On Wed, Oct 25, 2023 at 07:00:28PM +0200, Omar Polo wrote:
> On 2023/10/25 13:38:37 +0200, Alexander Bluhm wrote:
> > @@ -213,11 +214,27 @@ main(int argc, char *argv[])
> > perror("unveil");
> > my_exit(2);
> >
On Thu, Oct 19, 2023 at 04:04:26PM +0200, Jan Klemkow wrote:
> On Wed, Oct 18, 2023 at 08:53:44PM +0200, Alexander Bluhm wrote:
> > On Wed, Oct 18, 2023 at 08:19:29PM +0200, Mischa wrote:
> > > It's indeed something like that: ix -> vlan (tagged) -> veb
> >
&g
Hi,
Since 7.4 patch(1) does not work if an explicit patchfile is given on
command line.
https://marc.info/?l=openbsd-cvs&m=168941770509379&w=2
root@ot14:.../~# patch /usr/src/usr.bin/patch/patch.c patch-unveil.diff
Hmm... Looks like a unified diff to me...
The text leading up to this was:
-
Errata patches for X11 server and kernel mbuf handling have been
released for OpenBSD 7.3 and 7.4.
Binary updates for the amd64, arm64 and i386 platform are available
via the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.org/errata73.html
On Mon, Oct 16, 2023 at 10:17:50PM +0300, Vitaliy Makkoveev wrote:
> This socket comes from userland, so this flag is never set. This makes
> SB_NOINTR flag immutable, because we only set this bit on NFS client
> socket buffers for all it's lifetime.
>
> I want to do this because this flag modifie
On Wed, Oct 18, 2023 at 05:29:41PM +0200, Jan Klemkow wrote:
> This diff implements TCP Segmentation Offloading for ixl(4). I tested
> it successfully on amd64 and sparc64 with Intel X710. It should
> increase the TCP bulk performance to 10 Gbit/s. On sparc64 I got an
> increase from 600 MBit/s
Hi,
Now that syslogd handles delayed DNS lookups, I think we should
count dropped packets to UDP loghosts. Although not every dropped
UDP packet can be detected, the message makes the admin aware that
there may be a blind spot during startup.
syslogd[26493]: dropped 8 messages to udp loghost "@u
Hi,
If a packet is malformed, it is dropped by pf(4). The rule referenced
in pflog(4) is the default rule. As the default rule is a pass
rule, tcpdump prints "pass" although the packet is actually dropped.
I have reports from genua and OPNsense users who are confused by
the output.
With the dif
Hi,
The behaviour of the PFRULE_SRCTRACK and max_states check was
unintentionally changed by this commit.
revision 1.964
date: 2016/01/25 18:49:57; author: sashan; state: Exp; lines: +18 -10;
commitid: KeemoLxcm7FS1oYy;
- plugging massive pf_state_key leak
OK mp
Hi,
pf_pull_hdr() allows to pass an action pointer parameter as output
value. This is never used, all callers pass a NULL argument. Remove
ACTION_SET() entirely.
The logic if (fragoff >= len) in pf_pull_hdr() looks odd. One is
the offset in the IP packet, the latter the length of some header
w
On Fri, Oct 06, 2023 at 04:17:34PM +0300, Vitaliy Makkoveev wrote:
> On Fri, Oct 06, 2023 at 02:50:21PM +0200, Alexander Bluhm wrote:
> > On Fri, Oct 06, 2023 at 12:45:54PM +0300, Vitaliy Makkoveev wrote:
> > > On Thu, Oct 05, 2023 at 10:42:25PM +1000, David Gwynne wrote:
>
On Fri, Oct 06, 2023 at 03:47:31PM +0300, Vitaliy Makkoveev wrote:
> On Fri, Oct 06, 2023 at 02:14:52PM +0200, Alexander Bluhm wrote:
> > > @@ -718,11 +743,13 @@ softclock(void *arg)
> > > softclock_process_tick_timeout(to, new);
> > > }
On Fri, Oct 06, 2023 at 12:45:54PM +0300, Vitaliy Makkoveev wrote:
> On Thu, Oct 05, 2023 at 10:42:25PM +1000, David Gwynne wrote:
> > > On 5 Oct 2023, at 21:50, Vitaliy Makkoveev wrote:
> > >
> > > I don't like this unlocked if_flags check we have in ifq_start_task().
