This adds APIs to the POSIX sgt module to allow unit tests to create
sgt's and sgl's more easily. The new APIs allow the caller to pass in a
table of sgl's and creates the new sgt/sgl as needed.
Since this there is a new API to create an sgl, this also includes a new
nvgpu_sgl_free() API.
JIRA NVGPU-1563
Change-Id: I27ab7654289cbd2212a75625b6dd24d2b742e3bf
Signed-off-by: Philip Elcan <pelcan@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1966151
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@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>
Add separate new unit gr/ctxsw_prog that provides interface to access
h/w header files hw_ctxsw_prog_*.h
Add below chip specific files that access above h/w unit and provide
interface through g->ops.gr.ctxsw_prog.*() HAL for rest of the units
common/gr/ctxsw_prog/ctxsw_prog_gm20b.c
common/gr/ctxsw_prog/ctxsw_prog_gp10b.c
common/gr/ctxsw_prog/ctxsw_prog_gv11b.c
Remove all the h/w header includes from rest of the units and code.
Remove direct calls to h/w headers ctxsw_prog_*() and use HALs
g->ops.gr.ctxsw_prog.*() instead
In gr_gk20a_find_priv_offset_in_ext_buffer(), h/w header
ctxsw_prog_extended_num_smpc_quadrants_v() is only defined on gk20a
And since we don't support gk20a remove corresponding code
Add missing h/w header ctxsw_prog_main_image_pm_mode_ctxsw_f() for
some chips
Add new h/w header ctxsw_prog_gpccs_header_stride_v()
Jira NVGPU-1526
Change-Id: I170f5c0da26ada833f94f5479ff299c0db56a732
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1966111
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
MISRA Rule 10.4 only allows the usage of arithmetic operations on
operands of the same essential type category.
Adding "U" at the end of the integer literals or casting operands
to have same type of operands when an arithmetic operation is
performed.
This fixes violations where an arithmetic operation is performed on
signed and unsigned int types.
JIRA NVGPU-992
Change-Id: I27e3e59c3559c377b4bd3cbcfced90fdf90350f2
Signed-off-by: Sai Nikhil <snikhil@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1921459
Reviewed-by: svc-misra-checker <svc-misra-checker@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Adeel Raza <araza@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This is part of a move to 64KiB for usermode mapping to fix failures
when the system page size is 64KiB. When remapping or zapping the
vma, use the existing size, not hardcoded size. Also change the
verification of the size when creating the mapping to verify it is at
least as big as PAGE_SIZE. This allows 4KiB mappings to continue to
work until nvrm_gpu is changed to use 64KiB mappings.
Bug 2441531
Change-Id: I447ef8e9f84e6d70bbe96b527e267ec41c5630b8
Signed-off-by: Allen Martin <amartin@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1964687
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@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>
With MSI enabled and kernel 4.14, there are some occurrences of
interrupts not being launched (no issue on kernel 4.9). If
an interrupt is not served for a software method, we end up most
of the time with PBDMA busy, while engine is idle.
Disable MSI for dGPU on kernel 4.14
Bug 200460636
Change-Id: I3c5657d63195de5c714f44d331ac992199811d9f
Signed-off-by: Thomas Fleury <tfleury@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1968073
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Deepak Nibade <dnibade@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>
MISRA Rule 10.1 states that operands shall not be of an
inappropriate essential type.
For example, the use of bitwise OR on signed values is not
permitted.
This patch restructures return status assignment in the
gr_tu104_init_sw_bundle64() routine to eliminate the use
of bitwise OR when constructing the return value.
Rather than continue to attempt to submit bundles after an
idle failure the routine instead now returns immediately on
the first error.
This seems like a reasonable/better approach given the contents
of the context buffer are in a questionable state after the
first error anyway and the client that issued ALLOC_OBJ_CTX will
continue to properly see the error.
Also, this patch changes the type of the return variable
('err') from u32 to int to match the function prototype.
