IRQs can get triggered during nvgpu power-on due to MMU fault, invalid
PRIV ring or bus access etc. Handlers for those IRQs can't access the
full state related to the IRQ unless nvgpu is fully powered on.
In order to let the IRQ handlers know about the nvgpu power-on state
gk20a.power_on_state variable has to be protected through spinlock
to avoid the deadlock due to usage of earlier power_lock mutex.
Further the IRQs need to be disabled on local CPU while updating the
power state variable hence use spin_lock_irqsave and spin_unlock_-
irqrestore APIs for protecting the access.
JIRA NVGPU-1592
Change-Id: If5d1b5e2617ad90a68faa56ff47f62bb3f0b232b
Signed-off-by: Sagar Kamble <skamble@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2203860
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This patch adds unit testing for the mm.as unit including:
- feature tests
- error injection testing
- 100% line coverage and 96% branches (one missing branch that
cannot be tested or removed)
JIRA NVGPU-917
Change-Id: I54bdac21e56554d1d960955f1a140ab98c9f3e5e
Signed-off-by: Nicolas Benech <nbenech@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2194399
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>
While fixing CERT-C issues in the bitmap_allocator, a number of
asserts were added. In particular, if blk_size is 0 it will cause
a BUG(). Within the bitmap_allocator unit test, the init will try
to set blk_size to 0 to make sure it fails. Now the failure is a
BUG(), not a return code. So this fix adds EXPECT_BUG where needed.
JIRA NVGPU-906
Change-Id: I62ae725e494a7cf561ac8cac7685cffbbc8de38e
Signed-off-by: Nicolas Benech <nbenech@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2135823
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>
A separate bug caused EXPECT_BUG to always return true without actually
calling the corresponding test code. This hid some issues in the buddy
allocator unit where the fini() operation was called several times while
expecting it to call BUG(). Doing so caused the mutex unlock operation to
not be called, which caused a deadlock for all subsequent calls. The fix
is to explicitly release the mutex after each call to fini() that expects
a BUG().
JIRA NVGPU-3562
Change-Id: Ic26058a272c616d2a6052d319f38a4d4dc33ef1c
Signed-off-by: Nicolas Benech <nbenech@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2126874
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>
Unit tests in page_table were initializing the VM with a kernel_reserved
section that was slightly smaller than expected. Then, when the fixed
alloc to support semaphores was done, the fixed address used was actually
below the start of the kernel address space. As a result this caused an
overflow in the base shift in the buddy allocator responsible for
managing the fixed alloc.
Change-Id: I4e688d418262ac8d9d4b66b46bd32ca5456d95e8
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2130433
Reviewed-by: Philip Elcan <pelcan@nvidia.com>
Tested-by: Philip Elcan <pelcan@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This fixes MISRA rule 4.7 violations in the function nvgpu_vm_map(). The
violations were caused by trying to use ERR_PTR() to return error
information. Rather than try to return errors in a pointer, just change
the API to return an int and pass the pointer the arguments.
JIRA NVGPU-3332
Change-Id: I2852a6de808d9203b8c7826e2b8211bab97ccd16
Signed-off-by: Philip Elcan <pelcan@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2114027
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vinod Gopalakrishnakurup <vinodg@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
MISRA Rule 11.2 doesn't allow conversion to or from an incomplete type
pointer, as it may result incorrect point alignment and may further lead
to undefined behavior.
MISRA Rule 16.x requires all switch statements to be well-formed with
terminating break statement for every switch-clause.
This patch fixes 11.2 and 16.x violations in common.mm.nvgpu_mem.
Jira NVGPU-3339
Change-Id: I002393cc64d44826e6954d1bf6af71bd569e862f
Signed-off-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2113096
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
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>
MISRA rule 21.3 forbids from using calloc, malloc, realloc and free
identifiers for function or macro names. This patch renames nvgpu
allocator free operator to free_alloc to follow rule 21.3.
Jira NVGPU-3336
Change-Id: Ie9f48d567255a3e1dca70632fbe3d36b45023f3f
Signed-off-by: Vedashree Vidwans <vvidwans@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2111365
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
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>
Move the gv11b MMU fault handling code into a new mm.mmu_fault HAL.
Also move the existing gmmu_mmu_fault HAL code into this HAL as they
are basically the same logical entity.
JIRA NVGPU-2042
JIRA NVGPU-1313
Change-Id: I41d3e180c762f191d4de3237e9052bdc456f9e4c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2109693
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
There are many miscellaneous HALs for various MM related functionality.
This patch aims to migrate all the remaining MM code from the <chip>/
mm_<chip>.[ch] files in HAL files under hal/.
Much of this is fairly straightforward copy/paste and updates to the
HAL init files.
The exception to that is the move of the left over gv11b MMU fault
handling code in mm_gv11b.c. Having both a hal/mm/mm/mm_gv11b.c and
a gv11b/mm_gv11b.c file causes tmake to choke so the gv11b/mm_gv11b.c
file was moved to gv11b/mmu_fault_gv11b.c. This will be cleaned up in
a subsequent patch.
JIRA NVGPU-2042
Change-Id: I12896de865d890a61afbcb71159cff486119ffb8
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2109050
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
The alloc_insty_block() function in the MM HAL is not a HAL. It does
not abstract any HW accesses; instead it just wraps a dma allocation.
As such remove it from the HAL and move the single gk20a implementation
to common/mm/mm.c as nvgpu_alloc_inst_block().
JIRA NVGPU-2042
Change-Id: I0a586800a11cd230ca43b85f94a35de107f5d1e1
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2109049
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Move the gk20a_init_mm_setup_hw to common code. This function just
calls HALs to initialize various bits of HW. However, since this
code assumes that (some of) the HALs are present this code is treated
as a vGPU HAL (that is it's NULL for vGPU).