> > > Guess READ_ONCE() is
On Fri, Oct 06, 2023 at 02:12:31PM +0300, Vitaliy Makkoveev wrote:
> I reworked your diff for a little. At firts I use separate
> softclock_thread_mpsafe() for mpsafe timers. I don't think we need to
> bind this thread to the primary CPU.
I also think a separate thread is better. Makes no sense t
On Thu, Oct 05, 2023 at 07:15:23AM +0200, Kirill Miazine wrote:
> > This diff checks IFF_RUNNING flag within while (!ifq_empty()) loop of
> > wg_peer_destroy(). If the flag is not set queue will be purged and check
> > performed again. I intentionally keep netlock to prevent ifconfig
> > manipulati
On Sun, Sep 03, 2023 at 02:00:46AM +0200, Alexander Bluhm wrote:
> When DNS lookup for remote loghost in @ line in syslog.conf does
> not work at startup, retry in intervals.
Thanks Paul de Weerd for testing my diff. Together we improved the
debug output to make clear what is going on.
Hi,
This is a first step to unlock TCP syn cache. The timer function
is independent of the socket code. That makes it easy to start
there.
Introduce tcp syn cache mutex. Document fields protected by net
lock and mutex. Devide timer function in parts protected by mutex
and sending with netlock
On Wed, Oct 04, 2023 at 11:03:27PM +0300, Vitaliy Makkoveev wrote:
> On Wed, Oct 04, 2023 at 09:13:59PM +0200, Alexander Bluhm wrote:
> > On Wed, Oct 04, 2023 at 08:42:48PM +0200, Kirill Miazine wrote:
> > > > If it happns again, could you send an 'ps axlww | grep ifcon
On Wed, Oct 04, 2023 at 08:42:48PM +0200, Kirill Miazine wrote:
> > If it happns again, could you send an 'ps axlww | grep ifconifg'
> > output? Then we see the wait channel where it hangs in the kernel.
> >
> > $ ps axlww
> >UID PID PPID CPU PRI NI VSZ RSS WCHAN STAT TT TIM
On Wed, Oct 04, 2023 at 10:53:30AM -0400, Sonic wrote:
> When it happened to me back then (2022) I'm pretty sure I did a "down"
> followed by a "delete" and then the "destroy".
root@ot6:.../~# cd /usr/src/regress/sys/net/wg
root@ot6:.../wg# make ifconfig
...
root@ot6:.../wg# ifconfig wg11
wg11: fl
On Wed, Oct 04, 2023 at 10:08:01AM -0400, Sonic wrote:
> See the post:
> "Uninterruptible D State after ifconfig wg0 destroy" Oct. 31, 2022 in the
> Bugs archive.
I have a test regress/sys/net/wg that configures a wg(4), sends
some traffic, and destroys it. I have never seen this bug. There
must
Errata patches for X11 libraries libX11 and libXpm have been released
for OpenBSD 7.2 and 7.3.
Binary updates for the amd64, arm64 and i386 platform are available
via the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.org/errata72.html
h
On Tue, Aug 22, 2023 at 06:30:31AM +0200, Alexandr Nedvedicky wrote:
> Currently we have something like this:
>
> { mbuf, pcb } <-> state key <-> { state, state ... }
>
> with this diff we get to:
>
> { mbuf, pcb } <-> state <-> state key
>
> Basically when we do process
On Fri, Sep 29, 2023 at 01:46:40AM +0200, Tobias Heider wrote:
> Like with route messages we should really only forward pfkey messages
> that made it past the validation step. This fixes a lot of possible
> crashes in ipsecctl -m.
>
> ok?
OK bluhm@
> diff /home/user/got/co/src
> commit - 1ce2bc2
Hi,
There is a lock order problem in ixl(4) show by witness.
