From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Henderson To: Jim Blandy Cc: egcs@egcs.cygnus.com, gdb@sourceware.cygnus.com Subject: Re: IA32: printing FP register variables Date: Mon, 19 Jul 1999 23:41:00 -0000 Message-id: <19990719234100.C17063@cygnus.com> References: <9500.931826533@upchuck.cygnus.com> X-SW-Source: 1999-q3/msg00094.html On Mon, Jul 19, 1999 at 03:26:18PM -0500, Jim Blandy wrote: > Once the other internal work is done, would the EGCS folks be willing > to experiment with both, and give us some sense of how much difference > it makes? Yes. It seems a reasonable thing to try. r~ >From tm@netcom.com Tue Jul 20 12:14:00 1999 From: Toshiyasu Morita To: jtc@redback.com Cc: gdb@sourceware.cygnus.com Subject: Re: GDB with Hitachi Linker output? Date: Tue, 20 Jul 1999 12:14:00 -0000 Message-id: <199907201913.MAA22621@netcom13.netcom.com> References: <5mwvvw74ic.fsf@jtc.redbacknetworks.com> X-SW-Source: 1999-q3/msg00095.html Content-length: 1478 The SRCONV utility is actually pretty broken, through no fault of Cygnus. We ran into the problem three years ago that SRCONV-emitted files caused the Hitachi in-circuit emulator to crash. After much consultation with both sides, we determined that: 1) The Hitachi SYSROF spec only describes the output of the data from the compiler/linker, and do not specify implicit ordering dependencies, and 2) linker/debugger are fragile and do not emit proper error messages when a certain section is out of order. Therefore, it was impossible to debug the problem and we gave up on SRCONV. Toshi > > >>>>> "Bill" == Bill Gatliff writes: > Bill> Is anyone using GDB sh-hitachi-hms with executables supplied by > Bill> the Hitachi linker? > > I don't think so. > > AFAIK, the hitachi object format is sysroff. There is a sysroff to > COFF converter in binutils (srconv), but nothing to do the reverse. I > vaguely remember a receiving email from Ian Taylor two or three years > ago that said that support for whatever was missing in BFD that caused > srconv to be a separate program (instead of available through objcopy) > had been added. That project never became high enough priority to > follow up. > > It's possible that the email was referring to nlmconv instead of > srconv, since I worked on both programs in that time frame. I'm > sorry I can't remember any more details. > > --jtc > > -- > J.T. Conklin > RedBack Networks >