<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
<html lang="en">
<head>
<meta content="text/html; charset=US-ASCII" http-equiv="Content-Type">
<title>
GitLab
</title>
<style>img {
max-width: 100%; height: auto;
}
</style>
</head>
<body>
<div class="content">
<h3>
Ben Gamari pushed to branch wip/gc/nonmoving-nonconcurrent
at <a href="https://gitlab.haskell.org/ghc/ghc">Glasgow Haskell Compiler / GHC</a>
</h3>
<h4>
Commits:
</h4>
<ul>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/80a5d3e11acc58b69267ca940cfb5065cc149a29">80a5d3e1</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:30:19Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts/BlockAlloc: Allow aligned allocation requests
This implements support for block group allocations which are aligned to
an integral number of blocks.
This will be used by the nonmoving garbage collector, which uses the
block allocator to allocate the segments which back its heap. These
segments are a fixed number of blocks in size, with each segment being
aligned to the segment size boundary. This allows us to easily find the
segment metadata stored at the beginning of the segment.
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/530e682481943f99439650f432c95c06e53a9f18">530e6824</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:30:19Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">testsuite/testblockalloc: A bit of refactoring
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/7c0652b318c3f6e1dc58ba51db11aba3bf1f9b67">7c0652b3</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:30:19Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">testsuite/testblockalloc: Test aligned block group allocation
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/ac70542677d3a10b216f202727157314565657d1">ac705426</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:30:20Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts/BlockAlloc: Wibbles
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/8788c063a50146dd396b393e34aaf838156aebf1">8788c063</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:30:20Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts/BlockAlloc: Use allocLargeChunk in aligned block allocation
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/f3191083052fb4a03c593aa745e9d9610bb8a8ce">f3191083</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:32:01Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">Disallow allocating megablocks, update tests
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/67cbec4afac497f34c9e8f89af6ffe5ecaf28cc2">67cbec4a</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:32:01Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">Fix lint errors
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/59d32a9973a3170e2269acc4077fc5516797e044">59d32a99</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:32:01Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">Use allocLargeChunkOnNode to reduce splitting
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/ead67e400308d2ee48f20dd99e33ea03d906e3b2">ead67e40</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:32:01Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">Allow allocating megablocks in allocAlignedGroupOnNode
This is currently broken because freeGroup assumes integral number of
megablocks when freeing megablocks but we try to split the megablocks
returned by allocLargeChunkOnNode to smaller groups and free the rest.
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/5fab18475e34b14a4b308b7ff74b18dc195dea30">5fab1847</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:32:46Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">Merge branches 'wip/gc/misc-rts' and 'wip/gc/aligned-block-allocation' into wip/gc/preparation
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/7cb5b05c1b9aa5ae1c013fd906b6c6218d0a42e0">7cb5b05c</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:33:55Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts/StableName: Expose FOR_EACH_STABLE_NAME, freeSnEntry, SNT_size
These will be needed when we implement sweeping in the nonmoving
collector.
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/98c601f2b5050cb987a2bf6de6d1856e2f1b1ddd">98c601f2</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:33:55Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts: Disable aggregate-return warnings from gcc
This warning is a bit of a relic; there is little reason to avoid
aggregate return values in 2019.
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/225afe49453d0b47e125cd616cd6b80e9854b42c">225afe49</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:33:55Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts/Scav: Expose scavenging functions
To keep the non-moving collector nicely separated from the moving
collector its scavenging phase will live in another file,
`NonMovingScav.c`. However, it will need to use these functions so
let's expose them.
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/babf1fb84377389a1ecdd14602f67d0c768fcb71">babf1fb8</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:33:55Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts: Introduce flag to enable the nonmoving old generation
This flag will enable the use of a non-moving oldest generation.
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/f16399754201eadcd432c2362110195ac061e424">f1639975</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:33:55Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts: Introduce debug flag for non-moving GC
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/85a47315beb6a0b6cc56fc2b195f11b8d8de35db">85a47315</a></strong>
<div>
<span>by Ömer Sinan Ağacan</span>
<i>at 2019-06-19T00:33:55Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">rts: Non-concurrent mark and sweep
This implements the core heap structure and a serial mark/sweep
collector which can be used to manage the oldest-generation heap.
