[Git][ghc/ghc][ghc-9.0] Naming, value types and tests for Addr# atomics
Ben Gamari
gitlab at gitlab.haskell.org
Sun Oct 25 15:11:15 UTC 2020
Ben Gamari pushed to branch ghc-9.0 at Glasgow Haskell Compiler / GHC
Commits:
5a2400c6 by Viktor Dukhovni at 2020-10-23T20:51:00-04:00
Naming, value types and tests for Addr# atomics
The atomic Exchange and CAS operations on integral types are updated to
take and return more natural `Word#` rather than `Int#` values. These
are bit-block not arithmetic operations, and the sign bit plays no
special role.
Standardises the names to `atomic<OpType><ValType>Addr#`, where `OpType` is one
of `Cas` or `Exchange` and `ValType` is presently either `Word` or `Addr`.
Eventually, variants for `Word32` and `Word64` can and should be added,
once #11953 and related issues (e.g. #13825) are resolved.
Adds tests for `Addr#` CAS that mirror existing tests for
`MutableByteArray#`.
- - - - -
8 changed files:
- compiler/GHC/Builtin/primops.txt.pp
- compiler/GHC/StgToCmm/Prim.hs
- libraries/base/GHC/Ptr.hs
- libraries/ghc-prim/changelog.md
- testsuite/tests/codeGen/should_compile/cg011.hs
- testsuite/tests/codeGen/should_run/cgrun080.hs
- testsuite/tests/concurrent/should_run/AtomicPrimops.hs
- testsuite/tests/concurrent/should_run/AtomicPrimops.stdout
Changes:
=====================================
compiler/GHC/Builtin/primops.txt.pp
=====================================
@@ -2471,17 +2471,47 @@ primop WriteOffAddrOp_Word64 "writeWord64OffAddr#" GenPrimOp
with has_side_effects = True
can_fail = True
-primop InterlockedExchange_Addr "interlockedExchangeAddr#" GenPrimOp
+primop InterlockedExchange_Addr "atomicExchangeAddrAddr#" GenPrimOp
Addr# -> Addr# -> State# s -> (# State# s, Addr# #)
{The atomic exchange operation. Atomically exchanges the value at the first address
with the Addr# given as second argument. Implies a read barrier.}
with has_side_effects = True
+ can_fail = True
-primop InterlockedExchange_Int "interlockedExchangeInt#" GenPrimOp
- Addr# -> Int# -> State# s -> (# State# s, Int# #)
+primop InterlockedExchange_Word "atomicExchangeWordAddr#" GenPrimOp
+ Addr# -> Word# -> State# s -> (# State# s, Word# #)
{The atomic exchange operation. Atomically exchanges the value at the address
with the given value. Returns the old value. Implies a read barrier.}
with has_side_effects = True
+ can_fail = True
+
+primop CasAddrOp_Addr "atomicCasAddrAddr#" GenPrimOp
+ Addr# -> Addr# -> Addr# -> State# s -> (# State# s, Addr# #)
+ { Compare and swap on a word-sized memory location.
+
+ Use as: \s -> atomicCasAddrAddr# location expected desired s
+
+ This version always returns the old value read. This follows the normal
+ protocol for CAS operations (and matches the underlying instruction on
+ most architectures).
+
+ Implies a full memory barrier.}
+ with has_side_effects = True
+ can_fail = True
+
+primop CasAddrOp_Word "atomicCasWordAddr#" GenPrimOp
+ Addr# -> Word# -> Word# -> State# s -> (# State# s, Word# #)
+ { Compare and swap on a word-sized and aligned memory location.
+
+ Use as: \s -> atomicCasWordAddr# location expected desired s
+
+ This version always returns the old value read. This follows the normal
+ protocol for CAS operations (and matches the underlying instruction on
+ most architectures).
