simonmar (Simon Marlow)
User

Projects

User Details

User Since
Jun 8 2014, 1:59 AM (206 w, 6 d)

Recent Activity

Yesterday

simonmar committed rGHCc618732ebb23: isDllName: use Opt_ExternalDynamicRefs, not WayDyn (authored by simonmar).
isDllName: use Opt_ExternalDynamicRefs, not WayDyn
Fri, May 25, 4:09 AM
simonmar committed rGHC5b6ef59f9959: Add -fghci-leak-check to check for space leaks (authored by simonmar).
Add -fghci-leak-check to check for space leaks
Fri, May 25, 4:09 AM
simonmar added a comment to D4731: UNREG: put SRT into read-only data section.

But SRTs are writable - they have a static link field at the end which is modified by the GC. So I'm confused, how can this work? We should probably fix the declaration to remove the const instead.

Fri, May 25, 3:48 AM

Thu, May 24

simonmar requested changes to D4722: Mark system and internal symbols as private symbols in asm.

Had a chat with @last_g about this in person today. The overall goal is good and this will improve GHC's interaction with perf. Let's figure out the concrete steps to get this in:

Thu, May 24, 3:56 PM

Wed, May 23

simonmar committed rGHCd424d4a46a72: Fix a bug in SRT generation (authored by simonmar).
Fix a bug in SRT generation
Wed, May 23, 12:17 PM
simonmar closed D4721: Fix a bug in SRT generation.
Wed, May 23, 12:17 PM
simonmar added a comment to D4721: Fix a bug in SRT generation.

@Phyx thanks so much for this! I'll go ahead and push.

Wed, May 23, 11:48 AM
simonmar added a comment to D4722: Mark system and internal symbols as private symbols in asm.

@last_g I'm probably being dumb but I don't follow why D4713 would lead to more unknown addresses in the perf output, could you elaborate?

Wed, May 23, 4:38 AM
simonmar added a comment to D4715: Disable the SRT offset optimisation on MachO platforms.

Build error though:

Wed, May 23, 4:38 AM
simonmar accepted D4715: Disable the SRT offset optimisation on MachO platforms.

Thanks for doing this!

Wed, May 23, 4:38 AM
simonmar accepted D4719: base/TimerManager: Clamp timer expiration time to maxBound.

successfirsttry

Wed, May 23, 4:12 AM
simonmar created Image Macro "successfirsttry".
Wed, May 23, 4:12 AM
simonmar updated subscribers of D4721: Fix a bug in SRT generation.

Well, I've been waiting 17h for a build and it looks like it's still queued. If anyone is able to validate on Windows that would be great. cc @awson @Phyx

Wed, May 23, 3:43 AM

Tue, May 22

simonmar updated subscribers of D4722: Mark system and internal symbols as private symbols in asm.

I'm not sure yet whether this is the right way to go. I'd like to hear input from others (@bgamari, @niteria, @angerman, @olsner?)

Tue, May 22, 11:04 AM
simonmar added a comment to D4719: base/TimerManager: Clamp timer expiration time to maxBound.

Is it possible to make a test for this?

Tue, May 22, 10:32 AM
simonmar added inline comments to D4715: Disable the SRT offset optimisation on MachO platforms.
Tue, May 22, 10:16 AM
simonmar created D4721: Fix a bug in SRT generation.
Tue, May 22, 10:04 AM
simonmar added inline comments to D4715: Disable the SRT offset optimisation on MachO platforms.
Tue, May 22, 3:56 AM
simonmar added inline comments to D4719: base/TimerManager: Clamp timer expiration time to maxBound.
Tue, May 22, 3:39 AM

Mon, May 21

simonmar added inline comments to D4700: [RFC] base: Handle long timeouts.
Mon, May 21, 3:46 PM
simonmar added a comment to D4713: Mark code related symbols as @function not @object.

We ought to check what LLVM does and that whatever we do here doesn't break the LLVM backend.

Mon, May 21, 3:03 AM
simonmar added inline comments to D4700: [RFC] base: Handle long timeouts.
Mon, May 21, 2:49 AM
simonmar added inline comments to D4715: Disable the SRT offset optimisation on MachO platforms.
Mon, May 21, 2:42 AM

Sun, May 20

simonmar requested changes to D4647: RFC: Continuation arguments.

Partial review - this looks really cool! Definitely needs polishing + comments though.

