User Details
User Details
- User Since
- May 28 2014, 2:27 PM (601 w, 5 d)
Yesterday
Yesterday
glebius added a comment to D53872: bpf: modularize ifnet(9) part of bpf.
lltable: use own lock
glebius committed rG607f11055d2d: linux: store Linux Ethernet interface number in struct ifnet (authored by glebius).
linux: store Linux Ethernet interface number in struct ifnet
glebius committed rGfbf05d2147b1: linux: separate all ifnet(9) related code into linux_ifnet.c (authored by glebius).
linux: separate all ifnet(9) related code into linux_ifnet.c
Sun, Dec 7
Sun, Dec 7
glebius added a comment to D54109: netlink: use same cred as caller's cred similar to ifioctl.
Looks correct! Thanks. What did happen before? This command failed?
Sat, Dec 6
Sat, Dec 6
tcp: fix build with RSS
net: fix LINT-NOIP build
Fri, Dec 5
Fri, Dec 5
glebius committed rGbc14dcd113ae: tests/if_bridge: mark several tests as requring vlan(4) or gif(4) (authored by glebius).
tests/if_bridge: mark several tests as requring vlan(4) or gif(4)
glebius requested review of D54089: net: attach IPv4 and IPv6 stacks to an interface with EVENTHANDLER(9).
glebius requested review of D54086: lltable: use own lock.
glebius requested review of D54085: net: invoke ifnet_arrival_event handlers before linking in.
glebius committed rG8291ea8761c2: netinet6: use IF_ADDR_LOCK instead of IF_AFDATA_LOCK in defrtr_ipv6_only_ifp (authored by glebius).
netinet6: use IF_ADDR_LOCK instead of IF_AFDATA_LOCK in defrtr_ipv6_only_ifp
glebius committed rG59062fd0b2f1: netinet6: use IF_ADDR_LOCK instead of IF_AFDATA_LOCK (authored by glebius).
netinet6: use IF_ADDR_LOCK instead of IF_AFDATA_LOCK
glebius updated the diff for D54077: linux: store Linux Ethernet interface number in struct ifnet.
- Make the netlink_writer cookie an honest const ifnet pointer.
- Refactor some code to provide the new constness constraint.
Thu, Dec 4
Thu, Dec 4
glebius requested review of D54077: linux: store Linux Ethernet interface number in struct ifnet.
glebius requested review of D54076: linux: separate all ifnet(9) related code into linux_ifnet.c.
net: remove dom_ifmtu
net: routing table attach never fails
glebius committed rG27ff90cd3d8d: tests/carp: make a 0.2 second pause before configuring second jail (authored by glebius).
tests/carp: make a 0.2 second pause before configuring second jail
glebius committed rGe353cbffd600: tests/carp: make sleep interval in the wait loop smaller (authored by glebius).
tests/carp: make sleep interval in the wait loop smaller
sendfile: remove SF_SYNC in FreeBSD 16.0
pf: make unhandled_af() inline
glebius requested review of D54064: pf: make unhandled_af() inline.
glebius requested review of D54063: usb: don't create ifnet(9) for usbus devices.
glebius updated the diff for D53877: ipfw: create a bpf tap point for every log rule.
- Address the fact that many rules can have the same number.
- Support 32-bit rule numbers.
glebius updated the diff for D53872: bpf: modularize ifnet(9) part of bpf.
- Rebase + unsignigicant changes.
glebius updated the diff for D53871: bpf: don't recursively enter the network epoch in bpf taps.
- Address USB.
glebius updated the diff for D53870: bpf: convert several boolean natured fields of bpf_d to flags.
The d_promisc appears to be locked by the global BPF_LOCK(). Thus, remove it
from this changes. This addresses Mark's comments.
Wed, Dec 3
Wed, Dec 3
bpf: global bpf list doesn't need CK
glebius committed rG845896655815: bpf: calculate net.bpf.stats buffer size dynamically (authored by glebius).
bpf: calculate net.bpf.stats buffer size dynamically
bpf: retire struct bpf_if_ext
tcp: store flowid info in syncache
tcp: don't set flowid in tcp_input()
glebius added a comment to D54053: lacp: Sort port map by interface index.
My understanding. First, qsort() itself never passes NULL pointer. Second, could a map[] entry within p->pm_count be NULL or not? The new code assumes it can't be. Then we also have a case with lacp_port that has NULL lp_ifp, which is covered.
glebius added a comment to D54053: lacp: Sort port map by interface index.
If pointers in the map[] can be NULL, then you removed too much, didn't you?
minidlna: wrap up a sub-release
glebius added inline comments to D54053: lacp: Sort port map by interface index.
Tue, Nov 25
Tue, Nov 25
glebius added inline comments to D53870: bpf: convert several boolean natured fields of bpf_d to flags.
glebius updated the diff for D53868: bpf: calculate net.bpf.stats buffer size dynamically.
- Apply Mark's suggestion
glebius added inline comments to D53877: ipfw: create a bpf tap point for every log rule.
Mon, Nov 24
Mon, Nov 24
glebius added a comment to D53899: arp: Reduce lifetime of ARP entries.
