Fix siginfo_t for Illumos

[dylni]

The siginfo_t struct appears to be incorrect for Illumos. It currently has si_code before si_errno and appears to be too large. I'm not familiar with this platform, so please let me know if this fix is inaccurate.

siginfo_t:
https://github.com/illumos/illumos-gate/blob/74e98f497ae878ae481ccba1f91fd8dc26fa6627/usr/src/boot/sys/sys/signal.h#L196

siginfo.h:
https://github.com/illumos/illumos-gate/blob/74e98f497ae878ae481ccba1f91fd8dc26fa6627/usr/src/uts/common/sys/siginfo.h#L164

I am not sure if the packed() attribute should remain, but I copied it from the Solaris definition.

[rust-highfive]

Some changes occurred in solarish module

cc @jclulow,@pfmooney

[rust-highfive]

r? @Amanieu

(rust-highfive has picked a reviewer for you, use r? to override)

[jclulow]

siginfo_t: https://github.com/illumos/illumos-gate/blob/74e98f497ae878ae481ccba1f91fd8dc26fa6627/usr/src/boot/sys/sys/signal.h#L196

That's not the right header. Note that it's under usr/src/boot which was imported with our copy of the FreeBSD boot loader.

[jclulow]

Can I ask what lead you to make this change? Did something fail to compile or work correctly on an illumos system?

[dylni]

@jclulow Yes, the motivation is this issue:
dylni/process_control#12

libc currently does not define siginfo_t::si_status on Illumos, which is required by process_control. The signal codes are also required so were included with this change.

If you let me know the correct definition, I can certainly update this PR.

[pfmooney]

@jclulow Yes, the motivation is this issue: dylni/process_control#12

libc currently does not define siginfo_t::si_status on Illumos, which is required by process_control. The signal codes are also required so were included with this change.

If you let me know the correct definition, I can certainly update this PR.

I'm working on an update to expose the siginfo_t fields you need using function accessors like the other unix-y platforms have done. I should have it posted in the next day or so.

[dylni]

Thank you @pfmooney! Please also include CLD_EXITED and the related constants from this PR if possible.