Syncpoint IRQs are currently requested in a code path that runs
during resume. Due to this, we get multiple overlapping registered
interrupt handlers as host1x is suspended and resumed.
Rearrange interrupt code to only request IRQs during initialization.
Bug 4658418
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Change-Id: I25e129452aeb39c23aa2d6f6a54729f60390d088
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3145171
Reviewed-by: Santosh BS <santoshb@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Upstream commit 690da22dbfa8 ("asm-generic/io.h: kill vmalloc.h
dependency") removed the vmalloc.h header file from io.h and this breaks
building various drivers with the latest -next kernels. Fix this
by ensuring vmalloc.h is included and in most cases slab is not actually
needed and so remove this where possible. Note that it is fine to make
this change for all current supported kernels.
Bug 4593750
Change-Id: I003d1302bda226d356467e6ede99949b2716940a
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3141984
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Enable CRC generation feature on VIC, OFA, and NVENC where
possible. Whether CRC generation can be enabled depends on
whether the engine's SEC register aperture is writable, based
on platform configuration.
The firewall configuration is for now determined in a hacky
way using ioremap, pending design decisions on how to do it
in a cleaner way.
Bug 4273775
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Change-Id: Ia7507f4cce2b49703328d5dfdd74aaac8cff8153
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3120662
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
The dma-buf import code currently attaches and maps all imported
dma-bufs to the drm device to get their sgt for mapping to the
directly managed IOMMU domain.
In many cases, like for newer chips (Tegra186+), the directly
managed IOMMU domain is, however, not used. Mapping to the drm
device can also cause issues e.g. with swiotlb since it is not
a real device.
To improve the situation, only attach and map imported dma-bufs
when required.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Change-Id: If73cc4e9419a2eebf67269eadb34050986f73077
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3090413
Reviewed-by: Santosh BS <santoshb@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The functions nvhost_job_put() and nvhost_syncpt_set_minval() are no
longer used so remove these functions.
The flag CONFIG_TEGRA_HOST1X is always true for out-of-tree drivers and
so remove the legacy code for when this flag is not defined.
Bug 4475969
Change-Id: I88c89d30bfc9be4b8628f115ea96835cb5f6f295
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3091653
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
The OOT variants of the host1x and tegra-drm drivers originally had
a '-next' suffix in the name to differentiate these from the upstream
versions. Now that the OOT drivers are installed under the 'updates'
directory, the '-next' suffix can be dropped because 'kmod' will
prioritise loading the variant in 'updates' over those built from the
kernel tree.
Bug 4410785
Change-Id: I29287ddb935b681a5ee8a3326428eba82715da76
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3065465
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Currently, each process holding an open TegraDRM channel reserves
for itself one of the limited number of hardware memory contexts.
Attempting to allocate a channel when all contexts are in use
results in failure.
While we cannot have more contexts than the hardware supports in
active use, idle channels don't necessarily need to have a backing
memory context. As such, in this patch, we add another layer
to allow hardware memory contexts to be "stolen away" by channels
that are in active use, from idle processes.
The way this is implemented, is by keeping track of memory
mappings on each abstracted memory context. If we need to
steal that memory context's backing hardware context, we unmap
everything from it and give it away. When that abstracted
memory context is needed again (re-activated), we attempt
to allocate or steal another hardware context and re-map
the previously unmapped buffers.
Unfortunately, this means additional overhead and unpredictability
at submit time. Submit can fail if we cannot re-allocate a
backing memory context. Future work includes a provision for
un-stealable backing hardware memory contexts for processes
requiring more determinism, as well as optimization and cosmetic
improvements.
Bug 4403250
Bug 4399310
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Change-Id: I3d13e3476f1bff3c4757152254496cddaaafd76a
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3058905
Reviewed-by: Santosh BS <santoshb@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The following error messages are sometimes observed on boot ...
tegra-nvjpg 15380000.nvjpg: failed to get icc write handle
tegra-nvdec 15480000.nvdec: failed to get icc write handle
tegra-nvjpg 15540000.nvjpg: failed to get icc write handle
tegra-nvenc 154c0000.nvenc: failed to get icc write handle
tegra-vic 15340000.vic: failed to get icc write handle
tegra-nvjpg 15380000.nvjpg: failed to get icc write handle
The above messages are harmless because the ICC core is returning
-EPROBE_DEFER to indicate that the ICC provider is not available. When
-EPROBE_DEFER is returned the kernel will attempt to probe the device
again and so print an error on -EPROBE_DEFER can be misleading. Fix the
above by using the function dev_err_probe() to print error messages
because this function will only print an error message if the error code
is not -EPROBE_DEFER.
