Commit Graph

70 Commits

Author SHA1 Message Date
George Karpenkov 3c2ed8f338 [analyzer] Correct modelling of OSDynamicCast: eagerly state split
Previously, OSDynamicCast was modeled as an identity.

This is not correct: the output of OSDynamicCast may be zero even if the
input was not zero (if the class is not of desired type), and thus the
modeling led to false positives.

Instead, we are doing eager state split:
in one branch, the returned value is identical to the input parameter,
and in the other branch, the returned value is zero.

This patch required a substantial refactoring of canEval infrastructure,
as now it can return different function summaries, and not just true/false.

rdar://45497400

Differential Revision: https://reviews.llvm.org/D53624

llvm-svn: 345338
2018-10-25 23:38:07 +00:00
George Karpenkov d3e7675331 [analyzer] [NFC] Change scanReachableSymbols to use ranges
Remove unused overload. Clean up some usages.

Differential Revision: https://reviews.llvm.org/D53615

llvm-svn: 345101
2018-10-23 23:12:12 +00:00
George Karpenkov 081c47760c [analyzer] Do not stop tracking CXX methods touching OSObject.
Trust generalized annotations for OSObject.

Differential Revision: https://reviews.llvm.org/D53550

llvm-svn: 345100
2018-10-23 23:11:50 +00:00
George Karpenkov 48de582f65 [analyzer] Trust summaries for OSObject::retain and OSObject::release
Refactor the way in which summaries are consumed for safeMetaCast

Differential Revision: https://reviews.llvm.org/D53549

llvm-svn: 345099
2018-10-23 23:11:30 +00:00
Artem Dergachev 65b4d7ddd9 [analyzer] NFC: RetainCountChecker: Don't dump() symbols into program point tags.
We don't need a separate node for every symbol, because whenever the first
symbol leaks, a bug is emitted, the analysis is sinked, and the checker
callback immediately returns due to State variable turning into null,
so we never get to see the second leaking symbol.

Additionally, we are no longer able to break normal analysis while experimenting
with debug dumps.

Differential Revision: https://reviews.llvm.org/D52804

llvm-svn: 344538
2018-10-15 17:47:56 +00:00
George Karpenkov ca7a23e761 [analyzer] Avoid unneeded invalidation in RetainCountChecker
Differential Revision: https://reviews.llvm.org/D53168

llvm-svn: 344312
2018-10-11 22:59:38 +00:00
George Karpenkov 41dc8de6ae [analyzer] Retain count checker for OSObject: recognize OSDynamicCast
For now, tresting the cast as a no-op, and disregarding the case where
the output becomes null due to the type mismatch.

rdar://45174557

Differential Revision: https://reviews.llvm.org/D53156

llvm-svn: 344311
2018-10-11 22:59:16 +00:00
George Karpenkov 04553e530f [analyzer] Process state in checkEndFunction in RetainCountChecker
Modify the RetainCountChecker to perform state "adjustments" in
checkEndFunction, as performing work in PreStmt<ReturnStmt> does not
work with destructors.
The previous version made an implicit assumption that no code runs
after the return statement is executed.

rdar://43945028

Differential Revision: https://reviews.llvm.org/D52338

llvm-svn: 342770
2018-09-21 20:37:20 +00:00
George Karpenkov 6babf2ae16 [analyzer] [NFC] Prefer make_unique over "new"
Differential Revision: https://reviews.llvm.org/D52336

llvm-svn: 342767
2018-09-21 20:36:21 +00:00
George Karpenkov d5ef0d2a4b [analyzer] Better retain count rules for OSObjects
Differential Revision: https://reviews.llvm.org/D51184

llvm-svn: 340961
2018-08-29 20:28:33 +00:00
George Karpenkov ab0011ebc0 [analyzer] Preliminary version of retain count checking for OSObjects
Has quite a lot of false positives, disabled behind the flag.

Differential Revision: https://reviews.llvm.org/D50880

llvm-svn: 340502
2018-08-23 00:26:59 +00:00
George Karpenkov 6e9fd1377d [analyzer] [NFC] Fix minor formatting issues in RetainCountChecker
Differential Revision: https://reviews.llvm.org/D51072

llvm-svn: 340378
2018-08-22 01:17:09 +00:00
George Karpenkov 80c9e78e3b [analyzer] [NFC] Extract a method for creating RefVal from RetEffect in RetainCountChecker
Differential Revision: https://reviews.llvm.org/D51071

llvm-svn: 340377
2018-08-22 01:16:49 +00:00
George Karpenkov efef49cd2f [analyzer] [NFC] Split up RetainSummaryManager from RetainCountChecker - try #2
Turns out it can't be removed from the analyzer since it relies on CallEvent.

Moving to staticAnalyzer/core

Differential Revision: https://reviews.llvm.org/D51023

llvm-svn: 340247
2018-08-21 03:09:02 +00:00
Bruno Cardoso Lopes bb3b7cff96 Revert "[analyzer] [NFC] Split up RetainSummaryManager from RetainCountChecker"
This reverts commit a786521fa66c72edd308baff0c08961b6d964fb1.

Bots haven't caught up yet, but broke modules build with:

../tools/clang/include/clang/StaticAnalyzer/Checkers/MPIFunctionClassifier.h:18:10:
fatal error: cyclic dependency in module 'Clang_StaticAnalyzer_Core':
Clang_StaticAnalyzer_Core -> Clang_Analysis ->
Clang_StaticAnalyzer_Checkers -> Clang_StaticAnalyzer_Core
         ^

llvm-svn: 340117
2018-08-18 03:22:11 +00:00
George Karpenkov 0ac54fad53 [analyzer] [NFC] Split up RetainSummaryManager from RetainCountChecker
ARCMigrator is using code from RetainCountChecker, which is a layering
violation (and it also does it badly, by using a different header, and
then relying on implementation being present in a header file).

This change splits up RetainSummaryManager into a separate library in
lib/Analysis, which can be used independently of a checker.

Differential Revision: https://reviews.llvm.org/D50934

llvm-svn: 340114
2018-08-18 01:45:50 +00:00
George Karpenkov b1b791b9b1 [analyzer] [NFC] Minor refactoring of ISL-specific code in RetainCountChecker
Differential Revision: https://reviews.llvm.org/D50879

llvm-svn: 340098
2018-08-17 21:43:27 +00:00
George Karpenkov bc0cddf0c8 [analyzer] Re-instate support for MakeCollectable is RetainCountChecker
Differential Revision: https://reviews.llvm.org/D50872

llvm-svn: 340097
2018-08-17 21:42:59 +00:00
George Karpenkov c4d6b93dc7 [analyzer] [NFC] Move canEval function from RetainCountChecker to RetainCountSummaries
Differential Revision: https://reviews.llvm.org/D50863

llvm-svn: 340094
2018-08-17 21:42:05 +00:00
George Karpenkov 70c2ee30bc [analyzer] [NFC] Split up RetainCountChecker
At some point, staring at 4k+ LOC file becomes a bit hard.

Differential Revision: https://reviews.llvm.org/D50821

llvm-svn: 340092
2018-08-17 21:41:07 +00:00