Move parsing MMTK_ variables into Rust

gc/mmtk.c

@@ -77,60 +77,6 @@ RB_THREAD_LOCAL_SPECIFIER struct MMTk_GCThreadTLS *rb_mmtk_gc_thread_tls;
 
 #include <pthread.h>
 
-static size_t
-rb_mmtk_system_physical_memory(void)
-{
-#ifdef __linux__
-    const long physical_pages = sysconf(_SC_PHYS_PAGES);
-    const long page_size = sysconf(_SC_PAGE_SIZE);
-    if (physical_pages == -1 || page_size == -1) {
-        rb_bug("failed to get system physical memory size");
-    }
-    return (size_t)physical_pages * (size_t)page_size;
-#elif defined(__APPLE__)
-    int mib[2];
-    mib[0] = CTL_HW;
-    mib[1] = HW_MEMSIZE; // total physical memory
-    int64_t physical_memory;
-    size_t length = sizeof(int64_t);
-    if (sysctl(mib, 2, &physical_memory, &length, NULL, 0) == -1) {
-        rb_bug("failed to get system physical memory size");
-    }
-    return (size_t)physical_memory;
-#else
-#error no implementation of rb_mmtk_system_physical_memory on this platform
-#endif
-}
-
-static size_t
-rb_mmtk_parse_heap_limit(const char *argv, bool *had_error)
-{
-    char *endval = NULL;
-    int pow = 0;
-
-    size_t base = strtol(argv, &endval, 10);
-    if (base == 0) {
-        *had_error = true;
-    }
-
-    // if there were non-numbers in the string
-    // try and parse them as IEC units
-    if (*endval) {
-        if (strcmp(endval, "TiB") == 0)  {
-            pow = 40; // tebibytes. 2^40
-        } else if (strcmp(endval, "GiB") == 0)  {
-            pow = 30; // gibibytes. 2^30
-        } else if (strcmp(endval, "MiB") == 0)  {
-            pow = 20; // mebibytes. 2^20
-        } else if (strcmp(endval, "KiB") == 0)  {
-            pow = 10; // kibibytes. 2^10
-        }
-    }
-
-    *had_error = false;
-    return (base << pow);
-}
-
 static void
 rb_mmtk_init_gc_worker_thread(MMTk_VMWorkerThread gc_thread_tls)
 {
@@ -489,54 +435,6 @@ MMTk_Builder *
 rb_mmtk_builder_init(void)
 {
     MMTk_Builder *builder = mmtk_builder_default();
-
-    size_t mmtk_max_heap_size = RB_MMTK_DEFAULT_HEAP_MAX;
-    size_t mmtk_min_heap_size = RB_MMTK_DEFAULT_HEAP_MIN;
-    const char *mmtk_max_heap_size_arg;
-    const char *mmtk_min_heap_size_arg;
-    enum mmtk_heap_mode mode = RB_MMTK_DYNAMIC_HEAP;
-
-    // switch to fixed mode if defined, for any other value assume the default, which is dynamic
-    if (getenv("MMTK_HEAP_MODE") && !strcmp(getenv("MMTK_HEAP_MODE"), "fixed")) {
-        mode = RB_MMTK_FIXED_HEAP;
-    }
-
-    if ((mmtk_max_heap_size_arg = getenv("MMTK_HEAP_MAX"))) {
-        bool max_error = true;
-        mmtk_max_heap_size = rb_mmtk_parse_heap_limit(mmtk_max_heap_size_arg, &max_error);
-        if (max_error) {
-            rb_warn("MMTk maximum heap size invalid, using default");
-            mmtk_max_heap_size = RB_MMTK_DEFAULT_HEAP_MAX;
-        }
-    }
-
-    if ((mmtk_min_heap_size_arg = getenv("MMTK_HEAP_MIN"))) {
-        bool min_error;
-        mmtk_min_heap_size = rb_mmtk_parse_heap_limit(mmtk_min_heap_size_arg, &min_error);
-        if (min_error) {
-            rb_warn("MMTk minimum heap size invalid, using default");
-            mmtk_min_heap_size = RB_MMTK_DEFAULT_HEAP_MIN;
-        }
-    };
-
-    switch (mode) {
-      case RB_MMTK_FIXED_HEAP:
-        mmtk_set_fixed_heap(builder, mmtk_max_heap_size);
-        break;
-      case RB_MMTK_DYNAMIC_HEAP:
-        if (mmtk_min_heap_size >= mmtk_max_heap_size) {
-            rb_bug(
-                "MMTK Heap size invalid: please set max heap (%zu) larger than min heap (%zu)",
-                mmtk_max_heap_size, mmtk_min_heap_size
-            );
-        }
-        mmtk_set_dynamic_heap(builder, mmtk_min_heap_size, mmtk_max_heap_size);
-        break;
-      default:
-        rb_bug("Unreachable: Invalid MMTK Heap Mode");
-        break;
-    }
-
     return builder;
 }
 

