analyzer: use value_range and range_op in eval_condition [PR98447]

PR analyzer/98447 covers a false positive from -fanalyzer.
The root cause is that the analyzer's constraint-handling code
doesn't "know" that if X := Y % 64, then X < 64.

This patch adds some minimal usage of value_range and range_op to
-fanalyzer.  It adds a new svalue::maybe_get_value_range vfunc, which
attempts to get a value_range for an svalue (but bails out for awkward
cases), and if value_ranges are available, uses them in eval_condition
to determine if the result is known based on range_op.

Doing so fixes the above false positive, improves a couple of other
tests in the DejaGnu testsuite, and eliminates 31 false +ves in the
analyzer integration testsuite (out of ~2200), along with a slight
speedup to the latter (albeit testing with a debug build of gcc).

A deeper integration with ranger could probably be made, but that is
clearly stage 1 material.

gcc/analyzer/ChangeLog:
	PR analyzer/98447
	* common.h: Include "value-range.h".
	* region-model.cc: Include "value-relation.h" and "range-op.h".
	(region_model::eval_condition): Try using range_op to see if we
	can get a known boolean from the value_ranges of the operands.
	* svalue.cc: Include "value-relation.h" and "range-op.h".
	(type_can_have_value_range_p): New.
	(svalue::maybe_get_value_range): New.
	(constant_svalue::maybe_get_value_range): New.
	(unknown_svalue::maybe_get_value_range): New.
	(unaryop_svalue::maybe_get_value_range): New.
	(binop_svalue::maybe_get_value_range): New.
	* svalue.h (svalue::maybe_get_value_range): New vfunc decl.
	(constant_svalue::maybe_get_value_range): New decl.
	(unknown_svalue::maybe_get_value_range): New decl.
	(unaryop_svalue::maybe_get_value_range): New decl.
	(binop_svalue::maybe_get_value_range): New decl.

gcc/testsuite/ChangeLog:
	PR analyzer/98447
	* c-c++-common/analyzer/conditionals-pr98447-1.c: New test.
	* c-c++-common/analyzer/conditionals-pr98447-2.c: New test.
	* c-c++-common/analyzer/null-deref-pr108400-SoftEtherVPN-WebUi.c:
	Updated for false positive being fixed for C.
	* gcc.dg/analyzer/data-model-1.c: Update expected output to
	reflect improved output.

Signed-off-by: David Malcolm <dmalcolm@redhat.com>

0 files changed, 0 insertions, 0 deletions