- Mar 10, 2021
-
-
Related to #19381 #19359 #14702 After a spike in memory usage we have been conservative about returning allocated blocks to the OS in case we are still allocating a lot and would end up just reallocating them. The result of this was that up to 4 * live_bytes of blocks would be retained once they were allocated even if memory usage ended up a lot lower. For a heap of size ~1.5G, this would result in OS memory reporting 6G which is both misleading and worrying for users. In long-lived server applications this results in consistent high memory usage when the live data size is much more reasonable (for example ghcide) Therefore we have a new (2021) strategy which starts by retaining up to 4 * live_bytes of blocks before gradually returning uneeded memory back to the OS on subsequent major GCs which are NOT caused by a heap overflow. Each major GC which is NOT caused by heap overflow increases the consec_idle_gcs counter and the amount of memory which is retained is inversely proportional to this number. By default the excess memory retained is oldGenFactor (controlled by -F) / 2 ^ (consec_idle_gcs * returnDecayFactor) On a major GC caused by a heap overflow, the `consec_idle_gcs` variable is reset to 0 (as we could continue to allocate more, so retaining all the memory might make sense). Therefore setting bigger values for `-Fd` makes the rate at which memory is returned slower. Smaller values make it get returned faster. Setting `-Fd0` disables the memory return completely, which is the behaviour of older GHC versions. The default is `-Fd4` which results in the following scaling: > mapM print [(x, 1/ (2**(x / 4))) | x <- [1 :: Double ..20]] (1.0,0.8408964152537146) (2.0,0.7071067811865475) (3.0,0.5946035575013605) (4.0,0.5) (5.0,0.4204482076268573) (6.0,0.35355339059327373) (7.0,0.29730177875068026) (8.0,0.25) (9.0,0.21022410381342865) (10.0,0.17677669529663687) (11.0,0.14865088937534013) (12.0,0.125) (13.0,0.10511205190671433) (14.0,8.838834764831843e-2) (15.0,7.432544468767006e-2) (16.0,6.25e-2) (17.0,5.255602595335716e-2) (18.0,4.4194173824159216e-2) (19.0,3.716272234383503e-2) (20.0,3.125e-2) So after 13 consecutive GCs only 0.1 of the maximum memory used will be retained. Further to this decay factor, the amount of memory we attempt to retain is also influenced by the GC strategy for the oldest generation. If we are using a copying strategy then we will need at least 2 * live_bytes for copying to take place, so we always keep that much. If using compacting or nonmoving then we need a lower number, so we just retain at least `1.2 * live_bytes` for some protection. In future we might want to make this behaviour more aggressive, some relevant literature is > Ulan Degenbaev, Jochen Eisinger, Manfred Ernst, Ross McIlroy, and Hannes Payer. 2016. Idle time garbage collection scheduling. SIGPLAN Not. 51, 6 (June 2016), 570–583. DOI:https://doi.org/10.1145/2980983.2908106 which describes the "memory reducer" in the V8 javascript engine which on an idle collection immediately returns as much memory as possible.
-
- Mar 09, 2021
-
-
This fixes a regression that led to loss of location information in error messages about the use of tuple sections in patterns.
-
This fixes two classes of warnings that appear when bootstrapping with GHC 9.0: * `ghc-boot.cabal` was using `cabal-version: >=1.22`, which `cabal-install-3.4` now warns about, instead recommending the use of `cabal-version: 1.22`. * Several pattern matches were producing `Pattern match(es) are non-exhaustive` because of incorrect CPP. The pattern-match coverage checker _did_ become smarter in GHC 9.1, however, so I ended up needing to keep the CPP, adjusting them to use `#if __GLASGOW_HASKELL__ < 901` instead.
-
It's surprisingly tricky to deal with 'main' (#19397). This patch does quite bit of refactoring do to it right. Well, more-right anyway! The moving parts are documented in GHC.Tc.Module Note [Dealing with main] Some other oddments: * Rename tcRnExports to rnExports; no typechecking here! * rnExports now uses checkNoErrs rather than failIfErrsM; the former fails only if rnExports itself finds errors * Small improvements to tcTyThingCategory, which ultimately weren't important to the patch, but I've retained as a minor improvement.
-
The Ord instance was non-deterministic, but it's easy assume that it is deterministic. In fact, haddock-api used to do exactly that before haddock/7e8c7c3491f3e769368b8e6c767c62a33e996c80
-
Hadrian should behave well and not delete files created by configure with the clean command. With this patch hadrian now deletes the fs/mingw tarballs only with distclean. This fixes #19320. The main impact being that validate won't have to redownload the tarballs when re-run.
-
This resolves #19457 by making a note of breaking changes (introduced in GHC 9.2) to the way that GHC typechecks operator sections where the operator has nested `forall`s or contexts in its type signature.
-
Now that GHC 9.0.1 is released, it is time to drop support for bootstrapping with GHC 8.8, as we only support building with the previous two major GHC releases. As an added bonus, this allows us to remove several bits of CPP that are either always true or no longer reachable.
-
- Mar 08, 2021
-
-
If startEventlog is called after the program has already started running then quite a few useful events are missing from the eventlog because they are only posted when the program starts. This patch adds a mechanism to declare that an event should be reposted everytime the startEventlog function is called. Now in EventLog.c there is a global list of functions called `eventlog_header_funcs` which stores a list of functions which should be called everytime the eventlog starts. When calling `postInitEvent`, the event will not only be immediately posted to the eventlog but also added to the global list. When startEventLog is called, the list is traversed and the events reposted.
