From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0) Description of problem: Same problem as what everyone with these motherboards has had with 7.x, but the trick with adding apic to the kernel params doesnt work anymore (in 8.0). Using the SMP kernel doesnt fix it, using single cpu kernel with APIC option doesnt work either. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Use a Intel L440GX+ Motherboard 2.Install redhat 8.0 (may need to use APIC kernel option to get installer to start, I dont remember) 3. Start newly installed system. Actual Results: The aic7xxx module waits about 5-10 seconds, spits out hex garbage (not gonna put that here cause I dont want to type it all) and then says stuff like: Is not an active device Command already completed not ready or command retry failed after bus reset: host x channel x id x lun x Expected Results: System boot normally. Additional info:
I'm also trying to run this down. I have found that 8.0 runs fine if you have two processors. I tried to fake it out and remove one - NO GO. I suppose I could pick up a pair of PIII 600's on ebay for $150, but I also know that SuSE 8.0 and 8.1 run on these, with one processor, JUST FINE. From power up to completely installed - no problems. Several threads call this a 'bios bug' from Intel and write it off as "Linux 2.4 un-friendly". I don't buy it. If it were then SuSE would not work and RedHat 8.0 would not work with 2 processors. I went through this with 7.2 (the fake-out by pulling a processor worked on 7.2) and had much higher hopes for 8.0. Wasted money, I guess. C Roberts
All you have to do is recompile the kernel with the following changes: 1. compile for PIII support 2. Enable local APIC support for uniprocessors 3. Enable IO-APIC support for uniprocessors It works fine on 1 cpu for me and others. I even built rpms from RedHat's spec file for it, since I have multiple machines. Arjan has said in the past that Red Hat does not enable these options because it breaks a LOT of laptops. I guess SuSe does not care or else they know something Red Hat does not.
I have confirmed that this "fix" works for Red Hat 7.2 as well, on a uniprocessor L440GX+ board.
Shoot, forgot to mention that this was used to create an updated (2.4.18-27.7.x) kernel on an updated 7.2 system. Props to Tom Diehl for solving this one.
See also Bug 78234, which appears to be a duplicate of this bug, but has more information.
*** This bug has been marked as a duplicate of 79752 ***
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.