Sun, May 20, 5:13 PM
simonmar added a comment to D4644: Do not resurrect all dead threads at once.
In D4644#128693, @osa1 wrote:

The problem with this is that it's non-deterministic - it makes a difference which thread you resurrect first, and we have no way to tell which order to resurrect threads in.

Right, I didn't think that's a problem because scheduling and garbage collection are already non-deterministic, which makes BlockedIndefinitelyOnMVar non-deterministic as well. If this is not OK (why?) then perhaps we should close Trac #10241 and Trac #10793 (and any tickets in the future) as WONTFIX. Or do you think this can be done in a deterministic way?

Sun, May 20, 5:13 PM

Thu, May 17

simonmar added a comment to D4475: RFC: Add Int8# and Word8#.

(warning: this is a really old partial review that I only just discovered I forgot to submit)

Thu, May 17, 10:11 AM
simonmar committed rGHCf27e4f624fe1: Fix GHCi space leaks (#15111) (authored by simonmar).
Fix GHCi space leaks (#15111)
Thu, May 17, 9:45 AM
simonmar closed D4659: Fix GHCi space leaks (#15111).
Thu, May 17, 9:44 AM
simonmar updated the diff for D4709: Add -fghci-leak-check to check for space leaks.

retrying; database failure the first time

Thu, May 17, 9:44 AM
simonmar created D4709: Add -fghci-leak-check to check for space leaks.
Thu, May 17, 9:44 AM
simonmar added a comment to D4700: [RFC] base: Handle long timeouts.

This seems like quite an elaborate fix for an overflow problem. Why did the overflow occur in the first place, shouldn't the time remaining be less than maxBound?

Thu, May 17, 9:44 AM
simonmar accepted D4702: rts: Fix compaction of SmallMutArrPtrs.

Looks reasonable

Thu, May 17, 9:09 AM
simonmar added a comment to D4707: Fix retainer profiling after SRT overhaul.

Passes the profiling tests here

Thu, May 17, 8:53 AM
simonmar added a comment to rGHCeb8e692cab79: An overhaul of the SRT representation.

D4707, testing now.

Thu, May 17, 8:01 AM
simonmar accepted D4704: rts: Don't madvise if mmap failed.

nicecatch

Thu, May 17, 8:01 AM
simonmar created D4707: Fix retainer profiling after SRT overhaul.
Thu, May 17, 7:37 AM
simonmar committed rGHC5f15d53a98ad: Add /* fallthrough */ to fix -Wimplicit-fallthrough warning (authored by simonmar).
Add /* fallthrough */ to fix -Wimplicit-fallthrough warning
Thu, May 17, 5:24 AM
simonmar committed rGHCf2d27c1ad693: Comments and refactoring only (authored by simonmar).
Comments and refactoring only
Thu, May 17, 5:24 AM
simonmar committed rGHCb701e4754d1d: Update Cabal submodule (authored by simonmar).
Update Cabal submodule
Thu, May 17, 5:24 AM
simonmar added a comment to D4637: Merge FUN_STATIC closure with its SRT.

Thanks, will take a look.

Thu, May 17, 4:56 AM
simonmar added a comment to D4637: Merge FUN_STATIC closure with its SRT.

on gcc (Ubuntu 7.3.0-16ubuntu3) 7.3.0

Thu, May 17, 3:02 AM
simonmar added a comment to rGHC3310f7f14c0b: InfoTables: Fix #if uses introduced by D4634.

Thanks! Sorry for the breakage. (could we make Harbourmaster any faster? The iteration time of >24 hours means it's pretty painful to get to a clean validate on all platforms)

Thu, May 17, 2:54 AM

Wed, May 16

simonmar added a comment to D4632: An overhaul of the SRT representation.

Yes, though https://phabricator.haskell.org/rGHC2b0918c9834be1873728176e4944bec26271234a is responsible for the info table compression.

Wed, May 16, 1:03 PM
simonmar added a comment to D4637: Merge FUN_STATIC closure with its SRT.

I pushed this but forgot that there were still comments to address... sorry about that, will address them in a follow up.

