| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Rename get_physical_addr_bits and related functions to something that
more clearly conveys what they are doing. The basic idea of these
functions is to translate from a physical GPU address to a IOMMU GPU
address. To do that a particular bit (that varies from chip to chip)
is added to the physical address.
JIRA NVGPU-68
Change-Id: I536cc595c4397aad69a24f740bc74db03f52bc0a
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1542966
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Change license of OS independent source code files to MIT.
JIRA NVGPU-218
Change-Id: I1474065f4b552112786974a16cdf076c5179540e
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1565880
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The basic nvgpu_mem_sgl implementation provides support
for OS specific scatter-gather list implementations by
simply copying them node by node. This is inefficient,
taking extra time and memory.
This patch implements an nvgpu_mem_sgt struct to act as
a header which is inserted at the front of any scatter-
gather list implementation. This labels every struct
with a set of ops which can be used to interact with
the attached scatter gather list.
Since nvgpu common code only has to interact with these
function pointers, any sgl implementation can be used.
Initialization only requires the allocation of a single
struct, removing the need to copy or iterate through the
sgl being converted.
Jira NVGPU-186
Change-Id: I2994f804a4a4cc141b702e987e9081d8560ba2e8
Signed-off-by: Sunny He <suhe@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1541426
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The last major item preventing the core MM code in the nvgpu
driver from being platform agnostic is the usage of Linux
scattergather tables and scattergather lists. These data
structures are used throughout the mapping code to handle
discontiguous DMA allocations and also overloaded to represent
VIDMEM allocs.
The notion of a scatter gather table is crucial to a HW device
that can handle discontiguous DMA. The GPU has a MMU which
allows the GPU to do page gathering and present a virtually
contiguous buffer to the GPU HW. As a result it makes sense
for the GPU driver to use some sort of scatter gather concept
so maximize memory usage efficiency.
To that end this patch keeps the notion of a scatter gather
list but implements it in the nvgpu common code. It is based
heavily on the Linux SGL concept. It is a singly linked list
of blocks - each representing a chunk of memory. To map or
use a DMA allocation SW must iterate over each block in the
SGL.
This patch implements the most basic level of support for this
data structure. There are certainly easy optimizations that
could be done to speed up the current implementation. However,
this patches' goal is to simply divest the core MM code from
any last Linux'isms. Speed and efficiency come next.
Change-Id: Icf44641db22d87fa1d003debbd9f71b605258e42
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1530867
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove the mm.get_iova_addr() HAL and replace it with a new HAL
called mm.gpu_phys_addr(). This new HAL provides the real phys
address that should be passed to the GPU from a physical address
obtained from a scatter list. It also provides a mechanism by
which the HAL code can add extra bits to a GPU physical address
based on the attributes passed in. This is necessary during GMMU
page table programming.
Also remove the flags argument from the various address functions.
This flag was used for adding an IO coherence bit to the GPU
physical address which is not supported.
JIRA NVGPU-30
Change-Id: I69af5b1c6bd905c4077c26c098fac101c6b41a33
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1530864
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Support only VM pointers and ref-counting for maintaining VMs. This
dramatically reduces the complexity of the APIs, avoids the API
abuse that has existed, and ensures that future VM usage is
consistent with current usage.
Also remove the combined VM free/instance block deletion. Any place
where this was done is now replaced with an explict free of the
instance block and a nvgpu_vm_put().
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: Ib73e8d574ecc9abf6dad0b40a2c5795d6396cc8c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1480227
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Unify the initialization routines for the vGPU and regular GPU paths.
This helps avoid any further code divergence. This also assumes that
the code running on the regular GPU essentially works for the vGPU.
The only addition is that the regular GPU path calls an API in the
vGPU code that sends the necessary RM server message.
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: I37af1993fd8b50f666ae27524d382cce49cf28f7
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1480226
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
nvgpu_log/info/warn/err() internally add a \n to the end of the message.
Hence, callers should not include a \n at the end of the message. Doing
so results in duplicate \n being printed, which ends up creating empty
log messages. Remove the duplicate \n from all err/warn messages.
Bug 1928311
Change-Id: I99362c5327f36146f28ba63d4e68181589735c39
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-on: http://git-master/r/1487232
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove the VM init/de-init from the HAL and instead use a single
set of routines that init/de-init VMs. This prevents code divergence
between vGPUs and regular GPUs.