This patch also renames the MM HW init HAL.
Sicne the gv11b variant of this setup_hw HAL did have some differences
from the gk20a version the new common version required some work. The
gv11b code was copied into the common function but now the gv11b
specific calls were HAL'ified (since they will need to be anyway for
the MMU fault buf code) and protected by an if-condition.
JIRA NVGPU-2042
Change-Id: I959293394ee8793923d858ee37b7ce346096654f
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2104541
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Move the remaining GMMU HAL related code from the gm20b/, gp10b/,
and gv11b/ directories to new gmmu hal source files.
Also update all makefiles and HAL init code to refelct the new
location of the headers and source code.
JIRA NVGPU-2042
Change-Id: Ic9b85cc547bd0f994ad11042fc4093c517327399
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2103672
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Make a hal/mm/gmmu sub-unit for the GMMU HAL code. Also move the
gk20a specific HAL code there. gp10b will happen in the next patch.
This change also updates all the GMMU related HAL usage, of which
there is quite a bit. Generally the only change is a .gmmu needs to
be inserted into the HAL path. Each HAL init was also updated.
JIRA NVGPU-2042
Change-Id: I6c46bdfddb8e021f56103d9457fb3e2a226f8947
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2099693
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Rename the two native GPU GMMU map/unmap functions and update the
HAL initializations to reflect this:
gk20a_locked_gmmu_map -> nvgpu_gmmu_map_locked
gk20a_locked_gmmu_unmap -> nvgpu_gmmu_unmap_locked
This matches what other units do for handling vGPU "HAL" indirection.
Also move the function declarations to <nvgpu/gmmu.h> since these are
shared among all non-vGPU chips. But since these are still technically
HAL operations they should never be called directly. This is a bit of
an organixational issue that I have not thought through hwo to solve
yet.
Ideally they would go into a "hal/mm/gmmu/" include somewhere, but
that A) doesn't yet exist, and B) those are chip specific; these
functions are native specific. Ugh.
JIRA NVGPU-2042
Change-Id: Ibc614f2928630d12eafcec6ce73019628b44ad94
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2099692
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com>
Reviewed-by: Seshendra Gadagottu <sgadagottu@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Add a new MM HAL directory to contain all MM related HAL units.
As part of this change add cache unit to the MM HAL. This contains
several related fixes:
1. Move the cache code in gk20a/mm_gk20a.c and gv11b/mm_gv11b.c to
the new cache HAL. Update makefiles and header includes to take
this into account. Also rename gk20a_{read,write}l() to their
nvgpu_ variants.
2. Update the MM gops: move the cache related functions to the new
cache HAL and update all calls to this HAL to reflect the new
name.
3. Update some direct calls to gk20a MM cache ops to pass through
the HAL instead.
4. Update the unit tests for various MM related things to use the
new MM HAL locations.
This change accomplishes two architecture design goals. Firstly it
removes a multiple HW include from mm_gk20a.c (the flush HW header).
Secondly it moves code from the gk20a/ and gv11b/ directories into
more proper locations under hal/.
JIRA NVGPU-2042
Change-Id: I91e4bdca4341be4dbb46fabd72622b917769f4a6
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2095749
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Moved
-mmu_fault_pending mm ops to is_mmu_fault_pending mc ops
-mmu_fault_pending fb ops to is_mmu_fault_pending fb.intr ops. This
is needed to check if mmu fault intr is pending for volta onwards.
Added
is_mmu_fault_pending fifo ops. This is needed to check if mmu fault
interrupt is pending for chips prior to volta
JIRA NVGPU-1313
Change-Id: Ie8e778387cd486cb19b18c4aee734c581dcd9229
Signed-off-by: Seema Khowala <seemaj@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2094895
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Moved fb interrupt handling related code to fb intr sub-unit.
Moved following hals from fb hal to fb intr hal and renamed to:
void (*enable)(struct gk20a *g);
void (*disable)(struct gk20a *g);
void (*isr)(struct gk20a *g);
gk20a_readl/writel are replaced with nvgpu_read/writel.
Hals are populated with new function names and code is modified
to call new hals.
Moved ecc interrupt to gv11b_fb_intr_handle_ecc in a separate file:
fb_intr_ecc_gv11b.c/h
JIRA NVGPU-2034
Change-Id: I80c7110c902c4e082561cf7cbe65c20eb9acb661
Signed-off-by: Seshendra Gadagottu <sgadagottu@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2090070
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Added the following HALs
- ramin.base_shift
- ramin.alloc_base
Use above HALs in mm, instead of using hw definitions.
Defined nvgpu_inst_block_ptr to
- get inst_block address,
- shift if by base_shift
- assert upper 32 bits are 0
- return lower 32 bits
Added missing #include for <nvgpu/mm.h>
Jira NVGPU-3015
Change-Id: I558a6f4c9fbc6873a5b71f1557ea9ad8eae2778f
Signed-off-by: Thomas Fleury <tfleury@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2077840
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Add -t/--test-level option for the unit test framework. This
correlates with the test plan levels in GVS. L0 tests are a subset of
the L1 tests. Currently, only test levels 0 and 1 are supported.
The test output has been updated to display skipped tests. Skipped
tests are not included in the test dump.
JIRA NVGPU-2251
Change-Id: Icc2ff19a81529be8526e89f70983e96040390e00
Signed-off-by: Philip Elcan <pelcan@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2085280
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
The licenses used in the userspace folder were inconsistent.
This patch fixes the situation by ensuring all files are using the
MIT license.
JIRA NVGPU-2974
Change-Id: I8c89667f72732ff6f73e6cf4be1acd9e58a7e516
Signed-off-by: Nicolas Benech <nbenech@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2087125
GVS: Gerrit_Virtual_Submit
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
