#
5873c738 |
| 29-Oct-2024 |
dlg <dlg@openbsd.org> |
use nsecuptime instead of using nanouptime and doing a bunch of maths.
ok claudio@
|
#
df8d9afd |
| 02-Jan-2022 |
jsg <jsg@openbsd.org> |
spelling ok jmc@ reads ok tb@
|
#
0e8d4891 |
| 10-Dec-2020 |
dlg <dlg@openbsd.org> |
classify packets without a flowid into bucket 0, not a random bucket.
putting packets into random buckets means packets in a flow/connection will be reorderd. pf assigns a flowid if it's enabled, an
classify packets without a flowid into bucket 0, not a random bucket.
putting packets into random buckets means packets in a flow/connection will be reorderd. pf assigns a flowid if it's enabled, and you need pf to configure code, so it's reasonable to assume that most packets will have a flowid. using bucket 0 like this is what we do in most other places that bin packets with the flowid.
show more ...
|
#
5e41c075 |
| 18-Jun-2020 |
dlg <dlg@openbsd.org> |
add $OpenBSD$ tags
|
#
379654e8 |
| 17-Jun-2020 |
dlg <dlg@openbsd.org> |
make ph_flowid in mbufs 16bits by storing whether it's set in csum_flags.
i've been wanting to do this for a while, and now that we've got stoeplitz and it gives us 16 bits, it seems like the right
make ph_flowid in mbufs 16bits by storing whether it's set in csum_flags.
i've been wanting to do this for a while, and now that we've got stoeplitz and it gives us 16 bits, it seems like the right time.
show more ...
|
#
57e2f549 |
| 27-Nov-2017 |
sthen <sthen@openbsd.org> |
lenght->length, mostly in comments
|
#
b22f13b1 |
| 24-Jul-2017 |
mikeb <mikeb@openbsd.org> |
Fixup free list handling in fqcodel_deq_begin
We're growing an mbuf free list while iterating over flow queues and need to adjust our internal statistics on every iteration by using a portion of the
Fixup free list handling in fqcodel_deq_begin
We're growing an mbuf free list while iterating over flow queues and need to adjust our internal statistics on every iteration by using a portion of the free list corresponding to the current iteration.
show more ...
|
#
f18829cd |
| 28-Jun-2017 |
mikeb <mikeb@openbsd.org> |
Provide pf queue manager API
Discussed with and OK henning@ at d2k17 as a part of a larger diff.
|
#
255c5c6e |
| 24-May-2017 |
mikeb <mikeb@openbsd.org> |
Fail with EINVAL when asked to create a non-root queue instead of panicking
Prompted by a bug report from semarie@, thanks!
|
#
d7a2b5de |
| 15-May-2017 |
jsg <jsg@openbsd.org> |
clang warns on unused static const variables. Remove one such unused variable so the kernel will build on arm64 again.
|
#
76bd1156 |
| 12-May-2017 |
mikeb <mikeb@openbsd.org> |
Export delay statistics instead of bucket information
|
#
65ed68d6 |
| 11-May-2017 |
mikeb <mikeb@openbsd.org> |
Remove a leftover from the ifq_mfreeml conversion
|
#
50f8ee3e |
| 11-May-2017 |
mikeb <mikeb@openbsd.org> |
Add pf queue configuration glue for FQ-CoDel
|
#
269ba8f5 |
| 11-May-2017 |
mikeb <mikeb@openbsd.org> |
Increment drop counter on the enqueue failure
|
#
7a7c13ac |
| 07-May-2017 |
mikeb <mikeb@openbsd.org> |
Fix stage transition from the initial one to DROPPING
When the initial state is set to DROPPING, the code immediately jumps to a CONTROL state bypassing DROPPING. To fix this we start with an explic
Fix stage transition from the initial one to DROPPING
When the initial state is set to DROPPING, the code immediately jumps to a CONTROL state bypassing DROPPING. To fix this we start with an explicit INITIAL state so that we do an INITIAL->DROPPING transition right off the bat in the beginning of the loop and then perform a DROPPING->CONTROL and either CONTROL->DROPPING and restart the loop or CONTROL->RECOVERY/ACCEPTING and terminate.
show more ...
|
#
4fa13868 |
| 04-May-2017 |
mikeb <mikeb@openbsd.org> |
Implementation of the Flow Queue - Controlled Delay (FQ-CoDel)
The purpose of FQ-CoDel is to provide fair sharing of bandwidth between simultaneous connections and reduce latency differences among t
Implementation of the Flow Queue - Controlled Delay (FQ-CoDel)
The purpose of FQ-CoDel is to provide fair sharing of bandwidth between simultaneous connections and reduce latency differences among them.
OK mpi, sthen, visa
show more ...
|