[Differential] D12457: Complete INADDR_HASH lock protection

2017-09-22 Thread rwatson (Robert Watson)
rwatson added a subscriber: bz. rwatson added a comment. Many of these changes are not sufficient to address the underlying problems in this code, as the lock covers loop iteration but fails to protect the stability of the 'ifp' or 'ia' pointer outside of the loop. It could be that this prev

[Differential] D12453: mbuf: Remove UDP_IPV4_EX, which is never defined.

2017-09-22 Thread rwatson (Robert Watson)
rwatson accepted this revision. rwatson added a comment. This revision is now accepted and ready to land. This seems generally sensible. I'm not quite sure about keeping the "Non-standard RSS hash types" comment, as M_HASHTYPE_OPAQUE is basically intended to be a catch-all for any other

Re: svn commit: r304313 - head/sys/net

2016-08-21 Thread Robert Watson
On 21 Aug 2016, at 19:18, Andrey V. Elsukov wrote: > On 21.08.16 20:33, Adrian Chadd wrote: >>> Does this make sense? >> >> Right. Let me go and look into it a little more. I think we may want >> to revert the change (which just landed to -11, so maybe revert that >> too) so I can test both of

Re: Does FreeBSD have sendmmsg or recvmmsg system calls?

2016-01-03 Thread Robert Watson
On Sun, 3 Jan 2016, Rui Paulo wrote: NetBSD 7.0 has just introduced these two syscalls. And Linux also has them. Does FreeBSD have them? Or plan to support them in the future? FreeBSD does not have them.  It doesn't seem especially hard to implement, though.  Do you know any major applicatio

[Differential] [Accepted] D1805: [sockbuf] Don't access fields directly, use accessor functions

2015-02-12 Thread rwatson (Robert Watson)
rwatson accepted this revision. rwatson added a comment. Seems sensible. REVISION DETAIL https://reviews.freebsd.org/D1805 To: davide, kmacy, np, lstewart, rrs, julian, adrian, rwatson Cc: emaste, freebsd-net ___ freebsd-net@freebsd.org mailing list

[Differential] [Requested Changes To] D1809: [sockbuf] Don't expose lock details when isn't needed

2015-02-11 Thread rwatson (Robert Watson)
rwatson requested changes to this revision. rwatson added a comment. This revision now requires changes to proceed. Looks like some serious transcription errors are in this patch. INLINE COMMENTS sys/kern/uipc_sockbuf.c:225 Should this be so->so_rcv? sys/kern/uipc_sockbuf.c:234 Should this be

[Differential] [Updated] D1809: [sockbuf] Don't expose lock details when isn't needed

2015-02-11 Thread rwatson (Robert Watson)
rwatson added a comment. Seems reasonable. I'm slightly surprised that any code other than sbwait() actually implements the sleep on sb_acc, and wonder if either a use of abstraction improvement, or an actual abstraction improvement, is needed to fix t4_ddp.c. REVISION DETAIL https://reviews

[Differential] [Commented On] D1761: Extend LRO support to accumulate more than 65535 bytes

2015-02-03 Thread rwatson (Robert Watson)
rwatson added a comment. np@'s comments are good ones, both with respect to ACK clocking and BPF -- but this will also affect local firewalls that do state tracking and/or transformation, and likewise will run into problems. Given this discussion, I'm increasingly convinced that this is Not A G

[Differential] [Updated] D1761: Extend LRO support to accumulate more than 65535 bytes

2015-02-03 Thread rwatson (Robert Watson)
rwatson added a comment. Historically, I believe that we've allowed additional trailer content/padding to be included at the end of mbuf chains that isn't in the IP datagram, and is silently ignored, as it isn't included in the IP packet length. I may misunderstand what was going on there, but

Inlining struct m_hdr with struct mbuf; assertion changes

2015-01-11 Thread Robert Watson
Dear all: As you may have spotted from recent commits to 11-CURRENT, I've been working through the mbuf allocator and mbuf consumers (protocols, device drivers) trying to reduce consumer dependence on the specifics of the mbuf layout -- i.e., using M_SIZE() instead of caller assumptions about MLE

Re: Enabling VIMAGE in GENERIC

2014-12-04 Thread Robert Watson
On Mon, 1 Dec 2014, George Neville-Neil wrote: On a slight tangent. I ran VIMAGE kernels vs. non VIMAGE kernels for both a VANILLA kernel and a PF kernel (PF on but no rules) as a quick smoke test today. The raw forwarding performance was unchanged between kernels with and without VIMAGE on

[Differential] [Commented On] D1201: Allow UMA allocated memory to be freed when VNET jails are torn down.

