Page MenuHomeFreeBSD

tuexen (Michael Tuexen)
User

Projects

User Details

User Since
Feb 4 2016, 4:45 PM (428 w, 6 d)

Recent Activity

Sun, Apr 21

tuexen requested changes to D44889: tcp_bbr.4: Add a deprecation notice.

The decision is to keep BBR. See https://wiki.freebsd.org/TransportProtocols/18April2024.

Sun, Apr 21, 9:52 AM

Fri, Apr 19

tuexen abandoned D42084: Improve handling of detached inps.
Fri, Apr 19, 4:56 PM
tuexen closed D44746: tcp: drop data received after a FIN has been processed.

Committed in c9cd686bd4a0.

Fri, Apr 19, 4:55 PM

Thu, Apr 18

tuexen closed D44648: tcp rack: improve BBR_LOG_CWND event.
Thu, Apr 18, 8:28 PM
tuexen committed rG1941914d3bb4: tcp rack: improve BBR_LOG_CWND event (authored by tuexen).
tcp rack: improve BBR_LOG_CWND event
Thu, Apr 18, 8:28 PM
tuexen committed rGc9cd686bd4a0: tcp: drop data received after a FIN has been processed (authored by tuexen).
tcp: drop data received after a FIN has been processed
Thu, Apr 18, 8:22 PM

Wed, Apr 17

tuexen committed rG994f9c9e3c2c: tcp rack: fix sending (authored by tuexen).
tcp rack: fix sending
Wed, Apr 17, 2:20 PM
tuexen added a reverting change for rGb5ee7411bfab: tcp rack: fix sending: rG134884436681: Revert "tcp rack: fix sending".
Wed, Apr 17, 2:17 PM
tuexen committed rG134884436681: Revert "tcp rack: fix sending" (authored by tuexen).
Revert "tcp rack: fix sending"
Wed, Apr 17, 2:17 PM
tuexen added a reverting change for D44610: tcp rack: fix sending : rG134884436681: Revert "tcp rack: fix sending".
Wed, Apr 17, 2:17 PM
tuexen committed rGb5ee7411bfab: tcp rack: fix sending (authored by tuexen).
tcp rack: fix sending
Wed, Apr 17, 2:16 PM
tuexen committed rG14d778433227: tcp bbr: improve code consistency (authored by tuexen).
tcp bbr: improve code consistency
Wed, Apr 17, 2:03 PM
tuexen committed rG5344772af985: tcp: add some debug output (authored by tuexen).
tcp: add some debug output
Wed, Apr 17, 2:02 PM
tuexen committed rG90cde57af8c6: tcp: improve consistency (authored by tuexen).
tcp: improve consistency
Wed, Apr 17, 2:01 PM
tuexen committed rGc6770e8c99ab: tcp rack: fix memory corruption (authored by tuexen).
tcp rack: fix memory corruption
Wed, Apr 17, 2:00 PM
tuexen committed rG7a019565ff6d: uipc_socket: handle socket buffer locks in sopeeloff (authored by tuexen).
uipc_socket: handle socket buffer locks in sopeeloff
Wed, Apr 17, 1:59 PM
tuexen committed rG6572d12cfc67: tcp bblog: cleanup (authored by tuexen).
tcp bblog: cleanup
Wed, Apr 17, 1:56 PM
tuexen committed rG60898a7cef0b: tcp hpts: initialize variable (authored by tuexen).
tcp hpts: initialize variable
Wed, Apr 17, 1:56 PM
tuexen committed rGf3e2beb306c6: tcp: address a warning (authored by tuexen).
tcp: address a warning
Wed, Apr 17, 1:55 PM
tuexen committed rG7ebfafa813ce: tcp: fix conversion of rttvar (authored by tuexen).
tcp: fix conversion of rttvar
Wed, Apr 17, 1:55 PM
tuexen committed rGbe30b22fbb77: tcp: fix comment (authored by tuexen).
tcp: fix comment
Wed, Apr 17, 1:53 PM
tuexen committed rG8375db082c6b: tcp hpts: improve consistency (authored by tuexen).
tcp hpts: improve consistency
Wed, Apr 17, 1:52 PM
tuexen committed rGfbc8dfd0ed80: tcp bblog: use correct length (authored by tuexen).
tcp bblog: use correct length
Wed, Apr 17, 1:51 PM
tuexen committed rGfd2a580db248: tcp: no data on SYN segments unless doing TFO (authored by tuexen).
tcp: no data on SYN segments unless doing TFO
Wed, Apr 17, 1:50 PM
tuexen committed rGd69099e4330e: TCP LRO: add dtrace probe points (authored by tuexen).
TCP LRO: add dtrace probe points
Wed, Apr 17, 1:49 PM
tuexen committed rG8244b35ff88c: TCP LRO: disable mbuf queuing when packet filter hooks are in place (authored by tuexen).
TCP LRO: disable mbuf queuing when packet filter hooks are in place
Wed, Apr 17, 1:49 PM
tuexen committed rGbe7586fe468f: sctp: improve sending of packets containing an INIT ACK chunk (authored by tuexen).
sctp: improve sending of packets containing an INIT ACK chunk
Wed, Apr 17, 1:48 PM
tuexen committed rG49557689bf1e: sctp: improve consistency (authored by tuexen).
sctp: improve consistency
Wed, Apr 17, 1:47 PM
tuexen committed rGc5e78e553c66: dtrace: allow NULL interface pointer for ifinfo_t translator (authored by tuexen).
dtrace: allow NULL interface pointer for ifinfo_t translator
Wed, Apr 17, 1:47 PM
tuexen committed rG8b2c694d3836: RACK, BBR: handle EACCES like EPERM for IP output handling (authored by tuexen).
RACK, BBR: handle EACCES like EPERM for IP output handling
Wed, Apr 17, 1:46 PM
tuexen committed rG307daa4f1759: vnic: fix compilation of an IPv6 only kernel (authored by tuexen).
vnic: fix compilation of an IPv6 only kernel
Wed, Apr 17, 1:44 PM
tuexen committed rG2da45e304dba: TCP LRO: convert TCP header fields to host byte order earlier (authored by tuexen).
TCP LRO: convert TCP header fields to host byte order earlier
Wed, Apr 17, 1:43 PM