Here I replaced a wrong net lock with kernel lock.
revision 1.84
date: 2022/08/05 13:57:16; author: bluhm; state: Exp; lines: +6 -3;
commitid: sAcV6NsO35L03mLS;
The netlock for SIOCSIFMEDIA and SIOCGIFM
On Mon, Sep 18, 2023 at 02:12:21PM +0300, Vitaliy Makkoveev wrote:
> On Mon, Sep 18, 2023 at 02:03:08PM +0300, Vitaliy Makkoveev wrote:
> > Also use this mutex to protect `evqueue_head', `evqueue_tail' and
> > `evqueue_count'.
> >
>
> Sorry, the right diff:
OK bluhm@
> Index: sys/dev/hotplug.c
On Mon, Sep 18, 2023 at 11:40:28AM +0300, Vitaliy Makkoveev wrote:
> Context switch looks fine here.
OK bluhm@
> Index: sys/dev/pv/hypervic.c
> ===
> RCS file: /cvs/src/sys/dev/pv/hypervic.c,v
> retrieving revision 1.19
> diff -u -p
On Thu, Sep 14, 2023 at 01:12:22PM +0300, Vitaliy Makkoveev wrote:
> `log_mtx' mutex(9) already used for message buffer protection, so use it
> to protect `logread_filtops' too.
>
> ok?
OK bluhm@
> Index: sys/kern/subr_log.c
> ===
>
On Fri, Sep 22, 2023 at 12:21:42PM +0900, YASUOKA Masahiko wrote:
> A leak may happens when wgpeer is deleted.
>
> ok?
OK bluhm@
> The state queue should be freeed when wg_peer is destroyed.
> diff from IIJ.
>
> Index: sys/net/if_wg.c
> ==
Errata patches for npppd have been released for OpenBSD 7.2 and 7.3.
Binary updates for the amd64, arm64 and i386 platform are available
via the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.org/errata72.html
https://www.openbsd.org/err
On Fri, Sep 15, 2023 at 04:18:13PM +0200, Martin Pieuchot wrote:
> On 11/09/23(Mon) 21:05, Martin Pieuchot wrote:
> > On 06/09/23(Wed) 23:13, Alexander Bluhm wrote:
> > > On Wed, Sep 06, 2023 at 12:23:33PM -0500, Scott Cheloha wrote:
> > > > On Wed, Sep 06, 20
On Thu, Sep 07, 2023 at 10:32:58PM +0300, Vitaliy Makkoveev wrote:
> Remove the remnants of the leftover selinfo from vnode(9) layer. Just
> mechanical replacement because knote(9) API is already used. I don't
> want make klist MP safe with this diff.
>
> headers added where is was required. Disa
On Sun, Sep 03, 2023 at 02:00:46AM +0200, Alexander Bluhm wrote:
> Hi,
>
> When DNS lookup for remote loghost in @ line in syslog.conf does
> not work at startup, retry in intervals.
>
> Please test if you use the feature.
>
> ok?
anyone?
>
On Wed, Sep 06, 2023 at 12:23:33PM -0500, Scott Cheloha wrote:
> On Wed, Sep 06, 2023 at 01:04:19PM +0100, Martin Pieuchot wrote:
> > Debugging OOM is hard. UVM uses per-CPU counters and sadly
> > counters_read(9) needs to allocate memory. This is not acceptable in
> > ddb(4). As a result I cann
Hi,
ip_output() and ip6_output() should be MP safe when called with
NULL options.
ok?
bluhm
Index: netinet/ip_input.c
===
RCS file: /data/mirror/openbsd/cvs/src/sys/netinet/ip_input.c,v
retrieving revision 1.385
diff -u -p -r1.385
On Mon, Sep 04, 2023 at 07:22:03PM +0300, Vitaliy Makkoveev wrote:
> > On 4 Sep 2023, at 16:19, Alexander Bluhm wrote:
> >
> > Hi,
> >
> > Variable scs_use is basically counting packet insertions to syn
> > cache, so I would prefer type long to exclude over
Hi,
Variable scs_use is basically counting packet insertions to syn
cache, so I would prefer type long to exclude overflow on fast
machines. With the current limits int should be enough, but long
does not hurt.