+
+ Implies a full memory barrier.}
+ with has_side_effects = True
+ can_fail = True
------------------------------------------------------------------------
section "Mutable variables"
=====================================
compiler/GHC/StgToCmm/Prim.hs
=====================================
@@ -846,9 +846,14 @@ emitPrimOp dflags primop = case primop of
-- Atomic operations
InterlockedExchange_Addr -> \[src, value] -> opIntoRegs $ \[res] ->
emitPrimCall [res] (MO_Xchg (wordWidth platform)) [src, value]
- InterlockedExchange_Int -> \[src, value] -> opIntoRegs $ \[res] ->
+ InterlockedExchange_Word -> \[src, value] -> opIntoRegs $ \[res] ->
emitPrimCall [res] (MO_Xchg (wordWidth platform)) [src, value]
+ CasAddrOp_Addr -> \[dst, expected, new] -> opIntoRegs $ \[res] ->
+ emitPrimCall [res] (MO_Cmpxchg (wordWidth platform)) [dst, expected, new]
+ CasAddrOp_Word -> \[dst, expected, new] -> opIntoRegs $ \[res] ->
+ emitPrimCall [res] (MO_Cmpxchg (wordWidth platform)) [dst, expected, new]
+
-- SIMD primops
(VecBroadcastOp vcat n w) -> \[e] -> opIntoRegs $ \[res] -> do
checkVecCompatibility dflags vcat n w
=====================================
libraries/base/GHC/Ptr.hs
=====================================
@@ -171,7 +171,7 @@ castPtrToFunPtr (Ptr addr) = FunPtr addr
exchangePtr :: Ptr (Ptr a) -> Ptr b -> IO (Ptr c)
exchangePtr (Ptr dst) (Ptr val) =
IO $ \s ->
- case (interlockedExchangeAddr# dst val s) of
+ case (atomicExchangeAddrAddr# dst val s) of
(# s2, old_val #) -> (# s2, Ptr old_val #)
------------------------------------------------------------------------
=====================================
libraries/ghc-prim/changelog.md
=====================================
@@ -21,8 +21,13 @@
- Add primops for atomic exchange:
- interlockedExchangeAddr# :: Addr# -> Addr# -> State# s -> (# State# s, Addr# #)
- interlockedExchangeInt# :: Addr# -> Int# -> State# s -> (# State# s, Int# #)
+ atomicExchangeAddrAddr# :: Addr# -> Addr# -> State# s -> (# State# s, Addr# #)
+ atomicExchangeWordAddr# :: Addr# -> Word# -> State# s -> (# State# s, Word# #)
+
+- Add primops for atomic compare and swap at a given Addr#:
+
+ atomicCasAddrAddr# :: Addr# -> Addr# -> Addr# -> State# s -> (# State# s, Addr# #)
+ atomicCasWordAddr# :: Addr# -> Word# -> Word# -> State# s -> (# State# s, Word# #)
- Add an explicit fixity for `(~)` and `(~~)`:
=====================================
testsuite/tests/codeGen/should_compile/cg011.hs
=====================================
@@ -1,11 +1,11 @@
{-# LANGUAGE CPP, MagicHash, BlockArguments, UnboxedTuples #-}
--- Tests compilation for interlockedExchange primop.
+-- Tests compilation for atomicExchangeWordAddr# primop.
