[Git][ghc/ghc][wip/cfuneqcan-refactor] 151 commits: SMP.h: Add C11-style atomic operations

Richard Eisenberg gitlab at gitlab.haskell.org
Wed Nov 25 20:33:05 UTC 2020



Richard Eisenberg pushed to branch wip/cfuneqcan-refactor at Glasgow Haskell Compiler / GHC


Commits:
b9d4dd9c by Ben Gamari at 2020-10-24T20:44:17-04:00
SMP.h: Add C11-style atomic operations

- - - - -
ccf2d4b0 by Ben Gamari at 2020-10-24T20:59:39-04:00
rts: Infrastructure for testing with ThreadSanitizer

- - - - -
a61f66d6 by Ben Gamari at 2020-10-24T20:59:39-04:00
rts/CNF: Initialize all bdescrs in group

It seems wise and cheap to ensure that the whole bdescr of all blocks of
a compact group is valid, even if most cases only look at the flags
field.

- - - - -
65136c13 by Ben Gamari at 2020-10-24T20:59:39-04:00
rts/Capability: Intialize interrupt field

Previously this was left uninitialized.

Also clarify some comments.

- - - - -
b3ce6aca by Ben Gamari at 2020-10-24T20:59:39-04:00
rts/Task: Make comments proper Notes

- - - - -
d3890ac7 by Ben Gamari at 2020-10-24T20:59:39-04:00
rts/SpinLock: Move to proper atomics

This is fairly straightforward; we just needed to use relaxed operations
for the PROF_SPIN counters and a release store instead of a write
barrier.

- - - - -
ef88712f by Ben Gamari at 2020-10-24T20:59:39-04:00
rts/OSThreads: Fix data race

Previously we would race on the cached processor count. Avoiding this is
straightforward; just use relaxed operations.

- - - - -
33a719c3 by Ben Gamari at 2020-10-24T20:59:39-04:00
rts/ClosureMaros: Use relaxed atomics

- - - - -
f08951fd by Ben Gamari at 2020-10-24T20:59:39-04:00
configure: Bump minimum-supported gcc version to 4.7

Since the __atomic_* builtins are not supported until gcc 4.7. Given
that this version was released in 2012 I think this is acceptable.

- - - - -
d584923a by Ben Gamari at 2020-10-24T20:59:39-04:00
testsuite: Fix thread leak in hs_try_putmvar00[13]

- - - - -
bf1b0bc7 by Ben Gamari at 2020-10-24T20:59:39-04:00
rts: Introduce SET_HDR_RELEASE

Also ensure that we also store the info table pointer last to ensure
that the synchronization covers all stores.

- - - - -
1a2e9f5e by Ben Gamari at 2020-10-24T21:00:19-04:00
gitlab-ci: Add nightly-x86_64-linux-deb9-tsan job

- - - - -
58a5b0e5 by GHC GitLab CI at 2020-10-24T21:00:19-04:00
testsuite: Mark setnumcapabilities001 as broken with TSAN

Due to #18808.

- - - - -
d9bc7dea by GHC GitLab CI at 2020-10-24T21:00:19-04:00
testsuite: Skip divbyzero and derefnull under TSAN

ThreadSanitizer changes the output of these tests.

- - - - -
fcc42a10 by Ben Gamari at 2020-10-24T21:00:19-04:00
testsuite: Skip high memory usage tests with TSAN

ThreadSanitizer significantly increases the memory footprint of tests,
so much so that it can send machines into OOM.

- - - - -
cae4bb3e by Ben Gamari at 2020-10-24T21:00:19-04:00
testsuite: Mark hie002 as high_memory_usage

This test has a peak residency of 1GByte; this is large enough to
classify as "high" in my book.

- - - - -
dae1b86a by Ben Gamari at 2020-10-24T21:00:19-04:00
testsuite: Mark T9872[abc] as high_memory_usage

These all have a maximum residency of over 2 GB.

- - - - -
c5a0bb22 by Ben Gamari at 2020-10-24T21:00:19-04:00
gitlab-ci: Disable documentation in TSAN build

Haddock chews through enough memory to cause the CI builders to OOM and
there's frankly no reason to build documentation in this job anyways.

- - - - -
4cb1232e by Ben Gamari at 2020-10-24T21:00:19-04:00
TSANUtils: Ensure that C11 atomics are supported

- - - - -
7ed15f7f by Ben Gamari at 2020-10-24T21:00:19-04:00
testsuite: Mark T3807 as broken with TSAN

Due to #18883.

- - - - -
f7e6f012 by Ben Gamari at 2020-10-24T21:00:19-04:00
testsuite: Mark T13702 as broken with TSAN due to #18884

- - - - -
16b136b0 by Ben Gamari at 2020-10-24T21:00:36-04:00
rts: Factor out logic to identify a good capability for running a task

Not only does this make the control flow a bit clearer but it also
allows us to add a TSAN suppression on this logic, which requires
(harmless) data races.

- - - - -
2781d68c by Ben Gamari at 2020-10-24T21:00:36-04:00
rts: Annotate benign race in waitForCapability

- - - - -
f6b4b492 by Ben Gamari at 2020-10-24T21:00:36-04:00
rts: Clarify locking behavior of releaseCapability_

- - - - -
65219810 by Ben Gamari at 2020-10-24T21:00:36-04:00
rts: Add assertions for task ownership of capabilities

- - - - -
31fa87ec by Ben Gamari at 2020-10-24T21:00:36-04:00
rts: Use relaxed atomics on n_returning_tasks

This mitigates the warning of a benign race on n_returning_tasks in
shouldYieldCapability.

See #17261.