It can be negative as it starts at a positive limit and counts
backwards. After all
Hi,
Avoid a useless increment and decrement of of the tcp syn cache
refcount by unexpanding the SYN_CACHE_TIMER_ARM() macro in the timer
callback.
ok?
bluhm
Index: netinet/tcp_input.c
===
RCS file: /data/mirror/openbsd/cvs/src/sys/
Hi,
When DNS lookup for remote loghost in @ line in syslog.conf does
not work at startup, retry in intervals.
Please test if you use the feature.
ok?
bluhm
Index: syslogd.c
===
RCS file: /data/mirror/openbsd/cvs/src/usr.sbin/syslo
On Thu, Aug 17, 2023 at 12:02:35PM +1000, David Gwynne wrote:
> there are links between the pcb/socket layer and pf as an optimisation,
> and links on mbufs between both sides of a forwarded connection.
> these links let pf skip an rb tree lookup for outgoing packets.
>
> right now these links are
On Wed, Aug 09, 2023 at 12:41:18AM +0300, Vitaliy Makkoveev wrote:
> I think it's better to merge SO_BINDANY cases from both switch blocks.
> This time SO_LINGER case is separated, so there is no reason for
> dedicated switch block.
OK bluhm@
> Index: sys/kern/uipc_socket.c
>
On Fri, Aug 04, 2023 at 12:38:23AM +0300, Vitaliy Makkoveev wrote:
> @@ -1856,6 +1856,9 @@ sosetopt(struct socket *so, int level, i
> case SO_SNDLOWAT:
> case SO_RCVLOWAT:
> {
> + struct sockbuf *sb = (optname == SO_SNDBUF ||
> +
On Tue, Aug 01, 2023 at 01:50:52AM +0200, Alexandr Nedvedicky wrote:
> OK to commit?
OK bluhm@
> 8<---8<---8<--8<
> diff --git a/sys/net/pf_table.c b/sys/net/pf_table.c
> index 6f23a6f795d..c862c804f84 100644
> --- a/sys/net/pf_table.c
> +++
On Sat, Jul 22, 2023 at 05:42:49PM +0300, Vitaliy Makkoveev wrote:
> Thanks for testing. There is updated diff below.
OK bluhm@
> Index: sys/kern/uipc_socket.c
> ===
> RCS file: /cvs/src/sys/kern/uipc_socket.c,v
> retrieving revision
On Thu, Jul 27, 2023 at 09:15:48AM -0700, Chris Cappuccio wrote:
> Jan Klemkow [j.klem...@wemelug.de] wrote:
> > +#if NVLAN > 0
> > + if (ext.evh)
> > + hdrlen += ETHER_VLAN_ENCAP_LEN;
> > +#endif
> > if (ext.ip
On Wed, Jul 26, 2023 at 11:30:45AM +0200, Jan Klemkow wrote:
> Hi,
>
> I missed the vlan-tag size in the mss calculation of lro packets in
> ix(4). This diff add vlan-header detection in ether_extract_headers()
> and uses this information to calculate the right mss.
>
> This fixes forwarding of
Errata patches for guests running on hypervisors on AMD cpus have
been released for OpenBSD 7.2 and 7.3.
Binary updates for the amd64 and i386 platform are available via
the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.org/errata72.html
Errata patches for AMD cpus with their firmware updates and for
wscons have been released for OpenBSD 7.2 and 7.3.
Binary updates for the amd64, i386 and arm64 platform are available
via the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.o
On Thu, Jul 13, 2023 at 02:22:17AM +0300, Vitaliy Makkoveev wrote:
> This is a part of my standalone sblock() work. I need this movement
> because buffers related SO_SND* and SO_RCV* socket options modification
> should be protected with sblock(). However, standalone sblock() has
> different lock o
On Fri, Jul 21, 2023 at 03:05:41PM +0200, Claudio Jeker wrote:
> On Fri, Jul 21, 2023 at 03:17:35PM +0300, Vitaliy Makkoveev wrote:
> > On Thu, Jul 20, 2023 at 09:57:00PM +0200, Alexander Bluhm wrote:
> > > Hi,
> > >
> > > I wonder why UDP echo d
On Fri, Jul 21, 2023 at 11:35:21AM +0300, Vitaliy Makkoveev wrote:
> It contains SS_CANTSENDMORE, SS_ISSENDING, SS_CANTRCVMORE and
> SS_RCVATMARK bits. Also do `sb_flags' output as hex, it contains flags
> too.
