On Navi1x, the SPM golden settings are lost after GFXOFF
enter/exit, so reconfigure the golden settings after GFXOFF
exit.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On Navi1x, the SPM golden settings are lost after GFXOFF
enter/exit, so reconfiguration is needed. Make the
configuration code as an interface for future use.
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Before ras recovery is issued, user could operate this debugfs
node to enable/disable the harvest of all RAS IPs' ras error
count registers, which will help keep hardware's registers'
status instead of cleaning up them.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Once ras recovery is issued by ras sync flood interrupt or
ras controller interrupt, add this guard to bypass or execute
ras error count register harvest of all IPs.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Access the exported P2P dmabuf over XGMI, if available.
Otherwise, fall back to the existing PCIe method.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Arunpravin <apaneers@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
GFX10 KIQ will hang if we try below steps:
modprobe amdgpu
rmmod amdgpu
modprobe amdgpu sched_hw_submission=4
Due to KIQ is always living there even after KMD unloaded
thus when doing the realod KIQ will crash upon its register
being programed by different values with the previous loading
(the config like HQD addr, ring size, is easily changed if we alter
the sched_hw_submission)
the fix is we must inactive KIQ first before touching any
of its registgers
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Emily Deng <Emily.Deng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Some registers are not accessible to virtual function setup, so
skip their initialization when in VF-SRIOV mode.
v2: move SRIOV VF check into specify functions;
modify commit description and comment.
Signed-off-by: Liu ChengZhe <ChengZhe.Liu@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When '*sgt' is allocated, we must allocated 'sizeof(**sgt)' bytes instead
of 'sizeof(*sg)'.
The sizeof(*sg) is bigger than sizeof(**sgt) so this wastes memory but
it won't lead to corruption.
Fixes: f44ffd677f ("drm/amdgpu: add support for exporting VRAM using DMA-buf v3")
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Allows UMD to know if TMZ is supported and enabled.
This commit also bumps KMS_DRIVER_MINOR because if we don't
UMD can't tell if "ids_flags & AMDGPU_IDS_FLAGS_TMZ == 0" means
"tmz is not enabled" or "tmz may be enabled but the kernel doesn't
report it".
v2: use amdgpu_is_tmz() and reworded commit message.
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If multiple process share system memory through /dev/shm, KFD allocate
memory should not fail if it reaches the system memory limit because
one copy of physical system memory are shared by multiple process.
Add module parameter no_system_mem_limit to provide user option to
disable system memory limit check at runtime using sysfs or during
driver module init using kernel boot argument. By default the system
memory limit is on.
Print out debug message to warn user if KFD allocate memory failed
because system memory reaches limit.
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
I suspect the only reason this was set was to avoid touching
the display related registers on arcturus. Someone should
double check this on arcturus with S3.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This adds a new gmc callback to get the size reserved by the pre-OS
console and provides a helper function for use by gmc IP drivers.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This will allow us to split the allocation for systems
where we have to keep the stolen memory around to avoid
S3 issues. This way we don't waste as much memory and
still avoid any screen artifacts during the bios to
driver transition.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Just return early to match other bo_create functions.
v2: check if the bo_ptr is NULL rather than checking the size.
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> (v1)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
what:
the MQD's save and restore of KCQ (kernel compute queue)
cost lots of clocks during world switch which impacts a lot
to multi-VF performance
how:
introduce a paramter to control the number of KCQ to avoid
performance drop if there is no kernel compute queue needed
notes:
this paramter only affects gfx 8/9/10
v2:
refine namings
v3:
choose queues for each ring to that try best to cross pipes evenly.
v4:
fix indentation
some cleanupsin the gfx_compute_queue_acquire()
v5:
further fix on indentations
more cleanupsin gfx_compute_queue_acquire()
TODO:
in the future we will let hypervisor driver to set this paramter
automatically thus no need for user to configure it through
modprobe in virtual machine
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
During driver's probe, when it hits bad gpu tag in eeprom i2c
init calling(the tag was set when reported bad page reaches
bad page threshold in last driver's working loop), there are
some strategys to deal with the cases:
1. when the module parameter amdgpu_bad_page_threshold = 0,
that means page retirement feature is disabled, so just resetting
the eeprom is fine.
2. When amdgpu_bad_page_threshold is not 0, and moreover, user
sets one bigger valid data in order to make current boot up
succeeds, correct eeprom header tag and do not break booting.
3. For other cases, driver's probe will be broken.
v2: Just update eeprom header tag instead of resetting the whole
table header when user sets one bigger threshold data.
v3: Use dev_info/dev_err to print PCI device information, which
helps in mGPU case.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When amdgpu_bad_page_threshold = 0, bad page reservation stuffs
are skipped in either UMC ECC irq or page retirement calling of
sync flood isr.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Bad page information should not be exposed by sysfs when
bad page retirement is disabled, so decouple it from ras
sysfs group creating, and add one guard before creating.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add one definition for the RAS module's FS name. It's used
in both debugfs and sysfs cases.
v2: Use static variable instead of macro definition.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
RAS flags needs to be cleaned as well when user requires
one clean eeprom.
v2: RAS flags shall be restored after eeprom reset succeeds.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When GPU executes recovery and retriving bad GPU tag
from external eerpom device, the recovery will be broken
and error message is printed as well for user's awareness.
v2: Refine warning message in threshold reaching case, and
fix spelling typo.
v3: Fix explicit calling of bad gpu.
v4: Rename function names.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Once the bad page saved to eeprom reaches the configured
threshold, ras recovery will be issued to notify user.
v2: Fix spelling typo.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Once the ras recovery is issued from eeprom write itself,
bad page reservation should be ignored, otherwise, recursive
calling of writting to eeprom would happen.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When retrieving bad gpu tag from eeprom, GPU init should
fail as the GPU needs to be retired for further check.
v2: Fix spelling typo, correct the condition to detect
bad gpu tag and refine error message.
v3: Refine function argument name.
v4: Fix missing check of returning value of i2c
initialization error case.
v5: Use dev_err to print PCI information in dmesg instead
of DRM_ERROR.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Bad page threshold value should be valid in the range between
-1 and max records length of eeprom. It could determine when
saved bad pages exceed threshold value, and proceed corresponding
actions.
v2: When using the default typical value, it should be min
value between typical value and eeprom max records length.
v3: drop the case of setting bad_page_cnt_threshold to be
0xFFFFFFFF, as it confuses user.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
bad_page_threshold could be configured to enable/disable the
associated bad page retirement feature in RAS.
When it's -1, ras will use typical bad page failure value to
handle bad page retirement.
When it's 0, disable bad page retirement, and no bad page
will be recorded and saved.
For other valid value, driver will use this manual value
as the threshold value of totoal bad pages.
v2: correct documentation of this parameter.
v3: remove confused statement in documentation.
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
1. For Navi12, CHIP_SIENNA_CICHLID, skip tmr load operation;
2. Check pointer before release firmware.
v2: use CHIP_SIENNA_CICHLID instead
v3: remove local "bool ret"; fix grammer issue
v4: use my name instead of "root"
v5: fix grammer issue and indent issue
Signed-off-by: Liu ChengZhe <ChengZhe.Liu@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