JIRA NVGPU-650
Change-Id: I2aae0f85d0373ddf95b5899ce8c452352c6aa670
Signed-off-by: Scott Long <scottl@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1965496
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Deepak Nibade <dnibade@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>
In nvgpu_sim_esc_readl() we prepare a message, issue RPC and then copy back
the response from offset "data_offset + 0xc"
But while configuring the message we incorrectly set the response offset as
just "data_offset"
Fix this by correctly configuring the response offset as "data_offset + 0xc"
Jira NVGPUT-41
Change-Id: I855e140b97e7128367446d0962eec283b069f514
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1756844
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
If enabled, fecs trace updating happens from ucode
side even when there is no fecs trace dumper application
to consume it. Due to this, trace buffer will get
eventually full and ucode will trigger ctxsw_intr0.
Reset fecs_trace buffer to handle timestamp buffer full
ctxsw_intr0.
Bug 2361571
Bug 200472922
Change-Id: Ia26a17635fc6bd6e8663b8af983acc91839ecfcd
Signed-off-by: Seema Khowala <seemaj@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1965370
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
ctxsw_intr0 is triggered by ucode even if it
is not enabled by driver. Add handling
for processing ctxsw_intr0. fecs mailbox(6)
is used to report fecs/gpccs misc error codes.
Also dump falcon stats for unhandled fecs intr.
Bug 2361571
Bug 200472922
Change-Id: Iefb3c0d46ad1d08db07fd3c08cff91a77835908c
Signed-off-by: Seema Khowala <seemaj@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1966984
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
MISRA Rule 5.7 Definition: A tag name shall be a unique identifier.
Rule 5.7 violations can occur if a variable has the same name as a
struct. bios.c defines a "struct bit". "bit" is a very common name for
variables which causes a name conflict with "struct bit". Therefore,
change "struct bit" to "struct bios_bit" to resolve rule 5.7 violations.
Jira NVGPU-845
Change-Id: I02f2fa6cf1701c7de2e365635b18ecee0f8296fa
Signed-off-by: Adeel Raza <araza@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1965693
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: svc-misra-checker <svc-misra-checker@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Add a flag that let's userspace enable the unified VM functionality
on a selective bassis. This feature is working for all cases except
a single MODS trace. This will allow test coverage to be selectively
added in certain userspace tests as well to help prevent this feature
from bit rotting (as it has historically done).
Also update the unit test for the page table management in the GMMU
to reflect this new flag. It's been set to false since the target
platform for safety is currently not using unified address spaces.
Bug 200438879
Change-Id: Ibe005472910d1668e8372754be8dd792773f9d8c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1951864
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
The pd_cache header declarations were oriignally part of the
gmmu.h header. This is not good from a unit isolation perspective
so this patch moves all the pd_cache specifics over to a new
header file: <nvgpu/pd_cache.h>.
Also a couple of static inlines that were possible when the code
was part of gmmu.h were turned into real, first class functions.
This allowed the pd_cache.h header to not include the gmmu.h
header file.
Also fix an issue in the nvgpu_pd_write() function where the data
was being passed as a size_t for some reason. This has now been
changed to a u32.
JIRA NVGPU-1444
Change-Id: Ib9e9e5a54544de403bfcd8e11c30de05721ddbcc
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1966352
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
MISRA Rule 10.3 states that the value of an expression shall not
be assigned to an object with a narrower essential type or of a
different esseential type category.
For example, assigning an unsigned 32bit value (u32) to a signed
32bit value (int) is not permitted.
This patch modifies the gr_gk20a_init_golden_ctx_image() and
gk20a_init_sw_bundle() routines to use an int (instead of u32)
for return status handling making them consistent with the other
gr routines used in this part of the gr object allocation path.
JIRA NVGPU-647
Change-Id: I53c47d9a169bd0d4cdbce107bd4ad8e7978ae01d
Signed-off-by: Scott Long <scottl@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1965735
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-misra-checker <svc-misra-checker@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
MISRA Rule 11.8 states that a cast shall not remove any const or
volatile qualification from the type pointed to by a pointer.
The linux kernel's container_of() macro contains such a violation as
it generates a pointer to a caller-specified (and so possibly non-const
qualified) type by casting an internally declared const pointer.
The gk20a_from_pmu() uses the container_of() macro to convert
from a struct nvgpu_pmu pointer to a struct gk20a pointer.
The struct nvgpu_pmu has a back pointer to struct gk20a already
however and so this change modifies gk20a_from_gpu() to just
return this back pointer rather than use container_of().
