No need to check poison setting for each channel, check for umc0
channel0 is enough.
Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add basic support for GC 10.1.4,
it uses same IP blocks with GC 10.1.3
Signed-off-by: Lang Yu <Lang.Yu@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
"adev->gfx.rlc.rlcg_reg_access_supported = true;"
the above varible were set too late during driver initialization.
it will cause the driver to fail to write/read register during GMC hw init
in sriov mode.
move gfx_xxx_init_rlcg_reg_access_ctrl() function to gfx early init stage
to avoid this issue.
Fixes: 5d447e2967 ("drm/amdgpu: add helper for rlcg indirect reg access")
Signed-off-by: Yang Wang <KevinYang.Wang@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
After the buffer object is successfully mapped,
call amdgpu_bo_kunmap before the function returns.
Signed-off-by: zhanglianjie <zhanglianjie@uniontech.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cleanup the kfd code by removing the unused old debugger
implementation.
The address watch was only ever implemented in the upstream
driver for GFXv7 (Kaveri). The user mode tools runtime using
this API was never open-sourced. Work on the old debugger
prototype that used this API has been discontinued years ago.
Only a small piece of resetting wavefronts is kept and
is moved to kfd_device_queue_manager.c.
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
These have been at production level for a while. Drop
the flag.
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We want to have lockdep annotation here, so make sure that we reserve
the PD while removing PRTs even if it isn't strictly necessary since the
VM object is about to be destroyed anyway.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Since newly added BOs don't have any mappings it's ok to add them
without holding the VM lock. Only when we add per VM BOs the lock is
mandatory.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reported-by: Bhardwaj, Rajneesh <Rajneesh.Bhardwaj@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Bypass group programming (utcl2_harvest) aims to forbid UTCL2 to send
invalidation command to harvested SE/SA. Once invalidation command comes
into harvested SE/SA, SE/SA has no response and system hang.
This patch is to add checking if the GART table is already allocated before
invalidating TLB. The new procedure is as following:
1. Calling amdgpu_gtt_mgr_init() in amdgpu_ttm_init(). After this step GTT
BOs can be allocated, but GART mappings are still ignored.
2. Calling amdgpu_gart_table_vram_alloc() from the GMC code. This allocates
the GART backing store.
3. Initializing the hardware, and programming the backing store into VMID0
for all VMHUBs.
4. Calling amdgpu_gtt_mgr_recover() to make sure the table is updated with
the GTT allocations done before it was allocated.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Confirmed with hardware team, there is harvesting for gc 10.3.1.
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The parameter order in the list_add_tail is incorrect, it causes the
reuse of ras reserved page.
Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The commit d5e8ff5f7b ("drm/amdgpu: Fixed the defect of soft lock caused by infinite loop")
had fixed this defect.
Revert workaround
commit a2170b4af6 ("drm/amdgpu: Add judgement to avoid infinite loop").
Signed-off-by: yipechai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The FRU and RAS EEPROMs share the same I2C bus on Aldebaran and Vega 20
ASICs. Set the FRU bus "pointer" to this single bus, as access to the FRU
is sought through that bus "pointer" and not through the RAS bus "pointer".
Cc: Roy Sun <Roy.Sun@amd.com>
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Fixes: 2f60dd5076 ("drm/amd: Expose the FRU SMU I2C bus")
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Prevent random memory access in the FRU EEPROM code by passing the size of
the destination buffer to the reading routine, and reading no more than the
size of the buffer.
Cc: Kent Russell <kent.russell@amd.com>
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Read buffers no longer expose the I2C address, and so we don't need to
offset by two when we get the read data.
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Kent Russell <kent.russell@amd.com>
Cc: Andrey Grodzovsky <Andrey.Grodzovsky@amd.com>
Fixes: bd607166af ("drm/amdgpu: Enable reading FRU chip via I2C v3")
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Buffer is abbreviated "buf" (buf-fer), not "buff" (buff-er).
This is consistent with the rest of the kernel code.
Cc: Kent Russell <kent.russell@amd.com>
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This adds support to create userptr BOs on restore and introduces a new
ioctl op to restart memory notifiers for the restored userptr BOs.
When doing CRIU restore MMU notifications can happen anytime after we call
amdgpu_mn_register. Prevent MMU notifications until we reach stage-4 of the
restore process i.e. criu_resume ioctl op is received, and the process is
ready to be resumed. This ioctl is different from other KFD CRIU ioctls
since its called by CRIU master restore process for all the target
processes being resumed by CRIU.
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: David Yat Sin <david.yatsin@amd.com>
Signed-off-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This adds support to discover the buffer objects that belong to a
process being checkpointed. The data corresponding to these buffer
objects is returned to user space plugin running under criu master
context which then stores this info to recreate these buffer objects
during a restore operation.
