Source-Changes-D archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: CVS commit: src/lib/libc/net



On Sun, Dec 25, 2011 at 10:20:35AM +0900, Izumi Tsutsui wrote:
> > On Sun, Dec 25, 2011 at 09:54:37AM +0900, Izumi Tsutsui wrote:
> > > > On Sat, Dec 24, 2011 at 10:17:37PM +0900, Izumi Tsutsui wrote:
> > > > > > > Log Message:
> > > > > > > Specify "-fno-tree-ter" to getaddrinfo.c and gethnamaddr.c on arm 
> > > > > > > and gcc45
> > > > > > > for workaround of a possible optimazation bug.
> > > > > > 
> > > > > > Please make sure this only is applied if the active CC is GCC.
> > > > > 
> > > > > Isn't this enough?
> > > > > 
> > > > > ---
> > > > > .if defined(HAVE_GCC) && ${HAVE_GCC} == "45" && ${MACHINE_CPU} == 
> > > > > "arm"   
> > > > > COPTS.getaddrinfo.c+= -fno-tree-ter
> > > > > COPTS.gethnamaddr.c+= -fno-tree-ter
> > > > > .endif
> > > > 
> > > > No, this only checks that HAVE_GCC is supported. HAVE_GCC, HAVE_LLVM and
> > > > HAVE_PCC are non-exclusive with a fallback order.
> > > 
> > > So what's the proper test?
> > > 
> > > I don't see any proper examples:
> > > bin/csh/Makefile
> > > games/phantasia/Makefile
> > > lib/libedit/Makefile
> > > sbin/fsdb/Makefile
> > > etc.
> > > 
> > > Could you please fix them?
> > 
> > I will fix them at some point. I don't care too much about those, since
> > they apply to platforms that LLVM doesn't support (and isn't likely to
> > support soon) or they simply use standard options. -fno-tree-ter doesn't
> > qualify for that though.
> > 
> > For completeness, it should be:
> > 
> > .if ${HAVE_GCC:U} == "45" && ${MACHINE_CPU} == "arm"
> > COPTS.getaddrinfo.c+=       ${${ACTIVE_CC} == "gcc" :? -fno-tree-ter :}
> > COPTS.gethnameaddr.c+=      ${${ACTIVE_CC} == "gcc" :? -fno-tree-ter :}
> > .endif
> 
> Please commit.  I can't test LLVM right now.

The above won't break with LLVM, but please make sure that it does what
you want.

Joerg


Home | Main Index | Thread Index | Old Index