Mon, Apr 15

tuexen closed D44800: tcp bbr: improve code consistency.
Mon, Apr 15, 9:54 PM
tuexen committed rG605a00660ead: tcp bbr: improve code consistency (authored by tuexen).
tcp bbr: improve code consistency
Mon, Apr 15, 9:54 PM
tuexen retitled D44800: tcp bbr: improve code consistency from tcp bbr: fix memory corruption to tcp bbr: improve code consistency.
Mon, Apr 15, 8:10 PM
tuexen updated the diff for D44800: tcp bbr: improve code consistency.

Remove initialisation, which is not needed anymore.

Mon, Apr 15, 8:10 PM
tuexen requested review of D44800: tcp bbr: improve code consistency.
Mon, Apr 15, 8:01 PM
tuexen added a comment to D44746: tcp: drop data received after a FIN has been processed.
In D44746#1019854, @gbe wrote:

Should BBR not also be affected by this change?

Possibly. But BBR is not kept in sync, it is more deprecated...

Mon, Apr 15, 10:50 AM
tuexen updated the diff for D44746: tcp: drop data received after a FIN has been processed.

Same fix for the BBR stack.

Mon, Apr 15, 10:49 AM

Fri, Apr 12

tuexen added a comment to D44746: tcp: drop data received after a FIN has been processed.
In D44746#1019926, @gbe wrote:
In D44746#1019854, @gbe wrote:

Should BBR not also be affected by this change?

Possibly. But BBR is not kept in sync, it is more deprecated...

As the author the manual pages for both tcp_rack(4) and tcp_bbr(4), should I update the bbr man page and add a deprecation notice somehow?

Since 14.1-RELEASE is approaching, it would be good to have updated documentation, or a complete removal of BBR in place.

Fri, Apr 12, 8:41 AM

Thu, Apr 11

tuexen added a comment to D44746: tcp: drop data received after a FIN has been processed.
In D44746#1019854, @gbe wrote:

Should BBR not also be affected by this change?

Thu, Apr 11, 8:54 AM

Wed, Apr 10

tuexen requested review of D44746: tcp: drop data received after a FIN has been processed.
Wed, Apr 10, 10:30 PM
tuexen accepted D44734: Create little-endian hints files on all architectures.

I tested this patch on a BE system, the hints file is written in LE and can be used without problems.
I personally would commit the optimization (where there is not functional change intended) separately
from the change of the byte ordering of the hints file.

Wed, Apr 10, 2:33 PM
tuexen added a comment to D44734: Create little-endian hints files on all architectures.

I'll test the patch on a BE system and report back.