This patch also clears up the naming of the routines a little bit.
Since some VMs are used inplace and others are dynamically allocated
the APIs for freeing them were confusing. Also some free calls also
clean up an instance block (this is API abuse - but this is how it
currently exists).
The new API looks like this:
void __nvgpu_vm_remove(struct vm_gk20a *vm);
void nvgpu_vm_remove(struct vm_gk20a *vm);
void nvgpu_vm_remove_inst(struct vm_gk20a *vm,
struct nvgpu_mem *inst_block);
void nvgpu_vm_remove_vgpu(struct vm_gk20a *vm);
int nvgpu_init_vm(struct mm_gk20a *mm,
struct vm_gk20a *vm,
u32 big_page_size,
u64 low_hole,
u64 kernel_reserved,
u64 aperture_size,
bool big_pages,
bool userspace_managed,
char *name);
void nvgpu_deinit_vm(struct vm_gk20a *vm);
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: Ia4016384c54746bfbcaa4bdd0d29d03d5d7f7f1b
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1477747
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Refactor the API for initializing and cleaning up VMs.
This also involved moving a bunch of GMMU code out into the
gmmu code since part of initializing a VM involves initializing
the page tables for the VM.
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: I4710f08c26a6e39806f0762a35f6db5c94b64c50
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1477746
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This function is an internal function to the VM manager that allocates
virtual memory space in the GVA allocator. It is unfortunately used in
the vGPU code, though. In any event, this patch cleans up and moves the
implementation of these functions into the VM common code.
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: I24a3d29b5fcb12615df27d2ac82891d1bacfe541
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1477745
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The vm_reserve_va_node struct is essentially a special VM area that
can be used for sparse mappings and fixed mappings. The name of this
struct is somewhat confusing (as node is typically used for list
items). Though this struct is a part of a list it doesn't really
make sense to call this a list item since it's much more. Based on
that the struct has been renamed to nvgpu_vm_area to capture the
actual use of the struct more accurately.
This also moves all of the management code of vm areas to a new file
devoted solely to vm_area management.
Also add a brief overview of the VM architecture. This should help
other people follow along the hierachy of ownership and lifetimes in
the rather complex MM code.
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: If85e1cf868031d0dc265e7bed50b58a2aed2602e
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1477744
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch begins splitting out the VM implementation from mm_gk20a.c and
moves it to common/linux/vm.c and common/mm/vm.c. This split is necessary
because the VM code has two portions: first, an interface for the OS
specific code to use (i.e userspace mappings), and second, a set of APIs
for the driver to use (init, cleanup, etc) which are not OS specific.
This is only the beginning of the split - there's still a lot of things
that need to be carefully moved around.
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: I3b57cba245d7daf9e4326a143b9c6217e0f28c96
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1477743
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch begins the major rework of the GPU's virtual memory manager
(VMM). The VMM is the piece of code that handles the userspace interface
to buffers and their mappings into the GMMU. The core data structure is
the VM - for now still known as 'struct vm_gk20a'. Each one of these
structs represents one addres space to which channels or TSGs may bind
themselves to.
The VMM splits the interface up into two broad categories. First there's
the common, OS independent interfaces; and second there's the OS specific
interfaces.
OS independent
--------------
This is the code that manages the lifetime of VMs, the buffers inside
VMs (search, batch mapping) creation, destruction, etc.
OS Specific
-----------
This handles mapping of buffers represented as they are represented by
the OS (dma_buf's for example on Linux).
This patch is by no means complete. There's still Linux specific functions
scattered in ostensibly OS independent code. This is the first step. A
patch that rewrites everything in one go would simply be too big to
effectively review.
Instead the goal of this change is to simply separate out the basic
OS specific and OS agnostic interfaces into their own header files. The
next series of patches will start to pull the relevant implementations
into OS specific C files and common C files.
JIRA NVGPU-12
JIRA NVGPU-30
Change-Id: I242c7206047b6c769296226d855b7e44d5c4bfa8
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1464939
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add wrapper header file nvgpu/bug.h. It #includes <linux/bug.h>
in Linux.
JIRA NVGPU-13
Change-Id: I7bf02ba554333f7cbd79d72bd1cb423c81ebcb49
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/1461545
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
gk20a_err() and gk20a_warn() require a struct device pointer,
which is not portable across operating systems. The new nvgpu_err()
and nvgpu_warn() macros take struct gk20a pointer. Convert code
to use the more portable macros.
