Page MenuHomePhabricator

scott-0 (scott douglass)
User

Projects

User does not belong to any projects.

User Details

User Since
Oct 25 2013, 7:42 AM (320 w, 2 d)

Recent Activity

Nov 19 2015

scott-0 added a comment to D14653: [libcxx] Introduce the mechanism for fixing -fno-exceptions test failures..

It would be nice if the no-exceptions library called an intermediate helper, for example __libcxx_noexceptions_abort(), instead of calling abort() directly. Then the user and the tests could provide a replacement for __libcxx_noexceptions_abort() instead of abort() which has other uses.

Nov 19 2015, 7:55 AM

Oct 6 2015

scott-0 closed D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..
Oct 6 2015, 1:51 AM
scott-0 accepted D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..

Committed as r249322

Oct 6 2015, 1:49 AM

Oct 5 2015

scott-0 committed rL249322: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..
[ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM.
Oct 5 2015, 7:52 AM
scott-0 added a comment to D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..

Thanks for the review; I will do those final edits and commit.

Oct 5 2015, 7:45 AM
scott-0 added a comment to D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..

I think it'd be better to do the sorting when the instruction is expanded, it seems like a useful property to have if anyone wants to analyse an LDM/STM.

Oct 5 2015, 7:35 AM
scott-0 updated the diff for D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..

Addressed comments.

Oct 5 2015, 6:47 AM

Oct 2 2015

scott-0 added a comment to D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..

Thanks for the review. I'll put up a new version after the weekend.

Oct 2 2015, 12:46 AM

Oct 1 2015

scott-0 updated the diff for D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..

Thanks for the review.

Oct 1 2015, 9:36 AM
scott-0 added a comment to D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..

New version on it's way.

Oct 1 2015, 9:35 AM
scott-0 added a comment to D13240: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer.

Committed including changing from -mtriple ... to target triple = ...

Oct 1 2015, 5:08 AM
scott-0 committed rL249002: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer.
[ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer
Oct 1 2015, 4:58 AM
scott-0 closed D13240: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer by committing rL249002: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer.
Oct 1 2015, 4:58 AM

Sep 30 2015

scott-0 updated D13240: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer.
Sep 30 2015, 6:52 AM
scott-0 updated the diff for D13240: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer.

Yes, '-stop-after' makes it straight-forward to create a test case. Added. Thanks for the pointer.

Sep 30 2015, 6:51 AM

Sep 29 2015

scott-0 updated subscribers of D13240: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer.
Sep 29 2015, 3:49 AM
scott-0 retitled D13240: [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer from to [ARM] More care with Thumb1 writeback in ARMLoadStoreOptimizer.
Sep 29 2015, 3:32 AM
scott-0 retitled D13239: [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM. from to [ARM] Modify codegen for memcpy intrinsic to prefer LDM/STM..
Sep 29 2015, 3:22 AM

Aug 27 2015

scott-0 added a comment to D12366: Avoid unnecessarily storing vtable pointers in more destructor cases.

Thanks for the review and advice; I'll give undef a try. It's a much simpler approach.

Aug 27 2015, 12:43 AM

Aug 26 2015

scott-0 added a parent revision for D12366: Avoid unnecessarily storing vtable pointers in more destructor cases: D12365: Make SideEffectFinder publicly available (nfc).
Aug 26 2015, 8:23 AM
scott-0 added a child revision for D12365: Make SideEffectFinder publicly available (nfc): D12366: Avoid unnecessarily storing vtable pointers in more destructor cases.
Aug 26 2015, 8:23 AM
scott-0 retitled D12366: Avoid unnecessarily storing vtable pointers in more destructor cases from to Avoid unnecessarily storing vtable pointers in more destructor cases.
Aug 26 2015, 8:23 AM
scott-0 retitled D12365: Make SideEffectFinder publicly available (nfc) from to Make SideEffectFinder publicly available (nfc).
Aug 26 2015, 8:21 AM

Aug 24 2015

scott-0 committed rL245830: [ARM] Use AEABI helpers for i64 div and rem.
[ARM] Use AEABI helpers for i64 div and rem
Aug 24 2015, 2:18 AM
scott-0 closed D12232: [ARM] Use AEABI helpers for i64 div and rem by committing rL245830: [ARM] Use AEABI helpers for i64 div and rem.
Aug 24 2015, 2:18 AM
scott-0 committed rL245829: [ARM] Refactor LowerDivRem before adding LowerREM (nfc).
[ARM] Refactor LowerDivRem before adding LowerREM (nfc)
Aug 24 2015, 2:18 AM
scott-0 closed D12230: [ARM] Refactor LowerDivRem before adding LowerREM (nfc) by committing rL245829: [ARM] Refactor LowerDivRem before adding LowerREM (nfc).
Aug 24 2015, 2:18 AM