Wed, Apr 10, 8:50 AM

Sun, Apr 7

tuexen committed rGe8c149ab85c7: tcp: add some debug output (authored by tuexen).
tcp: add some debug output
Sun, Apr 7, 8:46 PM
tuexen closed D44669: tcp: add some debug output.
Sun, Apr 7, 8:46 PM
tuexen requested review of D44669: tcp: add some debug output.
Sun, Apr 7, 11:30 AM

Sat, Apr 6

tuexen requested review of D44648: tcp rack: improve BBR_LOG_CWND event.
Sat, Apr 6, 8:16 AM
tuexen closed D44645: tcp: improve consistency.
Sat, Apr 6, 8:06 AM
tuexen committed rG3e1c8a35f741: tcp: improve consistency (authored by tuexen).
tcp: improve consistency
Sat, Apr 6, 8:06 AM
tuexen closed D44647: tcp rack: fix memory corruption.
Sat, Apr 6, 8:02 AM
tuexen committed rGd902c8f55b8d: tcp rack: fix memory corruption (authored by tuexen).
tcp rack: fix memory corruption
Sat, Apr 6, 8:02 AM

Fri, Apr 5

tuexen requested review of D44647: tcp rack: fix memory corruption.
Fri, Apr 5, 8:22 PM
tuexen requested review of D44645: tcp: improve consistency.
Fri, Apr 5, 5:56 PM
tuexen committed rG681711b77cde: uipc_socket: handle socket buffer locks in sopeeloff (authored by tuexen).
uipc_socket: handle socket buffer locks in sopeeloff
Fri, Apr 5, 4:23 PM
tuexen closed D44640: uipc_socket: handle socket buffer locks in sopeeloff.
Fri, Apr 5, 4:23 PM
tuexen committed rG7df0ef5f48e1: tcp rack: fix sending (authored by tuexen).
tcp rack: fix sending
Fri, Apr 5, 3:51 PM
tuexen closed D44610: tcp rack: fix sending .
Fri, Apr 5, 3:51 PM
tuexen committed rG60bc195745b8: tcp bblog: cleanup (authored by tuexen).
tcp bblog: cleanup
Fri, Apr 5, 3:41 PM
tuexen closed D44616: tcp bblog: cleanup.
Fri, Apr 5, 3:41 PM
tuexen retitled D44616: tcp bblog: cleanup from tcp hpts: cleanup to tcp bblog: cleanup.
Fri, Apr 5, 3:36 PM
tuexen committed rGaaaa01c0c858: tcp hpts: initialize variable (authored by tuexen).
tcp hpts: initialize variable
Fri, Apr 5, 3:35 PM
tuexen closed D44584: tcp hpts: initialize variable.
Fri, Apr 5, 3:34 PM
tuexen added a comment to D44640: uipc_socket: handle socket buffer locks in sopeeloff.

Can we please add some basic SCTP regression tests to the tree to help catch problems like this in the future?

Fri, Apr 5, 2:24 PM

Thu, Apr 4

tuexen requested review of D44640: uipc_socket: handle socket buffer locks in sopeeloff.
Thu, Apr 4, 9:28 PM
tuexen accepted D43470: tcp: refactor cwnd during SACK transmissions and enable TSO.
Thu, Apr 4, 12:47 PM
tuexen accepted D43355: tcp: fix erroneous transmission selection after RTO w/ SACK incoming.
Thu, Apr 4, 12:45 PM
tuexen added a reviewer for D44584: tcp hpts: initialize variable: lstewart.
Thu, Apr 4, 9:33 AM
tuexen added a comment to D44632: sockets: Add hhook in sonewconn for inheriting OSD specific data.

I was about to ask "where is #ifdef TCP_HHOOKS?" But then I realized that socket_hhh isn't hidden under any ifdef.

Unrelated to this review, but may I ask if it is possible to hide socket_hhh and other stuff that came in 4ec7371233268 under TCP_HHOOKS? Or maybe SOCKET_HHOOKS, if these two are used as two independent features.

Thu, Apr 4, 8:12 AM
tuexen accepted D44632: sockets: Add hhook in sonewconn for inheriting OSD specific data.
Thu, Apr 4, 8:05 AM

Wed, Apr 3

tuexen closed D44619: tcp: address warning.
Wed, Apr 3, 10:19 PM
tuexen committed rG6b454da6bbaa: tcp: address a warning (authored by tuexen).
tcp: address a warning
Wed, Apr 3, 10:19 PM
tuexen requested review of D44619: tcp: address warning.
Wed, Apr 3, 8:17 PM
tuexen closed D44612: tcp: fix conversion of rttvar.
Wed, Apr 3, 4:44 PM
tuexen committed rGe0bd180130b8: tcp: fix conversion of rttvar (authored by tuexen).
tcp: fix conversion of rttvar
Wed, Apr 3, 4:44 PM
tuexen added a comment to D44610: tcp rack: fix sending .

if idle is used as a boolean, why not be consequent about it and declare it as such also? To prevent future misuses with numeric calculations...

