BinaryTieredBufferPool is a buffer pool that uses multiple sub-pools with power-of-two sizes.
| 50 | // BinaryTieredBufferPool is a buffer pool that uses multiple sub-pools with |
| 51 | // power-of-two sizes. |
| 52 | type BinaryTieredBufferPool struct { |
| 53 | // exponentToNextLargestPoolMap maps a power-of-two exponent (e.g., 12 for |
| 54 | // 4KB) to the index of the next largest sizedBufferPool. This is used by |
| 55 | // Get() to find the smallest pool that can satisfy a request for a given |
| 56 | // size. |
| 57 | exponentToNextLargestPoolMap []int |
| 58 | // exponentToPreviousLargestPoolMap maps a power-of-two exponent to the |
| 59 | // index of the previous largest sizedBufferPool. This is used by Put() |
| 60 | // to return a buffer to the most appropriate pool based on its capacity. |
| 61 | exponentToPreviousLargestPoolMap []int |
| 62 | sizedPools []bufferPool |
| 63 | fallbackPool bufferPool |
| 64 | maxPoolCap int // Optimization: Cache max capacity |
| 65 | } |
| 66 | |
| 67 | // NewBinaryTieredBufferPool returns a BufferPool backed by multiple sub-pools. |
| 68 | // This structure enables O(1) lookup time for Get and Put operations. |
nothing calls this directly
no outgoing calls
no test coverage detected