thu Aug 20 23:18:14 2020] Hardware name: GIGABYTE G482-Z53-YF/MZ52-G40-00, BIOS R12 05/13/2020
[Thu Aug 20 23:18:14 2020] Workqueue: events amdgpu_ras_do_recovery [amdgpu]
[Thu Aug 20 23:18:14 2020] RIP: 0010:evict_process_queues_cpsch+0xc9/0x130 [amdgpu]
[Thu Aug 20 23:18:14 2020] Code: 49 8d 4d 10 48 39 c8 75 21 eb 44 83 fa 03 74 36 80 78 72 00 74 0c 83 ab 68 01 00 00 01 41 c6 45 41 00 48 8b 00 48 39 c8 74 25 <80> 78 70 00 c6 40 6d 01 74 ee 8b 50 28 c6 40 70 00 83 ab 60 01 00
[Thu Aug 20 23:18:14 2020] RSP: 0018:ffffb29b52f6fc90 EFLAGS: 00010213
[Thu Aug 20 23:18:14 2020] RAX: 1c884edb0a118914 RBX: ffff8a0d45ff3c00 RCX: ffff8a2d83e41038
[Thu Aug 20 23:18:14 2020] RDX: 0000000000000000 RSI: 0000000000000082 RDI: ffff8a0e2e4178c0
[Thu Aug 20 23:18:14 2020] RBP: ffffb29b52f6fcb0 R08: 0000000000001b64 R09: 0000000000000004
[Thu Aug 20 23:18:14 2020] R10: ffffb29b52f6fb78 R11: 0000000000000001 R12: ffff8a0d45ff3d28
[Thu Aug 20 23:18:14 2020] R13: ffff8a2d83e41028 R14: 0000000000000000 R15: 0000000000000000
[Thu Aug 20 23:18:14 2020] FS:  0000000000000000(0000) GS:ffff8a0e2e400000(0000) knlGS:0000000000000000
[Thu Aug 20 23:18:14 2020] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[Thu Aug 20 23:18:14 2020] CR2: 000055c783c0e6a8 CR3: 00000034a1284000 CR4: 0000000000340ee0
[Thu Aug 20 23:18:14 2020] Call Trace:
[Thu Aug 20 23:18:14 2020]  kfd_process_evict_queues+0x43/0xd0 [amdgpu]
[Thu Aug 20 23:18:14 2020]  kfd_suspend_all_processes+0x60/0xf0 [amdgpu]
[Thu Aug 20 23:18:14 2020]  kgd2kfd_suspend.part.7+0x43/0x50 [amdgpu]
[Thu Aug 20 23:18:14 2020]  kgd2kfd_pre_reset+0x46/0x60 [amdgpu]
[Thu Aug 20 23:18:14 2020]  amdgpu_amdkfd_pre_reset+0x1a/0x20 [amdgpu]
[Thu Aug 20 23:18:14 2020]  amdgpu_device_gpu_recover+0x377/0xf90 [amdgpu]
[Thu Aug 20 23:18:14 2020]  ? amdgpu_ras_error_query+0x1b8/0x2a0 [amdgpu]
[Thu Aug 20 23:18:14 2020]  amdgpu_ras_do_recovery+0x159/0x190 [amdgpu]
[Thu Aug 20 23:18:14 2020]  process_one_work+0x20f/0x400
[Thu Aug 20 23:18:14 2020]  worker_thread+0x34/0x410

When GPU hang, user process will fail to create a compute queue whose
struct object will be freed later, but driver wrongly add this queue to
queue list of the proccess. And then kfd_process_evict_queues will
access a freed memory, which cause a system crash.

v2:
The failure to execute_queues should probably not be reported to
the caller of create_queue, because the queue was already created.
Therefore change to ignore the return value from execute_queues.

Reviewed-by: Felix Kuehling 
Signed-off-by: Dennis Li 
Signed-off-by: Alex Deucher 
Cc: stable@vger.kernel.org

drm/amdkfd: Support navy_flounder KFD

2020-07-15T16:46:55+00:00

Add KFD support for Navy Flounder.

