Matthew Dillon ac8f990bde This patch implements O_DIRECT about 80% of the way. It takes a patchset ...

Tor created a while ago, removes the raw I/O piece (that has cache coherency
problems), and adds a buffer cache / VM freeing piece.

Essentially this patch causes O_DIRECT I/O to not be left in the cache, but
does not prevent it from going through the cache, hence the 80%.  For
the last 20% we need a method by which the I/O can be issued directly to
buffer supplied by the user process and bypass the buffer cache entirely,
but still maintain cache coherency.

I also have the code working under -stable but the changes made to sys/file.h
may not be MFCable, so an MFC is not on the table yet.

Submitted by:	tegge, dillon

2001-05-24 07:22:27 +00:00

..

bus_if.m

…

device_if.m

…

genassym.sh

…

gensetdefs.pl

Generate useful error messages.

2001-04-13 09:37:25 +00:00

imgact_aout.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

imgact_elf.c

Lock the VM while twiddling the vmspace.

2001-05-23 22:05:08 +00:00

imgact_gzip.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

imgact_shell.c

…

inflate.c

…

init_main.c

- Lock the VM when initializing the vmspace for proc0.

2001-05-23 22:06:47 +00:00

init_sysent.c

Regen syscalls that were made mpsafe via vm_mtx

2001-05-19 01:37:12 +00:00

kern_acct.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_acl.c

o As part of the move to not maintaining copies of the vnode owning uid

2001-04-29 19:53:50 +00:00

kern_cap.c

…

kern_clock.c

Remove unneeded includes of sys/ipl.h and machine/ipl.h.

2001-05-15 23:22:29 +00:00

kern_condvar.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_conf.c

Call strlen() once instead of twice.

2001-04-14 21:33:58 +00:00

kern_descrip.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_environment.c

…

kern_event.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_exec.c

Axe unneeded spl()'s.

2001-05-21 18:30:50 +00:00

kern_exit.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

kern_fork.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

kern_idle.c

- Split out the support for per-CPU data from the SMP code. UP kernels

2001-05-10 17:45:49 +00:00

kern_intr.c

- Remove the global ithread_list_lock spin lock in favor of per-ithread

2001-05-17 22:43:26 +00:00

kern_jail.c

…

kern_kthread.c

Convert the allproc and proctree locks from lockmgr locks to sx locks.

2001-03-28 11:52:56 +00:00

kern_ktr.c

Switch from save/disable/restore_intr() to critical_enter/exit().

2001-03-28 03:06:10 +00:00

kern_ktrace.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_linker.c

o Actually extract version of interface and store it along with the name.

2001-03-22 08:58:45 +00:00

kern_lock.c

Instead of asserting that a mutex is not still locked after unlocking it,

2001-04-28 12:11:01 +00:00

kern_lockf.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_malloc.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_mib.c

Overhaul of the SMP code. Several portions of the SMP kernel support have

2001-04-27 19:28:25 +00:00

kern_module.c

…

kern_mutex.c

- Move state about lock objects out of struct lock_object and into a new

2001-05-04 17:15:16 +00:00

kern_ntptime.c

Update to the 2001-04-02 version of the nanokernel code from Dave Mills.

2001-04-16 13:05:05 +00:00

kern_physio.c

…

kern_proc.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_prot.c

o Modify access control checks in p_candebug() such that the policy is as

2001-05-17 21:48:44 +00:00

kern_random.c

…

kern_resource.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

kern_shutdown.c

Revert consequences of changes to mount.h, part 2.

2001-04-29 02:45:39 +00:00

kern_sig.c

- Remove unneeded include of sys/ipl.h.

2001-05-15 23:13:58 +00:00

kern_subr.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_switch.c

Catch up to header include changes:

2001-03-28 09:17:56 +00:00

kern_sx.c

- Move state about lock objects out of struct lock_object and into a new

2001-05-04 17:15:16 +00:00

kern_synch.c

Add in assertions to ensure that we always call msleep or mawait with

2001-05-23 19:38:26 +00:00

kern_syscalls.c

…

kern_sysctl.c

Add convenience function kernel_sysctlbyname() for kernel consumers,

2001-05-19 05:45:55 +00:00

kern_tc.c

…

kern_time.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

kern_timeout.c

Catch up to header include changes:

2001-03-28 09:17:56 +00:00

kern_xxx.c

This is kind of a hack, but it should work. Currently, world is broken

2001-03-24 04:40:49 +00:00

ksched.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

link_aout.c

…

link_elf_obj.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

link_elf.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

linker_if.m

…

Make.tags.inc

- FDESC, FIFO, NULL, PORTAL, PROC, UMAP and UNION file

2001-05-23 09:42:29 +00:00

Makefile

- FDESC, FIFO, NULL, PORTAL, PROC, UMAP and UNION file

2001-05-23 09:42:29 +00:00

makeobjops.pl

Use getopt instead of a home grown one

2001-04-07 20:51:24 +00:00

makesyscalls.sh

…

md4c.c

Import kernel part of SMB/CIFS requester.

