Add a "priv" fence struct type and use that in the fence type to
emphasize that the inner data is not meant to be seen.
The fence unit needs to have an outside-visible fence type so that
fences can be allocated directly as a struct field in job metadata for
performance and simplicity, so hiding the type entirely wouldn't work.
A couple of places need to touch the priv data directly in channel code.
Those can be thought to be technically fence unit's code scattered
outside the fence files, but they mean that the architecture is not
perfect yet.
Jira NVGPU-5773
Change-Id: Ifa3c95757ae31eef0e32f2605293e23e210b065f
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2395071
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Split up nvgpu_channel_clean_up_jobs() on the clean_all parameter so
that there's one version for the asynchronous ("deferred") cleanup and
another for the synchronous deterministic cleanup that occurs in the
submit path.
Forking another version like this adds some repetition, but this lets us
look at both versions clearly in order to come up with a coherent plan.
For example, it might be feasible to have the light cleanup of pooled
items in also the nondeterministic path, and deferring heavy cleanup to
another, entirely separated job queue.
Jira NVGPU-5493
Change-Id: I5423fd474e5b8f7b273383f12302126f47076bd3
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2346065
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
As the submit job metadata has been simplified, the fence pool for job
tracking fences is now just complex code for very simple purposes, so
delete it. It's enough to hold the fence memory in the job struct itself
instead of having separately allocated objects with different lifetimes.
Each channel is using preallocated job arrays based on the prespecified
inflight job count. The fences are used for tracking job completion, and
a new job cannot be submitted before a previous wait has completed.
This means that even with a ringbuffer with space for only one job, the
previous job memory cannot get reclaimed by a new submit because the
submits are ordered.
Jira NVGPU-5773
Change-Id: I0c777df700aa7cfda6f971efa47aa72c5462b53a
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2392704
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Unify the job metadata handling by deleting the parts that have handled
dynamically allocated job structs and fences. Now a channel can be in
one less mode than before which reduces branching in tricky places and
makes the submit/cleanup sequence easier to understand.
While preallocating all the resources upfront may increase average
memory consumption by some kilobytes, users of channels have to supply
the worst case numbers anyway and this preallocation has been already
done on deterministic channels.
Flip the channel_joblist_delete() call in nvgpu_channel_clean_up_jobs()
to be done after nvgpu_channel_free_job(). Deleting from the list (which
is a ringbuffer) makes it possible to reuse the job again, so the job
must be freed before that. The comment about using post_fence is no
longer valid; nvgpu_channel_abort() does not use fences.
This inverse order has not posed problems before because it's been buggy
only for deterministic channels, and such channels do not do the cleanup
asynchronously so no races are possible. With preallocated job list for
all channels, this would have become a problem.
Jira NVGPU-5492
Change-Id: I085066b0c9c2475e38be885a275d7be629725d64
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2346064
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-cert <svc-mobile-cert@nvidia.com>
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: Debarshi Dutta <ddutta@nvidia.com>
Reviewed-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Decouple the fence information needed for providing submit postfences to
userspace by adding a separate type for that and using it to pass fence
data to ioctls.
The data in struct nvgpu_fence_type is used in various places:
- job tracking needs to know when a post fence is expired
- job submitters within the driver (vidmem clears) need to be able to
wait for these fences
- userspace needs the fence as an id, value pair or as a file descriptor
created from an os fence
To keep object lifetimes strict, start decoupling the os fence data out
of struct nvgpu_fence_type: delete nvgpu_fence_install_fd() and add
nvgpu_fence_extract_user() to return a struct nvgpu_user_fence that
contains only the necessary information. Storing the os fence in job
tracking metadata is legacy code and not useful. Passing the os fence
from where it's created through the whole submit path inside this
combined fence type has been convenient, though.
The internally stored cde job fence in dmabuf compression metadata is
still nvgpu_fence_type to keep this patch simple.
Jira NVGPU-5248
Change-Id: I75b7da676fb6aa083828f888c55571bbf7645ef3
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2359064
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com>
Reviewed-by: svc-mobile-cert <svc-mobile-cert@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Add a generic profiler based on the channel kickoff profiler. This
aims to provide a mechanism to allow engineers to (more) easily profile
arbitrary software paths within nvgpu.
Usage of this profiler is still primarily through debugfs. Next up is
a generic debugfs interface for this profiler in the Linux code.
