Description of Problem:At boot time with the stock 2.4.7-10 kernel the machine comes up normally until just after the init acripts returns on on enabling swap. It then hangs for about a minute and a half. It then continues booting and all is normal. When I upgrade to the 2.4.9-7 kernel the machine hangs at the same spot but never comes up. I have waited as long as 5 minutes but it stays hung. This is an athlon machine. I tried noathlon but no change in the results. Before upgrading to 7.2 I was running an ac kernel with the athlon optimizations turned on and I have had no problems with the machine. Version-Release number of selected component (if applicable): 2.4.7-10 and 2.4.9-7 How Reproducible: Every time I boot. Steps to Reproduce: 1. Boot the machine and wait. 2. 3. Actual Results: Machine will hang. How long depends on the kernel. See above Expected Results: Machine should come up normally. Additional Information: The only thing I can find interesting in the logs is the following: Oct 29 00:01:16 tigger kernel: hdc: lost interrupt Oct 29 00:01:16 tigger kernel: hdc: lost interrupt Oct 29 00:01:16 tigger kernel: Vendor: HP Model: CD-Writer+ 9100 Rev: 1.0c Oct 29 00:01:16 tigger kernel: Type: CD-ROM ANSI SCSI revision: 02 The problem Is this is not where the hang is. The logs show nothing when it hangs. See below: Oct 28 23:59:22 tigger rc.sysinit: Checking filesystems succeeded Oct 28 23:59:23 tigger rc.sysinit: Mounting local filesystems: succeeded Oct 28 23:59:23 tigger rc.sysinit: Enabling local filesystem quotas: succeeded Oct 28 23:59:23 tigger rc.sysinit: Enabling swap space: succeeded Oct 29 00:01:01 tigger init: Entering runlevel: 5 It hangs right here! Oct 29 00:01:01 tigger kudzu: Updating /etc/fstab succeeded Oct 29 00:01:12 tigger kudzu: succeeded
can you please try adding "ide=nodma" to the kernel command line ? in lilo this can be done by typing "linux ide=nodma" in grub you need to type "e", then go to the second line, edit that and add ide=nodma to the end of that line, and then press enter, and "b" for booting into that kernel.
Tried booting with ide=nodma as you requested. There is no change I still get the lost interrupt errors and the hangs on both kernels.
I did a little more debugging on this. I added echo statments to rc.sysinit to see where it is hanging. That revealed that it was hanging trying to load the ide-scsi modules. I thought that maybe the cause of the hang is/was the lost interrupts on hdc. So I moved the 2 cdroms to hdg and hdh which is an hpt 3xx controller (not exactly sure which one. It is the one that does udma100 with raid). Now the machine boots both kernels without a problem. Moving the cdrom drives back to the via i(hdc and hde) controller reproduces the problem. FWIW this machine ran fine with the ac kernels seawolf and the cd drives on hdc and hde. I did not have this problem until after the upgrade. I would like to track this down some more if you have any ideas. I cannot program but I have no problem recompiling or reconfiguring the system (short of formatting the disks this is my personal workstation). Please let me know and thanks for the help.
I do not think this is valid any longer.