2014-12-02 Thread rwatson (Robert Watson)
rwatson added a comment. This patch is probably safe if limited to UDP. REVISION DETAIL https://reviews.freebsd.org/D1201 To: rodrigc, alfredperlstein, melifaro, glebius, hrs, wollman, bryanv, rpaulo, adrian, bz, gnn, hiren, rwatson Cc: freebsd-net, emaste, gnn, rwatson __

RSS patch merged (was: svn commit: r263198 - in head/sys: amd64/conf conf net netinet netinet6 sys (fwd))

2014-03-15 Thread Robert Watson
rt remain disabled by default (something we might think about revisting prior to 11.0). Robert -- Forwarded message -- Date: Sat, 15 Mar 2014 00:57:50 + (UTC) From: Robert Watson To: src-committ...@freebsd.org, svn-src-...@freebsd.org, svn-src-h...@freebsd.org Su

Re: [rfc] migrate lagg to an rmlock

2013-08-24 Thread Robert Watson
On Sat, 24 Aug 2013, Alexander V. Chernikov wrote: On 24.08.2013 00:54, Adrian Chadd wrote: I'd like to commit this to -10. It migrates the if_lagg locking from a rw lock to a rm lock. We see a bit of contention between the transmit and We're running lagg with rmlock on several hundred heavi

Re: Latency issues with buf_ring

2012-12-06 Thread Robert Watson
On Tue, 4 Dec 2012, Andre Oppermann wrote: For most if not all ethernet drivers from 100Mbit/s the TX DMA rings are so large that buffering at the IFQ level doesn't make sense anymore and only adds latency. So it could simply directly put everything into the TX DMA and not even try to soft-qu

Re: Latency issues with buf_ring

2012-12-06 Thread Robert Watson
On Tue, 4 Dec 2012, John Baldwin wrote: Q2: Are there any case studies or benchmarks for buf_ring, or it is just blindly being used because someone claimed it was better and offered it for free? One of the points of locking is to avoid race conditions, so the fact that you have races in a su

Re: protect common resources in kernel

2012-12-02 Thread Robert Watson
On Sun, 2 Dec 2012, Choupani wrote: I'm working on kernel in FreeBSD-9. I need to protect a common resource (for example a queue). There are 4 points for access (read/write) this common resource as bellows: 1. ether_input() – hardware interrupt 2. ip_input() & ip_output() – software interrupt

Re: netisr panic?

2012-11-17 Thread Robert Watson
Panics along these lines often occur if there is a concurrency bug in a device driver such that it modifies an mbuf after dispatching to the network stack. E.g., by freeing it, reusing it, an errant dereference, etc. Not guaranteed, but that is where I'd start. Robert On Sat, 17 Nov 2012, Ad

Re: Use of V_tcbinfo lock in tcp_do_segment()

2012-07-22 Thread Robert Watson
On Thu, 24 May 2012, Vijay Singh wrote: Folks, I am trying to understand the need to hold the V_tcbinfo lock in part of this function [code included below for reference]. At present this causes the socket upcall to be called with the V_tcbinfo lock held, which I'd like to avoid. We release t

Re: tcp_detach can return with inpcb lock held

2012-01-06 Thread Robert Watson
On Thu, 5 Jan 2012, Navdeep Parhar wrote: Looks like there's a case where tcp_detach could return with the inp lock held. I see an XXXRW comment questioning this possibility, but we should either add an assertion to verify that the case does not occur, or unlock the inpcb before returning.

Re: Panic in the udp_input() under heavy load

2012-01-03 Thread Robert Watson
On Fri, 30 Dec 2011, Maxim Sobolev wrote: On 12/30/2011 4:46 PM, Maxim Sobolev wrote: I see. Would you guys mind if I put that NULL pointer check into the code for the time being and turn it into some kind of big nasty warning in 8-stable branch only? I could also open a ticket, put all deb

Re: Panic in the udp_input() under heavy load

2011-11-07 Thread Robert Watson
On Mon, 7 Nov 2011, Maxim Sobolev wrote: On 11/7/2011 3:25 PM, Bjoern A. Zeeb wrote: Now if you are clever you'd also log the inp there as the above will only prove the case that something is wrong but still not help us in anything to figure out what. Good point, thank you Sir. Would that

Re: Kernel panic on FreeBSD 9.0-beta2

2011-09-30 Thread Robert Watson
On Wed, 28 Sep 2011, Mikolaj Golub wrote: On Mon, 26 Sep 2011 16:12:55 +0200 K. Macy wrote: KM> Sorry, didn't look at the images (limited bw), I've seen something KM> like this before in timewait. This "can't happen" with UDP so will be KM> interested in learning more about the bug. The pa

Re: FreeBSD I/OAT (QuickData now?) driver

2011-06-11 Thread Robert Watson
On Mon, 6 Jun 2011, grarpamp wrote: I know we've got polling. And probably MSI-X in a couple drivers. Pretty sure there is still one CPU doing the interrupt work? And none of the multiple queue thread spreading tech exists? Actually, with most recent 10gbps cards, and even 1gbps cards, we pr

Re: Connections not purged on address deletion

2011-02-05 Thread Robert Watson
On Fri, 4 Feb 2011, Prabhu Hariharan wrote: When I delete an IP-address from an interface, the TCP (and other) connections using that local IP-address are not getting purged. The telnet or ssh sessions on the other end just get hung, as FreeBSD address-deletion doesn't handle this situation

Re: TCP resident expert?

