denshooter/genode
1
0
Fork 0
mirror of https://github.com/mmueller41/genode.git synced 2026-01-21 12:32:56 +01:00
Code Issues Packages Projects Releases Wiki Activity

mxtasking: Made channel count independent from habitat size.

Browse Source
This commit is contained in:
Michael Mueller
2025-08-29 17:17:47 +02:00
parent 8cdefc4c68
commit 052026e86a
2 changed files with 14 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
repos/ealanos/src/lib/mx/tasking/scheduler.cpp
View File

@@ -1,5 +1,6 @@
#include "scheduler.h"
#include "mx/system/environment.h"
#include "mx/util/field_alloc.h"
#include "tukija/syscall-generic.h"
#include <cassert>
#include <mx/memory/global_heap.h>
@@ -25,6 +26,7 @@ Scheduler::Scheduler(const mx::util::core_set &core_set, const std::uint16_t pre
: _core_set(core_set), _count_channels(core_set.size()), _worker({}), _channel_numa_node_map({0U}),
_epoch_manager(core_set.size(), resource_allocator, _is_running), _statistic(_count_channels)
{
this->_vacant_channels_alloc = new (memory::GlobalHeap::allocate_cache_line_aligned(sizeof(util::Field_Allocator<config::max_cores()>))) util::Field_Allocator<config::max_cores()>(core_set.size());
this->_worker.fill(nullptr);
this->_channel_numa_node_map.fill(0U);
Genode::log("Initializing scheduler");
@@ -37,20 +39,23 @@ Scheduler::Scheduler(const mx::util::core_set &core_set, const std::uint16_t pre
auto ptr = memory::GlobalHeap::allocate(this->_channel_numa_node_map[worker_id], sizeof(Worker));
this->_worker[worker_id] =
new (ptr)
Worker(worker_id, core_id, this->_channel_numa_node_map[worker_id], this->_is_running, _vacant_channels_alloc, _remainder_channel_count,
Worker(worker_id, core_id, this->_channel_numa_node_map[worker_id], this->_is_running, *_vacant_channels_alloc, _remainder_channel_count,
prefetch_distance, this->_epoch_manager[worker_id], this->_epoch_manager.global_epoch(),
this->_statistic);
ptr = memory::GlobalHeap::allocate(this->_channel_numa_node_map[worker_id], sizeof(Channel));
this->_channels[worker_id] =
new (ptr) Channel(worker_id, this->_channel_numa_node_map[worker_id], prefetch_distance);
Genode::log("Channel ", worker_id, " created at ", _channels[worker_id]);
}
for (auto channel_id : core_set) {
auto ptr = memory::GlobalHeap::allocate(this->_channel_numa_node_map[channel_id], sizeof(Channel));
this->_channels[channel_id] =
new (ptr) Channel(channel_id, this->_channel_numa_node_map[channel_id], prefetch_distance);
Genode::log("Channel ", channel_id, " created at ", _channels[channel_id]);
}
Genode::log("Using ", _count_channels, " Channels");
Genode::log("CPU freq: ", mx::system::Environment::get_cpu_freq(), "kHz");
/* We need to state the actual number of channels here. But as _count_channels only denotes the
number of channels from the restricted core_set, defined by the application, we could end up reporting a number below the actual number of channels. This could lead to the queue stealing misbehaving, e.g. not all queues being taken by the workers. Furthermore, we must subtract the queue used by the foreman which is not stealable by definition.*/
Tukija::Cip::cip()->channel_info.count = worker_count - 1;
Tukija::Cip::cip()->channel_info.count = core_set.size() - 1;
}
Scheduler::~Scheduler() noexcept

6
repos/ealanos/src/lib/mx/tasking/scheduler.h
View File

@@ -83,8 +83,8 @@ public:
//Genode::log("Allocation before resume ", allocation);
unsigned current_cores = Tukija::Cip::cip()->cores_current.count();
allocate_cores(_count_channels - current_cores);
_is_running = true;
allocate_cores(_count_channels - current_cores);
_is_running = true;
}
[[nodiscard]] inline Worker *my_self() noexcept {
@@ -281,7 +281,7 @@ private:
alignas(64) std::array<Channel *, config::max_cores()> _channels{nullptr};
alignas(64) mx::util::Field_Allocator<config::max_cores()> _vacant_channels_alloc{63};
alignas(64) mx::util::Field_Allocator<config::max_cores()> *_vacant_channels_alloc{nullptr};
alignas(64) std::atomic<std::int32_t> _remainder_channel_count{0};
// Map of channel id to NUMA region id.