Aug 21 2015

scott-0 updated D12232: [ARM] Use AEABI helpers for i64 div and rem.
Aug 21 2015, 3:21 AM
scott-0 added a parent revision for D12232: [ARM] Use AEABI helpers for i64 div and rem: D12230: [ARM] Refactor LowerDivRem before adding LowerREM (nfc).
Aug 21 2015, 3:21 AM
scott-0 added a child revision for D12230: [ARM] Refactor LowerDivRem before adding LowerREM (nfc): D12232: [ARM] Use AEABI helpers for i64 div and rem.
Aug 21 2015, 3:21 AM
scott-0 retitled D12232: [ARM] Use AEABI helpers for i64 div and rem from to [ARM] Use AEABI helpers for i64 div and rem.
Aug 21 2015, 3:20 AM
scott-0 retitled D12230: [ARM] Refactor LowerDivRem before adding LowerREM (nfc) from to [ARM] Refactor LowerDivRem before adding LowerREM (nfc).
Aug 21 2015, 3:18 AM

Jul 30 2015

scott-0 added a comment to D11459: [Driver] Fix handling of -fbuiltin/-fcommon when combined with -mkernel.

The change looks right to me (but I'm not in a position to approve committing).

Jul 30 2015, 6:48 AM

Jul 13 2015

scott-0 committed rL242036: [ARM] Handle commutativity when converting to tADDhirr in Thumb2.
[ARM] Handle commutativity when converting to tADDhirr in Thumb2
Jul 13 2015, 8:32 AM
scott-0 closed D11132: [ARM] Handle commutativity when converting to tADDhirr in Thumb2 by committing rL242036: [ARM] Handle commutativity when converting to tADDhirr in Thumb2.
Jul 13 2015, 8:32 AM
scott-0 committed rL242035: [ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2.
[ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2
Jul 13 2015, 8:31 AM
scott-0 committed rL242034: [ARM] Small refactor of tryConvertingToTwoOperandForm (nfc).
[ARM] Small refactor of tryConvertingToTwoOperandForm (nfc)
Jul 13 2015, 8:31 AM
scott-0 closed D11131: [ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2 by committing rL242035: [ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2.
Jul 13 2015, 8:31 AM
scott-0 closed D11130: [ARM] Small refactor of tryConvertingToTwoOperandForm (nfc) by committing rL242034: [ARM] Small refactor of tryConvertingToTwoOperandForm (nfc).
Jul 13 2015, 8:31 AM
scott-0 updated the diff for D11132: [ARM] Handle commutativity when converting to tADDhirr in Thumb2.

Also, run thumb_rewrite.s tests in Thumb2 now that they pass.

Jul 13 2015, 8:20 AM
scott-0 updated the diff for D11131: [ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2.

Yes, that's a good idea; added.

Jul 13 2015, 8:19 AM
scott-0 added a comment to D11131: [ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2.

Thanks for the review!

Jul 13 2015, 6:24 AM
scott-0 retitled D11132: [ARM] Handle commutativity when converting to tADDhirr in Thumb2 from to [ARM] Handle commutativity when converting to tADDhirr in Thumb2.
Jul 13 2015, 1:35 AM
scott-0 retitled D11131: [ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2 from to [ARM] Add Thumb2 ADD with SP narrowing from 3 operand to 2.
Jul 13 2015, 1:33 AM
scott-0 retitled D11130: [ARM] Small refactor of tryConvertingToTwoOperandForm (nfc) from to [ARM] Small refactor of tryConvertingToTwoOperandForm (nfc).
Jul 13 2015, 1:32 AM

Jul 9 2015

scott-0 committed rL241802: [ARM] Thumb1 3 to 2 operand convertion for commutative operations.
[ARM] Thumb1 3 to 2 operand convertion for commutative operations
Jul 9 2015, 7:14 AM
scott-0 committed rL241801: [ARM] Don't be overzealous converting Thumb1 3 to 2 operands.
[ARM] Don't be overzealous converting Thumb1 3 to 2 operands
Jul 9 2015, 7:14 AM
scott-0 closed D11056: [ARM] Don't be overzealous converting Thumb1 3 to 2 operands by committing rL241801: [ARM] Don't be overzealous converting Thumb1 3 to 2 operands.
Jul 9 2015, 7:14 AM
scott-0 closed D11057: [ARM] Thumb1 3 to 2 operand convertion for commutative operations by committing rL241802: [ARM] Thumb1 3 to 2 operand convertion for commutative operations.
Jul 9 2015, 7:14 AM
scott-0 committed rL241800: [ARM] Add Thumb2 ADD with PC narrowing from 3 operand to 2.
[ARM] Add Thumb2 ADD with PC narrowing from 3 operand to 2
Jul 9 2015, 7:14 AM
scott-0 committed rL241799: [ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc).
[ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc)
Jul 9 2015, 7:14 AM
scott-0 closed D11055: [ARM] Add Thumb2 ADD with PC narrowing from 3 operand to 2 operand by committing rL241800: [ARM] Add Thumb2 ADD with PC narrowing from 3 operand to 2.
Jul 9 2015, 7:14 AM
scott-0 closed D11054: [ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc) by committing rL241799: [ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc).
Jul 9 2015, 7:14 AM
scott-0 committed rL241798: [ARM] Add ADD tests for Thumb2 narrowing (nfc).
[ARM] Add ADD tests for Thumb2 narrowing (nfc)
Jul 9 2015, 7:13 AM
scott-0 closed D11053: [ARM] Add ADD tests for Thumb2 narrowing (nfc) by committing rL241798: [ARM] Add ADD tests for Thumb2 narrowing (nfc).
Jul 9 2015, 7:13 AM
scott-0 added a comment to D11056: [ARM] Don't be overzealous converting Thumb1 3 to 2 operands.

