Re: [BUG] Kernel crash from DCCP Fwd: TTCP crash

I think we should hold back pushing any of these fixes into 2.6.14 until we've tracked them all down properly. Understand. Fine by me. Will learn about reference counting as don't know much about that yet. Ian - To unsubscribe from this list: send the line unsubscribe dccp in the body of a

Re: [BUG] Kernel crash from DCCP Fwd: TTCP crash

Folks, After this latest fix from Herbert we have another crash. This time it seems to be in skb_in-nh.raw being uninitialised in icmp.c if I understand everything I'm getting taught about debugging correctly. Not sure why this occurs but the information is a start. Regards, Ian Entry for

Re: [BUG] Kernel crash from DCCP Fwd: TTCP crash

Thanks. It appears that the stack is getting smashed by ip_options_echo. So please apply this debugging patch and give us the last stack trace you get before it crashes. OK. This will probably be tomorrow now unless Arnaldo wants to test sooner... Ian - To unsubscribe from this list: send

Re: [BUG] Kernel crash from DCCP Fwd: TTCP crash

I had just patched mine with all four corrected and had no problems but haven't tested with just two patched. I will if you tell me your logic though When I say no problems what I should say is Yeah!! The crash has gone away!! I didn't quite make that clear. Ian - To unsubscribe from

Re: [BUG] Kernel crash from DCCP Fwd: TTCP crash

On 18/10/05, Herbert Xu [EMAIL PROTECTED] wrote: On Tue, Oct 18, 2005 at 02:56:21PM +1300, Ian McDonald wrote: We also use ip_build_and_send_pkt and don't clear out that area before use. Going to do some tests to confirm my suspicision as I suspect it occurs when sending a reset packet

Re: OS/Application Latency issue

some variation on smaller packets you could quite easily get 25 msec. This is a large part of the reason for the proposed VoIP CCID which specifies 10 msec between packets. This may well suit your needs better... -- Ian McDonald http://wand.net.nz/~iam4 WAND Network Research Group University

Re: Please confirm this kernel crash using DCCP

, Ian -- Ian McDonald http://wand.net.nz/~iam4 WAND Network Research Group University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [NETFILTER]: DCCP conntrack/NAT

and run in bridge mode. It's great to have this out there though and appreciate it. Ian -- Ian McDonald http://wand.net.nz/~iam4 WAND Network Research Group University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL PROTECTED

[PATCH] 1/1 dccp: transmit buffering

Initial implementation of transmit buffering. Tested using netem with delay and loss also. Signed off by: Ian McDonald [EMAIL PROTECTED] diff --git a/include/linux/dccp.h b/include/linux/dccp.h index 088529f..4a2f845 100644 --- a/include/linux/dccp.h +++ b/include/linux/dccp.h @@ -411,6

Re: [PATCH] 1/1 dccp: transmit buffering

problems (I hit this in testing). I presume the best way is to create a sk_stop_timer_sync rather than use generic timer... Ian -- Ian McDonald http://wand.net.nz/~iam4 WAND Network Research Group University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp

[BUG] Problem with feature negotiation

39 ba e0 e8 69 b5 57 df 83 c4 10 8d 65 f4 31 c0 5b 5e 5f 5d c3 55 89 e5 57 8b 7d 08 56 53 8b 9f 78 03 00 00 8b 33 eb 46 8b 43 0c 85 c0 75 08 0f 0b ad 01 dd 37 ba e0 50 e8 -- Ian McDonald http://wand.net.nz/~iam4 WAND Network Research Group Department of Computer Science University of Waikato New

Re: [RFC] CCID infrastructure improvements

... Ian -- Ian McDonald http://wand.net.nz/~iam4 WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org

Re: panic in ccid3_hc_rx_packet_recv

On 2/17/06, Ian McDonald [EMAIL PROTECTED] wrote: On 2/17/06, Arnaldo Carvalho de Melo [EMAIL PROTECTED] wrote: Long session over a wi-fi link, tons of packets already transmitted then this panic, Ian, can you take a look? There are two divisions in ccid3_hc_rx_packet_recv... OK. Just

Re: [PATCH] 1/1 dccp: transmit buffering

as it uses ACK clocking in effect rather than being time based like CCID3. Comments/help really appreciated as I'm pulling my hair out and just want to use my extra lock Ian -- Ian McDonald http://wand.net.nz/~iam4 WAND Network Research Group Department of Computer Science University of Waikato

[BUG] DCCP : Suspected race in ackvec code

-- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL PROTECTED] More majordomo info

Re: net-2.6.17 rebased...

