This reverts commit 35b07e1e28.
Reason for revert: This commit caused a regression rocrtst memory
subtest: Maximum Single Allocation in Memory Pools failed.
Change-Id: I15330625603f893200a08cd8b5b097f9bf95361f
[ROCm/ROCR-Runtime commit: e515fd818b]
This fixes a build issue with kfdtest and the amdgpu pro driver build.
This was requested as kfdtest is needed for regular testing due to the
inclusion of the ROCr/KFD stack in the amdgpu pro driver (OSGSUP-199)
Change-Id: I224d2e9ee3f02065596890b4d8226484f4fac04f
Signed-off-by: Jeremy Newton <Jeremy.Newton@amd.com>
[ROCm/ROCR-Runtime commit: 8026ba250c]
We want to compare bits, not check if a defined value is true
Change-Id: Ie51ede96d18eae01aff6677d852a056ee12bd9c6
[ROCm/ROCR-Runtime commit: e34dfa8ebd]
There is no default case, and we were missing a few types defined from
hsakmttypes.h. This was found via clang
Change-Id: I26193cb111a9d8220b1eff21c7313fe060288f36
[ROCm/ROCR-Runtime commit: 761d9d84d2]
While the ternary is nice to read, strlen in general is an expensive
call, so call it once and check if the value is greater than our maximum
allowable string length and adjust accordingly
Signed-off-by: Kent Russell <kent.russell@amd.com>
Change-Id: Id744f2ba0eb81bb2b3c52eb69f38a615398a655d
[ROCm/ROCR-Runtime commit: 025036a662]
Don't update the vm_object if GPU mapping failed. Print an error message
to help diagnose underlying problems.
Change-Id: I801ab6fe6c155bd25e6c0358007c106a4a019480
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
[ROCm/ROCR-Runtime commit: c7e6f5a274]
Use MAP_POPULATE when allocating anonymous system memory for later
GPU mapping as a userptr. This can speed up large allocations by
more than factor 2. I suspect populating pages in this way is more
efficient than the CPU page fault code path triggered by
get_user_pages in the kernel.
Change-Id: I188bbc1462ccb650d48cbfb1080dbb8eb7ada8b5
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
[ROCm/ROCR-Runtime commit: 8f26c0c40c]
On gfx9, the maximum number of wavefronts per queue is the minimum of
40 waves per compute units, or 512 waves per shader engine. On gfx10,
there can only be 32 waves per compute units.
Signed-off-by: Laurent Morichetti <laurent.morichetti@amd.com>
Change-Id: I148d1a4fe6c07cdbfaa1f77939eb29311c81c008
[ROCm/ROCR-Runtime commit: 783e346777]
Reserve some space in the context save area for the debugger's
use. There should be 32 bytes per wave for a given queue.
Change-Id: I65ddb6123d0f6afd3149844617ad19023009101d
[ROCm/ROCR-Runtime commit: 2ed2e46b9b]
The queue control stack size cannot exceed 0x7000 on ASICs
gfx1010 through gfx1031. The lower limit is not achievable
with AQL so this should have no practical effect.
Fixes control stack size overflow on large ASICs.
Change-Id: Ib78cf6e4c5f096044bf8de24debe211689891caa
Signed-off-by: Jay Cornwall <jay.cornwall@amd.com>
[ROCm/ROCR-Runtime commit: 44f80d170d]
1. Add KFDEvictTest.* for gfx90c based on CI test results
2. Remove SDMA blacklist based on SDMA issue fixed:
Change-Id: I86910fc98a5141f29959b35248a900f0c098a6e8
[ROCm/ROCR-Runtime commit: 36249ddc0e]
This patch is the hot fix to fix the param number checking after remove
dgpu input.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Change-Id: Ic980588f78616f99076de742af580afb4273fb2f
[ROCm/ROCR-Runtime commit: 8fc816affe]
gfx90c should use GFX902 which is the same with gfx902.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Change-Id: Id24dc2c85c9f49f36b00889c3b8b1b19cce34e09
[ROCm/ROCR-Runtime commit: 8ea0d49337]
These are removed now that we've consolidated the dev package
information into CMakeLists.txt from hsakmt-dev.txt.
Change-Id: I49496ec5def85b0af7fa6b15110910528a8e0be0
[ROCm/ROCR-Runtime commit: 654ee83ac8]
Add extended descriptions and e-mail address to CMakeLists
A lintian error will remain regarding stripping the .so, as we
will not be doing this for Release versions of the hsakmt .so
Signed-off-by: Kent Russell <kent.russell@amd.com>
Change-Id: I41c768dee28c0564d92b9c103a6e2d97590e4589
[ROCm/ROCR-Runtime commit: 0a4b23d625]
Whether use dgpu path will check the props which exposed from kernel.
We won't need hard code in the ASIC table.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Change-Id: I0c018a26b219914a41197ff36dbec7a75945d452
[ROCm/ROCR-Runtime commit: ad87f38dad]
KFD already implemented the fallback path for APU. Thunk will use flag
which exposed by kfd to configure is_dgpu instead of hardcode before.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Change-Id: I445f6cf668f9484dd06cd9ae1bb3cfe7428ec7eb
[ROCm/ROCR-Runtime commit: 12813691a2]
Gfx10 need 12bytes/wave control stack
Change-Id: I6c6f2819572e6b43aa3140d4dbe79d930e4c1c9c
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
Signed-off-by: Jay Cornwall <Jay.Cornwall@amd.com>
[ROCm/ROCR-Runtime commit: 3d3b28b670]
This is needed to avoid additional references to mapped BOs in child
processes that can prevent freeing memory in the parent process and lead
to out-of-memory conditions.
