Compare commits

..

281 Commits

Author SHA1 Message Date
Chris Smowton
2aa5cf84ff Merge pull request #18215 from smowton/smowton/ke2/basic-generics
KE2: Upgrade to Kotlin 2.1.0; restore basic type parameter and type argument extraction
2024-12-13 15:39:54 +00:00
Chris Smowton
b60298d033 Rename classId 2024-12-11 15:58:44 +00:00
Chris Smowton
53593a39f0 Note Kotlin class ID distinction is a TODO 2024-12-11 15:57:09 +00:00
Chris Smowton
44e44dcce9 Cascade failures from failure to find a type-parameter parent. 2024-12-11 15:46:03 +00:00
Chris Smowton
504a630123 Note class of unexpected type arguments 2024-12-11 15:32:44 +00:00
Chris Smowton
448d3680f6 Add comment noting structure of argsIncludingOuterClasses 2024-12-11 15:31:45 +00:00
Chris Smowton
43576a169f Add comment noting Caffeiene dependency 2024-12-11 15:29:15 +00:00
Ian Lynagh
fbead0fd63 Merge pull request #18254 from igfoo/igfoo/extractExpressionStmt
KE2: extractExpressionStmt can be used with null statements
2024-12-10 11:25:16 +00:00
Ian Lynagh
0f2634a228 KE2: extractExpressionStmt can be used with null statements 2024-12-09 16:54:23 +00:00
Ian Lynagh
8c8599435e Merge pull request #18250 from igfoo/igfoo/extractExpressionExpr
KE2: extractExpressionExpr can take null expressions
2024-12-09 14:38:42 +00:00
Ian Lynagh
f4ae7f8e81 KE2: extractExpressionExpr can take null expressions 2024-12-09 14:31:43 +00:00
Ian Lynagh
b1683f7549 Merge pull request #18237 from igfoo/igfoo/ret
KE2: extractExpression always returns an ID
2024-12-09 11:48:34 +00:00
Ian Lynagh
ada6801a17 KE2: extractExpression always returns an ID
It used to sometimes return null, which could mean either it extracted a
statement or it failed to extract an expression. Also, what it returned
didn't take into account any ExprStmt or StmtExpr wrappers.

Now, it will always return an ID of the type that it StmtExprParent
corresponds to.
2024-12-06 15:23:42 +00:00
Ian Lynagh
d568d04357 Merge pull request #18210 from igfoo/igfoo/nullExpr
KE2: Start generating errorexprs/errorstmts
2024-12-05 13:19:28 +00:00
Ian Lynagh
d36fabf4ec KE2: Add a TODO 2024-12-05 13:02:09 +00:00
Chris Smowton
784a63b6d5 KE2: Restore the basics of type-parameter and type-argument extraction 2024-12-04 22:38:06 +00:00
Chris Smowton
2cc2d931f3 KE2: Upgrade to Kotlin 2.1.0 2024-12-04 22:38:05 +00:00
Chris Smowton
51a1ea52e1 Merge pull request #18150 from smowton/smowton/ke2/external-class-extraction
KE2: restore basic function label construction
2024-12-04 22:37:37 +00:00
Chris Smowton
0d39ab21c5 Address review comments 2024-12-04 22:37:09 +00:00
Ian Lynagh
7f6818042d KE2: Generate erors for null expressions/statements 2024-12-04 18:03:54 +00:00
Ian Lynagh
af1804380a KE2: Add support for generating error expressions and statements 2024-12-04 17:52:26 +00:00
Ian Lynagh
d8a9615c0c Merge pull request #18182 from igfoo/igfoo/bb
KE2: Remove some unnecessary !!s
2024-12-03 14:18:49 +00:00
Ian Lynagh
769a615de1 KE2: Remove some unnecessary !!s
useType already handles null types, and extracts an error type for them.

The error also includes info about where that came from via the `with`
stack, although we might want to make that finer grained in future.
2024-12-03 14:18:03 +00:00
Ian Lynagh
b1b8717718 Merge pull request #18183 from igfoo/igfoo/callable
KE2: Put 'callable' into 'StmtParent'
2024-12-03 12:42:50 +00:00
Ian Lynagh
703aee2ae6 KE2: Remove most redundant 'callable' args 2024-12-02 18:28:15 +00:00
Ian Lynagh
034f283c4f KE2: Tell StmtExprParent about callable
This should allow us to simplify everything that uses it.
2024-12-02 18:02:11 +00:00
Tamás Vajk
439e8f079c Merge pull request #18128 from tamasvajk/ke-constants
KE2: Extract `bool`, `char`, `float`, `double` constants
2024-12-02 14:59:01 +01:00
Tamas Vajk
98ab6213a4 Code quality improvement 2024-12-02 14:26:10 +01:00
Ian Lynagh
2490606cd1 Merge pull request #18168 from igfoo/igfoo/nofake
KE2: Remove the fakeOverride code
2024-12-02 12:18:39 +00:00
Tamas Vajk
6118253b14 Code quality improvements 2024-12-02 12:07:44 +01:00
Tamas Vajk
149136c2a4 KE2: Extract bool, char, float, double constants 2024-12-02 11:58:12 +01:00
Ian Lynagh
0ccf117bf7 KE2: Remove the fakeOverride code
As far as I can see, the analysis API isn't giving us fake overrides.
2024-11-29 17:49:13 +00:00
Ian Lynagh
7b4e830386 Merge pull request #18149 from igfoo/igfoo/with
KE2: Small method renaming
2024-11-28 15:44:26 +00:00
Chris Smowton
cf78938a0d KE2: restore basic function label construction 2024-11-28 15:38:13 +00:00
Ian Lynagh
078e292c74 Merge pull request #18148 from igfoo/igfoo/dollar
KE2: Simplify escaping a dollar in a string
2024-11-28 15:25:52 +00:00
Ian Lynagh
194a61945e KE2: Small method renaming 2024-11-28 14:50:02 +00:00
Ian Lynagh
4765917d34 KE2: Simplify escaping a dollar in a string 2024-11-28 14:38:51 +00:00
Ian Lynagh
51c79952f3 Merge pull request #18146 from igfoo/igfoo/fix
KE2: Fix build
2024-11-28 13:04:33 +00:00
Ian Lynagh
433f5d311b KE2: Fix build 2024-11-28 12:36:20 +00:00
Tamás Vajk
0572e28adc Merge pull request #18127 from tamasvajk/ke-null
KE2: Extract `null` literal
2024-11-28 09:11:05 +01:00
Chris Smowton
222b50cd5e Merge pull request #18134 from smowton/smowton/ke2/external-class-extraction
KE2: basic external class extraction
2024-11-27 18:15:33 +00:00
Chris Smowton
fe4dc296f5 Don't query non-Kt source elements for locations etc 2024-11-27 18:04:40 +00:00
Chris Smowton
54961ddc88 Fixups 2024-11-27 17:54:45 +00:00
Ian Lynagh
d46cb189d8 Merge pull request #18135 from igfoo/igfoo/priv_unused
KE2: Remove some debugging functions, and mark some others as private
2024-11-27 17:43:52 +00:00
Chris Smowton
d27b5ed96e Remove redundant comment 2024-11-27 17:12:22 +00:00
Chris Smowton
dd9d8720b0 Add doc comment 2024-11-27 17:12:21 +00:00
Chris Smowton
a3d78f1bad Neaten symbol-to-location 2024-11-27 17:12:20 +00:00
Chris Smowton
cc0a112ea6 Generalise warnElement and errorElement 2024-11-27 17:12:19 +00:00
Chris Smowton
97ecd18678 Merge duplicate functions 2024-11-27 17:12:18 +00:00
Chris Smowton
e29d9ddacb Restore location and name reporting for symbols 2024-11-27 17:12:17 +00:00
Chris Smowton
bfdb5e0b17 Add error function taking a throwable to LoggerBase 2024-11-27 17:12:16 +00:00
Chris Smowton
dfad8c8475 Don't bubble TODOs and other unchecked exceptions up to top level 2024-11-27 17:12:15 +00:00
Chris Smowton
1fc2a61f95 KE2: basic external class extraction 2024-11-27 17:12:12 +00:00
Tamas Vajk
def1916fd8 KE2: Extract null literal 2024-11-27 16:47:12 +01:00
Tamás Vajk
7e77ad2e71 Merge pull request #18110 from tamasvajk/ke2-lambda
KE2: Extract lambda expressions
2024-11-27 16:43:47 +01:00
Ian Lynagh
75f1c08ea2 KE2: Remove some debugging functions, and mark some others as private 2024-11-27 15:25:32 +00:00
Tamas Vajk
352e5d0c68 Remove unused code 2024-11-27 15:45:38 +01:00
Tamas Vajk
7d50eb5670 Fix review findings 2024-11-27 14:38:05 +01:00
Tamas Vajk
44e318546f KE2: Extract more constructs for lambda expressions 2024-11-27 14:38:05 +01:00
Tamas Vajk
b42fbde130 KE2: Extract generated class for lambda expressions 2024-11-27 14:37:58 +01:00
Ian Lynagh
5245dad3c1 Merge pull request #18118 from igfoo/igfoo/diag
KE2: Put diagnostics from the analysis API into the database
2024-11-27 10:57:22 +00:00
Ian Lynagh
cc0eb9ab36 KE2: Put diagnostics from the analysis API into the database 2024-11-26 15:42:38 +00:00
Ian Lynagh
48168bf66c Merge pull request #18096 from igfoo/igfoo/deprec
KE2: Don't actually deprecate WhenBranch.getCondition() yet
2024-11-26 14:48:27 +00:00
Ian Lynagh
661fb9ee58 Merge pull request #18095 from igfoo/igfoo/remove
KE2: Remove some old debugging code
2024-11-26 14:48:16 +00:00
Ian Lynagh
2c595417f1 KE2: Don't actually deprecate WhenBranch.getCondition() yet
It makes a lot of noise in the CFG QLL, that we aren't fixing yet
2024-11-25 17:14:35 +00:00
Ian Lynagh
0b529c92bc KE2: Remove some old debugging code 2024-11-25 17:04:27 +00:00
Ian Lynagh
86ddb3b6c1 Merge pull request #18081 from igfoo/igfoo/dbscheme_comments
KE2: Add more dbscheme comments
2024-11-25 12:06:18 +00:00
Tamás Vajk
0103711b47 Merge pull request #18058 from tamasvajk/ke2-when
KE2: Extract `when` expressions
2024-11-25 09:04:24 +01:00
Ian Lynagh
bb50bc0d85 Merge pull request #18075 from igfoo/igfoo/comp
KE2: Small refactoring
2024-11-22 15:34:28 +00:00
Ian Lynagh
37e950dcbf Merge pull request #18076 from igfoo/igfoo/werror
KE2: Add warnings-as-error to build system, but commented out for now
2024-11-22 15:33:51 +00:00
Ian Lynagh
b816c1f396 Merge pull request #18077 from igfoo/igfoo/stmt
KE2: Reenable more code for ExprParent.stmt
2024-11-22 15:33:38 +00:00
Ian Lynagh
bafee5ec10 Merge pull request #18079 from igfoo/igfoo/dc
KE2: Remove some dead code
2024-11-22 15:33:20 +00:00
Tamas Vajk
3abd9a755e Code quality improvements 2024-11-22 16:22:39 +01:00
Ian Lynagh
b3dbd73741 KE2: Remove some dead code 2024-11-22 14:10:47 +00:00
Ian Lynagh
19986f0307 KE2: Reenable more code for ExprParent.stmt 2024-11-22 14:04:02 +00:00
Ian Lynagh
cb8237fe67 KE2: Add warnings-as-error to build system, but commented out for now
Once we get closer to completion, it will be useful to have this on.
2024-11-22 13:42:31 +00:00
Ian Lynagh
d280a41062 KE2: Small refactoring
Avoids shadowing `trapWriterWriteExpr`, and removes the need to check
for an impossible case.
2024-11-22 13:39:55 +00:00
Ian Lynagh
05fa3328f0 Merge pull request #18064 from igfoo/igfoo/ke2_lang_ver
KE2: Use the right language version
2024-11-22 11:41:32 +00:00
Ian Lynagh
40006fc566 Merge pull request #18031 from igfoo/igfoo/kttypes
KE2: Start working on KtTypes
2024-11-22 11:41:01 +00:00
Tamas Vajk
6c8cb103fc Fix KE1 2024-11-22 11:37:09 +01:00
Tamas Vajk
052a243db6 Fix KE1 to extract the new when condition constructs 2024-11-22 10:16:41 +01:00
Ian Lynagh
6d990d47db KE2: Use the right language version
With this, if I make the testsuite driver use 1.7, then the test code

sealed interface ReadResult
data class Number(val number: Int) : ReadResult
data class Text(val text: String) : ReadResult
data object EndOfFile : ReadResult