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: David Yat Sin <david.yatsin@amd.com>
Signed-off-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
MESA polls for errors every 2-3 seconds. Printing with dev_info() causes
the dmesg log to fill up with the same message, e.g,
[18028.206676] amdgpu 0000:0b:00.0: amdgpu: df doesn't config ras function.
Make it dev_dbg_once(), as it isn't something correctible during boot or
thereafter, so printing just once is sufficient. Also sanitize the message.
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Hawking Zhang <Hawking.Zhang@amd.com>
Cc: John Clements <john.clements@amd.com>
Cc: Tao Zhou <tao.zhou1@amd.com>
Cc: yipechai <YiPeng.Chai@amd.com>
Fixes: 8b0fb0e967 ("drm/amdgpu: Modify gfx block to fit for the unified ras block data and ops")
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Some people complained about the name and this matches much
more Linux naming conventions for object functions.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Whenever a bo_va structure is added or removed the VM and eventually
added BO should be locked.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
trace_amdgpu_vm_update_ptes trace unable to log when nptes too large
Signed-off-by: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
dGPUs connected to Intel systems configured for suspend to idle
will not have the power rails cut at suspend and resetting the GPU
may lead to problematic behaviors.
Fixes: e25443d276 ("drm/amdgpu: add a dev_pm_ops prepare callback (v2)")
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1879
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We ran into the problem that clearing really larger buffer (60GiB) caused an
SDMA timeout.
Restructure the function to use the dst window instead of mapping the whole
buffer into the GART and then fill only 2MiB/256MiB chunks at a time.
v2: rebase on restructured window map.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Instead of limiting the size before we call the mapping
function let the function itself limit the size.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
That should never happen, but make sure that we only warn instead of
crash.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We probably never trigger this, but the logic inside the check is
inverted.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
A lot of below message are outputed in SRIOV case.
amdgpu: indirect registers access through rlcg is not supported
Also drop redundant ret set, as it's initialized to be false already.
Fixes: 29dbcac82f ("drm/amdgpu: add helper to query rlcg reg access flag")
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
1. The infinite loop causing soft lock occurs on multiple amdgpu cards
supporting ras feature.
2. This a workaround patch to fix 6492e1b07c.
It is valid for multiple amdgpu cards of the same type.
3. The root cause is that each GPU card device has a separate .ras_list
link header, but the instance and linked list node of each ras block
are unique. When each device is initialized, each ras instance will
repeatedly add link node to the device every time. In this way, only
the .ras_list of the last initialized device is completely correct.
the .ras_list->prev and .ras_list->next of the device initialzied
before can still point to the correct ras instance, but the prev
pointer and next pointer of the pointed ras instance both point to
the last initialized device's .ras_ list instead of the beginning
.ras_ list. When using list_for_each_entry_safe searches for
non-existent Ras nodes on devices other than the last device, the
last ras instance next pointer cannot always be equal to the
beginning .ras_list, so that the loop cannot be terminated, the
program enters a infinite loop.
BTW: Since the data and initialization process of each card are the same,
the link list between ras instances will not be destroyed every time
the device is initialized.
4. The soft locked logs are as follows:
[ 262.165690] CPU: 93 PID: 758 Comm: kworker/93:1 Tainted: G OE 5.13.0-27-generic #29~20.04.1-Ubuntu
[ 262.165695] Hardware name: Supermicro AS -4124GS-TNR/H12DSG-O-CPU, BIOS T20200717143848 07/17/2020
[ 262.165698] Workqueue: events amdgpu_ras_do_recovery [amdgpu]
[ 262.165980] RIP: 0010:amdgpu_ras_get_ras_block+0x86/0xd0 [amdgpu]
[ 262.166239] Code: 68 d8 4c 8d 71 d8 48 39 c3 74 54 49 8b 45 38 48 85 c0 74 32 44 89 fa 44 89 e6 4c 89 ef e8 82 e4 9b dc 85 c0 74 3c 49 8b 46 28 <49> 8d 56 28 4d 89 f5 48 83 e8 28 48 39 d3 74 25 49 89 c6 49 8b 45
[ 262.166243] RSP: 0018:ffffac908fa87d80 EFLAGS: 00000202
[ 262.166247] RAX: ffffffffc1394248 RBX: ffff91e4ab8d6e20 RCX: ffffffffc1394248
[ 262.166249] RDX: ffff91e4aa356e20 RSI: 000000000000000e RDI: ffff91e4ab8c0000
[ 262.166252] RBP: ffffac908fa87da8 R08: 0000000000000007 R09: 0000000000000001
[ 262.166254] R10: ffff91e4930b64ec R11: 0000000000000000 R12: 000000000000000e
[ 262.166256] R13: ffff91e4aa356df8 R14: ffffffffc1394320 R15: 0000000000000003
[ 262.166258] FS: 0000000000000000(0000) GS:ffff92238fb40000(0000) knlGS:0000000000000000
[ 262.166261] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 262.166264] CR2: 00000001004865d0 CR3: 000000406d796000 CR4: 0000000000350ee0
[ 262.166267] Call Trace:
[ 262.166272] amdgpu_ras_do_recovery+0x130/0x290 [amdgpu]
[ 262.166529] ? psi_task_switch+0xd2/0x250
[ 262.166537] ? __switch_to+0x11d/0x460
[ 262.166542] ? __switch_to_asm+0x36/0x70
[ 262.166549] process_one_work+0x220/0x3c0
[ 262.166556] worker_thread+0x4d/0x3f0
[ 262.166560] ? process_one_work+0x3c0/0x3c0
[ 262.166563] kthread+0x12b/0x150
[ 262.166568] ? set_kthread_struct+0x40/0x40
[ 262.166571] ret_from_fork+0x22/0x30
Fixes: 6492e1b07c ("drm/amdgpu: Unify ras block interface for each ras block")
Signed-off-by: yipechai <YiPeng.Chai@amd.com>
Reviewed-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
'linux/pci.h' included in 'amdgpu_device.c' is duplicated.
