Date: Thu, 19 Sep 2002 12:51:45 -1000 (HST) From: Vincent Poy <vince@oahu.WURLDLINK.NET> To: John Baldwin <jhb@FreeBSD.ORG> Cc: mckay@FreeBSD.ORG, <current@FreeBSD.ORG> Subject: Re: dc(4) patch Message-ID: <20020919123131.C4630-100000@oahu.WURLDLINK.NET> In-Reply-To: <XFMail.20020919165904.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 19 Sep 2002, John Baldwin wrote:
> A while ago I started having problems with a dc(4) cardbus card that
> I hadn't had before. Lots of failures to force tx and rx to idle
> state resulting in the card eventually hanging under load and
> basically being worthless until I ejected it and reinserted it.
> ifconfig up/down, etc. didn't help. So I bought a new dc(4) cardbus
> card and it had the same problems. First card:
>
> dc0: <Abocom FE2500 10/100BaseTX> port 0x1100-0x11ff mem 0x88002000-0x880023ff irq 11 at device 0.0
> on cardbus0
>
> (really a LinkSys PCMP100 or some such)
>
> Second card:
>
> dc0: <Accton EN2242 MiniPCI 10/100BaseTX> port 0x1100-0x11ff mem 0x88002000-0x880023ff irq 11 at
> device 0.0 on cardbus0
>
> (really a SpeedStream SS1012). The patch below fixed the hangs
> on both cards:
>
> --- if_dc.c 4 Sep 2002 18:14:17 -0000 1.77
> +++ if_dc.c 19 Sep 2002 20:57:03 -0000
> @@ -1366,7 +1370,8 @@
> for (i = 0; i < DC_TIMEOUT; i++) {
> isr = CSR_READ_4(sc, DC_ISR);
> if (isr & DC_ISR_TX_IDLE &&
> - (isr & DC_ISR_RX_STATE) == DC_RXSTATE_STOPPED)
> + ((isr & DC_ISR_RX_STATE) == DC_RXSTATE_STOPPED ||
> + (isr & DC_ISR_RX_STATE) == DC_RXSTATE_WAIT))
> break;
> DELAY(10);
> }
>
> The automatic TX underrun recovery stuff didn't make a bit of difference
> for the SpeedStream card FWIW. Also, unlike one of the PR's which claimed
> that the SpeedStream SS1020 (PCI adapter rather than cardbus card), the
> dc(4) driver read the right MAC address just fine w/o the need for any
> changes.
>
>
I think the first card is a LinkSys PCMP200 and not the PCMP100
since the 100 is a 16bit PCMCIA and runs as a ed1 adapter in 4.6.2-RELEASE
but doesn't get recognized at all under -current. I've tried both the
LinkSys PCMP200 v2.0/PCMPC200 v2.0 and the NetGear FA511 Cardbus NICs but
they don't get recognize under -current either. The SpeedStream SS1012
card is identical to the SMC Networks SMC8036TX card which is $US10
cheaper and comes with a lifetime warranty versus the SpeedStream 1 year
warranty. However, I noticed that using either the SS1012 or the
SMC8036TX which I am doing now is that there seems to be a lot of errors.