makes the extractor print

=== Diagnostics
--- Diagnostic:
WRONG_MODIFIER_TARGET
ERROR
Modifier 'data' is not applicable to 'standalone object'.
Location(startLine=5, startColumn=1, endLine=5, endColumn=4)
--- End diagnostics
2024-11-21 16:36:16 +00:00
Ian Lynagh
82c41316c6 KE2: Populate Kotlin type nullability and alias information 2024-11-21 16:00:01 +00:00
Ian Lynagh
7baeea6365 KE2: Use a more consistent TRAP label 2024-11-21 15:21:35 +00:00
Ian Lynagh
d17e3d521c KE2: Start working on KtTypes 2024-11-21 15:21:34 +00:00
Ian Lynagh
eae40dbc03 KE2: Keep KE1 building with the kt-type changes
It won't work, but it'll still compile
2024-11-21 15:21:33 +00:00
Tamas Vajk
a2d90ed0c6 KE2: Extract when expressions 2024-11-21 16:02:20 +01:00
Ian Lynagh
74ee483fa1 KE2: Add rules_jvm_external back to build system 2024-11-20 18:23:01 +00:00
Ian Lynagh
8fe48d6dce Merge commit 'e3990b7d04db2ca3ac99c029a0afc131e695db0b' into ke2
That is the repo ql as at the internal repo's
    git merge-base origin/rc/3.16 origin/main
2024-11-20 17:40:00 +00:00
Chris Smowton
e89e0f5c4a Merge pull request #18027 from igfoo/igfoo/fixbuild
KE2: Fix build
2024-11-20 16:48:36 +00:00
Tamás Vajk
1dbf54e9e7 Merge pull request #18028 from tamasvajk/ke2-if
KE2: Extract `if` expressions/statements
2024-11-19 18:40:45 +01:00
Tamas Vajk
39aefb8d17 Fix code review finding 2024-11-19 18:06:35 +01:00
Tamas Vajk
28a5634615 KE2: Extract if expressions/statements 2024-11-19 13:57:18 +01:00
Ian Lynagh
e4a82888c0 KE2: Fix build 2024-11-19 11:56:55 +00:00
Ian Lynagh
147f6a10e7 Merge pull request #18016 from igfoo/igfoo/file_numbers
KE2: Use the right file numbers
2024-11-19 11:03:55 +00:00
Tamás Vajk
750b8239e7 Merge pull request #18006 from tamasvajk/ke2-parens
KE2: Extract parenthesized expressions
2024-11-19 11:23:39 +01:00
Tamas Vajk
ea54eab376 Add todo comment 2024-11-19 08:37:38 +01:00
Ian Lynagh
7bda00cb5b KE2: Use the right file numbers
The thread that did the extraction could see the file number counter
after it had been incremented (possibly multiple times) by the main
thread. This fixes some consistency query failures in tests.
2024-11-18 18:37:56 +00:00
Paolo Tranquilli
5c1f413d44 Java: fix embedded kotlin extractor build 2024-11-18 17:47:24 +01:00
Tamas Vajk
9f3a0ca432 KE2: Extract parenthesized expressions 2024-11-18 13:41:05 +01:00
Ian Lynagh
c2dfe0ef4a Merge pull request #17978 from igfoo/igfoo/issafeaccess
KE2: Rename safeAccess to isSafeAccess
2024-11-14 11:28:23 +00:00
Tamás Vajk
c3324ee2f4 Merge pull request #17974 from tamasvajk/ke2-code-quality-01
KE2: Improve code quality in expression extraction
2024-11-14 08:36:11 +01:00
Chris Smowton
efe20b2452 Merge pull request #17884 from smowton/ke2/properties-and-variables
KE2: implement basic usage of properties, variables and flexible types
2024-11-13 15:32:44 +00:00
Chris Smowton
f12818a96d Apply review comments 2024-11-13 15:15:09 +00:00
Chris Smowton
fdaa6c5b4b KE2: implement basic usage of properties, variables and flexible types 2024-11-13 15:06:59 +00:00
Ian Lynagh
22096b1984 KE2: Rename safeAccess to isSafeAccess
To follow our standard naming convention.
2024-11-13 12:32:36 +00:00
Tamas Vajk
3ae58d072c KE2: Improve code quality in expression extraction 2024-11-13 09:38:48 +01:00
Tamás Vajk
fcde605569 Merge pull request #17939 from tamasvajk/ke2-binary-op-compareTo
KE2: Extract `compareTo` calls for binary comparisons
2024-11-13 09:09:59 +01:00
Tamas Vajk
70658bcd52 Fix review findings 2024-11-12 15:58:42 +01:00
Tamas Vajk
14150ea78d KE2: Extract compareTo calls for binary comparisons 2024-11-12 15:49:44 +01:00
Tamás Vajk
15468bcd11 Merge pull request #17874 from tamasvajk/ke2-unary-ops
KE2: Extract unary operators
2024-11-12 15:48:47 +01:00
Tamas Vajk
320905925b KE2: Extract unary operators 2024-11-12 12:32:51 +01:00
Tamás Vajk
a9e45d8609 Merge pull request #17761 from tamasvajk/ke2-binary-ops
KE2: extract binary operators
2024-11-12 12:30:39 +01:00
Ian Lynagh
83b3e8c7e5 Merge pull request #17961 from igfoo/igfoo/ke2_erasure
KE2: Remove erasure
2024-11-12 10:18:39 +00:00
Tamas Vajk
53460d7ca0 Add comment 2024-11-12 09:19:04 +01:00
Ian Lynagh
cfb269eba9 KE2: Remove erasure 2024-11-11 17:49:23 +00:00
Ian Lynagh
0249c49ce5 Java: Add up/downgrade scripts 2024-11-11 17:48:35 +00:00
Ian Lynagh
66be970b2e Java/Kotlin: Remove the erasure relation
It's no longer used
2024-11-11 17:48:30 +00:00
Ian Lynagh
b8b0fcad67 Kotlin: Don't write the erasure relation
It's no longer used
2024-11-11 17:48:24 +00:00
Ian Lynagh
4aed952c7d Java: Remove redundant getErasure overrides
The root definition covers these cases already
2024-11-11 17:48:17 +00:00
Tamas Vajk
db13b32285 Extract comparison operators 2024-11-08 13:53:59 +01:00
Tamas Vajk
a5fcfaf289 Add todo comment with missing binary operators 2024-11-08 13:53:58 +01:00
Tamas Vajk
227d30243c Extract reference equals 2024-11-08 13:53:58 +01:00
Tamas Vajk
bc35c509f0 Extract more numeric binary operators 2024-11-08 13:53:57 +01:00
Tamas Vajk
255d5c9942 KE2: Extract binary operators on numeric types 2024-11-08 13:53:57 +01:00
Tamás Vajk
212143ff45 Merge pull request #17881 from tamasvajk/ke2-safe-qualified-expr
KE2: Extract safe qualified expressions
2024-11-08 13:53:21 +01:00
Tamás Vajk
71931c38f2 Merge pull request #17885 from smowton/smowton/ke2-jar-sources
KE2: add source jars to intellij project
2024-11-07 15:04:39 +01:00
Chris Smowton
33a0e99347 KE2: add source jars to intellij project 2024-10-31 17:58:23 +00:00
Tamas Vajk
84166e8731 KE2: Extract safe qualified expressions 2024-10-31 13:14:07 +01:00
Chris Smowton
26d40a7e42 Merge pull request #17878 from smowton/smowton/ke2/debugger-support
KE2: Enable attaching debugger to extractor
2024-10-31 10:22:52 +00:00
Tamás Vajk
f57fe719c1 Merge pull request #17802 from tamasvajk/ke2-method-call
KE2: Extract simple method calls
2024-10-31 08:53:38 +01:00
Tamas Vajk
11975a1b25 Remove todo comment 2024-10-31 08:28:42 +01:00
Chris Smowton
3e4345e0aa Enable attaching debugger to ke2 extractor 2024-10-29 22:14:26 +00:00
Tamas Vajk
9dd37b0ede Fix typo 2024-10-29 13:26:37 +01:00
Tamas Vajk
c10a0e549a Handle named arguments in method call extraction 2024-10-29 12:12:16 +01:00
Tamas Vajk
4bf6280435 KE2: Extract simple method calls 2024-10-18 14:06:08 +02:00
Ian Lynagh
a922f97200 Merge pull request #17777 from igfoo/igfoo/log-sev
KE2: Log our verbosity level
2024-10-16 12:46:58 +01:00
Tamás Vajk
9a4cd2152a Merge pull request #17752 from tamasvajk/ke2-string-plus
KE2: Extract `String.plus` and `String?.plus` calls
2024-10-16 13:35:08 +02:00
Ian Lynagh
9b13368e23 KE2: Log our verbosity level
This happens at `info` level, which is logged by default.
2024-10-15 16:23:25 +01:00
Tamas Vajk
7b198da95f Improve code quality 2024-10-15 10:29:14 +02:00
Tamas Vajk
125797cd4f Improve code quality 2024-10-14 20:31:52 +02:00
Tamas Vajk
a3a93d826e KE2: Extract String.plus and String?.plus calls 2024-10-14 14:39:01 +02:00
Tamás Vajk
bc15f40f8f Merge pull request #17729 from tamasvajk/ke2-numeric-plus
KE2: Extract binary plus on numeric types
2024-10-11 13:26:41 +02:00
Tamas Vajk
ea688372bd Apply review findings 2024-10-11 10:17:16 +02:00
Ian Lynagh
4b73fed267 KE2: Add more dbscheme comments 2024-10-10 17:45:29 +01:00
Tamas Vajk
7e8b20d200 KE2: Extract binary plus on numeric types 2024-10-10 14:30:24 +02:00
Tamás Vajk
643419a32f Merge pull request #17707 from tamasvajk/ke2-vari
KE2: Extract local variable declarations
2024-10-10 12:32:11 +02:00
Tamas Vajk
e82b1762c0 Apply code review findings 2024-10-09 16:02:54 +02:00
Tamas Vajk
a471fa004a KE2: Extract local variable declarations 2024-10-09 15:19:42 +02:00
Tamás Vajk
01c71ba8d6 Merge pull request #17706 from tamasvajk/ke2-is-as
KE2: Extract `is` and `as` expression kinds
2024-10-09 15:17:50 +02:00
Ian Lynagh
e0596905f9 Merge pull request #17685 from igfoo/igfoo/types
KE2: Don't call buildClassType; once we get into symbol land, stay there
2024-10-09 13:27:28 +01:00
Tamas Vajk
7ff60f8081 Fix extracted child expression 2024-10-09 13:39:07 +02:00
Ian Lynagh
cdf96276c8 KE2: Add a TODO comment 2024-10-09 12:28:51 +01:00
Ian Lynagh
171f68f6d9 Merge pull request #17702 from igfoo/igfoo/dbscheme
KE2: Add more dbscheme comments
2024-10-09 12:13:36 +01:00
Tamas Vajk
a232fcab36 KE2: Extract is and as expression kinds 2024-10-09 09:40:24 +02:00
Ian Lynagh
2cb2aabceb Merge pull request #17698 from igfoo/igfoo/labels
KE2: Update github labeler config
2024-10-08 18:23:58 +01:00
Ian Lynagh
135ea99b65 KE2: Add more dbscheme comments 2024-10-08 17:32:23 +01:00
Ian Lynagh
5edf520439 Merge pull request #17695 from igfoo/igfoo/nulltype
KE2: Handle null types (emit errortypes)
2024-10-08 15:54:20 +01:00
Ian Lynagh
174e7f625d Merge pull request #17692 from igfoo/igfoo/unused
KE2: Remove some dead code
2024-10-08 15:54:02 +01:00
Ian Lynagh
141377a038 Merge pull request #17697 from igfoo/igfoo/callDescription
KE2: Fix use of the wrong variable in log output
2024-10-08 15:53:30 +01:00
Ian Lynagh
862293ae3e KE2: Update github labeler config 2024-10-08 15:37:07 +01:00
Ian Lynagh
780fc699fd KE2: Fix use of the wrong variable in log output 2024-10-08 15:33:28 +01:00
Ian Lynagh
565e780285 KE2: Handle null types (emit errortypes) 2024-10-08 15:18:15 +01:00
Ian Lynagh
b61799fc1d KE2: Remove some dead code 2024-10-08 14:26:55 +01:00
Ian Lynagh
5c76b43fa8 KE2: Don't call buildClassType; once we get into symbol land, stay there 2024-10-08 13:14:11 +01:00
Ian Lynagh
135e909d5e KE2: Remove some 'types' code from the 'class' file 2024-10-08 12:56:00 +01:00
Ian Lynagh
15348dc15b Merge pull request #17675 from igfoo/igfoo/comments
KE2: Add some Java dbscheme and library comments
2024-10-08 11:32:33 +01:00
Ian Lynagh
34557203a0 Merge pull request #17677 from igfoo/igfoo/types
KE2: Pull type extraction out as separate from class extraction
2024-10-08 11:32:22 +01:00
Tamás Vajk
7c3fb3262d Merge pull request #17664 from tamasvajk/ke2-extract-some-expr
KE2: Extract some expr/stmt kinds
2024-10-08 10:17:21 +02:00
Ian Lynagh
9ef185ad6f KE2: Fix build 2024-10-07 18:42:41 +01:00
Ian Lynagh
56fc16c9f5 KE2: Pull more type extraction out into Types.kt 2024-10-07 18:40:58 +01:00
Ian Lynagh
cc09d6da5f KE2: Pull type extraction out as separate from class extraction 2024-10-07 18:29:50 +01:00
Ian Lynagh
b003eb16cc KE2: Add some Java dbscheme and library comments 2024-10-07 16:35:46 +01:00
Ian Lynagh
b46be1b71a Merge pull request #17667 from igfoo/igfoo/conc
KE2: Be concurrency-safe (hopefully!) and enable concurrency
2024-10-07 12:04:19 +01:00
Ian Lynagh
3aaeefad92 KE2: Enable 8 threads 2024-10-04 16:20:21 +01:00
Ian Lynagh
fd3ac0b838 KE2: Use a semaphore to avoid more than maxThreads open TRAP files at once 2024-10-04 16:19:51 +01:00
Ian Lynagh
f5033d1e88 KE2: Make the shared stuff threadsafe 2024-10-04 16:11:26 +01:00
Tamas Vajk
aa5fa12b4f Add TODO comment 2024-10-04 16:38:09 +02:00
Tamas Vajk
cc1f1dd473 KE2: Extract some expr/stmt kinds 2024-10-04 13:35:30 +02:00
Tamás Vajk
8711099de2 Merge pull request #17662 from tamasvajk/ke2-expressions-separate
KE2: Move expr/stmt extraction to separate file
2024-10-04 12:48:58 +02:00
Tamas Vajk
bb32ebb304 KE2: Move expr/stmt extraction to separate file 2024-10-04 11:42:42 +02:00
Ian Lynagh
d6189073d6 Merge pull request #17645 from igfoo/igfoo/top
KE2: Refactor the top level a bit
2024-10-03 12:50:42 +01:00
Ian Lynagh
a1c4413563 KE2: Clarify a 2-stage TODO comment 2024-10-03 11:54:39 +01:00
Ian Lynagh
4701bc7aef KE2: Make concurrent extraction possible 2024-10-02 16:42:24 +01:00
Ian Lynagh
5be65ffead KE2: Only call analyze once, on the sourceModule 2024-10-02 16:29:56 +01:00
Ian Lynagh
f63273a531 Merge pull request #17622 from igfoo/igfoo/ke2-comments
Java/Kotlin: Add some dbscheme comments
2024-10-02 16:14:51 +01:00
Ian Lynagh
e0d157277c Java: Improve files/folder qldoc 2024-10-02 14:03:31 +01:00
Ian Lynagh
32be2296e6 Java/Kotlin: Add some dbscheme comments 2024-09-30 13:02:36 +01:00
Ian Lynagh
8196460da3 Merge pull request #17600 from igfoo/igfoo/ke2-constrs
KE2: Add bugfix from KE1's #17599
2024-09-27 12:18:09 +01:00
Ian Lynagh
97b56dbeb9 Merge pull request #17601 from igfoo/igfoo/ke2-owners
KE2: Add CODEOWNERS
2024-09-27 12:17:42 +01:00
Ian Lynagh
980dd04daa KE2: Add CODEOWNERS 2024-09-27 11:27:55 +01:00
Ian Lynagh
e52d3ba68f KE2: Add bugfix from KE1's #17599 2024-09-27 11:26:02 +01:00
Ian Lynagh
93cd6bb2cf Merge pull request #17594 from igfoo/igfoo/nodeclstack
KE2: Remove the declaration stack for now
2024-09-26 15:32:15 +01:00
Ian Lynagh
0c2aedbb55 KE2: Remove the declaration stack for now
Lets see if we still need it in KE2, or if there's a simpler way.
2024-09-26 14:38:35 +01:00
Tamas Vajk
52934ee5db Code quality improvements 2024-09-26 13:13:20 +01:00
Tamas Vajk
154e841de8 Use extension functions to group extractor functionality 2024-09-26 13:13:19 +01:00
Tamas Vajk
40c28f76f2 KE2 WIP: reintroduce source class extraction 2024-09-26 13:13:17 +01:00
Tamas Vajk
5766580037 KE2: WIP: Move function extraction to symbols 2024-09-26 13:13:16 +01:00
Tamas Vajk
c7f8596643 KE2: Format code in IDEA 2024-09-26 13:13:15 +01:00
Tamas Vajk
a794913b9e KE2: Change Kotlin compiler version in IDEA settings 2024-09-26 13:13:14 +01:00
Tamas Vajk
2bc1b46f9e KE2: Add IntelliJ IDEA settings 2024-09-26 13:13:13 +01:00
Tamas Vajk
1ecf685dfd KE2: Tolerate existing KotlinExtractorDbScheme.kt file in build script 2024-09-26 13:13:12 +01:00
Tamas Vajk
6e3e05dc67 KE2: Modify bazel script to include all java files 2024-09-26 13:13:11 +01:00
Ian Lynagh
1dc8f2594d bazel: Add rules_jvm_external dependency 2024-09-26 13:13:10 +01:00
Ian Lynagh
d85a39b781 KE2: Add classpath to analysis context 2024-09-26 13:13:09 +01:00
Ian Lynagh
8df542b2ce KE2: Print diagnostics reported by analysis API
Ultimately they ought to be in the database and/or logs.
2024-09-26 13:13:08 +01:00
Ian Lynagh
a09ed81b00 KE2: Reenable extractExprContext 2024-09-26 13:13:07 +01:00
Ian Lynagh
6ae4d225b1 KE2: Remove some old code 2024-09-26 13:13:06 +01:00
Ian Lynagh
186022e89c KE2: Emit truncated diagnostic info 2024-09-26 13:13:05 +01:00
Ian Lynagh
092290c066 KE2: Add diagnostic counts to the logger state 2024-09-26 13:13:04 +01:00
Ian Lynagh
e2c127b85f KE2: Pull out a LoggerState 2024-09-26 13:13:03 +01:00
Ian Lynagh
3c0ef3de51 KE2: Reenable extractorContextStack, but now it's in the file logger
This allows multiple threads to run on different files with their own stack.
2024-09-26 13:13:02 +01:00
Ian Lynagh
24c545c00b KE2: Use the FileLogger when making a FileTrapWriter 2024-09-26 13:13:01 +01:00
Ian Lynagh
ce45b0e1d7 KE2: TrapWriter: Use the BasicLogger interface
This will allow FileTrapWriters to log via their FileLogger, which means
it will have access to file-specific state
2024-09-26 13:13:00 +01:00
Ian Lynagh
9ce31cc2b9 KE2: Add a BasicLogger interface 2024-09-26 13:12:59 +01:00
Ian Lynagh
2e3addaf98 KE2: Remove redundant value 2024-09-26 13:12:58 +01:00
Ian Lynagh
b53c29152c KE2: Start handling literals 2024-09-26 13:12:56 +01:00
Ian Lynagh
4ac1c83fcf KE2: More return statement extraction 2024-09-26 13:12:55 +01:00
Ian Lynagh
482cf2f0ff KE2: Start extracting return statements 2024-09-26 13:12:54 +01:00
Ian Lynagh
9601b10734 KE2: Towards extracting expressions 2024-09-26 13:12:53 +01:00
Ian Lynagh
d105258363 KE2: Start extracting blocks 2024-09-26 13:12:52 +01:00
Ian Lynagh
35400d80e8 KE2: Start looking at function bodies 2024-09-26 13:12:51 +01:00
Ian Lynagh
16e182f7a8 KE2: Start extracting locations 2024-09-26 13:12:50 +01:00
Ian Lynagh
572b83cb90 KE2: Output something for classes to satisfy the db checks 2024-09-26 13:12:49 +01:00
Ian Lynagh
310f4e3491 KE2: Emit methods 2024-09-26 13:12:48 +01:00
Ian Lynagh
81f879f453 KE2: Start extracting methods 2024-09-26 13:12:47 +01:00
Ian Lynagh
d85f05be0c KE2: Start extracting method return types 2024-09-26 13:12:46 +01:00
Ian Lynagh
581fed8ae9 KE2: More type extraction 2024-09-26 13:12:45 +01:00
Ian Lynagh
dbf82d5225 KE2: Start looking at extracting types 2024-09-26 13:12:44 +01:00
Ian Lynagh
74d2b43bfb KE2: Make analysis info available to the extrator modules 2024-09-26 13:12:43 +01:00
Ian Lynagh
50e139f29c KE2: Implement CODEQL_EXTRACTOR_JAVA_KOTLIN_DUMP 2024-09-26 13:12:42 +01:00
Ian Lynagh
834f2c0dfb KE2: Tweak functino labels slightly 2024-09-26 13:12:41 +01:00
Ian Lynagh
770f2d6949 KE2: Get some kind of function ID written 2024-09-26 13:12:40 +01:00
Ian Lynagh
4e9a1ef925 KE2: Start extracting functions 2024-09-26 13:12:39 +01:00
Ian Lynagh
c98415631f KE2: Start extracting declaration parents 2024-09-26 13:12:38 +01:00
Ian Lynagh
75e78965f0 KE2: Towards parent decls 2024-09-26 13:12:37 +01:00
Ian Lynagh
429daa3f7c KE2: Start extracting declarations 2024-09-26 13:12:35 +01:00
Ian Lynagh
c47660ae70 KE2: Enable the internal-test-exception code 2024-09-26 13:12:34 +01:00
Ian Lynagh
90a73582ee KE2: Extract package info 2024-09-26 13:12:33 +01:00
Ian Lynagh
f9f766c508 KE2: Start turning KotlinUsesExtractor back on 2024-09-26 13:12:32 +01:00
Ian Lynagh
f3d41ba597 KE2: Actually make location labels 2024-09-26 13:12:31 +01:00
Ian Lynagh
0f1f53cc87 KE2: Ensure all log messages at least get written to the log file 2024-09-26 13:12:30 +01:00
Ian Lynagh
92a2b51be0 KE2: Pass the trap writer in to the file extractor 2024-09-26 13:12:29 +01:00
Ian Lynagh
30626ca7e4 KE2: Start getting deeper into KotlinFileExtractor 2024-09-26 13:12:28 +01:00
Ian Lynagh
e46e5e4cd8 KE2: Start on KotlinFileExtractor 2024-09-26 13:12:27 +01:00
Ian Lynagh
0e32446daa KE2: Remove the LighterAST LoC support 2024-09-26 13:12:26 +01:00
Ian Lynagh
f34b140e2f KE2: Extract file meta info 2024-09-26 13:12:25 +01:00
Ian Lynagh
99161bcb1e KE2: Start writing the actual TRAP files 2024-09-26 13:12:24 +01:00
Ian Lynagh
2c20072e88 KE1: Add some exception handling 2024-09-26 13:12:23 +01:00
Ian Lynagh
70926097df KE2: Remove unnecessary imports 2024-09-26 13:12:22 +01:00
Ian Lynagh
8ebd07e655 KE2: Get TrapFileWriter working 2024-09-26 13:12:20 +01:00
Ian Lynagh
9c4aa931d5 KE2: Move the context stack from LoggerBase to Logger
This will let us have different threads with their own contexts that
share a LoggerBase.
2024-09-26 13:12:19 +01:00
Ian Lynagh
6391ed9865 KE2: Towards TrapFileWriter 2024-09-26 13:12:18 +01:00
Ian Lynagh
4886602426 KE2: Pull a TrapFileWriter.kt out of KotlinExtractor.kt 2024-09-26 13:12:17 +01:00
Ian Lynagh
f54ff1176d KE2: Pass the trap directory through 2024-09-26 13:12:16 +01:00
Ian Lynagh
b903f05883 KE2: Populate source directory 2024-09-26 13:12:15 +01:00
Ian Lynagh
155da0b243 KE2: test-kotlin2/library-tests/files now has no consistency failures 2024-09-26 13:12:14 +01:00
Ian Lynagh
6073180e02 KE2: Emit compilation_finished 2024-09-26 13:12:13 +01:00
Ian Lynagh
f2e47fc09e KE2: More logging 2024-09-26 13:12:12 +01:00
Ian Lynagh
f3afedd510 KE2: We now create a Logger 2024-09-26 13:12:11 +01:00
Ian Lynagh
8b11b65292 KE2: Add the compilation properly 2024-09-26 13:12:10 +01:00
Ian Lynagh
0f12ec3a72 KE2: Start actually emitting some TRAP 2024-09-26 13:12:09 +01:00
Ian Lynagh
50c04b44ca KE2: Tweak LogCounter; now renamed to DiagnosticCounter 2024-09-26 13:12:08 +01:00
Ian Lynagh
88c40d52c8 KE2: Build all Kotlin source files 2024-09-26 13:12:07 +01:00
Ian Lynagh
57da1df4bb KE2: Get the test driver working 2024-09-26 13:12:06 +01:00
Ian Lynagh
d442a532ad KE2: Merge KotlinExtractorExtension into KotlinExtractor 2024-09-26 13:12:05 +01:00
Ian Lynagh
ca0ed61147 KE2: Add the top-level eror handling 2024-09-26 13:12:04 +01:00
Ian Lynagh
9a1b3dd2de KE2: Comment out KE1 code 2024-09-26 13:12:03 +01:00
Ian Lynagh
888c9bce44 KE2: Put the main source file into our package 2024-09-26 13:12:02 +01:00
Ian Lynagh
6ce74be717 KE2: Remove the KE1 resources 2024-09-26 13:12:01 +01:00
Ian Lynagh
1cfbc8e86d KE2: Handle multiple files 2024-09-26 13:12:00 +01:00
Ian Lynagh
1bd1789861 KE2: Get source file list from arguments 2024-09-26 13:11:59 +01:00
Paolo Tranquilli
b0a1475c10 KE2: package ke2 executable with wrapper scripts 2024-09-26 13:11:57 +01:00
Ian Lynagh
602ffb0516 KE2: More steps towards something working 2024-09-26 13:11:56 +01:00
Ian Lynagh
1fc01606ec KE2: More steps towards something working 2024-09-26 13:11:55 +01:00
Ian Lynagh
cd7b0e3757 KE2: Uncomment more imports 2024-09-26 13:11:54 +01:00
Paolo Tranquilli
7447474207 KE2: add some third party dependencies as maven artifacts 2024-09-26 13:11:53 +01:00
Paolo Tranquilli
dc51c5fc5b KE2: add bazel BUILD file 2024-09-26 13:11:52 +01:00
Ian Lynagh
5189f17e6f KE2: Remove old build system from the KE2 copy 2024-09-26 13:11:51 +01:00
Ian Lynagh
3c347317e5 KE2: Add trivial build 2024-09-26 13:11:50 +01:00
Ian Lynagh
8322e31148 KE2: Copy Kotlin extractor 1 to start Kotlin extractor 2
Sans deps.
2024-09-26 13:11:49 +01:00
2320 changed files with 64730 additions and 48388 deletions

