[PATCH v4 9/9] rust: percpu: cache per-CPU pointers in the dynamic case

Mitchell Levy posted 9 patches 1 month, 1 week ago
[PATCH v4 9/9] rust: percpu: cache per-CPU pointers in the dynamic case
Posted by Mitchell Levy 1 month, 1 week ago
Currently, the creation of a `PerCpuNumeric` requires a memory read via
the `Arc` managing the dynamic allocation. While the compiler might be
clever enough to consolidate these reads in some cases, the read must
happen *somewhere*, which, when we're concerning ourselves with
individual instructions, is a very high burden.

Instead, cache the `PerCpuPointer` inside the `DynamicPerCpu` structure;
then, the `Arc` is used solely to manage the allocation.

Signed-off-by: Mitchell Levy <levymitchell0@gmail.com>
---
 rust/kernel/percpu/dynamic.rs | 22 ++++++++++++----------
 rust/kernel/percpu/numeric.rs |  4 ++--
 2 files changed, 14 insertions(+), 12 deletions(-)

diff --git a/rust/kernel/percpu/dynamic.rs b/rust/kernel/percpu/dynamic.rs
index 99acbf6363f5..dcf3e1c4f7a9 100644
--- a/rust/kernel/percpu/dynamic.rs
+++ b/rust/kernel/percpu/dynamic.rs
@@ -72,6 +72,9 @@ pub struct DynamicPerCpu<T> {
     // INVARIANT: The memory location in each CPU's per-CPU area pointed at by the alloc is
     // initialized.
     alloc: Option<Arc<PerCpuAllocation<T>>>,
+    // INVARIANT: `ptr` is the per-CPU pointer managed by `alloc`, which does not change for the
+    // lifetime of `self`.
+    pub(super) ptr: PerCpuPtr<T>,
 }
 
 impl<T: Zeroable> DynamicPerCpu<T> {
@@ -83,9 +86,13 @@ impl<T: Zeroable> DynamicPerCpu<T> {
     pub fn new_zero(flags: Flags) -> Option<Self> {
         let alloc: PerCpuAllocation<T> = PerCpuAllocation::new_zero()?;
 
+        let ptr = alloc.0;
         let arc = Arc::new(alloc, flags).ok()?;
 
-        Some(Self { alloc: Some(arc) })
+        Some(Self {
+            alloc: Some(arc),
+            ptr,
+        })
     }
 }
 
@@ -115,15 +122,10 @@ pub fn new_with(val: &T, flags: Flags) -> Option<Self> {
 
         let arc = Arc::new(alloc, flags).ok()?;
 
-        Some(Self { alloc: Some(arc) })
-    }
-}
-
-impl<T> DynamicPerCpu<T> {
-    /// Gets the allocation backing this per-CPU variable.
-    pub(crate) fn alloc(&self) -> &Arc<PerCpuAllocation<T>> {
-        // SAFETY: This type's invariant ensures that `self.alloc` is `Some`.
-        unsafe { self.alloc.as_ref().unwrap_unchecked() }
+        Some(Self {
+            alloc: Some(arc),
+            ptr,
+        })
     }
 }
 
diff --git a/rust/kernel/percpu/numeric.rs b/rust/kernel/percpu/numeric.rs
index e76461f05c66..23a7a09216d0 100644
--- a/rust/kernel/percpu/numeric.rs
+++ b/rust/kernel/percpu/numeric.rs
@@ -22,7 +22,7 @@ impl DynamicPerCpu<$ty> {
             pub fn num(&mut self) -> PerCpuNumeric<'_, $ty> {
                 // The invariant is satisfied because `DynamicPerCpu`'s invariant guarantees that
                 // this pointer is valid and initialized on all CPUs.
-                PerCpuNumeric { ptr: &self.alloc().0 }
+                PerCpuNumeric { ptr: &self.ptr }
             }
         }
         impl StaticPerCpu<$ty> {
@@ -78,7 +78,7 @@ impl DynamicPerCpu<$ty> {
             pub fn num(&mut self) -> PerCpuNumeric<'_, $ty> {
                 // The invariant is satisfied because `DynamicPerCpu`'s invariant guarantees that
                 // this pointer is valid and initialized on all CPUs.
-                PerCpuNumeric { ptr: &self.alloc().0 }
+                PerCpuNumeric { ptr: &self.ptr }
             }
         }
         impl StaticPerCpu<$ty> {

-- 
2.34.1