2011-01-16 Thread Robert Watson
On Sat, 15 Jan 2011, William Allen Simpson wrote: Who's the kernel expert on TCP around here? ISC wants me to port TCPCT to FreeBSD. Although I've joined this list (some time ago), I've not seen any traffic discussing TCP'ish things. Need somebody willing to walk me through the processes a

Re: Future of netnatm: looking for testers

2010-12-17 Thread Robert Watson
On Wed, 15 Dec 2010, Bjoern A. Zeeb wrote: I would request two things: 1) the extra couple of months; this will not prevent the evitable removal yet only defer it. Sounds good to me -- my goal is not to remove NETNATM, rather, the remove code that doesn't compile or work. I'm happy to si

Re: bjoern just received the itojun award at the ietf

2010-11-10 Thread Robert Watson
On Wed, 10 Nov 2010, Randy Bush wrote: bjoern zeeb just received the itojun award. congratulations, bjoern. and than you for all the hard work on the ipv6 stack. Indeed -- many congratulations, Bjoern! The slow road from an experimental protocol in an experimental network stack to one widel

Re: Question on TCP reassembly counter

2010-10-24 Thread Robert Watson
On Sat, 23 Oct 2010, Lawrence Stewart wrote: One observation though: net.inet.tcp.reass.cursegments was non-zero (it was just 1) after 30 rounds, where each round is (as earlier) 15-concurrent instances of netperf for 20s. This was on the netserver side. And, it was zero before the netperf ru

Re: [PATCH] Netdump for review and testing -- preliminary version

2010-10-15 Thread Robert Watson
On Thu, 14 Oct 2010, Attilio Rao wrote: No, what I'm saying is: UMA needs to not call its drain handlers, and ideally not call into VM to fill slabs, from the dumping context. That's easy to implement and will cause the dump to fail rather than causing the system to hang. My point is, howeve

Re: [PATCH] Netdump for review and testing -- preliminary version

2010-10-08 Thread Robert Watson
On Fri, 8 Oct 2010, Attilio Rao wrote: GENERAL FRAMEWORK ARCHITECTURE Netdump is composed, right now, by an userland "server" and a kernel "client". The former is run on the target machine (where the dump will phisically happen) and it is responsible for receiving  the packets containing cor

Re: mbuf changes

2010-10-02 Thread Robert Watson
On Thu, 30 Sep 2010, Julian Elischer wrote: On 9/30/10 10:49 AM, Ryan Stone wrote: It's not a big thing but it would be nice to replace the m_next and m_nextpkt fields with queue.h macros. funny, I've never even thought of that.. I have, and it's a massive change touching code all over the

Re: Removal of deprecated implied connect for TCP

2010-09-09 Thread Robert Watson
On Tue, 31 Aug 2010, Garrett Wollman wrote: In article <4c7d02bb.40...@freebsd.org> an...@freebsd.org writes: sendto() will not be touched or modified. It's just that on a TCP socket the tcp protocol will not perform an implied connect anymore. The only thing that changes is TCP dropping a

Re: Removal of deprecated implied connect for TCP

2010-08-31 Thread Robert Watson
On Tue, 31 Aug 2010, Andre Oppermann wrote: I'm not entirely comfortable with this change, and would like a chance to cogitate on it a bit more. While I'm not aware of any applications depending on the semantic for TCP, I know that we do use it for UNIX domain sockets. I don't have any plan

Re: Removal of deprecated implied connect for TCP

2010-08-31 Thread Robert Watson
On Sun, 29 Aug 2010, Andre Oppermann wrote: When T/TCP RFC1644 support was introduced in r6283 by wollman 15 years ago the semantics of sendto(2) with regard to TCP sockets were changed. It became possible directly do a sendto(2) call with the target address in the *to argument instead of do

Re: Sending packets at Kernel level

2010-08-30 Thread Robert Watson
On Sun, 29 Aug 2010, Sadish Kulasekere wrote: I need to know how to send packet from the kernel level. Can someone please point me to any documentation? Hi Sadish: Take a look at the socket(9) man page, which is the KPI used by in-kernel consumers of sockets such as the NFS client and serve

Future of netnatm: volunteer wanted -- and/or -- removal notice

2010-07-28 Thread Robert Watson
Dear all: When the new link layer framework was introduced in 8.0, one of our ATM stacks, netnatm, was left behind. As a result, it neither compiles nor runs in 8.x and 9.x. This e-mail serves two purposes: (1) To solicit a volunteer who can work on the netnatm stack in 9.x, with potentia

Re: [PATCH] Remove if_watchdog and if_timer

2009-11-24 Thread Robert Watson
On Tue, 24 Nov 2009, John Baldwin wrote: Now that no drivers in the tree use if_watchdog and if_timer, this patch just removes them completely. Since if_timer was a short that was adjacent to if_index, removing if_timer would have still left a padding hole in the form of a short on all of ou