JIRA NVGPU-16
Change-Id: I071e8c50959bfa81730ca964d912bc69f9c7e6ad
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/1457355
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Make an nvgpu DMA API include file so that the intricacies of the
Linux DMA API can be hidden from the calling code.
Also document the nvgpu DMA API.
JIRA NVGPU-12
Change-Id: I7578e4c726ad46344b7921179d95861858e9a27e
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1323326
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Renaming was done with the following command:
$ find -type f | \
xargs sed -i 's/struct mem_desc/struct nvgpu_mem/g'
Also rename mem_desc.[ch] to nvgpu_mem.[ch].
JIRA NVGPU-12
Change-Id: I69395758c22a56aa01e3dffbcded70a729bf559a
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1325547
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Rename the functions used for mem_desc access to nvgpu_mem_*.
JIRA NVGPU-12
Change-Id: Ibfdc1112d43f0a125e4487c250e3f977ffd2cd75
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1323325
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use nvgpu rbtree instead of linux rbtree to store
mapped buffers for each VM
Move to use "struct nvgpu_rbtree_node" instead of
"struct rb_node"
And similarly use rbtree APIs from <nvgpu/rbtree.h>
instead of linux APIs
Jira NVGPU-13
Change-Id: Id96ba76e20fa9ecad016cd5d5a6a7d40579a70f2
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1453043
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use nvgpu list APIs instead of linux list APIs
for reserved VA list and buffer VA list
Jira NVGPU-13
Change-Id: I83c02345d54bca03b00270563567227510cfce6b
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1454013
Reviewed-by: svccoveritychecker <svccoveritychecker@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Use the new kmem API functions in vgpu/*. Also reshuffle the order
of some allocs in the vgpu init code to allow usage of the nvgpu
kmem APIs.
Bug 1799159
Bug 1823380
Change-Id: I6c6dcff03b406a260dffbf89a59b368d31a4cb2c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1318318
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move all programming of FB to fb_*.c files, and remove the inclusion
of FB hardware headers from other files.
TLB invalidate function took previously a pointer to VM, but the new
API takes only a PDB mem_desc, because FB does not need to know about
higher level VM.
GPC MMU is programmed from the same function as FB MMU, so added
dependency to GR hardware header to FB.
GP106 ACR was also triggering a VPR fetch, but that's not applicable
to dGPU, so removed that call.
Change-Id: I4eb69377ac3745da205907626cf60948b7c5392a
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/1321516
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
API DEFINE_MUTEX() is defined in Linux and might
not be available in other OSs.
Hence remove its usage from nvgpu
Declare and explicitly initialize below mutexes
for both nvgpu and vgpu
g->mm.priv_lock
g->mm.tlb_lock
Jira NVGPU-13
Change-Id: If72885a6da0227a1552303206172f1f2b751471d
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1298042
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of using Linux APIs for mutex and spinlocks
directly, use new APIs defined in <nvgpu/lock.h>
Replace Linux specific mutex/spinlock declaration,
init, lock, unlock APIs with new APIs
e.g
struct mutex is replaced by struct nvgpu_mutex and
mutex_lock() is replaced by nvgpu_mutex_acquire()
And also include <nvgpu/lock.h> instead of including
<linux/mutex.h> and <linux/spinlock.h>
Add explicit nvgpu/lock.h includes to below
files to fix complilation failures.
gk20a/platform_gk20a.h
include/nvgpu/allocator.h
Jira NVGPU-13
Change-Id: I81a05d21ecdbd90c2076a9f0aefd0e40b215bd33
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1293187
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Guest doesn't explicitly send command to the RM server
to invalidate tlb which is done implicitly when mapping
or unmapping buffer. Remove support for this call.
Bug 1665111
Change-Id: Icf2edae7feffa35b1dbf87c227b3e98b506e6519
Signed-off-by: Aparna Das <aparnad@nvidia.com>
Reviewed-on: http://git-master/r/1287728
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move semaphore_gk20a.c drivers/gpu/nvgpu/common/ since the semaphore
code is common to all chips.
Move the semaphore_gk20a.h header file to drivers/gpu/nvgpu/include/nvgpu
and rename it to semaphore.h. Also update all places where the header
is inluced to use the new path.
