Newer HW doesn't appear to send this event, which will cause long delays
in runlist updates if they don't complete immediately.
RM doesn't use these events anywhere, and an NVGPU commit message notes
that polling is the preferred method even on HW that supports the event.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
We didn't used to be aware that runlist/engine IDs weren't the same thing,
or that there was such variability in configuration between GPUs.
By exposing this information to a client, and giving it explicit control
of which runlist it's allocating a channel on, we're able to make better
choices.
The immediate effect of this is that on GPUs where CE0 is the "GRCE", we
will now be allocating a copy engine running asynchronously to GR for BO
migrations - as intended.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
We have a need to fetch data from GPU-specific sub-devices that is not
tied to any particular engine object.
This commit provides the framework to support such queries.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This will be required to support Volta, but also allows us to remove code
that's duplicated for each channel type already.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Introduces a new method of defining channels available from the display,
common to all channel types, allowing for more flexibility in available
channel types/counts, and reducing the amount of boiler-plate required.
This will be required to support Volta.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Engines are initialised on an as-needed basis, so this results in the
same behaviour, whilst allowing us to simplify things a bit.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
We should be reading registers to determine which subunits are really
present on a given board, and this needs to be done after DEVINIT.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Likely a rebase bug. Should have no impact in default configuration due
to using per-instance setting by default.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
In preparation to enabling -Wvla, remove VLA. In this particular
case directly use macro NVKM_MSGQUEUE_CMDLINE_SIZE instead of local
variable cmdline_size. Also, remove cmdline_size as it is not
actually useful anymore.
The use of stack Variable Length Arrays needs to be avoided, as they
can be a vector for stack exhaustion, which can be both a runtime bug
or a security flaw. Also, in general, as code evolves it is easy to
lose track of how big a VLA can get. Thus, we can end up having runtime
failures that are hard to debug.
Also, fixed as part of the directive to remove all VLAs from
the kernel: https://lkml.org/lkml/2018/3/7/621
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
gcc points out a buffer that is clearly too small to be used
in a meaningful way, as the 'sizeof(*args) + argc > sizeof(stack)'
will always fail:
In function 'memcpy',
inlined from 'nvif_vmm_map' at drivers/gpu/drm/nouveau/nvif/vmm.c:55:2:
include/linux/string.h:353:9: error: '__builtin_memcpy' offset 40 is out of the bounds [0, 16] of object 'stack' with type 'u8[16]' {aka 'unsigned char[16]'} [-Werror=array-bounds]
return __builtin_memcpy(p, q, size);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/nouveau/nvif/vmm.c: In function 'nvif_vmm_map':
drivers/gpu/drm/nouveau/nvif/vmm.c:40:5: note: 'stack' declared here
This makes the buffer large enough so it should serve the purpose
that the author presumably had in mind. Alternatively we could
just get rid of it completely and simplify the code at the cost
of always doing the kmalloc (as we do in the current version).
Fixes: 920d2b5ef2 ("drm/nouveau/mmu: define user interfaces to mmu vmm opertaions")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This patch adds rampatch download compatibility for ROME >= 3.2.
Starting with ROME 3.2, the 'download mode' field of the rampatch
header indicates if the controller acknowledges (or not) the received
rampatch segments. If not, we need to send all the segments without
expecting any event from the controller (except for the last segment).
Goal is (I assume) to speed-up rampatch download.
This fixes BT on Dragonboard-600c P2 which includes the following BT
controller:
hci0: ROME Patch Version Request
hci0: Product:0x00000008
hci0: Patch :0x00000111
hci0: ROM :0x00000302
hci0: SOC :0x00000023
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
This function allows to send a HCI command without expecting any
controller event/response in return. This is allowed for vendor-
specific commands only.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
When both CONFIG_BT_HCIUART_INTEL and CONFIG_BT_HCIUART_BCM are not
selected, sparse complains like this:
drivers/bluetooth/hci_ldisc.c:437:9: warning: switch with no cases
Fix the sparse warning by proving a default switch case.
Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
I've seen timeout errors from HCI commands where it looks like
schedule_timeout() has returned immediately; additional logging for the
error case gives:
req_status=1 req_result=0 remaining=10000 jiffies
so the device is still in state HCI_REQ_PEND and the value returned by
schedule_timeout() is the same as the original timeout (HCI_INIT_TIMEOUT
on a system with HZ=1000).
Use wait_event_interruptible_timeout() instead of open-coding similar
behaviour which is subject to the spurious failure described above.
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
There are some controllers sending out advertising data with illegal
length value which is longer than HCI_MAX_AD_LENGTH, causing the
buffer last_adv_data overflows. To avoid these controllers from
overflowing the buffer, we do not process the advertisement data
if its length is incorrect.
Signed-off-by: Chriz Chow <chriz.chow@aminocom.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Initialize hw_name to "BCM", this avoids the need for a number of NULL
checks on hw_name later.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
btbcm_setup_patchram() starts with initializing the controller (and
getting the firmware filename) and then after loading the firmware,
does a re-init. This almost entirely duplicates the code in
btbcm_initialize(), use that function instead.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
btbcm_finalize() does a re-init of the controller, which is almost the
same as the initial init. Modify btbcm_initialize() so that it can be
used for this re-init and modify btbcm_finalize() to use it.
As an added bonus this also makes the dev_info from btbcm_finalize()
use the proper hw_name instead of always printing "BCM".
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Make btbcm_initialize() also work for USB connected device,
btbcm_initialize() and btbcm_setup_patchram() are quite similar,
this is a preparation patch for making btbcm_setup_patchram() use
btbcm_initialize() to remove the code duplication.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
We are using the same loop in both the UART and USB bus cases, refactor
things a bit to share the loop.
This is mostly meant to improve readability.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
btbcm_setup_patchram() was using the upper nibble of the revision code to
determine if we are dealing with an uart or USB connected bcm-bt device,
but just as btbcm_initialize() has started accepting 1 and 2 as uart
connected devices, I've now encountered an USB connected device (0a5c:216c)
which has 0 in the upper nibble. So it seems that the upper nibble is not
really a reliable indicator of the bus type.
Instead check hdev->bus which does give us a reliable indication. This
fixes the patchram code trying to load the patchram by the fallback BCM.hcd
filename, now it correctly requests BCM43142A0-0a5c-216c.hcd.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Add support for Qualcomm serial slave devices. Probe the serial device,
retrieve its maximum speed and register a new hci uart device.
Signed-off-by: Thierry Escande <thierry.escande@linaro.org>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Add binding document for serial bluetooth chips using Qualcomm protocol.
Signed-off-by: Thierry Escande <thierry.escande@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Add a new serial node for the Qualcomm BT controller QCA6174. This
allows automatic probing and hci registration through the serdev
framework instead of relying on the userspace helpers.
Signed-off-by: Thierry Escande <thierry.escande@linaro.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>