Signed-off-by: Chengming Gui 
Reviewed-by: Tao Zhou 
Signed-off-by: Alex Deucher

drm/amdkfd: Fix circular locking dependency warning

2020-07-01T05:59:27+00:00

[  150.887733] ======================================================
[  150.893903] WARNING: possible circular locking dependency detected
[  150.905917] ------------------------------------------------------
[  150.912129] kfdtest/4081 is trying to acquire lock:
[  150.917002] ffff8f7f3762e118 (&mm->mmap_sem#2){++++}, at:
                                 __might_fault+0x3e/0x90
[  150.924490]
               but task is already holding lock:
[  150.930320] ffff8f7f49d229e8 (&dqm->lock_hidden){+.+.}, at:
                                destroy_queue_cpsch+0x29/0x210 [amdgpu]
[  150.939432]
               which lock already depends on the new lock.

[  150.947603]
               the existing dependency chain (in reverse order) is:
[  150.955074]
               -> #3 (&dqm->lock_hidden){+.+.}:
[  150.960822]        __mutex_lock+0xa1/0x9f0
[  150.964996]        evict_process_queues_cpsch+0x22/0x120 [amdgpu]
[  150.971155]        kfd_process_evict_queues+0x3b/0xc0 [amdgpu]
[  150.977054]        kgd2kfd_quiesce_mm+0x25/0x60 [amdgpu]
[  150.982442]        amdgpu_amdkfd_evict_userptr+0x35/0x70 [amdgpu]
[  150.988615]        amdgpu_mn_invalidate_hsa+0x41/0x60 [amdgpu]
[  150.994448]        __mmu_notifier_invalidate_range_start+0xa4/0x240
[  151.000714]        copy_page_range+0xd70/0xd80
[  151.005159]        dup_mm+0x3ca/0x550
[  151.008816]        copy_process+0x1bdc/0x1c70
[  151.013183]        _do_fork+0x76/0x6c0
[  151.016929]        __x64_sys_clone+0x8c/0xb0
[  151.021201]        do_syscall_64+0x4a/0x1d0
[  151.025404]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  151.030977]
               -> #2 (&adev->notifier_lock){+.+.}:
[  151.036993]        __mutex_lock+0xa1/0x9f0
[  151.041168]        amdgpu_mn_invalidate_hsa+0x30/0x60 [amdgpu]
[  151.047019]        __mmu_notifier_invalidate_range_start+0xa4/0x240
[  151.053277]        copy_page_range+0xd70/0xd80
[  151.057722]        dup_mm+0x3ca/0x550
[  151.061388]        copy_process+0x1bdc/0x1c70
[  151.065748]        _do_fork+0x76/0x6c0
[  151.069499]        __x64_sys_clone+0x8c/0xb0
[  151.073765]        do_syscall_64+0x4a/0x1d0
[  151.077952]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  151.083523]
               -> #1 (mmu_notifier_invalidate_range_start){+.+.}:
[  151.090833]        change_protection+0x802/0xab0
[  151.095448]        mprotect_fixup+0x187/0x2d0
[  151.099801]        setup_arg_pages+0x124/0x250
[  151.104251]        load_elf_binary+0x3a4/0x1464
[  151.108781]        search_binary_handler+0x6c/0x210
[  151.113656]        __do_execve_file.isra.40+0x7f7/0xa50
[  151.118875]        do_execve+0x21/0x30
[  151.122632]        call_usermodehelper_exec_async+0x17e/0x190
[  151.128393]        ret_from_fork+0x24/0x30
[  151.132489]
               -> #0 (&mm->mmap_sem#2){++++}:
[  151.138064]        __lock_acquire+0x11a1/0x1490
[  151.142597]        lock_acquire+0x90/0x180
[  151.146694]        __might_fault+0x68/0x90
[  151.150879]        read_sdma_queue_counter+0x5f/0xb0 [amdgpu]
[  151.156693]        update_sdma_queue_past_activity_stats+0x3b/0x90 [amdgpu]
[  151.163725]        destroy_queue_cpsch+0x1ae/0x210 [amdgpu]
[  151.169373]        pqm_destroy_queue+0xf0/0x250 [amdgpu]
[  151.174762]        kfd_ioctl_destroy_queue+0x32/0x70 [amdgpu]
[  151.180577]        kfd_ioctl+0x223/0x400 [amdgpu]
[  151.185284]        ksys_ioctl+0x8f/0xb0
[  151.189118]        __x64_sys_ioctl+0x16/0x20
[  151.193389]        do_syscall_64+0x4a/0x1d0
[  151.197569]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  151.203141]
               other info that might help us debug this:

[  151.211140] Chain exists of:
                 &mm->mmap_sem#2 --> &adev->notifier_lock --> &dqm->lock_hidden

[  151.222535]  Possible unsafe locking scenario:

[  151.228447]        CPU0                    CPU1
[  151.232971]        ----                    ----
[  151.237502]   lock(&dqm->lock_hidden);
[  151.241254]                                lock(&adev->notifier_lock);
[  151.247774]                                lock(&dqm->lock_hidden);
[  151.254038]   lock(&mm->mmap_sem#2);

This commit fixes the warning by ensuring get_user() is not called
while reading SDMA stats with dqm_lock held as get_user() could cause a
page fault which leads to the circular locking scenario.

Signed-off-by: Mukul Joshi 
Reviewed-by: Felix Kuehling 
Signed-off-by: Alex Deucher

drm/amdkfd: label internally used symbols as static

2020-07-01T05:59:23+00:00

Used sparse(make C=1) to find these loose ends.

v2:
removed unwanted extra line

Signed-off-by: Nirmoy Das 
Reviewed-by: Alex Deucher 
Signed-off-by: Alex Deucher

drm/amdkfd: Support Sienna_Cichlid KFD v4

2020-07-01T05:59:11+00:00

v4: drop get_tile_config, comment out other callbacks

Signed-off-by: Yong Zhao 
Signed-off-by: Alex Deucher

drm/amdkfd: Track SDMA utilization per process

2020-05-28T18:00:50+00:00

Track SDMA usage on a per process basis and report it through sysfs.
The value in the sysfs file indicates the amount of time SDMA has
been in-use by this process since the creation of the process.
This value is in microsecond granularity.

Signed-off-by: Mukul Joshi 
Reviewed-by: Felix Kuehling 
Signed-off-by: Alex Deucher

drm/amdkfd: Use a systematic method to calculate queue mask bit

2020-05-01T19:19:08+00:00

The queue mask used for set_resources always assumes the queue number
per pipe is 8, so KFD needs to align with that by using function
amdgpu_queue_mask_bit_to_set_resource_bit().

Signed-off-by: Yong Zhao 
Signed-off-by: Alex Deucher

drm/amdkfd: Enable over-subscription with >1 GWS queue

2020-04-28T20:20:30+00:00

The current GWS usage model will only allows a single GWS-enabled
process to be active on the GPU at once. This ensures that a
barrier-using kernel gets a known amount of GPU hardware, to
prevent deadlock due to inability to go beyond the GWS barrier.

The HWS watches how many GWS entries are assigned to each process,
and goes into over-subscription mode when two processes need more
than the 64 that are available. The current KFD method for working
with this is to allocate all 64 GWS entries to each GWS-capable
process.

When more than one GWS-enabled process is in the runlist, we must
make sure the runlist is in over-subscription mode, so that the
HWS gets a chained RUN_LIST packet and continues scheduling
kernels.

Signed-off-by: Joseph Greathouse 
Reviewed-by: Felix Kuehling 
Signed-off-by: Felix Kuehling 
Signed-off-by: Alex Deucher

drm/amdkfd: change SDMA MQD memory type

2020-02-28T21:59:20+00:00

SDMA MQD memory type is NC that causes MQD data overwritten
accidentally by an old stable cache line. Changing it to UC
default for GART will fix the issue.

The mqd_gfx9 parameter is meant for control stacks that are
allocated together with user mode queue MQDs. Setting
mqd_gfx9 to true maps the control stack pages as NC.
Here it was accidentally applied to SDMA MQDs,
which are allocated together with the HIQ MQD. Setting
the mqd_gfx9 to false avoids that.

Signed-off-by: Eric Huang 
Acked-by: Yong Zhao 
Reviewed-by: Felix Kuehling 
Signed-off-by: Alex Deucher