This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC 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]

Re: [PATCH, PR 41089] Do not scalarize va_lists in early SRA


On Wed, Sep 9, 2009 at 6:21 PM, Martin Jambor <mjambor@suse.cz> wrote:

> Uros reported that this fixes some but not all alpha stdarg failures,
> there is still one execution test failing and one dump match not
> matching. ?However, if I managed to cross-compile it correctly, no SRA
> does anything at all in the execution testcase and similarly the early
> SRA does nothing in the dump scan testcase (late SRA happens once but
> it is too late to affect the dump).

Perhaps I was not clear enough, only stdarg-1.c still fails on alpha, that is:

FAIL: gcc.c-torture/execute/stdarg-1.c execution,  -O3 -fomit-frame-pointer
FAIL: gcc.c-torture/execute/stdarg-1.c execution,  -O3 -g

Following failures were fixed by your patch, including dump match in stdarg-2:

FAIL: gcc.c-torture/execute/stdarg-4.c execution,  -O1
FAIL: gcc.c-torture/execute/stdarg-4.c execution,  -O2
FAIL: gcc.c-torture/execute/stdarg-4.c execution,  -O3 -fomit-frame-pointer
FAIL: gcc.c-torture/execute/stdarg-4.c execution,  -O3 -g
FAIL: gcc.c-torture/execute/stdarg-4.c execution,  -Os
...
FAIL: gcc.dg/tree-ssa/stdarg-2.c scan-tree-dump stdarg "f15: va_list escapes 1,
needs to save all GPR units and all FPR units"

Uros.


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