dGPUs in all versions of ROCm and APUs starting with ROCM 2.2 can
under-allocate scratch resources. This patch adds support for
the CP to trigger a recoverable error so that the host can attempt to
re-allocate scratch to satisfy the currently stalled kernel.
Note that this patch does not include a mechanism to handle dynamic
scratch allocation for queues with in-flight kernels, as these queues
would first need to be drained and descheduled, which would require some
additional effort in the hsaPP and HW queue scheduler. If the CP
encounters this scenerio it will assert. I suspect this is not a
particularly common occurence in most of our applications so it is left
as a TODO.
This patch also fixes a few memory leaks and updates the old DMA callback
object interface to use a much cleaner c++11 lambda interface.
Change-Id: Ica8a5fc88888283415507544d6cc49fa748fe84d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/42201
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Matt Sinclair <mattdsinclair@gmail.com>
Maintainer: Matt Sinclair <mattdsinclair@gmail.com>
In the HSAQueueDescriptor ptr function, we mod the index by numElts, but
numElts was previously just set to size, which was the raw size of the
queue. This lead to indexing past the queue. We fix this by dividing by
the size by the AQL packet size to get the actual number of elements the
queue can hold.
We also add an assert for indexing into the queue, as there is a
scenario where the queue reports a larger size than it actually is.
Change-Id: Ie5e699379f303255305c279e58a34dc783df86a0
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/42423
Reviewed-by: Matt Sinclair <mattdsinclair@gmail.com>
Reviewed-by: Matthew Poremba <matthew.poremba@amd.com>
Maintainer: Matt Sinclair <mattdsinclair@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
This patch is adding an extra parameter to the Ruby.create_system
function. The idea is to remove any assumption about cpu configuration
in the ruby scripts.
At the moment the scripts are assuming a flat list of cpu assigned
to the system object. Unfortunately this is not standardized, as
some systems might empoloy a different layout of cpus, like grouping
them in cluster objects.
With this patch we are allowing client scripts to provide the cpu list
as an extra argument
This has the extra benefit of removing the indexing hack
if len(system.cpu) == 1:
which was present in most scripts
Change-Id: Ibc06b920273cde4f7c394d61c0ca664a7143cd27
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/43287
Maintainer: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Jason Lowe-Power <power.jg@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <power.jg@gmail.com>
Most ISAs used that constant exactly once, when setting up a Process.
This change just propogates the constant to the one place it's used. In
MIPS, the endianness is hard coded as little. There were some checks
which would change the behavior if the endianness was big. This change
removes that dead code. If someone wants to add support for big endian
MIPS, they can go back and add in the small bits of code that would be
required. It's likely the existing big endian support was incomplete and
not tested, so it's probably best for someone interested in it to start
fresh anyway.
Change-Id: Ife6ffcf4bca40001d5d9126f7d795f954f66bb22
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/40178
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
The systemc dir was not included in this fix.
First it was identified that there were only occurrences
at 0, 1, 2 and 3 levels of indentation, using:
grep -nrE --exclude-dir=systemc \
"^ *union [A-Za-z].* {$" src/
Then the following commands were run to replace:
<indent level>union X ... {
by:
<indent level>union X ...
<indent level>{
Level 0:
grep -nrl --exclude-dir=systemc \
"^union [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^union ([A-Za-z].*) \{$/union \1\n\{/g'
Level 1:
grep -nrl --exclude-dir=systemc \
"^ union [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^ union ([A-Za-z].*) \{$/ union \1\n \{/g'
and so on.
Change-Id: I066854eb27a8acd2cc2dfa41596bb1b1f66c71b1
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/43328
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Gabe Black <gabe.black@gmail.com>
The systemc dir was not included in this fix.
First it was identified that there were only occurrences
at 0, 1, and 2 levels of indentation (and 2 of 2 spaces,
1 of 3 spaces and 2 of 12 spaces), using:
grep -nrE --exclude-dir=systemc \
"^ *enum [A-Za-z].* {$" src/
Then the following commands were run to replace:
<indent level>enum X ... {
by:
<indent level>enum X ...
<indent level>{
Level 0:
grep -nrl --exclude-dir=systemc \
"^enum [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^enum ([A-Za-z].*) \{$/enum \1\n\{/g'
Level 1:
grep -nrl --exclude-dir=systemc \
"^ enum [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^ enum ([A-Za-z].*) \{$/ enum \1\n \{/g'
and so on.