OK bluhm@
> Index: sys/kern/uipc_socket.c
> ==
Hi,
I wonder why UDP echo does not work with inetd on 127.0.0.1.
Note that it is default off. One of my regress machines has it
enabled for other tests. There perl dist/Net-Ping/t/510_ping_udp.t
expects that UDP echo works on 127.0.0.1.
It was disabled with this commit:
---
Hi,
When running tcpdump on loopback, I see a currupted packet for each
valid packet.
root@v74:.../~# tcpdump -ni lo0 -X -vvv
tcpdump: listening on lo0, link-type LOOP
21:01:07.645880
: e161 ff01 7f00 0001 7f00 0001 .a..
0010: 0800 00df 3a4f d257 38fa 0236 0129
Errata patches for ssh-agent have been released for
OpenBSD 7.2 and 7.3.
Binary updates for the amd64, i386 and arm64 platform are available
via the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.org/errata72.html
https://www.openbsd.org
On Mon, Jul 10, 2023 at 02:07:01PM +0200, Jan Klemkow wrote:
> On Sat, Jul 08, 2023 at 05:15:26PM +0300, Alexander Bluhm wrote:
> > I am not aware of any more limitations when enabling LRO for TCP
> > in the network drivers. The feature allows to receive agregated
> > packet
Hi,
While testing my ART reference couting fix, I discovered a rtentry
leak that is triggered by regress/sbin/route and detected with
btrace(8) refcnt.
The reference returned by rtalloc() must be freed with rtfree() in
all cases.
ok?
bluhm
Index: netinet6/in6_ifattach.c
===
Hi,
I am not aware of any more limitations when enabling LRO for TCP
in the network drivers. The feature allows to receive agregated
packets larger than the MTU. Receiving TCP streams becomes much
faster.
As the network hardware is not aware whether a packet is received
locally or forwarded, ev
On Fri, Jul 07, 2023 at 12:31:02PM +0300, YASUOKA Masahiko wrote:
> Hi,
>
> I'd like to expand the counters in struct mbstat from u_short to u_long.
>
> When I was debugging a mbuf leak, I saw the result of "netstat -m"
> ---
> 28647 mbufs in use:
> 28551 mbufs allocated to data
>
On Fri, Jul 07, 2023 at 12:07:36PM +0300, YASUOKA Masahiko wrote:
> Hi,
>
> This netstat diff is already needed to avoid compiler warnings since
> it uses struct tcpcb directly.
>
> ok?
OK bluhm@
> Index: usr.bin/netstat/inet.c
> =
On Fri, Jul 07, 2023 at 10:43:21AM +0200, Claudio Jeker wrote:
> On Fri, Jul 07, 2023 at 11:38:58AM +0300, YASUOKA Masahiko wrote:
> > Hi,
> >
> > Does using 64 bit for timer in tcpcb require this?
>
> Not sure about this but one comment below.
When we introduced struct tcp_info, we discussed if
Hi,
When we preserve M_TCP_TSO we also must keep ph_mss. In lo(4)
LRO/TSO this logic was missing. As this may be relevant only for
weird pf configs that forward from loopback and ifconfig tcplro is
disabled by default, it is not relevant yet.
ok?
bluhm
Index: net/if.c
On Thu, Jul 06, 2023 at 08:49:03PM +0200, Jan Klemkow wrote:
> > @@ -109,6 +109,9 @@
> > #include
> > #include
> > #include
>
> I think is a merge bug, isn't it?
>
> > +#include
> > +#include
> > +#include
Right.
> > + error = tcp_if_output_tso(ifp, mp, dst, rt, ifcap, mtu);
> > +
On Tue, Jul 04, 2023 at 06:38:46PM +0300, Alexander Bluhm wrote:
> On Sat, May 20, 2023 at 06:59:14PM -0700, Andrew Hewus Fresh wrote:
> > Here's a patch to bring perl in base from 5.36.0 to 5.36.1, there are
> > only a few fixes in it and at least one of them I had previ
Hi,
When forwaring in packets parallel while removing the corresponding
ARP entry in a loop like this
while :; do arp -nd 10.10.21.1; arp -nd 10.10.22.4; done
it was quite easy to crash the kernel.