Wed, May 16, 8:21 AM
simonmar committed rGHC838b69032566: Merge FUN_STATIC closure with its SRT (authored by simonmar).
Merge FUN_STATIC closure with its SRT
Wed, May 16, 7:37 AM
simonmar closed D4637: Merge FUN_STATIC closure with its SRT.
Wed, May 16, 7:37 AM
simonmar committed rGHC2b0918c9834b: Save a word in the info table on x86_64 (authored by simonmar).
Save a word in the info table on x86_64
Wed, May 16, 7:37 AM
simonmar closed D4634: Save a word in the info table on x86_64.
Wed, May 16, 7:36 AM
simonmar committed rGHCfbd28e2c6b5f: Allow CmmLabelDiffOff with different widths (authored by simonmar).
Allow CmmLabelDiffOff with different widths
Wed, May 16, 7:36 AM
simonmar closed D4633: Allow CmmLabelDiffOff with different widths.
Wed, May 16, 7:36 AM
simonmar committed rGHCeb8e692cab79: An overhaul of the SRT representation (authored by simonmar).
An overhaul of the SRT representation
Wed, May 16, 7:36 AM
simonmar closed D4632: An overhaul of the SRT representation.
Wed, May 16, 7:36 AM
simonmar updated the diff for D4659: Fix GHCi space leaks (#15111).

rebase

Wed, May 16, 3:39 AM
simonmar committed rGHCa18e7dfabd23: Force findPtr to be included in the binary (authored by simonmar).
Force findPtr to be included in the binary
Wed, May 16, 3:25 AM
simonmar closed D4683: Force findPtr to be included in the binary.
Wed, May 16, 3:25 AM

Tue, May 15

simonmar updated the diff for D4634: Save a word in the info table on x86_64.

fix comment

Tue, May 15, 11:19 AM
simonmar accepted D4693: Make finalizers more reliable..

Thanks for this!

Tue, May 15, 7:59 AM

Mon, May 14

simonmar updated the diff for D4637: Merge FUN_STATIC closure with its SRT.

update

Mon, May 14, 3:08 AM
simonmar updated the diff for D4634: Save a word in the info table on x86_64.

update

Mon, May 14, 3:07 AM
simonmar updated the diff for D4633: Allow CmmLabelDiffOff with different widths.

update

Mon, May 14, 3:07 AM
simonmar updated the diff for D4632: An overhaul of the SRT representation.

update

Mon, May 14, 3:06 AM
simonmar updated the diff for D4659: Fix GHCi space leaks (#15111).

comment fix

Mon, May 14, 2:46 AM

Sun, May 13

simonmar updated the diff for D4659: Fix GHCi space leaks (#15111).

update haddock perf stats

Sun, May 13, 3:20 PM
simonmar accepted D4636: Fix another batch of `./validate --slow` failures.

The reason that validate --slow might not be exactly the same as what we see in CI is that validate --slow enables DEBUG for stage 2, whereas we don't do that in CI because we're building binaries for distribution and DEBUG would make them slower. The warning messages you're seeing in the Cabal tests are only emitted when the compiler is built with DEBUG. (apologies if this was pointed out elsewhere, I haven't read all the surrounding tickets here)

Sun, May 13, 3:06 AM
simonmar updated the diff for D4659: Fix GHCi space leaks (#15111).

add a hack to fix backpack

Sun, May 13, 2:49 AM

Thu, May 10

simonmar accepted D4642: ghc-pkg: Configure handle encodings.

LGTM

Thu, May 10, 12:51 PM
simonmar accepted D4666: Fix unwinding of C -> Haskell FFI calls with -threaded.

LGTM

Thu, May 10, 12:51 PM
simonmar committed rGHC87e169a30dae: Revert "Add -fghci-leak-check to check for space leaks" (authored by simonmar).
Revert "Add -fghci-leak-check to check for space leaks"
Thu, May 10, 12:09 PM
simonmar added a reverting change for rGHC5fe6aaa3756c: Add -fghci-leak-check to check for space leaks: rGHC87e169a30dae: Revert "Add -fghci-leak-check to check for space leaks".
Thu, May 10, 12:09 PM
simonmar added a comment to D4658: Add -fghci-leak-check to check for space leaks.

Sorry! Of course it broke things because I haven't committed the fix for these yet. D'oh. I'll back it out.

Thu, May 10, 11:38 AM
simonmar updated subscribers of D4659: Fix GHCi space leaks (#15111).
Thu, May 10, 7:41 AM
simonmar accepted D4680: Fix #15038.

LGTM, thanks for straightening this out. Should we make a ticket for the issue with the regression test?

Thu, May 10, 2:51 AM
simonmar added a comment to D4683: Force findPtr to be included in the binary.