AFAIU, when arptimer() finds entry in state ARP_LLINFO_REACHABLE it gives it arp_rexmit time (1 second) to get feedback from the forwarding path and sets its state to ARP_LLINFO_VERIFY. In 1 second if feedback was collected, it will issue arprequest(). My concern is that the jitter is bigger than 1 second. I could be wrong, and my concern is erroneous. I just wanted to make sure that this specific behavior is checked to persist. I don't think we have a regression test for it.
glebius added a comment to D53899: arp: Reduce lifetime of ARP entries.
In the past we used to have a logic that would renew ARP entires before they expire. The ARP request is sent before an entry is about to be expired. The goal is to avoid to have a doubled RTT on normal traffic every time we renew the entry. I did a quick look around, and failed to find the logic. Is it still there?
Sat, Nov 22
Sat, Nov 22
glebius updated the summary of D51792: tcp: store flowid info in syncache.
glebius requested review of D53886: tcp: don't set flowid in tcp_input().
glebius updated the diff for D51792: tcp: store flowid info in syncache.
- Rebase on main.
- Separate tcp_input() change into its own commit.
glebius added a reviewer for D53877: ipfw: create a bpf tap point for every log rule: lytboris_gmail.com.
glebius updated the diff for D53872: bpf: modularize ifnet(9) part of bpf.
- Rebase
- Build with MAC
- Address bz's comment
bpf: remove DDB code
Fri, Nov 21
Fri, Nov 21
glebius added a reviewer for D53876: ifconfig: print warning and return success on ipfw0, ipfwlog0 cloning: ae.
glebius added a reviewer for D53875: ipfw: create "ipfw0" and "ipfwlog0" bpf tapping points without ifnet(9): ae.
glebius requested review of D53877: ipfw: create a bpf tap point for every log rule.
glebius requested review of D53876: ifconfig: print warning and return success on ipfw0, ipfwlog0 cloning.
glebius requested review of D53875: ipfw: create "ipfw0" and "ipfwlog0" bpf tapping points without ifnet(9).
glebius requested review of D53874: libpcap: obtain list of bpf tap points with BIOCGETIFLIST.
glebius requested review of D53873: bpf: add BIOCGETIFLIST ioctl that returns all available tap points.
glebius requested review of D53872: bpf: modularize ifnet(9) part of bpf.
glebius requested review of D53871: bpf: don't recursively enter the network epoch in bpf taps.
glebius requested review of D53870: bpf: convert several boolean natured fields of bpf_d to flags.
glebius requested review of D53869: bpf: global bpf list doesn't need CK.
glebius requested review of D53868: bpf: calculate net.bpf.stats buffer size dynamically.
glebius requested review of D53867: bpf: retire struct bpf_if_ext.
glebius committed rGfd91012ebff2: bpf: leave only locked version of bpf_detachd() (authored by glebius).
bpf: leave only locked version of bpf_detachd()
glebius committed rGff3ccf6f1a02: bpf: refactor buffer pre-allocation for BIOCSETIF (authored by glebius).
bpf: refactor buffer pre-allocation for BIOCSETIF
bpf: remove dead code
glebius committed rG88b38d43f557: ipfw: add extra parenthesis around ACTION_PTR() macro (authored by glebius).
ipfw: add extra parenthesis around ACTION_PTR() macro
mbuf: allow const pointer for m_rcvif()
tests/net: add some bpf(4) tests
glebius committed rGcf95eabb5572: netlink: in snl_init_writer() don't overwrite error in case of failure (authored by Bruno Silvestre <bruno.silvestre@gmail.com>).
netlink: in snl_init_writer() don't overwrite error in case of failure
Fri, Nov 14
Fri, Nov 14
glebius committed rG82d8a5029a80: unix/stream: fix instant panic w/o INVARIANTS (authored by glebius).
unix/stream: fix instant panic w/o INVARIANTS
glebius committed rG69f61cee2efb: unix/stream: fix a race with MSG_PEEK on SOCK_SEQPACKET with MSG_EOR (authored by glebius).
unix/stream: fix a race with MSG_PEEK on SOCK_SEQPACKET with MSG_EOR
Thu, Nov 13
Thu, Nov 13
glebius added inline comments to D53517: Add ecn(9) manual.
glebius added inline comments to D53516: Update ip_ecn to RFC 6040.
Tue, Nov 11
Tue, Nov 11
glebius added inline comments to D53632: unix/stream: fix a race with MSG_PEEK on SOCK_SEQPACKET with MSG_EOR.
glebius updated the diff for D53632: unix/stream: fix a race with MSG_PEEK on SOCK_SEQPACKET with MSG_EOR.
- Keep the error case freeing loop under the lock.
- Assert before subtracting.
Mon, Nov 10
Mon, Nov 10
glebius committed rGa3cd3a4fd68c: netlink: plug lock leak in genl_register_family() (authored by Bruno Silvestre <bruno.silvestre@gmail.com>).
netlink: plug lock leak in genl_register_family()
Sun, Nov 9
Sun, Nov 9
glebius updated the diff for D53632: unix/stream: fix a race with MSG_PEEK on SOCK_SEQPACKET with MSG_EOR.
Fix pointer for partial mbuf in peek case.
Nov 8 2025
Nov 8 2025
glebius updated the summary of D53632: unix/stream: fix a race with MSG_PEEK on SOCK_SEQPACKET with MSG_EOR.
glebius updated the diff for D53632: unix/stream: fix a race with MSG_PEEK on SOCK_SEQPACKET with MSG_EOR.
New version.