From Bugzilla Helper: User-Agent: Mozilla/4.76 [en] (X11; U; Linux 2.4.3-12 i686) Description of problem: When attempting a basic install from boot.img, the installer hangs when trying to load the aic7xxx module. This appears to be related to bug #29555. Hardware info below: Motherboard: Intel L440GX+ On-board SCSI controller: Adaptec AIC-7896 VT4 output is very similar as shown in bug #29555. Lots of "aborting command due to timeout" Is this going to be fixed in 7.2? How reproducible: Always Steps to Reproduce: 1. Insert floppy with boot.img from roswell 2. Boot the machine 3. Hit enter at the boot prompt. 4. The install will hang when trying to load the aic7xxx module. Actual Results: The install hung. Expected Results: The module should have loaded fine, and I should have been able to go on the the rest of the install. Additional info:
What happens if you boot the kernel with the magic flag "apic", e.g., "linux apic"?
That fixed the problem with booting. That fixed the problem with booting, I should have read bug #29555 a bit more thoroughly. However, I still think this falls in the description of "bug" as 29555 hasn't really been fixed yet, as I understand.
Reassigning to glen - probably ought to dupe to the other related bug.
Um, since I don't own code, I think a developer would be a better person to own this defect. Arjan, please take a look at this bug and the tretise that transpires in bug #29555 and make a call as to whether (a) this defect should be closed as dup *and* (b) whether bug #29555 needs to be re-opened, and (c) if you re-open the older bug, what the disposition of the bug is (and let me know the latter in a separate e-mail, please).
We (Red Hat) really need to fix this defect before next release.
"apic" is automatically enabled on "known broken 440GX bioses". It seems your bios is not yet in this list, so I hereby ask you to download the program from http://people.redhat.com/arjanv/dmidecode.c and then (as root) do gcc dmidecode.c -o dmidecode ./dmidecode | mail -s "needs apic" hardwarebugs-list the later command will send an email with your bios identifiers to the list we have to gather this information. I would recommend you run ./dmidecode first without the mail so you can see what information will be sent out. Thank you for your cooperation.
OK, I've sent off the output. Please verify receipt. Thanks for your quick response to all this. I'm very impressed with the beta so far.
Identifier added to the list; kernels with it will appear in rawhide in a few days.
On a somewhat related note, can anybody point me to docs on how the whole beta/rawhide/final process goes? Do changes/bugfixes, etc in Roswell make it back into the Roswell stuff I see on the beta mirrors, or does everything go into rawhide until the official release? I.e., Is rawhide essentially an "improving-with-time" version of RedHat based off the latest beta?
rawhide will remain a random snapshot, but I believe the "up2date" mechanism is going to be used to upgrade Roswell with stable packages.