View File

@@ -1,5 +1,4 @@
{
"image": "mcr.microsoft.com/devcontainers/base:ubuntu-24.04",
"extensions": [
"rust-lang.rust-analyzer",
"bungcip.better-toml",

1
.gitattributes vendored
View File

@@ -50,6 +50,7 @@
*.dll -text
*.pdb -text
/maven_install.json linguist-generated=true
/java/ql/test/stubs/**/*.java linguist-generated=true
/java/ql/test/experimental/stubs/**/*.java linguist-generated=true
/java/kotlin-extractor/deps/*.jar filter=lfs diff=lfs merge=lfs -text

3
.github/labeler.yml vendored
View File

@@ -11,7 +11,7 @@ Go:
- change-notes/**/*go.*
Java:
- any: [ 'java/**/*', '!java/kotlin-extractor/**/*', '!java/ql/test/kotlin/**/*' ]
- any: [ 'java/**/*', '!java/kotlin-extractor/**/*', '!java/kotlin-extractor2/**/*', '!java/ql/test-kotlin*/**/*' ]
- change-notes/**/*java.*
JS:
@@ -20,6 +20,7 @@ JS:
Kotlin:
- java/kotlin-extractor/**/*
- java/kotlin-extractor2/**/*
- java/ql/test-kotlin*/**/*
Python:

View File

@@ -7,11 +7,6 @@ on:
- "rc/*"
- "codeql-cli-*"
pull_request:
paths:
- '**.ql'
- '**.qll'
- '**/qlpack.yml'
- '**.dbscheme'
permissions:
contents: read
@@ -38,9 +33,9 @@ jobs:
# run with --check-only if running in a PR (github.sha != main)
if : ${{ github.event_name == 'pull_request' }}
shell: bash
run: codeql query compile -q -j0 */ql/{src,examples} --keep-going --warnings=error --check-only --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}" --compilation-cache-size=500 --ram=56000
run: codeql query compile -q -j0 */ql/{src,examples} --keep-going --warnings=error --check-only --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}" --compilation-cache-size=500
- name: compile queries - full
# do full compile if running on main - this populates the cache
if : ${{ github.event_name != 'pull_request' }}
shell: bash
run: codeql query compile -q -j0 */ql/{src,examples} --keep-going --warnings=error --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}" --compilation-cache-size=500 --ram=56000
run: codeql query compile -q -j0 */ql/{src,examples} --keep-going --warnings=error --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}" --compilation-cache-size=500

View File

@@ -48,7 +48,7 @@ jobs:
- name: "Build Swift extractor using Bazel"
run: |
bazel clean --expunge
bazel run //swift:install --nouse_action_cache --noremote_accept_cached --noremote_upload_local_results --spawn_strategy=local
bazel run //swift:create-extractor-pack --nouse_action_cache --noremote_accept_cached --noremote_upload_local_results --spawn_strategy=local
bazel shutdown
- name: Perform CodeQL Analysis

View File

@@ -5,10 +5,8 @@ on:
paths:
- "csharp/**"
- "shared/**"
- "misc/bazel/**"
- .github/actions/fetch-codeql/action.yml
- codeql-workspace.yml
- "MODULE.bazel"
branches:
- main
- "rc/*"
@@ -16,11 +14,9 @@ on:
paths:
- "csharp/**"
- "shared/**"
- "misc/bazel/**"
- .github/workflows/csharp-qltest.yml
- .github/actions/fetch-codeql/action.yml
- codeql-workspace.yml
- "MODULE.bazel"
branches:
- main
- "rc/*"

View File

@@ -72,7 +72,7 @@ repos:
- id: rust-codegen
name: Run Rust checked in code generation
files: ^misc/codegen/|^rust/(prefix\.dbscheme|schema/|codegen/|.*/generated/|ql/lib/(rust\.dbscheme$|codeql/rust/elements)|\.generated.list)
files: ^misc/codegen/|^rust/(schema.py$|codegen/|.*/generated/|ql/lib/(rust\.dbscheme$|codeql/rust/elements)|\.generated.list)
language: system
entry: bazel run //rust/codegen -- --quiet
pass_filenames: false

87
.vscode/tasks.json vendored
View File

@@ -38,93 +38,6 @@
"command": "${config:python.pythonPath}",
},
"problemMatcher": []
},
{
"label": "Create query change note",
"type": "process",
"command": "python3",
"args": [
"misc/scripts/create-change-note.py",
"${input:language}",
"src",
"${input:name}",
"${input:categoryQuery}"
],
"presentation": {
"reveal": "never",
"close": true
},
"problemMatcher": []
},
{
"label": "Create library change note",
"type": "process",
"command": "python3",
"args": [
"misc/scripts/create-change-note.py",
"${input:language}",
"lib",
"${input:name}",
"${input:categoryLibrary}"
],
"presentation": {
"reveal": "never",
"close": true
},
"problemMatcher": []
}
],
"inputs": [
{
"type": "pickString",
"id": "language",
"description": "Language",
"options":
[
"go",
"java",
"javascript",
"cpp",
"csharp",
"python",
"ruby",
"rust",
"swift",
]
},
{
"type": "promptString",
"id": "name",
"description": "Short name (kebab-case)"
},
{
"type": "pickString",
"id": "categoryQuery",
"description": "Category (query change)",
"options":
[
"breaking",
"deprecated",
"newQuery",
"queryMetadata",
"majorAnalysis",
"minorAnalysis",
"fix",
]
},
{
"type": "pickString",
"id": "categoryLibrary",
"description": "Category (library change)",
"options":
[
"breaking",
"deprecated",
"feature",
"majorAnalysis",
"minorAnalysis",
"fix",
]
}
]
}

View File

@@ -1,4 +0,0 @@
---
category: majorAnalysis
---
* Added support for TypeScript 5.7.

View File

@@ -10,6 +10,7 @@
/swift/ @github/codeql-swift
/misc/codegen/ @github/codeql-swift
/java/kotlin-extractor/ @github/codeql-kotlin
/java/kotlin-extractor2/ @github/codeql-kotlin
/java/ql/test-kotlin1/ @github/codeql-kotlin
/java/ql/test-kotlin2/ @github/codeql-kotlin
@@ -42,6 +43,3 @@ MODULE.bazel @github/codeql-ci-reviewers
# Misc
/misc/scripts/accept-expected-changes-from-ci.py @RasmusWL
/misc/scripts/generate-code-scanning-query-list.py @RasmusWL
# .devcontainer
/.devcontainer/ @github/codeql-ci-reviewers

13
Cargo.lock generated
View File

@@ -381,10 +381,8 @@ version = "0.1.0"
dependencies = [
"anyhow",
"argfile",
"chrono",
"clap",
"codeql-extractor",
"dunce",
"figment",
"glob",
"itertools 0.13.0",
@@ -406,7 +404,6 @@ dependencies = [
"ra_ap_vfs",
"rust-extractor-macros",
"serde",
"serde_json",
"serde_with",
"stderrlog",
"triomphe",
@@ -543,12 +540,6 @@ version = "0.1.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9bda8e21c04aca2ae33ffc2fd8c23134f3cac46db123ba97bd9d3f3b8a4a85e1"
[[package]]
name = "dunce"
version = "1.0.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "92773504d58c093f6de2459af4af33faa518c13451eb8f2b5698ed3d36e7c813"
[[package]]
name = "either"
version = "1.13.0"
@@ -2043,9 +2034,9 @@ dependencies = [
[[package]]
name = "serde_json"
version = "1.0.133"
version = "1.0.132"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "c7fceb2473b9166b2294ef05efcb65a3db80803f0b03ef86a5fc88a2b85ee377"
checksum = "d726bfaff4b320266d395898905d0eba0345aae23b54aee3a737e260fd46db03"
dependencies = [
"itoa",
"memchr",

View File

@@ -28,6 +28,7 @@ bazel_dep(name = "gazelle", version = "0.38.0")
bazel_dep(name = "rules_dotnet", version = "0.17.4")
bazel_dep(name = "googletest", version = "1.14.0.bcr.1")
bazel_dep(name = "rules_rust", version = "0.52.2")
bazel_dep(name = "rules_jvm_external", version = "6.2")
bazel_dep(name = "buildifier_prebuilt", version = "6.4.0", dev_dependency = True)
@@ -68,7 +69,7 @@ use_repo(py_deps, "vendor__anyhow-1.0.44", "vendor__cc-1.0.70", "vendor__clap-2.
# deps for ruby+rust
# keep in sync by running `misc/bazel/3rdparty/update_cargo_deps.sh`
tree_sitter_extractors_deps = use_extension("//misc/bazel/3rdparty:tree_sitter_extractors_extension.bzl", "r")
use_repo(tree_sitter_extractors_deps, "vendor__anyhow-1.0.93", "vendor__argfile-0.2.1", "vendor__chrono-0.4.38", "vendor__clap-4.5.20", "vendor__dunce-1.0.5", "vendor__encoding-0.2.33", "vendor__figment-0.10.19", "vendor__flate2-1.0.34", "vendor__glob-0.3.1", "vendor__globset-0.4.15", "vendor__itertools-0.10.5", "vendor__itertools-0.13.0", "vendor__lazy_static-1.5.0", "vendor__log-0.4.22", "vendor__num-traits-0.2.19", "vendor__num_cpus-1.16.0", "vendor__proc-macro2-1.0.89", "vendor__quote-1.0.37", "vendor__ra_ap_base_db-0.0.232", "vendor__ra_ap_cfg-0.0.232", "vendor__ra_ap_hir-0.0.232", "vendor__ra_ap_hir_def-0.0.232", "vendor__ra_ap_hir_expand-0.0.232", "vendor__ra_ap_ide_db-0.0.232", "vendor__ra_ap_intern-0.0.232", "vendor__ra_ap_load-cargo-0.0.232", "vendor__ra_ap_parser-0.0.232", "vendor__ra_ap_paths-0.0.232", "vendor__ra_ap_project_model-0.0.232", "vendor__ra_ap_span-0.0.232", "vendor__ra_ap_syntax-0.0.232", "vendor__ra_ap_vfs-0.0.232", "vendor__rand-0.8.5", "vendor__rayon-1.10.0", "vendor__regex-1.11.1", "vendor__serde-1.0.214", "vendor__serde_json-1.0.133", "vendor__serde_with-3.11.0", "vendor__stderrlog-0.6.0", "vendor__syn-2.0.87", "vendor__tracing-0.1.40", "vendor__tracing-subscriber-0.3.18", "vendor__tree-sitter-0.24.4", "vendor__tree-sitter-embedded-template-0.23.2", "vendor__tree-sitter-json-0.24.8", "vendor__tree-sitter-ql-0.23.1", "vendor__tree-sitter-ruby-0.23.1", "vendor__triomphe-0.1.14", "vendor__ungrammar-1.16.1")
use_repo(tree_sitter_extractors_deps, "vendor__anyhow-1.0.93", "vendor__argfile-0.2.1", "vendor__chrono-0.4.38", "vendor__clap-4.5.20", "vendor__encoding-0.2.33", "vendor__figment-0.10.19", "vendor__flate2-1.0.34", "vendor__glob-0.3.1", "vendor__globset-0.4.15", "vendor__itertools-0.10.5", "vendor__itertools-0.13.0", "vendor__lazy_static-1.5.0", "vendor__log-0.4.22", "vendor__num-traits-0.2.19", "vendor__num_cpus-1.16.0", "vendor__proc-macro2-1.0.89", "vendor__quote-1.0.37", "vendor__ra_ap_base_db-0.0.232", "vendor__ra_ap_cfg-0.0.232", "vendor__ra_ap_hir-0.0.232", "vendor__ra_ap_hir_def-0.0.232", "vendor__ra_ap_hir_expand-0.0.232", "vendor__ra_ap_ide_db-0.0.232", "vendor__ra_ap_intern-0.0.232", "vendor__ra_ap_load-cargo-0.0.232", "vendor__ra_ap_parser-0.0.232", "vendor__ra_ap_paths-0.0.232", "vendor__ra_ap_project_model-0.0.232", "vendor__ra_ap_span-0.0.232", "vendor__ra_ap_syntax-0.0.232", "vendor__ra_ap_vfs-0.0.232", "vendor__rand-0.8.5", "vendor__rayon-1.10.0", "vendor__regex-1.11.1", "vendor__serde-1.0.214", "vendor__serde_json-1.0.132", "vendor__serde_with-3.11.0", "vendor__stderrlog-0.6.0", "vendor__syn-2.0.87", "vendor__tracing-0.1.40", "vendor__tracing-subscriber-0.3.18", "vendor__tree-sitter-0.24.4", "vendor__tree-sitter-embedded-template-0.23.2", "vendor__tree-sitter-json-0.24.8", "vendor__tree-sitter-ql-0.23.1", "vendor__tree-sitter-ruby-0.23.1", "vendor__triomphe-0.1.14", "vendor__ungrammar-1.16.1")
dotnet = use_extension("@rules_dotnet//dotnet:extensions.bzl", "dotnet")
dotnet.toolchain(dotnet_version = "9.0.100")
@@ -162,6 +163,35 @@ use_repo(
"kotlin-stdlib-2.1.0-Beta1",
)
maven = use_extension("@rules_jvm_external//:extensions.bzl", "maven")
# run
# REPIN=1 bazel run @maven_deps//:pin
# from this directory after modifying the following to update maven_install.json
maven.install(
name = "maven_deps",
# The Caffeine version needs to match https://github.com/JetBrains/kotlin/blob/master/gradle/libs.versions.toml
# See also https://youtrack.jetbrains.com/issue/KT-73751/Analysis-API-Caffeine-dependency which seeks a better
# way of including the needed dependency.
artifacts = [
"org.jetbrains.kotlin:%s:2.1.0" % kotlin_lib
for kotlin_lib in ("kotlin-annotation-processing", "kotlin-compiler")
] + [ "com.github.ben-manes.caffeine:caffeine:2.9.3" ] ,
lock_file = "//:maven_install.json",
repositories = [
"https://repo1.maven.org/maven2",
# some of these URLs might be needed at some point
# "https://maven.pkg.jetbrains.space/kotlin/p/kotlin/bootstrap",
# "https://maven.pkg.jetbrains.space/kotlin/p/kotlin/kotlin-ide-plugin-dependencies",
# "https://www.jetbrains.com/intellij-repository/releases",
# "https://cache-redirector.jetbrains.com/intellij-third-party-dependencies",
],
)
use_repo(
maven,
"maven_deps",
)
go_sdk = use_extension("@rules_go//go:extensions.bzl", "go_sdk")
go_sdk.download(version = "1.23.1")

View File

@@ -2,8 +2,19 @@ load("//misc/bazel:pkg.bzl", "codeql_pack")
package(default_visibility = ["//visibility:public"])
codeql_pack(
name = "actions",
srcs = ["//actions/extractor"],
experimental = True,
)
[
codeql_pack(
name = "-".join(parts),
srcs = [
"//actions/extractor",
],
pack_prefix = "/".join(parts),
)
for parts in (
[
"experimental",
"actions",
],
["actions"],
)
]

View File

@@ -1,4 +1,58 @@
{
"DataFlow Java/C++/C#/Go/Python/Ruby/Swift Legacy Configuration": [
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl1.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl2.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl3.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl4.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl5.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl6.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl1.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl2.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl3.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl4.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImplLocal.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl1.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl2.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl3.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl4.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl1.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl2.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl3.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl4.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl5.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl1.qll",
"go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl2.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl1.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl2.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl3.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl4.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImpl1.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImpl2.qll",
"swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImpl1.qll"
],
"TaintTracking Legacy Configuration Java/C++/C#/Go/Python/Ruby/Swift": [
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
"cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking2/TaintTrackingImpl.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking2/TaintTrackingImpl.qll",
"cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking3/TaintTrackingImpl.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/tainttracking2/TaintTrackingImpl.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/tainttracking3/TaintTrackingImpl.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/tainttracking4/TaintTrackingImpl.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/tainttracking5/TaintTrackingImpl.qll",
"go/ql/lib/semmle/go/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
"go/ql/lib/semmle/go/dataflow/internal/tainttracking2/TaintTrackingImpl.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/tainttracking2/TaintTrackingImpl.qll",
"java/ql/lib/semmle/code/java/dataflow/internal/tainttracking3/TaintTrackingImpl.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/tainttracking1/TaintTrackingImpl.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/tainttracking2/TaintTrackingImpl.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/tainttracking3/TaintTrackingImpl.qll",
"python/ql/lib/semmle/python/dataflow/new/internal/tainttracking4/TaintTrackingImpl.qll",
"ruby/ql/lib/codeql/ruby/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
"swift/ql/lib/codeql/swift/dataflow/internal/tainttracking1/TaintTrackingImpl.qll"
],
"SsaReadPosition Java/C#": [
"java/ql/lib/semmle/code/java/dataflow/internal/rangeanalysis/SsaReadPositionCommon.qll",
"csharp/ql/lib/semmle/code/csharp/dataflow/internal/rangeanalysis/SsaReadPositionCommon.qll"

View File

@@ -1,3 +0,0 @@
description: Implement compilation_build_mode/2
compatibility: full
compilation_build_mode.rel: delete

View File

@@ -1,17 +1,3 @@
## 3.0.0
### Breaking Changes
* Deleted the old deprecated data flow API that was based on extending a configuration class. See https://github.blog/changelog/2023-08-14-new-dataflow-api-for-writing-custom-codeql-queries for instructions on migrating your queries to use the new API.
### Deprecated APIs
* The `NonThrowing` class (`semmle.code.cpp.models.interfaces.NonThrowing`) has been deprecated. Please use the `NonCppThrowingFunction` class instead.
## 2.1.1
No user-facing changes.
## 2.1.0
### New Features

View File

@@ -0,0 +1,4 @@
---
category: minorAnalysis
---
* The `Guards` library (`semmle.code.cpp.controlflow.Guards`) has been improved to recognize more guard conditions. Additionally, the guards library no longer considers guards in static local initializers or global initializers as `GuardCondition`s.