This is the first step towards a concurrent mark-and-sweep collector
aimed at low-latency applications.
The full design of the collector implemented here is described in detail
in a technical note
B. Gamari. "A Concurrent Garbage Collector For the Glasgow Haskell
Compiler" (2018)
The basic heap structure used in this design is heavily inspired by
K. Ueno & A. Ohori. "A fully concurrent garbage collector for
functional programs on multicore processors." /ACM SIGPLAN Notices/
Vol. 51. No. 9 (presented by ICFP 2016)
This design is intended to allow both marking and sweeping
concurrent to execution of a multi-core mutator. Unlike the Ueno design,
which requires no global synchronization pauses, the collector
introduced here requires a stop-the-world pause at the beginning and end
of the mark phase.
To avoid heap fragmentation, the allocator consists of a number of
fixed-size /sub-allocators/. Each of these sub-allocators allocators into
its own set of /segments/, themselves allocated from the block
allocator. Each segment is broken into a set of fixed-size allocation
blocks (which back allocations) in addition to a bitmap (used to track
the liveness of blocks) and some additional metadata (used also used
to track liveness).
This heap structure enables collection via mark-and-sweep, which can be
performed concurrently via a snapshot-at-the-beginning scheme (although
concurrent collection is not implemented in this patch).
The mark queue is a fairly straightforward chunked-array structure.
The representation is a bit more verbose than a typical mark queue to
accomodate a combination of two features:
* a mark FIFO, which improves the locality of marking, reducing one of
the major overheads seen in mark/sweep allocators (see [1] for
details)
* the selector optimization and indirection shortcutting, which
requires that we track where we found each reference to an object
in case we need to update the reference at a later point (e.g. when
we find that it is an indirection). See Note [Origin references in
the nonmoving collector] (in `NonMovingMark.h`) for details.
Beyond this the mark/sweep is fairly run-of-the-mill.
[1] R. Garner, S.M. Blackburn, D. Frampton. "Effective Prefetch for
Mark-Sweep Garbage Collection." ISMM 2007.
Co-Authored-By: Ben Gamari <ben@well-typed.com>
</pre>
</li>
<li>
<strong><a href="https://gitlab.haskell.org/ghc/ghc/commit/044531d5b1a0011e631a614dad52c8cfbb38be1b">044531d5</a></strong>
<div>
<span>by Ben Gamari</span>
<i>at 2019-06-19T00:33:55Z</i>
</div>
<pre class="commit-message" style="white-space: pre-wrap; margin: 0;">testsuite: Add nonmoving WAY
This simply runs the compile_and_run tests with `-xn`, enabling the
nonmoving oldest generation.
</pre>
</li>
</ul>
<h4>29 changed files:</h4>
<ul>
<li class="file-stats">
<a href="#6ae923e173536bf0ddb39552416a7e21767287db">
docs/users_guide/runtime_control.rst
</a>
</li>
<li class="file-stats">
<a href="#241971342cd0c394b24e67a781cf7e3d2f01aa72">
includes/rts/Flags.h
</a>
</li>
<li class="file-stats">
<a href="#0dcc7b3b05fd9aac5ea350885b06f0718bcf70c2">
includes/rts/storage/Block.h
</a>
</li>
<li class="file-stats">
<a href="#c184941110ec11e4f595256c2354bed617511571">
libraries/base/GHC/RTS/Flags.hsc
</a>
</li>
<li class="file-stats">
<a href="#33249795fee267712d5c3a7ecfa9f2edea260401">
rts/Capability.c