] ccid3: Divide by zero fix In rare circumstances 0 is returned by dccp_li_hist_calc_i_mean which leads to a divide by zero in ccid3_hc_rx_packet_recv. Explicitly check for zero return now. Update copyright notice at same time. Found by Arnaldo. Signed-off-by: Ian McDonald [EMAIL PROTECTED

Re: net-2.6.17 rebased...

F**k - just pasted in the wrong file. Trying again On 3/2/06, Ian McDonald [EMAIL PROTECTED] wrote: On 3/2/06, David S. Miller [EMAIL PROTECTED] wrote: This tree was getting crufty, so I rebased it today. It was actually a lot easier than I had anticipated. master.kernel.org

[BUG] Performance issues if DCCP packets are routed

at at some stage (along with updating tx buffering, CCID2 not working with loss/delay) but interested in seeing if others have similar problems? Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University

Updated version of iperf for DCCP

/-X option Help is fixed to display protocol help Can this please be updated on the iperf page? At some stage I think it would be worth merging some of the changes upstream but not entirely sure of the procedure. Can somebody clarify? Thanks, Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog

Re: [PATCH 1/1] DCCP Ackvec: fix soft lockup in ackvec handling code

for me. Signed-off-by: Ian McDonald [EMAIL PROTECTED] Thanks, Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand diff --git a/net/dccp/ackvec.c b/net/dccp/ackvec.c index

Locking validator output on DCCP

is occurring second and is actually creating a new socket so they are trying to lock on different sockets. Can someone tell me whether I am correct in my thinking or not? If I am then I will work out how to tell the lock validator not to worry about it. Thanks, Ian -- Ian McDonald Web: http

[PATCH 1/1] [DCCP]: Fix default sequence window size

but it is a step in the right direction. Compile tested only for IPv6 but not particularly complex change. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index f2c011f..3119804 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -32,6 +32,8

Re: [PATCH 11/11] CCID2: Add network profiling code for monitoring state

at http://vger.kernel.org/majordomo-info.html -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body

Re: [PATCH 07/11] CCID2: Tell DCCP to quickly check whether cwnd is available

majordomo info at http://vger.kernel.org/majordomo-info.html -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp

[PATCH 1/1] [DCCP]: Fix default sequence window size (revised)

but it is a step in the right direction. Compile tested only for IPv6 but not particularly complex change. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/feat.h b/net/dccp/feat.h index 6048373..b44c455 100644 --- a/net/dccp/feat.h +++ b/net/dccp/feat.h @@ -26,4 +26,6

[PATCH 1/1] [DCCP]: Shift sysctls

This applies on top of previous patch and shifts further sysctls into feat.h. No change in functionality - shifting code only. Tested. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/feat.h b/net/dccp/feat.h index b44c455..cee553d 100644 --- a/net/dccp/feat.h +++ b/net

Re: Locking validator output on DCCP

On 6/22/06, Ian McDonald [EMAIL PROTECTED] wrote: On 6/21/06, Arjan van de Ven [EMAIL PROTECTED] wrote: On Wed, 2006-06-21 at 10:34 +1000, Herbert Xu wrote: As I read this it is not a recursive lock as sk_clone is occurring second and is actually creating a new socket so they are trying

Re: mail voor linux mailinglist

. Nobody is working fulltime on this code so it can take a while to fix - feel free to help though! One note - when using CCID3 have you turned off ackvecs via the sysctl as that is really for CCID2... It's probably something we should do in the code actually. Ian -- Ian McDonald Web: http

