tcp retransmission concept.
Hi, Is any one know how tcp retransmission works inside kernel ? regards varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: gso vs tso
If both are ON then which one works tso or gso ? On Wed, Nov 19, 2014 at 12:22 PM, Varun Sharma wrote: > If both are ON then which one works tso or gso ? > > > > On Wed, Nov 19, 2014 at 12:07 PM, Raj Ravi wrote: >> TSO is a hardware feature whereas GSO is a software feature. >> >> In the sense, TSO needs "the device" to break the frame into MTU sized >> segments. >> BUT, GSO is a strategy followed by kernel to avoid processing smaller >> packets throughout the stack. >> >> Refer this: >> http://blogs.gnome.org/markmc/2008/05/28/checksums-scatter-gather-io-and-segmentation-offload/ >> >> <<< >> GSO is a generalisation of this in the kernel. The idea is that you >> delay segmenting a packet until the latest possible moment. In the >> case where a device doesn’t support TSO, this would be just before >> passing the skb to the driver. If the device does support TSO, the >> unsegmented skb would be passed to the driver. >> <<< >> >> >> On Wed, Nov 19, 2014 at 11:40 AM, Varun Sharma wrote: >>> Hi, >>> >>> I have basic question about gso(generic segmentation offload) vs >>> tso(TCP segmentation offload) >>> >>> 1. How both are different to each other? >>> >>> 2. If both are ON using ethtool which one is perform segmentation gso or >>> tso ? >>> >>> >>> Regards >>> varun >>> >>> ___ >>> Kernelnewbies mailing list >>> Kernelnewbies@kernelnewbies.org >>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
gso vs tso
Hi, I have basic question about gso(generic segmentation offload) vs tso(TCP segmentation offload) 1. How both are different to each other? 2. If both are ON using ethtool which one is perform segmentation gso or tso ? Regards varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
skb headroom and tailroom
Hi, I have few basic question regarding headroom and tailroom in skb. 1. At time of skb allocation sk->sk_prot->max_header value is 272 . If we add tcp header+ip header+mac header outcome will always less than 272 .Then why allocate more ? Is it for future case ? 2. Why space left in headroom and tailroom ? Regards Varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
packet drop
Hi All, I am sending packet between two back to back connected machine.when I use same network (192.168.7.x) packet reach upto NF_IP_LOCAL_IN (IPTABLES INPUT) hook.It is accepted. But when I spoof ip address network(192.168.8.x) it pass NF_IP_PRE_ROUTING hook but dropped before NF_IP_LOCAL_IN hook. case 1: 192.168.7.12 --> 192.168.7.19 Accepted. case 2: 192.168.8.25 --> 192.168.7.19 Dropped. I didn't get why it dropped ? How network change effect on routing inside kernel tcp/ip stack ? Regards Varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
how tso (tcp segmentation offload) effect TCP bidirectional test sending throughput.
Hi, if we off tso (tcp segmentation offload) on sending side , *TCP Bidirectional* test sending throughput decrease as compare to *TCP Unidirectional *test sending throughput . whereas if we on tso(tcp segmentation offload) on sending side , both testgive almost same sending throughput result. is anyone know how tso(tcp segmentation offload) effect TCP Bidirectional sending throughput ? i am using iperf-2.0.5 for this case. Thanks in advance. Regards varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Sending rate decrease in TCP bidirectional test
Hi, I am using iperf v 2.0.5 for testing 10G card. Its a dual port card. Two 16 core machines with 64GB RAM are connected back to back. In *TCP Bidirectional test* case sending throughput decrease as compare to *TCP Unidirectional test* case sending . All cases use default setting . No extra parameter is set. In case of 4 Client threads Output comes :-- Unidirectional Send Process --- *9.6* Gbps Bidirectional Send Process *4.9* Gbps In case of 8 Client threads Output comes :-- Unidirectional Send Process --- *9.7* Gbps Bidirectional Send Process -- *6.4* Gbps In case of 16 Client threads Output comes :-- Unidirectional Send Process ---* 9.7* Gbps Bidirectional Send Process --*8* Gbps Is this problem regarding TCP/IP stack or nic card ? Regards varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: IPFIX discussion