The race is exposed by putting a delay() here:
Index: net/rtable.c
===
On Wed, Jul 05, 2023 at 03:41:31AM +, Visa Hankala wrote:
> Initialize stack-based mutexed using mtx_init(). This removes the need
> of the NOWITNESS kludge and lets the lock checker do its job with these
> mutexes.
>
> At the moment, static initialization of locks inside functions does not
>
anyone?
On Fri, May 26, 2023 at 06:44:25PM +0200, Alexander Bluhm wrote:
> Hi,
>
> in_ifcap_cksum() checks ifp == NULL
> in_hdr_cksum_out() sets ip_sum = 0
> in_proto_cksum_out() and in6_proto_cksum_out() always write
> th_sum if M_TCP_CSUM_OUT is set and proto is IP
Hi,
After changing tcp now tick to milliseconds, it will wrap around
after 49 days of uptime. That may be a problem in some places of
our stack. Better use a 64 bit counter.
As timestamp option is 32 bit in TCP protocol, we have to use the
lower 32 bit there. There are casts to 32 bits that sh
On Mon, Jul 03, 2023 at 11:12:17PM +0200, Jan Klemkow wrote:
> bluhm pointed out that the ether_brport_isset() check it just allowed on
> ethernet devices. Thus, I put an additional ethernet check in the
> condition. This also fixes EBUSY errors of "ifconfig lo0 tcplro" calls
> in my setup.
>
>
Hi,
As final step before making LRO (Large Receive Offload) the default,
we have to fix path MTU discovery when forwarding.
The drivers, currently ix(4) and lo(4) only, record an upper bound
of the size of the original packets in ph_mss. When sending we
must chop the packets with TSO (TCP Segmen
On Mon, Jul 03, 2023 at 04:48:36PM +1000, Jonathan Matthew wrote:
> This adds kstats for the hardware counters available in bge(4) devices,
> BCM5705
> and newer. The main complication is that some of the counters are already
> used
> in bge_stats_update_regs() as part of a hardware bug workarou
anyone?
On Fri, Jun 23, 2023 at 06:06:16PM +0200, Alexander Bluhm wrote:
> Hi,
>
> Claudio@ mentioned the idea to use TSO and LRO on the loopback
> interface to transfer TCP faster.
>
> I see a performance effect with this diff, but more importantly it
> gives us more test
On Tue, Jun 27, 2023 at 01:55:23PM +0200, Mark Kettenis wrote:
> > Date: Tue, 27 Jun 2023 11:09:32 +
> > From: Klemens Nanni
> >
> > On Tue, Jun 27, 2023 at 01:32:37PM +0300, Vitaliy Makkoveev wrote:
> > > M_TEMP seems unreasonable for interface groups data allocations.
> >
> > After claudio
On Sat, Jun 24, 2023 at 11:20:50AM +1000, David Gwynne wrote:
> maybe it's time to re-evaluate siphash?
In one of our products we have replaced SipHash with xxHash.
https://xxhash.com/
https://github.com/Cyan4973/xxHash
bluhm
Hi,
Claudio@ mentioned the idea to use TSO and LRO on the loopback
interface to transfer TCP faster.
I see a performance effect with this diff, but more importantly it
gives us more test coverage. Currently LRO on lo(4) is default
off.
Future plan is:
- Fix some corner cases for LRO/TSO with TC
Hi,
I am working on a diff to run UDP input in parallel. Btrace kstack
analysis shows that SIP hash for PCB lookup is quite expensive.
When running in parallel we get lock contention on the PCB table
mutex.
So it results in better performance to calculate the hash value
before taking the mutex.
On Mon, Jun 12, 2023 at 03:46:28PM +0200, Alexander Bluhm wrote:
> I found a little inconsistency in IPv6 forwarding with TSO.