This revealed an odd location for the enum gk20a_mem_rw_flag. This should
be in the mm headers. As a result many places that did not need anything
semaphore related had to include the semaphore header file. Fixing this
oddity allowed the semaphore include to be removed from many C files that
did not need it.
Bug 1799159
Change-Id: Ie017219acf34c4c481747323b9f3ac33e76e064c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1284627
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Simplify ref-counting on VMs: take a ref when a VM is bound to a
channel and drop a ref when a channel is freed.
Previously ref-counts were scattered over the driver. Also the CE
and CDE code would bind channels with custom rolled code. This was
because the gk20a_vm_bind_channel() function took an as_share as
the VM argument (the VM was then inferred from that as_share).
However, it is trivial to abtract that bit out and allow a central
bind channel function that just takes a VM and a channel.
Bug 1846718
Change-Id: I156aab259f6c7a2fa338408c6c4a3a464cd44a0c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1261886
Reviewed-by: Richard Zhao <rizhao@nvidia.com>
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove the special VMA that could be used for allocating fixed
addresses. This feature was never used and is not worth maintaining.
Bug 1396644
Bug 1729947
Change-Id: I06f92caa01623535516935acc03ce38dbdb0e318
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1265302
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The basic structure of this patch is to make the small page allocator
and the large page allocator into pointers (where they used to be just
structs). Then assign each of those pointers to the same actual
allocator since the buddy allocator has supported mixed page sizes
since its inception.
For the rest of the driver some changes had to be made in order to
actually support mixed pages in a single address space.
1. Unifying the allocation page size determination
Since the allocation and map operations happen at distinct
times both mapping and allocation of GVA space must agree
on page size. This is because the allocation has to separate
allocations into separate PDEs to avoid the necessity of
supporting mixed PDEs.
To this end a function __get_pte_size() was introduced which
is used both by the balloc code and the core GPU MM code. It
determines page size based only on the length of the mapping/
allocation.
2. Fixed address allocation + page size
Similar to regular mappings/GVA allocations fixed address
mapping page size determination had to be modified. In the
past the address of the mapping determined page size since
the address space split was by address (low addresses were
small pages, high addresses large pages). Since that is no
longer the case the page size field in the reserve memory
ioctl is now honored by the mapping code. When, for instance,
CUDA makes a memory reservation it specifies small or large
pages. When CUDA requests mappings to be made within that
address range the page size is then looked up in the reserved
memory struct.
Fixed address reservations were also modified to now always
allocate at a PDE granularity (64M or 128M depending on
large page size. This prevents non-fixed allocations from
ending up in the same PDE and causing kernel panics or GMMU
faults.
3. The rest...
The rest of the changes are just by products of the above.
Lots of places required minor updates to use a pointer to
the GVA allocator struct instead of the struct itself.
Lastly, this change is not truly complete. More work remains to be
done in order to fully remove the notion that there was such a thing
as separate address spaces for different page sizes. Basically after
this patch what remains is cleanup and proper documentation.
Bug 1396644
Bug 1729947
Change-Id: If51ab396a37ba16c69e434adb47edeef083dce57
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1265300
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move the GPU allocators to common/mm/ since the allocators are common
code across all GPUs. Also rename the allocator code to move away from
gk20a_ prefixed structs and functions.
This caused one issue with the nvgpu_alloc() and nvgpu_free() functions.
There was a function for allocating either with kmalloc() or vmalloc()
depending on the size of the allocation. Those have now been renamed to
nvgpu_kalloc() and nvgpu_kfree().
Bug 1799159
Change-Id: Iddda92c013612bcb209847084ec85b8953002fa5
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1274400
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
vgpu_vm_alloc_share() wants to return -EINVAL if VMA areas requested
do not fulfill the criteria. The error code gets overwritten by a
call to vgpu_comm_sendrecv(), which makes vgpu_vm_alloc_share() always
return 0.
Change-Id: I93f56025f963d1d4ad2f9b06139fce742d3be41b
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/1249961
GVS: Gerrit_Virtual_Submit
Reviewed-by: Aingara Paramakuru <aparamakuru@nvidia.com>
Reviewed-by: Seema Khowala <seemaj@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove a global debugfs variable and instead save the allocator
debugfs root node in the gk20a struct.