2001-04-10 07:59:06 +00:00

md5c.c

…

p1003_1b.c

o First step in cleaning up authorization code for the posix4

2001-05-06 16:15:42 +00:00

posix4_mib.c

…

subr_acl_posix1e.c

o As part of the move to not maintaining copies of the vnode owning uid

2001-04-29 19:53:50 +00:00

subr_autoconf.c

…

subr_blist.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

subr_bus.c

…

subr_clist.c

Make cblock_alloc_cblocks() spell its own name

2001-03-27 10:21:26 +00:00

subr_devstat.c

…

subr_disk.c

Always initialize bio_resid from bio_bcount in the disk mini-layer so

2001-05-08 08:24:54 +00:00

subr_disklabel.c

…

subr_diskmbr.c

…

subr_diskslice.c

…

subr_eventhandler.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

subr_kobj.c

…

subr_log.c

syslogd gets kernel log messages only once every 30 seconds or

2001-05-23 19:02:50 +00:00

subr_mchain.c

…

subr_module.c

…

subr_param.c

…

subr_pcpu.c

- Split out the support for per-CPU data from the SMP code. UP kernels

2001-05-10 17:45:49 +00:00

subr_prf.c

…

subr_prof.c

Remove unneeded includes of sys/ipl.h and machine/ipl.h.

2001-05-15 23:22:29 +00:00

subr_rman.c

…

subr_sbuf.c

Rewrite of the CAM error recovery code.

2001-03-27 05:45:52 +00:00

subr_scanf.c

…

subr_smp.c

- Split out the support for per-CPU data from the SMP code. UP kernels

2001-05-10 17:45:49 +00:00

subr_taskqueue.c

remove include of ipl.h because it no longer exists

2001-05-16 02:52:06 +00:00

subr_trap.c

Don't acquire Giant just to call trap_fatal(), we are about to panic

2001-05-23 22:58:09 +00:00

subr_turnstile.c

- Move state about lock objects out of struct lock_object and into a new

2001-05-04 17:15:16 +00:00

subr_witness.c

- Don't panic on a try lock operation for a sleep lock if we hold a spin

2001-05-17 22:44:56 +00:00

subr_xxx.c

…

sys_generic.c

Back out scanning file descriptors with holding a process lock.

2001-05-15 10:19:57 +00:00

sys_pipe.c

aquire vm_mutex a little bit earlier to protect a pmap call.

2001-05-23 10:26:36 +00:00

sys_process.c

- FDESC, FIFO, NULL, PORTAL, PROC, UMAP and UNION file

2001-05-23 09:42:29 +00:00

sys_socket.c

…

syscalls.c

Regen syscalls that were made mpsafe via vm_mtx

2001-05-19 01:37:12 +00:00

syscalls.master

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

sysv_ipc.c

…

sysv_msg.c

…

sysv_sem.c

…

sysv_shm.c

Correct style bugs with regards to long lines and comments.

2001-05-23 23:38:05 +00:00

tty_compat.c

…

tty_conf.c

…

tty_cons.c

Send the remains (such as I have located) of "block major numbers" to

2001-03-26 12:41:29 +00:00

tty_pty.c

Send the remains (such as I have located) of "block major numbers" to

2001-03-26 12:41:29 +00:00

tty_snoop.c

Convert this driver to (ab?)use line disciplines to get the input it

2001-05-22 22:13:14 +00:00

tty_subr.c

Make cblock_alloc_cblocks() spell its own name

2001-03-27 10:21:26 +00:00

tty_tty.c

Support /dev/ctty again

2001-05-15 18:12:38 +00:00

tty.c

Unifdef DEV_SNP; snp(4) no longer requires these ugly hacks.

2001-05-22 22:16:18 +00:00

uipc_accf.c

…

uipc_domain.c

…

uipc_mbuf2.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

uipc_mbuf.c

Increment mbstat.m_mpfail, not mbstat.m_mcfail, when m_pullup() fails.

2001-05-23 20:44:54 +00:00

uipc_proto.c

…

uipc_sockbuf.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

uipc_socket2.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

uipc_socket.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

uipc_syscalls.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

uipc_usrreq.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

vfs_acl.c

o As part of the move to not maintaining copies of the vnode owning uid

2001-04-29 19:53:50 +00:00

vfs_aio.c

Check validity of signal callback requested via aio routines.

2001-04-18 22:18:39 +00:00

vfs_bio.c

This patch implements O_DIRECT about 80% of the way. It takes a patchset

2001-05-24 07:22:27 +00:00

vfs_cache.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

vfs_cluster.c

This patch implements O_DIRECT about 80% of the way. It takes a patchset

2001-05-24 07:22:27 +00:00

vfs_conf.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

vfs_default.c

Don't release the vm lock just to turn around and grab it again.

2001-05-23 19:51:12 +00:00

vfs_export.c

vfs_subr.c is getting rather fat. The underlying repocopy and this

2001-04-26 20:47:14 +00:00

vfs_extattr.c

Don't release Giant around vm_oject_page_clean() in fsync() as the pager

2001-05-23 22:55:13 +00:00

vfs_init.c

Revert consequences of changes to mount.h, part 2.

2001-04-29 02:45:39 +00:00

vfs_lookup.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

vfs_mount.c

Undo part of the tangle of having sys/lock.h and sys/mutex.h included in

2001-05-01 08:13:21 +00:00

vfs_subr.c

Introduce a global lock for the vm subsystem (vm_mtx).

2001-05-19 01:28:09 +00:00

vfs_syscalls.c

Don't release Giant around vm_oject_page_clean() in fsync() as the pager

2001-05-23 22:55:13 +00:00

vfs_vnops.c

This patch implements O_DIRECT about 80% of the way. It takes a patchset

2001-05-24 07:22:27 +00:00

vnode_if.pl

replace calls to non-existant bail() subroutine with calls to

2001-03-23 11:48:50 +00:00

vnode_if.src

VOP_BALLOC was never really a VOP in the first place, so convert it

2001-04-29 12:36:52 +00:00