module M where
-import GHC.Exts (interlockedExchangeInt#, Int#, Addr#, State# )
+import GHC.Exts (atomicExchangeWordAddr#, Word#, Addr#, State# )
-swap :: Addr# -> Int# -> State# s -> (# #)
-swap ptr val s = case (interlockedExchangeInt# ptr val s) of
+swap :: Addr# -> Word# -> State# s -> (# #)
+swap ptr val s = case (atomicExchangeWordAddr# ptr val s) of
(# s2, old_val #) -> (# #)
=====================================
testsuite/tests/codeGen/should_run/cgrun080.hs
=====================================
@@ -25,8 +25,8 @@ import GHC.Types
main = do
alloca $ \ptr_i -> do
- poke ptr_i (1 :: Int)
- w1 <- newEmptyMVar :: IO (MVar Int)
+ poke ptr_i (1 :: Word)
+ w1 <- newEmptyMVar :: IO (MVar Word)
forkIO $ do
v <- swapN 50000 2 ptr_i
putMVar w1 v
@@ -37,15 +37,14 @@ main = do
-- Should be [1,2,3]
print $ sort [v0,v1,v2]
-swapN :: Int -> Int -> Ptr Int -> IO Int
+swapN :: Word -> Word -> Ptr Word -> IO Word
swapN 0 val ptr = return val
swapN n val ptr = do
val' <- swap ptr val
swapN (n-1) val' ptr
-swap :: Ptr Int -> Int -> IO Int
-swap (Ptr ptr) (I# val) = do
- IO $ \s -> case (interlockedExchangeInt# ptr val s) of
- (# s2, old_val #) -> (# s2, I# old_val #)
-
+swap :: Ptr Word -> Word -> IO Word
+swap (Ptr ptr) (W# val) = do
+ IO $ \s -> case (atomicExchangeWordAddr# ptr val s) of
+ (# s2, old_val #) -> (# s2, W# old_val #)
=====================================
testsuite/tests/concurrent/should_run/AtomicPrimops.hs
=====================================
@@ -6,6 +6,8 @@ module Main ( main ) where
import Control.Concurrent
import Control.Concurrent.MVar
import Control.Monad (when)
+import Foreign.Marshal.Alloc
+import Foreign.Ptr
import Foreign.Storable
import GHC.Exts
import GHC.IO
@@ -22,6 +24,7 @@ main = do
fetchOrTest
fetchXorTest
casTest
+ casTestAddr
readWriteTest
-- | Test fetchAddIntArray# by having two threads concurrenctly
@@ -54,12 +57,14 @@ fetchXorTest = do
work mba 0 val = return ()
work mba n val = fetchXorIntArray mba 0 val >> work mba (n-1) val
- -- Initial value is a large prime and the two patterns are 1010...
- -- and 0101...
+ -- The two patterns are 1010... and 0101... The second pattern is larger
+ -- than maxBound, avoid warnings by initialising as a Word.
(n0, t1pat, t2pat)
| sizeOf (undefined :: Int) == 8 =
- (0x00000000ffffffff, 0x5555555555555555, 0x9999999999999999)
- | otherwise = (0x0000ffff, 0x55555555, 0x99999999)
+ ( 0x00000000ffffffff, 0x5555555555555555
+ , fromIntegral (0x9999999999999999 :: Word))
+ | otherwise = ( 0x0000ffff, 0x55555555
+ , fromIntegral (0x99999999 :: Word))
expected
| sizeOf (undefined :: Int) == 8 = 4294967295
| otherwise = 65535
@@ -90,13 +95,15 @@ fetchOpTest op expected name = do
-- | Initial value and operation arguments for race test.
--
--- Initial value is a large prime and the two patterns are 1010...
--- and 0101...
+-- The two patterns are 1010... and 0101... The second pattern is larger than
+-- maxBound, avoid warnings by initialising as a Word.
n0, t1pat, t2pat :: Int
(n0, t1pat, t2pat)
| sizeOf (undefined :: Int) == 8 =
- (0x00000000ffffffff, 0x5555555555555555, 0x9999999999999999)
- | otherwise = (0x0000ffff, 0x55555555, 0x99999999)
+ ( 0x00000000ffffffff, 0x5555555555555555
+ , fromIntegral (0x9999999999999999 :: Word))
+ | otherwise = ( 0x0000ffff, 0x55555555
+ , fromIntegral (0x99999999 :: Word))
fetchAndTest :: IO ()
fetchAndTest = fetchOpTest fetchAndIntArray expected "fetchAndTest"
@@ -120,8 +127,10 @@ fetchNandTest = do
fetchOrTest :: IO ()
fetchOrTest = fetchOpTest fetchOrIntArray expected "fetchOrTest"
where expected
- | sizeOf (undefined :: Int) == 8 = 15987178197787607039
- | otherwise = 3722313727
+ | sizeOf (undefined :: Int) == 8
+ = fromIntegral (15987178197787607039 :: Word)
+ | otherwise
+ = fromIntegral (3722313727 :: Word)
-- | Test casIntArray# by using it to emulate fetchAddIntArray# and
-- then having two threads concurrenctly increment a counter,
@@ -131,7 +140,7 @@ casTest = do
tot <- race 0
(\ mba -> work mba iters 1)
(\ mba -> work mba iters 2)
- assertEq 3000000 tot "casTest"
+ assertEq (3 * iters) tot "casTest"
where
work :: MByteArray -> Int -> Int -> IO ()
work mba 0 val = return ()
@@ -179,6 +188,45 @@ race n0 thread1 thread2 = do
mapM_ takeMVar [done1, done2]
readIntArray mba 0
+-- | Test atomicCasWordAddr# by having two threads concurrenctly increment a
+-- counter, checking the sum at the end.