- - - - -
6517a2ea by Ben Gamari at 2020-10-24T21:00:36-04:00
rts: Mitigate races in capability interruption logic

- - - - -
2e9ba3f2 by Ben Gamari at 2020-10-24T21:00:36-04:00
rts/Capability: Use relaxed operations for last_free_capability

- - - - -
e10dde37 by Ben Gamari at 2020-10-24T21:00:37-04:00
rts: Use relaxed operations for cap->running_task (TODO)

This shouldn't be necessary since only the owning thread of the capability
should be touching this.

- - - - -
855325cd by Ben Gamari at 2020-10-24T21:00:37-04:00
rts/Schedule: Use relaxed operations for sched_state

- - - - -
811f915d by Ben Gamari at 2020-10-24T21:00:37-04:00
rts: Accept data race in work-stealing implementation

This race is okay since the task is owned by the capability pushing it.
By Note [Ownership of Task] this means that the capability is free to
write to `task->cap` without taking `task->lock`.

Fixes #17276.

- - - - -
8d2b3c3d by Ben Gamari at 2020-10-24T21:00:37-04:00
rts: Eliminate data races on pending_sync

- - - - -
f8871018 by Ben Gamari at 2020-10-24T21:00:37-04:00
rts/Schedule: Eliminate data races on recent_activity

We cannot safely use relaxed atomics here.

- - - - -
d079b943 by Ben Gamari at 2020-10-24T21:00:37-04:00
rts: Avoid data races in message handling

- - - - -
06f80497 by Ben Gamari at 2020-10-24T21:00:37-04:00
rts/Messages: Drop incredibly fishy write barrier

executeMessage previously had a write barrier at the beginning of its
loop apparently in an attempt to synchronize with another thread's
writes to the Message. I would guess that the author had intended to use
a load barrier here given that there are no globally-visible writes done
in executeMessage.

I've removed the redundant barrier since the necessary load barrier is
now provided by the ACQUIRE_LOAD.

- - - - -
d4a87779 by Ben Gamari at 2020-10-24T21:00:38-04:00
rts/ThreadPaused: Avoid data races

- - - - -
56778ab3 by Ben Gamari at 2020-10-24T21:00:38-04:00
rts/Schedule: Eliminate data races in run queue management

- - - - -
086521f7 by Ben Gamari at 2020-10-24T21:00:38-04:00
rts: Eliminate shutdown data race on task counters

- - - - -
abad9778 by Ben Gamari at 2020-10-24T21:00:38-04:00
rts/Threads: Avoid data races (TODO)

Replace barriers with appropriate ordering. Drop redundant barrier in
tryWakeupThread (the RELEASE barrier will be provided by sendMessage's
mutex release).

We use relaxed operations on why_blocked and the stack although it's not
clear to me why this is necessary.

- - - - -
2f56be8a by Ben Gamari at 2020-10-24T21:00:39-04:00
rts/Messages: Annotate benign race

- - - - -
7c0cdab1 by Ben Gamari at 2020-10-24T21:00:39-04:00
rts/RaiseAsync: Synchronize what_next read

- - - - -
6cc2a8a5 by Ben Gamari at 2020-10-24T21:00:39-04:00
rts/Task: Move debugTrace to avoid data race

Specifically, we need to hold all_tasks_mutex to read taskCount.

- - - - -
bbaec97d by Ben Gamari at 2020-10-24T21:00:39-04:00
Disable flawed assertion

- - - - -
dd175a92 by Ben Gamari at 2020-10-24T21:00:39-04:00
Document schedulePushWork race

- - - - -
3416244b by Ben Gamari at 2020-10-24T21:00:40-04:00
Capabiliity: Properly fix data race on n_returning_tasks

There is a real data race but can be made safe by using proper atomic
(but relaxed) accesses.

- - - - -
dffd9432 by Ben Gamari at 2020-10-24T21:00:40-04:00
rts: Make write of to_cap->inbox atomic

This is necessary since emptyInbox may read from to_cap->inbox without
taking cap->lock.

- - - - -
1f4cbc29 by Ben Gamari at 2020-10-24T21:00:57-04:00
rts/BlockAlloc: Use relaxed operations

- - - - -
d0d07cff by Ben Gamari at 2020-10-24T21:00:57-04:00
rts: Rework handling of mutlist scavenging statistics

- - - - -
9e5c7f6d by Ben Gamari at 2020-10-24T21:00:57-04:00
rts: Avoid data races in StablePtr implementation

This fixes two potentially problematic data races in the StablePtr
implementation:

 * We would fail to RELEASE the stable pointer table when enlarging it,
   causing other cores to potentially see uninitialized memory.

 * We would fail to ACQUIRE when dereferencing a stable pointer.

- - - - -
316add67 by Ben Gamari at 2020-10-24T21:00:57-04:00
rts/Storage: Use atomics

- - - - -
5c23bc4c by Ben Gamari at 2020-10-24T21:00:58-04:00
rts/Updates: Use proper atomic operations

- - - - -
3d0f033c by Ben Gamari at 2020-10-24T21:00:58-04:00
rts/Weak: Eliminate data races

By taking all_tasks_mutex in stat_exit. Also better-document the fact
that the task statistics are protected by all_tasks_mutex.

- - - - -
edb4b92b by Ben Gamari at 2020-10-24T21:01:18-04:00
rts/WSDeque: Rewrite with proper atomics

After a few attempts at shoring up the previous implementation, I ended
up turning to the literature and now use the proven implementation,

> N.M. Lê, A. Pop, A.Cohen, and F.Z. Nardelli. "Correct and Efficient
> Work-Stealing for Weak Memory Models". PPoPP'13, February 2013,
> ACM 978-1-4503-1922/13/02.

Note only is this approach formally proven correct under C11 semantics
but it is also proved to be a bit faster in practice.

