looks ok to me. please try a symbol of array type and see what happens. A note here, i actaully expected "floppy" be promoted to 'auto-local". Sun
On Sat, Feb 19, 2011 at 5:45 AM, Feng Zhou <fengzho...@gmail.com> wrote: > Hello, all > > Can gatekeeper review the attached patch for bug #713 > (https://bugs.open64.net/show_bug.cgi?id=713) please? Thank you. > -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- > Bug #713 can be reproduced with the following test-case (also attached > to bug #713's bug report): > > #include <stdio.h> > > typedef struct _GFloppy { > char *device; > } GFloppy; > > static GFloppy floppy; > > typedef struct _GDevice { > void *ptr; > } GDevice; > > int main(int argc, char *argv[]) > { > static GDevice dev = { &(floppy.device) }; > > printf("%d\n", *(int*)(dev.ptr)); > return 0; > } > > This bug only happens when -Ofast is specified. With IPA, two .I files > are generated: symtab.I and 1.I (which contains code of main > function). The static symbol `floppy' is merged into global symbol > table (symtab.I/G) and emitted into symtab.s. It is not exported > because it is a local symbol. On the other hand, when generating the > initialization code for `dev' from INITO table during compiling the > main function, code emission generates references to `floppy' > directly. This caused the linking problem since `floppy' is not > exported. The fix is, instead of generating references to local > symbols directly when IPA is enabled, generate references using the > section start + offset. > > -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- > > Feng Zhou > > ------------------------------------------------------------------------------ > The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE: > Pinpoint memory and threading errors before they happen. > Find and fix more than 250 security defects in the development cycle. > Locate bottlenecks in serial and parallel code that limit performance. > http://p.sf.net/sfu/intel-dev2devfeb > _______________________________________________ > Open64-devel mailing list > Open64-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/open64-devel > > ------------------------------------------------------------------------------ The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE: Pinpoint memory and threading errors before they happen. Find and fix more than 250 security defects in the development cycle. Locate bottlenecks in serial and parallel code that limit performance. http://p.sf.net/sfu/intel-dev2devfeb _______________________________________________ Open64-devel mailing list Open64-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open64-devel