+casTestAddr :: IO ()
+casTestAddr = do
+ tot <- raceAddr 0
+ (\ addr -> work addr (fromIntegral iters) 1)
+ (\ addr -> work addr (fromIntegral iters) 2)
+ assertEq (3 * fromIntegral iters) tot "casTestAddr"
+ where
+ work :: Ptr Word -> Word -> Word -> IO ()
+ work ptr 0 val = return ()
+ work ptr n val = add ptr val >> work ptr (n-1) val
+
+ -- Fetch-and-add implemented using CAS.
+ add :: Ptr Word -> Word -> IO ()
+ add ptr n = peek ptr >>= go
+ where
+ go old = do
+ old' <- atomicCasWordPtr ptr old (old + n)
+ when (old /= old') $ go old'
+
+ -- | Create two threads that mutate the byte array passed to them
+ -- concurrently. The array is one word large.
+ raceAddr :: Word -- ^ Initial value of array element
+ -> (Ptr Word -> IO ()) -- ^ Thread 1 action
+ -> (Ptr Word -> IO ()) -- ^ Thread 2 action
+ -> IO Word -- ^ Final value of array element
+ raceAddr n0 thread1 thread2 = do
+ done1 <- newEmptyMVar
+ done2 <- newEmptyMVar
+ ptr <- asWordPtr <$> callocBytes (sizeOf (undefined :: Word))
+ forkIO $ thread1 ptr >> putMVar done1 ()
+ forkIO $ thread2 ptr >> putMVar done2 ()
+ mapM_ takeMVar [done1, done2]
+ peek ptr
+ where
+ asWordPtr :: Ptr a -> Ptr Word
+ asWordPtr = castPtr
+
------------------------------------------------------------------------
-- Test helper
@@ -254,3 +302,13 @@ casIntArray :: MByteArray -> Int -> Int -> Int -> IO Int
casIntArray (MBA mba#) (I# ix#) (I# old#) (I# new#) = IO $ \ s# ->
case casIntArray# mba# ix# old# new# s# of
(# s2#, old2# #) -> (# s2#, I# old2# #)
+
+------------------------------------------------------------------------
+-- Wrappers around Addr#
+
+-- Should this be added to Foreign.Storable? Similar to poke, but does the
+-- update atomically.
+atomicCasWordPtr :: Ptr Word -> Word -> Word -> IO Word
+atomicCasWordPtr (Ptr addr#) (W# old#) (W# new#) = IO $ \ s# ->
+ case atomicCasWordAddr# addr# old# new# s# of
+ (# s2#, old2# #) -> (# s2#, W# old2# #)
=====================================
testsuite/tests/concurrent/should_run/AtomicPrimops.stdout
=====================================
@@ -4,4 +4,5 @@ fetchNandTest: OK
fetchOrTest: OK
fetchXorTest: OK
casTest: OK
+casTestAddr: OK
readWriteTest: OK
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/5a2400c6570e4069b04f8d727c9058620bb99f3c
--
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/5a2400c6570e4069b04f8d727c9058620bb99f3c
You're receiving this email because of your account on gitlab.haskell.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20201025/0c8899cf/attachment-0001.html>
More information about the ghc-commits
mailing list