This is the mail archive of the systemtap@sourceware.org mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug translator/16974] New: excessive context size after {stmt;}->stmt; optimization


https://sourceware.org/bugzilla/show_bug.cgi?id=16974

            Bug ID: 16974
           Summary: excessive context size after {stmt;}->stmt;
                    optimization
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: translator
          Assignee: systemtap at sourceware dot org
          Reporter: fche at redhat dot com

When dead_stmtexpr_remover::visit_block unwraps statements
from unnecessary { } nesting levels, it has an unintended
consequence on context memory allocation.  Namely,
c_tmpcounter::visit_* only special cases the visit_block
type for special handling, wherein distinct statements
may have overlapping temporaries.  So for constructs such
as


if (foo=="1") foreach (a in b) ... 
else if (foo=="2") foreach (c in d) ...
else ...

all the temporary variables for all blocks (now nested
singleton statements) are expanded as siblings in the
function context struct.

Instead, all of these tree-like staptree nodes (if_statement,
for*_statement, perhaps others) should do what
c_tmpcounter::visit_block does, namely wrap individual
subtree branches into a common anonymous union {}.

-- 
You are receiving this mail because:
You are the assignee for the bug.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]