i am using ipfix for exporting flow record .when i am sending flow record to collector , on collector side one flow record show as triple record .{one flow record contain source ip,dest ip source port ,dest port ,protocol } .all this keys shows as sourceipv4 address in collector output.even i am sending correct element id of flow keys . do u know where is the problem ? i check template format also On Wed, Sep 11, 2013 at 8:45 PM, Robert Clove wrote: > ya i lot a little bit. > What are you doing on IPFIX? > what's your queries. > > > On Wed, Sep 11, 2013 at 1:31 AM, Varun Sharma wrote: > >> Hi, >> >> Is there any one who knows about IPFIX ? I want to discuss about some >> question on ipfix . >> >> Varun >> >> ___ >> Kernelnewbies mailing list >> Kernelnewbies@kernelnewbies.org >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >> >> > ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
IPFIX discussion
Hi, Is there any one who knows about IPFIX ? I want to discuss about some question on ipfix . Varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
IPFIX collector
Hi , Is there any open source IPFIX collector available for testing purpose . Regards Varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
ipfix exporter source code
Hi , Is there any open source ipfix exporter available with source code ? varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: package compile stop due to make error
makefile bulid after ./configure in both palce. one in src folder and anotherone in fprobe folder .how to compare these two makefile. On Tue, Jul 16, 2013 at 2:12 PM, Youssouf Said wrote: > You need a makefile in src with the correct target > Le 16 juil. 2013 12:32, "Varun Sharma" a écrit : > >> Hi, >> >> I try to compile fprobe package .After ./configure command , when i run >> make command compile process failed >> >> # make >> >> make all-recursive >> make[1]: Entering directory `/home/fprobe-1.1' >> Making all in src >> make[2]: Entering directory `/home/fprobe-1.1/src' >> make[2]: Nothing to be done for `all'. >> make[2]: Leaving directory `/home/fprobe-1.1/src' >> make[2]: Entering directory `/home/fprobe-1.1' >> make[2]: Nothing to be done for `all-am'. >> make[2]: Leaving directory `/home/fprobe-1.1' >> make[1]: Leaving directory `/home/fprobe-1.1' >> >> can anyone knows how to resolve this problem . >> >> Regards >> Varun Sharma >> >> ___ >> Kernelnewbies mailing list >> Kernelnewbies@kernelnewbies.org >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >> >> ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
package compile stop due to make error
Hi, I try to compile fprobe package .After ./configure command , when i run make command compile process failed # make make all-recursive make[1]: Entering directory `/home/fprobe-1.1' Making all in src make[2]: Entering directory `/home/fprobe-1.1/src' make[2]: Nothing to be done for `all'. make[2]: Leaving directory `/home/fprobe-1.1/src' make[2]: Entering directory `/home/fprobe-1.1' make[2]: Nothing to be done for `all-am'. make[2]: Leaving directory `/home/fprobe-1.1' make[1]: Leaving directory `/home/fprobe-1.1' can anyone knows how to resolve this problem . Regards Varun Sharma ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: tcpdump hang and system gone in freeze state
As such no change in code .I just write printk statement in kernel code flow .specfically in send and recv system call flow . On Fri, Jul 12, 2013 at 10:29 AM, Youssouf Said wrote: > Can you list what changes have you done ? > > Varun Sharma a écrit : > > > Hi, > > I build new kernel (2.6.32.60). when i execute tcpdump on this build > kernel ,tcpdump gone in hang state and system gone in freeze state . > > > ltrace tcpdump output is--- > > __libc_start_main(0x452220, 1, 0x7fff623b6158, 0x462d70, 0x462d60 > > strrchr("tcpdump", '/')= NULL > getopt(1, 0x7fff623b6158, "aAB:c:C:dDeE:fF:G:i:IKlLm:M:nNOp"...) = -1 > time(NULL) = 1373602325 > gmtime(0x7fff623b5e68) = 0x313eba03e0 > localtime(0x7fff623b5e68) = 0x313eba03e0 > getuid() = 0 > pcap_lookupdev(0x7fff623b5f30) = "eth0" > pcap_create(0x314d636140, 0x7fff623b5f30, 0x11f6690, 0x313eb9ce90, 11) = > 0x11f6ac0 > pcap_set_snaplen(0x11f6ac0, 65535, 0x8038a0, 0x306874, 0) = 0 > pcap_set_promisc(0x11f6ac0, 1, 0, 0x306874, 0) = 0 > pcap_set_timeout(0x11f6ac0, 1000, 0, 0x306874, 0) = 0 > pcap_activate(0x11f6ac0, 0, 0, 0x306874, 0 > > > tcpdump hang in pcap_activate . whereas if run tcpdump in original kernel > it works fine.can anybody know the reason? > ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