- - - - -
d39bbd3d by Ben Gamari at 2020-10-24T21:01:33-04:00
rts: Use relaxed atomics for whitehole spin stats

- - - - -
8f802f38 by Ben Gamari at 2020-10-24T21:01:33-04:00
rts: Avoid lock order inversion during fork

Fixes #17275.

- - - - -
cef667b0 by GHC GitLab CI at 2020-10-24T21:01:34-04:00
rts: Use proper relaxe operations in getCurrentThreadCPUTime

Here we are doing lazy initialization; it's okay if we do the check more
than once, hence relaxed operation is fine.

- - - - -
8cf50eb1 by Ben Gamari at 2020-10-24T21:01:54-04:00
rts/STM: Use atomics

This fixes a potentially harmful race where we failed to synchronize
before looking at a TVar's current_value.

Also did a bit of refactoring to avoid abstract over management of
max_commits.

- - - - -
88a7ce38 by Ben Gamari at 2020-10-24T21:01:54-04:00
rts/stm: Strengthen orderings to SEQ_CST instead of volatile

Previously the `current_value`, `first_watch_queue_entry`, and
`num_updates` fields of `StgTVar` were marked as `volatile` in an
attempt to provide strong ordering. Of course, this isn't sufficient.

We now use proper atomic operations. In most of these cases I strengthen
the ordering all the way to SEQ_CST although it's possible that some
could be weakened with some thought.

- - - - -
f97c59ce by Ben Gamari at 2020-10-24T21:02:11-04:00
Mitigate data races in event manager startup/shutdown

- - - - -
c7c3f8aa by Ben Gamari at 2020-10-24T21:02:22-04:00
rts: Accept benign races in Proftimer

- - - - -
5a98dfca by Ben Gamari at 2020-10-24T21:02:22-04:00
rts: Pause timer while changing capability count

This avoids #17289.

- - - - -
01d95525 by Ben Gamari at 2020-10-24T21:02:22-04:00
Fix #17289

- - - - -
9a528985 by Ben Gamari at 2020-10-24T21:02:23-04:00
suppress #17289 (ticker) race

- - - - -
1726ec41 by Ben Gamari at 2020-10-24T21:02:23-04:00
rts: Fix timer initialization

Previously `initScheduler` would attempt to pause the ticker and in so
doing acquire the ticker mutex. However, initTicker, which is
responsible for initializing said mutex, hadn't been called
yet.

- - - - -
bfbe4366 by Ben Gamari at 2020-10-24T21:02:23-04:00
rts: Fix races in Pthread timer backend shudown

We can generally be pretty relaxed in the barriers here since the timer
thread is a loop.

- - - - -
297acc71 by Ben Gamari at 2020-10-24T21:02:44-04:00
rts/Stats: Hide a few unused unnecessarily global functions

- - - - -
aad1f803 by Ben Gamari at 2020-10-30T00:41:14-04:00
rts/GC: Use atomics

- - - - -
d0bc0517 by Ben Gamari at 2020-10-30T00:41:14-04:00
rts: Use RELEASE ordering in unlockClosure

- - - - -
d44f5232 by Ben Gamari at 2020-10-30T00:41:14-04:00
rts/Storage: Accept races on heap size counters

- - - - -
4e4a7386 by Ben Gamari at 2020-10-30T00:41:14-04:00
rts: Join to concurrent mark thread during shutdown

Previously we would take all capabilities but fail to join on the thread
itself, potentially resulting in a leaked thread.

- - - - -
a80cc857 by GHC GitLab CI at 2020-10-30T00:41:14-04:00
rts: Fix race in GC CPU time accounting

Ensure that the GC leader synchronizes with workers before calling
stat_endGC.

- - - - -
105d43db by Ben Gamari at 2020-10-30T14:02:19-04:00
rts/SpinLock: Separate out slow path

Not only is this in general a good idea, but it turns out that GCC
unrolls the retry loop, resulting is massive code bloat in critical
parts of the RTS (e.g. `evacuate`).

- - - - -
f7b45cde by Ben Gamari at 2020-10-30T14:02:19-04:00
rts: Use relaxed ordering on spinlock counters

- - - - -
b1dda153 by Ben Gamari at 2020-11-01T12:58:36-05:00
rts/Stats: Protect with mutex

While on face value this seems a bit heavy, I think it's far better than
enforcing ordering on every access.

- - - - -
5c2e6bce by Ben Gamari at 2020-11-01T12:58:36-05:00
rts: Tear down stats_mutex after exitHeapProfiling

Since the latter wants to call getRTSStats.

- - - - -
ef25aaa1 by Ben Gamari at 2020-11-01T13:02:11-05:00
rts: Annotate hopefully "benign" races in freeGroup

- - - - -
3a181553 by Ben Gamari at 2020-11-01T13:02:18-05:00
Strengthen ordering in releaseGCThreads

- - - - -
af474f62 by Ben Gamari at 2020-11-01T13:05:38-05:00
Suppress data race due to close

This suppresses the other side of a race during shutdown.