[PATCH 1/1] [DCCP]: Introduce tx buffering

help Andrea's work with CCID2. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/include/linux/dccp.h b/include/linux/dccp.h index 676333b..2d7671c 100644 --- a/include/linux/dccp.h +++ b/include/linux/dccp.h @@ -438,6 +438,7 @@ struct dccp_ackvec; * @dccps_role - Role of this sock, one

Re: mail voor linux mailinglist - Performance measurements with the latest version of the DCCP (CCID 3)

without further reading whether we should be storing more history - it would help but I can't find it quickly reading whether it is a requirement. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University

[PATCH 1/1] [DCCP]: Fix default sequence window size

about 2.8 Mbits/sec to 3.5. This is still far too slow but it is a step in the right direction. Compile tested only for IPv6 but not particularly complex change. Signed off by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/feat.h b/net/dccp/feat.h index 6048373..b44c455 100644 --- a/net

[PATCH 1/1] [DCCP]: Shift sysctls

For 2.6.19 unless considered trivial enough. This applies on top of previous patch and shifts further sysctls into feat.h. No change in functionality - shifting code only. Tested. Signed off by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/feat.h b/net/dccp/feat.h index b44c455

Patches in progress

. Andrea - you have a lot of good patches at the moment - are you able to put yours on a web server easily as well? If not I'll just pull them all out of e-mail... Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer

[BUG] CCID3 - loss not detected if in order

therefore does not get set to anything but 0. This causes CCID3 to send packets too fast which explains some of the behaviour I'm seeing. I am thinking about how to fix this at the moment but as always others are free to dive in! Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http

Re: Help: how to get the latest code?

patches - in particular Andrea Bittau has written a lot of CCID2 patches to improve that. Also have a look at: http://linux-net.osdl.org/index.php/DCCP Regards, Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer

Re: Help: how to get the latest code?

, and the moving of congestion window. Which files am I suppose to look at? I'm not the CCID2 expert. Andra might be able to help you with it. Please try to keep the discussion on the dccp mailing list as I'm not the only (and certainly not the best!!) dccp progammer. -- Ian McDonald Web: http

Re: Help: how to get the latest code?

:) Also, thanks to Ian for the previous answer. malin For most patches save to a text file, change into the top level of your source code tree and then type: patch -p1 patchname -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department

Re: dccp: dead code?

of questions to dccp@vger.kernel.org as Arnaldo and I are often busy and others can help also. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list

Fwd: dccp: dead code?

Should have replied to list as well... -- Forwarded message -- From: Ian McDonald [EMAIL PROTECTED] Date: Jul 30, 2006 8:41 AM Subject: Re: dccp: dead code? To: Venkat Yekkirala [EMAIL PROTECTED] On 7/27/06, Venkat Yekkirala [EMAIL PROTECTED] wrote: Hi, I am trying to figure

Re: how to choose ccid3 in ttcp

). CCID3 performance has some issues which I'm working on at present too. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line

[PATCH 3/7] [DCCP]: Introduces follows48 function

This adds a new function to see if two sequence numbers follow each other. Signed off by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index b8931d3..84b477d 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -81,6 +81,14 @@ static inline u64 max48(const

[PATCH 4/7] [DCCP]: Introduce dccp_rx_hist_find_entry

This adds a new function dccp_rx_hist_find_entry. Signed off by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c index 7b6b03e..1c68182 100644 --- a/net/dccp/ccids/lib/packet_history.c +++ b/net/dccp/ccids/lib

[PATCH 5/7] [DCCP]: Fix CCID3 to correct performance

so box just acting as router with 1.2 msec RTT. The performance with this is the same with or without the patch at around 30 Mbit/s. Signed off by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index 0f85970..dad20c9 100644 --- a/net/dccp/ccids

[PATCH 6/7] [DCCP]: Shift sysctls into feat.h

This shifts further sysctls into feat.h. No change in functionality - shifting code only. Signed off by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/feat.h b/net/dccp/feat.h index b44c455..cee553d 100644 --- a/net/dccp/feat.h +++ b/net/dccp/feat.h @@ -27,5 +27,10 @@ extern int

