Remove Softdep.

Softdep has been a no-op for some time now, this removes it to get
it out of the way.

Flensing mostly done in Talinn, with some help from krw@

ok deraadt@

remove unused prototypes from buf.h

bufcache_getanycleanbuf() removed in vfs_bio.c 1.181
buf_acquire_unmapped() removed in vfs_biomem.c 1.22
buf_print() removed in vfs_bio.c 1.47
cluster_write() rem

remove unused prototypes from buf.h

bufcache_getanycleanbuf() removed in vfs_bio.c 1.181
buf_acquire_unmapped() removed in vfs_biomem.c 1.22
buf_print() removed in vfs_bio.c 1.47
cluster_write() removed along with vfs_cluster.c in 2016

show more ...

Re commit what was committed in version 1.43 with a fix added to
ensure we handle the uvm_objects of bread_cluster buffers correctly.

Original commit message:
Fix the buffer cache code to not use a

Re commit what was committed in version 1.43 with a fix added to
ensure we handle the uvm_objects of bread_cluster buffers correctly.

Original commit message:
Fix the buffer cache code to not use a giant uvm obj of all pages
when a small one on each buf is all that is needed. reduces the
cost of large frees by about 25%.

Again, lots of assistence from kettenis and mlarkin

still ok kettenis@

show more ...

back out the buffer cache uvm_obj change for now.

the bread_cluster code has confused even me and mark,
we need to handle the buffer slice and dice case better
for bread_cluster.

Fix the buffer cache code to not use a giant uvm obj of all pages
when a small one on each buf is all that is needed. reduces the
cost of large frees by about 25%.
ok kettenis@

getblk(9): tsleep(9) -> tsleep_nsec(9); ok visa@

Make sure we clear all valid bytes in clrbuf.

ok deraadt@

Simplify the startup of the cleaner, reaper and update threads by
passing the main function directly to kthread_create(9). The start_*
functions are mere stepping stones nowadays and can be pruned.
T

Simplify the startup of the cleaner, reaper and update threads by
passing the main function directly to kthread_create(9). The start_*
functions are mere stepping stones nowadays and can be pruned.
They used to contain more logic in the pre-kthread era.

While here, set `cleanerproc' and `syncerproc' during the thread
creation rather than expect the threads to set the proc pointer.
Also, rename `sched_sync' to `syncer_thread' to reduce confusion
with the scheduler-related functions.

OK kettenis@, deraadt@, mpi@

show more ...

Flip previously warm pages in the buffer cache to memory above the DMA
region if uvm tells us it is available.
nits from deraadt@
ok krw@ guenther@ kettenis@

Switch geteblks()'s size argument from int to size_t. It's called with
unsigned variables as argument in most places anyway. Decrease the
chance of signedness/range mismatch issues.

ok stefan

move buf_rb_bufs from RB macros to RBT functions

i had to shuffle the order of some header bits cos RBT_PROTOTYPE
needs to see what RBT_HEAD produces.

VOP_REALLOCBLKS() and related code is unused since the removal of
cluster_write().
ok beck zhuk

beck@ forgot to commit this with sys/kern/vfs_bio.c r1.100.

> Start some refactoring in here. this gets bufadjust out
> of the hibernate path and starts preparing for some other work in here

Only declare buf_daemon() once

move buffer size adjustment to buf_adjcnt - from Walter Neto
ok mpi@

Use two 2q caches for the buffer cache, moving previously warm buffers from the
first queue to the second.
Mark the first queue as DMA in preparation for being able to use more memory
by flipping. Fl

Use two 2q caches for the buffer cache, moving previously warm buffers from the
first queue to the second.
Mark the first queue as DMA in preparation for being able to use more memory
by flipping. Flipper code currently only sets and clears the flag.
ok tedu@ guenther@

show more ...

disksort is dead, and now so is any code that relied on its compat
in bufqs.

it's only taken us 6 years.

ok millert@ krw@

save the bufq pointer from the buf before we turn it loose so it won't
change on us. also, remove unused second arg to bufq_wait.
from pedro at bitrig via david hill.
ok beck kettenis

rearrange fields for better packing. reduces size to 256 on amd64.
ok deraadt

replace LRU bufcache with something originally modelled after 2Q.
this should provide a degree of scan resistance, and also serves as a
midway point for further development of multi queue algorithms.

replace LRU bufcache with something originally modelled after 2Q.
this should provide a degree of scan resistance, and also serves as a
midway point for further development of multi queue algorithms.
i've tried to minimize the risk and degree of regressions.
probably ok beck

show more ...

pull the bufcache freelist code out into separate functions to allow new
algorithms to be tested. in the process, drop support for unused B_AGE and
b_synctime options.
previous versions ok beck deraa

pull the bufcache freelist code out into separate functions to allow new
algorithms to be tested. in the process, drop support for unused B_AGE and
b_synctime options.
previous versions ok beck deraadt

show more ...

remove #define b_errcnt b_resid. nothing uses it, and nothing should.

remove the #define b_cylinder b_resid from bufs. i hated the
overloading of that thing.

the only hardware that seems to care about cylinders in our tree
are floppy drives, and the drivers for those

remove the #define b_cylinder b_resid from bufs. i hated the
overloading of that thing.

the only hardware that seems to care about cylinders in our tree
are floppy drives, and the drivers for those calculate their own
cylinders from logical block addresses and ignore whatever the rest
of the kernel thought b_cylinders should be.

most of this diff is moving the floppy drivers to using b_resid as
a resid and using that as part of the calculation for real cylinder
values.

the rest of the diff is getting rid of the useless assignments to
b_cylinder that dont get used by anything (now that disksort is
gone).

ok miod@

show more ...

now that all the direct users of disksort have been removed, we can now
safely remove disksort.

most hardware and pretty much all of the kernel has moved to logical
block addressing when dealing wit

now that all the direct users of disksort have been removed, we can now
safely remove disksort.

most hardware and pretty much all of the kernel has moved to logical
block addressing when dealing with disks, so the assumptions disksort
was built against arent useful these days. it also has bad edge cases
with lots of sequential writes being able to starve other io requests
in the system. these issues have been addressed by becks nscan
implementation, which disksort was previously deprecated in favour
of.

this removes the guts of disksort and the bufq wrapper around it.

ok miod@

show more ...

remove some abuse of bufqs where they were overloaded to store workq_tasks
for completing swap io against files. this in turn bloated struct buf.

this moves from using workq_tasks to the task api, b

remove some abuse of bufqs where they were overloaded to store workq_tasks
for completing swap io against files. this in turn bloated struct buf.

this moves from using workq_tasks to the task api, but stores the task
handler in vndbuf which is allocated specially for every io in this path
anyway.

ok kettenis@ beck@

show more ...