Re: dummynet dropping too many packets

2009-10-17 Thread Robert Watson
On Sat, 17 Oct 2009, rihad wrote: Just rebooted with the "ifp->if_snd.ifq_drv_maxlen = 1024;" kernel, all ok so far. There's currenlty only 1000 or so entries in the ipfw table and around 350-400 net mbps load, so I'll wait a few hours for the numbers to grow to >2000 and 460-480 respectively

Re: dummynet dropping too many packets

2009-10-17 Thread Robert Watson
On Sat, 17 Oct 2009, rihad wrote: P.S.: BTW, there's a small admin-type inconsistency in FreeBSD 7.1: /etc/rc.firewall gets executed before values set by /etc/sysctl.conf are in effect, so "queue 2000" isn't allowed in ipfw pipe rules (as net.inet.ip.dummynet.pipe_slot_limit is only 100 by de

Re: dummynet dropping too many packets

2009-10-15 Thread Robert Watson
On Thu, 15 Oct 2009, rihad wrote: meaning that USABLE_TX_BD is expected to be smaller than MAX_TX_BD. What if MAX_TX_BD is itself way smaller than 1024, which I'll eventually set ifq_drv_maxlen to? Can a driver guru please comment on this? In a few days I'm going to try it anyway, and if the

Re: Page fault in IFNET_WLOCK_ASSERT [if.c and pccbb.c]

2009-10-11 Thread Robert Watson
On Sun, 11 Oct 2009, Harsha Srinath wrote: I'm running an updated HEAD kernel and got a page fault in ifindex_alloc_locked() in if.c. I figured that the problem was caused by the (pluggable) network card of my laptop and found that during the initialization of the interface, cb_event_thread()

Re: dummynet dropping too many packets

2009-10-07 Thread Robert Watson
On Wed, 7 Oct 2009, rihad wrote: Suggestions like increasing timer resolution are intended to spread out the injection of packets by dummynet to attempt to reduce the peaks of burstiness that occur when multiple queues inject packets in a burst that exceeds the queue depth supported by combin

Re: dummynet dropping too many packets

2009-10-07 Thread Robert Watson
On Wed, 7 Oct 2009, rihad wrote: Robert Watson wrote: On Wed, 7 Oct 2009, rihad wrote: snapshot of the top -SH output in the steady state? Let top run for a few minutes and then copy/paste the first 10-20 lines into an e-mail. Sure. Mind you: now there's only 1800 entries in each o

Re: dummynet dropping too many packets

2009-10-07 Thread Robert Watson
On Wed, 7 Oct 2009, rihad wrote: snapshot of the top -SH output in the steady state? Let top run for a few minutes and then copy/paste the first 10-20 lines into an e-mail. Sure. Mind you: now there's only 1800 entries in each of the two ipfw tables, so any drops have stopped. But it only t

Re: dummynet dropping too many packets

2009-10-07 Thread Robert Watson
On Wed, 7 Oct 2009, rihad wrote: rihad wrote: I've yet to test how this direct=0 improves extensive dummynet drops. Ooops... After a couple of minutes, suddenly: net.inet.ip.intr_queue_drops: 1284 Bumped it up a bit. Yes, I was going to suggest that moving to deferred dispatch has probabl

Re: dummynet dropping too many packets

2009-10-06 Thread Robert Watson
On Wed, 7 Oct 2009, Eugene Grosbein wrote: On Tue, Oct 06, 2009 at 08:28:35PM +0500, rihad wrote: I don't think net.inet.ip.intr_queue_maxlen is relevant to this problem, as net.inet.ip.intr_queue_drops is normally zero or very close to it at all times. When net.isr.direct is 1, this queue

Re: in_pcb crash on 7.2 RELEASE 64bits

2009-06-29 Thread Robert Watson
On Mon, 22 Jun 2009, Jerry Toung wrote: may be someone has seen this  already before I dig into the issue myself. A consistent crash with the following trace.       Unread portion of the kernel message buffer: This is a NULL pointer dereference -- could you tell me what specific version of

Re: panic in sbflush

2009-06-06 Thread Robert Watson
On Fri, 5 Jun 2009, Barney Cordoba wrote: I'm getting a panic in sbflush where mbcnt is 0 and sb_mb is not empty. Any clues as to what might cause this? It happening during a load test. sbflush() panics are typically symptoms of bugs elsewhere in the network stack or kernel, often race condit

Re: panic in soabort

2009-04-25 Thread Robert Watson
On Fri, 24 Apr 2009, pluknet wrote: 2009/4/23 Robert Watson : On Thu, 23 Apr 2009, pluknet wrote: Please, give me comment on this. The panic is on 6.2-REL. Is it known to be fixed in the latter releases? It may well be -- there have been quite significant architectural improvements to

Re: kern/132734: panic in net/if_mib.c

2009-04-23 Thread Robert Watson
The following reply was made to PR kern/132734; it has been noted by GNATS. From: Robert Watson To: Mikolaj Golub Cc: Alexey Illarionov , bug-follo...@freebsd.org Subject: Re: kern/132734: panic in net/if_mib.c Date: Thu, 23 Apr 2009 20:33:43 +0100 (BST) On Thu, 23 Apr 2009, Mikolaj Golub