Re: [PATCH 0/7] [DCCP]: Fixes and enhancements

be verified through iperf but we should implement what the RFC says. Basically the implementation in the DCCP code was buggy and was transmitting too fast so I have made it conform to the RFC much closer. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND

Re: low tx speed in ccid3 after loss

recalculates loss every time there is one and if you only get one then you could get in trouble... Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from

Re: low tx speed in ccid3 after loss

= max(min(X_calc, min_rate), s/t_mbi); If p doesn't change then X_calc doesn't change and therefore you can't go above X_calc Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New

Re: [PATCH 0/7] [DCCP]: Fixes and enhancements

On 8/27/06, David Miller [EMAIL PROTECTED] wrote: From: Ian McDonald [EMAIL PROTECTED] Date: Sun, 27 Aug 2006 16:57:17 +1200 Yes I see that now. However I can't see #5 in net-2.6.git in your tree or Linus' where 1-4 made it in... Resend it to me privately and I'll figure out what happened

[PATCH 0/1] [DCCP]: Tidy up code slightly

Dave, Arnaldo has pointed this one out to me in latest series of patches. Can this go into 2.6.18 please? (And I've checked for white space too!) Ian - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL PROTECTED] More majordomo info at

Re: [PATCH 0/1] [DCCP]: Tidy up code slightly

On 8/28/06, David Miller [EMAIL PROTECTED] wrote: From: Ian McDonald [EMAIL PROTECTED] Date: Mon, 28 Aug 2006 16:34:50 +1200 Arnaldo has pointed this one out to me in latest series of patches. Can this go into 2.6.18 please? It's not a bug fix, so we'll defer it to 2.6.19 I guess that's

Re: Fwd: CCID2 patches

about half a dozen times. Don't let pride get in your way. If you want to see why just google for reiserfs4. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe

Re: [PATCH 0/1] [DCCP]: Tidy up code slightly

On 8/28/06, David Miller [EMAIL PROTECTED] wrote: From: Ian McDonald [EMAIL PROTECTED] Date: Mon, 28 Aug 2006 16:34:50 +1200 Arnaldo has pointed this one out to me in latest series of patches. Can this go into 2.6.18 please? It's not a bug fix, so we'll defer it to 2.6.19 I haven't seen

Re: [PATCH 6/10]: [DCCP] CCID2: Tell DCCP to quickly check whether cwnd is available

. Remember Dave merged my tx queueing recently so this is available now Applied, but this is expensive. Consider having 100,000 DCCP sockets, some large percentage of which are polling in this manner, it simply doesn't scale. Agree. -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http

[PATCH 0/7] [DCCP]: Further fixes and enhancements

Here is my latest set of patches for DCCP. If possible I would like these to go into 2.6.19. I have tested against 2.6.18rc5 and latest net-2.6.19 git tree of Dave M as well. Dave - Patches 1 and 2 are trivial and just introducing constants and using them. Patch 4 is shifting some code into a

[PATCH 6/7] [DCCP]: Fix setting of packet size in CCID3

Set initial packet size to defaults as existing code doesn't work as set_sockopt occurs after initialisation so dccps_packet_size is of no use really. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index 7b4699a..e6c8e4c 100644

[PATCH 2/7] [DCCP]: Use constants for CCIDs

With constants for CCID numbers this now uses them in some places. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/ccids/ccid2.c b/net/dccp/ccids/ccid2.c index 457dd3d..2efb505 100644 --- a/net/dccp/ccids/ccid2.c +++ b/net/dccp/ccids/ccid2.c @@ -808,7 +808,7 @@ static void

[PATCH 3/7] [DCCP]: Introduce dccp_probe

This adds DCCP probing shamelessly ripped off from TCP probes by Stephen Hemminger. I've put in here support for further CCID3 variables as well. Andrea/Arnaldo might look to extend for CCID2. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/Kconfig b/net/dccp/Kconfig

Re: [PATCH] net/dccp: Allow default/fallback service code

DCCP_SOCKOPT_PACKET_SIZE is used for CCID3 to set default packet size for calculations. -DCCP_SOCKOPT_SERVICE sets the service. This is compulsory as per the -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science