Change-Id: I25c90510a14dde515cc23ea5dc1f68e8d7e37a66
Signed-off-by: Philip Cox <Philip.Cox@amd.com>
[ROCm/ROCR-Runtime commit: f7a3427c99]
strlen(src) should not be used as the length in strncpy. Use memcpy
since we know the length of the string, and ensure that we
NULL-terminate regardless of length
Signed-off-by: Kent Russell <kent.russell@amd.com>
Change-Id: I21cc6d106510c69464e7ac9d3fc7da3a1e6d1a68
[ROCm/ROCR-Runtime commit: 04f6b9e16b]
The option to use kfd_fd for cpu mapping is for very old broken KFD
version, it is not used in upstreaming process. This causes issue when
multiple process uses shared system memory because the GTT address is
over 40 bits.
Change to always use render node fd to create CPU mapping.
Change-Id: Id7e7b2a2e2f13c6e62c5de170589abfff4d456b0
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
[ROCm/ROCR-Runtime commit: 9e9771a7d9]
This reverts commit c0a0ada18b.
Reason for revert: Change was submitted by accident
Change-Id: If05c705e22296fd3ca789f269737d379a933361d
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
[ROCm/ROCR-Runtime commit: fec3780c1a]
Since CMAKE_MODULE_PATH can already be set by another project,
we should just append the libhsamkt cmake module directory to it.
Change-Id: I999dc52a2862e4bbff02e0a8e8b39530f4dae2cd
Signed-off-by: Vlad Sytchenko <vladislav.sytchenko@amd.com>
[ROCm/ROCR-Runtime commit: 5fb771a195]
Default is OFF to conform to latest cmake standard (3.15) and
because this feature can cause some confusion for unaware developers.
Change-Id: I6fdbd5ab76c90b73ac6451a383334a83cabef438
Signed-off-by: Sean Keely <Sean.Keely@amd.com>
[ROCm/ROCR-Runtime commit: f0b0555996]
Immediate benefit is that this enables cmake native find_package
support including transitive passthrough of link dependencies,
options, and include paths.
It does require a bump to cmake 3.6 to enable proper package file
names.
Change-Id: I3f4e2f2db745c04a7f686111b080098a0d92fbf5
Signed-off-by: Sean Keely <Sean.Keely@amd.com>
[ROCm/ROCR-Runtime commit: 06add334b9]
Signed-off-by: Gang Ba <gaba@amd.com>
Change-Id: I5c23a8dacf9bc50c740908aabe391432f2c7112e
Signed-off-by: Gang Ba <gaba@amd.com>
[ROCm/ROCR-Runtime commit: d675d1cce1]
KFD now passes the ASIC revision to user level through some bits
in the HSA topology's capability field. Some user-level software
wants this because different ASIC revisions may require user-level
software to do different things (e.g. patch code for things that
are changed in later hardware revisions).
Change-Id: I16f2a15ae0875edd01ebdb1f1685ec7865f7049e
[ROCm/ROCR-Runtime commit: 5ddd8fb68b]
Using static_assert breaks in "Many Linux" build environment. It is not
supported by that libc version. _Static_assert is a compiler built-in
and does not depend on the libc version.
Change-Id: I37cf0ad10de94d8f6fc8cefc4fdda55c9520d599
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
[ROCm/ROCR-Runtime commit: df16950a0c]
Update the non-upstream ioctl numbers to align with the change in the
kernel.
Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Change-Id: Ie0ddccb343a023b55eb18477c59341acaa666e99
[ROCm/ROCR-Runtime commit: a37a88ddcb]
The purpose of this KFDTest is to investigate the behaviour of an
SDMA queue when an invalid memory address is used.
v2: Don't wait for SDMA queue to finish - it won't finish because of
the gpuvm fault.
v3: Create kfd event before SDMA queue submission. This fix the issue
that gpuvm fault happens earlier than kfd event is created then KFD
exception handler can't find the kfd event (to wake up kfd test)
v4: Instead of using 0x12345678 as the invalid VA, map one page of
FB to gpu and unmap it. Use the mapped GPUVA as the
invalid address
Change-Id: I58af1511f75d869adddede302b238c2725f3fe5a
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
[ROCm/ROCR-Runtime commit: 46f5e83066]
We need HostAccess for large-bar XGMI configs for this test,
but we can't use it on small-bar XGMI configs, so deal with it
appropriately
Change-Id: Ic16b810de03adfc63de21a00c55e5f3ea1ce66d1
[ROCm/ROCR-Runtime commit: dc784ba482]
Felix Kuehling
Jeremy Newton
Kent Russell
Laurent Morichetti
Chengming Gui
Jay Cornwall
Harish Kasiviswanathan
Jinzhou.Su
Huang Rui
Oak Zeng
Philip Cox
Philip Yang
Ramesh Errabolu
Yong Zhao
Cole Nelson
Gang Ba
Vlad Sytchenko
Sean Keely
Eric Huang
Joseph Greathouse
Ranieri Althoff
Amber Lin
Ori Messinger