View File

@@ -1,3 +0,0 @@
## 2.1.1
No user-facing changes.

View File

@@ -1,9 +0,0 @@
## 3.0.0
### Breaking Changes
* Deleted the old deprecated data flow API that was based on extending a configuration class. See https://github.blog/changelog/2023-08-14-new-dataflow-api-for-writing-custom-codeql-queries for instructions on migrating your queries to use the new API.
### Deprecated APIs
* The `NonThrowing` class (`semmle.code.cpp.models.interfaces.NonThrowing`) has been deprecated. Please use the `NonCppThrowingFunction` class instead.

View File

@@ -1,2 +1,2 @@
---
lastReleaseVersion: 3.0.0
lastReleaseVersion: 2.1.0

View File

@@ -1,5 +1,5 @@
name: codeql/cpp-all
version: 3.0.0
version: 2.1.1-dev
groups: cpp
dbscheme: semmlecode.cpp.dbscheme
extractor: cpp

View File

@@ -112,7 +112,4 @@ class Compilation extends @compilation {
* termination, but crashing due to something like a segfault is not.
*/
predicate normalTermination() { compilation_finished(this, _, _) }
/** Holds if this compilation was compiled using the "none" build mode. */
predicate buildModeNone() { compilation_build_mode(this, 0) }
}

View File

@@ -5,6 +5,7 @@
import cpp
import semmle.code.cpp.ir.IR
private import semmle.code.cpp.ir.ValueNumbering
private import semmle.code.cpp.ir.implementation.raw.internal.TranslatedExpr
private import semmle.code.cpp.ir.implementation.raw.internal.InstructionTag
@@ -59,7 +60,91 @@ class MatchValue extends AbstractValue, TMatchValue {
}
/**
* A Boolean condition in the AST that guards one or more basic blocks.
* A value number such that at least one of the instructions is
* a `CompareInstruction`.
*/
private class CompareValueNumber extends ValueNumber {
CompareInstruction cmp;
CompareValueNumber() { cmp = this.getAnInstruction() }
/** Gets a `CompareInstruction` belonging to this value number. */
CompareInstruction getCompareInstruction() { result = cmp }
/**
* Gets the left and right operands of a `CompareInstruction` that
* belong to this value number.
*/
predicate hasOperands(Operand left, Operand right) {
left = cmp.getLeftOperand() and
right = cmp.getRightOperand()
}
}
private class CompareEQValueNumber extends CompareValueNumber {
override CompareEQInstruction cmp;
}
private class CompareNEValueNumber extends CompareValueNumber {
override CompareNEInstruction cmp;
}
private class CompareLTValueNumber extends CompareValueNumber {
override CompareLTInstruction cmp;
}
private class CompareGTValueNumber extends CompareValueNumber {
override CompareGTInstruction cmp;
}
private class CompareLEValueNumber extends CompareValueNumber {
override CompareLEInstruction cmp;
}
private class CompareGEValueNumber extends CompareValueNumber {
override CompareGEInstruction cmp;
}
/**
* A value number such that at least one of the instructions provides
* the integer value controlling a `SwitchInstruction`.
*/
private class SwitchConditionValueNumber extends ValueNumber {
SwitchInstruction switch;
pragma[nomagic]
SwitchConditionValueNumber() { this.getAnInstruction() = switch.getExpression() }
/** Gets an expression that belongs to this value number. */
Operand getExpressionOperand() { result = switch.getExpressionOperand() }
Instruction getSuccessor(CaseEdge kind) { result = switch.getSuccessor(kind) }
}
private class BuiltinExpectCallValueNumber extends ValueNumber {
BuiltinExpectCallInstruction instr;
BuiltinExpectCallValueNumber() { this.getAnInstruction() = instr }
ValueNumber getCondition() { result.getAnInstruction() = instr.getCondition() }
Operand getAUse() { result = instr.getAUse() }
}
private class LogicalNotValueNumber extends ValueNumber {
LogicalNotInstruction instr;
LogicalNotValueNumber() { this.getAnInstruction() = instr }
ValueNumber getUnary() { result.getAnInstruction() = instr.getUnary() }
}
/**
* A Boolean condition in the AST that guards one or more basic blocks. This includes
* operands of logical operators but not switch statements.
*
* For performance reasons conditions inside static local initializers or
* global initializers are not considered `GuardCondition`s.
*/
cached
class GuardCondition extends Expr {
@@ -369,6 +454,9 @@ private predicate nonExcludedIRAndBasicBlock(IRBlock irb, BasicBlock controlled)
*
* Note that `&&` and `||` don't have an explicit representation in the IR,
* and therefore will not appear as IRGuardConditions.
*
* For performance reasons conditions inside static local initializers or
* global initializers are not considered `IRGuardCondition`s.
*/
cached
class IRGuardCondition extends Instruction {
@@ -517,7 +605,7 @@ class IRGuardCondition extends Instruction {
cached
predicate comparesLt(Operand left, Operand right, int k, boolean isLessThan, boolean testIsTrue) {
exists(BooleanValue value |
compares_lt(this, left, right, k, isLessThan, value) and
compares_lt(valueNumber(this), left, right, k, isLessThan, value) and
value.getValue() = testIsTrue
)
}
@@ -528,7 +616,7 @@ class IRGuardCondition extends Instruction {
*/
cached
predicate comparesLt(Operand op, int k, boolean isLessThan, AbstractValue value) {
compares_lt(this, op, k, isLessThan, value)
compares_lt(valueNumber(this), op, k, isLessThan, value)
}
/**
@@ -538,7 +626,8 @@ class IRGuardCondition extends Instruction {
cached
predicate ensuresLt(Operand left, Operand right, int k, IRBlock block, boolean isLessThan) {
exists(AbstractValue value |
compares_lt(this, left, right, k, isLessThan, value) and this.valueControls(block, value)
compares_lt(valueNumber(this), left, right, k, isLessThan, value) and
this.valueControls(block, value)
)
}
@@ -549,7 +638,8 @@ class IRGuardCondition extends Instruction {
cached
predicate ensuresLt(Operand op, int k, IRBlock block, boolean isLessThan) {
exists(AbstractValue value |
compares_lt(this, op, k, isLessThan, value) and this.valueControls(block, value)
compares_lt(valueNumber(this), op, k, isLessThan, value) and
this.valueControls(block, value)
)
}
@@ -562,7 +652,7 @@ class IRGuardCondition extends Instruction {
Operand left, Operand right, int k, IRBlock pred, IRBlock succ, boolean isLessThan
) {
exists(AbstractValue value |
compares_lt(this, left, right, k, isLessThan, value) and
compares_lt(valueNumber(this), left, right, k, isLessThan, value) and
this.valueControlsEdge(pred, succ, value)
)
}
@@ -574,7 +664,7 @@ class IRGuardCondition extends Instruction {
cached
predicate ensuresLtEdge(Operand left, int k, IRBlock pred, IRBlock succ, boolean isLessThan) {
exists(AbstractValue value |
compares_lt(this, left, k, isLessThan, value) and
compares_lt(valueNumber(this), left, k, isLessThan, value) and
this.valueControlsEdge(pred, succ, value)
)
}
@@ -583,7 +673,7 @@ class IRGuardCondition extends Instruction {
cached
predicate comparesEq(Operand left, Operand right, int k, boolean areEqual, boolean testIsTrue) {
exists(BooleanValue value |
compares_eq(this, left, right, k, areEqual, value) and
compares_eq(valueNumber(this), left, right, k, areEqual, value) and
value.getValue() = testIsTrue
)
}
@@ -591,7 +681,7 @@ class IRGuardCondition extends Instruction {
/** Holds if (determined by this guard) `op == k` evaluates to `areEqual` if this expression evaluates to `value`. */
cached
predicate comparesEq(Operand op, int k, boolean areEqual, AbstractValue value) {
unary_compares_eq(this, op, k, areEqual, false, value)
unary_compares_eq(valueNumber(this), op, k, areEqual, false, value)
}
/**
@@ -601,7 +691,8 @@ class IRGuardCondition extends Instruction {
cached
predicate ensuresEq(Operand left, Operand right, int k, IRBlock block, boolean areEqual) {
exists(AbstractValue value |
compares_eq(this, left, right, k, areEqual, value) and this.valueControls(block, value)
compares_eq(valueNumber(this), left, right, k, areEqual, value) and
this.valueControls(block, value)
)
}
@@ -612,7 +703,8 @@ class IRGuardCondition extends Instruction {
cached
predicate ensuresEq(Operand op, int k, IRBlock block, boolean areEqual) {
exists(AbstractValue value |
unary_compares_eq(this, op, k, areEqual, false, value) and this.valueControls(block, value)
unary_compares_eq(valueNumber(this), op, k, areEqual, false, value) and
this.valueControls(block, value)
)
}
@@ -625,7 +717,7 @@ class IRGuardCondition extends Instruction {
Operand left, Operand right, int k, IRBlock pred, IRBlock succ, boolean areEqual
) {
exists(AbstractValue value |
compares_eq(this, left, right, k, areEqual, value) and
compares_eq(valueNumber(this), left, right, k, areEqual, value) and
this.valueControlsEdge(pred, succ, value)
)
}
@@ -637,7 +729,7 @@ class IRGuardCondition extends Instruction {
cached
predicate ensuresEqEdge(Operand op, int k, IRBlock pred, IRBlock succ, boolean areEqual) {
exists(AbstractValue value |
unary_compares_eq(this, op, k, areEqual, false, value) and
unary_compares_eq(valueNumber(this), op, k, areEqual, false, value) and
this.valueControlsEdge(pred, succ, value)
)
}
@@ -725,13 +817,20 @@ class IRGuardCondition extends Instruction {
}
private Instruction getBranchForCondition(Instruction guard) {
result.(ConditionalBranchInstruction).getCondition() = guard
// There are a lot of guards inside global or static local initializers,
// and on certain databases this can make the `ensures*` predicates
// blow up.
// These guards are likely not super important anyway.
guard.getEnclosingFunction() instanceof Function and
(
result.(ConditionalBranchInstruction).getCondition() = guard
or
result.(SwitchInstruction).getExpression() = guard
)
or
exists(LogicalNotInstruction cond |
result = getBranchForCondition(cond) and cond.getUnary() = guard
)
or
result.(SwitchInstruction).getExpression() = guard
}
/**
@@ -740,7 +839,7 @@ private Instruction getBranchForCondition(Instruction guard) {
* Beware making mistaken logical implications here relating `areEqual` and `testIsTrue`.
*/
private predicate compares_eq(
Instruction test, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
ValueNumber test, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
) {
/* The simple case where the test *is* the comparison so areEqual = testIsTrue xor eq. */
exists(AbstractValue v | simple_comparison_eq(test, left, right, k, v) |
@@ -759,10 +858,10 @@ private predicate compares_eq(
or
/* (x is true => (left == right + k)) => (!x is false => (left == right + k)) */
exists(AbstractValue dual | value = dual.getDualValue() |
compares_eq(test.(LogicalNotInstruction).getUnary(), left, right, k, areEqual, dual)
compares_eq(test.(LogicalNotValueNumber).getUnary(), left, right, k, areEqual, dual)
)
or
compares_eq(test.(BuiltinExpectCallInstruction).getCondition(), left, right, k, areEqual, value)
compares_eq(test.(BuiltinExpectCallValueNumber).getCondition(), left, right, k, areEqual, value)
}
/**
@@ -801,12 +900,10 @@ private predicate compares_eq(
* latter.
*/
private predicate unary_compares_eq(
Instruction test, Operand op, int k, boolean areEqual, boolean inNonZeroCase, AbstractValue value
ValueNumber test, Operand op, int k, boolean areEqual, boolean inNonZeroCase, AbstractValue value
) {
/* The simple case where the test *is* the comparison so areEqual = testIsTrue xor eq. */
exists(AbstractValue v |
unary_simple_comparison_eq(test, k, inNonZeroCase, v) and op.getDef() = test
|
exists(AbstractValue v | unary_simple_comparison_eq(test, op, k, inNonZeroCase, v) |
areEqual = true and value = v
or
areEqual = false and value = v.getDualValue()
@@ -817,7 +914,7 @@ private predicate unary_compares_eq(
/* (x is true => (op == k)) => (!x is false => (op == k)) */
exists(AbstractValue dual, boolean inNonZeroCase0 |
value = dual.getDualValue() and
unary_compares_eq(test.(LogicalNotInstruction).getUnary(), op, k, inNonZeroCase0, areEqual, dual)
unary_compares_eq(test.(LogicalNotValueNumber).getUnary(), op, k, inNonZeroCase0, areEqual, dual)
|
k = 0 and inNonZeroCase = inNonZeroCase0
or
@@ -827,82 +924,95 @@ private predicate unary_compares_eq(
// ((test is `areEqual` => op == const + k2) and const == `k1`) =>
// test is `areEqual` => op == k1 + k2
inNonZeroCase = false and
exists(int k1, int k2, ConstantInstruction const |
exists(int k1, int k2, Instruction const |
compares_eq(test, op, const.getAUse(), k2, areEqual, value) and
int_value(const) = k1 and
k = k1 + k2
)
or
unary_compares_eq(test.(BuiltinExpectCallInstruction).getCondition(), op, k, areEqual,
unary_compares_eq(test.(BuiltinExpectCallValueNumber).getCondition(), op, k, areEqual,
inNonZeroCase, value)
}
/** Rearrange various simple comparisons into `left == right + k` form. */
private predicate simple_comparison_eq(
CompareInstruction cmp, Operand left, Operand right, int k, AbstractValue value
CompareValueNumber cmp, Operand left, Operand right, int k, AbstractValue value
) {
left = cmp.getLeftOperand() and
cmp instanceof CompareEQInstruction and
right = cmp.getRightOperand() and
cmp instanceof CompareEQValueNumber and
cmp.hasOperands(left, right) and
k = 0 and
value.(BooleanValue).getValue() = true
or
left = cmp.getLeftOperand() and
cmp instanceof CompareNEInstruction and
right = cmp.getRightOperand() and
cmp instanceof CompareNEValueNumber and
cmp.hasOperands(left, right) and
k = 0 and
value.(BooleanValue).getValue() = false
}
/**
* Rearrange various simple comparisons into `op == k` form.
* Holds if `op` is an operand that is eventually used in a unary comparison
* with a constant.
*/
private predicate isRelevantUnaryComparisonOperand(Operand op) {
// Base case: `op` is an operand of a `CompareEQInstruction` or `CompareNEInstruction`,
// and the other operand is a constant.
exists(CompareInstruction eq, Instruction instr |
eq.hasOperands(op, instr.getAUse()) and
exists(int_value(instr))
|
eq instanceof CompareEQInstruction
or
eq instanceof CompareNEInstruction
)
or
// C doesn't have int-to-bool conversions, so `if(x)` will just generate:
// r2_1(glval<int>) = VariableAddress[x]
// r2_2(int) = Load[x] : &:r2_1, m1_6
// v2_3(void) = ConditionalBranch : r2_2
exists(ConditionalBranchInstruction branch | branch.getConditionOperand() = op)
or
// If `!x` is a relevant unary comparison then so is `x`.
exists(LogicalNotInstruction logicalNot |
isRelevantUnaryComparisonOperand(unique( | | logicalNot.getAUse())) and
logicalNot.getUnaryOperand() = op
)
or
// If `y` is a relevant unary comparison and `y = x` then so is `x`.
not op.isDefinitionInexact() and
exists(CopyInstruction copy |
isRelevantUnaryComparisonOperand(unique( | | copy.getAUse())) and
op = copy.getSourceValueOperand()
)
or
// If phi(x1, x2) is a relevant unary comparison then so are `x1` and `x2`.
not op.isDefinitionInexact() and
exists(PhiInstruction phi |
isRelevantUnaryComparisonOperand(unique( | | phi.getAUse())) and
op = phi.getAnInputOperand()
)
or
// If `__builtin_expect(x)` is a relevant unary comparison then so is `x`.
exists(BuiltinExpectCallInstruction call |
isRelevantUnaryComparisonOperand(unique( | | call.getAUse())) and
op = call.getConditionOperand()
)
}
/** Rearrange various simple comparisons into `op == k` form. */
private predicate unary_simple_comparison_eq(
Instruction test, int k, boolean inNonZeroCase, AbstractValue value
ValueNumber test, Operand op, int k, boolean inNonZeroCase, AbstractValue value
) {
exists(SwitchInstruction switch, CaseEdge case |
test = switch.getExpression() and
exists(CaseEdge case, SwitchConditionValueNumber condition |
condition = test and
op = condition.getExpressionOperand() and
case = value.(MatchValue).getCase() and
exists(switch.getSuccessor(case)) and
exists(condition.getSuccessor(case)) and
case.getValue().toInt() = k and
inNonZeroCase = false
)
or
// Any instruction with an integral type could potentially be part of a
// check for nullness when used in a guard. So we include all integral
// typed instructions here. However, since some of these instructions are
// already included as guards in other cases, we exclude those here.
// These are instructions that compute a binary equality or inequality
// relation. For example, the following:
// ```cpp
// if(a == b + 42) { ... }
// ```
// generates the following IR:
// ```
// r1(glval<int>) = VariableAddress[a] :
// r2(int) = Load[a] : &:r1, m1
// r3(glval<int>) = VariableAddress[b] :
// r4(int) = Load[b] : &:r3, m2
// r5(int) = Constant[42] :
// r6(int) = Add : r4, r5
// r7(bool) = CompareEQ : r2, r6
// v1(void) = ConditionalBranch : r7
// ```
// and since `r7` is an integral typed instruction this predicate could
// include a case for when `r7` evaluates to true (in which case we would
// infer that `r6` was non-zero, and a case for when `r7` evaluates to false
// (in which case we would infer that `r6` was zero).
// However, since `a == b + 42` is already supported when reasoning about
// binary equalities we exclude those cases here.
not test.isGLValue() and
not simple_comparison_eq(test, _, _, _, _) and
not simple_comparison_lt(test, _, _, _) and
not test = any(SwitchInstruction switch).getExpression() and
(
test.getResultIRType() instanceof IRAddressType or
test.getResultIRType() instanceof IRIntegerType or
test.getResultIRType() instanceof IRBooleanType
) and
isRelevantUnaryComparisonOperand(op) and
op.getDef() = test.getAnInstruction() and
(
k = 1 and
value.(BooleanValue).getValue() = true and
@@ -919,10 +1029,12 @@ private class BuiltinExpectCallInstruction extends CallInstruction {
BuiltinExpectCallInstruction() { this.getStaticCallTarget().hasName("__builtin_expect") }
/** Gets the condition of this call. */
Instruction getCondition() {
Instruction getCondition() { result = this.getConditionOperand().getDef() }
Operand getConditionOperand() {
// The first parameter of `__builtin_expect` has type `long`. So we skip
// the conversion when inferring guards.
result = this.getArgument(0).(ConvertInstruction).getUnary()
result = this.getArgument(0).(ConvertInstruction).getUnaryOperand()
}
}
@@ -932,23 +1044,23 @@ private class BuiltinExpectCallInstruction extends CallInstruction {
* `__builtin_expect(left == right + k, _)` to `0`.
*/
private predicate builtin_expect_eq(
CompareInstruction cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
CompareValueNumber cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
) {
exists(BuiltinExpectCallInstruction call, Instruction const, AbstractValue innerValue |
exists(BuiltinExpectCallValueNumber call, Instruction const, AbstractValue innerValue |
int_value(const) = 0 and
cmp.hasOperands(call.getAUse(), const.getAUse()) and
compares_eq(call.getCondition(), left, right, k, areEqual, innerValue)
|
cmp instanceof CompareNEInstruction and
cmp instanceof CompareNEValueNumber and
value = innerValue
or
cmp instanceof CompareEQInstruction and
cmp instanceof CompareEQValueNumber and
value.getDualValue() = innerValue
)
}
private predicate complex_eq(
CompareInstruction cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
ValueNumber cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
) {
sub_eq(cmp, left, right, k, areEqual, value)
or
@@ -962,24 +1074,24 @@ private predicate complex_eq(
* an instruction that compares the value of `__builtin_expect(op == k, _)` to `0`.
*/
private predicate unary_builtin_expect_eq(
CompareInstruction cmp, Operand op, int k, boolean areEqual, boolean inNonZeroCase,
CompareValueNumber cmp, Operand op, int k, boolean areEqual, boolean inNonZeroCase,
AbstractValue value
) {
exists(BuiltinExpectCallInstruction call, Instruction const, AbstractValue innerValue |
exists(BuiltinExpectCallValueNumber call, Instruction const, AbstractValue innerValue |
int_value(const) = 0 and
cmp.hasOperands(call.getAUse(), const.getAUse()) and
unary_compares_eq(call.getCondition(), op, k, areEqual, inNonZeroCase, innerValue)
|
cmp instanceof CompareNEInstruction and
cmp instanceof CompareNEValueNumber and
value = innerValue
or
cmp instanceof CompareEQInstruction and
cmp instanceof CompareEQValueNumber and
value.getDualValue() = innerValue
)
}
private predicate unary_complex_eq(
Instruction test, Operand op, int k, boolean areEqual, boolean inNonZeroCase, AbstractValue value
ValueNumber test, Operand op, int k, boolean areEqual, boolean inNonZeroCase, AbstractValue value
) {
unary_sub_eq(test, op, k, areEqual, inNonZeroCase, value)
or
@@ -995,7 +1107,7 @@ private predicate unary_complex_eq(
/** Holds if `left < right + k` evaluates to `isLt` given that test is `testIsTrue`. */
private predicate compares_lt(
Instruction test, Operand left, Operand right, int k, boolean isLt, AbstractValue value
ValueNumber test, Operand left, Operand right, int k, boolean isLt, AbstractValue value
) {
/* In the simple case, the test is the comparison, so isLt = testIsTrue */
simple_comparison_lt(test, left, right, k) and
@@ -1008,23 +1120,22 @@ private predicate compares_lt(
or
/* (x is true => (left < right + k)) => (!x is false => (left < right + k)) */
exists(AbstractValue dual | value = dual.getDualValue() |
compares_lt(test.(LogicalNotInstruction).getUnary(), left, right, k, isLt, dual)
compares_lt(test.(LogicalNotValueNumber).getUnary(), left, right, k, isLt, dual)
)
}
/** Holds if `op < k` evaluates to `isLt` given that `test` evaluates to `value`. */
private predicate compares_lt(Instruction test, Operand op, int k, boolean isLt, AbstractValue value) {
unary_simple_comparison_lt(test, k, isLt, value) and
op.getDef() = test
private predicate compares_lt(ValueNumber test, Operand op, int k, boolean isLt, AbstractValue value) {
unary_simple_comparison_lt(test, op, k, isLt, value)
or
complex_lt(test, op, k, isLt, value)
or
/* (x is true => (op < k)) => (!x is false => (op < k)) */
exists(AbstractValue dual | value = dual.getDualValue() |
compares_lt(test.(LogicalNotInstruction).getUnary(), op, k, isLt, dual)
compares_lt(test.(LogicalNotValueNumber).getUnary(), op, k, isLt, dual)
)
or
exists(int k1, int k2, ConstantInstruction const |
exists(int k1, int k2, Instruction const |
compares_lt(test, op, const.getAUse(), k2, isLt, value) and
int_value(const) = k1 and
k = k1 + k2
@@ -1033,42 +1144,38 @@ private predicate compares_lt(Instruction test, Operand op, int k, boolean isLt,
/** `(a < b + k) => (b > a - k) => (b >= a + (1-k))` */
private predicate compares_ge(
Instruction test, Operand left, Operand right, int k, boolean isGe, AbstractValue value
ValueNumber test, Operand left, Operand right, int k, boolean isGe, AbstractValue value
) {
exists(int onemk | k = 1 - onemk | compares_lt(test, right, left, onemk, isGe, value))
}
/** Rearrange various simple comparisons into `left < right + k` form. */
private predicate simple_comparison_lt(CompareInstruction cmp, Operand left, Operand right, int k) {
left = cmp.getLeftOperand() and
cmp instanceof CompareLTInstruction and
right = cmp.getRightOperand() and
private predicate simple_comparison_lt(CompareValueNumber cmp, Operand left, Operand right, int k) {
cmp.hasOperands(left, right) and
cmp instanceof CompareLTValueNumber and
k = 0
or
left = cmp.getLeftOperand() and
cmp instanceof CompareLEInstruction and
right = cmp.getRightOperand() and
cmp.hasOperands(left, right) and
cmp instanceof CompareLEValueNumber and
k = 1
or
right = cmp.getLeftOperand() and
cmp instanceof CompareGTInstruction and
left = cmp.getRightOperand() and
cmp.hasOperands(right, left) and
cmp instanceof CompareGTValueNumber and
k = 0
or
right = cmp.getLeftOperand() and
cmp instanceof CompareGEInstruction and
left = cmp.getRightOperand() and
cmp.hasOperands(right, left) and
cmp instanceof CompareGEValueNumber and
k = 1
}
/** Rearrange various simple comparisons into `op < k` form. */
private predicate unary_simple_comparison_lt(
Instruction test, int k, boolean isLt, AbstractValue value
SwitchConditionValueNumber test, Operand op, int k, boolean isLt, AbstractValue value
) {
exists(SwitchInstruction switch, CaseEdge case |
test = switch.getExpression() and
exists(CaseEdge case |
test.getExpressionOperand() = op and
case = value.(MatchValue).getCase() and
exists(switch.getSuccessor(case)) and
exists(test.getSuccessor(case)) and
case.getMaxValue() > case.getMinValue()
|
// op <= k => op < k - 1
@@ -1081,7 +1188,7 @@ private predicate unary_simple_comparison_lt(
}
private predicate complex_lt(
CompareInstruction cmp, Operand left, Operand right, int k, boolean isLt, AbstractValue value
ValueNumber cmp, Operand left, Operand right, int k, boolean isLt, AbstractValue value
) {
sub_lt(cmp, left, right, k, isLt, value)
or
@@ -1089,7 +1196,7 @@ private predicate complex_lt(
}
private predicate complex_lt(
Instruction test, Operand left, int k, boolean isLt, AbstractValue value
ValueNumber test, Operand left, int k, boolean isLt, AbstractValue value
) {
sub_lt(test, left, k, isLt, value)
or
@@ -1099,7 +1206,7 @@ private predicate complex_lt(
// left - x < right + c => left < right + (c+x)
// left < (right - x) + c => left < right + (c-x)
private predicate sub_lt(
CompareInstruction cmp, Operand left, Operand right, int k, boolean isLt, AbstractValue value
ValueNumber cmp, Operand left, Operand right, int k, boolean isLt, AbstractValue value
) {
exists(SubInstruction lhs, int c, int x |
compares_lt(cmp, lhs.getAUse(), right, c, isLt, value) and
@@ -1130,7 +1237,7 @@ private predicate sub_lt(
)
}
private predicate sub_lt(Instruction test, Operand left, int k, boolean isLt, AbstractValue value) {
private predicate sub_lt(ValueNumber test, Operand left, int k, boolean isLt, AbstractValue value) {
exists(SubInstruction lhs, int c, int x |
compares_lt(test, lhs.getAUse(), c, isLt, value) and
left = lhs.getLeftOperand() and
@@ -1149,7 +1256,7 @@ private predicate sub_lt(Instruction test, Operand left, int k, boolean isLt, Ab
// left + x < right + c => left < right + (c-x)
// left < (right + x) + c => left < right + (c+x)
private predicate add_lt(
CompareInstruction cmp, Operand left, Operand right, int k, boolean isLt, AbstractValue value
ValueNumber cmp, Operand left, Operand right, int k, boolean isLt, AbstractValue value
) {
exists(AddInstruction lhs, int c, int x |
compares_lt(cmp, lhs.getAUse(), right, c, isLt, value) and
@@ -1192,7 +1299,7 @@ private predicate add_lt(
)
}
private predicate add_lt(Instruction test, Operand left, int k, boolean isLt, AbstractValue value) {
private predicate add_lt(ValueNumber test, Operand left, int k, boolean isLt, AbstractValue value) {
exists(AddInstruction lhs, int c, int x |
compares_lt(test, lhs.getAUse(), c, isLt, value) and
(
@@ -1217,7 +1324,7 @@ private predicate add_lt(Instruction test, Operand left, int k, boolean isLt, Ab
// left - x == right + c => left == right + (c+x)
// left == (right - x) + c => left == right + (c-x)
private predicate sub_eq(
CompareInstruction cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
ValueNumber cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
) {
exists(SubInstruction lhs, int c, int x |
compares_eq(cmp, lhs.getAUse(), right, c, areEqual, value) and
@@ -1250,7 +1357,7 @@ private predicate sub_eq(
// op - x == c => op == (c+x)
private predicate unary_sub_eq(
Instruction test, Operand op, int k, boolean areEqual, boolean inNonZeroCase, AbstractValue value
ValueNumber test, Operand op, int k, boolean areEqual, boolean inNonZeroCase, AbstractValue value
) {
inNonZeroCase = false and
exists(SubInstruction sub, int c, int x |
@@ -1272,7 +1379,7 @@ private predicate unary_sub_eq(
// left + x == right + c => left == right + (c-x)
// left == (right + x) + c => left == right + (c+x)
private predicate add_eq(
CompareInstruction cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
ValueNumber cmp, Operand left, Operand right, int k, boolean areEqual, AbstractValue value
) {
exists(AddInstruction lhs, int c, int x |
compares_eq(cmp, lhs.getAUse(), right, c, areEqual, value) and
@@ -1317,7 +1424,7 @@ private predicate add_eq(
// left + x == right + c => left == right + (c-x)
private predicate unary_add_eq(
Instruction test, Operand left, int k, boolean areEqual, boolean inNonZeroCase,
ValueNumber test, Operand left, int k, boolean areEqual, boolean inNonZeroCase,
AbstractValue value
) {
inNonZeroCase = false and
@@ -1351,6 +1458,4 @@ private class IntegerOrPointerConstantInstruction extends ConstantInstruction {
}
/** The int value of integer constant expression. */
private int int_value(Instruction i) {
result = i.(IntegerOrPointerConstantInstruction).getValue().toInt()
}
private int int_value(IntegerOrPointerConstantInstruction i) { result = i.getValue().toInt() }