The backpack failures are from D4659 (I stacked this diff on top of that one in my tree)

Thu, May 10, 2:45 AM
simonmar added a comment to D4679: RTS: Auto-size the allocation area depending on CPU cache size.

The only way to decide whether this is a good idea is to get data (and plenty of it). nofib on multiple different hardware types, the nofib/parallel suite too.

Thu, May 10, 2:43 AM

Wed, May 9

simonmar updated the diff for D4683: Force findPtr to be included in the binary.

also update rts.cabal.in

Wed, May 9, 3:04 PM
simonmar abandoned D4657: Turn off -ffunction-sections for Printer.c.

See D4683

Wed, May 9, 3:02 PM
simonmar created D4683: Force findPtr to be included in the binary.
Wed, May 9, 3:01 PM
simonmar added a comment to D4659: Fix GHCi space leaks (#15111).

I wanted to check the impact on compiler perf, so full nofib results: P180 (ignore the runtime results, I think my laptop was busy doing other things during one of the runs)

Wed, May 9, 8:20 AM
simonmar committed rGHC5fe6aaa3756c: Add -fghci-leak-check to check for space leaks (authored by simonmar).
Add -fghci-leak-check to check for space leaks
Wed, May 9, 7:47 AM
simonmar updated the diff for D4659: Fix GHCi space leaks (#15111).
  • rebase
  • accept new stats result for T13701
Wed, May 9, 7:46 AM
simonmar closed D4658: Add -fghci-leak-check to check for space leaks.
Wed, May 9, 7:46 AM

Tue, May 8

simonmar requested changes to D4666: Fix unwinding of C -> Haskell FFI calls with -threaded.

Just a few nits

Tue, May 8, 4:25 AM
simonmar added a comment to D4657: Turn off -ffunction-sections for Printer.c.

I suppose I could add these to the list of -u symbols we already have in package.conf.in.

Tue, May 8, 3:27 AM
simonmar added a comment to D4617: [RFC] rts/posix: Load libnuma dynamically.

I have mixed feelings about this.

Tue, May 8, 3:25 AM
simonmar accepted D4675: rts: Compile with gcc -Og.

Ooh, I learned something today.

Tue, May 8, 3:17 AM
simonmar added a comment to D4652: Make error Ids non-CAFFY (fixes #15038).

(Personally I think entering an unused sum field should be a panic rather than an exception)

Tue, May 8, 3:10 AM

Sat, May 5

simonmar requested review of D4657: Turn off -ffunction-sections for Printer.c.
Sat, May 5, 12:35 PM

Fri, May 4

simonmar requested changes to D4652: Make error Ids non-CAFFY (fixes #15038).

Hmm, I'm not keen on bumping binary sizes by nearly 1%, given that this change is unforced.

Fri, May 4, 2:51 PM
simonmar updated the diff for D4637: Merge FUN_STATIC closure with its SRT.

address comments

Fri, May 4, 7:55 AM
simonmar updated the diff for D4634: Save a word in the info table on x86_64.

address comments

Fri, May 4, 7:55 AM
simonmar updated the diff for D4633: Allow CmmLabelDiffOff with different widths.

address comments

Fri, May 4, 7:55 AM
simonmar updated the diff for D4633: Allow CmmLabelDiffOff with different widths.

address comments

Fri, May 4, 7:39 AM
simonmar updated the diff for D4632: An overhaul of the SRT representation.

address comments

Fri, May 4, 7:39 AM

Thu, May 3

simonmar accepted D4653: rts: Allow profiling by closure type in prof way.

Sure, let's do it.

Thu, May 3, 3:13 PM
simonmar added inline comments to D4617: [RFC] rts/posix: Load libnuma dynamically.
Thu, May 3, 1:56 PM
simonmar added inline comments to D4632: An overhaul of the SRT representation.
Thu, May 3, 1:38 PM
simonmar updated the diff for D4659: Fix GHCi space leaks (#15111).

rebase

Thu, May 3, 4:41 AM
simonmar updated the diff for D4658: Add -fghci-leak-check to check for space leaks.

update test output

Thu, May 3, 4:41 AM
simonmar added a comment to D4606: Correctly add unwinding info in manifestSp and makeFixupBlocks.

LGTM, I'll let @bgamari sign off on it.

Thu, May 3, 4:35 AM
simonmar added inline comments to D4659: Fix GHCi space leaks (#15111).
Thu, May 3, 4:35 AM