Change-Id: Ib186cf379049098ceaec20dfe4d1edcedd5f940d
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/43326
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Gabe Black <gabe.black@gmail.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
In most ISAs except MIPS and Power, this was implemented as
inst->advancePC(). It works just fine to call this function all the
time, but the idea had originally been that for ISAs which could simply
advance the PC using the PC itself, they could save the virtual function
call. Since the only ISAs which could skip the call were MIPS and Power,
and neither is at the point where that level of performance tuning
matters, this function can be collapsed with little downside.
If this turns out to be a performance bottleneck in the future, the way
the PC is managed could be revisited to see if we can factor out this
trip to the instruction object in the first place.
Change-Id: I533d1ad316e5c936466c529b7f1238a9ab87bd1c
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/39335
Maintainer: Gabe Black <gabe.black@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Alex Dutu <alexandru.dutu@amd.com>
The function had been introduced in the past when we needed to
instantiate either an ArmSystem or a LinuxArmSystem depending on the
workload. Now that the workload object has been introduced in gem5, we
always instantiate an ArmSystem in FS mode, hence we don't need a
function to generate the System object
Change-Id: I79ccf31087b84521cce32da71bc835ff202dc432
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/43285
Reviewed-by: Jason Lowe-Power <power.jg@gmail.com>
Maintainer: Jason Lowe-Power <power.jg@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
The systemc dir was not included in this fix.
First it was identified that there were only occurrences
at 0, 1, 2 and 3 levels of indentation (and a single
occurrence of 2 and 3 spaces), using:
grep -nrE --exclude-dir=systemc \
"^ *struct [A-Za-z].* {$" src/
Then the following commands were run to replace:
<indent level>struct X ... {
by:
<indent level>struct X ...
<indent level>{
Level 0:
grep -nrl --exclude-dir=systemc
"^struct [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^struct ([A-Za-z].*) \{$/struct \1\n\{/g'
Level 1:
grep -nrl --exclude-dir=systemc \
"^ struct [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^ struct ([A-Za-z].*) \{$/ struct \1\n \{/g'
and so on.
Change-Id: I362ef58c86912dabdd272c7debb8d25d587cd455
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/39017
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
The systemc dir was not included in this fix.
First it was identified that there were only occurrences
at 0, 1, and 2 levels of indentation, using:
grep -nrE --exclude-dir=systemc \
"^ *class [A-Za-z].* {$" src/
Then the following commands were run to replace:
<indent level>class X ... {
by:
<indent level>class X ...
<indent level>{
Level 0:
grep -nrl --exclude-dir=systemc
"^class [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^class ([A-Za-z].*) \{$/class \1\n\{/g'
Level 1:
grep -nrl --exclude-dir=systemc \
"^ class [A-Za-z].* {$" src/ | \
xargs sed -Ei \
's/^ class ([A-Za-z].*) \{$/ class \1\n \{/g'
and so on.
Change-Id: I17615ce16a333d69867b27c7bae0f4fdafd8b2eb
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/39015
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
We are adding a controller method to MemInterface objects making
them able to generate the appropriate memory controller.
This will bring the following benefits
a) Semplification: It will simplify MemConfig.config_mem
b) Reusability: Scripts not using config_mem
won't have to duplicate the if...else checks
c) Modularity: Users will be able to define their own
dram interfaces without needing to handle the mem_ctrl
mapping in the shared MemConfig.py module
Change-Id: I4b836fd7c91675cf7aacc644f25989484d5be3ec
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/42074
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Wendy Elsasser <wendy.elsasser@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
There is no way to make this sort of template work with more than one
ISA at a time, and it's also more complex than it needs to be,
particularly since the methods within it are never used in performance
critical code. Using virtual functions is also simpler and uses less
code.
Change-Id: I0baa1a651fa656420f6f90776572f8700a6d7cab
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/40106
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
The only difference between the NormalLogger and Logger is
a simple implementation for log(), which is then called by
the other loggers. Since this is common to everybody, move
this implementation to Logger and remove NormalLogger.
This makes it possible to test the NormalLoggers using the
current gtest logging framework.
Change-Id: I6805fa14f58ddc7d37b00fcd7fcacb32e0b5d456
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/41395
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Bobby R. Bruce <bbruce@ucdavis.edu>
In fs simulation, the kernel is loaded to physical address first and
then it would relocate itself to virtual address. The address which
using by kernel symbol table is virtual address. To debug the process
before kernel relocated to virutal memory, we need another copy of
symbol table for physical address.