This one?
@ FIXME: ARMARM says 'add sp, sp, #32'
@ CHECK: add sp, #32 @ encoding: [0x08,0xb0]

Jul 9 2015, 6:13 AM
scott-0 added a comment to D11055: [ARM] Add Thumb2 ADD with PC narrowing from 3 operand to 2 operand.

Thanks!

Jul 9 2015, 6:08 AM
scott-0 added a comment to D11054: [ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc).

Thanks for the review!

Jul 9 2015, 6:06 AM
scott-0 added a comment to D11053: [ARM] Add ADD tests for Thumb2 narrowing (nfc).

Thanks for the review!

Jul 9 2015, 6:01 AM
scott-0 retitled D11057: [ARM] Thumb1 3 to 2 operand convertion for commutative operations from to [ARM] Thumb1 3 to 2 operand convertion for commutative operations.
Jul 9 2015, 3:36 AM
scott-0 retitled D11056: [ARM] Don't be overzealous converting Thumb1 3 to 2 operands from to [ARM] Don't be overzealous converting Thumb1 3 to 2 operands.
Jul 9 2015, 3:33 AM
scott-0 retitled D11055: [ARM] Add Thumb2 ADD with PC narrowing from 3 operand to 2 operand from to [ARM] Add Thumb2 ADD with PC narrowing from 3 operand to 2 operand.
Jul 9 2015, 3:27 AM
scott-0 retitled D11054: [ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc) from Refactor converting Thumb1 from 3 to 2 operand (nfc) to [ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc).
Jul 9 2015, 3:23 AM
scott-0 retitled D11054: [ARM] Refactor converting Thumb1 from 3 to 2 operand (nfc) from to Refactor converting Thumb1 from 3 to 2 operand (nfc).
Jul 9 2015, 3:23 AM
scott-0 retitled D11053: [ARM] Add ADD tests for Thumb2 narrowing (nfc) from to [ARM] Add ADD tests for Thumb2 narrowing (nfc).
Jul 9 2015, 3:20 AM

Jul 1 2015

scott-0 accepted D10770: Expand Phabricator docs slightly.

committed as r241161

Jul 1 2015, 6:42 AM
scott-0 committed rL241161: Expand Phabricator docs slightly.
Expand Phabricator docs slightly
Jul 1 2015, 6:42 AM
scott-0 updated the diff for D10770: Expand Phabricator docs slightly.

Thanks for the reviews; comments addressed.

Jul 1 2015, 6:18 AM
scott-0 added a reviewer for D10770: Expand Phabricator docs slightly: klimek.
Jul 1 2015, 5:51 AM

Jun 26 2015

scott-0 retitled D10770: Expand Phabricator docs slightly from to Expand Phabricator docs slightly.
Jun 26 2015, 10:21 AM

Jun 15 2015

scott-0 added a comment to D10395: [ARM] Keep track of previous changes to the bit pattern stored HW_FP.

Thanks; I don't see any other problems.

Jun 15 2015, 3:25 AM
scott-0 added a comment to D10395: [ARM] Keep track of previous changes to the bit pattern stored HW_FP.

One minor comment.

Jun 15 2015, 3:13 AM

Jun 10 2015

scott-0 committed rL239476: some StmtExprs do not have side-effects.
some StmtExprs do not have side-effects
Jun 10 2015, 8:22 AM
scott-0 closed D10211: some StmtExprs do not have side-effects by committing rL239476: some StmtExprs do not have side-effects.
Jun 10 2015, 8:22 AM
scott-0 committed rL239474: add ConstEvaluatedExprVisitor.
add ConstEvaluatedExprVisitor
Jun 10 2015, 6:57 AM
scott-0 closed D10210: add ConstEvaluatedExprVisitor by committing rL239474: add ConstEvaluatedExprVisitor.
Jun 10 2015, 6:57 AM

Jun 5 2015

scott-0 updated the diff for D10211: some StmtExprs do not have side-effects.