tcpdump hang and system gone in freeze state
Hi, I build new kernel (2.6.32.60). when i execute tcpdump on this build kernel ,tcpdump gone in hang state and system gone in freeze state . ltrace tcpdump output is--- __libc_start_main(0x452220, 1, 0x7fff623b6158, 0x462d70, 0x462d60 strrchr("tcpdump", '/')= NULL getopt(1, 0x7fff623b6158, "aAB:c:C:dDeE:fF:G:i:IKlLm:M:nNOp"...) = -1 time(NULL) = 1373602325 gmtime(0x7fff623b5e68) = 0x313eba03e0 localtime(0x7fff623b5e68) = 0x313eba03e0 getuid() = 0 pcap_lookupdev(0x7fff623b5f30) = "eth0" pcap_create(0x314d636140, 0x7fff623b5f30, 0x11f6690, 0x313eb9ce90, 11) = 0x11f6ac0 pcap_set_snaplen(0x11f6ac0, 65535, 0x8038a0, 0x306874, 0) = 0 pcap_set_promisc(0x11f6ac0, 1, 0, 0x306874, 0) = 0 pcap_set_timeout(0x11f6ac0, 1000, 0, 0x306874, 0) = 0 pcap_activate(0x11f6ac0, 0, 0, 0x306874, 0 tcpdump hang in pcap_activate . whereas if run tcpdump in original kernel it works fine.can anybody know the reason? ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
neigh_hh_output function
Hi, can anyone briefly help me to understand which value is copy in skb through neigh_hh_output(include/net/neighbour.h) ? I print that value but that is not mac address . I see the hh_len comes from struct hh_cache is 14 .Due to this i think mac_header bulid here .But on exploring i found that mac header build in eth_header function .Now What is value copy in skb through neigh_hh_output function ? Regards Varun ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
tcp window concept
Hi, I am sending 5000 bytes of data in send system call. struct tcp_sock has max_window and snd_wnd two variable.Which are increase as packet sent out . whereas tcp header window variable value is fix for all packets.can anyone tell what is relationship b/w these windows ? and why max_window and snd_wnd size increase after packets sent out ? ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
kdb understanding
Hi, I am using 2.6.32.60 kernel . I want to use kdb on this kernel. Pls suggest me some tutorials or links that help me to understand kdb. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: assembly of packets
I see the receiving side also .Packets are not assembled by ip-defrag . now where packets are assembled at transport layer ? if yes then how they are assembled ? On Tue, Jun 4, 2013 at 7:24 PM, Craig Jackson wrote: > > -Original Message- > > From: Varun Sharma [mailto:vsd...@gmail.com] > > Sent: Tuesday, June 04, 2013 2:45 AM > > To: Craig Jackson; kernelnewbies@kernelnewbies.org; > > valdis.kletni...@vt.edu > > Subject: Re: assembly of packets > > > > On receving side upto tcp layer skb->len=1480 now where is packet > > assembled on tcp layer ? > > or it is directly copies into user space buffer. > > > > I am sending len = 1 in send system call. > > > >> On Mon, Jun 3, 2013 at 9:45 PM, Craig Jackson > >> wrote: > >> In my experience, the second paragraph isn't quite true. What you see > >> with TSO is the pre-segmentation "packet", up to 65k. (By this I mean > >> the set of data which is given to the offload hardware to segment.) So > >> you need to make sure that your "-s" value is big enough to see > >> everything. > >> > >> (Speaking as someone who was bitten by bugs in the early versions of > >> TSO.) > >> > >> Craig > > > >>> -Original Message- > >>> From: kernelnewbies-bounces+cjackson=ebscohost@kernelnewbies.org > >>> [mailto:kernelnewbies-bounces+cjackson=ebscohost@kernelnewbies.org > ] > >>> On Behalf Of valdis.kletni...