Remove some warnings that only appear on i386 (not on amd64) and that
for some reason best known to others are deemed fatal for i386.
Making this code 'pass lint' does absolutely nothing for its readability (etc).
- Addresses the issue described in PR/38828.
- Some simplification in threading and sleepq subsystems.
- Eliminates pmap_collect() and, as a side note, allows pmap optimisations.
- Eliminates XS_CTL_DATA_ONSTACK in scsipi code.
- Avoids few scans on LWP list and thus potentially long holds of proc_lock.
- Cuts ~1.5k lines of code. Reduces amd64 kernel size by ~4k.
- Removes __SWAP_BROKEN cases.
Tested on x86, mips, acorn32 (thanks <mpumford>) and partly tested on
acorn26 (thanks to <bjh21>).
Discussed on <tech-kern>, reviewed by <ad>.
to pass any real configuration data. This benefits things like
rump_smbfs which need to init rump before running the userspace
mount routines (because they make rump system calls) but cannot
run p2k_setup() at that stage because it requires data provided by
the userspace mount routines. And p2k_setup() in turn forks (to
detach), so it cannot be run after rump has been inited and threads
have been created.
Note: We use the storage of puffs_cache_write from puffs_ops for
this purpose. It's not issued by the kernel and hence currently
unused, and this saves us from the trouble of bumping the lib major
version.
mutex, it's a good idea to convert the unlock path too so that the
recursive mutex gets released as many times as it was locked.
Fixes pthread + cxa deadlock.
systems should be mounted with ukfs_mount_disk() from now on. The
partition argument specifies which label is being mounted (or the
entire image). E.g. partition 4 should be label 'e'.
reference counting and not release nodes based just on puffs'
impression of if they are free.
This also allows us to reclaim vnodes already in inactive if the
file system so desires. Some file systems, most notably ffs, change
file state already in inactive. This could lead to a deadlock in
the middle of inactive and reclaim if some other puffs operation
was processed in between (as exposed by haad's open(at) test
program).
Also, properly thread the componentname from lookup to the actual
vnode operation. This required the changes the rump componentname
routines. Yes, the rename case is truly mindbogglingly disgusting.
Puke for yourself.
dsl