root@bigbang [3:36pm][/usr/home/vince] >> netstat -s
tcp:
505575 packets sent
120454 data packets (143749591 bytes)
76 data packets (34453 bytes) retransmitted
6 data packets unnecessarily retransmitted
0 resends initiated by MTU discovery
276069 ack-only packets (1497 delayed)
0 URG only packets
0 window probe packets
106651 window update packets
2325 control packets
678552 packets received
67126 acks (for 143703608 bytes)
623 duplicate acks
0 acks for unsent data
612664 packets (686204656 bytes) received in-sequence
67 completely duplicate packets (48111 bytes)
0 old duplicate packets
19 packets with some dup. data (3318 bytes duped)
1262 out-of-order packets (1474923 bytes)
0 packets (0 bytes) of data after window
0 window probes
367 window update packets
8 packets received after close
0 discarded for bad checksums
0 discarded for bad header offset fields
0 discarded because packet too short
1823 connection requests
307 connection accepts
6 bad connection attempts
0 listen queue overflows
557 connections established (including accepts)
2305 connections closed (including 51 drops)
38 connections updated cached RTT on close
38 connections updated cached RTT variance on close
13 connections updated cached ssthresh on close
1564 embryonic connections dropped
66488 segments updated rtt (of 65063 attempts)
66 retransmit timeouts
1 connection dropped by rexmit timeout
0 persist timeouts
0 connections dropped by persist timeout
23 keepalive timeouts
23 keepalive probes sent
0 connections dropped by keepalive
45309 correct ACK header predictions
606682 correct data packet header predictions
309 syncache entries added
0 retransmitted
2 dupsyn
0 dropped
307 completed
0 bucket overflow
0 cache overflow
2 reset
0 stale
0 aborted
0 badack
0 unreach
0 zone failures
0 cookies sent
0 cookies received
udp:
44761 datagrams received
0 with incomplete header
0 with bad data length field
0 with bad checksum
9 with no checksum
903 dropped due to no socket
14443 broadcast/multicast datagrams dropped due to no socket
0 dropped due to full socket buffers
0 not for hashed pcb
29415 delivered
48858 datagrams output
ip:
873532 total packets received
0 bad header checksums
0 with size smaller than minimum
0 with data size < data length
0 with ip length > max ip packet size
0 with header length < data size
0 with data length < header length
0 with bad options
0 with incorrect version number
0 fragments received
0 fragments dropped (dup or out of space)
0 fragments dropped after timeout
0 packets reassembled ok
725596 packets for this host
770 packets for unknown/unsupported protocol
147070 packets forwarded (0 packets fast forwarded)
93 packets not forwardable
0 packets received for unknown multicast group
0 redirects sent
556438 packets sent from this host
1730 packets sent with fabricated ip header
0 output packets dropped due to no bufs, etc.
17295 output packets discarded due to no route
0 output datagrams fragmented
0 fragments created
0 datagrams that can't be fragmented
0 tunneling packets that can't find gif
0 datagrams with bad address in header
icmp:
906 calls to icmp_error
0 errors not generated 'cuz old message was icmp
Output histogram:
echo reply: 398
destination unreachable: 906
0 messages with bad code fields
0 messages < minimum length
0 bad checksums
0 messages with bad length
0 multicast echo requests ignored
0 multicast timestamp requests ignored
Input histogram:
echo reply: 174
destination unreachable: 925
echo: 398
time exceeded: 1554
398 message responses generated
0 invalid return addresses
0 no return routes
ICMP address mask responses are disabled
My HP ProCurve 2424M reports the following errors and I've
verified the cable is good as it works fine with a Intel Pro1000T Gigabit
NIC on a desktop machine...
HP ProCurve Switch 2424M HP2424M1-sfo 19-Sep-2002
15:51:03
===========================- TELNET - MANAGER MODE -============================
W 09/19/02 15:06:00 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:09:30 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:10:54 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:15:06 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:16:30 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:28:03 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:29:27 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:31:33 FFI: port 5-Excessive CRC/alignment errors. See help.
W 09/19/02 15:36:06 FFI: port 5-Excessive CRC/alignment errors. See help.
I 09/19/02 15:41:48 mgr: SME TELNET Session - MANAGER Mode established
W 09/19/02 15:50:06 FFI: port 5-Excessive CRC/alignment errors. See help.
I 09/19/02 15:50:13 mgr: SME TELNET Session - MANAGER Mode established
I 09/19/02 15:50:44 mgr: SME TELNET Session - MANAGER Mode established
---- Log events stored in memory 1-1000. Log events on screen 988-1000.
Cheers,
Vince - vince@WURLDLINK.NET - Vice President ________ __ ____
Unix Networking Operations - FreeBSD-Real Unix for Free / / / / | / |[__ ]
WurldLink Corporation / / / / | / | __] ]
San Francisco - Honolulu - Hong Kong / / / / / |/ / | __] ]
HongKong Stars/Gravis UltraSound Mailing Lists Admin /_/_/_/_/|___/|_|[____]
Almighty1@IRC - oahu.DAL.NET Hawaii's DALnet IRC Network Server Admin
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020919123131.C4630-100000>