Bug 3436156
Bug 4496044
Change-Id: I47b77e5a0f2bdb817a832daa305246c8803f456b
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3075237
(cherry picked from commit 61319aef4d)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3075849
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Remove the legacy NVHOST support from the nvhost.h file that is no
longer needed. Eventually we will remove this file completely, but it is
currently still needed as it provides a shim layer for working with the
upstream host1x driver.
Bug 4475969
Change-Id: Id4d97b41de2590800e321e4eeb9c03b434e00dc5
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3065976
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
For Linux v6.8, commit ef175b29a242 ("of: Stop circularly including
of_device.h and of_platform.h") updated the OF headers included by these
header files. This breaks the build for various drivers and so fix this
by including the headers that are actually needed for each driver.
Bug 4448428
Change-Id: Ia40ab13f865d5631c96855ecc49145848f99c996
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3032442
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
Tested-by: Laxman Dewangan <ldewangan@nvidia.com>
Building the Tegra DRM driver with Linux v6.8 fails with the following
error ...
drivers/gpu/drm/tegra/hdmi.c:623:25: error: implicit declaration of
function ‘drm_eld_size’; did you mean ‘mm_cid_size’?
[-Werror=implicit-function-declaration]
623 | size_t length = drm_eld_size(hdmi->output.connector.eld), i;
| ^~~~~~~~~~~~
Commit ("8eb80946ab0c drm/edid: split out drm_eld.h from drm_edid.h")
added the header file 'drm/drm_eld.h' which now needs to be included to
build the Tegra DRM driver. Fix the build issue by using conftest to
detect the presence of the header 'drm/drm_eld.h' and include this
header if present.
Bug 4448428
Change-Id: Ifb17648d1dc721e1f0de3f864f4cc76450eea394
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3027485
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
During channel debug information dump, when printing CDMA
opcodes, the circular nature of the CDMA pushbuffer wasn't being
taken into account, sometimes accessing past the end. Change
the printing to take this into account.
Bug 4398831
Bug 4386806
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Change-Id: I3a24da2c310e9414882f7cabbbda5158b6da00a2
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3034409
Reviewed-by: Rongrong Zhou <rongrongz@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The compiler option -Wmissing-prototypes is being enabled globally in
the upstream Linux kernel and this causes build failures for various
drivers. The build failures occur because either the driver is missing
an include file which has the prototype or because the function is not
declared statically when it should be (ie. there are no external users).
Fix the various build failures and enable -Wmissing-prototypes to
prevent any new instances from occurring. Note that the only driver that
is not fixed at the moment is rtl8822ce due to large number of failures
and so build this with '-Wno-missing-prototypes' for now, which is not
different to how it was being compiled prior to this change.
Bug 4404965
Change-Id: Ie5572d23659e0346fa035d645d9043b0a6da5fdc
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3027488
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Instead of using kernel version checks in the Tegra DRM driver for the
tegra_fb_ops structure, just check if the appropriate definition is
defined and if so use it. This fixes building the Tegra DRM driver
against some 3rd party kernels that include backports from upstream.
Note that this also aligns the driver closer to the upstream code by
using the same definitions that the upstream driver uses.
Bug 4346767
Change-Id: I8cb40899d0bf7cdc233e5f4cee7b6afb6f12d015
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3027832
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
When BPMP BWMGR receives ICC avg_bw requests from different memory
clients, it will sum up all avg_bw values together and use it to
determine the final EMC frequency.
Transitioning to ICC peak_bw requests will cause BPMP to evaluate
the bandwidth requirements of each memory client individually,
selecting the maximum bandwidth request from among all the clients
to determine the final EMC frequency.
This modification prevents excessive memory bandwidth allocation
when multiple host1x clients collaborate for data processing.
Currently, host1x clients request the full theoretical 100% data
bandwidth, even though the system typically doesn't fully utilize
that amount during runtime.
To address the issue of insufficient memory bandwidth when multiple
host1x clients are used together, we can reduce the boost_up_threshold
value of cactmon DFS. This adjustment ensures that when actual memory
bandwidth utilization surpasses the specified boost-up bandwidth
threshold, EMC frequency will be further scaled by one step further
to alleviate the problem.
Bug 4328471
Signed-off-by: Johnny Liu <johnliu@nvidia.com>
Change-Id: I826a374666f38718652c5cae449c0aadeabfbdb5
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2996561
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The FB deferred IO ops helper name is changed in Linux 6.6 with
below mainline kernel change:
---
commit 744d35d3436fbeed9131f6ab717aeb3f775b19c4
Author: Thomas Zimmermann <tzimmermann@suse.de>
Date: Sat Jul 29 21:26:49 2023 +0200
fbdev: Align deferred I/O with naming of helpers
---
Use the new macros name.