The end goal for this is to profile the recovery code and generate
interesting statistics.
JIRA NVGPU-5606
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Change-Id: I99783ec7e5143855845bde4e98760ff43350456d
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2355319
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Implement empty stubs of the channel watchdog functions for when
watchdog is disabled from build. Add some forward declarations that were
missing. Now most call sites don't need #idefs for the build flag.
Add error checks for the wdt alloc failure.
Jira NVGPU-5494
Jira NVGPU-5493
Change-Id: I2d42e8ab4c5e045cd280b2e1f254396127bd154b
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2352050
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Split out the max value increment and syncpt interrupt registration out
of nvgpu_channel_sync_incr*(). This API is called in the submit path to
prepare buffers and tracking resources, but later on in the submit path
errors can still occur so that the increment wouldn't happen (unless
artificially forced by sw).
The increment and irq registration cannot easily be undone and it makes
more sense to do these at the moment when the prepared job is finally
ready, so add a new nvgpu_channel_sync_mark_progress() API to be called
later in the submit path to signal that progress shall eventually happen
on the sync. Without this, the max value would stay too large after an
unsuccessful submit until the channel gets closed.
The sync object (syncpt or semaphore) is always exclusively owned by the
channel that allocated it, so nonatomically reading the max value first
in sync_incr() and incrementing it later in mark_progress() is racefree;
all submits per channel are serialized.
Change the channel syncpoint to client managed from host managed so that
nvhost-exported sync fences behave correctly with the temporary state
where the fence threshold is over the max value. Ideally we'd always
track nvgpu-owned syncpts' max values internally, but this is enough for
now.
Jira NVGPU-5491
Change-Id: Idf0bda7ac93d7f2f114cdeb497fe6b5369d21c95
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2340465
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Move preallocation of priv cmdbuf metadata structs to the priv cmdbuf
level and do it always, not only on deterministic channels. This makes
job tracking simpler and loosens dependencies from jobs to cmdbuf
internals. The underlying dma memory for the cmdbuf data has always been
preallocated.
Rename the priv cmdbuf functions to have a consistent prefix.
Refactor the channel sync wait and incr ops to free any priv cmdbufs
they allocate. They have been depending on the caller to free their
resources even on error conditions, requiring the caller to know how
they work.
The error paths that could occur after a priv cmdbuf has been allocated
have likely been wrong for a long time. Usually the cmdbuf queue allows
allocating only from one end and freeing from only the other end, as
that's natural with the hardware job queue. However, in error conditions
the just recently allocated entries need to be put back. Improve the
interface for this.
[not part of the cherry-pick:] Delete the error prints about not enough
priv cmd buffer space. That is not an error. When obeying the
user-provided job sizes more strictly, momentarily running out of job
tracking resources is possible when the job cleanup thread does not
catch up quickly enough. In such a case the number of inflight jobs on
the hardware could be less than the maximum, but the inflight job count
that nvgpu sees via the consumed resources could reach the maximum.
Also remove the wrong translation to -EINVAL from err from one call to
nvgpu_priv_cmdbuf_alloc() - the -EAGAIN from the failed allocation is
important.
[not part of the cherry-pick: a bunch of MISRA mitigations.]
Jira NVGPU-4548
Change-Id: I09d02bd44d50a5451500d09605f906d74009a8a4
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2329657
(cherry picked from commit 25412412f31436688c6b45684886f7552075da83)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2332506
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com>
Reviewed-by: svc-mobile-cert <svc-mobile-cert@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Avoid repetitive branching on the c->deterministic flag and on build
time flags by breaking the submit function on the runtime flag into two
functions of which one gets called.
In deterministic mode the job tracking conditions are simpler, there are
a few extra prechecks to guarantee deterministic latency and the
railgate corner case, and deferred cleanup is never done.
In nondeterministic mode job tracking has more conditions, a power
reference is taken for the job lifetime, and deferred cleanup is
assumed.
These two paths still share some common code. Split it to two more
functions to act as easy building blocks so that the main logic is
apparent.
Jira NVGPU-4548
Change-Id: I64f91dcf09acb16f409dc04a12ad1e144d0cce56
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2333728
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Aggressive sync destroy is used on some platforms where the amount of
syncpoints is limited. It can cause sync objects to get allocated and
freed in the submit path and when jobs are cleaned up, so require
deferred cleanup. Allocations do not belong to job tracking in a
deterministic submit path.