Thanks for the reviews! I have addressed the review comments. (But I still need a review on the prereq differential http://reviews.llvm.org/D10210)

Jun 5 2015, 8:53 AM

Jun 3 2015

scott-0 added a comment to D10211: some StmtExprs do not have side-effects.

I see two typos in comments already "side-efect" and "warnig" -- I'll be sure to fix those.

Jun 3 2015, 3:48 AM
scott-0 retitled D10211: some StmtExprs do not have side-effects from to some StmtExprs do not have side-effects.
Jun 3 2015, 3:44 AM
scott-0 retitled D10210: add ConstEvaluatedExprVisitor from to add ConstEvaluatedExprVisitor.
Jun 3 2015, 3:39 AM

May 15 2015

scott-0 added a comment to D9747: [AArch64] Mark SMIN/SMAX/UMIN/UMAX nodes as legal and add patterns for them.

LGTM

May 15 2015, 4:46 AM

Apr 8 2015

scott-0 committed rL234421: [ARM] make vminnm/vmaxnm work with ?le, ?ge and no-nans-fp-math.
[ARM] make vminnm/vmaxnm work with ?le, ?ge and no-nans-fp-math
Apr 8 2015, 10:21 AM
scott-0 closed D8725: [ARM] make vminnm/vmaxnm work with ?le, ?ge and no-nans-fp-math by committing rL234421: [ARM] make vminnm/vmaxnm work with ?le, ?ge and no-nans-fp-math.
Apr 8 2015, 10:21 AM

Apr 1 2015

scott-0 closed D8720: [docs] add cross-reference.

Closing

Apr 1 2015, 3:06 AM

Mar 31 2015

scott-0 retitled D8725: [ARM] make vminnm/vmaxnm work with ?le, ?ge and no-nans-fp-math from to [ARM] make vminnm/vmaxnm work with ?le, ?ge and no-nans-fp-math .
Mar 31 2015, 9:47 AM
scott-0 added a comment to D8720: [docs] add cross-reference.

Thanks! Committed in r233699.

Mar 31 2015, 9:16 AM
scott-0 committed rL233699: [docs] add cross-reference.
[docs] add cross-reference
Mar 31 2015, 8:10 AM
scott-0 added a comment to D8720: [docs] add cross-reference.

Yes, that's essentially what I did before I created the review. [I used 'make -f Makefile.sphinx' which does the same thing.]

Mar 31 2015, 7:58 AM
scott-0 added a comment to D8720: [docs] add cross-reference.

It did when I tried it with sphinx, but it is my first experience creating a ReST link.

Mar 31 2015, 7:39 AM
scott-0 retitled D8720: [docs] add cross-reference from to [docs] add cross-reference.
Mar 31 2015, 7:34 AM

Mar 23 2015

scott-0 committed rL232940: handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob.
handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob
Mar 23 2015, 3:57 AM
scott-0 closed D8196: handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob by committing rL232940: handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob.
Mar 23 2015, 3:57 AM
scott-0 committed rL232939: Convert cascading if-else-if to switch. NFC.
Convert cascading if-else-if to switch. NFC
Mar 23 2015, 3:49 AM
scott-0 closed D8485: Convert cascading if-else-if to switch. NFC by committing rL232939: Convert cascading if-else-if to switch. NFC.
Mar 23 2015, 3:49 AM

Mar 20 2015

scott-0 added a comment to D8485: Convert cascading if-else-if to switch. NFC.

Title should say "Convert cascading if-else-if to switch. NFC"

Mar 20 2015, 12:28 PM
scott-0 updated the diff for D8196: handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob.

rebased to be after cleanup in http://reviews.llvm.org/D8485

Mar 20 2015, 10:19 AM
scott-0 retitled D8485: Convert cascading if-else-if to switch. NFC from to covert cascading if-else-if to switch.
Mar 20 2015, 10:18 AM

Mar 16 2015

scott-0 added a comment to D8197: Add commit message guidelines to developer policy.

Is it worth adding a mention of "Differential Revision: <URL>" as per http://llvm.org/docs/Phabricator.html#committing-a-change

Mar 16 2015, 4:58 AM

Mar 10 2015

scott-0 added a comment to D8196: handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob.

Ok, thanks. I'd rather do the whole function as a separate, no-functional-change commit. Are you happy with this as-is?

Mar 10 2015, 9:23 AM
scott-0 updated the diff for D8196: handle armeb/thumb/thumbeb consistently in gnutools::Assemble::ConstructJob.

Is this what you had in mind?

Mar 10 2015, 8:55 AM