On 2017-12-03, at 3:17 PM, Helge Deller wrote: > Overall I think we are in pretty good shape with the port, > but here is a small list where work would be helpful: > > - gcc: Make gccgo work (still crashes) > - kernel/gdb: For core files written by kernel, gdb is reporting: BFD: > Warning: /tmp/core.http.1655.ls3011.1465554447 is truncated: expected core > file size >= 2187264, found: 2121728. > - kernel/glibc: Maybe implement real kernel VDSO support ? > - gcc: _mcount can be optimized (smaller) by not subtracting offset > - gcc: put PLABELS into read-only section (gcc function pa_reloc_rw_mask() > needs modification, see "section mismatches" mails)
Regarding the PLABELS, I believe the dynamic linker would have to be changed to support putting them into a read-only section. We have been slowly loosing packages due to a lack of porting effort. - Porting openjdk would be a huge help to package count - isn't there a generic implementation - git build fails due to testsuite fail - vim/newvim have testsuite issues (maybe gcc bug?) - gtk+3.0 (two tests fails) - erlang packages - many builds fail due to a segv - desktoptojson faults building krdc - gdb is pretty fragile There's probably more but I agree things are much better than before. Dave -- John David Anglin [email protected]