</a>
</li>
<li class="file-stats">
<a href="#62ec5569a8af1e443ae952b393d15b9dd1cea199">
rts/Capability.h
</a>
</li>
<li class="file-stats">
<a href="#640fc6f2f536ba83f1694f9a7ef3d43e2d0428d3">
rts/RtsFlags.c
</a>
</li>
<li class="file-stats">
<a href="#9ed11d0519762dae04656481b089dbb5b05acf98">
rts/RtsStartup.c
</a>
</li>
<li class="file-stats">
<a href="#7d6961cb073b9fe3df3a81f0ad5cf36eed293c42">
rts/Schedule.c
</a>
</li>
<li class="file-stats">
<a href="#52526efc348b52718c69d51f8df6e5425ff68d48">
rts/Schedule.h
</a>
</li>
<li class="file-stats">
<a href="#2f69bf0632a148614b9aceae17e7e627784cf567">
rts/StableName.c
</a>
</li>
<li class="file-stats">
<a href="#a6a45e2bced459a8c9ae14a62849f69ae436b0e9">
rts/StableName.h
</a>
</li>
<li class="file-stats">
<a href="#af7520a7c19cb7a894a90a944c97a4ed95f12b78">
rts/Trace.h
</a>
</li>
<li class="file-stats">
<a href="#8a78f03f80f0544ead476d53dd0f1db7cbd68d1b">
rts/Weak.c
</a>
</li>
<li class="file-stats">
<a href="#22d6a1257dd9727593b678661c672529539427ba">
rts/ghc.mk
</a>
</li>
<li class="file-stats">
<a href="#a9accda65187a56e0d42dd7fb7a717af6b41d918">
rts/sm/BlockAlloc.c
</a>
</li>
<li class="file-stats">
<a href="#37a93df244a7789fddf557443b6398d06763e9cc">
rts/sm/Evac.c
</a>
</li>
<li class="file-stats">
<a href="#6477e10756faf038741e63d1ad499a1df809fe10">
rts/sm/GC.c
</a>
</li>
<li class="file-stats">
<a href="#06e70a7de800b2de9d75c2d5b5b554929f6b9bcb">
rts/sm/GC.h
</a>
</li>
<li class="file-stats">
<a href="#a481abfdf0106c53321542436ad78f6df542b1b8">
rts/sm/GCAux.c
</a>
</li>
<li class="file-stats">
<a href="#26a3faea0f0e66eab204fe0639219eea4729a9d3">
rts/sm/GCThread.h
</a>
</li>
<li class="file-stats">
<a href="#b4a38dad10ba126e05a0b6adc99aa0c2eb742875">
<span class="new-file">
+
rts/sm/NonMoving.c
</span>
</a>
</li>
<li class="file-stats">
<a href="#371d3c3c9fad484c1163aa733d70507ce33084c9">
<span class="new-file">
+
rts/sm/NonMoving.h
</span>
</a>
</li>
<li class="file-stats">
<a href="#4f5ba3ce18ccc17ad367193f379aa61f4c9180a0">
<span class="new-file">
+
rts/sm/NonMovingMark.c
</span>
</a>
</li>
<li class="file-stats">
<a href="#bcae703553e66a4a73c4aa4a930513cc4aa923cf">
<span class="new-file">
+
rts/sm/NonMovingMark.h
</span>
</a>
</li>
<li class="file-stats">
<a href="#7893c76945f74c60b62878fa12bb5de75b8e1340">
<span class="new-file">
+
rts/sm/NonMovingScav.c
</span>
</a>
</li>
<li class="file-stats">
<a href="#9cbb27e8dac63b5209040e177a71a14a21cfeadc">
<span class="new-file">
+
rts/sm/NonMovingScav.h
</span>
</a>
</li>
<li class="file-stats">
<a href="#78077cd665e91d3877a41e287033df6d2884b2ce">
<span class="new-file">
+
rts/sm/NonMovingSweep.c
</span>
</a>
</li>
<li class="file-stats">
<a href="#e0df3d3aff01c04aefc7478eff33fb15fd9f98a6">
<span class="new-file">
+
rts/sm/NonMovingSweep.h
</span>
</a>
</li>
</ul>
<h5>The diff was not included because it is too large.</h5>
</div>
<div class="footer" style="margin-top: 10px;">
<p style="font-size: small; color: #777;">
—
<br>
<a href="https://gitlab.haskell.org/ghc/ghc/compare/01a98ff3ef62c10980b5845640b1529fd387af7d...044531d5b1a0011e631a614dad52c8cfbb38be1b">View it on GitLab</a>.
<br>
You're receiving this email because of your account on gitlab.haskell.org.
If you'd like to receive fewer emails, you can
adjust your notification settings.
</p>
</div>
</body>
</html>