Bug 4346767
Change-Id: Idc2e2319994fd2ed7040d8d5528a0d6074257420
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3008092
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The functions for fd/handle conversion are unexported as it is
used internally to drm_prime based on following change:
---
Author: Thomas Zimmermann <tzimmermann@suse.de>
Date: Tue Jun 20 09:59:59 2023 +0200
drm/prime: Unexport helpers for fd/handle conversion
---
Remove the initialisation of these functions pointer.
Bug 4346767
Change-Id: I821f4e63663a7f35e5fa39ed84383765ed8599fa
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3008091
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Since kernel 6.1, merging SYNC_FILE FDs will coalesce
signaled fences into a stub fence - no longer a host1x
fence. To avoid breakage on userspace side, detect such
situations here and instead of returning an error, omit
such fences when processing.
Bug 4199972
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Change-Id: I22451ffa22c388bbeeb3528e4760730c3f83bc1b
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3002691
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
When actmon reports the client usage information, it will call the
get_rate callback provided by the client, and divide the number of
sampled active clock cycles from the actmon counter register with the
current client rate to determine the client usage value.
Bug 4338396
Signed-off-by: Johnny Liu <johnliu@nvidia.com>
Change-Id: I5e3f69b8ed1d90407fe0b8202e7c9cde75a0dd30
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3000330
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
In Linux v6.5, commit 62aeaeaa1b26 ("drm/aperture: Remove primary
argument") removed the 'primary' argument from the function
drm_aperture_remove_framebuffers(). Update the test in conftest script
for drm_aperture_remove_framebuffers() to test for this and use the
generated definition accordlingly in the Tegra DRM driver to fix the
build for Linux v6.5.
Bug 4221847
Change-Id: I7a7bb3a029b236e0ce42c1a404085757d95374d1
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2990528
In Linux v6.4, the various drm_scdc_get/set functions were updated to
take an argument of type 'struct drm_connector'. Add a new test to the
conftest script that checks if this argument is used and use the
definition created by conftest to select which version of the function
is used.
Bug 4221847
Change-Id: Id6b6396bbe875a0268f8fe2f68ae08f78a32c954
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2989035
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
In Linux v6.4, the module pointer argument is removed from the
class_create() function. Add a test to the conftest script that checks
if this argument for the class_create() function has been removed and
use the definition created by conftest to select which version of the
function is used.
Bug 4183168
Bug 4221847
Change-Id: I440e4b318001886cd0319bb3499ba33178475e8c
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2989020
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
If probing of the VIC engine fails then the following crash is observed
...
tegra-vic 15340000.vic: devfreq_add_device: Unable to find governor for
the device
tegra-vic 15340000.vic: failed to init devfreq: -22
tegra-vic: probe of 15340000.vic failed with error -22
tegra-nvdec 15480000.nvdec: Adding to iommu group 45
list_add corruption. prev->next should be next
(ffff000094505f80), but was 0000000000000000. (prev=ffff000095245cf0).
------------[ cut here ]------------
kernel BUG at lib/list_debug.c:26!
Internal error: Oops - BUG: 00000000f2000800 [#1] SMP
...
pc : __list_add_valid+0x80/0x84
lr : __list_add_valid+0x80/0x84
...
Call trace:
__list_add_valid+0x80/0x84
host1x_subdev_register+0x80/0x170 [host1x_next]
__host1x_client_register+0xac/0x160 [host1x_next]
nvdec_probe+0x1c8/0x58c [tegra_drm_next]
platform_probe+0x6c/0xe0
really_probe+0xc4/0x3dc
__driver_probe_device+0x114/0x190
driver_probe_device+0x44/0x114
__driver_attach+0xcc/0x1f0
bus_for_each_dev+0x78/0xd0
driver_attach+0x28/0x30
bus_add_driver+0xbc/0x220
driver_register+0x70/0x120
__platform_register_drivers+0x70/0x140
host1x_drm_init+0x58/0x1000 [tegra_drm_next]
do_one_initcall+0x48/0x2d0
do_init_module+0x5c/0x270
load_module+0xb08/0xc70
__do_sys_init_module+0x1dc/0x214
__arm64_sys_init_module+0x20/0x2c
invoke_syscall.constprop.0+0x7c/0xd0
el0_svc_common.constprop.0+0x140/0x150
do_el0_svc+0x38/0xa0
el0_svc+0x38/0x18c
el0t_64_sync_handler+0xb4/0x130
el0t_64_sync+0x17c/0x180
Code: aa0103e2 910fa000 aa0403e1 941b9509 (d4210000)
---[ end trace f8f20f933fe03c7e ]---
The problem is that some of the engines are not cleaning up correctly on
probe failure and so fix this by adding the necessary calls clean up if
probe fails.
Bug 4303860
Change-Id: Ife1e0ce7429458d187422162a298cc140f0e3bc4
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2986790
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