Re: PACKET_SIZE option and congestion control on variable-length packets

On 10/4/06, Ian McDonald [EMAIL PROTECTED] wrote: On 10/3/06, Gerrit Renker [EMAIL PROTECTED] wrote: | Strategy | -- | 1/ Remove the PACKET_SIZE socket options as they don't help with the problem; | I have therefore updated Ian's patch to be used standalone [attached

Re: [PATCH 0/7] [DCCP]: Further fixes and enhancements

have an opinion on the style of these. These three patches need to be applied in order. Please don't merge patches 5-7 now as per other threads going on. These aren't needed I believe. I will still verify that is the case but can't yet do that. -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog

[PATCH 1/1] DCCP: Stop logs being filled

and causes logs to be filled. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index 0a21be4..47e3c2e 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -58,6 +58,9 @@ #define DCCP_RESOURCE_PROBE_INTERVAL ((u #define DCCP_RTO_MAX ((unsigned)(120 * HZ

Re: 2.6.19-rcX vs 2.6.20

it for the number of patches I do for the effort I would need to put in? I would like to see bug fixes in 2.6.19 too if they are obvious fixes e.g. the one I sent today. I won't get into the debate about other cosmetic ones Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http

DCCP into Dave M's net-2.6.20 tree?

Arnaldo, Are you able to merge Gerrit's tree containing many patches soon so we don't miss out on 2.6.20 and start to merge some of the growing backlog? Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science

Re: Sequence numbers in Response packets (Linux 2.6.17)

once you've sent your patch. You're doing well to work through these. Just wish I had more time to help! Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand

Re: [PATCH 0/7]: Retransmission Sequence Numbers

based rather than byte based as it makes more sense for a datagram based protocol. I'm not sure when I'll pick this up next. Of course you are free to take my code and use if you want. It is ugly though at present. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com

Re: CCID3: dead code in ccid3_hc_rx_packet_recv?

what it was... Anyway I'll have a look at it as I can but feel free to make patches. If you're removing chunks of code I'd prefer it commented out in the meantime rather than removed. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group

Re: [RFC] [PATCH]: use explicit enums for CCID 3 states

pile which makes no progress due to Real Life. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body

[RFC PATCH 1/2]: DCCP Introduce tx queue limits

I ask to merge) Appreciate feedback. This implements tx buffer sysctls for DCCP. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/include/linux/sysctl.h b/include/linux/sysctl.h index d98562f..4db45ec 100644 --- a/include/linux/sysctl.h +++ b/include/linux/sysctl.h @@ -612,6

[BUG] DCCP performance stops on CCID3 at times

Folks, Was chatting to Arnaldo about how I was working on this and he suggested to post to list. I'm still working on it but anybody else can feel free to help! I was getting the following in my logs sometimes when testing and the connection would basically stop: Nov 1 15:16:47 localhost

Re: [BUG] DCCP performance stops on CCID3 at times

which crippled my testing and I was glad to get sorted. I also removed some of the cc_val/window stuff at this point as I figured 80% working was better than 0% working as regards to congestion control... Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND

Re: [RFC PATCH 1/2]: DCCP Introduce tx queue limits

patches shortly. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL

[PATCH 1/2] [DCCP]: Implement tx_qlen sysctl

This implements tx buffer sysctls for DCCP. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/Documentation/networking/dccp.txt b/Documentation/networking/dccp.txt index c2328c8..b452ff8 100644 --- a/Documentation/networking/dccp.txt +++ b/Documentation/networking/dccp.txt @@ -104,6

[PATCH 0/2] [DCCP]: Introduce limit on tx_buffer

Folks, Given this a good testing and happy with it. Apologies for not being against very latest tree (I can't sync trees easily at moment). Gerrit has volunteered to fix up otherwise I will look at next week. This brings tx in DCCP up to a good state in my opinion. This should have been done