View File

@@ -29,5 +29,5 @@ deprecated module DataFlow {
private import semmle.code.cpp.dataflow.internal.DataFlowImplSpecific
private import codeql.dataflow.DataFlow
import DataFlowMake<Location, CppOldDataFlow>
import Public
import semmle.code.cpp.dataflow.internal.DataFlowImpl1
}

View File

@@ -0,0 +1,22 @@
/**
* Provides a `DataFlow2` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.dataflow.DataFlow` for the full documentation.
*/
import cpp
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.DataFlow2` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow2 {
import semmle.code.cpp.dataflow.internal.DataFlowImpl2
}

View File

@@ -0,0 +1,22 @@
/**
* Provides a `DataFlow3` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.dataflow.DataFlow` for the full documentation.
*/
import cpp
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.DataFlow3` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow3 {
import semmle.code.cpp.dataflow.internal.DataFlowImpl3
}

View File

@@ -0,0 +1,22 @@
/**
* Provides a `DataFlow4` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.dataflow.DataFlow` for the full documentation.
*/
import cpp
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.DataFlow4` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
deprecated module DataFlow4 {
import semmle.code.cpp.dataflow.internal.DataFlowImpl4
}

View File

@@ -793,27 +793,28 @@ private Element interpretElement0(
) {
(
// Non-member functions
funcHasQualifiedName(result, namespace, name) and
elementSpec(namespace, type, subtypes, name, signature, _) and
subtypes = false and
type = "" and
(
elementSpecMatchesSignature(result, namespace, type, subtypes, name, signature)
or
signature = "" and
elementSpec(namespace, type, subtypes, name, signature, _)
elementSpec(namespace, type, subtypes, name, "", _) and
funcHasQualifiedName(result, namespace, name)
)
or
// Member functions
exists(Class namedClass, Class classWithMethod |
hasClassAndName(classWithMethod, result, name) and
classHasQualifiedName(namedClass, namespace, type)
|
(
elementSpecMatchesSignature(result, namespace, type, subtypes, name, signature)
elementSpecMatchesSignature(result, namespace, type, subtypes, name, signature) and
hasClassAndName(classWithMethod, result, name)
or
signature = "" and
elementSpec(namespace, type, subtypes, name, "", _)
elementSpec(namespace, type, subtypes, name, "", _) and
hasClassAndName(classWithMethod, result, name)
) and
classHasQualifiedName(namedClass, namespace, type) and
(
// member declared in the named type or a subtype of it
subtypes = true and

View File

@@ -0,0 +1,39 @@
/**
* DEPRECATED: Recursion through `DataFlow::Configuration` is impossible in
* any supported tooling. There is no need for this module because it's
* impossible to accidentally depend on recursion through
* `DataFlow::Configuration` in current releases.
*
* When this module is imported, recursive use of `DataFlow::Configuration` is
* disallowed. Importing this module will guarantee the absence of such
* recursion, which is unsupported and will be unconditionally disallowed in a
* future release.
*
* Recursive use of `DataFlow{2..4}::Configuration` is always disallowed, so no
* import is needed for those.
*/
import cpp
private import semmle.code.cpp.dataflow.DataFlow
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
* Four copies are available: `DataFlow` through `DataFlow4`.
*/
abstract private class ConfigurationRecursionPrevention extends DataFlow::Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(DataFlow::Node source, DataFlow::Node sink) {
strictcount(DataFlow::Node n | this.isSource(n)) < 0
or
strictcount(DataFlow::Node n | this.isSink(n)) < 0
or
strictcount(DataFlow::Node n1, DataFlow::Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
super.hasFlow(source, sink)
}
}

View File

@@ -16,6 +16,7 @@
*/
import semmle.code.cpp.dataflow.DataFlow
import semmle.code.cpp.dataflow.DataFlow2
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.TaintTracking` instead.
@@ -24,9 +25,10 @@ import semmle.code.cpp.dataflow.DataFlow
* global (inter-procedural) taint-tracking analyses.
*/
deprecated module TaintTracking {
import semmle.code.cpp.dataflow.internal.TaintTrackingUtil
import semmle.code.cpp.dataflow.internal.tainttracking1.TaintTrackingParameter::Public
private import semmle.code.cpp.dataflow.internal.DataFlowImplSpecific
private import semmle.code.cpp.dataflow.internal.TaintTrackingImplSpecific
private import codeql.dataflow.TaintTracking
import TaintFlowMake<Location, CppOldDataFlow, CppOldTaintTracking>
import semmle.code.cpp.dataflow.internal.tainttracking1.TaintTrackingImpl
}

View File

@@ -0,0 +1,22 @@
/**
* Provides a `TaintTracking2` module, which is a copy of the `TaintTracking`
* module. Use this class when data-flow configurations or taint-tracking
* configurations must depend on each other. Two classes extending
* `DataFlow::Configuration` should never depend on each other, but one of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`. The
* `TaintTracking::Configuration` class extends `DataFlow::Configuration`, and
* `TaintTracking2::Configuration` extends `DataFlow2::Configuration`.
*
* See `semmle.code.cpp.dataflow.TaintTracking` for the full documentation.
*/
/**
* DEPRECATED: Use `semmle.code.cpp.dataflow.new.TaintTracking2` instead.
*
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
deprecated module TaintTracking2 {
import semmle.code.cpp.dataflow.internal.tainttracking2.TaintTrackingImpl
}