>
> Sending with TSO should only done if the large packet does not fit
> in the interface MTU. In case tcp_if_output_tso() does not process
> the packet,
Errata patches for libX11 CVE-2023-3138 have been released for
OpenBSD 7.2 and 7.3.
Binary updates for the amd64, i386 and arm64 platform are available
via the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.org/errata72.html
https://www.
On Thu, Jun 15, 2023 at 11:36:21AM +0200, Stefan Fritsch wrote:
> >From FreeBSD commits
>
> commit c0db7289c3de290d821311942d5533f2284af77f
> Author: pfg
> Date: Wed Aug 8 15:08:22 2018 +
>
> msdosfs: fixes for Undefined Behavior.
>
> and
>
>
On Fri, Jun 09, 2023 at 08:10:22PM +0200, Jan Klemkow wrote:
> On Fri, Jun 09, 2023 at 06:59:57PM +0200, Jan Klemkow wrote:
> > On Fri, Jun 09, 2023 at 06:11:38PM +0200, Jan Klemkow wrote:
> > > TSO packets are limited to MAXMCLBYTES (64k). Thus, we don't need to
> > > allocate IXGBE_TSO_SIZE (256
Hi,
I found a little inconsistency in IPv6 forwarding with TSO.
Sending with TSO should only done if the large packet does not fit
in the interface MTU. In case tcp_if_output_tso() does not process
the packet, we should send an ICMP6 error. Rearrange the code that
it looks more like other calls
On Wed, Jun 07, 2023 at 01:36:23PM +0300, Vitaliy Makkoveev wrote:
> In this point the interface is already removed from the list of all
> interfaces and from the interface index map and all possible
> concurrent ioctl() threads finished. Remove this dead code.
Should we set ifp->if_ioctl to NULL?
On Wed, Jun 07, 2023 at 01:15:46PM +0300, Vitaliy Makkoveev wrote:
> In this point, the interface is disconnected from everywhere. No need to
> hold netlock for dummy 'nd_ifinfo' release. Netlock is also not needed
> for TAILQ_EMPTY(&ifp->if_*hooks) assertions.
OK bluhm@
> Index: sys/net/if.c
> =
On Wed, Jun 07, 2023 at 12:59:11PM +0300, Vitaliy Makkoveev wrote:
> On Wed, Jun 07, 2023 at 10:19:32AM +1000, David Gwynne wrote:
> >
> >
> > > On 7 Jun 2023, at 06:33, Vitaliy Makkoveev wrote:
> > >
> > >> On 6 Jun 2023, at 20:29, Alexander Bluh
On Tue, Jun 06, 2023 at 11:33:36PM +0300, Vitaliy Makkoveev wrote:
> > On 6 Jun 2023, at 20:29, Alexander Bluhm wrote:
> >
> > On Tue, Jun 06, 2023 at 05:54:31PM +0300, Vitaliy Makkoveev wrote:
> >> On Tue, Jun 06, 2023 at 02:31:52PM +0200, Alexander Bluhm wrote:
>
On Tue, Jun 06, 2023 at 05:54:31PM +0300, Vitaliy Makkoveev wrote:
> On Tue, Jun 06, 2023 at 02:31:52PM +0200, Alexander Bluhm wrote:
> > Hi,
> >
> > I would suggest to rename ifconfig tcprecvoffload to tcplro. Maybe
> > it's just because I had to type that lo
Hi,
I would suggest to rename ifconfig tcprecvoffload to tcplro. Maybe
it's just because I had to type that long name too often.
With that we have consistent naming:
# ifconfig ix0 tcplro
# sysctl net.inet.tcp.tso=1
Also the coresponding flag are named LRO.
# ifconfig ix1 hwfeatures
ix1: flags=
On Wed, May 31, 2023 at 09:36:22AM +1000, David Gwynne wrote:
> we could export these csum flags as part of the bpf header so we can
> teach tcpdump to shut up in this situation.
Linux does not do that and I want to keep pcap file format compatible.