- - - - -
b4686bff by Ben Gamari at 2020-11-01T13:09:59-05:00
Merge branch 'wip/tsan/ci' into wip/tsan/all

- - - - -
b8e66e0e by Ben Gamari at 2020-11-01T13:10:01-05:00
Merge branch 'wip/tsan/storage' into wip/tsan/all

- - - - -
375512cf by Ben Gamari at 2020-11-01T13:10:02-05:00
Merge branch 'wip/tsan/wsdeque' into wip/tsan/all

- - - - -
65ebf07e by Ben Gamari at 2020-11-01T13:10:03-05:00
Merge branch 'wip/tsan/misc' into wip/tsan/all

- - - - -
55c375d0 by Ben Gamari at 2020-11-01T13:10:04-05:00
Merge branch 'wip/tsan/stm' into wip/tsan/all

- - - - -
a9f75fe2 by Ben Gamari at 2020-11-01T13:10:06-05:00
Merge branch 'wip/tsan/event-mgr' into wip/tsan/all

- - - - -
8325d658 by Ben Gamari at 2020-11-01T13:10:24-05:00
Merge branch 'wip/tsan/timer' into wip/tsan/all

- - - - -
07e82ba5 by Ben Gamari at 2020-11-01T13:10:35-05:00
Merge branch 'wip/tsan/stats' into wip/tsan/all

- - - - -
638f38c5 by Ben Gamari at 2020-11-08T09:29:16-05:00
Merge remote-tracking branch 'origin/wip/tsan/all'

- - - - -
22888798 by Ben Gamari at 2020-11-08T12:08:40-05:00
Fix haddock submodule

The previous merge mistakenly reverted it.

- - - - -
d445cf05 by Ben Gamari at 2020-11-10T10:26:20-05:00
rts/linker: Fix relocation overflow in PE linker

Previously the overflow check for the IMAGE_REL_AMD64_ADDR32NB
relocation failed to account for the signed nature of the value.
Specifically, the overflow check was:

    uint64_t v;
    v = S + A;
    if (v >> 32) { ... }

However, `v` ultimately needs to fit into 32-bits as a signed value.
Consequently, values `v > 2^31` in fact overflow yet this is not caught
by the existing overflow check.

Here we rewrite the overflow check to rather ensure that
`INT32_MIN <= v <= INT32_MAX`. There is now quite a bit of repetition
between the `IMAGE_REL_AMD64_REL32` and `IMAGE_REL_AMD64_ADDR32` cases
but I am leaving fixing this for future work.

This bug was first noticed by @awson.

Fixes #15808.

- - - - -
4c407f6e by Sylvain Henry at 2020-11-10T10:27:00-05:00
Export SPEC from GHC.Exts (#13681)

- - - - -
7814cd5b by David Eichmann at 2020-11-10T10:27:35-05:00
ghc-heap: expose decoding from heap representation

Co-authored-by: Sven Tennie <sven.tennie at gmail.com>
Co-authored-by: Matthew Pickering <matthewtpickering at gmail.com>
Co-authored-by: Ben Gamari <bgamari.foss at gmail.com>

- - - - -
fa344d33 by Richard Eisenberg at 2020-11-10T10:28:10-05:00
Add test case for #17186.

This got fixed sometime recently; not worth it trying to
figure out which commit.

- - - - -
2e63a0fb by David Eichmann at 2020-11-10T10:28:46-05:00
Add code comments for StgInfoTable and StgStack structs

- - - - -
fcfda909 by Ben Gamari at 2020-11-11T03:19:59-05:00
nativeGen: Make makeImportsDoc take an NCGConfig rather than DynFlags

It appears this was an oversight as there is no reason the full DynFlags
is necessary.

- - - - -
6e23695e by Ben Gamari at 2020-11-11T03:19:59-05:00
Move this_module into NCGConfig

In various places in the NCG we need the Module currently being
compiled. Let's move this into the environment instead of chewing threw
another register.

- - - - -
c6264a2d by Ben Gamari at 2020-11-11T03:20:00-05:00
codeGen: Produce local symbols for module-internal functions

It turns out that some important native debugging/profiling tools (e.g.
perf) rely only on symbol tables for function name resolution (as
opposed to using DWARF DIEs). However, previously GHC would emit
temporary symbols (e.g. `.La42b`) to identify module-internal
entities. Such symbols are dropped during linking and therefore not
visible to runtime tools (in addition to having rather un-helpful unique
names). For instance, `perf report` would often end up attributing all
cost to the libc `frame_dummy` symbol since Haskell code was no covered
by any proper symbol (see #17605).

We now rather follow the model of C compilers and emit
descriptively-named local symbols for module internal things. Since this
will increase object file size this behavior can be disabled with the
`-fno-expose-internal-symbols` flag.

With this `perf record` can finally be used against Haskell executables.
Even more, with `-g3` `perf annotate` provides inline source code.

- - - - -
584058dd by Ben Gamari at 2020-11-11T03:20:00-05:00
Enable -fexpose-internal-symbols when debug level >=2

This seems like a reasonable default as the object file size increases
by around 5%.

- - - - -
c34a4b98 by Ömer Sinan Ağacan at 2020-11-11T03:20:35-05:00
Fix and enable object unloading in GHCi

Fixes #16525 by tracking dependencies between object file symbols and
marking symbol liveness during garbage collection

See Note [Object unloading] in CheckUnload.c for details.

- - - - -
2782487f by Ray Shih at 2020-11-11T03:20:35-05:00
Add loadNativeObj and unloadNativeObj

(This change is originally written by niteria)

This adds two functions:
* `loadNativeObj`
* `unloadNativeObj`
and implements them for Linux.

They are useful if you want to load a shared object with Haskell code
using the system linker and have GHC call dlclose() after the
code is no longer referenced from the heap.

Using the system linker allows you to load the shared object
above outside the low-mem region. It also loads the DWARF sections
in a way that `perf` understands.

`dl_iterate_phdr` is what makes this implementation Linux specific.

- - - - -
7a65f9e1 by GHC GitLab CI at 2020-11-11T03:20:35-05:00
rts: Introduce highMemDynamic

- - - - -
e9e1b2e7 by GHC GitLab CI at 2020-11-11T03:20:35-05:00
Introduce test for dynamic library unloading

This uses the highMemDynamic flag introduced earlier to verify that
dynamic objects are properly unloaded.