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -729,39 +729,41 @@ private predicate exprToDefinitionByReferenceStep(Expr exprIn, Expr argOut) {
private module FieldFlow {
private import DataFlowImplCommon
private import DataFlowImplLocal
private import DataFlowPrivate
private import semmle.code.cpp.dataflow.DataFlow
/**
* A configuration for finding local-only flow through fields.
* A configuration for finding local-only flow through fields. This uses the
* `Configuration` class in the dedicated `DataFlowImplLocal` copy of the
* shared library that's not user-exposed directly.
*
* To keep the flow local to a single function, we put barriers on parameters
* and return statements. Sources and sinks are the values that go into and
* out of fields, respectively.
*/
private module FieldConfig implements DataFlow::ConfigSig {
predicate isSource(Node source) {
private class FieldConfiguration extends Configuration {
FieldConfiguration() { this = "FieldConfiguration" }
override predicate isSource(Node source) {
storeStep(source, _, _)
or
// Also mark `foo(a.b);` as a source when `a.b` may be overwritten by `foo`.
readStep(_, _, any(Node node | node.asExpr() = source.asDefiningArgument()))
}
predicate isSink(Node sink) { readStep(_, _, sink) }
override predicate isSink(Node sink) { readStep(_, _, sink) }
predicate isBarrier(Node node) { node instanceof ParameterNode }
override predicate isBarrier(Node node) { node instanceof ParameterNode }
predicate isBarrierOut(Node node) {
override predicate isBarrierOut(Node node) {
node.asExpr().getParent() instanceof ReturnStmt
or
node.asExpr().getParent() instanceof ThrowExpr
}
}
private module Flow = DataFlow::Global<FieldConfig>;
predicate fieldFlow(Node node1, Node node2) {
Flow::flow(node1, node2) and
exists(FieldConfiguration cfg | cfg.hasFlow(node1, node2)) and
// This configuration should not be able to cross function boundaries, but
// we double-check here just to be sure.
getNodeEnclosingCallable(node1) = getNodeEnclosingCallable(node2)

View File

@@ -0,0 +1,168 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides an implementation of global (interprocedural) taint tracking.
* This file re-exports the local (intraprocedural) taint-tracking analysis
* from `TaintTrackingParameter::Public` and adds a global analysis, mainly
* exposed through the `Configuration` class. For some languages, this file
* exists in several identical copies, allowing queries to use multiple
* `Configuration` classes that depend on each other without introducing
* mutual recursion among those configurations.
*/
import TaintTrackingParameter::Public
private import TaintTrackingParameter::Private
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural taint tracking analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the taint tracking library must define its own unique extension of
* this abstract class.
*
* A taint-tracking configuration is a special data flow configuration
* (`DataFlow::Configuration`) that allows for flow through nodes that do not
* necessarily preserve values but are still relevant from a taint tracking
* perspective. (For example, string concatenation, where one of the operands
* is tainted.)
*
* To create a configuration, extend this class with a subclass whose
* characteristic predicate is a unique singleton string. For example, write
*
* ```ql
* class MyAnalysisConfiguration extends TaintTracking::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isSanitizer`.
* // Optionally override `isSanitizerIn`.
* // Optionally override `isSanitizerOut`.
* // Optionally override `isSanitizerGuard`.
* // Optionally override `isAdditionalTaintStep`.
* }
* ```
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but it is unsupported to depend on
* another `TaintTracking::Configuration` or a `DataFlow::Configuration` in the
* overridden predicates that define sources, sinks, or additional steps.
* Instead, the dependency should go to a `TaintTracking2::Configuration` or a
* `DataFlow2::Configuration`, `DataFlow3::Configuration`, etc.
*/
abstract deprecated class Configuration extends DataFlow::Configuration {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant taint source.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source) { none() }
/**
* Holds if `source` is a relevant taint source with the given initial
* `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source, DataFlow::FlowState state) { none() }
/**
* Holds if `sink` is a relevant taint sink
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink) { none() }
/**
* Holds if `sink` is a relevant taint sink accepting `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink, DataFlow::FlowState state) { none() }
/** Holds if the node `node` is a taint sanitizer. */
predicate isSanitizer(DataFlow::Node node) { none() }
final override predicate isBarrier(DataFlow::Node node) {
this.isSanitizer(node) or
defaultTaintSanitizer(node)
}
/**
* Holds if the node `node` is a taint sanitizer when the flow state is
* `state`.
*/
predicate isSanitizer(DataFlow::Node node, DataFlow::FlowState state) { none() }
final override predicate isBarrier(DataFlow::Node node, DataFlow::FlowState state) {
this.isSanitizer(node, state)
}
/** Holds if taint propagation into `node` is prohibited. */
predicate isSanitizerIn(DataFlow::Node node) { none() }
final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
/** Holds if taint propagation out of `node` is prohibited. */
predicate isSanitizerOut(DataFlow::Node node) { none() }
final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
*/
predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
this.isAdditionalTaintStep(node1, node2) or
defaultAdditionalTaintStep(node1, node2, _)
}
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalTaintStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
none()
}
final override predicate isAdditionalFlowStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
this.isAdditionalTaintStep(node1, state1, node2, state2)
}
override predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
(
this.isSink(node) or
this.isSink(node, _) or
this.isAdditionalTaintStep(node, _) or
this.isAdditionalTaintStep(node, _, _, _)
) and
defaultImplicitTaintRead(node, c)
}
/**
* Holds if taint may flow from `source` to `sink` for this configuration.
*/
// overridden to provide taint-tracking specific qldoc
override predicate hasFlow(DataFlow::Node source, DataFlow::Node sink) {
super.hasFlow(source, sink)
}
}

View File

@@ -0,0 +1,10 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*/
import semmle.code.cpp.dataflow.internal.TaintTrackingUtil as Public
module Private {
import semmle.code.cpp.dataflow.DataFlow::DataFlow as DataFlow
import semmle.code.cpp.dataflow.internal.DataFlowImpl as DataFlowInternal
}

View File

@@ -0,0 +1,168 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides an implementation of global (interprocedural) taint tracking.
* This file re-exports the local (intraprocedural) taint-tracking analysis
* from `TaintTrackingParameter::Public` and adds a global analysis, mainly
* exposed through the `Configuration` class. For some languages, this file
* exists in several identical copies, allowing queries to use multiple
* `Configuration` classes that depend on each other without introducing
* mutual recursion among those configurations.
*/
import TaintTrackingParameter::Public
private import TaintTrackingParameter::Private
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural taint tracking analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the taint tracking library must define its own unique extension of
* this abstract class.
*
* A taint-tracking configuration is a special data flow configuration
* (`DataFlow::Configuration`) that allows for flow through nodes that do not
* necessarily preserve values but are still relevant from a taint tracking
* perspective. (For example, string concatenation, where one of the operands
* is tainted.)
*
* To create a configuration, extend this class with a subclass whose
* characteristic predicate is a unique singleton string. For example, write
*
* ```ql
* class MyAnalysisConfiguration extends TaintTracking::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isSanitizer`.
* // Optionally override `isSanitizerIn`.
* // Optionally override `isSanitizerOut`.
* // Optionally override `isSanitizerGuard`.
* // Optionally override `isAdditionalTaintStep`.
* }
* ```
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but it is unsupported to depend on
* another `TaintTracking::Configuration` or a `DataFlow::Configuration` in the
* overridden predicates that define sources, sinks, or additional steps.
* Instead, the dependency should go to a `TaintTracking2::Configuration` or a
* `DataFlow2::Configuration`, `DataFlow3::Configuration`, etc.
*/
abstract deprecated class Configuration extends DataFlow::Configuration {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant taint source.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source) { none() }
/**
* Holds if `source` is a relevant taint source with the given initial
* `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source, DataFlow::FlowState state) { none() }
/**
* Holds if `sink` is a relevant taint sink
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink) { none() }
/**
* Holds if `sink` is a relevant taint sink accepting `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink, DataFlow::FlowState state) { none() }
/** Holds if the node `node` is a taint sanitizer. */
predicate isSanitizer(DataFlow::Node node) { none() }
final override predicate isBarrier(DataFlow::Node node) {
this.isSanitizer(node) or
defaultTaintSanitizer(node)
}
/**
* Holds if the node `node` is a taint sanitizer when the flow state is
* `state`.
*/
predicate isSanitizer(DataFlow::Node node, DataFlow::FlowState state) { none() }
final override predicate isBarrier(DataFlow::Node node, DataFlow::FlowState state) {
this.isSanitizer(node, state)
}
/** Holds if taint propagation into `node` is prohibited. */
predicate isSanitizerIn(DataFlow::Node node) { none() }
final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
/** Holds if taint propagation out of `node` is prohibited. */
predicate isSanitizerOut(DataFlow::Node node) { none() }
final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
*/
predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
this.isAdditionalTaintStep(node1, node2) or
defaultAdditionalTaintStep(node1, node2, _)
}
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalTaintStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
none()
}
final override predicate isAdditionalFlowStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
this.isAdditionalTaintStep(node1, state1, node2, state2)
}
override predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
(
this.isSink(node) or
this.isSink(node, _) or
this.isAdditionalTaintStep(node, _) or
this.isAdditionalTaintStep(node, _, _, _)
) and
defaultImplicitTaintRead(node, c)
}
/**
* Holds if taint may flow from `source` to `sink` for this configuration.
*/
// overridden to provide taint-tracking specific qldoc
override predicate hasFlow(DataFlow::Node source, DataFlow::Node sink) {
super.hasFlow(source, sink)
}
}

View File

@@ -0,0 +1,9 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*/
import semmle.code.cpp.dataflow.internal.TaintTrackingUtil as Public
module Private {
import semmle.code.cpp.dataflow.DataFlow2::DataFlow2 as DataFlow
}

View File

@@ -29,5 +29,5 @@ module DataFlow {
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import codeql.dataflow.DataFlow
import DataFlowMake<Location, CppDataFlow>
import Public
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
}

View File

@@ -0,0 +1,20 @@
/**
* Provides a `DataFlow2` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.dataflow.new.DataFlow` for the full documentation.
*/
import cpp
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
module DataFlow2 {
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl2
}

View File

@@ -0,0 +1,20 @@
/**
* Provides a `DataFlow3` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.dataflow.new.DataFlow` for the full documentation.
*/
import cpp
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
module DataFlow3 {
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl3
}

View File

@@ -0,0 +1,20 @@
/**
* Provides a `DataFlow4` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.dataflow.new.DataFlow` for the full documentation.
*/
import cpp
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) data flow analyses.
*/
module DataFlow4 {
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl4
}

View File

@@ -16,16 +16,18 @@
*/
import semmle.code.cpp.dataflow.new.DataFlow
import semmle.code.cpp.dataflow.new.DataFlow2
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
module TaintTracking {
import semmle.code.cpp.ir.dataflow.internal.TaintTrackingUtil
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingParameter::Public
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import semmle.code.cpp.ir.dataflow.internal.TaintTrackingImplSpecific
private import codeql.dataflow.TaintTracking
private import semmle.code.cpp.Location
import TaintFlowMake<Location, CppDataFlow, CppTaintTracking>
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingImpl
}

View File

@@ -0,0 +1,20 @@
/**
* Provides a `TaintTracking2` module, which is a copy of the `TaintTracking`
* module. Use this class when data-flow configurations or taint-tracking
* configurations must depend on each other. Two classes extending
* `DataFlow::Configuration` should never depend on each other, but one of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`. The
* `TaintTracking::Configuration` class extends `DataFlow::Configuration`, and
* `TaintTracking2::Configuration` extends `DataFlow2::Configuration`.
*
* See `semmle.code.cpp.dataflow.new.TaintTracking` for the full documentation.
*/
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
module TaintTracking2 {
import semmle.code.cpp.ir.dataflow.internal.tainttracking2.TaintTrackingImpl
}

View File

@@ -0,0 +1,20 @@
/**
* Provides a `TaintTracking3` module, which is a copy of the `TaintTracking`
* module. Use this class when data-flow configurations or taint-tracking
* configurations must depend on each other. Two classes extending
* `DataFlow::Configuration` should never depend on each other, but one of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`. The
* `TaintTracking::Configuration` class extends `DataFlow::Configuration`, and
* `TaintTracking2::Configuration` extends `DataFlow2::Configuration`.
*
* See `semmle.code.cpp.dataflow.new.TaintTracking` for the full documentation.
*/
/**
* Provides classes for performing local (intra-procedural) and
* global (inter-procedural) taint-tracking analyses.
*/
module TaintTracking3 {
import semmle.code.cpp.ir.dataflow.internal.tainttracking3.TaintTrackingImpl
}

View File

@@ -25,5 +25,5 @@ module DataFlow {
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import codeql.dataflow.DataFlow
import DataFlowMake<Location, CppDataFlow>
import Public
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
}

View File

@@ -0,0 +1,16 @@
/**
* Provides a `DataFlow2` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.ir.dataflow.DataFlow` for the full documentation.
*/
import cpp
module DataFlow2 {
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl2
}

View File

@@ -0,0 +1,16 @@
/**
* Provides a `DataFlow3` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.ir.dataflow.DataFlow` for the full documentation.
*/
import cpp
module DataFlow3 {
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl3
}

View File

@@ -0,0 +1,16 @@
/**
* Provides a `DataFlow4` module, which is a copy of the `DataFlow` module. Use
* this class when data-flow configurations must depend on each other. Two
* classes extending `DataFlow::Configuration` should never depend on each
* other, but one of them should instead depend on a
* `DataFlow2::Configuration`, a `DataFlow3::Configuration`, or a
* `DataFlow4::Configuration`.
*
* See `semmle.code.cpp.ir.dataflow.DataFlow` for the full documentation.
*/
import cpp
module DataFlow4 {
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl4
}

View File

@@ -16,11 +16,13 @@
*/
import semmle.code.cpp.ir.dataflow.DataFlow
import semmle.code.cpp.ir.dataflow.DataFlow2
module TaintTracking {
import semmle.code.cpp.ir.dataflow.internal.TaintTrackingUtil
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingParameter::Public
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplSpecific
private import semmle.code.cpp.ir.dataflow.internal.TaintTrackingImplSpecific
private import codeql.dataflow.TaintTracking
import TaintFlowMake<Location, CppDataFlow, CppTaintTracking>
import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingImpl
}

View File

@@ -0,0 +1,15 @@
/**
* Provides a `TaintTracking2` module, which is a copy of the `TaintTracking`
* module. Use this class when data-flow configurations or taint-tracking
* configurations must depend on each other. Two classes extending
* `DataFlow::Configuration` should never depend on each other, but one of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`. The
* `TaintTracking::Configuration` class extends `DataFlow::Configuration`, and
* `TaintTracking2::Configuration` extends `DataFlow2::Configuration`.
*
* See `semmle.code.cpp.ir.dataflow.TaintTracking` for the full documentation.
*/
module TaintTracking2 {
import semmle.code.cpp.ir.dataflow.internal.tainttracking2.TaintTrackingImpl
}

View File

@@ -0,0 +1,15 @@
/**
* Provides a `TaintTracking3` module, which is a copy of the `TaintTracking`
* module. Use this class when data-flow configurations or taint-tracking
* configurations must depend on each other. Two classes extending
* `DataFlow::Configuration` should never depend on each other, but one of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`. The
* `TaintTracking::Configuration` class extends `DataFlow::Configuration`, and
* `TaintTracking2::Configuration` extends `DataFlow2::Configuration`.
*
* See `semmle.code.cpp.ir.dataflow.TaintTracking` for the full documentation.
*/
module TaintTracking3 {
import semmle.code.cpp.ir.dataflow.internal.tainttracking3.TaintTrackingImpl
}

