linux/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c, branch v5.9

drm/amdgpu: fix NULL pointer access issue when unloading driver

2020-08-18T21:00:46Z

When unloading driver by "modprobe -r amdgpu", one NULL pointer dereference bug occurs in ras debugfs releasing. The cause is the duplicated debugfs_remove, as drm debugfs_root dir has been cleaned up already by drm_minor_unregister. BUG: kernel NULL pointer dereference, address: 00000000000000a0 PGD 0 P4D 0 Oops: 0002 [#1] SMP PTI CPU: 11 PID: 1526 Comm: modprobe Tainted: G OE 5.6.0-guchchen #1 Hardware name: System manufacturer System Product Name/TUF Z370-PLUS GAMING II, BIOS 0411 09/21/2018 RIP: 0010:down_write+0x15/0x40 Code: eb de e8 7e 17 72 ff cc cc cc cc cc cc cc cc cc cc cc cc cc cc 0f 1f 44 00 00 53 48 89 fb e8 92 d8 ff ff 31 c0 ba 01 00 00 00 48 0f b1 13 75 0f 65 48 8b 04 25 c0 8b 01 00 48 89 43 08 5b c3 RSP: 0018:ffffb1590386fcd0 EFLAGS: 00010246 RAX: 0000000000000000 RBX: 00000000000000a0 RCX: 0000000000000000 RDX: 0000000000000001 RSI: ffffffff85b2fcc2 RDI: 00000000000000a0 RBP: ffffb1590386fd30 R08: ffffffff85b2fcc2 R09: 000000000002b3c0 R10: ffff97a330618c40 R11: 00000000000005f6 R12: ffff97a3481beb40 R13: 00000000000000a0 R14: ffff97a3481beb40 R15: 0000000000000000 FS: 00007fb11a717540(0000) GS:ffff97a376cc0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00000000000000a0 CR3: 00000004066d6006 CR4: 00000000003606e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: simple_recursive_removal+0x63/0x370 ? debugfs_remove+0x60/0x60 debugfs_remove+0x40/0x60 amdgpu_ras_fini+0x82/0x230 [amdgpu] ? __kernfs_remove.part.17+0x101/0x1f0 ? kernfs_name_hash+0x12/0x80 amdgpu_device_fini+0x1c0/0x580 [amdgpu] amdgpu_driver_unload_kms+0x3e/0x70 [amdgpu] amdgpu_pci_remove+0x36/0x60 [amdgpu] pci_device_remove+0x3b/0xb0 device_release_driver_internal+0xe5/0x1c0 driver_detach+0x46/0x90 bus_remove_driver+0x58/0xd0 pci_unregister_driver+0x29/0x90 amdgpu_exit+0x11/0x25 [amdgpu] __x64_sys_delete_module+0x13d/0x210 do_syscall_64+0x5f/0x250 entry_SYSCALL_64_after_hwframe+0x44/0xa9 Signed-off-by: Guchun Chen Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher

drm/amdgpu: add printing after executing page reservation to eeprom

2020-08-06T20:24:11Z

This will tell users if the faulty page has been written to external eeprom device in dmesg log. Signed-off-by: Guchun Chen Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher

drm/amdgpu: RAS emergency restart logic refine

2020-07-15T16:41:47Z

If we are in RAS triggered situation and BACO isn't support, emergency restart is needed, and this code is only needed for some specific cases(vega20 with given smu fw version). After we add smu mode1 reset for sienna cichlid, we need to share AMD_RESET_METHOD_MODE1 with psp mode1 reset, so in amdgpu_device_gpu_recover, we need differentiate which mode1 reset we are using, then decide if it's a full reset and then decide if emergency restart is needed, the logic will become much more complex. After discussion with Hawking, move emergency restart logic to an independent function. Signed-off-by: Likun Gao Signed-off-by: Wenhui Sheng Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher

drm/amdgpu: label internally used symbols as static

2020-07-01T05:59:23Z

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/amdgpu: remove useless code in RAS

2020-06-02T20:47:43Z

Module parameter amdgpu_ras_mask has been involved in the calculation of ras support capability, so drop this redundant code. Signed-off-by: Guchun Chen Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher

drm/amdgpu: fix RAS memory leak in error case

2020-06-02T20:47:20Z

RAS context memory needs to freed in failure case. Signed-off-by: Guchun Chen Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher

drm/amdgpu: print warning when input address is invalid

2020-05-28T18:00:49Z

This will assist debug in error injection case. Signed-off-by: Guchun Chen Reviewed-by: Tao Zhou Signed-off-by: Alex Deucher

drm/amdgpu: Update RAS XGMI error inject sequence

2020-05-14T21:42:35Z

Disable XGMI link power down prior to issuing a XGMI RAS error Reviewed-by: Hawking Zhang Signed-off-by: John Clements Signed-off-by: Alex Deucher

drm/amdgpu: allocate large structures dynamically

2020-05-05T17:12:55Z

After the structure was padded to 1024 bytes, it is no longer suitable for being a local variable, as the function surpasses the warning limit for 32-bit architectures: drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:587:5: error: stack frame size of 1072 bytes in function 'amdgpu_ras_feature_enable' [-Werror,-Wframe-larger-than=] int amdgpu_ras_feature_enable(struct amdgpu_device *adev, ^ Use kzalloc() instead to get it from the heap. Fixes: a0d254820f43 ("drm/amdgpu: update RAS TA to Host interface") Acked-by: Christian König Signed-off-by: Arnd Bergmann Signed-off-by: Alex Deucher

drm/amdgpu: update RAS error handling

2020-04-30T20:48:20Z

Parse return status from TA to determine error severity Reviewed-by: Hawking Zhang Signed-off-by: John Clements Signed-off-by: Alex Deucher