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Changcheng Deng <deng.changcheng@zte.com.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We observed a GPU hang when querying GMC CG state(i.e.,
cat amdgpu_pm_info) on cyan skillfish. Acctually, cyan
skillfish doesn't support any CG features.
Just prevent it from accessing GMC CG registers.
Signed-off-by: Lang Yu <Lang.Yu@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This will cause misconfigured systems to not run the GPU suspend
routines.
* In APUs that are properly configured system will go into s2idle.
* In APUs that are intended to be S3 but user selects
s2idle the GPU will stay fully powered for the suspend.
* In APUs that are intended to be s2idle and system misconfigured
the GPU will stay fully powered for the suspend.
* In systems that are intended to be s2idle, but AMD dGPU is also
present, the dGPU will go through S3
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This will be used to help make decisions on what to do in
misconfigured systems.
v2: squash in semicolon fix from Stephen Rothwell
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On some OEM setups users can configure the BIOS for S3 or S2idle.
When configured to S3 users can still choose 's2idle' in the kernel by
using `/sys/power/mem_sleep`. Before commit 6dc8265f98 ("drm/amdgpu:
always reset the asic in suspend (v2)"), the GPU would crash. Now when
configured this way, the system should resume but will use more power.
As such, adjust the `amdpu_acpi_is_s0ix function` to warn users about
potential power consumption issues during their first attempt at
suspending.
Reported-by: Bjoren Dasse <bjoern.daase@gmail.com>
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1824
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wrong order for config and counter_id parameters was passed, when calling df_v3_6_pmc_set_deferred and df_v3_6_pmc_is_deferred functions.
Signed-off-by: huangqu <jinsdb@126.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add a new CTX ioctl operation to set stable pstates for profiling.
When creating traces for tools like RGP or using SPM or doing
performance profiling, it's required to enable a special
stable profiling power state on the GPU. These profiling
states set fixed clocks and disable certain other power
features like powergating which may impact the results.
Historically, these profiling pstates were enabled via sysfs,
but this adds an interface to enable it via the CTX ioctl
from the application. Since the power state is global
only one application can set it at a time, so if multiple
applications try and use it only the first will get it,
the ioctl will return -EBUSY for others. The sysfs interface
will override whatever has been set by this interface.
Mesa MR: https://gitlab.freedesktop.org/mesa/drm/-/merge_requests/207
v2: don't default r = 0;
v3: rebase on Evan's PM cleanup
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Enable the FRU EEPROM I2C bus for Sienna Cichlid
server boards, for which it is enabled by checking
the VBIOS version.
Cc: Roy Sun <Roy.Sun@amd.com>
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Expose both SMU I2C buses. Some boards use the same bus for both the RAS
and FRU EEPROMs and others use different buses. This enables the
additional I2C bus and sets the right buses to use for RAS and FRU EEPROM
access.
Cc: Roy Sun <Roy.Sun@amd.com>
Co-developed-by: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alex Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Use IP version rather than codename for athub.
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On ALDEBARAN, the umc channel bits are not original values, they
are hashed.
Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On ALDEBARAN, we need to traverse all column bits higher than
BIT11(C4C3C2) in a row, the shift of R14 bit should be also taken
into account. Retire all pages we find.
Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
One piece of umc normalizing address can be mapped to 16 pieces of
physical address in each umc channel on ALDEBARAN.
Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Create common amdgpu_umc_fill_error_record function for all versions
of UMC and clean up related codes.
Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On GPUs with RAS, poison can propagate between processes if VRAM is not
cleared when it is freed or allocated. The reason is, that not all write
accesses clear RAS poison. 32-byte writes by the SDMA engine do clear RAS
poison. Clearing memory in the background when it is freed should avoid
major performance impact. KFD has been doing this already for a long time.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
