FC2 screaming interrupt

Ralph Jones ralph at nomeking.com
Fri Aug 20 03:13:03 UTC 2004


I seem to have an instance of the Fedora Core 2 "screaming interrupt" 
problem, although it isn't acting precisely as others have described it.

My FC2 box is on a small network with several Windows boxes, where it 
serves as a Samba data/backup server. It runs normally, except that 
whenever I logout of an X session (under any username), it disables 
interrupt 11 and takes down the network card. Logging back in (as the same 
or another user) does not make it recover; the only thing that will restore 
network function is a reboot. The messages file contains:

Aug 15 09:43:51 heloise gdm(pam_unix)[2908]: session closed for user root
Aug 15 09:43:51 heloise gconfd (root-3058): Exiting
Aug 15 09:43:53 heloise kernel: irq 11: nobody cared! (screaming interrupt?)
Aug 15 09:43:53 heloise kernel: irq 11: Please try booting with acpi=off 
and report a bug
Aug 15 09:43:53 heloise kernel: Stack pointer is garbage, not printing trace
Aug 15 09:43:53 heloise kernel: handlers:
Aug 15 09:43:53 heloise kernel: [<128be97c>] (tulip_interrupt+0x0/0x774 
[tulip])
Aug 15 09:43:53 heloise kernel: Disabling IRQ #11
Aug 15 09:43:53 heloise kernel: agpgart: Found an AGP 2.0 compliant device 
at 0000:00:00.0.
Aug 15 09:43:53 heloise kernel: agpgart: Putting AGP V2 device at 
0000:00:00.0 into 1x mode
Aug 15 09:43:53 heloise kernel: agpgart: Putting AGP V2 device at 
0000:01:00.0 into 1x mode
Aug 15 09:44:25 heloise gdm(pam_unix)[2908]: session opened for user ralph 
by (uid=0)
Aug 15 09:44:27 heloise gconfd (ralph-9657): starting (version 2.6.0), pid 
9657 user 'ralph'
Aug 15 09:44:27 heloise gconfd (ralph-9657): Resolved address 
"xml:readonly:/etc/gconf/gconf.xml.mandatory" to a read-only config source 
at position 0

Booting with acpi=off as suggested by the error message has no effect. 
Suggestions?

rj 






More information about the fedora-list mailing list