gc/mmtk.h

@@ -98,10 +98,6 @@ void mmtk_init_binding(MMTk_Builder *builder,
                        const struct MMTk_RubyUpcalls *upcalls,
                        MMTk_ObjectReference weak_reference_dead_value);
 
-void mmtk_set_fixed_heap(MMTk_Builder *builder, size_t max_heap_size);
-
-void mmtk_set_dynamic_heap(MMTk_Builder *builder, size_t min_heap_size, size_t max_heap_size);
-
 void mmtk_initialize_collection(MMTk_VMThread tls);
 
 MMTk_Mutator *mmtk_bind_mutator(MMTk_VMMutatorThread tls);

gc/mmtk/Cargo.toml

@@ -24,6 +24,7 @@ env_logger = "0.11.3"
 once_cell = "1.17.0"
 atomic_refcell = "0.1.9"
 probe = "0.5"
+sysinfo = "0.32.0"
 
 [dependencies.mmtk]
 features = ["is_mmtk_object", "object_pinning", "sticky_immix_non_moving_nursery"]

gc/mmtk/src/api.rs

@@ -9,6 +9,7 @@ use crate::binding::RubyBinding;
 use crate::mmtk;
 use crate::Ruby;
 use crate::RubySlot;
+use crate::utils::default_heap_max;
 use mmtk::memory_manager;
 use mmtk::memory_manager::mmtk_init;
 use mmtk::util::constants::MIN_OBJECT_SIZE;
@@ -40,6 +41,30 @@ pub extern "C" fn mmtk_builder_default() -> *mut MMTKBuilder {
     let mut builder = MMTKBuilder::new_no_env_vars();
     builder.options.no_finalizer.set(true);
 
+
+    const DEFAULT_HEAP_MIN: usize = 1 << 20;
+
+    let mut mmtk_heap_min = match std::env::var("MMTK_HEAP_MIN") {
+        Ok(mode) if (mode.parse::<usize>().is_ok()) => mode.parse::<usize>().unwrap(),
+        Ok(_) | Err(_) => DEFAULT_HEAP_MIN
+    };
+
+    let mut mmtk_heap_max = match std::env::var("MMTK_HEAP_MAX") {
+        Ok(mode) if (mode.parse::<usize>().is_ok()) => mode.parse::<usize>().unwrap(),
+        Ok(_) | Err(_) => default_heap_max()
+    };
+
+    if mmtk_heap_min >= mmtk_heap_max {
+        println!("MMTK_HEAP_MIN({}) >= MMTK_HEAP_MAX({}). Using default values.", mmtk_heap_min, mmtk_heap_max);
+        mmtk_heap_min = DEFAULT_HEAP_MIN;
+        mmtk_heap_max = default_heap_max();
+    }
+
+    let mmtk_mode = match std::env::var("MMTK_HEAP_MODE") {
+        Ok(mode) if (mode == "fixed") => GCTriggerSelector::FixedHeapSize(mmtk_heap_max),
+        Ok(_) | Err(_) => GCTriggerSelector::DynamicHeapSize(mmtk_heap_min, mmtk_heap_max)
+    };
+
     // Parse the env var, if it's not found set the plan name to MarkSweep
     let plan_name = std::env::var("MMTK_PLAN")
         .unwrap_or(String::from("MarkSweep"));
@@ -51,7 +76,7 @@ pub extern "C" fn mmtk_builder_default() -> *mut MMTKBuilder {
     builder.options.plan.set(plan_selector);
 
     // Between 1MiB and 500MiB
-    builder.options.gc_trigger.set(GCTriggerSelector::DynamicHeapSize(1 << 20, 500 << 20));
+    builder.options.gc_trigger.set(mmtk_mode);
 