[PATCH 2/2] [DCCP]: Put limits on tx queue length

This puts a limit on transmit queue length and sends back EAGAIN if the buffer is full. Signed-off-by: Ian McDonald [EMAIL PROTECTED] --- diff --git a/net/dccp/proto.c b/net/dccp/proto.c index 2729fdc..6e3c311 100644 --- a/net/dccp/proto.c +++ b/net/dccp/proto.c @@ -633,6 +633,14 @@ int

Status of kernel.org servers??

without success. Can anybody point to whats going on as well at present and a timeline/plan to resolve? Thanks, Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand

[PATCH 0/3] DCCP: Minor tidyups

A few more minor changes that can hopefully make it into 2.6.20 tree Ian - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html

[PATCH 3/3] DCCP: Remove unneeded variable

diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index 2fa0c6d..55b972b 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -982,7 +982,6 @@ static void ccid3_hc_rx_packet_recv(stru const struct dccp_options_received *opt_recv; struct dccp_rx_hist_entry

Re: [PATCH 1/10]: Avoid congestion control on zero-sized data packets

On 11/25/06, Gerrit Renker [EMAIL PROTECTED] wrote: [CCID 3]: Avoid congestion control on zero-sized data packets Signed-off-by: Ian McDonald [EMAIL PROTECTED] I think in both of these functions we should change len to unsigned or make more explicit checks as we can have problems

Re: [PATCH 0/10] [RFC]: CCID 3 updates

obviously correct. Others I'm acking to say they look ok. I'll do some regression testing hopefully later in the week but I don't want to hold things up in case people want my opinion (as if I am that important... ha ha) Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com

Re: [PATCH 3/10]: Resolve small FIXME

On 11/25/06, Gerrit Renker [EMAIL PROTECTED] wrote: [CCID 3]: Resolve small FIXME Signed-off-by: Ian McDonald [EMAIL PROTECTED] -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New

Re: [PATCH 5/10]: Consistently update t_nom, t_ipi, t_delta

On 11/25/06, Gerrit Renker [EMAIL PROTECTED] wrote: [CCID 3] Consistently update t_nom, t_ipi, t_delta Acked-by: Ian McDonald [EMAIL PROTECTED] * ties updating these parameters to updating the sending rate X, exploiting that all three parameters in turn depend on X; and using a small

Re: [PATCH 9/10]: Add documentation for TFRC structures

On 11/25/06, Gerrit Renker [EMAIL PROTECTED] wrote: [NET]: Add documentation for TFRC structures Signed-off-by: Ian McDonald [EMAIL PROTECTED] -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University

Re: [RFC] [PATCH 1/2]: Remove small typo

On 11/29/06, Gerrit Renker [EMAIL PROTECTED] wrote: [CCID 3]: Remove small typo Acked-by: Ian McDonald [EMAIL PROTECTED] -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand

Re: [PATCH 0/2] [RFC]: Clean-ups

that the above is reason for it... Your thoughts Gerrit? (I haven't re-read the code or checked any of it out properly) Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand

Re: [PATCH 2/2]: Use `unsigned' for packet lengths

understand it and therefore we would break programs... Regards, Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp

Re: [PATCH 0/2] [RFC]: Clean-ups

On 11/29/06, Gerrit Renker [EMAIL PROTECTED] wrote: Quoting Ian McDonald: | The code here predates me but I think the reason for it might be to do | with RFC4342 (remember to read 3448 in conjunction with this). | | If the sender never receives a feedback packet from the receiver

Re: [PATCH 2/2]: Use `unsigned' for packet lengths

On 11/29/06, Gerrit Renker [EMAIL PROTECTED] wrote: Quoting Ian McDonald: | I think I didn't explain my point well here. You can't change to u32 | but need to be unsigned int (not u64). Don't get this: u32 is a 32-bit unsigned value and therefore looks sufficient - and you are proposing

Re: [RFC] [PATCH 1/2]: Remove small typo

On 11/29/06, Gerrit Renker [EMAIL PROTECTED] wrote: This is a Re-sent after discusssion with Ian. I have renamed the variable also, as the term got too long for the 80 columns fixed linelength - hope `t_nfb' is ok as mnemonic for the nofeedback timeout. Signed-off-by: Ian McDonald [EMAIL