Although this has been technically allowed before, deterministic
channels have likely not been a priority on those old platforms with
aggressive sync destroy set.
Update virtualized gp10b platform data to match on a gp10b-vgpu compat
string instead of gk20a-vgpu. gk20a (Tegra T124) hasn't been supported
for a long time. Delete the aggressive sync destroy field from this
platform. It's got enough syncpoints to not dynamically allocate them;
having this property set for gp10b-vgpu has likely been a mistake.
This is not a completely pure cherry-pick: also extend the gpu
characteristics to not advertise full deterministic submit support when
aggressive sync destroy is off. This platform flag cannot be adjusted by
the user unlike many other flags.
Jira NVGPU-4548
Change-Id: I283f546d48b79ac94b943d88e5dce55710858330
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2322042
(cherry picked from commit b1ba2b997b2174e365bcb0782ef3e67260ff9e57)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2328411
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Add an accessor function in the priv cmdbuf object for gva and size to
be written in a gpfifo entry once the cmdbuf build is finished. This
helps in eventually hiding the struct priv_cmd_entry as an
implementation detail.
Add a sanity check to verify that the buffer has been filled exactly to
the requested size. The cmdbufs are used to hold wait and increment
commands for syncpoints or gpu semaphores. A prefence buffer can hold a
number of wait commands of equal size, and the postfence buffer holds
exactly one increment.
Jira NVGPU-4548
Change-Id: I83132bf6de52794ecc419e033e9f4599e488fd68
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2325102
(cherry picked from commit d1831463a487666017c4c80fab0292a0b85c7d83)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2331339
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com>
Reviewed-by: svc-mobile-cert <svc-mobile-cert@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
A corner case has existed since ancient times for syncpoint-backed
prefences to not cause a gpu wait if the fence is found to be completed
in the submit path. This adds some unnecessary complexity, so don't
check for completion in software. Let the gpu "wait" for these
known-to-be-trivial waits too. Necessary priv cmdbuf space has been
allocated anyway.
Originally nvhost had 16-bit fences which would wrap around relatively
quickly, so waiting for an old fence could have looked like waiting for
a fence that will expire long in the future. With 32-bit thresholds,
this hasn't been the case for several Tegra generations anymore, and
nvhost doesn't ignore waits like this either.
The wait priv cmdbuf in submit path can still be missing even with a
prefence supplied because the Android sync framework supports sync fds
that contain zero fences inside; this can happen at least when merging
fences that have all been expired. In such conditions the wait cmdbuf
wouldn't even get allocated.
[this is squashed with commit 8b3b0cb12d118 (gpu: nvgpu: allow no wait
cmd with valid input fence) from
https://git-master.nvidia.com/r/c/linux-nvgpu/+/2325677]
Jira NVGPU-4548
Change-Id: Ie81fd8735c2614d0fedb7242dc9869d0961610eb
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2321762
(cherry picked from commit 8f3dac44934eb727b1bf4fb853f019cf4c15a5cd)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2324254
Reviewed-by: automaticguardword <automaticguardword@nvidia.com>
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com>
Reviewed-by: svc-mobile-cert <svc-mobile-cert@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Refactor user managed syncpoints out of the channel sync infrastructure
that deals with jobs submitted via the kernel api. The user syncpt only
needs to expose the id and gpu address of the reserved syncpoint. None
of the rest (fences, priv cmdbufs) is needed for that, so it hasn't been
ideal to couple with the user-allocated syncpts.
With user syncpts now provided by channel_user_syncpt, remove the
user_managed flag from the kernel sync api.
This allows moving all the kernel submit sync code to be conditionally
compiled in only when needed, and separates the user sync functionality
in a more clear way from the rest with a minimal API.
[this is squashed with commit 5111caea601a (gpu: nvgpu: guard user
syncpt with nvhost config) from
https://git-master.nvidia.com/r/c/linux-nvgpu/+/2325009]
Jira NVGPU-4548
Change-Id: I99259fc9cbd30bbd478ed86acffcce12768502d3
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2321768
(cherry picked from commit 1095ad353f5f1cf7ca180d0701bc02a607404f5e)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2319629
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
If the os fence is the only kind that's supported, fail a submit if the
user wants fences but doesn't explicitly request sync fences, expecting
syncpoints. Syncpoint support is advertised to userspace in the gpu
characteristics, so userspace already has the knowledge to request the
correct sync type.
