Thomas Zimmermann asked to backmerge -rc6 for drm-misc branches,
there's a few same-area-changed conflicts (xe and amdgpu mostly) that
are getting a bit too annoying.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This tag contains habanalabs driver and accel changes for v6.9.
The notable changes are:
- New features and improvements:
- Configure interrupt affinity according to NUMA nodes for the MSI-X interrupts that are
assigned to the userspace application which acquires the device.
- Move the HBM MMU page tables to reside inside the HBM to minimize latency when doing
page-walks.
- Improve the device reset mechanism when consecutive heartbeat failures occur (firmware
fails to ack on heartbeat message).
- Check also extended errors in the PCIe addr_dec interrupt information.
- Rate limit the error messages that can be printed to dmesg log by userspace actions.
- Firmware related fixes:
- Handle requests from firmware to reserve device memory
- Bug fixes and code cleanups:
- constify the struct device_type usage in accel (accel_sysfs_device_minor).
- Fix the PCI health check by reading uncached register.
- Fix reporting of drain events.
- Fix debugfs files permissions.
- Fix calculation of DRAM BAR base address.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Oded Gabbay <ogabbay@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/ZdxJprop0EniVQtf@ogabbay-vm-u22.habana-labs.com
Since commit aed65af1cc ("drivers: make device_type const"), the driver
core can properly handle constant struct device_type. Move the
accel_sysfs_device_minor variable to be a constant structure as well,
placing it into read-only memory which can not be modified at runtime.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ricardo B. Marliere <ricardo@marliere.net>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Today we read PCI VENDOR-ID in order to make sure PCI link is
healthy. Apparently the VENDOR-ID might be stored on host and
hence, when we read it we might not access the PCI bus.
In order to make sure PCI health check is reliable, we will start
checking the DEVICE-ID instead.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The reserved memory for FW is currently saved in an ASIC property in
units of MB, just like the value that comes from FW.
Except the fact that it is not clear from the property's name, it means
also that a calculation to actual size is required everywhere that it is
used.
Modify the property to hold the size in bytes.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Currently the reserved memory request from FW is handled when running
with preboot only, but this request is relevant also when running with
full FW.
Modify to always handle this reservation request.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Fetching sensor data can fail due to various reasons. In order
not to pollute the kernel log, those error prints must be
rate limited.
Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The glbl error cause handling has a wrong assumption that all error
bits are consecutive.
Fix the handling to check all relevant error bits per ASIC.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The FW interrupt info for a PCIe addr_dec event is set correctly, so
check for either global errors or razwi according to the indications
there.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Skip loading a linux FW image into the device with the current supported
ASICs is done for test purposes only.
Moreover, for future supported ASICs it is possible that there won't be
a need to load such an image.
The print in such a case is therefore not needed in most cases, so
replace the used dev_info() with dev_dbg().
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The hop size related properties is a MMU properties and not
asic properties.
As for PMMU and HMMU we could have different sizes.
Signed-off-by: Farah Kassabri <fkassabri@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Reviewed-by: Carl Vanderlip <quic_carlv@quicinc.com>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
As noted in the "Deprecated Interfaces, Language Features, Attributes,
and Conventions" documentation [1], size calculations (especially
multiplication) should not be performed in memory allocator (or similar)
function arguments due to the risk of them overflowing. This could lead
to values wrapping around and a smaller allocation being made than the
caller was expecting. Using those allocations could lead to linear
overflows of heap memory and other misbehaviors.
So, use the purpose specific kcalloc() function instead of the argument
size * count in the kzalloc() function.
Link: https://www.kernel.org/doc/html/next/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments [1]
Link: https://github.com/KSPP/linux/issues/162
Signed-off-by: Erick Archer <erick.archer@gmx.com>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The pointer input is assigned a value that is not read, it is
being re-assigned again later with the same value. Resolve this
by moving the declaration to input into the if block.
Cleans up clang scan build warning:
warning: Value stored to 'input' during its initialization is never
read [deadcode.DeadStores]
Signed-off-by: Colin Ian King <colin.i.king@intel.com>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
In newer kernel versions, irq_set_affinity_hint() is deprecated.
Instead, use the newer version which is irq_set_affinity_and_hint().
Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The coding style in the Linux kernel prefers not to use
braces for single-statement if conditions.
This patch removes the unnecessary braces from an if statement
in the file drivers/accel/habanalabs/common/command_submission.c,
which also resolves a coding style warning.
Signed-off-by: Malkoot Khan <engr.mkhan1990@gmail.com>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Currently the HMMU page tables reside in the host memory,
which will cause host access from the device for every page walk.
This can affect PCIe bandwidth in certain scenarios.
To prevent that problem, HMMU page tables will be moved to the device
memory so the miss transaction will read the hops from there instead of
going to the host.
Signed-off-by: Farah Kassabri <fkassabri@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
The mechanism of aborting device reset for consecutive fatal errors is
currently only for fatal errors that are reported by FW.
A non-responsive FW and consecutive heartbeat failures is also
considered fatal, so add them as well to this mechanism to avoid
recurring device reset in such a case.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
When the DRAM region size in the BAR is not a power of 2, calculating
the corresponding BAR base address should be done using the offset from
the DRAM start address, and not using directly the DRAM address.
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Fix a warning of a buffer overflow:
‘snprintf’ output between 38 and 47 bytes into a destination of size 32
Signed-off-by: Koby Elbaz <kelbaz@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
User interrupts are MSIx interrupts coming from Gaudi2, that have
specific range of IDs and are assigned to the sole use of the user
process that opened the Gaudi2 device (reminder: there can be only
a single user process running on Gaudi2 at any given time).
The interrupts are allocated and managed by the driver and therefore,
the user expects the driver to initialize them properly, which also
includes setting the affinity to the related CPU cores of the
device's NUMA node to get maximum performance.
Signed-off-by: Dani Liberman <dliberman@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Issue IP reset before shutdown in order to
complete all upstream requests to the SOC.
Without this DevTLB is complaining about
incomplete transactions and NPU cannot resume from
suspend.
This problem is only happening on recent IFWI
releases.
IP reset in rare corner cases can mess up PCI
configuration, so save it before the reset.
After this happens it is also impossible to
issue PLL requests and D0->D3->D0 cycle is needed
to recover the NPU. Add WP 0 request on power up,
so the PUNIT is always notified about NPU reset.
Use D0/D3 cycle for recovery as it can recover
from failed IP reset and FLR cannot.
Fixes: 3f7c063492 ("accel/ivpu/37xx: Fix hangs related to MMIO reset")
Signed-off-by: Jacek Lawrynowicz <jacek.lawrynowicz@linux.intel.com>
Reviewed-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240207102446.3126981-1-jacek.lawrynowicz@linux.intel.com
