- User Since
- Aug 20 2015, 4:19 PM (191 w, 15 h)
Thu, Apr 11
Wed, Apr 10
Integrated David's comments.
Tue, Apr 2
Mar 12 2019
Mar 11 2019
Integrated Ruiu's review comments.
Mar 8 2019
using utils/update_test_checks.py to generate the test checks.
changed test case based on spatel's comments.
add one missing updates in last version of the patch.
Integrated Teresa's comments.
Mar 6 2019
Mar 5 2019
Mar 4 2019
Added usage example to UserManual.rst suggested by David.
Mar 1 2019
Integrated Teresa's suggestion to change the err() to assert.
Added documentation change suggested by David.
Feb 28 2019
Feb 27 2019
BTW, I'm in the process of committing D54175. After that patch, PGO instrumentation can be called after the main inlining. I don't think it will conflict anything in this patch
Feb 26 2019
Feb 12 2019
Removed the unrelated instructions (comments) from the tests.
Feb 8 2019
Updated the patch to sync with D54175
Forgot to include one more test:
Fixed the missing changes David mentioned.
Feb 7 2019
Integrated the review comments from Teresa and David.
Added unit tests for
(2) context-sensitive profile summary
(3) opt new options and passes
(4) gold plugin
(5) using lto2 to test csgen and csuse.
Feb 5 2019
Jan 28 2019
Fixed a pass ordering issue in legacy pass manager for ThinLTO: Need to place CS Instrumentation after EliminateAvailableExternallyPass. This can cause duplicate profile variables that leads to bad profile for online profile merge.
Jan 23 2019
Integrated Teresa's review comments.
Jan 22 2019
Jan 15 2019
This cleanup looks good to me.
As Teresa suggested, split unrelated change to r351257 (NFC)
Jan 14 2019
An update: I now use a separated pass to create the comdat variables before LTO/ThinLTO. @ruir: You can discard this patch now. Thanks the help from ruir and pcc.
Fixed the issue with lld (where lld is not happy with the COMDAT variables created
Update the patch to sync with
Update the patch to sync with https://reviews.llvm.org/D54175
Jan 9 2019
@kristina: This patch does not functionally change anything in current code base. We think this patch makes more sense than the original one (which also contributed by us)
Current code works fine (as all the callers to getPGOFuncName() are before LTO). Tests for stripping behavior is already in the test suite.
It will break once that functions is called in LTO or after LTO with unknown function error. This will show in patch here https://reviews.llvm.org/D54175.