Re: Packet size s Re: [PATCH 2/2]: Use `unsigned' for packet lengths

this with a few changes. I'd say merge average patch, I'll update mine to go on top (tomorrow??) and allow a mechanism to switch to use it. Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University

Re: Packet size s Re: [PATCH 2/2]: Use `unsigned' for packet lengths

. No deprecate it still. My earlier patches split it into two - a tx and rx packet size. -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send

Re: Packet size s Re: [PATCH 2/2]: Use `unsigned' for packet lengths

have regressed... Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer Science University of Waikato New Zealand - To unsubscribe from this list: send the line unsubscribe dccp in the body of a message to [EMAIL

Re: Packet size s Re: [PATCH 2/2]: Use `unsigned' for packet lengths

On 11/29/06, Arnaldo Carvalho de Melo [EMAIL PROTECTED] wrote: On 11/28/06, Ian McDonald [EMAIL PROTECTED] wrote: ttcp-r: accept from localhost.ghostprotocols.net ttcp-t: connect ttcp-t: 256000 bytes in 22.98 real seconds = 10.88 KB/sec +++ ttcp-t: 1000 I/O calls, msec/call = 23.53

Re:

situation of Burak, I started working with DCCP again and facing similar problems. Any clue? Leandro. 2006/11/29, Ian McDonald [EMAIL PROTECTED]: Burak, There are a lot of changes going on at present in the tree. I'm going to test later today to verify them. Are you able to give sample code

Setup of CCIDs

net.dccp.default.tx_ccid=20 My opinion is that the send_ackvec should automatically change in the code when you change CCIDs. The default_seq_window increases from the default specified in the RFC of 100. Ian On 11/30/06, Ian McDonald [EMAIL PROTECTED] wrote: Hi Leandro, I'm doing some tests

Re:

Folks, Another thing to note is that with a limit on tx queue sizes now you will get EAGAIN if the queue is full... Ian On 11/30/06, Ian McDonald [EMAIL PROTECTED] wrote: Hi Leandro, I'm doing some tests again today. Have you got some code/results you can share? Also can you both try

[BUG] CCID3 no longer works!

: [ 645.678678] ccid3_hc_tx_packet_recv: Illegal ACK received - no packet has been sent Will start by look at that patch and if necessary then start bisecting. Regards, Ian -- Ian McDonald Web: http://wand.net.nz/~iam4 Blog: http://imcdnzl.blogspot.com WAND Network Research Group Department of Computer

Re: [BUG] CCID3 no longer works!

On 11/30/06, Ian McDonald [EMAIL PROTECTED] wrote: There are some clues at least on jandi3 log: Nov 30 11:57:49 localhost kernel: [ 645.677995] ccid3_hc_tx_packet_recv: Illegal ACK received - no packet has been sent Nov 30 11:57:49 localhost kernel: [ 645.678678] ccid3_hc_tx_packet_recv

Re: [PATCH] [RFC]: Fine-grained sending rate resolution / division errors

with that too when I converted the code from FreeBSD to Linux (and they had floating point as well...) I never got further than what went into the code. This all looks good in theory. Still trying to resolve earlier DCCP CCID3 issues. Acked-by: Ian McDonald [EMAIL PROTECTED] Ian -- Web: http

Re: [DCCP]: Fix BUG in send time calculation

On 12/1/06, Gerrit Renker [EMAIL PROTECTED] wrote: I think I found the bug, see description below. The problem was clearly in CCID 3, with CCID 2 performance was good. I think it makes sense to use the other patch (RTO) as well. CCID 3 needs some more work, but I think with this the reported

Re: [DCCP]: Fix BUG in send time calculation

On 12/1/06, Gerrit Renker [EMAIL PROTECTED] wrote: Commit Message - [DCCP]: Fix BUG in retransmission delay calculation Signed-off-by: Ian McDonald [EMAIL PROTECTED] I can confirm this fixes the problem in this case. Performance still sucks

  1   2   3   4   5   >