Mercurial > hg > CbC > CbC_gcc
comparison gcc/testsuite/gcc.dg/tree-ssa/ssa-dse-5.c @ 111:04ced10e8804
gcc 7
| author | kono |
|---|---|
| date | Fri, 27 Oct 2017 22:46:09 +0900 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| 68:561a7518be6b | 111:04ced10e8804 |
|---|---|
| 1 /* { dg-do compile } */ | |
| 2 /* { dg-options "-O2 -fdump-tree-optimized" } */ | |
| 3 | |
| 4 int x; | |
| 5 | |
| 6 int | |
| 7 f1 (int i, int j, int k) | |
| 8 { | |
| 9 int *p = k ? &i : &j; | |
| 10 i = 3; | |
| 11 *p = 5; | |
| 12 x = j; | |
| 13 } | |
| 14 | |
| 15 /* The assignment "i = 3" is partially dead. Our DSE pass doesn't handle | |
| 16 detection of partially dead assignments. | |
| 17 | |
| 18 There's two outputs which would indicate that the optimization was | |
| 19 performed. | |
| 20 | |
| 21 If we used block copying to detect and eliminate the partially dead | |
| 22 store, then we should see an assignment "i = 5" somewhere in the | |
| 23 dump file. | |
| 24 | |
| 25 Another approach would be to redirect the path from the true arm | |
| 26 of the first conditional so that it reaches the statement *p = 5 | |
| 27 rather than i = 3. */ | |
| 28 | |
| 29 /* { dg-final { scan-tree-dump-times "i = 5" 1 "optimized" { xfail *-*-* }} } */ | |
| 30 /* { dg-final { scan-tree-dump-times "<L.*>:;\[\n\t \]*\\*p = 5" 1 "optimized" { xfail *-*-*}} } */ | |
| 31 | |
| 32 |