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -0,0 +1,361 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides a `Configuration` class backwards-compatible interface to the data
* flow library.
*/
private import DataFlowImplCommon
private import DataFlowImplSpecific::Private
import DataFlowImplSpecific::Public
private import DataFlowImpl
import DataFlowImplCommonPublic
deprecated import FlowStateString
private import codeql.util.Unit
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural data flow analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the global data flow library must define its own unique extension
* of this abstract class. To create a configuration, extend this class with
* a subclass whose characteristic predicate is a unique singleton string.
* For example, write
*
* ```ql
* class MyAnalysisConfiguration extends DataFlow::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isBarrier`.
* // Optionally override `isAdditionalFlowStep`.
* }
* ```
* Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
* the edges are those data-flow steps that preserve the value of the node
* along with any additional edges defined by `isAdditionalFlowStep`.
* Specifying nodes in `isBarrier` will remove those nodes from the graph, and
* specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
* and/or out-going edges from those nodes, respectively.
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but two classes extending
* `DataFlow::Configuration` should never depend on each other. One of them
* should instead depend on a `DataFlow2::Configuration`, a
* `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
*/
abstract deprecated class Configuration extends string {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant data flow source.
*/
predicate isSource(Node source) { none() }
/**
* Holds if `source` is a relevant data flow source with the given initial
* `state`.
*/
predicate isSource(Node source, FlowState state) { none() }
/**
* Holds if `sink` is a relevant data flow sink.
*/
predicate isSink(Node sink) { none() }
/**
* Holds if `sink` is a relevant data flow sink accepting `state`.
*/
predicate isSink(Node sink, FlowState state) { none() }
/**
* Holds if data flow through `node` is prohibited. This completely removes
* `node` from the data flow graph.
*/
predicate isBarrier(Node node) { none() }
/**
* Holds if data flow through `node` is prohibited when the flow state is
* `state`.
*/
predicate isBarrier(Node node, FlowState state) { none() }
/** Holds if data flow into `node` is prohibited. */
predicate isBarrierIn(Node node) { none() }
/** Holds if data flow out of `node` is prohibited. */
predicate isBarrierOut(Node node) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
*/
predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
/**
* Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
none()
}
/**
* Holds if an arbitrary number of implicit read steps of content `c` may be
* taken at `node`.
*/
predicate allowImplicitRead(Node node, ContentSet c) { none() }
/**
* Gets the virtual dispatch branching limit when calculating field flow.
* This can be overridden to a smaller value to improve performance (a
* value of 0 disables field flow), or a larger value to get more results.
*/
int fieldFlowBranchLimit() { result = 2 }
/**
* Gets a data flow configuration feature to add restrictions to the set of
* valid flow paths.
*
* - `FeatureHasSourceCallContext`:
* Assume that sources have some existing call context to disallow
* conflicting return-flow directly following the source.
* - `FeatureHasSinkCallContext`:
* Assume that sinks have some existing call context to disallow
* conflicting argument-to-parameter flow directly preceding the sink.
* - `FeatureEqualSourceSinkCallContext`:
* Implies both of the above and additionally ensures that the entire flow
* path preserves the call context.
*
* These features are generally not relevant for typical end-to-end data flow
* queries, but should only be used for constructing paths that need to
* somehow be pluggable in another path context.
*/
FlowFeature getAFeature() { none() }
/** Holds if sources should be grouped in the result of `hasFlowPath`. */
predicate sourceGrouping(Node source, string sourceGroup) { none() }
/** Holds if sinks should be grouped in the result of `hasFlowPath`. */
predicate sinkGrouping(Node sink, string sinkGroup) { none() }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*/
predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
/**
* Holds if data may flow from `source` to `sink` for this configuration.
*
* The corresponding paths are generated from the end-points and the graph
* included in the module `PathGraph`.
*/
predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
/**
* Holds if data may flow from some source to `sink` for this configuration.
*/
predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
/**
* Holds if hidden nodes should be included in the data flow graph.
*
* This feature should only be used for debugging or when the data flow graph
* is not visualized (for example in a `path-problem` query).
*/
predicate includeHiddenNodes() { none() }
}
/**
* This class exists to prevent mutual recursion between the user-overridden
* member predicates of `Configuration` and the rest of the data-flow library.
* Good performance cannot be guaranteed in the presence of such recursion, so
* it should be replaced by using more than one copy of the data flow library.
*/
abstract deprecated private class ConfigurationRecursionPrevention extends Configuration {
bindingset[this]
ConfigurationRecursionPrevention() { any() }
override predicate hasFlow(Node source, Node sink) {
strictcount(Node n | this.isSource(n)) < 0
or
strictcount(Node n | this.isSource(n, _)) < 0
or
strictcount(Node n | this.isSink(n)) < 0
or
strictcount(Node n | this.isSink(n, _)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
or
strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
or
super.hasFlow(source, sink)
}
}
deprecated private FlowState relevantState(Configuration config) {
config.isSource(_, result) or
config.isSink(_, result) or
config.isBarrier(_, result) or
config.isAdditionalFlowStep(_, result, _, _) or
config.isAdditionalFlowStep(_, _, _, result)
}
private newtype TConfigState =
deprecated TMkConfigState(Configuration config, FlowState state) {
state = relevantState(config) or state instanceof FlowStateEmpty
}
deprecated private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
deprecated private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
deprecated private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
deprecated private module Config implements FullStateConfigSig {
class FlowState = TConfigState;
predicate isSource(Node source, FlowState state) {
getConfig(state).isSource(source, getState(state))
or
getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
}
predicate isSink(Node sink) { none() }
predicate isSink(Node sink, FlowState state) {
getConfig(state).isSink(sink, getState(state))
or
getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
}
predicate isBarrier(Node node) { none() }
predicate isBarrier(Node node, FlowState state) {
getConfig(state).isBarrier(node, getState(state)) or
getConfig(state).isBarrier(node)
}
predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
predicate isBarrierIn(Node node, FlowState state) { none() }
predicate isBarrierOut(Node node, FlowState state) { none() }
predicate isAdditionalFlowStep(Node node1, Node node2, string model) {
singleConfiguration() and
any(Configuration config).isAdditionalFlowStep(node1, node2) and
model = ""
}
predicate isAdditionalFlowStep(
Node node1, FlowState state1, Node node2, FlowState state2, string model
) {
getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
getConfig(state2) = getConfig(state1) and
model = ""
or
not singleConfiguration() and
getConfig(state1).isAdditionalFlowStep(node1, node2) and
state2 = state1 and
model = ""
}
predicate allowImplicitRead(Node node, ContentSet c) {
any(Configuration config).allowImplicitRead(node, c)
}
predicate neverSkip(Node node) { none() }
int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
int accessPathLimit() { result = 5 }
FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
predicate observeDiffInformedIncrementalMode() { none() }
}
deprecated private import Impl<Config> as I
/**
* A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
* Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
*/
deprecated class PathNode instanceof I::PathNode {
/** Gets a textual representation of this element. */
final string toString() { result = super.toString() }
/**
* Gets a textual representation of this element, including a textual
* representation of the call context.
*/
final string toStringWithContext() { result = super.toStringWithContext() }
/**
* Holds if this element is at the specified location.
* The location spans column `startcolumn` of line `startline` to
* column `endcolumn` of line `endline` in file `filepath`.
* For more information, see
* [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
*/
final predicate hasLocationInfo(
string filepath, int startline, int startcolumn, int endline, int endcolumn
) {
super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
}
/** Gets the underlying `Node`. */
final Node getNode() { result = super.getNode() }
/** Gets the `FlowState` of this node. */
deprecated final FlowState getState() { result = getState(super.getState()) }
/** Gets the associated configuration. */
deprecated final Configuration getConfiguration() { result = getConfig(super.getState()) }
/** Gets a successor of this node, if any. */
final PathNode getASuccessor() { result = super.getASuccessor() }
/** Holds if this node is a source. */
final predicate isSource() { super.isSource() }
/** Holds if this node is a grouping of source nodes. */
final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
/** Holds if this node is a grouping of sink nodes. */
final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
}
deprecated module PathGraph = I::PathGraph;
deprecated private predicate hasFlow(Node source, Node sink, Configuration config) {
exists(PathNode source0, PathNode sink0 |
hasFlowPath(source0, sink0, config) and
source0.getNode() = source and
sink0.getNode() = sink
)
}
deprecated private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
I::flowPath(source, sink) and source.getConfiguration() = config
}
deprecated private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
deprecated predicate flowsTo = hasFlow/3;

View File

@@ -545,7 +545,7 @@ module ProductFlow {
private predicate outImpl1(Flow1::PathNode pred1, Flow1::PathNode succ1, DataFlowCall call) {
Flow1::PathGraph::edges(pred1, succ1, _, _) and
exists(ReturnKindExt returnKind |
succ1.getNode() = getAnOutNodeExt(call, returnKind) and
succ1.getNode() = returnKind.getAnOutNode(call) and
returnKind = getParamReturnPosition(_, pred1.asParameterReturnNode()).getKind()
)
}
@@ -573,7 +573,7 @@ module ProductFlow {
private predicate outImpl2(Flow2::PathNode pred2, Flow2::PathNode succ2, DataFlowCall call) {
Flow2::PathGraph::edges(pred2, succ2, _, _) and
exists(ReturnKindExt returnKind |
succ2.getNode() = getAnOutNodeExt(call, returnKind) and
succ2.getNode() = returnKind.getAnOutNode(call) and
returnKind = getParamReturnPosition(_, pred2.asParameterReturnNode()).getKind()
)
}

View File

@@ -0,0 +1,168 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides an implementation of global (interprocedural) taint tracking.
* This file re-exports the local (intraprocedural) taint-tracking analysis
* from `TaintTrackingParameter::Public` and adds a global analysis, mainly
* exposed through the `Configuration` class. For some languages, this file
* exists in several identical copies, allowing queries to use multiple
* `Configuration` classes that depend on each other without introducing
* mutual recursion among those configurations.
*/
import TaintTrackingParameter::Public
private import TaintTrackingParameter::Private
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural taint tracking analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the taint tracking library must define its own unique extension of
* this abstract class.
*
* A taint-tracking configuration is a special data flow configuration
* (`DataFlow::Configuration`) that allows for flow through nodes that do not
* necessarily preserve values but are still relevant from a taint tracking
* perspective. (For example, string concatenation, where one of the operands
* is tainted.)
*
* To create a configuration, extend this class with a subclass whose
* characteristic predicate is a unique singleton string. For example, write
*
* ```ql
* class MyAnalysisConfiguration extends TaintTracking::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isSanitizer`.
* // Optionally override `isSanitizerIn`.
* // Optionally override `isSanitizerOut`.
* // Optionally override `isSanitizerGuard`.
* // Optionally override `isAdditionalTaintStep`.
* }
* ```
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but it is unsupported to depend on
* another `TaintTracking::Configuration` or a `DataFlow::Configuration` in the
* overridden predicates that define sources, sinks, or additional steps.
* Instead, the dependency should go to a `TaintTracking2::Configuration` or a
* `DataFlow2::Configuration`, `DataFlow3::Configuration`, etc.
*/
abstract deprecated class Configuration extends DataFlow::Configuration {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant taint source.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source) { none() }
/**
* Holds if `source` is a relevant taint source with the given initial
* `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source, DataFlow::FlowState state) { none() }
/**
* Holds if `sink` is a relevant taint sink
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink) { none() }
/**
* Holds if `sink` is a relevant taint sink accepting `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink, DataFlow::FlowState state) { none() }
/** Holds if the node `node` is a taint sanitizer. */
predicate isSanitizer(DataFlow::Node node) { none() }
final override predicate isBarrier(DataFlow::Node node) {
this.isSanitizer(node) or
defaultTaintSanitizer(node)
}
/**
* Holds if the node `node` is a taint sanitizer when the flow state is
* `state`.
*/
predicate isSanitizer(DataFlow::Node node, DataFlow::FlowState state) { none() }
final override predicate isBarrier(DataFlow::Node node, DataFlow::FlowState state) {
this.isSanitizer(node, state)
}
/** Holds if taint propagation into `node` is prohibited. */
predicate isSanitizerIn(DataFlow::Node node) { none() }
final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
/** Holds if taint propagation out of `node` is prohibited. */
predicate isSanitizerOut(DataFlow::Node node) { none() }
final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
*/
predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
this.isAdditionalTaintStep(node1, node2) or
defaultAdditionalTaintStep(node1, node2, _)
}
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalTaintStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
none()
}
final override predicate isAdditionalFlowStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
this.isAdditionalTaintStep(node1, state1, node2, state2)
}
override predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
(
this.isSink(node) or
this.isSink(node, _) or
this.isAdditionalTaintStep(node, _) or
this.isAdditionalTaintStep(node, _, _, _)
) and
defaultImplicitTaintRead(node, c)
}
/**
* Holds if taint may flow from `source` to `sink` for this configuration.
*/
// overridden to provide taint-tracking specific qldoc
override predicate hasFlow(DataFlow::Node source, DataFlow::Node sink) {
super.hasFlow(source, sink)
}
}

View File

@@ -0,0 +1,6 @@
import semmle.code.cpp.ir.dataflow.internal.TaintTrackingUtil as Public
module Private {
import semmle.code.cpp.ir.dataflow.DataFlow::DataFlow as DataFlow
import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl as DataFlowInternal
}

View File

@@ -0,0 +1,168 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides an implementation of global (interprocedural) taint tracking.
* This file re-exports the local (intraprocedural) taint-tracking analysis
* from `TaintTrackingParameter::Public` and adds a global analysis, mainly
* exposed through the `Configuration` class. For some languages, this file
* exists in several identical copies, allowing queries to use multiple
* `Configuration` classes that depend on each other without introducing
* mutual recursion among those configurations.
*/
import TaintTrackingParameter::Public
private import TaintTrackingParameter::Private
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural taint tracking analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the taint tracking library must define its own unique extension of
* this abstract class.
*
* A taint-tracking configuration is a special data flow configuration
* (`DataFlow::Configuration`) that allows for flow through nodes that do not
* necessarily preserve values but are still relevant from a taint tracking
* perspective. (For example, string concatenation, where one of the operands
* is tainted.)
*
* To create a configuration, extend this class with a subclass whose
* characteristic predicate is a unique singleton string. For example, write
*
* ```ql
* class MyAnalysisConfiguration extends TaintTracking::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isSanitizer`.
* // Optionally override `isSanitizerIn`.
* // Optionally override `isSanitizerOut`.
* // Optionally override `isSanitizerGuard`.
* // Optionally override `isAdditionalTaintStep`.
* }
* ```
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but it is unsupported to depend on
* another `TaintTracking::Configuration` or a `DataFlow::Configuration` in the
* overridden predicates that define sources, sinks, or additional steps.
* Instead, the dependency should go to a `TaintTracking2::Configuration` or a
* `DataFlow2::Configuration`, `DataFlow3::Configuration`, etc.
*/
abstract deprecated class Configuration extends DataFlow::Configuration {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant taint source.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source) { none() }
/**
* Holds if `source` is a relevant taint source with the given initial
* `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source, DataFlow::FlowState state) { none() }
/**
* Holds if `sink` is a relevant taint sink
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink) { none() }
/**
* Holds if `sink` is a relevant taint sink accepting `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink, DataFlow::FlowState state) { none() }
/** Holds if the node `node` is a taint sanitizer. */
predicate isSanitizer(DataFlow::Node node) { none() }
final override predicate isBarrier(DataFlow::Node node) {
this.isSanitizer(node) or
defaultTaintSanitizer(node)
}
/**
* Holds if the node `node` is a taint sanitizer when the flow state is
* `state`.
*/
predicate isSanitizer(DataFlow::Node node, DataFlow::FlowState state) { none() }
final override predicate isBarrier(DataFlow::Node node, DataFlow::FlowState state) {
this.isSanitizer(node, state)
}
/** Holds if taint propagation into `node` is prohibited. */
predicate isSanitizerIn(DataFlow::Node node) { none() }
final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
/** Holds if taint propagation out of `node` is prohibited. */
predicate isSanitizerOut(DataFlow::Node node) { none() }
final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
*/
predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
this.isAdditionalTaintStep(node1, node2) or
defaultAdditionalTaintStep(node1, node2, _)
}
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalTaintStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
none()
}
final override predicate isAdditionalFlowStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
this.isAdditionalTaintStep(node1, state1, node2, state2)
}
override predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
(
this.isSink(node) or
this.isSink(node, _) or
this.isAdditionalTaintStep(node, _) or
this.isAdditionalTaintStep(node, _, _, _)
) and
defaultImplicitTaintRead(node, c)
}
/**
* Holds if taint may flow from `source` to `sink` for this configuration.
*/
// overridden to provide taint-tracking specific qldoc
override predicate hasFlow(DataFlow::Node source, DataFlow::Node sink) {
super.hasFlow(source, sink)
}
}

View File

@@ -0,0 +1,5 @@
import semmle.code.cpp.ir.dataflow.internal.TaintTrackingUtil as Public
module Private {
import semmle.code.cpp.ir.dataflow.DataFlow2::DataFlow2 as DataFlow
}

View File

@@ -0,0 +1,168 @@
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* Provides an implementation of global (interprocedural) taint tracking.
* This file re-exports the local (intraprocedural) taint-tracking analysis
* from `TaintTrackingParameter::Public` and adds a global analysis, mainly
* exposed through the `Configuration` class. For some languages, this file
* exists in several identical copies, allowing queries to use multiple
* `Configuration` classes that depend on each other without introducing
* mutual recursion among those configurations.
*/
import TaintTrackingParameter::Public
private import TaintTrackingParameter::Private
/**
* DEPRECATED: Use `Global` and `GlobalWithState` instead.
*
* A configuration of interprocedural taint tracking analysis. This defines
* sources, sinks, and any other configurable aspect of the analysis. Each
* use of the taint tracking library must define its own unique extension of
* this abstract class.
*
* A taint-tracking configuration is a special data flow configuration
* (`DataFlow::Configuration`) that allows for flow through nodes that do not
* necessarily preserve values but are still relevant from a taint tracking
* perspective. (For example, string concatenation, where one of the operands
* is tainted.)
*
* To create a configuration, extend this class with a subclass whose
* characteristic predicate is a unique singleton string. For example, write
*
* ```ql
* class MyAnalysisConfiguration extends TaintTracking::Configuration {
* MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
* // Override `isSource` and `isSink`.
* // Optionally override `isSanitizer`.
* // Optionally override `isSanitizerIn`.
* // Optionally override `isSanitizerOut`.
* // Optionally override `isSanitizerGuard`.
* // Optionally override `isAdditionalTaintStep`.
* }
* ```
*
* Then, to query whether there is flow between some `source` and `sink`,
* write
*
* ```ql
* exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
* ```
*
* Multiple configurations can coexist, but it is unsupported to depend on
* another `TaintTracking::Configuration` or a `DataFlow::Configuration` in the
* overridden predicates that define sources, sinks, or additional steps.
* Instead, the dependency should go to a `TaintTracking2::Configuration` or a
* `DataFlow2::Configuration`, `DataFlow3::Configuration`, etc.
*/
abstract deprecated class Configuration extends DataFlow::Configuration {
bindingset[this]
Configuration() { any() }
/**
* Holds if `source` is a relevant taint source.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source) { none() }
/**
* Holds if `source` is a relevant taint source with the given initial
* `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSource(DataFlow::Node source, DataFlow::FlowState state) { none() }
/**
* Holds if `sink` is a relevant taint sink
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink) { none() }
/**
* Holds if `sink` is a relevant taint sink accepting `state`.
*
* The smaller this predicate is, the faster `hasFlow()` will converge.
*/
// overridden to provide taint-tracking specific qldoc
override predicate isSink(DataFlow::Node sink, DataFlow::FlowState state) { none() }
/** Holds if the node `node` is a taint sanitizer. */
predicate isSanitizer(DataFlow::Node node) { none() }
final override predicate isBarrier(DataFlow::Node node) {
this.isSanitizer(node) or
defaultTaintSanitizer(node)
}
/**
* Holds if the node `node` is a taint sanitizer when the flow state is
* `state`.
*/
predicate isSanitizer(DataFlow::Node node, DataFlow::FlowState state) { none() }
final override predicate isBarrier(DataFlow::Node node, DataFlow::FlowState state) {
this.isSanitizer(node, state)
}
/** Holds if taint propagation into `node` is prohibited. */
predicate isSanitizerIn(DataFlow::Node node) { none() }
final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
/** Holds if taint propagation out of `node` is prohibited. */
predicate isSanitizerOut(DataFlow::Node node) { none() }
final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
*/
predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
this.isAdditionalTaintStep(node1, node2) or
defaultAdditionalTaintStep(node1, node2, _)
}
/**
* Holds if taint may propagate from `node1` to `node2` in addition to the normal data-flow and taint steps.
* This step is only applicable in `state1` and updates the flow state to `state2`.
*/
predicate isAdditionalTaintStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
none()
}
final override predicate isAdditionalFlowStep(
DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
DataFlow::FlowState state2
) {
this.isAdditionalTaintStep(node1, state1, node2, state2)
}
override predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
(
this.isSink(node) or
this.isSink(node, _) or
this.isAdditionalTaintStep(node, _) or
this.isAdditionalTaintStep(node, _, _, _)
) and
defaultImplicitTaintRead(node, c)
}
/**
* Holds if taint may flow from `source` to `sink` for this configuration.
*/
// overridden to provide taint-tracking specific qldoc
override predicate hasFlow(DataFlow::Node source, DataFlow::Node sink) {
super.hasFlow(source, sink)
}
}

View File

@@ -0,0 +1,5 @@
import semmle.code.cpp.ir.dataflow.internal.TaintTrackingUtil as Public
module Private {
import semmle.code.cpp.ir.dataflow.DataFlow3::DataFlow3 as DataFlow
}