Re: panic in soabort

2009-04-23 Thread Robert Watson
On Thu, 23 Apr 2009, pluknet wrote: Please, give me comment on this. The panic is on 6.2-REL. Is it known to be fixed in the latter releases? It may well be -- there have been quite significant architectural improvements to socket life cycle (etc) between 6.2 and 7.x releases, which may well

Re: Forwarding w/o promisc on 6.4

2009-04-19 Thread Robert Watson
On Sun, 19 Apr 2009, Jon Otterholm wrote: I have a router running 6.4R that does not forward packets if I disable PROMISC on the interface. Hardware is a Dell PE with two Intel 82541EI chipsets (if_em). I have a number (~100) of vlan-interfaces on em0. Everything works as aexpected if I turn o

IFF_NEEDSGIANT now gone from 8.x (was: svn commit: r191253 - head/sys/net (fwd))

2009-04-18 Thread Robert Watson
:39:18 + (UTC) From: Robert Watson To: src-committ...@freebsd.org, svn-src-...@freebsd.org, svn-src-h...@freebsd.org Subject: svn commit: r191253 - head/sys/net Author: rwatson Date: Sat Apr 18 20:39:17 2009 New Revision: 191253 URL: http://svn.freebsd.org/changeset/base/191253 Log

Re: A Quick Question

2009-04-18 Thread Robert Watson
On Thu, 16 Apr 2009, Narek Gharibyan wrote: I would like to know is there any solution to problem show below, because we use FreeBSD 7.0 in our network structure and we are meeting face to face to this problem everyday\ Hi Narek: As Kip mentions, this panic message (a fatal trap in a softwar

Re: m_tag, malloc vs uma

2009-04-12 Thread Robert Watson
On Sat, 11 Apr 2009, Karim Fodil-Lemelin wrote: I think it would be desirable to make a change to more flexible m_tag types for 8.0, but I'm not sure I have time to implement/test it. Is this something you might be interested in working on? I'm thinking of basically replacing the m_tag_free

Re: m_tag, malloc vs uma

2009-04-10 Thread Robert Watson
On Fri, 10 Apr 2009, Karim Fodil-Lemelin wrote: Thank you for the answer, clear and concise. I asked the question because I had modified pf_get_mtag() to use uma directly in the hope that it would be faster then calling malloc. But since pf_mtag is 20bytes, malloc will end up using a fixed 32b

Re: m_tag, malloc vs uma

2009-04-10 Thread Robert Watson
On Fri, 10 Apr 2009, Karim Fodil-Lemelin wrote: Is there any plans on getting the mbuf tags sub-system integrated with the universal memory allocator? Getting tags for mbufs is still calling malloc in uipc_mbuf.c ... What would be the benefits of using uma instead? Hi Karim: Right now there

Re: Advice on a multithreaded netisr patch?

2009-04-08 Thread Robert Watson
On Wed, 8 Apr 2009, Barney Cordoba wrote: Is there any work being done on lighter weight locks for queues? It seems ridiculous to avoid using queues because of lock contention when the locks are only protecting a couple lines of code. My reading is that there are two, closely related, things

Re: Advice on a multithreaded netisr patch?

2009-04-07 Thread Robert Watson
On Tue, 7 Apr 2009, Barney Cordoba wrote: When I enabled LOCK_PROFILING my side modules, such as if_ibg, stopped working. It seems that the ifnet structure or something changed with that option enabled. Is there a way to sync this without having to integrate everything into a specific kernel

Re: Advice on a multithreaded netisr patch?

2009-04-07 Thread Robert Watson
On Tue, 7 Apr 2009, Barney Cordoba wrote: Have you tried LOCK_PROFILING? It would quickly tell you if driver locks were a source of significant contention. It works quite well... When I enabled LOCK_PROFILING my side modules, such as if_ibg, stopped working. It seems that the ifnet structu

Re: Advice on a multithreaded netisr patch?

2009-04-07 Thread Robert Watson
On Tue, 7 Apr 2009, Sepherosa Ziehau wrote: On Sun, Apr 5, 2009 at 9:34 PM, Ivan Voras wrote: Robert Watson wrote: On Sun, 5 Apr 2009, Ivan Voras wrote: I thought this has something to deal with NIC moderation (em) but can't really explain it. The bad performance part (not the jum

Re: Advice on a multithreaded netisr patch?

2009-04-07 Thread Robert Watson
On Mon, 6 Apr 2009, Julian Elischer wrote: while this is true, m_pullup ALWAYS does things so in fact you want to always put it in a test to see if it is really needed.. Then m_pullup() should be fixed? Keeping the expression of the pullup short makes the network code a lot more compact, wh

Re: Advice on a multithreaded netisr patch?

