I am able to successfully connect to Internet Navigator (http://www.inav.net), but after a period of time, often less than a minute, performance drops to unacceptable levels. I have connected to Internet Navigator using previous versions of Red Hat and I have not experienced these difficulties. I have incorportated the suggested fix in http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=7180 but it did not help. Here is a transcript of my /var/log/messages file, with username and password elided. Nov 26 23:33:14 localhost ifup-ppp: pppd started for ppp0 on /dev/ttyS1 at 115200 Nov 26 23:33:14 localhost modprobe: can't locate module char-major-108 Nov 26 23:33:14 localhost pppd[11834]: pppd 2.3.10 started by root, uid 0 Nov 26 23:33:15 localhost chat[11843]: abort on (BUSY) Nov 26 23:33:15 localhost chat[11843]: abort on (ERROR) Nov 26 23:33:15 localhost chat[11843]: abort on (NO CARRIER) Nov 26 23:33:15 localhost chat[11843]: abort on (NO DIALTONE) Nov 26 23:33:15 localhost chat[11843]: abort on (Invalid Login) Nov 26 23:33:15 localhost chat[11843]: abort on (Login incorrect) Nov 26 23:33:15 localhost chat[11843]: send (ATZ^M) Nov 26 23:33:15 localhost chat[11843]: expect (OK) Nov 26 23:33:16 localhost chat[11843]: ATZ^M^M Nov 26 23:33:16 localhost chat[11843]: OK Nov 26 23:33:16 localhost chat[11843]: -- got it Nov 26 23:33:16 localhost chat[11843]: send (ATDT6268888^M) Nov 26 23:33:16 localhost chat[11843]: expect (CONNECT) Nov 26 23:33:16 localhost chat[11843]: ^M Nov 26 23:33:48 localhost chat[11843]: ATDT6268888^M^M Nov 26 23:33:48 localhost chat[11843]: CONNECT Nov 26 23:33:48 localhost chat[11843]: -- got it Nov 26 23:33:48 localhost chat[11843]: send (^M) Nov 26 23:33:48 localhost chat[11843]: expect (ogin:) Nov 26 23:33:48 localhost chat[11843]: 38666 V42bis^M Nov 26 23:33:50 localhost chat[11843]: ^M Nov 26 23:33:50 localhost chat[11843]: ^M Nov 26 23:33:50 localhost chat[11843]: login: Nov 26 23:33:50 localhost chat[11843]: -- got it Nov 26 23:33:50 localhost chat[11843]: send (pomije^M) Nov 26 23:33:50 localhost chat[11843]: expect (assword:) Nov 26 23:33:50 localhost chat[11843]: ^M Nov 26 23:33:50 localhost chat[11843]: login: ******^M Nov 26 23:33:50 localhost chat[11843]: Password: Nov 26 23:33:50 localhost chat[11843]: -- got it Nov 26 23:33:50 localhost chat[11843]: send (********^M) Nov 26 23:33:51 localhost pppd[11834]: Serial connection established. Nov 26 23:33:51 localhost pppd[11834]: Using interface ppp0 Nov 26 23:33:51 localhost pppd[11834]: Connect: ppp0 <--> /dev/ttyS1 Nov 26 23:33:52 localhost pppd[11834]: local IP address 205.160.208.221 Nov 26 23:33:52 localhost pppd[11834]: remote IP address 205.160.208.248 Nov 26 23:33:52 localhost pppd[11834]: primary DNS address 199.120.107.101 Nov 26 23:33:52 localhost pppd[11834]: secondary DNS address 199.120.107.102 Nov 26 23:34:03 localhost gnome-name-server[11924]: starting Nov 26 23:34:03 localhost gnome-name-server[11924]: name server starting Nov 26 23:35:40 localhost kernel: hdc: lost interrupt Nov 26 23:35:50 localhost kernel: hdc: lost interrupt Nov 26 23:35:56 localhost gnome-name-server[11924]: input condition is: 0x10, exiting Nov 26 23:36:00 localhost kernel: hdc: lost interrupt Nov 26 23:36:30 localhost last message repeated 3 times Nov 26 23:39:41 localhost gnome-name-server[12022]: starting Nov 26 23:39:41 localhost gnome-name-server[12022]: name server starting Nov 26 23:41:05 localhost kernel: hdc: lost interrupt Nov 26 23:41:45 localhost last message repeated 4 times Nov 26 23:41:49 localhost gnome-name-server[12022]: input condition is: 0x10, exiting Nov 26 23:41:55 localhost kernel: hdc: lost interrupt Nov 26 23:42:15 localhost last message repeated 2 times I think that I start getting messages like Nov 26 23:41:45 localhost last message repeated 4 times when the slowdowns occur. Here is my ifcfp-ppp0 file DEVICE="ppp0" ONBOOT="no" USERCTL="no" MODEMPORT="/dev/ttyS1" LINESPEED="115200" PERSIST="yes" DEFABORT="yes" DEBUG="yes" INITSTRING="ATZ" DEFROUTE="yes" HARDFLOWCTL="yes" ESCAPECHARS="no" PPPOPTIONS="" PAPNAME="" REMIP="" NETMASK="" IPADDR="" MRU="" MTU="" DISCONNECTTIMEOUT="" RETRYTIMEOUT="5" BOOTPROTO="none" PPPOPTIONS="noccp"
I just opened Bug#:7393. While I agree that adding the 'noccp' option aleviates the problem of the connection "hanging", the performance is unacceptable in this mode. I believe the problem lies in the CCP compression code. Please see #7393 for additional info.
Commanding rpm -e -nodeps magicdev as root while X was shut down stopped both the bad network performance and the lost irq messages.
Some hardware seems to have strange problems with magicdev that somehow windows doesn't exhibit.
No problems here. Reports are sporadic and unreproducible. If specific information comes in, this can be fixed in the kernel (not magicdev).