Change-Id: I38107ff94b301df1a5170dd98774df88cfb02298
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/43104
Reviewed-by: Gabe Black <gabe.black@gmail.com>
Maintainer: Gabe Black <gabe.black@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
In some cases, we want to move and copy the symbol table to another
address. However, the name of symbol table should be unique. This rename
helper provides a way to modify the name of symbol. Developers can use
it to solve the conflict with this helper.
Change-Id: I4627e06da3a03da57009d613188be117c75750a0
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/43105
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Maintainer: Bobby R. Bruce <bbruce@ucdavis.edu>
Tested-by: kokoro <noreply+kokoro@google.com>
This patch add a new Ruby cache coherence protocol based on Arm' AMBA5
CHI specification. The CHI protocol defines and implements two state
machine types:
- Cache_Controller: generic cache controller that can be configured as:
- Top-level L1 I/D cache
- A intermediate level (L2, L3, ...) private or shared cache
- A CHI home node (i.e. the point of coherence of the system and
has the global directory)
- A DMA requester
- Memory_Controller: implements a CHI slave node and interfaces with
gem5 memory controller. This controller has the functionality of a
Directory_Controller on the other Ruby protocols, except it doesn't
have a directory.
The Cache_Controller has multiple cache allocation/deallocation
parameters to control the clusivity with respect to upstream caches.
Allocation can be completely disabled to use Cache_Controller as a
DMA requester or as a home node without a shared LLC.
The standard configuration file configs/ruby/CHI.py provides a
'create_system' compatible with configs/example/fs.py and
configs/example/se.py and creates a system with private L1/L2 caches
per core and a shared LLC at the home nodes. Different cache topologies
can be defined by modifying 'create_system' or by creating custom
scripts using the structures defined in configs/ruby/CHI.py.
This patch also includes the 'CustomMesh' topology script to be used
with CHI. CustomMesh generates a 2D mesh topology with the placement
of components manually defined in a separate configuration file using
the --noc-config parameter.
The example in configs/example/noc_config/2x4.yaml creates a simple 2x4
mesh. For example, to run a SE mode simulation, with 4 cores,
4 mem ctnrls, and 4 home nodes (L3 caches):
build/ARM/gem5.opt configs/example/se.py \
--cmd 'tests/test-progs/hello/bin/arm/linux/hello' \
--ruby --num-cpus=4 --num-dirs=4 --num-l3caches=4 \
--topology=CustomMesh --noc-config=configs/example/noc_config/2x4.yaml
If one doesn't care about the component placement on the interconnect,
the 'Crossbar' and 'Pt2Pt' may be used and they do not require the
--noc-config option.
Additional authors:
Joshua Randall <joshua.randall@arm.com>
Pedro Benedicte <pedro.benedicteillescas@arm.com>
Tuan Ta <tuan.ta2@arm.com>
JIRA: https://gem5.atlassian.net/browse/GEM5-908
Change-Id: I856524b0afd30842194190f5bd69e7e6ded906b0
Signed-off-by: Tiago Mück <tiago.muck@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/42563
Reviewed-by: Jason Lowe-Power <power.jg@gmail.com>
Maintainer: Jason Lowe-Power <power.jg@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Rather than print the errno, print the result of strerror so the user
doesn't have to look up what each error code does. Don't duplicate
descriptors needlessly which have been returned by pipe(). Use panic_if
instead of if () panic. Check each call to a standard library function
instead of calling multiple and then only knowing that one of them
failed (but not which one). Close the far side of our pipes for both
the gem5 process, and the process that will become diod. Remove some
stray #undef-s which undefine macros that were never defined. Don't
try to force the descriptors going to diod to be particular numbers.
Slightly reduce nesting in the if which checks the results of fork.
Drop unnecessary \n-s in calls to panic, inform, etc. Minor spacing
related style fixes. Use nullptr instead of NULL.
Change-Id: I48d93778a1e139ef624876a0b316486aac774d7f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/43083
Reviewed-by: Yu-hsin Wang <yuhsingw@google.com>
Maintainer: Gabe Black <gabe.black@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
The WalkCache stats are wrongly using the legacy framework.
With this patch we are registering those to the hierarchical structure.
As we need to pass the Stats::Group parent at construction time,
we are replacing 2d arrays with Vector2d for count stats and using a flat
vector of pointers for the Formula stats
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Change-Id: I8992bc262a376e4e81a4d608c11dff6902e0a01d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/42984
Maintainer: Jason Lowe-Power <power.jg@gmail.com>
Reviewed-by: Jason Lowe-Power <power.jg@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