- - - - -
5506f134 by Krzysztof Gogolewski at 2020-11-11T03:21:14-05:00
Force argument in setIdMult (#18925)

- - - - -
787e93ae by Ben Gamari at 2020-11-11T23:14:11-05:00
testsuite: Add testcase for #18733

- - - - -
5353fd50 by Ben Gamari at 2020-11-12T10:05:30-05:00
compiler: Fix recompilation checking

In ticket #18733 we noticed a rather serious deficiency in the current
fingerprinting logic for recursive groups. I have described the old
fingerprinting story and its problems in Note [Fingerprinting recursive
groups] and have reworked the story accordingly to avoid these issues.

Fixes #18733.

- - - - -
63fa3997 by Sebastian Graf at 2020-11-13T14:29:39-05:00
Arity: Rework `ArityType` to fix monotonicity (#18870)

As we found out in #18870, `andArityType` is not monotone, with
potentially severe consequences for termination of fixed-point
iteration. That showed in an abundance of "Exciting arity" DEBUG
messages that are emitted whenever we do more than one step in
fixed-point iteration.

The solution necessitates also recording `OneShotInfo` info for
`ABot` arity type. Thus we get the following definition for `ArityType`:

```
data ArityType = AT [OneShotInfo] Divergence
```

The majority of changes in this patch are the result of refactoring use
sites of `ArityType` to match the new definition.

The regression test `T18870` asserts that we indeed don't emit any DEBUG
output anymore for a function where we previously would have.
Similarly, there's a regression test `T18937` for #18937, which we
expect to be broken for now.

Fixes #18870.

- - - - -
197d59fa by Sebastian Graf at 2020-11-13T14:29:39-05:00
Arity: Emit "Exciting arity" warning only after second iteration (#18937)

See Note [Exciting arity] why we emit the warning at all and why we only
do after the second iteration now.

Fixes #18937.

- - - - -
de7ec9dd by David Eichmann at 2020-11-13T14:30:16-05:00
Add rts_listThreads and rts_listMiscRoots to RtsAPI.h

These are used to find the current roots of the garbage collector.

Co-authored-by: Sven Tennie's avatarSven Tennie <sven.tennie at gmail.com>
Co-authored-by: Matthew Pickering's avatarMatthew Pickering <matthewtpickering at gmail.com>
Co-authored-by: default avatarBen Gamari <bgamari.foss at gmail.com>

- - - - -
24a86f09 by Ben Gamari at 2020-11-13T14:30:51-05:00
gitlab-ci: Cache cabal store in linting job

- - - - -
0a7e592c by Ben Gamari at 2020-11-15T03:35:45-05:00
nativeGen/dwarf: Fix procedure end addresses

Previously the `.debug_aranges` and `.debug_info` (DIE) DWARF
information would claim that procedures (represented with a
`DW_TAG_subprogram` DIE) would only span the range covered by their entry
block. This omitted all of the continuation blocks (represented by
`DW_TAG_lexical_block` DIEs), confusing `perf`. Fix this by introducing
a end-of-procedure label and using this as the `DW_AT_high_pc` of
procedure `DW_TAG_subprogram` DIEs

Fixes #17605.

- - - - -
1e19183d by Ben Gamari at 2020-11-15T03:35:45-05:00
nativeGen/dwarf: Only produce DW_AT_source_note DIEs in -g3

Standard debugging tools don't know how to understand these so let's not
produce them unless asked.

- - - - -
ad73370f by Ben Gamari at 2020-11-15T03:35:45-05:00
nativeGen/dwarf: Use DW_AT_linkage instead of DW_AT_MIPS_linkage

- - - - -
a2539650 by Ben Gamari at 2020-11-15T03:35:45-05:00
gitlab-ci: Add DWARF release jobs for Debian 10, Fedora27

- - - - -
d61adb3d by Ryan Scott at 2020-11-15T03:36:21-05:00
Name (tc)SplitForAll- functions more consistently

There is a zoo of `splitForAll-` functions in `GHC.Core.Type` (as well as
`tcSplitForAll-` functions in `GHC.Tc.Utils.TcType`) that all do very similar
things, but vary in the particular form of type variable that they return. To
make things worse, the names of these functions are often quite misleading.
Some particularly egregious examples:

* `splitForAllTys` returns `TyCoVar`s, but `splitSomeForAllTys` returns
  `VarBndr`s.
* `splitSomeForAllTys` returns `VarBndr`s, but `tcSplitSomeForAllTys` returns
  `TyVar`s.
* `splitForAllTys` returns `TyCoVar`s, but `splitForAllTysInvis` returns
  `InvisTVBinder`s. (This in particular arose in the context of #18939, and
  this finally motivated me to bite the bullet and improve the status quo
  vis-à-vis how we name these functions.)

In an attempt to bring some sanity to how these functions are named, I have
opted to rename most of these functions en masse to use consistent suffixes
that describe the particular form of type variable that each function returns.
In concrete terms, this amounts to:

* Functions that return a `TyVar` now use the suffix `-TyVar`.
  This caused the following functions to be renamed:
  * `splitTyVarForAllTys` -> `splitForAllTyVars`
  * `splitForAllTy_ty_maybe` -> `splitForAllTyVar_maybe`
  * `tcSplitForAllTys` -> `tcSplitForAllTyVars`
  * `tcSplitSomeForAllTys` -> `tcSplitSomeForAllTyVars`
* Functions that return a `CoVar` now use the suffix `-CoVar`.
  This caused the following functions to be renamed:
  * `splitForAllTy_co_maybe` -> `splitForAllCoVar_maybe`
* Functions that return a `TyCoVar` now use the suffix `-TyCoVar`.
  This caused the following functions to be renamed:
  * `splitForAllTy` -> `splitForAllTyCoVar`
  * `splitForAllTys` -> `splitForAllTyCoVars`
  * `splitForAllTys'` -> `splitForAllTyCoVars'`
  * `splitForAllTy_maybe` -> `splitForAllTyCoVar_maybe`
* Functions that return a `VarBndr` now use the suffix corresponding to the
  most relevant type synonym. This caused the following functions to be renamed:
  * `splitForAllVarBndrs` -> `splitForAllTyCoVarBinders`
  * `splitForAllTysInvis` -> `splitForAllInvisTVBinders`
  * `splitForAllTysReq` -> `splitForAllReqTVBinders`
  * `splitSomeForAllTys` -> `splitSomeForAllTyCoVarBndrs`
  * `tcSplitForAllVarBndrs` -> `tcSplitForAllTyVarBinders`
  * `tcSplitForAllTysInvis` -> `tcSplitForAllInvisTVBinders`
  * `tcSplitForAllTysReq` -> `tcSplitForAllReqTVBinders`
  * `tcSplitForAllTy_maybe` -> `tcSplitForAllTyVarBinder_maybe`

Note that I left the following functions alone:

* Functions that split apart things besides `ForAllTy`s, such as `splitFunTys`
  or `splitPiTys`. Thankfully, there are far fewer of these functions than
  there are functions that split apart `ForAllTy`s, so there isn't much of a
  pressing need to apply the new naming convention elsewhere.
* Functions that split apart `ForAllCo`s in `Coercion`s, such as
  `GHC.Core.Coercion.splitForAllCo_maybe`. We could theoretically apply the new
  naming convention here, but then we'd have to figure out how to disambiguate
  `Type`-splitting functions from `Coercion`-splitting functions. Ultimately,
  the `Coercion`-splitting functions aren't used nearly as much as the
  `Type`-splitting functions, so I decided to leave the former alone.

This is purely refactoring and should cause no change in behavior.

- - - - -
645444af by Ryan Scott at 2020-11-15T03:36:21-05:00
Use tcSplitForAllInvisTyVars (not tcSplitForAllTyVars) in more places

The use of `tcSplitForAllTyVars` in `tcDataFamInstHeader` was the immediate
cause of #18939, and replacing it with a new `tcSplitForAllInvisTyVars`
function (which behaves like `tcSplitForAllTyVars` but only splits invisible
type variables) fixes the issue. However, this led me to realize that _most_
uses of `tcSplitForAllTyVars` in GHC really ought to be
`tcSplitForAllInvisTyVars` instead. While I was in town, I opted to replace
most uses of `tcSplitForAllTys` with `tcSplitForAllTysInvis` to reduce the
likelihood of such bugs in the future.

I say "most uses" above since there is one notable place where we _do_ want
to use `tcSplitForAllTyVars`: in `GHC.Tc.Validity.forAllTyErr`, which produces
the "`Illegal polymorphic type`" error message if you try to use a higher-rank
`forall` without having `RankNTypes` enabled. Here, we really do want to split
all `forall`s, not just invisible ones, or we run the risk of giving an
inaccurate error message in the newly added `T18939_Fail` test case.

I debated at some length whether I wanted to name the new function
`tcSplitForAllInvisTyVars` or `tcSplitForAllTyVarsInvisible`, but in the end,
I decided that I liked the former better. For consistency's sake, I opted to
rename the existing `splitPiTysInvisible` and `splitPiTysInvisibleN` functions
to `splitInvisPiTys` and `splitPiTysInvisN`, respectively, so that they use the
same naming convention. As a consequence, this ended up requiring a `haddock`
submodule bump.

Fixes #18939.

- - - - -
8887102f by Moritz Angermann at 2020-11-15T03:36:56-05:00
AArch64/arm64 adjustments

This addes the necessary logic to support aarch64 on elf, as well
as aarch64 on mach-o, which Apple calls arm64.

We change architecture name to AArch64, which is the official arm
naming scheme.

- - - - -
fc644b1a by Ben Gamari at 2020-11-15T03:37:31-05:00
ghc-bin: Build with eventlogging by default

We now have all sorts of great facilities using the
eventlog which were previously unavailable without
building a custom GHC. Fix this by linking with
`-eventlog` by default.
- - - - -
52114fa0 by Sylvain Henry at 2020-11-16T11:48:47+01:00
Add Addr# atomic primops (#17751)

This reuses the codegen used for ByteArray#'s atomic primops.

- - - - -
8150f654 by Sebastian Graf at 2020-11-18T23:38:40-05:00
PmCheck: Print types of uncovered patterns (#18932)

In order to avoid confusion as in #18932, we display the type of the
match variables in the non-exhaustiveness warning, e.g.

```
T18932.hs:14:1: warning: [-Wincomplete-patterns]
    Pattern match(es) are non-exhaustive
    In an equation for ‘g’:
        Patterns of type  ‘T a’, ‘T a’, ‘T a’ not matched:
            (MkT2 _) (MkT1 _) (MkT1 _)
            (MkT2 _) (MkT1 _) (MkT2 _)
            (MkT2 _) (MkT2 _) (MkT1 _)
            (MkT2 _) (MkT2 _) (MkT2 _)
            ...
   |
14 | g (MkT1 x) (MkT1 _) (MkT1 _) = x
   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
```

It also allows us to omit the type signature on wildcard matches which
we previously showed in only some situations, particularly
`-XEmptyCase`.

Fixes #18932.

