Dotaz: zahada, alebo kde mam chybu? :prolem so sietou na architekture arm-pxa

[vlk]

ahoj,

davam dohromady jeden taky hardware colibri..

"embedded" modul linuxovym kernelom priamo nieje podporovany (aj ked vsetky jeho sucasti podporovane su), takze bolo potrebne doplnit mikroarchitekturu pre kernel .. z nejakych existujucich patchov pre starsie kernely aj nejakych novsich neuplnych patchov sa mi podarilo dat dohromady funkcny kernel, busybox a nejak dat dohromady root filesystem.

funkcne bolo hadam vsetko zakladne co som potreboval (flash, ram, siet, usb, grafika, ...) okrem jednej hadam najdolezitejsej veci, ale tazko budem tvrdit ze nefunkcna, pretoze ono to nejak funguje.

konkretne funguju len pingy a NFS spojenie naviazane pocas bootovania kernelu root-nfs, kedy som ziskal aj IP adresu cez dhcp.

s konzoly modulu colibri to vyzera nasledovne:

# ifconfig
eth0      Link encap:Ethernet  HWaddr 11:22:33:44:55:66
          inet addr:172.16.0.2  Bcast:172.16.0.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:1999 errors:0 dropped:0 overruns:0 frame:0
          TX packets:931 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:2420001 (2.3 MiB)  TX bytes:222558 (217.3 KiB)
          Interrupt:178

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:25 errors:0 dropped:0 overruns:0 frame:0
          TX packets:25 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:1884 (1.8 KiB)  TX bytes:1884 (1.8 KiB)

# route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
172.16.0.0      0.0.0.0         255.255.255.0   U     0      0        0 eth0
0.0.0.0         172.16.0.1      0.0.0.0         UG    0      0        0 eth0

# httpd -h /www/

# wget http://172.16.0.2/index.html
Connecting to 172.16.0.2 (172.16.0.2:80)
index.html           100% |*******************************|     6  --:--:-- ETA

vsetko vyzera uplne normalne a aj normalne funguje. dokonca aj web server je funkcny z lokalnej strany..

ine je to z druhej strany z PC...

# ifconfig
eth0      Link encap:Ethernet  HWaddr 00:0d:56:ec:70:1f
          inet addr:172.16.0.1  Bcast:172.16.0.255  Mask:255.255.255.0
          inet6 addr: fe80::20d:56ff:feec:701f/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:189937 errors:1 dropped:0 overruns:0 frame:2
          TX packets:239993 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:49157002 (46.8 MiB)  TX bytes:184870456 (176.3 MiB)
          Interrupt:11
...

# route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
172.16.0.0      0.0.0.0         255.255.255.0   U     0      0        0 eth0
192.168.1.0     0.0.0.0         255.255.255.0   U     0      0        0 wlan0
0.0.0.0         192.168.1.1     0.0.0.0         UG    0      0        0 wlan0

# ping 172.16.0.2
PING 172.16.0.2 (172.16.0.2) 56(84) bytes of data.
64 bytes from 172.16.0.2: icmp_seq=1 ttl=64 time=2.77 ms
64 bytes from 172.16.0.2: icmp_seq=2 ttl=64 time=0.284 ms
--- 172.16.0.2 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1002ms
rtt min/avg/max/mdev = 0.284/1.529/2.774/1.245 ms

# wget http://172.16.0.2/index.html
--22:18:42--  http://172.16.0.2/index.html
           => `index.html'
Connecting to 172.16.0.2:80... failed: Connection timed out.

# nmap 172.16.0.2
Starting Nmap 4.50 ( http://insecure.org ) at 2008-01-06 22:25 CET
Note: Host seems down. If it is really up, but blocking our ping probes, try -PN
Nmap done: 1 IP address (0 hosts up) scanned in 2.127 seconds

nmap nenasiel nic ani s parametrom -PN ... len ze vsetky porty su filtrovane, podobne to skoncilo aj pri teste UDP

skusal som rozne finty, ale stale to bolo podobne, aj som kernel funkcny a rootfs napalil do internej flash, potom sa cez program udhcpc snazil ziskat IP adresu, alebo nastavit IP rucne... hadam vsetko ... a stale to neslo.

... vyzera to asi na chybu hardwaru, ale potom preco funguju pingy a spominane NFS ? .. takze skor to bude asi chyba ovladacov v kerneli.. (v povodnom windows CE som to nestihol vyskusat)

nema stym (aj niecim podobnym) nahodou niekto skusenosti ??

dik za kazdu radu.

palo.

Odpovědi

[matoo]

nazdar,
size s architekturou arm-pxa nemam ziadne skusenosti, ale tak ked sa spytam:

-o ako vyzera netstat -nat na 0.2 masine ? co hovori tcpdump ked sa snazis wgetnut ten index? IMHO telnet je asi praktickejsi na vyskusanie funkcnosti az po 7 vrstvu

-o kedze si sa dopingal z oboch stran na masiny, problem teda po 3vrstvu by si mat nemal .. naahodou nebezi na 0.2 masine FW? resp. nie je nastaveny nejaky iny wrapper, ktory blokuje incomming traffic? (okrem pingov teda)

[vlk]

no netstat vyzera nasledovne ...

