Implement Concurrency6 package -- split out Concurrency7 #791
Conversation
| not isExcluded(expr, Concurrency6Package::atomicAggregateObjectDirectlyAccessedQuery()) and | ||
| not expr.isUnevaluated() and | ||
| ( | ||
| exists(FieldAccess fa | |
There was a problem hiding this comment.
Are the atomic_.. access functions alway guaranteed to be function calls and not macros? Could we get false positives for this query if a standard library decided to implement a macro which did access a field? In general the C library can be implemented with functions or macros. Or, in practice do we expect all standard libraries to implement this using a function call?
There was a problem hiding this comment.
I think we don't have to worry about this in practice, for a few reasons. Here's part 2 of a very interesting blog post from RedHat about locking vs lock free atomics, and some of the features in atomic.h.
Let's now see why the macro wouldn't be useful ... since atomic types aren't structs but basic types, the macro would ... employ some compiler magic to access the hidden mutex and value members of the struct
It seems all current atomic implementations are lock free, and a locking implementation may never exist. It's easier to make lock free atomic pointers. Additionally, an atomic s is of type s so something like s->mutex would require extra effort to support in the language, and implementations would probably instead use a magic builtin compiler function e.g. __get_mutex(s) than add support for a magic s->mutex on an int type, etc.
I'm happy to make this more defensive, though, so it doesn't bite us.
| ) | ||
| or | ||
| // Everything else: not a memory order constant or an integer valued literal, also exclude | ||
| // variables and functions, things that flow further back. |
There was a problem hiding this comment.
What else would this be? And if we cannot determine the value for it, presumably we can't tell whether it's in violation or not?
| atomic_s1.x; // NON-COMPLIANT | ||
| ptr_atomic_s1->x; // NON-COMPLIANT | ||
| atomic_s1.x = 0; // NON-COMPLIANT | ||
| ptr_atomic_s1->x = 0; // NON-COMPLIANT |
There was a problem hiding this comment.
Can we add some user macro-wrapped cases?
…currency6-package
…ackage' into michaelrfairhurst/concurrency6-package
Description
Implement Concurrency6 package -- split out Concurrency7.
I had added support for
pthread_create, and left it in. Should I enable support for pthreads in the control flow graph inConcurrency.qll?Change request type
.ql,.qll,.qlsor unit tests)Rules with added or modified queries
DIR-5-2DIR-5-3RULE-12-6RULE-21-25RULE-22-11Release change checklist
A change note (development_handbook.md#change-notes) is required for any pull request which modifies:
If you are only adding new rule queries, a change note is not required.
Author: Is a change note required?
🚨🚨🚨
Reviewer: Confirm that format of shared queries (not the .qll file, the
.ql file that imports it) is valid by running them within VS Code.
Reviewer: Confirm that either a change note is not required or the change note is required and has been added.
Query development review checklist
For PRs that add new queries or modify existing queries, the following checklist should be completed by both the author and reviewer:
Author
As a rule of thumb, predicates specific to the query should take no more than 1 minute, and for simple queries be under 10 seconds. If this is not the case, this should be highlighted and agreed in the code review process.
Reviewer
As a rule of thumb, predicates specific to the query should take no more than 1 minute, and for simple queries be under 10 seconds. If this is not the case, this should be highlighted and agreed in the code review process.