2009-04-07 Thread Robert Watson
On Tue, 7 Apr 2009, Sepherosa Ziehau wrote: This issue is almost entirely independent from things like the cache line miss issue, unless you hit the uncommon case of having to do work in m_pullup(), in which case life sucks. It would be useful to use DTrace to profile a number of the workfull

Re: Advice on a multithreaded netisr patch?

2009-04-06 Thread Robert Watson
On Mon, 6 Apr 2009, Ivan Voras wrote: I think we're talking slightly at cross purposes. There are two transfers of interest: (1) DMA of the packet data to main memory from the NIC (2) Servicing of CPU cache misses to access data in main memory By the time you receive an interrupt, the DMA is

Re: Advice on a multithreaded netisr patch?

2009-04-06 Thread Robert Watson
On Mon, 6 Apr 2009, Barney Cordoba wrote: Is there a way to give a kernel thread exclusive use of a core? I know you can pin a kernel thread with sched_bind(), but is there a way to keep other threads from using the core? On an 8 core system it almost seems that the randomness of more cores i

Re: Advice on a multithreaded netisr patch?

2009-04-06 Thread Robert Watson
On Mon, 6 Apr 2009, Ivan Voras wrote: I'd like to understand more. If (in netisr) I have a mbuf with headers, is this data already transfered from the card or is it magically "not here yet"? A lot depends on the details of the card and driver. The driver will take cache misses on the descri

Re: Advice on a multithreaded netisr patch?

2009-04-05 Thread Robert Watson
On Sun, 5 Apr 2009, Ivan Voras wrote: The argument is not that they are slower (although they probably are a bit slower), rather that they introduce serialization bottlenecks by requiring synchronization between CPUs in order to distribute the work. Certainly some of the scalability issues in

Re: Advice on a multithreaded netisr patch?

2009-04-05 Thread Robert Watson
On Sun, 5 Apr 2009, Barney Cordoba wrote: I'm curious as to your assertion that hardware transmit queues are a big win. You're really just loading a transmit ring well ahead of actual transmission; there's no need to force a "start" for each packet queued. You then have more overheard managing

Re: Advice on a multithreaded netisr patch?

2009-04-05 Thread Robert Watson
On Sun, 5 Apr 2009, Ivan Voras wrote: I thought this has something to deal with NIC moderation (em) but can't really explain it. The bad performance part (not the jump) is also visible over the loopback interface. FYI, if you want high performance, you really want a card supporting multiple

Re: Advice on a multithreaded netisr patch?

2009-04-05 Thread Robert Watson
On Sun, 5 Apr 2009, Ivan Voras wrote: I thought this has something to deal with NIC moderation (em) but can't really explain it. The bad performance part (not the jump) is also visible over the loopback interface. FYI, if you want high performance, you really want a card supporting multiple

Re: Advice on a multithreaded netisr patch?

2009-04-05 Thread Robert Watson
On Sun, 5 Apr 2009, Ivan Voras wrote: I'm developing an application that needs a high rate of small TCP transactions on multi-core systems, and I'm hitting a limit where a kernel task, usually swi:net (but it depends on the driver) hits 100% of a CPU at some transactions/s rate and blocks fur

Re: Freebsd7.0 - errors in compilation

2009-03-22 Thread Robert Watson
On Fri, 20 Mar 2009, rajasooriyar lakshman wrote: I am new to FreeBSDMy question is regarding a compilation of a routing protocol in FreeBSD 7.0. I am trying to run a adhoc routing protocol called, Dynamic Source Routing (DSR) on my freebsd router (7.0).. I have got the implementation fr

Re: HEADSUP: tcpdump updated to 4.0.0

2009-03-22 Thread Robert Watson
On Sat, 21 Mar 2009, Rui Paulo wrote: I just updated the base system tcpdump to 4.0.0. Most noticeably this new version includes a new output format. Please report any regressions. Cool, thanks for working on this! Robert N M Watson Computer Laboratory University of Cambridge __

Re: HEADS UP: IFF_NEEDSGIANT consumers to be disabled, removed (was: Re: Wiki page for non-MPSAFE network stack de-orbit scheduling)

2009-03-20 Thread Robert Watson
On Mon, 16 Feb 2009, Robert Watson wrote: The following schedule is proposed, assuming nothing goes horribly wrong with the new USB code in the next few weeks, and remaining nits relating to USB network and 802.11 drivers are handled: 16 February 2009HEADS UP to lists (this e-mail

HEADS UP: if_ar, if_sr, *old* USB stack network driver users

2009-03-15 Thread Robert Watson
On Sun, 15 Mar 2009, Robert Watson wrote: 16 February 2009HEADS UP to lists (this e-mail) 01 March 2009 Disable build of all IFF_NEEDSGIANT drivers in 8.x Slightly delayed while USB NDIS was sorted out, I will be moving ahead with disabling the build of all IFF_NEEDSGIANT drivers

Re: HEADS UP: IFF_NEEDSGIANT consumers to be disabled, removed (was: Re: Wiki page for non-MPSAFE network stack de-orbit scheduling)