# netstat -nat
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address           Foreign Address         State
tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN

v pripade pokusu pripojit sa cez telnet tak to skonci vzdy bez odozvy ... podla tcpdump (na masine .0.1) nieje ziadna odpoved na ten packet ... vyzera to ako keby (.0.2) ignoroval kazde spojenie ... ako keby tam bol firewall, ale firewall nemam v kerneli nakonfigurovany, a ani zapnuty (zapnuty nieje ani na .0.1), takze tymto asi nebude.

pokus pripojit sa cez telnet:

# telnet 172.16.0.2 80
Trying 172.16.0.2...
telnet: Unable to connect to remote host: Connection timed out

tcpdump pocas pripajania na http z pocitaca .0.1 vyzera nasledovne (napriklad je viediet ze arp dotazy funguju OK)

# tcpdump -ni eth0 host 172.16.0.2
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes
10:17:01.134663 arp who-has 172.16.0.1 tell 172.16.0.2
10:17:01.134689 arp reply 172.16.0.1 is-at 00:0d:56:ec:70:1f
10:17:07.079324 IP 172.16.0.1.59663 > 172.16.0.2.80: S 3756444208:3756444208(0) win 5840 <mss 1460,sackOK,timestamp 1064553 0,nop,wscale 6>
10:17:10.075795 IP 172.16.0.1.59663 > 172.16.0.2.80: S 3756444208:3756444208(0) win 5840 <mss 1460,sackOK,timestamp 1065303 0,nop,wscale 6>
10:17:16.075784 IP 172.16.0.1.59663 > 172.16.0.2.80: S 3756444208:3756444208(0) win 5840 <mss 1460,sackOK,timestamp 1066803 0,nop,wscale 6>
10:17:28.075794 IP 172.16.0.1.59663 > 172.16.0.2.80: S 3756444208:3756444208(0) win 5840 <mss 1460,sackOK,timestamp 1069803 0,nop,wscale 6>
10:17:52.075799 IP 172.16.0.1.59663 > 172.16.0.2.80: S 3756444208:3756444208(0) win 5840 <mss 1460,sackOK,timestamp 1075803 0,nop,wscale 6>

tiez som skusal sa pripojit opacnou cestou z .0.2 na .0.1 (web server som si ce pusteny nemal, takze pripajal som sa na ssh cez telnet) a vyzeralo to nasledovne:

# telnet 172.16.0.1 22
telnet: cannot connect to remote host (172.16.0.1): Connection timed out

v tcpdumpe bolo vidiet ze packety z .0.1 dorazili, aj pocitac nane odpovedal, ale .0.1 odpovede odignoroval (vo vypise bol este traffic od nfs, ktory som nepastoval):

10:30:18.732415 arp who-has 172.16.0.1 tell 172.16.0.2
10:30:18.732459 arp reply 172.16.0.1 is-at 00:0d:56:ec:70:1f
10:30:18.923553 IP 172.16.0.2.48417 > 172.16.0.1.22: S 279310633:279310633(0) win 5840 <mss 1460,sackOK,timestamp 79536 0,nop,wscale 1>
10:30:18.923588 IP 172.16.0.1.22 > 172.16.0.2.48417: S 3274126934:3274126934(0) ack 279310634 win 5792 <mss 1460,sackOK,timestamp 1262514 79536,nop,wscale 6>
10:30:21.922706 IP 172.16.0.2.48417 > 172.16.0.1.22: S 279310633:279310633(0) win 5840 <mss 1460,sackOK,timestamp 79836 0,nop,wscale 1>
10:30:21.922744 IP 172.16.0.1.22 > 172.16.0.2.48417: S 3274126934:3274126934(0) ack 279310634 win 5792 <mss 1460,sackOK,timestamp 1263264 79536,nop,wscale 6>
10:30:22.519794 IP 172.16.0.1.22 > 172.16.0.2.48417: S 3274126934:3274126934(0) ack 279310634 win 5792 <mss 1460,sackOK,timestamp 1263414 79536,nop,wscale 6>
10:30:23.731785 arp who-has 172.16.0.2 tell 172.16.0.1
10:30:23.731975 arp reply 172.16.0.2 is-at 11:22:33:44:55:66

v podstate podla toho mozem tvrdit ze niektore prichodzie packety to ignoruje ... ale nechapem preco :(

[matoo]

hmm .. to je fakt zaujimave .. mno podla tohto vypisu to vyzera, ako keby bol zakazany akykolvek traffic IN na 0.2 sietovke .. ale vravis, ze fw nie je zapnuty.. pozeral si aj wrappery ? hosts.allow ?
lebo je zaujimave, ze sa masiny dopingaju navzajom ..

[vlk]

ahoj, trosku sa to posunulo ... ale stale neviem kde moze byt problem ...

podarilo sa mi nainstalovat tcpdump aj do toho arm modulu..

a podla vsetkeho packet dorazi, len nezareaguje daemon - neodosle odpoved.

rozmyslam ci mi este nechyba nejaka sluzba, lebo linux som si spravil na principe "from scratch" a spustam len init, mountnem /proc /sys a spustim /bin/sh

[vlk]

dik vsetkym ... chyba bola (tradicne) medzi klavesnicou a stolickou ... zla MAC adresa ... som si vymyslel prave taku ktora bola v rozsahu multicastovych a preto to neslo :(

[matoo]

aax - aax.. toto ma az mrzi, ze som si z tcpdump-u nevsimol :( ..