This is the mail archive of the crossgcc@cygnus.com mailing list for the crossgcc project.
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
I recently lost a disk on my SparcStation and decided to upgrade all my GNU software as part of rebuilding everything. This may have been a big mistake, since a couple of things that used to work seem to no longer be accepted. Specifically, I am doing MIPS cross tools: gcc 2.8.1 binutils 2.9.1 newlib 1.8.1 with cross-gcc patch, of course. Oh, right, configuration: target=mips-unknown-elf64 --with-gnu-as --with-gnu-ld --with-newlib --prefix=/usr/local/solaris/mips --program-prefix=mips Binutils and gcc build just fine. Even work, mostly. The problems I am seeing have to do with acceptance of options. 1) When building newlib (specifically libgloss), The Makefiles generated by configure include the option -mno_mips16. This seems reasonable to me, but not to the built cross compiler. The build dies as soon as it hits one. I could walk the subtree and edit the Makefiles, but this could take a really long time. 2) The -m4010 switch is accepted by gas, but the 4010-specific instructions are rejected. The old version I still have around (2.6) accepts these just fine. There is a disturbing comment in one of the source files - "Nobody's using the 4010 any more..." I wonder if LSI Logic knows about this? They might have a different opinion. This suggests that the 4010 implementation may have been deliberately broken. 3) ld now complains about mixed ISA levels. That's OK, but it also refuses to link them. I have no problem with a warning, but the MIPS architecture is deliberately backwards compatible, so this should NOT be an error. I think this is a really bad change, and seriously interferes with my work. Does anyone have any good workarounds for these, short of going back to an older version that does what I need? I actually know how to fix the last one, but I'd like to see this get reversed in some future distribution. The others have me perplexed - where in gcc and gas is the code that determines what -m options are accepted? And who decided the 4010 (an embedded core, if you are not familiar with it) was obsolete? Thanks, Ken Kenneth F. Greenberg, President California Advanced Software Tools, Inc. http://www.calast.com/ _______________________________________________ New CrossGCC FAQ: http://www.objsw.com/CrossGCC _______________________________________________ To remove yourself from the crossgcc list, send mail to crossgcc-request@cygnus.com with the text 'unsubscribe' (without the quotes) in the body of the message.