| Age | Commit message (Collapse) | Author |
|---|
|
functions as apart of initial analysis
This is off by default for the obvious reasons, we may also in the future want to expand upon this with a "matched view" such that the first step a user can take after loading is to validate what functions were matched and decide if they want to keep any of that data.
This fetch happens as apart of the initial analysis, such that a headless script will not return from the load function until it has finished fetching. We may want to add a configurable timeout, and/or run the fetch out of band (let load return before fetching finishes).
|
|
|
|
|
|
- Respects views fetch batch size and allowed tags
- Saves and loads from the view settings instead of qt settings
|
|
This then allows us to group the settings into more understandable subgroups such as fetcher & matcher
|
|
This fixes an issue where the commit dialog would not fill in the newly added network source
|
|
This helps when you have analyzed large binaries and want to create signatures after-the-fact, skipping analysis. This can speed up the time to create large datasets drastically.
|
|
Fixes https://github.com/Vector35/binaryninja-api/issues/7268
|
|
Also makes it so that constraints with an exact offset match are displayed differently.
|
|
|
|
|
|
This makes them immediately available, and works around the issue of invalidating the entire cache to allow for retrieving of locally pushed data
|
|
|
|
|
|
- Improve error messages to include what exactly caused the parsing to fail
- Relaxed the parsing validation to allow for more files to be parsed
- Added a form dialog to improve discoverability of load settings
We still need to relax optional fields failing to parse if they have no child, as that is considered an error, even with no validation.
|
|
operations
A new activity is added that will remove calls to `objc_retain`,
`objc_release`, `objc_autorelease`, and related functions. It is
disabled by default due to the fact it changes the semantics of the
code. It can be enabled on a per-function basis via the Function
Settings context menu or command palette entry. Alternatively, it can be
enabled in Open with Options or in user settings if a user would prefer
it be on for an entire file or for all files they open.
For now the activity is only eligible within arm64 binaries. Supporting
x86_64 will require matching some slightly different IL patterns.
|
|
|
|
|
|
|
|
LowLevelILFunction objects can be constructed without a owner
function. BNGetLowLevelILOwnerFunction can return NULL and must be
checked before creating a Function object.
|
|
|
|
|
|
Calls to `objc_msgSendSuper2` with a selector in the `init` family are
detected and their arguments are analyzed to determine the receiver
class. A call type adjustment is added to update the return type.
This handles most calls to `[super init]` from Objective-C code as the
compiler generates a straightforward code sequence for these calls. Some
calls from Swift are handled, but the Swift compiler generates more varied
code so additional work is needed for complete coverage.
|
|
implementations
|
|
|
|
|
|
Improves the speed by moving chunks of functions into worker threads, because of how the functions and possible functions are gathered we have many locations to insert fast fails, which is also partially addressed by this commit (see `maximum_possible_functions`).
|
|
of all the functions
This was caused by the internal flatbuffer verifier reaching the max number of tables, this is fine and the solution is to split the large chunk into smaller ones, effectively creating more flatbuffer "views" of tables.
This is fine from a performance perspective because we already have optimized for many small chunks (e.g. we have a lookup table stored next to each chunk).
I left some comments for future improvements but this should be good now.
Also alongside this change we improved generating performance by 2x on some pathological binaries.
|
|
It will just end up freezing the UI for an unreasonable time, just open the generated report in a real web browser or text editor, QTextBrowser falls over and fixing it would be more effort than its worth.
|
|
given
Fixes the case where we are generating many smaller chunks for a given view and then unintentionally merging
them back together, throwing away all data.
|
|
It is important that we only retrieve the medium-level IL if the function has
any user-defined variables, otherwise, we will possibly be generating MLIL for no reason.
For the above reason, we do a filter on user-defined variables first.
|
|
We do not need to consult the lifted IL if we have already cached the function GUID in the function metadata
|
|
This is functionally equivalent to the previous workflow_objc, with the
following changes:
1. It mutates the `core.function.metaAnalysis` workflow rather than
registering a new named workflow. The activities now all check for
the presence of the Objective-C metadata added by `ObjCProcessor` to
determine whether they should do work, rather than relying on
`MachoView` to override the function workflow when Objective-C
metadata is present. This fixes
https://github.com/Vector35/binaryninja-api/issues/6779.
2. The auto-inlining of `objc_msgSend` selector stub functions is
performed in a separate activity from the processing of
`objc_msgSend` call sites. The selector stub inlining activity is
configured so that it does not run in `DSCView` as the shared cache
needs different behavior for stub functions more generally that
`SharedCacheWorkflow` already provides.
3. The way that types like `id` and `SEL` are referenced is fixed so
that they show up as `id` rather than `objc_struct*`.
This also replaces the Objective-C portion of the shared cache's
workflow, and incorporates several bug fixes that had been applied to it
but not the standalone Objective-C workflow.
|
|
projects
|
|
|
|
Calls to `Workflow::Instance` that were looking up a built-in workflow
name are updated to use `Workflow::Get`. Others use `Workflow::GetOrCreate`.
|
|
|
|
C++ plug-ins now consistently use the `plugin_rpath` or
`ui_plugin_rpath` macros to ensure they have `SKIP_BUILD_RPATH` set when
building on Mac (i.e., no `LC_RPATH` is added).
Rust plug-ins have their build.rs updated to only specify `-Wl,-rpath`
when building for Linux. It is not needed on macOS. On macOS we instead
explicitly specify an `@rpath`-relative install name. This doesn't
change any behavior, but avoids leaving an absolute path as the
library's install name and is consistent with CMake's behavior for C++
plug-ins.
|
|
|
|
|
|
`Workflow::new` is replaced by `Workflow::build` that returns a
`Builder` type that supports the operations that mutate a workflow, such
as registering activities.
`Workflow::instance` is replaced by `Workflow::get` to make clear that
it is intended to look up an existing workflow.
`Workflow::cloned` is introduced to wrap the common pattern of
retrieving an existing workflow and cloning it with the same name in
order to modify it.
`Builder::activity_before` / `Builder::activity_after` are introduced to
wrap the common pattern of registering an activity then inserting it
before or after a given activity.
|
|
|
|
inserted
Instead of applying symbols and comments in the applier step, we will do it when the matched function is identified.
This has the side effect that if you turn off the apply activity names and comments will still be applied, more work to be done later.
|
|
|
|
|
|
message
|
|
This ensures that the Python source files are only generated and copied
into the output directory if inputs have changed, rather than being done
unconditionally.
|
|
This ensures that CMake detects when files that match the glob are added
or removed.
|
|
|
|
|
