[PATCH 0/3] Add support for RISC-V 32-bit

[Dmitry V. Levin]

On Fri, Feb 14, 2020 at 11:40:25AM -0800, Alistair Francis wrote:
> This adds initial support for 32-bit RISC-V. RV32 is a 32-bit architecture
> with a 64-bit time_t. It requires the kernel 5.4+ and the glibc support
> is currently being upstreamed.
> 
> This is the result from make check on RV32, so there are some errors that
> need to be fixed:
> 
> FAIL: clock.gen
> FAIL: ioctl_v4l2.gen
> FAIL: ipc_msg.gen
> FAIL: ipc_msg-Xabbrev.gen
> FAIL: ipc_msg-Xraw.gen
> FAIL: ipc_msg-Xverbose.gen
> FAIL: ipc_sem.gen
> FAIL: ipc_sem-Xabbrev.gen
> FAIL: ipc_sem-Xraw.gen
> FAIL: ipc_sem-Xverbose.gen
> FAIL: ksysent.gen
> FAIL: mmap-Xabbrev.gen
> FAIL: mmap-Xraw.gen
> FAIL: mmap-Xverbose.gen
> FAIL: mmap64.gen
> FAIL: mmap64-Xabbrev.gen
> FAIL: mmap64-Xraw.gen
> FAIL: mmap64-Xverbose.gen
> FAIL: printpath-umovestr.gen
> FAIL: regex.gen
> FAIL: sched.gen
> FAIL: semop.gen
> FAIL: semtimedop.gen
> FAIL: sockopt-timestamp.gen
> FAIL: trace_creds.gen
> FAIL: trace_fstat.gen
> FAIL: trace_fstatfs.gen
> FAIL: trace_lstat.gen
> FAIL: trace_personality_32.gen
> FAIL: trace_personality_regex_32.gen
> FAIL: trace_question.gen
> FAIL: trace_stat.gen
> FAIL: trace_stat_like.gen
> FAIL: trace_statfs.gen
> FAIL: trace_statfs_like.gen
> FAIL: wait4.gen
> FAIL: wait4-v.gen
> FAIL: xetitimer.gen
> FAIL: mmap
> FAIL: delay
> FAIL: filtering_syscall-syntax
> FAIL: qual_fault-syscall
> FAIL: qual_fault
> FAIL: strace-k
> FAIL: strace-k-p

This means that either strace itself or its tests (or both) are broken on
this platform.  I suppose it's mostly tests because they still use such
definitions as struct timespec.  Anyway, I'd be very reluctant to merge
something that fails the test suite so badly.  Could you investigate
or at least provide more details about these test failures, please?


-- 
ldv