| Message ID | b9c26ef6-a5fa-717c-98bd-cdcadaf0107d@gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | avoid ice due to inconsistent argument types to fold_build (PR 90662) | expand |
On 6/13/19 1:10 PM, Martin Sebor wrote: > Attached is a fix for the fold_build call with inconsistent > argument types introduced in a recent commit of mine. > > Tested on x86_64-linux. > > Martin > > gcc-90662.diff > > PR tree-optimization/90662 - strlen of a string in a vla plus offset not folded > > gcc/ChangeLog: > > PR tree-optimization/90662 > * tree-ssa-strlen.c (get_stridx): Convert fold_build2 operands > to the same type. > > gcc/testsuite/ChangeLog: > > PR tree-optimization/90662 > * gcc.dg/pr90866-2.c: New test. > * gcc.dg/pr90866.c: Ditto. OK jeff
On Fri, 14 Jun 2019 at 03:35, Jeff Law <law@redhat.com> wrote: > > On 6/13/19 1:10 PM, Martin Sebor wrote: > > Attached is a fix for the fold_build call with inconsistent > > argument types introduced in a recent commit of mine. > > > > Tested on x86_64-linux. > > > > Martin > > > > gcc-90662.diff > > > > PR tree-optimization/90662 - strlen of a string in a vla plus offset not folded > > > > gcc/ChangeLog: > > > > PR tree-optimization/90662 > > * tree-ssa-strlen.c (get_stridx): Convert fold_build2 operands > > to the same type. > > > > gcc/testsuite/ChangeLog: > > > > PR tree-optimization/90662 > > * gcc.dg/pr90866-2.c: New test. > > * gcc.dg/pr90866.c: Ditto. > OK > jeff Hi, I've noticed that pr90866-2.c fails on arm-none-eabi: /gcc/testsuite/gcc.dg/pr90866-2.c:17:5: error: conflicting types for 'i' /gcc/testsuite/gcc.dg/pr90866-2.c:17:1: note: an argument type that has a default promotion cannot match an empty parameter name list declaration /gcc/testsuite/gcc.dg/pr90866-2.c:16:5: note: previous declaration of 'i' was here compiler exited with status 1 FAIL: gcc.dg/pr90866-2.c (test for excess errors) Removing 'int i();' makes the test pass, but I'm wondering why is passes on other targets without this change? Christophe
On 6/18/19 2:38 AM, Christophe Lyon wrote: > On Fri, 14 Jun 2019 at 03:35, Jeff Law <law@redhat.com> wrote: >> >> On 6/13/19 1:10 PM, Martin Sebor wrote: >>> Attached is a fix for the fold_build call with inconsistent >>> argument types introduced in a recent commit of mine. >>> >>> Tested on x86_64-linux. >>> >>> Martin >>> >>> gcc-90662.diff >>> >>> PR tree-optimization/90662 - strlen of a string in a vla plus offset not folded >>> >>> gcc/ChangeLog: >>> >>> PR tree-optimization/90662 >>> * tree-ssa-strlen.c (get_stridx): Convert fold_build2 operands >>> to the same type. >>> >>> gcc/testsuite/ChangeLog: >>> >>> PR tree-optimization/90662 >>> * gcc.dg/pr90866-2.c: New test. >>> * gcc.dg/pr90866.c: Ditto. >> OK >> jeff > > Hi, > > I've noticed that pr90866-2.c fails on arm-none-eabi: > /gcc/testsuite/gcc.dg/pr90866-2.c:17:5: error: conflicting types for 'i' > /gcc/testsuite/gcc.dg/pr90866-2.c:17:1: note: an argument type that > has a default promotion cannot match an empty parameter name list > declaration > /gcc/testsuite/gcc.dg/pr90866-2.c:16:5: note: previous declaration of > 'i' was here > compiler exited with status 1 > FAIL: gcc.dg/pr90866-2.c (test for excess errors) > > Removing 'int i();' makes the test pass, but I'm wondering why is > passes on other targets without this change? I'm sure it's because of the difference in type promotion rules on the target. The function declaration shouldn't have an impact on the test case so I have removed it in r272418. Thanks Martin