@vt.edu > >>> Sent: Monday, June 03, 2013 12:03 PM > >>> To: Varun Sharma > >>> Cc: kernelnewbies@kernelnewbies.org > >>> Subject: Re: assembly of packets > >>> > >>> On Mon, 03 Jun 2013 14:48:41 +0530, Varun Sharma said: > >>> > >>> > If TSO(tcp segmentation offload ) is enabled then nic card handle > >>> > segmentation then where is assembly of packets happens ? Is it > >>> > tcp_rcv_established function ? > > > >>> The whole *point* of TSO is so the NIC does all the segmentation > >>> reassembly and DMA, and wake the kernel up when all the data is already > >>> stashed in buffers fully processed. > > > >>> Incidentally, this is why if you run tcpdump on an interface that has > >>> TSO enabled, you'll only see the first 3 handshaking packets and the > >>> final FIN packets - the other packets wake up the TCP stack at a point > >>> after where tcpdump's tap would have seen the packet. > > Sorry for top-posting before. It's incredibly hard to get Outlook 2007 to > do things right. > > Varun, I now see we may not have answered the question you have. TSO (TCP > Segementation Offload) > is associated with sending packets, not receiving them. They are > re-assembled at the receiving > end in the standard manner. > > Are you actually thinking of GRO (Generic Receive Offload)? > > Craig > ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
Re: assembly of packets
On receving side upto tcp layer skb->len=1480 now where is packet assembled on tcp layer ? or it is directly copies into user space buffer. I am sending len = 1 in send system call. On Mon, Jun 3, 2013 at 9:45 PM, Craig Jackson wrote: > In my experience, the second paragraph isn't quite true. What you see with > TSO is the pre-segmentation "packet", up to 65k. (By this I mean the set of > data which is given to the offload hardware to segment.) So you need to > make sure that your "-s" value is big enough to see everything. > > (Speaking as someone who was bitten by bugs in the early versions of TSO.) > > Craig > > -Original Message- > From: kernelnewbies-bounces+cjackson=ebscohost@kernelnewbies.org[mailto: > kernelnewbies-bounces+cjackson=ebscohost@kernelnewbies.org] On Behalf > Of valdis.kletni...@vt.edu > Sent: Monday, June 03, 2013 12:03 PM > To: Varun Sharma > Cc: kernelnewbies@kernelnewbies.org > Subject: Re: assembly of packets > > On Mon, 03 Jun 2013 14:48:41 +0530, Varun Sharma said: > > > If TSO(tcp segmentation offload ) is enabled then nic card handle > > segmentation then where is assembly of packets happens ? Is it > > tcp_rcv_established function ? > > The whole *point* of TSO is so the NIC does all the segmentation > reassembly and DMA, and wake the kernel up when all the data is already > stashed in buffers fully processed. > > Incidentally, this is why if you run tcpdump on an interface that has TSO > enabled, you'll only see the first 3 handshaking packets and the final FIN > packets - the other packets wake up the TCP stack at a point after where > tcpdump's tap would have seen the packet. > ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
assembly of packets
Hi, If TSO(tcp segmentation offload ) is enabled then nic card handle segmentation then where is assembly of packets happens ? Is it tcp_rcv_established function ? ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
tracing send and recv system call
Hi, I am tracing send and recv system call .From user space to kernel space switching send in kernel space not going into sys_send .It directly going into sys_sendto . I am using linux 2.6.32.60 for this operation.my printk statement is printed in sys_sendto but not in sys_send. Is it right way of calling. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
which function perform gso
Hi, which function perform Generic segmentation offload in kernel. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
c library wrapper function
Hi, I am tracing send system call . before going to kernel space system call no. conversion happens in c library wrapper function.how to get the name of that library which hold wrapper function of send system call ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
about ssh
Hi, I want to know how two independent ssh session use same port for communication .At what level difference is come. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
mss for tcp segment
Hi, What is the value of mss for tcp segment. ___ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies