This set of changes includes a rework of the custom syncpoint interrupt
code to take better advantage of existing DRM/KMS infrastructure.
There's also various bits of cleanup and fixes included.
-----BEGIN PGP SIGNATURE-----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=PcBB
-----END PGP SIGNATURE-----
Merge tag 'drm/tegra/for-6.3-rc1' of https://gitlab.freedesktop.org/drm/tegra into drm-next
drm/tegra: Changes for v6.3-rc1
This set of changes includes a rework of the custom syncpoint interrupt
code to take better advantage of existing DRM/KMS infrastructure.
There's also various bits of cleanup and fixes included.
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thierry Reding <thierry.reding@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230127170119.495943-1-thierry.reding@gmail.com
drm/i915 feature pull #2 v6.3:
Features and functionality:
- Enable HF-EEODB by switching HDMI, DP and LVDS to use struct drm_edid (Jani)
- Start using unversioned DMC firmware paths for new platforms (Gustavo)
Refactoring and cleanups:
- ELD refactor: Stop using hardware buffer, precompute ELD, and wire up ELD in
the state checker (Ville)
- Use generics for debugfs device parameters (Jani)
- DSB refactoring and fixes (Ville)
- Header refactoring, add new intel_display_limits.h (Jani)
- Split out GMCH code to a new file (Jani)
- Split out vblank code to a new file (Jani)
- i915_drv.h and struct drm_i915_private cleanups (Jani)
- Simplify FBC and DRRS debug attributes (Deepak R Varma)
- Remove some single-use macros (Rodrigo)
Fixes:
- Fix scaler limits for display versions 12 and 13 (Luca)
- Fix plane source size check for zero height (Drew Davenport)
- Implement PSR2 selective fetch workaround (Jouni)
- Expand a PSR workaound to more platforms and pipes (Jouni)
- Expand an HDMI infoframe workaround to all MTL steppings (Jouni)
- Enable PIPEDMC whenever its corresponding pipe is enabled (Imre)
Merges:
- Backmerge drm-next (Jani)
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/87tu0c44gv.fsf@intel.com
- Moved the driver to the accel subsystem. Currently only the files were
moved (including the uapi file which was also renamed). This doesn't
include registering to the accel subsystem. This will probably be only
in the next kernel version.
- In case of decoder error (axi error) in Gaudi2, we can now find the exact
IP that initiated the erroneous transaction and print the details for
better debug.
- Add more trace events. We now can trace mmio transactions and communication
with the preboot firmware.
- Add to Gaudi2 support for abrupt reset that is done by the firmware. This
was support so far only for Gaudi1.
- Add uAPI to flush memory transactions (to the device memory). This is
needed by the communications library in case of doing p2p with a host NIC
which access our HBM directly through the PCI BAR.
- Add uAPI to pass-through a request from user-space to firmware and get the
result back to user-space. This will allow the driver code to avoid the
need to add new packet (in the communication channel with the firmware) for
every new request type.
- Remove the option to export dma-buf by memory allocation handle in our uAPI.
This was planned for Gaudi2 but was never used. Instead, we will do export
by memory address (same as Gaudi1). In addition, we added the option to
specify an offset to the address. This is needed in Gaudi2 because there
the user allocates the entire HBM in one allocation, but would like to
export only small part of it.
- Multiple bug fixes, refactors and small optimizations.
-----BEGIN PGP SIGNATURE-----
iQEzBAABCgAdFiEE7TEboABC71LctBLFZR1NuKta54AFAmPS7fYACgkQZR1NuKta
54Bu/wf+I4dySZslvWhjqom9zYlubnwKLpLxuKGeR2OeMZxe5AJSyNCohmMXjl4S
ac+2qktRbecYXwNHCZBFjybKUgqBN5VuxRNRjV3/It+koON82QkMvBB48CeD7W/P
9y5b38LYOxqFixd5U5e7MpNhU1/AAKGnqEtQZgub0h0IeOrZM0RkTL6PJoUdQAcn
ap3pNXm6cpNKo5EZYlyRwEJRw1YlUKt3lcsKNnBDjij1j+IOtQ0sPA7Ii0YMdtqh
/gL81JkyZmgQBq9aGjr8vGyILvVTfm8qJu7t4hGEZ165bAP4bQaU9W+YMDgpKmgM
Bc+NghtCz13gmKPY/gudP5fDbPHUdA==
=fOxf
-----END PGP SIGNATURE-----
Merge tag 'drm-habanalabs-next-2023-01-26' of https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux into drm-next
This tag contains habanalabs driver and accel changes for v6.3:
- Moved the driver to the accel subsystem. Currently only the files were
moved (including the uapi file which was also renamed). This doesn't
include registering to the accel subsystem. This will probably be only
in the next kernel version.
- In case of decoder error (axi error) in Gaudi2, we can now find the exact
IP that initiated the erroneous transaction and print the details for
better debug.
- Add more trace events. We now can trace mmio transactions and communication
with the preboot firmware.
- Add to Gaudi2 support for abrupt reset that is done by the firmware. This
was support so far only for Gaudi1.
- Add uAPI to flush memory transactions (to the device memory). This is
needed by the communications library in case of doing p2p with a host NIC
which access our HBM directly through the PCI BAR.
- Add uAPI to pass-through a request from user-space to firmware and get the
result back to user-space. This will allow the driver code to avoid the
need to add new packet (in the communication channel with the firmware) for
every new request type.
- Remove the option to export dma-buf by memory allocation handle in our uAPI.
This was planned for Gaudi2 but was never used. Instead, we will do export
by memory address (same as Gaudi1). In addition, we added the option to
specify an offset to the address. This is needed in Gaudi2 because there
the user allocates the entire HBM in one allocation, but would like to
export only small part of it.
- Multiple bug fixes, refactors and small optimizations.
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Oded Gabbay <ogabbay@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20230126213317.GA1520525@ogabbay-vm-u20.habana-labs.com
Use the newly implemented tegra_dev_iommu_get_stream_id() helper to
encapsulate and centralize the IOMMU stream ID access.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Use the newly implemented tegra_dev_iommu_get_stream_id() helper to
encapsulate and centralize the IOMMU stream ID access.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Use the newly implemented tegra_dev_iommu_get_stream_id() helper to
encapsulate and centralize the IOMMU stream ID access.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Use the newly implemented tegra_dev_iommu_get_stream_id() helper to
encapsulate and centralize the IOMMU stream ID access.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Use the RUNTIME_PM_OPS() and pm_ptr() macros to handle the
.runtime_suspend/.runtime_resume callbacks.
These macros allow the suspend and resume functions to be automatically
dropped by the compiler when CONFIG_PM is disabled, without having
to use #ifdef guards.
This has the advantage of always compiling these functions in,
independently of any Kconfig option. Thanks to that, bugs and other
regressions are subsequently easier to catch.
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Fix the following coccicheck warning:
./drivers/gpu/drm/tegra/submit.c:689:2-7: WARNING:
NULL check before some freeing functions is not needed.
Signed-off-by: Yushan Zhou <katrinzhou@tencent.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Currently all fences have a 30 second timeout to ensure they are
cleaned up if the fence never completes otherwise. However, this
one size fits all solution doesn't actually fit in every case,
such as syncpoint waiting where we want to be able to have timeouts
longer than 30 seconds. As such, we want to be able to give control
over fence cancellation to the caller (and maybe eventually get rid
of the internal timeout altogether).
Here we add this cancellation mechanism by essentially adding a
function for entering the timeout path by function call, and changing
the syncpoint wait function to use it.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Move from the old, complex intr handling code to a new implementation
based on dma_fences. While there is a fair bit of churn to get there,
the new implementation is much simpler and likely faster as well due
to allowing signaling directly from interrupt context.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
In anticipation of removal of the intr API, implement job tracking
using DMA fences instead. The main two things about this are
making cdma_update schedule the work since fence completion can
now be called from interrupt context, and some complication in
ensuring the callback is not running when we free the fence.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
In anticipation of removal of the intr API, move host1x_syncpt_wait
to use DMA fences instead. As of this patch, this means that waits
have a 30 second maximum timeout because of the implicit timeout
we have with fences, but that will be lifted in a follow-up patch.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The code to write the syncpoint channel assignment register
incorrectly skips the write if hypervisor registers are not available.
The register, however, is within the guest aperture so remove the
check and assign syncpoints properly even on virtualized systems.
Fixes: c3f52220f2 ("gpu: host1x: Enable Tegra186 syncpoint protection")
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
On Tegra186+, the syncpoint ID has 10 bits of space. To allow
using more than 256 syncpoints, fix the mask.
Fixes: 9abdd497cd ("gpu: host1x: Tegra234 device data and headers")
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The Tegra DRM tree moved to freedesktop.org's gitlab a few releases ago,
so update the MAINTAINERS entry accordingly.
Signed-off-by: Thierry Reding <treding@nvidia.com>
It's a bit confusing to have two cached EDIDs in struct intel_connector
with slightly different purposes. Make the distinction a bit clearer by
moving the EDID cached for eDP and LVDS panels at connector init time to
struct intel_panel, and name it fixed_edid. That's what it is, a fixed
EDID for the panels.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/328350ef918638928a8286cdbab3107c8258332d.1674643465.git.jani.nikula@intel.com
Convert all the connectors that use cached connector edid and
detect_edid to drm_edid.
Since drm_get_edid() calls drm_connector_update_edid_property() while
drm_edid_read*() do not, we need to call drm_edid_connector_update()
separately, in part due to the EDID caching behaviour in HDMI and
DP. Especially DP depends on the details parsed from EDID. (The big
behavioural change conflating EDID reading with parsing and property
update was done in commit 5186421cbf ("drm: Introduce epoch counter to
drm_connector"))
v6: Rebase on drm_edid_connector_add_modes()
v5: Fix potential uninitialized var use (kernel test robot <lkp@intel.com>)
v4: Call drm_edid_connector_update() after reading HDMI/DP EDID
v3: Don't leak vga switcheroo EDID in LVDS init (Ville)
v2: Don't leak opregion fallback EDID (Ville)
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/eabb4de932841b38b34cc2818ea9fbf7c10224fd.1674643465.git.jani.nikula@intel.com
Stephen Rothwell reported htmldocs warnings when merging accel tree:
Documentation/ABI/testing/sysfs-driver-habanalabs:201: ERROR: Unexpected indentation.
Documentation/ABI/testing/sysfs-driver-habanalabs:201: WARNING: Block quote ends without a blank line; unexpected unindent.
Documentation/ABI/testing/sysfs-driver-habanalabs:201: ERROR: Unexpected indentation.
Documentation/ABI/testing/sysfs-driver-habanalabs:201: WARNING: Block quote ends without a blank line; unexpected unindent.
Fix these by fixing alignment of list of card status returned by
/sys/class/habanalabs/hl<n>/status.
Link: https://lore.kernel.org/linux-next/20230120130634.61c3e857@canb.auug.org.au/
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The device specific directory in debugfs does not have "accel". For
example, the documentation says device 0 should have a debugfs entry as
/sys/kernel/debug/accel/accel0/ but in reality the entry is
/sys/kernel/debug/accel/0/
Fix the documentation to match the implementation.
Fixes: 8c5577a5cc ("doc: add documentation for accel subsystem")
Signed-off-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
When a decode error happens, we often don't know the exact root
cause (the erroneous address that was accessed) and the exact engine
that created the erroneous transaction.
To find out, we need to go over all the relevant register blocks
in the ASIC. Once we find the relevant engine, we print its details
and the offending address.
This helps tremendously when debugging an error that was created
by running a user workload.
Signed-off-by: Koby Elbaz <kelbaz@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
This is required in order to allow the kernel to control relevant
configuration space via load and store instructions.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
If resetting device upon release while the release watchdog work is
scheduled, the compute reset is replaced with hard reset.
In this case, need to clear the in_compute_reset indication in the
device reset information structure.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
If device memory scrubbing from hl_device_reset() fails, we return with
an error code but not perform error handling code.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
This commit enhances the following error messages to also provide the
type of error occurred, this in order to ease debugging of errors
detected during firmware-load.
Signed-off-by: Moti Haimovski <mhaimovski@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Completion timestamp is taken during the actual command submission
release. As the release happens in a work queue, the timestamp taken
is not accurate. Hence, we will take the timestamp in the interrupt
handler itself while propagating it to the release function.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
In order to support more user interrupt types in the future, we
enumerate the user interrupt type instead of using a boolean.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Handling edma razwi is different than all other engines since edma
uses sft routers. For hbw transactions sft router contain separate
interface for each edma and for lbw there is common interface for
both edma engines of the same dcore.
To handle the razwi correctly we need to:
1. Simplify the calculation of the sft router address.
2. Add razwi handling for edma qm errors, since edma qman doesn't
reports axi error response.
Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
A device with status malfunction indicates that it can't be used.
In such a case we do not support certain reset types, e.g.,
all kinds of soft-resets (compute reset, inference soft-reset),
and reset upon device release.
A hard-reset is the only way that an unusable device can change its
status. All other reset procedures can't put the device in a reset
procedure, which might ultimately cause the device to change its
status, unintentionally, to become operational again.
Such a scenario has recently occurred, when a user requested
a hard-reset while another heavy user workload was ongoing (reset
request is queued).
Since the workload couldn't finish within reset's timeout limits, the
reset has failed and set a device status malfunction.
Eventually, when the user released the FD, an unsuccessful soft-reset
occurred, hence followed by an additional hard-reset that changed the
ASICs status back to be operational.
Signed-off-by: Koby Elbaz <kelbaz@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
AXI transaction id holds information about the initiator which caused
the page fault. In the future it will be translated automatically by
driver to an initiator name.
Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
As device status was changed recently, we must update the
documentation as well.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
In reviewing the ivpu driver, DEFINE_DRM_ACCEL_FOPS could have been used
if DRM_ACCEL_FOPS defined .mmap to be drm_gem_mmap. Lets add that since
accel drivers are a variant of drm drivers, modern drm drivers are
expected to use GEM, and mmap() is a common operation that is expected
to be heavily used in accel drivers thus the common accel driver should
be able to just use DEFINE_DRM_ACCEL_FOPS() for convenience.
Signed-off-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
get_maintainer.pl does not suggest Oded Gabbay, the DRM COMPUTE
ACCELERATORS DRIVERS AND FRAMEWORK maintainer for changes that touch
the Accel Subsystem header - drm_accel.h. This is because that file is
missing from the Accel Subsystem entry. Fix this.
Signed-off-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Captured addresses of low b/w razwi information contains only the
offset from the cfg base. To make it more user readable, add the cfg
base to it.
Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
In gaudi2 there night be different routers for low b/w and high b/w
transactions. But in the code that collects razwi information, we used
the same router for high b/w and low b/w.
Fixed it by reading the information also from low b/w routers.
Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Protect re-using the same timestamp buffer record before actually
adding it to the to interrupt wait list.
Mark ts buff offset as in use in the spinlock protection area of the
interrupt wait list to avoid getting in the re-use section in
ts_buff_get_kernel_ts_record before adding the node to the list.
this scenario might happen when multiple threads are racing on
same offset and one thread could set data in the ts buff in
ts_buff_get_kernel_ts_record then the other thread takes over
and get to ts_buff_get_kernel_ts_record and we will try
to re-use the same ts buff offset then we will try to
delete a non existing node from the list.
Signed-off-by: farah kassabri <fkassabri@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
use argument instead of fixed GFP value for allocation
in Timestamps buffers alloc function.
change data type of size to size_t.
Fixes: 9158bf69e7 ("habanalabs: Timestamps buffers registration")
Signed-off-by: farah kassabri <fkassabri@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Make sure all reserved/pad fields in uapi input structures
are set to 0.
Signed-off-by: farah kassabri <fkassabri@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
data is a void * type and does not require a cast.
Signed-off-by: XU pengfei <xupengfei@nfschina.com>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
This commit attaches the PCI device address to driver fatal messages
in order to ease debugging in multi-device setups.
Signed-off-by: Moti Haimovski <mhaimovski@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Because f/w does not update razwi info when sending events, remove the
use of it.
The driver is responsible to check if razwi happened and to
collect razwi data.
Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