- - - - -
165352a2 by Krzysztof Gogolewski at 2020-11-20T02:08:36-05:00
Export indexError from GHC.Ix (#18579)

- - - - -
b57845c3 by Kamil Dworakowski at 2020-11-20T02:09:16-05:00
Clarify interruptible FFI wrt masking state

- - - - -
321d1bd8 by Sebastian Graf at 2020-11-20T02:09:51-05:00
Fix strictness signatures of `prefetchValue*#` primops

Their strictness signatures said the primops are strict in their first
argument, which is wrong: Handing it a thunk will prefetch the pointer
to the thunk, but not evaluate it. Hence not strict.

The regression test `T8256` actually tests for laziness in the first
argument, so GHC apparently never exploited the strictness signature.

See also https://gitlab.haskell.org/ghc/ghc/-/issues/8256#note_310867,
where this came up.

- - - - -
0aec78b6 by Sebastian Graf at 2020-11-20T02:09:51-05:00
Demand: Interleave usage and strictness demands (#18903)

As outlined in #18903, interleaving usage and strictness demands not
only means a more compact demand representation, but also allows us to
express demands that we weren't easily able to express before.

Call demands are *relative* in the sense that a call demand `Cn(cd)`
on `g` says "`g` is called `n` times. *Whenever `g` is called*, the
result is used according to `cd`". Example from #18903:

```hs
h :: Int -> Int
h m =
  let g :: Int -> (Int,Int)
      g 1 = (m, 0)
      g n = (2 * n, 2 `div` n)
      {-# NOINLINE g #-}
  in case m of
    1 -> 0
    2 -> snd (g m)
    _ -> uncurry (+) (g m)
```

Without the interleaved representation, we would just get `L` for the
strictness demand on `g`. Now we are able to express that whenever
`g` is called, its second component is used strictly in denoting `g`
by `1C1(P(1P(U),SP(U)))`. This would allow Nested CPR to unbox the
division, for example.

Fixes #18903.
While fixing regressions, I also discovered and fixed #18957.

Metric Decrease:
    T13253-spj

- - - - -
3a55b3a2 by Sebastian Graf at 2020-11-20T02:09:51-05:00
Update user's guide entry on demand analysis and worker/wrapper

The demand signature notation has been undocumented for a long time.
The only source to understand it, apart from reading the `Outputable`
instance, has been an outdated wiki page.

Since the previous commits have reworked the demand lattice, I took
it as an opportunity to also write some documentation about notation.

- - - - -
fc963932 by Greg Steuck at 2020-11-20T02:10:31-05:00
Find hadrian location more reliably in cabal-install output

Fix #18944

- - - - -
9f40cf6c by Ben Gamari at 2020-11-20T02:11:07-05:00
rts/linker: Align bssSize to page size when mapping symbol extras

We place symbol_extras right after bss. We also need
to ensure that symbol_extras can be mprotect'd independently from the
rest of the image. To ensure this we round up the size of bss to a page
boundary, thus ensuring that symbol_extras is also page-aligned.

- - - - -
b739c319 by Ben Gamari at 2020-11-20T02:11:43-05:00
gitlab-ci: Add usage message to ci.sh

- - - - -
802e9180 by Ben Gamari at 2020-11-20T02:11:43-05:00
gitlab-ci: Add VERBOSE environment variable

And change the make build system's default behavior to V=0, greatly
reducing build log sizes.

- - - - -
2a8a979c by Ben Gamari at 2020-11-21T01:13:26-05:00
users-guide: A bit of clean-up in profiling flag documentation

- - - - -
56804e33 by Ben Gamari at 2020-11-21T01:13:26-05:00
testsuite: Refactor CountParserDeps

- - - - -
53ad67ea by Ben Gamari at 2020-11-21T01:13:26-05:00
Introduce -fprof-callers flag

This introducing a new compiler flag to provide a convenient way to
introduce profiler cost-centers on all occurrences of the named
identifier.

Closes #18566.

- - - - -
ecfd0278 by Sylvain Henry at 2020-11-21T01:14:09-05:00
Move Plugins into HscEnv (#17957)

Loaded plugins have nothing to do in DynFlags so this patch moves them
into HscEnv (session state).

"DynFlags plugins" become "Driver plugins" to still be able to register
static plugins.

Bump haddock submodule

- - - - -
72f2257c by Sylvain Henry at 2020-11-21T01:14:09-05:00
Don't initialize plugins in the Core2Core pipeline

Some plugins can be added via TH (cf addCorePlugin). Initialize them in
the driver instead of in the Core2Core pipeline.

- - - - -
ddbeeb3c by Ryan Scott at 2020-11-21T01:14:44-05:00
Add regression test for #10504

This issue was fixed at some point between GHC 8.0 and 8.2. Let's add a
regression test to ensure that it stays fixed.

Fixes #10504.

- - - - -
a4a6dc2a by Ben Gamari at 2020-11-21T01:15:21-05:00
dwarf: Apply info table offset consistently

Previously we failed to apply the info table offset to the aranges and
DIEs, meaning that we often failed to unwind in gdb. For some reason
this only seemed to manifest in the RTS's Cmm closures. Nevertheless,
now we can unwind completely up to `main`

- - - - -
69bfbc21 by Ben Gamari at 2020-11-21T01:15:56-05:00
hadrian: Disable stripping when debug information is enabled

- - - - -
7e93ae8b by Ben Gamari at 2020-11-21T13:13:29-05:00
rts: Post ticky entry counts to the eventlog

We currently only post the entry counters, not the other global
counters as in my experience the former are more useful. We use the heap
profiler's census period to decide when to dump.

Also spruces up the documentation surrounding ticky-ticky a bit.

- - - - -
bc9c3916 by Ben Gamari at 2020-11-22T06:28:10-05:00
Implement -ddump-c-backend argument

To dump output of the C backend.

- - - - -
901bc220 by Ben Gamari at 2020-11-22T12:39:02-05:00
Bump time submodule to 1.11.1

Also bumps directory, Cabal, hpc, time, and unix submodules.