     Box::into_raw(Box::new(builder))
 }
@@ -77,18 +102,6 @@ pub extern "C" fn mmtk_init_binding(
         .unwrap_or_else(|_| panic!("Binding is already initialized"));
 }
 
-#[no_mangle]
-pub extern "C" fn mmtk_set_fixed_heap(builder: *mut MMTKBuilder, max_heap_size: usize) {
-    let builder = unsafe { &mut *builder };
-    builder.options.gc_trigger.set(GCTriggerSelector::FixedHeapSize(max_heap_size));
-}
-
-#[no_mangle]
-pub extern "C" fn mmtk_set_dynamic_heap(builder: *mut MMTKBuilder, min_heap_size: usize, max_heap_size: usize) {
-    let builder = unsafe { &mut *builder };
-    builder.options.gc_trigger.set(GCTriggerSelector::DynamicHeapSize(min_heap_size, max_heap_size));
-}
-
 #[no_mangle]
 pub extern "C" fn mmtk_initialize_collection(tls: VMThread) {
     memory_manager::initialize_collection(mmtk(), tls)

gc/mmtk/src/utils.rs

@@ -3,6 +3,7 @@ use std::sync::atomic::{AtomicUsize, Ordering};
 use atomic_refcell::AtomicRefCell;
 use mmtk::scheduler::{GCWork, GCWorker, WorkBucketStage};
 
+use sysinfo::System;
 use crate::Ruby;
 
 pub struct ChunkedVecCollector<T> {
@@ -86,3 +87,73 @@ impl AfterAll {
         }
     }
 }
+
+pub fn default_heap_max() -> usize {
+    let mut s = System::new();
+    s.refresh_memory();
+    s.total_memory()
+        .checked_mul(80)
+        .and_then(|v| v.checked_div(100))
+        .expect("Invalid Memory size") as usize
+}
+
+pub fn parse_capacity(input: String) -> usize {
+    let trimmed = input.trim();
+
+    const KIBIBYTE: usize = 1024;
+    const MEBIBYTE: usize = 1024 * KIBIBYTE;
+    const GIBIBYTE: usize = 1024 * MEBIBYTE;
+
+    let (val, suffix) = if let Some(pos) = trimmed.find(|c: char| !c.is_numeric()) {
+        (&trimmed[..pos], &trimmed[pos..])
+    } else {
+        (trimmed, "")
+    };
+
+    // 1MiB is the default heap size
+    match (val, suffix) {
+        (number, "GiB") => number.parse::<usize>().unwrap() * GIBIBYTE,
+        (number, "MiB") => number.parse::<usize>().unwrap() * MEBIBYTE,
+        (number, "KiB") => number.parse::<usize>().unwrap() * KIBIBYTE,
+        (number, suffix) if suffix.is_empty() => number.parse::<usize>().unwrap(),
+        (_, _) => default_heap_max()
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_parse_capacity_parses_bare_bytes() {
+        assert_eq!(1234, parse_capacity(String::from("1234")));
+    }
+
+    #[test]
+    fn test_parse_capacity_parses_kibibytes() {
+        assert_eq!(10240, parse_capacity(String::from("10KiB")))
+    }
+
+    #[test]
+    fn test_parse_capacity_parses_mebibytes() {
+        assert_eq!(10485760, parse_capacity(String::from("10MiB")))
+    }
+
+    #[test]
+    fn test_parse_capacity_parses_gibibytes() {
+        assert_eq!(10737418240, parse_capacity(String::from("10GiB")))
+    }
+
+    #[test]
+    fn test_parses_nonsense_value_as_default_max() {
+        let mut s = System::new();
+        s.refresh_memory();
+        let eighty_percent_mem = (s.total_memory() / 100 * 80) as usize;
+
+        assert_eq!(eighty_percent_mem, parse_capacity(String::from("notanumber")));
+        assert_eq!(eighty_percent_mem, parse_capacity(String::from("5tartswithanumber")));
+        assert_eq!(eighty_percent_mem, parse_capacity(String::from("number1nthemiddle")));
+        assert_eq!(eighty_percent_mem, parse_capacity(String::from("numberattheend111")));
+        assert_eq!(eighty_percent_mem, parse_capacity(String::from("mult1pl3numb3r5")));
+    }
+}