JIRA NVGPU-862
Change-Id: If0e2481c1cf104c2fa6b89334e20e75705bf9c44
Signed-off-by: Scott Long <scottl@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1955540
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
All the 16.x MISRA rules are relevant to switch statement
formatting and hence addressed in single patch
As per MISRA 16.1, all switch statements should be well formatted.
16.3 fixes:
Add unconditional break statements to all the switch-clauses
to adhere to MISRA rule 16.3. Also do not allow fall-through
(even the intentional ones) from one switch-clause to next one.
16.4 fixes:
Make sure all "default" clauses in the switch statements are
non-empty.
16.6 fixes:
Fix all switch statement formatting to fix MISRA 16.6 violations
which requires all the switch clauses to be conforming.
JIRA NVGPU-1496
JIRA NVGPU-1533
JIRA NVGPU-1550
JIRA NVGPU-1558
Change-Id: I7f373e99491eb860ca7c9edfeb43a32ad0b07baa
Signed-off-by: tkudav <tkudav@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1961694
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Adeel Raza <araza@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vijayakumar Subbu <vsubbu@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
preempt_channel needs to use the channel to pass it to other
public functions, get access to a tsg etc. This qualifies it to take a
pointer to a channel as an input parameter instead of a chid.
Increment the channel ref counter using the function
gk20a_channel_from_id in functions where we get the chid from the h/w
registers directly. Once the prempt_channel function call is done,
use a gk20a_channel_put on the referenced channel.
Jira NVGPU-1461
Change-Id: I6c87c8104cfcb418d468c8c590087fd4aeabf4bd
Signed-off-by: Debarshi Dutta <ddutta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1963200
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
gk20a_fifo_recover_channel takes a reference to the channel via its
chid before passing the channel pointer to other public functions such
as gk20a_channel_abort and gk20a_fifo_error_ch. This qualifies the
gk20a_fifo_recover_channel to take a pointer to a channel instead of
only chid.
Jira NVGPU-1461
Change-Id: I338a12a05e5ccee785a202fea7848db5201a3a39
Signed-off-by: Debarshi Dutta <ddutta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1963199
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Fifo scheduling APIs require the HW reg mask accessor
fifo_sched_disable_runlist_m() to be used even from high-level logic.
Restructure the APIs to take in an explicit bitmap of runlist IDs and
translate the bitmap to units of fifo_sched_disable_runlist_m() (which
happens to be an identical bitmap) only just before accessing hardware.
Jira NVGPU-1309
Change-Id: I5d6ce5b719ef467172c07c8d7589d83942365025
Signed-off-by: Konsta Holtta <kholtta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1960225
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This reverts commit 15603b9fd5.
Causes a build break in the PD cache unit test. Not sure how this
passed GVS - must have been a race or something? Unclear.
Change-Id: Ia484a801d098d69441326fa1dd40a1c86e2e23ce
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1966335
The pd_cache header declarations were originally part of the
gmmu.h header. This is not good from a unit isolation perspective
so this patch moves all the pd_cache specifics over to a new
header file: <nvgpu/pd_cache.h>.
Also a couple of static inlines that were possible when the code
was part of gmmu.h were turned into real, first class functions.
This allows the pd_cache.h header to not include the gmmu.h
header file.
Also fix an issue in the nvgpu_pd_write() function where the data
was being passed as a size_t for some reason. This has now been
changed to a u32.
JIRA NVGPU-1444
Change-Id: Iead9a0d998396d2289ffcb3b48765d770400397b
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1965271
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
MISRA Rule 10.4 only allows the usage of arithmetic operations on
operands of the same essential type category.
Adding "U" at the end of the integer literals to have same type of
operands when an arithmetic operation is performed.
This fixes violations where an arithmetic operation is performed on
signed and unsigned int types.
JIRA NVGPU-992
Change-Id: I3b725e60f1908a4b3a308736d02600f86929cdd3
Signed-off-by: Sai Nikhil <snikhil@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1958306
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Adeel Raza <araza@nvidia.com>
Reviewed-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vijayakumar Subbu <vsubbu@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
The switch statement "switch (interleave_level)" has no conforming
switch clauses as none of the clauses end with unconditional break
statement.
The above switch statement is now fixed in accordance to misra-c
standards.
Jira NVGPU-1555
Change-Id: Id2ea98826b5fff51f42eed83a597d8e0e273ebde
Signed-off-by: Debarshi Dutta <ddutta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1962545
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