Do this check at the ioctl level. The in-kernel stuff that needs submits
(cde, copyengine) can work without syncpoints and sync fences are used
only in userspace.
Fail a submit also if CONFIG_SYNC is not set and sync fences are
requested. Lack of kernel support doesn't guarantee that userspace would
still wrongly want that.
Clarify the deferred cleanup requirements. The sync framework is needed
only for post sync fences, but deferred cleanup is still always needed
with semaphores because the internal tracking is done with dynamically
allocated (although small) objects.
Jira NVGPU-4548
Change-Id: I2e5a6554930cb413b2bb46ddfe388e41390bc7e4
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2321715
(cherry picked from commit d870956170906eae1088846ec05266c859669771)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2318157
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reduce the number of branches and make the code flow more
straightforward by having two complete paths for the gpfifo entry
writes: one when job tracking is done and another when not. Although
this adds some very minor duplication (of the user gpfifo append call),
this way it's easier to read what happens to the job metadata, and when
do we even have one.
Jira NVGPU-4548
Change-Id: I6be8bc5afaf139e7c49d5e44837e04f642dd5721
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2321761
(cherry picked from commit 9a3d3c8d556d563b9d67b370636791d6a1dd57ee)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2324253
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reduce complexity of the big gpfifo submit function by adding another
function to perform channel-global and driver-global sanity checks that
don't depend on submit parameters.
The nvgpu_channel_check_unserviceable() check was in the middle of the
submit function because there used to be a blocking wait just before it
when the hw gpfifo would be full. The blocking wait could exit with the
channel recovered from a timeout. Now it's ok to check this only once in
the beginning because the submit is non-blocking.
Jira NVGPU-4548
Change-Id: Idf19a560ca58a4f7da776c420dc9c6299cd7f7e7
Signed-off-by: Konsta Hölttä <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2321760
(cherry picked from commit 5359a2180f13505f57c62b9f639344913716370a)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2324252
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
- "include/trace/events/gk20a.h" file was having GPL2 license
(which should not used for QNX code). This file was used for
compiling linux userspace driver("libnvgpu-drv.so") and was used for
unit testing on QNX.
- This patch removes stubs in "include/trace/events/gk20a.h" file.
(which were used for linux userspace driver.)
- For QNX driver, "nvgpu_rmos/trace/events/gk20a.h" was used.
This patch moves that file to "include/nvgpu/posix/trace_gk20a.h" and
does relevant license change. This same file will be used for linux
userspace driver.
- This patch also creates a new file "include/nvgpu/trace.h" which
selects proper trace file depending on the config.
Bug 2802414
Change-Id: Icdfb251e5698073f986753a969e804161af3ecc5
Signed-off-by: Vaibhav Kachore <vkachore@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2286388
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
This change eliminates MISRA Advisory Rule 18.4 violations in the
following cases:
* nvgpu_submit_append_gpfifo_user_direct()
* nvgpu_submit_append_gpfifo_common()
- use array-indexing to access gpfifo entry lists
* gv11b_gr_intr_record_sm_error_state()
- use array-indexing to access sm_error_states table
Advisory Rule 18.4 states that the +, -, +=, and -= operators should
not be applied to an expression of pointer type.
JIRA NVGPU-3798
Change-Id: I736930e4ba09a88888b0ef48f62496c4082ea5a1
Signed-off-by: Scott Long <scottl@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2210173
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
CTS test dEQP-VK.api.object_management.max_concurrent.device_group
crashes with invalid userspace memory access.
Currently, nvgpu_submit_prepare_syncs() races with
nvgpu_channel_clean_up_jobs() and this race condition is exposed when
aggressive_sync_destroy_thresh is set to non-zero value.
nvgpu_submit_prepare_syncs() gets ref for c->sync to submit job and
releases channel sync_lock immediately. Meanwhile,
nvgpu_worker_poll_work() triggers nvgpu_channel_clean_up_jobs(), which
destroys ref'd c->sync pointer.
Channel sync is deleted by nvgpu_channel_clean_up_jobs() only if
aggressive_sync_destroy_thresh is non-zero.
So, nvgpu_channel_clean_up_jobs() and nvgpu_submit_prepare_syncs() will
race only in this scenario.
