♟ glebius

In pfi_initialize_vnet() rely on IFNET_RLOCK() which is sx(9) for stability of ifnet and ifgroup lists. First allocate enough pfi_kif structs, then enter epoch and attach them.
pfi_get_ifaces() NET_EPOCH_ENTER since pfi_skip_if() traverses ifgroup list.

With r343051 UMA switched from atomic counts to counter(9) and now kernel

Imaginary cat jumped my keyboard!

For 32-bit machines rollback the default number of vnode pager pbufs

The plan is to not be in epoch during full ifattach/ifclone event since these events have many places that allocate memory. Needs to be more fine grained. I'll ponder more over that.

Enter epoch in DAD timer.
Revert entering epoch for setsockopt. This needs to be fine grained.

Remove remnants of byte order manipulation, back when FreeBSD stack

Allow some nesting of ng_iface(4) interfaces and add a configuration knob.

Now that there is only one way to allocate a slab, remove uz_slab method.

Report cache zones in UMA stats sysctl, that 'vmstat -z' uses. This

Teach pfil_ioctl() about VIMAGE.

Return PFIL_CONSUMED if packet was consumed. While here gather all

Repair siftr(4): PFIL_IN and PFIL_OUT are defines of some value, relying

Unbreak call to ipf_check(): it expects the out parameter to be 0 or 1.

Hopefully fix compilation by other compilers.

Fix build without INET6.

Add new m_ext type for data for M_NOFREE mbufs, which doesn't actually do

Rebase on top of r343619.

Revert r316461: Remove "IPFW static rules" rmlock, and use pfil's global lock.

Remove "all rights reserved" from new files.

Obsolete MLINKS.
My copyrights.

In zone_alloc_bucket() max argument was calculated based on uz_count.

Address ae@ review, fix multiple mismerges and bugs in ip_fw_pfil.

Set max names length to MAXPFILNAME equal to 64.
Go through all pfil_run_hooks() and properly compare against PFIL_PASS

In D18951#406615, @kristof wrote:

Did you happen to do any benchmarking on this? I'd have expected "Sync pfil hooks epoch(9)" to give us a (small) performance improvement, but my initial test shows a small reduction in forwarding performance (with pf loaded).

Improve wording. [1]
Fix two .Nd present. Since mdoc(7) says that .Nd doesn't accept

Very happy to see this.

Create /dev/pfil only for the default VNET.
Corrections from Mateusz.
mandoc -T lint
igor

This panics during the netipsec and pf regression tests:

Remove sbytes from the API.

To be clear, this approach induces a certain amount of pain, and it should not be undertaken unless the benefits are worth it. Some justification based on actual users of sendfile() is needed. If it's not worth it, then this problem is just a wart.

Fix mistake in r343030: move nswbuf calculation back to

Do not reserve KVA for paging bufs in vm_ksubmap_init(), since now

Fix compilation failures on different arches that have vm_machdep.c not

style(9): break long line.

Remove harmless leftover from code that cycles over zone's kegs. Just use +

This was missed in r343051: make uz_allocs, uz_frees and uz_fails counter(9).

Only do uz_items accounting for zones that have a limit set in uz_max_items.

Make uz_allocs, uz_frees and uz_fails counter(9). This removes some

Fix compilation on 32-bit.

Remove pbuf(9) manual page that is now nearly 100% outdated.

Allocate pager bufs from UMA instead of 80-ish mutex protected linked list.

Add flag LK_NEW for lockinit() that is converted to LO_NEW and passed

o Move zone limit from keg level up to zone level. This means that now

Remove recursive NET_EPOCH_ENTER() from sysctl_ifmalist(), missed in r342872.

Remove support for FreeBSD 9 kernel, which used to change byte order

Remove unnecessary ifdef. With INVARIANTS all KASSERTs are empty statements,

Bring the comment up to date.

Simplify sosetopt() so that function has single return point. No

Timer interval is correctly counted in ticks (1/100 s) in

Mechanical cleanup of epoch(9) usage in network stack.

Remove part of comment that doesn't match reality.

Remove patch that I missed to add to commit index in r489641.

Upgrade minidlna to current snapshot in my repo.

Fix regression in r331368, that broke dumping of UMA startup pages

Use mbufq instead of ifqueue to queue mbufs.

I would like to see a generic code in netgraph that marks hooks with DLT. So, no special messages to be needed, nodes will autosense what's connected to them. However, this is just a wish not a blocker for this change.

Move the sleep into zone_alloc_item_locked() and put it into a loop.

As said, I'm not seeking for a review & approval so that I commit it. I'm not a ports committer. I'm seeking for a person who can finalize this ugly patch and commit. Thanks!

Rewrite uma_zalloc_arg() so that there is a single sleeping point.

Improve comment on why uk_lock needs to be first.

Address Mark's review.

Fix yet another edge case in uma_startup_count(). If zone size fits into

For not offpage zones the slab is placed at the end of page. Keg's uk_pgoff

Move assertion to beginning of the function.
Remove unlock avoidance in uma_zfree_arg(). Can't remember why it was

P.S. In pf itself, the dtrace probing should just substitute existing printf(9) debugging.

Committed years ago.

Merge remote-tracking branch 'FreeBSD/master' into pbuf to take
in and resolve with recent D16666 check-in.

Initialize compatibility epoch tracker for thread0. Fixes

Fix build on some architectures after r340413. On amd64 epoch.h

style(9), mostly adjusting overly long lines.

With epoch not inlined, there is no point in using _lite KPI. While here,

Plug build break after r340411.

glebius (Gleb Smirnoff)
User

Projects

User Details

Recent Activity
View All

Today

Yesterday

Mon, Feb 18

Fri, Feb 15

Mon, Feb 11

Sat, Feb 9

Fri, Feb 8

Thu, Feb 7

Sun, Feb 3

Sat, Feb 2

Fri, Feb 1

Thu, Jan 31

Tue, Jan 29

Fri, Jan 25

Thu, Jan 24

Jan 16 2019

Jan 15 2019

Jan 14 2019

Jan 11 2019

Jan 10 2019

Jan 9 2019

Jan 8 2019

Jan 7 2019

Dec 18 2018

Dec 14 2018

Dec 13 2018

Dec 7 2018

Dec 5 2018

Dec 4 2018

Dec 3 2018

Nov 28 2018

Nov 27 2018

Nov 26 2018

Nov 14 2018

Nov 13 2018

glebius (Gleb Smirnoff)User