-
The way in which allocatePinned took blocks out of the nursery was leading to horrible fragmentation in some workloads. The strategy now is that a separate free block list is reserved for each capability and blocks are taken from there. When it's empty the global SM lock is taken and a fresh block of size PINNED_EMPTY_SIZE is allocated. Fixes #19481
-
-
-
-
The BLOCKS_SIZE event reports the size of the currently allocated blocks in bytes. It is like the HEAP_SIZE event, but reports about the blocks rather than megablocks. You can work out the current heap fragmentation by looking at the difference between HEAP_SIZE and BLOCKS_SIZE. Fixes #19357
-
See #19357 The event reports the * Current number of megablocks allocated * The number that the RTS thinks it needs * The number is managed to return to the OS When current > need then the difference is returned to the OS, the successful number of returned mblocks is reported by 'returned'. In a fragmented heap current > need but returned < current - need.
-
Fixes #19409
-
- Mar 07, 2021
-
-
Ben Gamari authored
This implements the BoxedRep proposal, refactoring the `RuntimeRep` hierarchy from: ```haskell data RuntimeRep = LiftedPtrRep | UnliftedPtrRep | ... ``` to ```haskell data RuntimeRep = BoxedRep Levity | ... data Levity = Lifted | Unlifted ``` Updates binary, haddock submodules. Closes #17526. Metric Increase: T12545
-
-
- Mar 06, 2021
-
-
This mirrors the make build system and ensures that we don't end up with references to the build directory in the final executable. Fixes #19485.
-
Using .gitatttributes, we don't require users to set git's core.autocrlf setting to false on Windows to be able to checkout a working tree.
-
- Mar 05, 2021
-
-
This enables a registerised build for the riscv64 architecture.
-
This fixes #19484. In detail we: * Bump the index-state of hackage. * Require alex-3.2.6, as alex-3.2.5 doesn't build with 9.0. * Allow Cabal-3.4 as 3.2 doesn't build with ghc 9.0. * Allow a newer QuickCheck version that accepts the new base version. * Some code changes to account for Cabal changes.
-
-
This addresses points (1a) and (1b) of #19165. - Move mkFailExpr to HsToCore/Utils, as it can be shared - Desugar incomplete patterns and holes to an empty case, as in Note [Incompleteness and linearity] - Enable linear linting of desugarer output - Mark MultConstructor as broken. It fails Lint, but I'd like to fix this separately. Metric Decrease: T6048 -
Previously this test did nothing to prevent GHC from reading .ghci due to the `-e` arguments. Consequently it could fail due to multiple reloadings of DynFlags while evaluating .ghci.
-
Commit 2a942285 dramatically simplified the implementation and improved the performance of COMPLETE sets while making them applicable in more scenarios at the same time. But it turned out that there was a change in semantics that (to me unexpectedly) broke users' expectations (see #14422): They relied on the "type signature" of a COMPLETE pragma to restrict the scrutinee types of a pattern match for which they are applicable. This patch brings back that filtering, so the semantics is the same as it was in GHC 9.0. See the updated Note [Implementation of COMPLETE pragmas]. There are a few testsuite output changes (`completesig13`, `T14422`) which assert this change. Co-authored-by:
Sebastian Graf <sebastian.graf@kit.edu>
-
Fixes #19455.
-
markLiveObject is called by GC worker threads and therefore must be thread-safe. This was a rather egregious oversight which the testsuite missed. (cherry picked from commit fe28a062)
-
-
This was fixed as a result of #19181.
-
This produces much more detailed ticky profiles which include names of constructors. Related !3340 !2098 Fixes #19403
-
Part of #17804.
-
- Mar 03, 2021
-
-
Matthew Pickering authored
During testing it was observed that quite a few info tables were not being given locations (due to not being assigned source locations, because they were not enclosed by a source note). We can at least give the module name and type for such closures even if no more accurate source information. Especially for constructors this helps find them in the STG dumps.
-
Matthew Pickering authored
-
Matthew Pickering authored
-
Matthew Pickering authored
-
Matthew Pickering authored
-
Matthew Pickering authored
The `whereFrom` function provides a Haskell interface for using the information created by `-finfo-table-map`. Given a Haskell value, the info table address will be passed to the `lookupIPE` function in order to attempt to find the source location information for that particular closure. At the moment it's not possible to distinguish the absense of the map and a failed lookup.
-
Matthew Pickering authored
The `-fdistinct-constructor-tables` flag will generate a fresh info table for the usage of any data constructor. This is useful for debugging as now by inspecting the info table, you can determine which usage of a constructor caused that allocation rather than the old situation where the info table always mapped to the definition site of the data constructor which is useless. In conjunction with `-hi` and `-finfo-table-map` this gives a more fine grained understanding of where constructor allocations arise from in a program.
-
Matthew Pickering authored
This new flag embeds a lookup table from the address of an info table to information about that info table. The main interface for consulting the map is the `lookupIPE` C function > InfoProvEnt * lookupIPE(StgInfoTable *info) The `InfoProvEnt` has the following structure: > typedef struct InfoProv_{ > char * table_name; > char * closure_desc; > char * ty_desc; > char * label; > char * module; > char * srcloc; > } InfoProv; > > typedef struct InfoProvEnt_ { > StgInfoTable * info; > InfoProv prov; > struct InfoProvEnt_ *link; > } InfoProvEnt; The source positions are approximated in a similar way to the source positions for DWARF debugging information. They are only approximate but in our experience provide a good enough hint about where the problem might be. It is therefore recommended to use this flag in conjunction with `-g<n>` for more accurate locations. The lookup table is also emitted into the eventlog when it is available as it is intended to be used with the `-hi` profiling mode. Using this flag will significantly increase the size of the resulting object file but only by a factor of 2-3x in our experience.
-