Bug 1799159
Change-Id: If4eed34fa24775e962001e34840b334658f2321c
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1225611
(cherry picked from commit 1908fde10bb1fb60ce898ea329f5a441a3e4297a)
Reviewed-on: http://git-master/r/1242390
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move vgpu private data to a dedicated structure and allocate it
at probe time. Also add virt_handle helper function which is used
everywhere.
JIRA VFND-2103
Change-Id: I125911420be72ca9be948125d8357fa85d1d3afd
Signed-off-by: Richard Zhao <rizhao@nvidia.com>
Reviewed-on: http://git-master/r/1185206
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vladislav Buzov <vbuzov@nvidia.com>
|
|
|
|
|
|
|
|
|
| |
Move to a more generic name of GPU_ALLOC_*.
Change-Id: Icbbd366847a9d74f83f578e4d9ea917a6e8ea3e2
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1176445
Reviewed-by: Yu-Huan Hsu <yhsu@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Support multiple types of allocation backends. Currently there is
only one allocator implementation available: a buddy allocator.
Buddy allocators have certain limitations though. For one the
allocator requires metadata to be allocated from the kernel's
system memory. This causes a given buddy allocation to potentially
sleep on a kmalloc() call.
This patch has been created so that a new backend can be created
which will avoid any dynamic system memory management routines
from being called.
Bug 1781897
Change-Id: I98d6c8402c049942f13fee69c6901a166f177f65
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/1172115
GVS: Gerrit_Virtual_Submit
Reviewed-by: Konsta Holtta <kholtta@nvidia.com>
Reviewed-by: Yu-Huan Hsu <yhsu@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Propagate the buffer aperture flag in gk20a_locked_gmmu_map up so that
buffers represented as a mem_desc and present in vidmem can be mapped to
gpu.
JIRA DNVGPU-18
JIRA DNVGPU-76
Change-Id: I46cf87e27229123016727339b9349d5e2c835b3e
Signed-off-by: Konsta Holtta <kholtta@nvidia.com>
Reviewed-on: http://git-master/r/1169308
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
JIRA VFND-1005
Bug 1594604
Change-Id: Ic159a1aff9cee508194f1f5dff7a16eb0e47ad64
Signed-off-by: Richard Zhao <rizhao@nvidia.com>
Reviewed-on: http://git-master/r/833498
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Aingara Paramakuru <aparamakuru@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Implement NVGPU_GPU_IOCTL_ALLOC_AS_FLAGS_USERSPACE_MANAGED, which
enables creating userspace-managed GPU address spaces.
When an address space is marked as userspace-managed, the following
changes are in effect:
- Only fixed-address mappings are allowed.
- VA space allocation for fixed-address mappings is not required,
except to mark space as sparse.
- Maps and unmaps are always immediate. In particular, the mapping
ref increments at kickoffs and decrements at job completion are
skipped.
Bug 1614735
Bug 1623949
Bug 1660392
Change-Id: I834fe19b3f65e9b02c268952383eddee0e465759
Signed-off-by: Sami Kiminki <skiminki@nvidia.com>
Reviewed-on: http://git-master/r/738558
Reviewed-on: http://git-master/r/833253
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
JIRA VFND-890
Change-Id: I8eba041b663cead94f2cc3d75d6458d472f1a755
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Signed-off-by: Richard Zhao <rizhao@nvidia.com>
Reviewed-on: http://git-master/r/815378
(cherry picked from commit 4b52329e955758ec4368abcb463ce4e3a2653237)
Reviewed-on: http://git-master/r/820499
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add vgpu framework and build for T18x.
Bug 1677153
JIRA VFND-693
Change-Id: Icf9fd8e0b5769228aee59c54f9b000b992e5fcca
Signed-off-by: Aingara Paramakuru <aparamakuru@nvidia.com>
Reviewed-on: http://git-master/r/792559
Reviewed-on: http://git-master/r/806178
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Separate the kernel and userspace regions in the GPU virtual address
space. Do this by reserving the last part of the GPU VA aperture for
the kernel, and extend GPU VA aperture accordingly for regular address
spaces. This prevents the kernel polluting the userspace-visible GPU
VA regions, and thus, makes the success of fixed-address mapping more
predictable.
