On Tue, Feb 28, 2012 at 4:22 PM, Peter Gavin <[email protected]> wrote: > On Tue, Feb 28, 2012 at 5:02 PM, Julius Baxter <[email protected]> > wrote: >> >> >> Hi Pete, >> >> Nice work! I knew you were improving it but didn't know you got the >> kernel compiling already. Well done! > > > Thanks :) > > >> >> Are the GCC and binutils ports tracking the upstream tree? > > > Yes, both. The last time I merged changes from them was a few days ago. > >> >> There's two things worth pursuing at this point I think - first is the >> GDB port, and second is the regression testing of the tool chain (and >> third, the first's sim doing the second.) >> >> Let me explain that a little better. I was doing the CGEN port for a >> number of reasons, but one was to get the basic simulator in GDB >> working. So I'd be most interested in polishing off this good work >> with setting that up properly and ensuring the tool chain is correctly >> tested and passing. > > > Ok, I'll take a look at gdb. I think the semantics in the .cpu files > will need some more improvement to get that working, though.
Not much more is needed to get it to the level of running the GCC regression suite. We don't need any exception handling or peripherals modelled, just the bits of the ISA that GCC knows about. The idea being that the sim in GDB will be largely for the GCC regression suite execution tests and or1ksim is for simulation of fully-featured systems to run things like the Linux kernel port. So I don't think we need to go to too much effort in the GDB sim. > >> >> Also, how about the or1ksim test suite using this new compiler? Is that >> running? > > > Yes, the entire or1ksim testsuite passes. Nice work! > > Also, the entire binutils testsuite passes if you configure with > --target=or1k-elf. If you use --target=or1k-linux, it tries to do the > shared/PIC tests which don't pass. > > The gcc testsuite doesn't pass. I haven't looked too hard at it though. Cool, well I'll start looking into it. Hopefully it won't be too much work to get it up to scratch. Again, good work mate! Jules _______________________________________________ OpenRISC mailing list [email protected] http://lists.openrisc.net/listinfo/openrisc