2009-03-15 Thread Robert Watson
On Sun, 15 Mar 2009, Robert Watson wrote: On Mon, 16 Feb 2009, Robert Watson wrote: 16 February 2009HEADS UP to lists (this e-mail) 01 March 2009 Disable build of all IFF_NEEDSGIANT drivers in 8.x Slightly delayed while USB NDIS was sorted out, I will be moving ahead with

Re: HEADS UP: IFF_NEEDSGIANT consumers to be disabled, removed (was: Re: Wiki page for non-MPSAFE network stack de-orbit scheduling)

2009-03-15 Thread Robert Watson
On Mon, 16 Feb 2009, Robert Watson wrote: 16 February 2009HEADS UP to lists (this e-mail) 01 March 2009 Disable build of all IFF_NEEDSGIANT drivers in 8.x Slightly delayed while USB NDIS was sorted out, I will be moving ahead with disabling the build of all IFF_NEEDSGIANT

Re: kern/129846: [panic] /usr/sbin/ppp causes panic "Sleeping thread owns a non-sleepable lock

2009-03-13 Thread Robert Watson
On Thu, 12 Mar 2009, Richard Tector wrote: Robert Watson wrote: On Tue, 10 Mar 2009, Richard Tector wrote: Just a me too on 7.1-STABLE amd64 (as of 2009-03-05 at least). Note this does not happen with a 7.1-RELEASE kernel, so I'm assuming it's not userland. This problem is beli

Re: kern/129846: [panic] /usr/sbin/ppp causes panic "Sleeping thread owns a non-sleepable lock

2009-03-11 Thread Robert Watson
On Tue, 10 Mar 2009, Richard Tector wrote: Just a me too on 7.1-STABLE amd64 (as of 2009-03-05 at least). Note this does not happen with a 7.1-RELEASE kernel, so I'm assuming it's not userland. This problem is believed fixed as of r189531, committed 8 March -- could you try updating and see

Re: spliting kernel ipfw source ? (also involves sctp)

2009-03-02 Thread Robert Watson
On Mon, 2 Mar 2009, Rui Paulo wrote: On 1 Mar 2009, at 21:26, Julian Elischer wrote: Luigi Rizzo wrote: Hi, I am planning to split netinet/ip_fw2.c in a number of smaller files to make it more manageable, and while i do this I would also like to move the files related to ipfw2 (namely ip_fw

Re: HEADS UP: IFF_NEEDSGIANT consumers to be disabled, removed (was: Re: Wiki page for non-MPSAFE network stack de-orbit scheduling)

2009-02-24 Thread Robert Watson
On Mon, 16 Feb 2009, Robert Watson wrote: The following schedule is proposed, assuming nothing goes horribly wrong with the new USB code in the next few weeks, and remaining nits relating to USB network and 802.11 drivers are handled: 16 February 2009HEADS UP to lists (this e-mail

Re: bad usage of the shutdown system call produce a packet with null ip addresses

2009-02-23 Thread Robert Watson
On Mon, 23 Feb 2009, david guéluy wrote: By using a PFIL_HOOK on FreeBSD 7.1-prerelease, I notice that I receive some packets from 0.0.0.0 to 0.0.0.0. A buggy program in userland produce these packets when the shutdown system call is used on a socket which is not connected. Even if it's a

Re: panic: _rw_wlock_hard

2009-02-23 Thread Robert Watson
On Sat, 21 Feb 2009, Robert Watson wrote: Something like the following may help -- build-tested but not not run-time tested: I've now committed this patch to head as r188962 -- thanks for the report! Robert N M Watson Computer Laboratory University of Cambridge Index: in_

Re: panic: _rw_wlock_hard

2009-02-21 Thread Robert Watson
On Wed, 18 Feb 2009, dikshie wrote: does anyone see this panic on -HEAD: panic: _rw_wlock_hard: recursing but non-recursive rw radix node head @ /usr/src/sys/net/route.c:831 kdb_enter() at kdb_enter+0x3a panic() at panic+0x136 _rw_wlock_hard() at _rw_wlock_hard+0x66 _rw_wlock() at _rw_wlock+

Re: HEADS UP: IFF_NEEDSGIANT consumers to be disabled, removed

2009-02-17 Thread Robert Watson
On Tue, 17 Feb 2009, Krassimir Slavchev wrote: if_sr and if_ar are really simple and could probably be converted "trivially".. especially if their netgraph code is used. however I wonder if anyone still has that hardware (they are drivers for two sync serial cards). I still have such Digi/

HEADS UP: IFF_NEEDSGIANT consumers to be disabled, removed (was: Re: Wiki page for non-MPSAFE network stack de-orbit scheduling)

2009-02-16 Thread Robert Watson
(Bcc to arch@) On Mon, 26 May 2008, Robert Watson wrote: Just to keep track of things: http://wiki.freebsd.org/NONMPSAFE_DEORBIT Delayed by about six months, the merge and switch to the new USB stack in 8.x means that we're now fairly close to being able to pick up this project