Bug 200077571
Change-Id: I63f0e73d4c815a4a9fa4a9ce568709974690ef0f
Signed-off-by: Sami Kiminki <skiminki@nvidia.com>
Reviewed-on: http://git-master/r/747191
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- add hal initializaiton
- create folders vgpu/gk20a and vgpu/gm20b for specific code
Bug 1653185
Change-Id: If94d45e22a1d73d2e4916673736cc29751be4e40
Signed-off-by: Richard Zhao <rizhao@nvidia.com>
Reviewed-on: http://git-master/r/774148
GVS: Gerrit_Virtual_Submit
Reviewed-by: Aingara Paramakuru <aparamakuru@nvidia.com>
Reviewed-by: Ken Adams <kadams@nvidia.com>
|
|
|
|
|
|
|
|
| |
Implement support for privileged pages. Use them for kernel allocated buffers.
Change-Id: I720fc441008077b8e2ed218a7a685b8aab2258f0
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/761919
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add batch support for mapping and unmapping. Batching essentially
helps transform some per-map/unmap overhead to per-batch overhead,
namely gk20a_busy()/gk20a_idle() calls, GPU L2 flushes, and GPU TLB
invalidates. Batching with size 64 has been measured to yield >20x
speed-up in low-level fixed-address mapping microbenchmarks.
Bug 1614735
Bug 1623949
Change-Id: Ie22b9caea5a7c3fc68a968d1b7f8488dfce72085
Signed-off-by: Sami Kiminki <skiminki@nvidia.com>
Reviewed-on: http://git-master/r/733231
(cherry picked from commit de4a7cfb93e8228a4a0c6a2815755a8df4531c91)
Reviewed-on: http://git-master/r/763812
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 2e5803d0f2b7d7a1577a40f45ab9f3b22ef2df80 since
the issue seen with bug 200106514 is fixed with change
http://git-master/r/#/c/752080/.
Bug 200112195
Change-Id: I588151c2a7ea74bd89dc3fd48bb81ff2c49f5a0a
Signed-off-by: Bharat Nihalani <bnihalani@nvidia.com>
Reviewed-on: http://git-master/r/752503
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit ce1cf06b9a8eb6314ba0ca294e8cb430e1e141c0 since
it causes GPU pbdma interrupt to be generated.
Bug 200106514
Change-Id: If3ed9a914c4e3e7f3f98c6609c6dbf57e1eb9aad
Signed-off-by: Bharat Nihalani <bnihalani@nvidia.com>
Reviewed-on: http://git-master/r/749291
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 7eb42bc239dbd207208ff491c3fb65c3d83274d8.
The original commit was actually fine.
Change-Id: I564ce6530ac73fcfad17dcec9c53f0353b4f02d4
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/743300
(cherry picked from commit e99aa2485f8992eabe3556f3ebcb57bdc8ad91ff)
Reviewed-on: http://git-master/r/743301
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 2e235ac150fa4af8632c9abf0f109a10973a0bf5.
Change-Id: I3aa745152124c2bc09c6c6dc5aeb1084ae7e08a4
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/741469
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Hiroshi Doyu <hdoyu@nvidia.com>
Tested-by: Hiroshi Doyu <hdoyu@nvidia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Implement a new buddy allocation scheme for the GPU's VA space.
The bitmap allocator was using too much memory and is not a scaleable
solution as the GPU's address space keeps getting bigger. The buddy
allocation scheme is much more memory efficient when the majority
of the address space is not allocated.
The buddy allocator is not constrained by the notion of a split
address space. The bitmap allocator could only manage either small
pages or large pages but not both at the same time. Thus the bottom
of the address space was for small pages, the top for large pages.
Although, that split is not removed quite yet, the new allocator
enables that to happen.
The buddy allocator is also very scalable. It manages the relatively
small comptag space to the enormous GPU VA space and everything in
between. This is important since the GPU has lots of different sized
spaces that need managing.
Currently there are certain limitations. For one the allocator does
not handle the fixed allocations from CUDA very well. It can do so
but with certain caveats. The PTE page size is always set to small.
This means the BA may place other small page allocations in the
buddies around the fixed allocation. It does this to avoid having
large and small page allocations in the same PDE.
Change-Id: I501cd15af03611536490137331d43761c402c7f9
Signed-off-by: Alex Waterman <alexw@nvidia.com>
Reviewed-on: http://git-master/r/740694
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Terje Bergstrom <tbergstrom@nvidia.com>
Tested-by: Terje Bergstrom <tbergstrom@nvidia.com>
|