View File

@@ -364,14 +364,10 @@ class TranslatedFunctionCall extends TranslatedCallExpr, TranslatedDirectCall {
final override predicate mayThrowException() {
expr.getTarget().(ThrowingFunction).mayThrowException(_)
or
expr.getTarget() instanceof AlwaysSehThrowingFunction
}
final override predicate mustThrowException() {
expr.getTarget().(ThrowingFunction).mayThrowException(true)
or
expr.getTarget() instanceof AlwaysSehThrowingFunction
}
}

View File

@@ -196,8 +196,6 @@ private predicate isInvalidFunction(Function func) {
expr.getEnclosingFunction() = func and
not exists(expr.getType())
)
or
count(func.getEntryPoint().getLocation()) > 1
}
/**

View File

@@ -49,4 +49,3 @@ private import implementations.PostgreSql
private import implementations.System
private import implementations.StructuredExceptionHandling
private import implementations.ZMQ
private import implementations.Win32CommandExecution

View File

@@ -16,7 +16,7 @@ import semmle.code.cpp.models.interfaces.NonThrowing
* `__builtin___memcpy_chk`.
*/
private class MemcpyFunction extends ArrayFunction, DataFlowFunction, SideEffectFunction,
AliasFunction, NonCppThrowingFunction
AliasFunction, NonThrowingFunction
{
MemcpyFunction() {
// memcpy(dest, src, num)

View File

@@ -11,7 +11,7 @@ import semmle.code.cpp.models.interfaces.SideEffect
import semmle.code.cpp.models.interfaces.NonThrowing
private class MemsetFunctionModel extends ArrayFunction, DataFlowFunction, AliasFunction,
SideEffectFunction, NonCppThrowingFunction
SideEffectFunction, NonThrowingFunction
{
MemsetFunctionModel() {
this.hasGlobalOrStdOrBslName("memset")

View File

@@ -6,6 +6,6 @@ import semmle.code.cpp.models.interfaces.NonThrowing
*
* Note: The `throw` specifier was deprecated in C++11 and removed in C++17.
*/
class NoexceptFunction extends NonCppThrowingFunction {
class NoexceptFunction extends NonThrowingFunction {
NoexceptFunction() { this.isNoExcept() or this.isNoThrow() }
}

View File

@@ -13,7 +13,7 @@ import semmle.code.cpp.models.interfaces.NonThrowing
/**
* The standard functions `printf`, `wprintf` and their glib variants.
*/
private class Printf extends FormattingFunction, AliasFunction, NonCppThrowingFunction {
private class Printf extends FormattingFunction, AliasFunction, NonThrowingFunction {
Printf() {
this instanceof TopLevelFunction and
(
@@ -37,7 +37,7 @@ private class Printf extends FormattingFunction, AliasFunction, NonCppThrowingFu
/**
* The standard functions `fprintf`, `fwprintf` and their glib variants.
*/
private class Fprintf extends FormattingFunction, NonCppThrowingFunction {
private class Fprintf extends FormattingFunction, NonThrowingFunction {
Fprintf() {
this instanceof TopLevelFunction and
(
@@ -55,7 +55,7 @@ private class Fprintf extends FormattingFunction, NonCppThrowingFunction {
/**
* The standard function `sprintf` and its Microsoft and glib variants.
*/
private class Sprintf extends FormattingFunction, NonCppThrowingFunction {
private class Sprintf extends FormattingFunction, NonThrowingFunction {
Sprintf() {
this instanceof TopLevelFunction and
(
@@ -98,9 +98,7 @@ private class Sprintf extends FormattingFunction, NonCppThrowingFunction {
/**
* Implements `Snprintf`.
*/
private class SnprintfImpl extends Snprintf, AliasFunction, SideEffectFunction,
NonCppThrowingFunction
{
private class SnprintfImpl extends Snprintf, AliasFunction, SideEffectFunction, NonThrowingFunction {
SnprintfImpl() {
this instanceof TopLevelFunction and
(
@@ -207,7 +205,7 @@ private class StringCchPrintf extends FormattingFunction {
/**
* The standard function `syslog`.
*/
private class Syslog extends FormattingFunction, NonCppThrowingFunction {
private class Syslog extends FormattingFunction, NonThrowingFunction {
Syslog() {
this instanceof TopLevelFunction and
this.hasGlobalName("syslog") and

View File

@@ -15,7 +15,7 @@ import semmle.code.cpp.models.interfaces.NonThrowing
* Does not include `strlcat`, which is covered by `StrlcatFunction`
*/
class StrcatFunction extends TaintFunction, DataFlowFunction, ArrayFunction, SideEffectFunction,
NonCppThrowingFunction
NonThrowingFunction
{
StrcatFunction() {
this.hasGlobalOrStdOrBslName([

View File

@@ -13,7 +13,7 @@ import semmle.code.cpp.models.interfaces.NonThrowing
* The standard function `strcpy` and its wide, sized, and Microsoft variants.
*/
class StrcpyFunction extends ArrayFunction, DataFlowFunction, TaintFunction, SideEffectFunction,
NonCppThrowingFunction
NonThrowingFunction
{
StrcpyFunction() {
this.hasGlobalOrStdOrBslName([

View File

@@ -1,7 +1,9 @@
import semmle.code.cpp.models.interfaces.Throwing
class WindowsDriverExceptionAnnotation extends AlwaysSehThrowingFunction {
WindowsDriverExceptionAnnotation() {
class WindowsDriverFunction extends ThrowingFunction {
WindowsDriverFunction() {
this.hasGlobalName(["RaiseException", "ExRaiseAccessViolation", "ExRaiseDatatypeMisalignment"])
}
final override predicate mayThrowException(boolean unconditional) { unconditional = true }
}

View File

@@ -1,56 +0,0 @@
private import semmle.code.cpp.models.interfaces.CommandExecution
/** The `ShellExecute` family of functions from Win32. */
class ShellExecute extends Function {
ShellExecute() { this.hasGlobalName("ShellExecute" + ["", "A", "W"]) }
}
private class ShellExecuteModel extends ShellExecute, CommandExecutionFunction {
override predicate hasCommandArgument(FunctionInput input) { input.isParameterDeref(2) }
}
/** The `WinExec` function from Win32. */
class WinExec extends Function {
WinExec() { this.hasGlobalName("WinExec") }
}
private class WinExecModel extends WinExec, CommandExecutionFunction {
override predicate hasCommandArgument(FunctionInput input) { input.isParameterDeref(0) }
}
/** The `CreateProcess` family of functions from Win32. */
class CreateProcess extends Function {
CreateProcess() { this.hasGlobalName("CreateProcess" + ["", "A", "W"]) }
}
private class CreateProcessModel extends CreateProcess, CommandExecutionFunction {
override predicate hasCommandArgument(FunctionInput input) { input.isParameterDeref(0) }
}
/** The `CreateProcessAsUser` family of functions from Win32. */
class CreateProcessAsUser extends Function {
CreateProcessAsUser() { this.hasGlobalName("CreateProcessAsUser" + ["", "A", "W"]) }
}
private class CreateProcessAsUserModel extends CreateProcessAsUser, CommandExecutionFunction {
override predicate hasCommandArgument(FunctionInput input) { input.isParameterDeref(1) }
}
/** The `CreateProcessWithLogonW` function from Win32. */
class CreateProcessWithLogonW extends Function {
CreateProcessWithLogonW() { this.hasGlobalName("CreateProcessWithLogonW") }
}
private class CreateProcessWithLogonModel extends CreateProcessWithLogonW, CommandExecutionFunction {
override predicate hasCommandArgument(FunctionInput input) { input.isParameterDeref(4) }
}
/** The `CreateProcessWithTokenW` function from Win32. */
class CreateProcessWithTokenW extends Function {
CreateProcessWithTokenW() { this.hasGlobalName("CreateProcessWithTokenW") }
}
private class CreateProcessWithTokenWModel extends CreateProcessWithTokenW, CommandExecutionFunction
{
override predicate hasCommandArgument(FunctionInput input) { input.isParameterDeref(2) }
}

View File

@@ -5,16 +5,7 @@
import semmle.code.cpp.Function
import semmle.code.cpp.models.Models
/**
* A function that is guaranteed to never throw a C++ exception
*
* The function may still raise a structured exception handling (SEH) exception.
*/
abstract class NonCppThrowingFunction extends Function { }
/**
* A function that is guaranteed to never throw.
*
* DEPRECATED: use `NonCppThrowingFunction` instead.
*/
deprecated class NonThrowingFunction = NonCppThrowingFunction;
abstract class NonThrowingFunction extends Function { }

View File

@@ -11,7 +11,7 @@ import semmle.code.cpp.models.Models
import semmle.code.cpp.models.interfaces.FunctionInputsAndOutputs
/**
* A function that is known to raise an exception.
* A class that models the exceptional behavior of a function.
*/
abstract class ThrowingFunction extends Function {
/**
@@ -20,8 +20,3 @@ abstract class ThrowingFunction extends Function {
*/
abstract predicate mayThrowException(boolean unconditional);
}
/**
* A function that unconditionally raises a structured exception handling (SEH) exception.
*/
abstract class AlwaysSehThrowingFunction extends Function { }

View File

@@ -46,22 +46,6 @@ compilation_args(
string arg : string ref
);
/**
* Optionally, record the build mode for each compilation.
*/
compilation_build_mode(
unique int id : @compilation ref,
int mode : int ref
);
/*
case @compilation_build_mode.mode of
0 = @build_mode_none
| 1 = @build_mode_manual
| 2 = @build_mode_auto
;
*/
/**
* The source files that are compiled by a compiler invocation.
* If `id` is for the compiler invocation

File diff suppressed because it is too large Load Diff

View File

@@ -1,2 +0,0 @@
description: Implement compilation_build_mode/2
compatibility: backwards

View File

@@ -1,47 +0,0 @@
/**
* @name Guarded Free
* @description NULL-condition guards before function calls to the memory-deallocation
* function free(3) are unnecessary, because passing NULL to free(3) is a no-op.
* @kind problem
* @problem.severity recommendation
* @precision very-high
* @id cpp/guarded-free
* @tags maintainability
* readability
*/
import cpp
import semmle.code.cpp.controlflow.Guards
class FreeCall extends FunctionCall {
FreeCall() { this.getTarget().hasGlobalName("free") }
}
predicate blockContainsPreprocessorBranches(BasicBlock bb) {
exists(PreprocessorBranch ppb, Location bbLoc, Location ppbLoc |
bbLoc = bb.(Stmt).getLocation() and ppbLoc = ppb.getLocation()
|
bbLoc.getFile() = ppb.getFile() and
bbLoc.getStartLine() < ppbLoc.getStartLine() and
ppbLoc.getEndLine() < bbLoc.getEndLine()
)
}
from GuardCondition gc, FreeCall fc, Variable v, BasicBlock bb
where
gc.ensuresEq(v.getAnAccess(), 0, bb, false) and
fc.getArgument(0) = v.getAnAccess() and
bb = fc.getBasicBlock() and
(
// No block statement: if (x) free(x);
bb = fc.getEnclosingStmt()
or
// Block statement with a single nested statement: if (x) { free(x); }
strictcount(bb.(BlockStmt).getAStmt()) = 1
) and
strictcount(BasicBlock bb2 | gc.ensuresEq(_, 0, bb2, _) | bb2) = 1 and
not fc.isInMacroExpansion() and
not blockContainsPreprocessorBranches(bb) and
not (gc instanceof BinaryOperation and not gc instanceof ComparisonOperation) and
not exists(CommaExpr c | c.getAChild*() = fc)
select gc, "unnecessary NULL check before call to $@", fc, "free"

View File

@@ -1,17 +1,3 @@
## 1.3.0
### New Queries
* Added a new high-precision quality query, `cpp/guarded-free`, which detects useless NULL pointer checks before calls to `free`. A variation of this query was originally contributed as an [experimental query by @mario-campos](https://github.com/github/codeql/pull/16331).
### Minor Analysis Improvements
* The "Call to function with fewer arguments than declared parameters" query (`cpp/too-few-arguments`) query no longer produces results if the function has been implicitly declared.
## 1.2.7
No user-facing changes.
## 1.2.6
### Minor Analysis Improvements

View File

@@ -8,7 +8,7 @@
<p>
This rule finds accesses through a pointer of a memory location that has already been freed (i.e. through a dangling pointer).
Such memory blocks have already been released to the dynamic memory manager, and modifying them can lead to anything
from a segfault to memory corruption that would cause subsequent calls to the dynamic memory manager to behave
from a segfault to memory corruption that would cause subsequent calls to the dynamic memory manger to behave
erratically, to a possible security vulnerability.
</p>

View File

@@ -51,7 +51,5 @@ predicate tooFewArguments(FunctionCall fc, Function f) {
hasDefiniteNumberOfParameters(fde)
|
fde.getNumberOfParameters() > fc.getNumberOfArguments()
) and
// Don't report on implicit function declarations, as these are likely extraction errors.
not f.getADeclarationEntry().isImplicit()
)
}

View File

@@ -19,6 +19,7 @@ import semmle.code.cpp.security.Security
import semmle.code.cpp.valuenumbering.GlobalValueNumbering
import semmle.code.cpp.ir.IR
import semmle.code.cpp.ir.dataflow.TaintTracking
import semmle.code.cpp.ir.dataflow.TaintTracking2
import semmle.code.cpp.security.FlowSources
import semmle.code.cpp.models.implementations.Strcat
import ExecTaint::PathGraph

View File

@@ -45,7 +45,7 @@ predicate deleteMayThrow(DeleteOrDeleteArrayExpr deleteExpr) {
* like it might throw an exception, and the function does not have a `noexcept` or `throw()` specifier.
*/
predicate functionMayThrow(Function f) {
not f instanceof NonCppThrowingFunction and
not f instanceof NonThrowingFunction and
(not exists(f.getBlock()) or stmtMayThrow(f.getBlock()))
}

View File

@@ -1,3 +0,0 @@
## 1.2.7
No user-facing changes.

View File

@@ -1,9 +0,0 @@
## 1.3.0
### New Queries
* Added a new high-precision quality query, `cpp/guarded-free`, which detects useless NULL pointer checks before calls to `free`. A variation of this query was originally contributed as an [experimental query by @mario-campos](https://github.com/github/codeql/pull/16331).
### Minor Analysis Improvements
* The "Call to function with fewer arguments than declared parameters" query (`cpp/too-few-arguments`) query no longer produces results if the function has been implicitly declared.

View File

@@ -1,2 +1,2 @@
---
lastReleaseVersion: 1.3.0
lastReleaseVersion: 1.2.6

View File

@@ -1,28 +1,18 @@
<!DOCTYPE qhelp PUBLIC
"-//Semmle//qhelp//EN"
"qhelp.dtd">
<!DOCTYPE qhelp SYSTEM "qhelp.dtd">
<qhelp>
<overview>
<p>The <code>free</code> function, which deallocates heap memory, may accept a NULL pointer and take no action. Therefore, it is unnecessary to check the argument for the value of NULL before a function call to <code>free</code>. As such, these guards may hinder performance and readability.</p>
<p>The <code>free</code> function, which deallocates heap memory, may accept a NULL pointer and take no action. Therefore, it is unnecessary to check its argument for the value of NULL before a function call to <code>free</code>. As such, these guards may hinder performance and readability.</p>
</overview>
<recommendation>
<p>A function call to <code>free</code> should not depend upon the value of its argument. Delete the condition preceding a function call to <code>free</code> when its only purpose is to check the value of the pointer to be freed.</p>
<p>A function call to <code>free</code> should not depend upon the value of its argument. Delete the <code>if</code> condition preceeding a function call to <code>free</code> when its only purpose is to check the value of the pointer to be freed.</p>
</recommendation>
<example>
<sample src = "GuardedFree.cpp" />
<p>In this example, the condition checking the value of <code>foo</code> can be deleted.</p>
</example>
<references>
<li>
The Open Group Base Specifications Issue 7, 2018 Edition:
<a href="https://pubs.opengroup.org/onlinepubs/9699919799/functions/free.html">free - free allocated memory</a>.
<a href="https://pubs.opengroup.org/onlinepubs/9699919799/functions/free.html">free - free allocated memory</a>
</li>
</references>
</qhelp>
</qhelp>

View File

@@ -0,0 +1,26 @@
/**
* @name Guarded Free
* @description NULL-condition guards before function calls to the memory-deallocation
* function free(3) are unnecessary, because passing NULL to free(3) is a no-op.
* @kind problem
* @problem.severity recommendation
* @precision very-high
* @id cpp/guarded-free
* @tags maintainability
* readability
* experimental
*/
import cpp
import semmle.code.cpp.controlflow.Guards
class FreeCall extends FunctionCall {
FreeCall() { this.getTarget().hasGlobalName("free") }
}
from GuardCondition gc, FreeCall fc, Variable v, BasicBlock bb
where
gc.ensuresEq(v.getAnAccess(), 0, bb, false) and
fc.getArgument(0) = v.getAnAccess() and
bb = fc.getEnclosingStmt()
select gc, "unnecessary NULL check before call to $@", fc, "free"

View File

@@ -1,5 +1,5 @@
name: codeql/cpp-queries
version: 1.3.0
version: 1.2.7-dev
groups:
- cpp
- queries

View File

@@ -17,6 +17,7 @@
| test.cpp:49:12:49:12 | Load: x | test.cpp:46:22:46:22 | ValueNumberBound | -1 | true | CompareLT: ... < ... | test.cpp:48:9:48:13 | test.cpp:48:9:48:13 |
| test.cpp:49:12:49:12 | Load: x | test.cpp:46:29:46:29 | ValueNumberBound | -2 | true | CompareLT: ... < ... | test.cpp:48:9:48:13 | test.cpp:48:9:48:13 |
| test.cpp:54:12:54:12 | Load: x | test.cpp:46:22:46:22 | ValueNumberBound | -1 | true | CompareLT: ... < ... | test.cpp:52:7:52:11 | test.cpp:52:7:52:11 |
| test.cpp:54:12:54:12 | Load: x | test.cpp:46:29:46:29 | ValueNumberBound | -2 | true | CompareLT: ... < ... | test.cpp:52:7:52:11 | test.cpp:52:7:52:11 |
| test.cpp:62:10:62:13 | Load: iter | test.cpp:60:17:60:17 | ValueNumberBound | 0 | false | NoReason | file://:0:0:0:0 | file://:0:0:0:0 |
| test.cpp:62:10:62:13 | Load: iter | test.cpp:60:17:60:17 | ValueNumberBound | 3 | true | CompareLT: ... < ... | test.cpp:61:32:61:51 | test.cpp:61:32:61:51 |
| test.cpp:62:10:62:13 | Load: iter | test.cpp:61:39:61:51 | ValueNumberBound | -1 | true | CompareLT: ... < ... | test.cpp:61:32:61:51 | test.cpp:61:32:61:51 |

Some files were not shown because too many files have changed in this diff Show More