Wed, Apr 3, 4:37 PM
tuexen closed D44611: tcp: fix comment.
Wed, Apr 3, 4:29 PM
tuexen committed rG5a268d868890: tcp: fix comment (authored by tuexen).
tcp: fix comment
Wed, Apr 3, 4:29 PM
tuexen requested review of D44616: tcp bblog: cleanup.
Wed, Apr 3, 3:18 PM
tuexen requested review of D44612: tcp: fix conversion of rttvar.
Wed, Apr 3, 12:02 PM
tuexen requested review of D44611: tcp: fix comment.
Wed, Apr 3, 12:00 PM
tuexen requested review of D44610: tcp rack: fix sending .
Wed, Apr 3, 10:01 AM

Mon, Apr 1

tuexen requested review of D44584: tcp hpts: initialize variable.
Mon, Apr 1, 10:39 PM
tuexen committed rGb600644fdd6c: tcp hpts: improve consistency (authored by tuexen).
tcp hpts: improve consistency
Mon, Apr 1, 7:57 PM
tuexen closed D44527: TCP HPTS: improve consistency.
Mon, Apr 1, 7:57 PM
tuexen accepted D44222: ipsec_accel: handle TSO if supported.

Can't test it, but it looks OK.

Mon, Apr 1, 6:04 PM
tuexen added inline comments to D44222: ipsec_accel: handle TSO if supported.
Mon, Apr 1, 6:00 PM
tuexen added inline comments to D44222: ipsec_accel: handle TSO if supported.
Mon, Apr 1, 12:38 PM
tuexen added inline comments to D43213: tcp: retry CWR transmission if TSO doesn't support ECN.
Mon, Apr 1, 12:36 PM

Wed, Mar 27

tuexen requested review of D44527: TCP HPTS: improve consistency.
Wed, Mar 27, 2:06 PM
tuexen added a comment to D44510: tcp bblog: use correct length.

A sidenote: nginx practice to prevent this kind of bugs is this:

strlcpy(entry->tldl_reason, "UNKNOWN", sizeof("UNKNOWN"));

I am confused: isn't the last parameter of strlcpy the size of the destination buffer, not the length of the source. So if you write something like

strlcpy(entry->tldl_reason, "This string is much longer than the buffer size of thirtytwo bytes", sizeof("This string is much longer than the buffer size of thirtytwo bytes"));

will result in a buffer overflow.
Something like

strlcpy(entry->tldl_reason, "This string is much longer than the buffer size of thirtytwo bytes", TCP_LOG_REASON_LEN);

would truncate the string correctly and assure that tldl_reason is NUL terminated.

I would suggest to further improve this:

#define STRWSIZE0(s) #s, sizeof(#s)
#define STRWSIZE(s) #s, sizeof(#s) - 1

to prevent probability of typing a long string twice differently.

Wed, Mar 27, 2:01 PM
tuexen committed rGed505f893ab0: tcp bblog: use correct length (authored by tuexen).
tcp bblog: use correct length
Wed, Mar 27, 1:36 PM
tuexen closed D44510: tcp bblog: use correct length.
Wed, Mar 27, 1:35 PM
tuexen accepted D44420: Optimize HPTS so that little work is done until we have a hpts thread that is over the connection threshold.
Wed, Mar 27, 1:18 PM

Tue, Mar 26

tuexen requested review of D44510: tcp bblog: use correct length.
Tue, Mar 26, 12:20 PM

Mar 24 2024

tuexen accepted D44491: inpcb: fully retire inp_ppcb pointer.
Mar 24 2024, 6:17 PM

Mar 23 2024

tuexen accepted D44482: icmp6: bring rate limiting on a par with IPv4.
Mar 23 2024, 2:23 PM
tuexen accepted D44479: icmp6: make icmp6_ratelimit() responsible to update the stats counter.
Mar 23 2024, 2:22 PM
tuexen accepted D44480: icmp6: rate limit our echo replies.
Mar 23 2024, 2:21 PM
tuexen accepted D44475: icmp: do not store per-VNET identical array of strings.
Mar 23 2024, 2:20 PM