Closes #18847.

- - - - -
92c0afbf by Ben Gamari at 2020-11-22T12:39:38-05:00
hadrian: Dump STG when ticky is enabled

This changes the "ticky" modifier to enable dumping of final STG as this
is generally needed to make sense of the ticky profiles.

- - - - -
d23fef68 by Ben Gamari at 2020-11-22T12:39:38-05:00
hadrian: Introduce notion of flavour transformers

This extends Hadrian's notion of "flavour", as described in #18942.

- - - - -
179d0bec by Ben Gamari at 2020-11-22T12:39:38-05:00
hadrian: Add a viaLlvmBackend modifier

Note that this also slightly changes the semantics of these flavours as
we only use LLVM for >= stage1 builds.

- - - - -
d4d95e51 by Ben Gamari at 2020-11-22T12:39:38-05:00
hadrian: Add profiled_ghc and no_dynamic_ghc modifiers

- - - - -
6815603f by Ben Gamari at 2020-11-22T12:39:38-05:00
hadrian: Drop redundant flavour definitions

Drop the profiled, LLVM, and ThreadSanitizer flavour definitions as
these can now be realized with flavour transformers.

- - - - -
f88f4339 by Ben Gamari at 2020-11-24T02:43:20-05:00
rts: Flush eventlog buffers from flushEventLog

As noted in #18043, flushTrace failed flush anything beyond the writer.
This means that a significant amount of data sitting in capability-local
event buffers may never get flushed, despite the users' pleads for us to
flush.

Fix this by making flushEventLog flush all of the event buffers before
flushing the writer.

Fixes #18043.

- - - - -
7c03cc50 by Ben Gamari at 2020-11-24T02:43:55-05:00
gitlab-ci: Run LLVM job on appropriately-labelled MRs

Namely, those marked with the ~"LLVM backend" label

- - - - -
9b95d815 by Ben Gamari at 2020-11-24T02:43:55-05:00
gitlab-ci: Run LLVM builds on Debian 10

The current Debian 9 image doesn't provide LLVM 7.

- - - - -
2ed3e6c0 by Ben Gamari at 2020-11-24T02:43:55-05:00
CmmToLlvm: Declare signature for memcmp

Otherwise `opt` fails with:

    error: use of undefined value '@memcmp$def'

- - - - -
9f281886 by Richard Eisenberg at 2020-11-25T15:26:41-05:00
Move core flattening algorithm to Core.Unify

This sets the stage for a later change, where this
algorithm will be needed from GHC.Core.InstEnv.

This commit also splits GHC.Core.Map into
GHC.Core.Map.Type and GHC.Core.Map.Expr,
in order to avoid module import cycles
with GHC.Core.

- - - - -
40309d89 by Richard Eisenberg at 2020-11-25T15:26:42-05:00
Bump the # of commits searched for perf baseline

The previous value of 75 meant that a feature branch with
more than 75 commits would get spurious CI passes.

This affects #18692, but does not fix that ticket, because
if a baseline cannot be found, we should fail, not succeed.

- - - - -
6012b3ce by Richard Eisenberg at 2020-11-25T15:32:39-05:00
Write commit message here.

- - - - -
3272ecf4 by Richard Eisenberg at 2020-11-25T15:32:45-05:00
Rename the flattener to become the rewriter.

Now that flattening doesn't produce flattening variables,
it's not really flattening anything: it's rewriting. This
change also means that the rewriter can no longer be confused
the core flattener (in GHC.Core.Unify), which is sometimes used
during type-checking.

- - - - -


30 changed files:

- .gitlab-ci.yml
- .gitlab/ci.sh
- aclocal.m4
- compiler/GHC.hs
- compiler/GHC/Builtin/primops.txt.pp
- compiler/GHC/Cmm/CLabel.hs
- compiler/GHC/Cmm/CommonBlockElim.hs
- compiler/GHC/Cmm/Dataflow/Label.hs
- compiler/GHC/Cmm/Info/Build.hs
- compiler/GHC/CmmToAsm.hs
- compiler/GHC/CmmToAsm/Config.hs
- compiler/GHC/CmmToAsm/Dwarf.hs
- compiler/GHC/CmmToAsm/Dwarf/Constants.hs
- compiler/GHC/CmmToAsm/Dwarf/Types.hs
- compiler/GHC/CmmToAsm/Monad.hs
- compiler/GHC/CmmToAsm/PIC.hs
- compiler/GHC/CmmToAsm/Reg/Graph/TrivColorable.hs
- compiler/GHC/CmmToAsm/Reg/Linear.hs
- compiler/GHC/CmmToAsm/Reg/Linear/FreeRegs.hs
- compiler/GHC/CmmToAsm/Reg/Target.hs
- compiler/GHC/CmmToAsm/X86/CodeGen.hs
- compiler/GHC/CmmToAsm/X86/Ppr.hs
- compiler/GHC/CmmToC.hs
- compiler/GHC/CmmToLlvm/Base.hs
- compiler/GHC/Core/Coercion.hs
- compiler/GHC/Core/Coercion.hs-boot
- compiler/GHC/Core/Coercion/Axiom.hs
- compiler/GHC/Core/Coercion/Opt.hs
- compiler/GHC/Core/FamInstEnv.hs
- compiler/GHC/Core/InstEnv.hs


The diff was not included because it is too large.


View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/57609069ed365114b38487c97706f3151064b1fa...3272ecf4142476803b1f62ec67cb9d999a59bdcd

-- 
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/57609069ed365114b38487c97706f3151064b1fa...3272ecf4142476803b1f62ec67cb9d999a59bdcd
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/20201125/ca971c24/attachment-0001.html>


More information about the ghc-commits mailing list