In preparation for switching linuxulator to the use the native 1:1 threads refactor kern_sched_rr_get_interval() and sys_sched_rr_get_interval(). Add a kern_sched_rr_get_interval() counterpart which takes a targettd parameter to allow specify...

[dchagin]

Summary

...target thread directly by callee (new Linuxulator).

Linuxulator temporarily uses first thread in proc.

Move linux_sched_rr_get_interval() to the MI part.

While here rename e to a better name error.

[trasz]

sys/compat/linux/linux_misc.c

1934 ↗

(On Diff #2162)

Is the size and layout of l_timespec the same for 32 and 64 bit Linux binaries?

sys/kern/p1003_1b.c

314 ↗

(On Diff #2162)

Hm, another case of "asymetrical locking" - kern_sched_rr_get_interval_td() called with proc lock held, and dropping it, instead of leaving it to the caller.

[dchagin]

sys/compat/linux/linux_misc.c

1934 ↗

(On Diff #2162)

no, l_timespec machine dependent, declared in MD linux.h

[dchagin]

Fixed asymetrical locking, leave unlocking to the caller.

[trasz]

sys/compat/linux/linux_misc.c

1934 ↗

(On Diff #2162)

How does it work for both 32 and 64 bit Linux binaries at the same time, then?

[dchagin]

sys/compat/linux/linux_misc.c

1934 ↗

(On Diff #2162)

not sure what understand the question correctly

1. see lines 1952, 1953
2. linux_misc.c compiled twice, for linux64.ko and linux.ko. and for the linux.ko COMPAT_LINUX32 defined.

[trasz]

Ah, it's built twice; that explains. Thanks, the change looks fine.