Hence, if aggressive_sync_destroy_thresh value is non-zero, this patch
protects channel's sync pointer by holding channel sync_lock
during complete execution of nvgpu_submit_prepare_syncs().
Bug 2613870
Change-Id: I030d8df7af10d4ed86f921b5cf60de2b1d60e5d3
Signed-off-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2181360
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This patch adds nvgpu API in linux and posix to query vpr resize.
The new API nvgpu_is_vpr_resize_enabled() is used in
nvgpu_submit_channel_gpfifo().
Previously, if non-deterministic channel has timeout disabled and
GPU cannot railgate on some platform, then channel doesn't power ref
count and results in video freeze. To resolve non-determinstic channel
job tracking needs to be enabled if vpr resize is supported or if GPU
can railgate.
Bug 200532122
Change-Id: Icfbff6253762b195b2f5955749343974b1a7a269
Signed-off-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2171093
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
CTS test dEQP-VK.api.object_management.max_concurrent.device_group
crashes with invalid userspace memory access.
Currently, nvgpu_submit_prepare_syncs() races with
nvgpu_channel_clean_up_jobs() and this race condition is exposed when
aggressive_sync_destroy_thresh is set to non-zero value.
nvgpu_submit_prepare_syncs() gets ref for c->sync to submit job and
releases channel sync_lock. Meanwhile, nvgpu_worker_poll_work()
triggers nvgpu_channel_clean_up_jobs(), which destroys ref'd c->sync
pointer.
This patch protects channel's sync pointer by holding channel sync_lock
during complete execution of nvgpu_submit_prepare_syncs().
Bug 2613870
Change-Id: I6f3d48aff361d1cb38c30d2ce5de276d0c55fb6f
Signed-off-by: Seema Khowala <seemaj@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2176929
Reviewed-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-by: Debarshi Dutta <ddutta@nvidia.com>
GVS: Gerrit_Virtual_Submit
Tested-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Safety build does not support vidmem. This patch compiles out vidmem
related changes - vidmem, dma alloc, cbc/acr/pmu alloc based on
vidmem and corresponding tests like pramin, page allocator &
gmmu_map_unmap_vidmem..
As vidmem is applicable only in case of DGPUs the code is compiled
out using CONFIG_NVGPU_DGPU.
JIRA NVGPU-3524
Change-Id: Ic623801112484ffc071195e828ab9f290f945d4d
Signed-off-by: Sagar Kamble <skamble@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2132773
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vaibhav Kachore <vkachore@nvidia.com>
Reviewed-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Renamed gk20a_channel_* APIs to nvgpu_channel_* APIs.
Removed unused channel API int gk20a_wait_channel_idle
Renamed nvgpu_channel_free_usermode_buffers in os/linux-channel.c to
nvgpu_os_channel_free_usermode_buffers to avoid conflicts with the API
with the same name in channel unit.
Jira NVGPU-3248
Change-Id: I21379bd79e64da7e987ddaf5d19ff3804348acca
Signed-off-by: Debarshi Dutta <ddutta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2121902
Reviewed-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Check return value of below function and add void to ignore
the return value
update_gp_get
Rename
nvgpu_get_gp_free_count -> nvgpu_channel_update_gpfifo_get_and_get_free_count
nvgpu_gp_free_count -> nvgpu_channel_get_gpfifo_free_count
JIRA NVGPU-3388
Change-Id: I6e2265882c1f34e3bb47eaeac7a2c5a9fbe9b4eb
Signed-off-by: Seema Khowala <seemaj@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2115784
Reviewed-by: Thomas Fleury <tfleury@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Renamed local variable
- sync_fence -> flag_sync_fence
As "sync_fence" is already used to represent a type.
Also, renamed for consistency:
- fence_wait -> flag_fence_wait
- fence_get -> flag_fence_get
Jira NVGPU-3384
Change-Id: Ib40d068f0ebda985303a85a385f4123955d3b4c9
Signed-off-by: Thomas Fleury <tfleury@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2111613
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
(!c->deterministic) ||
(nvgpu_is_enabled(g, NVGPU_CAN_RAILGATE) && !c->deterministic)
is equivalent to
(!c->deterministic)
Remove second condition in nvgpu_submit_channel_gpfifo
Jira NVGPU-3384
Change-Id: Icf3e460d4fe9d310d94a21895832bbfae595df28
Signed-off-by: Thomas Fleury <tfleury@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2111610
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