Also smart tools make debugging real checksum pr
On Thu, Jun 01, 2023 at 04:38:53PM +, Peter Stuge wrote:
> David Gwynne wrote:
> > > Currently packets sent over loopback interface get their checksum
> > > calculated twice. In the output path it is set and during TCP/IP
> > > input it is calculated again to be compared with the previous valu
On Thu, May 25, 2023 at 10:40:51PM +0200, Jan Klemkow wrote:
> On Wed, May 24, 2023 at 05:28:58PM +0200, Alexander Bluhm wrote:
> > On Tue, May 23, 2023 at 02:14:57PM +0200, Jan Klemkow wrote:
> > > This diff sets needed offloading flags and the calculated mss to LRO
> >
Hi,
Currently packets sent over loopback interface get their checksum
calculated twice. In the output path it is set and during TCP/IP
input it is calculated again to be compared with the previous value.
This can be avoided by claiming that lo(4) supports hardware checksum
offloading. For each
anyone?
On Mon, May 22, 2023 at 10:17:31PM +0200, Alexander Bluhm wrote:
> Hi,
>
> I noticed a missing checksum count in netstat tcp packets sent
> software-checksummed. Turns out that our syn cache does the checksum
> calculation by hand, instead of the established mechanism
Hi,
in_ifcap_cksum() checks ifp == NULL
in_hdr_cksum_out() sets ip_sum = 0
in_proto_cksum_out() and in6_proto_cksum_out() always write
th_sum if M_TCP_CSUM_OUT is set and proto is IPPROTO_TCP.
ok?
bluhm
Index: netinet/ip_output.c
=
On Wed, May 17, 2023 at 01:02:58PM +0300, Vitaliy Makkoveev wrote:
> mrt_rtwalk_mfcsysctl() performs read-only access to protected data, so
> rtable_walk() could be called with shared netlock.
While I think the NET_LOCK_SHARED() is not sufficent, you can move
the NET_LOCK() into mrt_sysctl_mfc().
On Fri, May 26, 2023 at 06:25:46PM +0300, Vitaliy Makkoveev wrote:
> On Fri, May 26, 2023 at 05:08:06PM +0200, Alexander Bluhm wrote:
> > On Fri, May 26, 2023 at 05:29:58PM +0300, Vitaliy Makkoveev wrote:
> > > On Wed, May 17, 2023 at 01:02:58PM +0300, Vital
On Fri, May 26, 2023 at 05:29:58PM +0300, Vitaliy Makkoveev wrote:
> On Wed, May 17, 2023 at 01:02:58PM +0300, Vitaliy Makkoveev wrote:
> > mrt_rtwalk_mfcsysctl() performs read-only access to protected data, so
> > rtable_walk() could be called with shared netlock.
> >
>
> Regardless on sysctl(2)
Errata patches for rpki-client and LibreSSL libssl have been released
for OpenBSD 7.2 and 7.3.
Binary updates for the amd64, i386 and arm64 platform are available
via the syspatch utility. Source code patches can be found on the
respective errata page:
https://www.openbsd.org/errata72.html
h
On Tue, May 23, 2023 at 02:14:57PM +0200, Jan Klemkow wrote:
> Hi,
>
> This diff sets needed offloading flags and the calculated mss to LRO
> mbufs in ix(4). Thus, we can forward this packets and process them via
> tcp_if_output_tso(). This diff also uses tcp_if_output_tso() in
> ip6_forward().
Hi,
I noticed a missing checksum count in netstat tcp packets sent
software-checksummed. Turns out that our syn cache does the checksum
calculation by hand, instead of the established mechanism in ip
output.
Just set the flag M_TCP_CSUM_OUT and let in_proto_cksum_out() do
the work later.
While
On Sat, May 20, 2023 at 02:46:27PM +0200, Claudio Jeker wrote:
> On Fri, May 19, 2023 at 07:58:47PM +0200, Jan Klemkow wrote:
> > Hi,
> >
> > We use the wrong interface and mtu in tcp_mss() to calculate the mss if
> > the destination address points is a local address. In ip_output() we
> > use th
Hi,
When sending TCP packets with software TSO to the local address of
a physical interface, the TCP checksum was miscalculated.
This bug was triggered on loopback interface, when sending to the
local interface address of a physical interface. Due to another
bug, the smaller MTU of the physical
1 - 100 of 1116 matches
Mail list logo
