public inbox for gcc-cvs@sourceware.org help / color / mirror / Atom feed
From: David Malcolm <dmalcolm@gcc.gnu.org> To: gcc-cvs@gcc.gnu.org Subject: [gcc r12-8169] analyzer: fix escaping of pointer arithmetic [PR105264] Date: Thu, 14 Apr 2022 22:43:43 +0000 (GMT) [thread overview] Message-ID: <20220414224343.0C71F385803E@sourceware.org> (raw) https://gcc.gnu.org/g:a358e4b60815b41e27f3508014ceb592f86b9b45 commit r12-8169-ga358e4b60815b41e27f3508014ceb592f86b9b45 Author: David Malcolm <dmalcolm@redhat.com> Date: Thu Apr 14 09:52:00 2022 -0400 analyzer: fix escaping of pointer arithmetic [PR105264] PR analyzer/105264 reports that the analyzer can fail to treat (PTR + IDX) and PTR[IDX] as referring to the same memory under some situations. There are various ways in which this can happen when IDX is a symbolic value, due to having several ways in which such memory regions can be referred to symbolically. I attempted to fix this by being smarter when folding svalues and regions, but this fix seems too fiddly to attempt in stage 4. Instead, this less ambitious patch fixes a false positive from -Wanalyzer-use-of-uninitialized-value by making the analyzer's escape analysis smarter, so that it treats *PTR as escaping when (PTR + OFFSET) is passed to an external function, and thus it treats *PTR as possibly-initialized (the "passing &PTR[IDX]" case was already working). gcc/analyzer/ChangeLog: PR analyzer/105264 * region-model-reachability.cc (reachable_regions::handle_parm): Use maybe_get_deref_base_region rather than just region_svalue, to handle pointer arithmetic also. * svalue.cc (svalue::maybe_get_deref_base_region): New. * svalue.h (svalue::maybe_get_deref_base_region): New decl. gcc/testsuite/ChangeLog: PR analyzer/105264 * gcc.dg/analyzer/torture/symbolic-10.c: New test. Signed-off-by: David Malcolm <dmalcolm@redhat.com> Diff: --- gcc/analyzer/region-model-reachability.cc | 8 ++--- gcc/analyzer/svalue.cc | 42 ++++++++++++++++++++++ gcc/analyzer/svalue.h | 2 ++ .../gcc.dg/analyzer/torture/symbolic-10.c | 40 +++++++++++++++++++++ 4 files changed, 86 insertions(+), 6 deletions(-) diff --git a/gcc/analyzer/region-model-reachability.cc b/gcc/analyzer/region-model-reachability.cc index b876b8f0363..12d09c3e500 100644 --- a/gcc/analyzer/region-model-reachability.cc +++ b/gcc/analyzer/region-model-reachability.cc @@ -252,12 +252,8 @@ reachable_regions::handle_parm (const svalue *sval, tree param_type) m_mutable_svals.add (sval); else m_reachable_svals.add (sval); - if (const region_svalue *parm_ptr - = sval->dyn_cast_region_svalue ()) - { - const region *pointee_reg = parm_ptr->get_pointee (); - add (pointee_reg, is_mutable); - } + if (const region *base_reg = sval->maybe_get_deref_base_region ()) + add (base_reg, is_mutable); /* Treat all svalues within a compound_svalue as reachable. */ if (const compound_svalue *compound_sval = sval->dyn_cast_compound_svalue ()) diff --git a/gcc/analyzer/svalue.cc b/gcc/analyzer/svalue.cc index 536bc288dbf..a1403f0fbef 100644 --- a/gcc/analyzer/svalue.cc +++ b/gcc/analyzer/svalue.cc @@ -651,6 +651,48 @@ svalue::all_zeroes_p () const return false; } +/* If this svalue is a pointer, attempt to determine the base region it points + to. Return NULL on any problems. */ + +const region * +svalue::maybe_get_deref_base_region () const +{ + const svalue *iter = this; + while (1) + { + switch (iter->get_kind ()) + { + default: + return NULL; + + case SK_REGION: + { + const region_svalue *region_sval + = as_a <const region_svalue *> (iter); + return region_sval->get_pointee ()->get_base_region (); + } + + case SK_BINOP: + { + const binop_svalue *binop_sval + = as_a <const binop_svalue *> (iter); + switch (binop_sval->get_op ()) + { + case POINTER_PLUS_EXPR: + /* If we have a symbolic value expressing pointer arithmetic, + use the LHS. */ + iter = binop_sval->get_arg0 (); + continue; + + default: + return NULL; + } + return NULL; + } + } + } +} + /* class region_svalue : public svalue. */ /* Implementation of svalue::dump_to_pp vfunc for region_svalue. */ diff --git a/gcc/analyzer/svalue.h b/gcc/analyzer/svalue.h index 4bbe8588b8d..29ea2ee6408 100644 --- a/gcc/analyzer/svalue.h +++ b/gcc/analyzer/svalue.h @@ -175,6 +175,8 @@ public: per-type and thus it's meaningless for them to "have state". */ virtual bool can_have_associated_state_p () const { return true; } + const region *maybe_get_deref_base_region () const; + protected: svalue (complexity c, tree type) : m_complexity (c), m_type (type) diff --git a/gcc/testsuite/gcc.dg/analyzer/torture/symbolic-10.c b/gcc/testsuite/gcc.dg/analyzer/torture/symbolic-10.c new file mode 100644 index 00000000000..b2f3a8a1d86 --- /dev/null +++ b/gcc/testsuite/gcc.dg/analyzer/torture/symbolic-10.c @@ -0,0 +1,40 @@ +/* Verify that -fanalyzer considers that mmfs escapes when passing either: + *(mmfs + i) + and + (&mmfs[i]) + to an external function (for symbolic i). */ + +typedef struct s_mmfile { + char *ptr; + long size; +} mmfile_t; + +void init_mmfile(mmfile_t *ptr); + +long test__init_via_ptr_arith__read_via_array_idx(int i) +{ + mmfile_t mmfs[3]; + init_mmfile(mmfs + i); + return mmfs[i].size; /* { dg-bogus "uninit" } */ +} + +long test__init_via_array_idx__read_via_ptr_arith(int i) +{ + mmfile_t mmfs[3]; + init_mmfile(&mmfs[i]); + return (mmfs + i)->size; /* { dg-bogus "uninit" } */ +} + +long test__ptr_arith_for_both(int i) +{ + mmfile_t mmfs[3]; + init_mmfile(mmfs + i); + return (mmfs + i)->size; /* { dg-bogus "uninit" } */ +} + +long test__array_idx_for_both(int i) +{ + mmfile_t mmfs[3]; + init_mmfile(&mmfs[i]); + return mmfs[i].size; /* { dg-bogus "uninit" } */ +}
reply other threads:[~2022-04-14 22:43 UTC|newest] Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20220414224343.0C71F385803E@sourceware.org \ --to=dmalcolm@gcc.gnu.org \ --cc=gcc-cvs@gcc.gnu.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).