On Thu, 30 Aug 2018, Stafford Horne wrote: > On Wed, Aug 29, 2018 at 04:38:45PM -0600, Jeff Law wrote: > > ps. Any plans for glibc? How complete is the qemu support? I'm having > > reasonable success building little chroot filesystems, then using qemu > > to do bootstrap testing within those chroots. > > Hello Juff, > > I was planning to start to look into glibc. We have a port here: > > https://github.com/openrisc/or1k-glibc > > I believe the contributor's all have FSF copyright signoff in place. However, > the port could not be upstreamed a few years ago due to the GCC port not being > able to go upstream. > > Once GCC is resolved I think it shouldn't be too much effort to get glibc back > in shape.
You'll need to update it for all global changes in glibc since whenever work on that port started / whenever you last had it up to date with such cross-port changes in glibc. Also, note it's expected for new ports to have a shlib-versions file setting the minimum symbol version to that of the upstream glibc version the port actually goes in (and as already noted that file should also set an ABI-specific dynamic linker name, so for both reasons you have an ABI break from the previous glibc version). Look at review comments on other ports posted in the past year or two (RISC-V, ARC, NDS32, C-SKY) for an idea of various issues to watch out for and expectations for new ports and new port submissions (which include the upstream tools and Linux kernel port having everything required for building the glibc port with good test results). -- Joseph S. Myers jos...@codesourcery.com