On Tue, 18 Jun 2019 at 15:07, Martin Sebor <msebor@gmail.com> wrote: > > On 6/18/19 2:38 AM, Christophe Lyon wrote: > > On Fri, 14 Jun 2019 at 03:35, Jeff Law <law@redhat.com> wrote: > >> > >> On 6/13/19 1:10 PM, Martin Sebor wrote: > >>> Attached is a fix for the fold_build call with inconsistent > >>> argument types introduced in a recent commit of mine. > >>> > >>> Tested on x86_64-linux. > >>> > >>> Martin > >>> > >>> gcc-90662.diff > >>> > >>> PR tree-optimization/90662 - strlen of a string in a vla plus offset not folded > >>> > >>> gcc/ChangeLog: > >>> > >>> PR tree-optimization/90662 > >>> * tree-ssa-strlen.c (get_stridx): Convert fold_build2 operands > >>> to the same type. > >>> > >>> gcc/testsuite/ChangeLog: > >>> > >>> PR tree-optimization/90662 > >>> * gcc.dg/pr90866-2.c: New test. > >>> * gcc.dg/pr90866.c: Ditto. > >> OK > >> jeff > > > > Hi, > > > > I've noticed that pr90866-2.c fails on arm-none-eabi: > > /gcc/testsuite/gcc.dg/pr90866-2.c:17:5: error: conflicting types for 'i' > > /gcc/testsuite/gcc.dg/pr90866-2.c:17:1: note: an argument type that > > has a default promotion cannot match an empty parameter name list > > declaration > > /gcc/testsuite/gcc.dg/pr90866-2.c:16:5: note: previous declaration of > > 'i' was here > > compiler exited with status 1 > > FAIL: gcc.dg/pr90866-2.c (test for excess errors) > > > > Removing 'int i();' makes the test pass, but I'm wondering why is > > passes on other targets without this change? > > I'm sure it's because of the difference in type promotion rules > on the target. OK, it's surprising it work on arm-linux* and not on arm-none-eabi? > The function declaration shouldn't have an impact > on the test case so I have removed it in r272418. OK thanks > > Thanks > Martin
PR tree-optimization/90662 - strlen of a string in a vla plus offset not folded gcc/ChangeLog: PR tree-optimization/90662 * tree-ssa-strlen.c (get_stridx): Convert fold_build2 operands to the same type. gcc/testsuite/ChangeLog: PR tree-optimization/90662 * gcc.dg/pr90866-2.c: New test. * gcc.dg/pr90866.c: Ditto. Index: gcc/testsuite/gcc.dg/pr90866-2.c =================================================================== --- gcc/testsuite/gcc.dg/pr90866-2.c (nonexistent) +++ gcc/testsuite/gcc.dg/pr90866-2.c (working copy) @@ -0,0 +1,24 @@ +/* PR tree-optimization/90866 - ICE in fold_binary_loc, at fold-const.c:9827 + { dg-do compile } + { dg-options "-O2 -fsanitize=thread" } */ + +typedef enum { a } b; +typedef struct { + int c[0]; +} d; +typedef struct { + int *data; +} e; +typedef struct { + e buffer; +} f; +int g, h; +int i(); +int i(f *j, d *k, b l, int m) { + if (l) + if (m) { + h = j->buffer.data[0]; + k->c[g] = k->c[g] * 8; + } + return 0; +} Index: gcc/testsuite/gcc.dg/pr90866.c =================================================================== --- gcc/testsuite/gcc.dg/pr90866.c (nonexistent) +++ gcc/testsuite/gcc.dg/pr90866.c (working copy) @@ -0,0 +1,18 @@ +/* PR tree-optimization/90866 - ICE in fold_binary_loc, at fold-const.c:9827 + { dg-do compile } + { dg-options "-O3 -Wall -fno-tree-loop-optimize" } */ + +int a[1024], b[1024]; + +void f (void) +{ + int i = 0; + for ( ; ; i++) + { + b[16] = a[i * 16 + 10]; + b[i * 16 + 11] = a[i * 16 + 11] * 3; + b[i * 16 + 12] = a[i * 16 + 12] * 4; + b[i * 16 + 13] = a[i * 16 + 13] * 4; + b[i * 16 + 14] = a[i * 16 + 14] * 3; + } +} Index: gcc/tree-ssa-strlen.c =================================================================== --- gcc/tree-ssa-strlen.c (revision 272252) +++ gcc/tree-ssa-strlen.c (working copy) @@ -322,11 +322,17 @@ get_stridx (tree exp) if (TREE_CODE (ptr) == ARRAY_REF) { off = TREE_OPERAND (ptr, 1); - /* Scale the array index by the size of the element - type (normally 1 for char). */ - off = fold_build2 (MULT_EXPR, TREE_TYPE (off), off, - eltsize); ptr = TREE_OPERAND (ptr, 0); + if (!integer_onep (eltsize)) + { + /* Scale the array index by the size of the element + type in the rare case that it's greater than + the typical 1 for char, making sure both operands + have the same type. */ + eltsize = fold_convert (ssizetype, eltsize); + off = fold_convert (ssizetype, off); + off = fold_build2 (MULT_EXPR, ssizetype, off, eltsize); + } } else off = integer_zero_node;