aboutsummaryrefslogtreecommitdiffstats
path: root/lib/compiler/src
diff options
context:
space:
mode:
authorBjörn Gustavsson <[email protected]>2018-06-04 06:14:19 +0200
committerBjörn Gustavsson <[email protected]>2018-06-04 10:41:21 +0200
commit7eb06ed5ac1687d38245db2e0aef2756cb43b1ae (patch)
tree8f796f9c5bb2ec8f059a9ba7ef37aea9fd9a6cda /lib/compiler/src
parent7f3a501cada228c2fedbe34d0d30080e98560665 (diff)
downloadotp-7eb06ed5ac1687d38245db2e0aef2756cb43b1ae.tar.gz
otp-7eb06ed5ac1687d38245db2e0aef2756cb43b1ae.tar.bz2
otp-7eb06ed5ac1687d38245db2e0aef2756cb43b1ae.zip
sys_core_fold: Fix name capture problem
sys_core_fold could do unsafe transformations on the code from the old inliner (invoked using the compiler option `{inline,[{F/A}]}` to request inlining of specific functions). To explain the bug, let's first look at an example that sys_core_fold handles correctly. Consider this code: 'foo'/2 = fun (Arg1,Arg2) -> let <B> = Arg2 in let <A,B> = <B,Arg1> in {A,B} In this example, the lets can be completely eliminated, since the arguments for the lets are variables (as opposed to expressions). Since the variable B is rebound in the inner let, `sys_core_fold` must take special care when doing the substitutions. Here is the correct result: 'foo'/2 = fun (Arg1, Arg2) -> {Arg2,Arg1} Consider a slight modifictation of the example: 'bar'/2 = fun (Arg1,Arg2) -> let <B> = [Arg2] in let <A,B> = <B,[Arg1]> in {A,B} Here some of the arguments for the lets are expressions, so the lets must be kept. sys_core_fold does not handle this example correctly: 'bar'/2 = fun (Arg1,Arg2) -> let <B> = [Arg2] in let <B> = [Arg1] in {B,B} In the inner let, the variable A has been eliminated and replaced with the variable B in the body (the first B in the tuple). Since the B in the outer let is never used, the outer let will be eliminated, giving: 'bar'/2 = fun (Arg1,Arg2) -> let <B> = [Arg1] in {B,B} To handle this example correctly, sys_core_fold must rename the variable B in the inner let like this to avoid capturing B: 'bar'/2 = fun (Arg1,Arg2) -> let <B> = [Arg2] in let <NewName> = [Arg1] in {B,NewName} (Note: The `v3_kernel` pass alreday handles those examples correctly in case `sys_core_fold` has been disabled.)
Diffstat (limited to 'lib/compiler/src')
-rw-r--r--lib/compiler/src/sys_core_fold.erl37
1 files changed, 11 insertions, 26 deletions
diff --git a/lib/compiler/src/sys_core_fold.erl b/lib/compiler/src/sys_core_fold.erl
index a13bdedaf9..47042c2393 100644
--- a/lib/compiler/src/sys_core_fold.erl
+++ b/lib/compiler/src/sys_core_fold.erl
@@ -1275,13 +1275,18 @@ let_subst_list([], [], _) -> {[],[],[]}.
%%pattern(Pat, Sub) -> pattern(Pat, Sub, Sub).
pattern(#c_var{}=Pat, Isub, Osub) ->
- case sub_is_val(Pat, Isub) of
+ case sub_is_in_scope(Pat, Isub) of
true ->
+ %% This variable either has a substitution or is used in
+ %% the variable list of an enclosing `let`. In either
+ %% case, it must be renamed to an unused name to avoid
+ %% name capture problems.
V1 = make_var_name(),
Pat1 = #c_var{name=V1},
{Pat1,sub_set_var(Pat, Pat1, sub_add_scope([V1], Osub))};
false ->
- {Pat,sub_del_var(Pat, Osub)}
+ %% This variable has never been used. Add it to the scope.
+ {Pat,sub_add_scope([Pat#c_var.name], Osub)}
end;
pattern(#c_literal{}=Pat, _, Osub) -> {Pat,Osub};
pattern(#c_cons{anno=Anno,hd=H0,tl=T0}, Isub, Osub0) ->
@@ -1460,8 +1465,8 @@ is_subst(_) -> false.
%% sub_set_name(Name, Value, #sub{}) -> #sub{}.
%% sub_del_var(Var, #sub{}) -> #sub{}.
%% sub_subst_var(Var, Value, #sub{}) -> [{Name,Value}].
-%% sub_is_val(Var, #sub{}) -> boolean().
-%% sub_add_scope(#sub{}) -> #sub{}
+%% sub_is_in_scope(Var, #sub{}) -> boolean().
+%% sub_add_scope([Var], #sub{}) -> #sub{}
%% sub_subst_scope(#sub{}) -> #sub{}
%%
%% We use the variable name as key so as not have problems with
@@ -1496,18 +1501,6 @@ sub_set_name(V, Val, #sub{v=S,s=Scope,t=Tdb0}=Sub) ->
Tdb = copy_type(V, Val, Tdb1),
Sub#sub{v=orddict:store(V, Val, S),s=cerl_sets:add_element(V, Scope),t=Tdb}.
-sub_del_var(#c_var{name=V}, #sub{v=S,s=Scope,t=Tdb}=Sub) ->
- %% Profiling shows that for programs with many record operations,
- %% sub_del_var/2 is a bottleneck. Since the scope contains all
- %% variables that are live, we know that V cannot be present in S
- %% if it is not in the scope.
- case cerl_sets:is_element(V, Scope) of
- false ->
- Sub#sub{s=cerl_sets:add_element(V, Scope)};
- true ->
- Sub#sub{v=orddict:erase(V, S),t=kill_types(V, Tdb)}
- end.
-
sub_subst_var(#c_var{name=V}, Val, #sub{v=S0}) ->
%% Fold chained substitutions.
[{V,Val}] ++ [ {K,Val} || {K,#c_var{name=V1}} <- S0, V1 =:= V].
@@ -1533,16 +1526,8 @@ sub_subst_scope_1([H|T], Key, Acc) ->
sub_subst_scope_1(T, Key-1, [{Key,#c_var{name=H}}|Acc]);
sub_subst_scope_1([], _, Acc) -> Acc.
-sub_is_val(#c_var{name=V}, #sub{v=S,s=Scope}) ->
- %% When the bottleneck in sub_del_var/2 was eliminated, this
- %% became the new bottleneck. Since the scope contains all
- %% live variables, a variable V can only be the target for
- %% a substitution if it is in the scope.
- cerl_sets:is_element(V, Scope) andalso v_is_value(V, S).
-
-v_is_value(Var, [{_,#c_var{name=Var}}|_]) -> true;
-v_is_value(Var, [_|T]) -> v_is_value(Var, T);
-v_is_value(_, []) -> false.
+sub_is_in_scope(#c_var{name=V}, #sub{s=Scope}) ->
+ cerl_sets:is_element(V, Scope).
%% warn_no_clause_match(CaseOrig, CaseOpt) -> ok
%% Generate a warning if none of the user-specified clauses