Re: Multicast source address in recvfrom()

2009-02-03 Thread Robert Watson
* Robert Watson [090202 17:06]: On Sun, 1 Feb 2009, Derek Tattersall wrote: In order to become familiar with multicast implementation using FreeBSD, I found via Google a pair of test programs which multicast sent a simple text message and received the text message. I added some code to

Re: read() returns ETIMEDOUT

2009-02-02 Thread Robert Watson
On Sun, 1 Feb 2009, Mitar wrote: Is there any progress on this error reported: http://freebsd.monkey.org/freebsd-net/200805/msg00026.html I have the same or very similar issue. On my server large HTTP uploads are failing because there are only one direction data transmissions (when reading/

Re: Multicast source address in recvfrom()

2009-02-02 Thread Robert Watson
On Sun, 1 Feb 2009, Derek Tattersall wrote: In order to become familiar with multicast implementation using FreeBSD, I found via Google a pair of test programs which multicast sent a simple text message and received the text message. I added some code to report the source address, because non

Re: kern/129719: Panic during shutdown, tcp_ctloutput: inp == NULL

2009-02-02 Thread Robert Watson
The following reply was made to PR kern/129719; it has been noted by GNATS. From: Robert Watson To: Dan Nelson Cc: freebsd-gnats-sub...@freebsd.org, freebsd-b...@freebsd.org, freebsd-net@FreeBSD.org Subject: Re: kern/129719: Panic during shutdown, tcp_ctloutput: inp == NULL Date: Mon, 2

Re: kern/129719: Panic during shutdown, tcp_ctloutput: inp == NULL

2009-02-02 Thread Robert Watson
On Wed, 17 Dec 2008, Dan Nelson wrote: I've been trying to solve an intermittent connectivity problem where a server stops seeing incoming packets. It happened today, and when the system was shutting down, it paniced and rebooted. The gdb stack trace is a little mangled due to inlined functi

Re: TCP gets special treatment?

2009-02-01 Thread Robert Watson
On Wed, 28 Jan 2009, Per Hurtig (work) wrote: How differently are TCP packets treated compared to e.g. SCTP packets, while traversing the FreeBSD network stack (up to and including the IP-layer when using ipfw)?. I do not assume that the firewall (ipfw) is explicitly configured to check for e

Re: Still trying to debug mbuf leak.. Anyone familiar with uma_*?

2009-01-11 Thread Robert Watson
On Sun, 11 Jan 2009, Jason DiCioccio wrote: So I've got various vmcores here from kmem_malloc failures.. I already know that it's an excess of mbufs that's causing the crash, I'm just trying to figure out what the mbufs contain so that I can try and figure out where they came from.. Current

Re: TCP packet out-of-order problem

2009-01-06 Thread Robert Watson
the output of uname -a on the system? Thanks, Robert N M Watson Computer Laboratory University of Cambridge On Mon, Jan 5, 2009 at 9:13 PM, Robert Watson wrote: On Fri, 2 Jan 2009, Lin Jui-Nan Eric wrote: After running "netstat -s -p tcp", we found that lots of packets are discar

Re: TCP packet out-of-order problem

2009-01-05 Thread Robert Watson
On Fri, 2 Jan 2009, Lin Jui-Nan Eric wrote: After running "netstat -s -p tcp", we found that lots of packets are discarded due to memory problems. We googled for it, and found that sysctl oid "net.inet.tcp.reass.maxsegments" became 0, therefore packets never reassembled. Then we checked our

RE: net.inet.udp.blackhole issue

2009-01-03 Thread Robert Watson
On Tue, 30 Dec 2008, Yony Yossef wrote: I'm facing lots of UDP "Connection refused" errors while running multistream iperf test. Analyzing it with wireshark showed several "ICMP Port Unreachable" problems. I've overriden it with "sysctl net.inet.udp.blackhole=1", but I'm not sure this is the

Re: kernel network

2008-12-30 Thread Robert Watson
On Tue, 30 Dec 2008, Max Laier wrote: On Tuesday 30 December 2008 12:49:55 Ferner Cilloniz wrote: I do not think I could ever be more tired of this topic but I cannot seem to understand what to do. I have tried more about a month now to send arbitrary UDP packets from a kernel module but cann

Re: net.inet.udp.blackhole issue

2008-12-29 Thread Robert Watson
On Wed, 24 Dec 2008, Yony Yossef wrote: I'm facing lots of UDP "Connection refused" errors while running multistream iperf test. Analyzing it with wireshark showed several "ICMP Port Unreachable" problems. I've overriden it with "sysctl net.inet.udp.blackhole=1", but I'm not sure this is th

Re: NATM hardware available

2008-12-24 Thread Robert Watson
On Wed, 24 Dec 2008, Bruce Simpson wrote: Hartmut Brandt wrote: In any case there is still an Todo on my side: the routing information for NETNATM is currently lost somewhere between L2 and L3 :-) I guess I come back to you in the new year to fix this issue... Have to fetch my ATM equipment

  1   2   3   4   5   >