nanosleep: plug a kernel memory disclosure

[vangyzen]

Summary

nanosleep() updates rmtp on EINVAL. In that case, kern_nanosleep()
has not updated rmt, so sys_nanosleep() updates the user-space rmtp
by copying garbage from its stack frame. This is a kernel memory
disclosure. It's also not POSIX-compliant. Fix it to update rmtp
only on EINTR.

I found this while working on D10020, and I made it public before
realizing the possible security implications.
The Security Officer suggested that I go ahead and commit the fix.

[vangyzen]

Thankfully, this is likely not a severe disclosure.

[dchagin]

For linuxulator this does not make sense now as we check timespec values in linux_to_native_timespec before calling kern_nanosleep()

[vangyzen]

In D10044#207644, @dchagin wrote:

For linuxulator this does not make sense now as we check timespec values in linux_to_native_timespec before calling kern_nanosleep()

kern_nanosleep checks for negative nanoseconds, but I don't see where linux_to_native_timespec checks for that.

[dchagin]

In D10044#207661, @vangyzen wrote:

In D10044#207644, @dchagin wrote:

For linuxulator this does not make sense now as we check timespec values in linux_to_native_timespec before calling kern_nanosleep()

kern_nanosleep checks for negative nanoseconds, but I don't see where linux_to_native_timespec checks for that.

yes, this is error, linux timespec_valid() function casts tv_usec to unsigned before comparison. I'll fix